22.2.1 Preoperative Evaluation

The general assessment of the patient is discussed in Chapter 5. Fazio et al ¹ described a dedicated prognostic index for quantifying operative risk in colorectal carcinoma surgery from data collected from 5,034 consecutive patients undergoing major surgery. Primary end point was 30-day operative mortality. The patients’ median age was 66 years. Operative mortality was 2.3% with no significant variability between surgeons or through time. Multivariate analysis identified the following independent risk factors: age (odds ratio [OR] = 1.5 per 10-year increase), American Society of Anesthesiologists (ASA) grade (OR for ASA II, III, IV–V vs. I = 2.6, 4.3, 6.8), tumor/node/metastasis (TNM) staging (OR for stage IV vs. I–III = 2.6), mode of surgery (OR for urgent vs. nonurgent = 2.1), no-carcinoma resection versus carcinoma resection (OR = 4.5), and hematocrit level. The model has implications in everyday practice, because it may be used as an adjunct in the process of informed consent and for monitoring surgical performance through time. ²

22.2.2 Bowel Preparation

Preoperative bowel preparation has been the subject of considerable controversy. The belief that adequate mechanical preparation and antibiotic preparation are necessary has recently been questioned. As discussed in Chapter 5, most surgeons currently use some form of mechanical cleansing and systemic prophylactic antibiotics. Many are returning to the use of oral antibiotics as well.

22.2.3 Exploration of Abdomen

Before operation, a bladder catheter is routinely inserted into the bladder. A nasogastric tube is not necessary in the vast majority of cases, but an oral gastric tube is placed to decompress the stomach. Incisions should be made in a manner that provides maximum exposure for the planned resection. All operations can be performed through a midline incision and this is the access preferred by most surgeons. For a patient about to undergo a right hemicolectomy, an oblique right-sided abdominal incision is usually adequate and can be extended as needed. For transverse colon lesions, a supraumbilical transverse incision places access immediately in the area of the planned operation, and the incision can be extended in either direction if there is any difficulty encountered in taking down the flexures. For splenic flexure lesions, Rubin et al ³ advocate the use of a left subcostal transverse incision combined with the right lateral position. For left-sided colonic lesions, a subumbilical transverse incision can be used. This permits adequate exposure for even a low anterior resection. For descending colon lesions, an oblique incision may prove very convenient. The use of paramedian incisions appears quite antiquated. For emergency operations, a midline incision seems the access of choice.

When the abdomen has been opened, attention should be directed to ruling out the presence of metastatic disease, with special attention given to the liver and the pelvis. A relatively new technique advocated for the detection of occult hepatic metastases is intraoperative contact ultrasonography. ⁴ Lesions greater than 1 cm in diameter can be detected in 95% of cases, and those between 0.5 and 1.0 cm in 66% of cases. After assessment of the abdomen, attention is focused on the primary lesion to determine its resectability.

22.2.4 Principles of Resection

Dogma abounds with respect to the technical aspects of operation for colorectal carcinoma. The general principles advocated for all operations for carcinoma include removal of the primary lesion with adequate margin, including the areas of lymphatic drainage. The definition of an adequate margin especially for rectal carcinoma remains controversial. Approximately one-half of the patients seeking operative treatment already have metastatic disease spread to the regional lymph nodes. Controversy exists as to the appropriate extent of lymphatic dissection. Is a segmental resection adequate therapy? Or with a sigmoid carcinoma, for example, should a formal left hemicolectomy be performed? For the most part, the literature suggests that no survival advantage can be attributed to extended lymph node dissections for left colon and rectal carcinoma. ^{5 , 6 , 7 , 8} An exception is the isolated report by Enker et al. ⁹ Without doubt, this operation does result in increased morbidity, with patients often suffering from impotence, bladder difficulties, and potential vascular problems. Any marginal improvement is outweighed by the considerable morbidity. The principle of en bloc resection of involved structures is firmly established. Continued controversy surrounds radical lymph node dissection, luminal ligation, oophorectomy, and the “no-touch technique.” What is becoming increasingly evident is that differences in outcome among different surgeons suggest that technique is important. Whether a properly performed lymphadenectomy may produce a therapeutic benefit or whether it is simply a more accurate staging procedure is unknown.

For lesions located in the cecum or the ascending colon, a right hemicolectomy to encompass the bowel served by the ileocolic, right colic, and right branch of the midcolic vessels is recommended (▶ Fig. 22.1). For lesions involving the hepatic flexure, a more extended resection of the transverse colon is indicated (▶ Fig. 22.2). For lesions in the transverse colon, depending on the portion involved, a segment of bowel is removed as shown in ▶ Fig. 22.3. Splenic flexure lesions require removal of the distal half of the transverse colon and the descending colon (▶ Fig. 22.4). Sigmoid lesions are appropriately treated by excision of the sigmoid colon (▶ Fig. 22.5). Some surgeons prefer more radical excisions, but there is no convincing evidence to suggest that prolonged survival or decreased local recurrence will result (▶ Fig. 22.6). Indeed, operative mortality and postoperative complications are reportedly higher. ⁵ For patients who have synchronous carcinomas in different portions of the colon, a subtotal colectomy seems appropriate (▶ Fig. 22.7). Other suggested indications for subtotal colectomy include associated polyps (not removed by colonoscopy), acute or subacute obstruction, associated sigmoid diverticulosis (symptomatic), prior transverse colostomy for obstruction, young patient age (< 50 years) with a positive family history, and adherence of the sigmoid colon to a cecal carcinoma. ¹⁰

The techniques described in the following section pertain to good-risk patients. For poor-risk patients or patients undergoing palliative resection, segmental resections are more appropriate. Certain intraoperative precautions have been proposed to eliminate or at least minimize the dissemination of malignant cells. Concern has been expressed that manipulation of the carcinoma results in blood-borne metastases. There is also the risk of exfoliated malignant cells adjacent to the primary lesion becoming implanted at the suture line, in the peritoneal cavity, or in the wound. It has been postulated that handling of the primary lesion early in the operation promotes such dissemination. This thesis was supported by the demonstration of malignant cell in the circulation, ^{11 , 12} a finding that led Turnbull et al ¹³ to popularize the no-touch technique in which lymphovascular channels were ligated prior to any manipulation of the primary lesion. After using this maneuver, they reported improved survival rates for Dukes’ C lesions. However, this was not a controlled trial, and Turnbull’s results have not been duplicated. Therefore, the technique has not been adopted by most surgeons as standard therapy. In an effort to avoid implantation of malignant cells shed from the primary carcinoma, Cole et al ¹² recommended encirclement of the bowel lumen proximal and distal to the primary lesion. This is a simple addition to the operation and can usually be performed easily. Wound edges can be covered to prevent malignant cell implantation in the wound. In an effort to diminish the risk of implantation of malignant cells distal to the lesion, a host of cytotoxic agents (e.g., Dakin’s solution or bichloride of mercury) have been used to irrigate the distal bowel; distilled water has also been used for this purpose. Since none of these cytotoxic agents has been used in a clinical trial setting, their value is in question. As a means of diminishing suture line implantation, iodized catgut was popular for a brief time. ¹⁴ Another effort is the intraluminal installation of diluted formalin, in which case local recurrence was reportedly reduced to 2.6 from 14.3% for untreated patients. ¹⁵

None of the techniques tried thus far, including irrigation of the peritoneal cavity or the use of iodized catgut, has been proven of value. However, proximal and distal ligation of the bowel would appear to be a harmless practice. Adjuvant chemotherapy administered directly into the bowel lumen at the time of operation has been used but not in a trial setting.

Technique

Right Colectomy

With the appropriate retractor in place and the small bowel packed toward the left side of the abdomen, the procedure is begun by incising the parietal peritoneum from just below the terminal ileum toward the hepatic flexure (▶ Fig. 22.8). This can be done with Metzenbaum scissors or preferably by use of diathermy. If feasible, the colon is encircled above and below the carcinoma with umbilical tapes. This procedure is performed as soon as the lesion is appropriately mobile. The right colon is elevated from the retroperitoneum, with care taken not to injure the ureter, gonadal vessels, or inferior vena cava (▶ Fig. 22.9). As dissection is carried toward the hepatic flexure, attention is given to avoiding any injury to the duodenum. Peritoneal division is continued around the hepatic flexure and horizontally along the upper border of the transverse colon, with division of any adhesions to the gallbladder. As dissection continues, the lesser sac is opened by dividing the gastrocolic ligament. Mobilization is continued for as far as the resection is planned. During this stage of the operation, the second and third portions of the duodenum are exposed, and caution should be exercised to prevent injury to this structure. Next, the greater omentum is transected vertically (▶ Fig. 22.10). The medial aspect of the peritoneum then is incised along the planned area of resection. Mesenteric attachments are divided until the vascular anatomy is clear. The vessels are now displayed, and their trunks are clamped, divided, and the remaining end is ligated (▶ Fig. 22.11). The ileocolic, right colic, and right branches of the middle colic artery are divided in turn. The small vessels adjacent to the small bowel and transverse colon at the level of the proposed transection are divided between clamps, and hemostasis is secured. The two ends of bowel are now ready for division (▶ Fig. 22.12). The technique the surgeon adopts—stapling or suturing—will direct how the bowel is handled.

Surgeons who advocate the no-touch isolation technique ligate the lymphovascular structures as the initial maneuver of the operation. An incision is made in the root of the mesentery, and the trunks of the vessels are identified, divided, and ligated prior to any mobilization. The major concern with this method is the potential need to deal with the ureter, gonadal vessels, and duodenum without the benefit of adequate exposure, which may keep these structures out of harm’s way. In terms of long-term survival advantage, the efficacy of this method has not been supported in surgical trials. In a multicenter prospective randomized controlled trial, Wiggers et al ¹⁶ compared the no-touch isolation technique to a conventional technique. Both overall and corrected survival data did not differ between the two groups, although there was a tendency toward reduction in the number of occurrences and the length of time to the development of liver metastases with the no-touch technique.

For the re-establishment of intestinal continuity, the authors’ preference is to use staplers. A functional end-to-end anastomosis is created in the following way. ¹⁷ Once the site of transection has been selected, a small area (less than that necessary for hand-sutured anastomoses) is cleared. Enough fat is cleared from the edge so that when the bowel opening is closed, there is room for the linear stapler to be applied without inclusion of mesenteric fat or appendices epiploicae. Obesity per se is not a contraindication to the use of staplers. In fact, since less clearing is necessary, staplers may have an advantage in such circumstances. The linear cutting instrument (75 mm in length) is applied in the mesenteric–antimesenteric plane (▶ Fig. 22.13a). If the bowel diameter is too large to fit within the jaws of the instrument, such as with the transverse colon, the instrument is placed so that the tips of the instrument are on the antimesenteric border. Then if the bowel is not completely transected and stapled closed, the open area will occur where an opening would have been made to create the functional end-to-end anastomosis. The advantage of a stapled transection is that little or no devascularization of the bowel is necessary. With manual anastomoses, too little clearing may make anastomotic suturing insecure, and too much may jeopardize the viability of the bowel ends. Caution, however, must be exercised in patients with diverticular disease.

It is often easier to close the mesentery before the anastomosis is constructed, especially with a right hemicolectomy and with patients who are overweight. Completing this step prior to constructing the anastomosis also will diminish the possibility of the error of rotation of the ileum (▶ Fig. 22.13b).

The antimesenteric borders are aligned, and at each corner an amount of tissue just adequate to insert the fork of the anastomosing instrument is excised (▶ Fig. 22.13c). The instrument is inserted to its full length to create a large anastomosis (▶ Fig. 22.13d). The halves of the instrument are joined and the bowel is drawn up, which ensures an anastomosis that is the full length that the instrument is capable of creating (▶ Fig. 22.13e). The anastomosis should be checked to see that it is being created near the antimesenteric border and that no fat, omentum, mesentery, sponge, or viscus is trapped. The halves of the instrument should be gently separated because excessive force may result in disruption of the anastomosis. Some bleeding at the suture line is not uncommon and may be controlled by sponge compression, light cautery, or suture ligature. Excessive cautery may result in a weakened anastomosis and may predispose to leakage. Heavy bleeding should be controlled with fine sutures.

The anastomotic suture lines are held apart in preparation for the application of the linear stapler (▶ Fig. 22.13f). Welter et al ¹⁸ showed that with the functional end-to-end anastomosis the area of the anastomosis can be increased by up to one-third of the original bowel lumen if the linear stapler closing staple line is applied so as to hold the anastomosing staple lines in an open V position. There is also the very remote possibility that if the suture lines remain in apposition, unwanted healing may occur from one to the other.

If a stapling instrument is used to close the bowel defect, the staple lines should be slightly staggered when the linear stapler is applied (▶ Fig. 22.13g). This modification was suggested by Chassin et al ¹⁹ to avoid too many intersecting staple lines, which may create an ischemic point with potential for a leak. Allis clamps are applied to the tissues being approximated to prevent a portion of the bowel circumference from slipping back as the jaws of the instrument compress the tissue (▶ Fig. 22.13h). The instrument is fired, and the excess tissue is cut away prior to release of the instrument to avoid injury to the anastomosis. A fine ooze of blood is reassuring of a good blood supply to the anastomosis, but more brisk bleeding should be controlled with gentle cautery or a fine suture. Alternatively, the bowel opening can be closed with the application of the linear cutting instrument. This technical variation has been suggested to reduce the cost of the anastomosis. If adopted, care should be taken to ensure that the instrument application does not compromise the size of the anastomosis. The anastomosis is checked to ensure that it is complete and no leak is present (▶ Fig. 22.13i).

Another option (preferred by DEB) is to close the open defect in layers with two running sutures (polyglactin and polyester). This takes only a little more time, is less expensive than using another stapler, produces an inverted anastomosis, and provides the largest diameter anastomosis (as no bowel tissue is excised.

In a series of 223 anastomoses performed in 205 patients, Kyzer and Gordon reviewed their experience with the use of staples for construction of anastomoses following colonic resection. ²⁰ Indications for operation included malignancy, benign neoplasms, inflammatory bowel disease, and several miscellaneous entities. A functional end-to-end anastomosis using the standard GIA cartridge and the TA 55 instruments was performed. The operative mortality rate was 1.5%, with none of the deaths related to the anastomosis. Intraoperative complications encountered included bleeding, leak (one), tissue fracture (one), instrument failure (four), and technical error (three). Early postoperative complications related to or potentially related to the anastomosis included bleeding (five), pelvic abscess (one), fistula (one), peritonitis (two), and ischemia of the anastomosis (one). Late complications included five patients with small bowel obstruction, two of whom required operation. Anastomotic recurrences developed in 5.9% of patients. Our experience with stapling instruments has shown them to be a reliable method for performing anastomoses in the colon in a safe and expeditious manner. Complications after functional end-to-end anastomoses reported by other authors are depicted in ▶ Table 22.1.

Table 22.1 Complications after functional end-to-end anastomosis

Author(s)	No. of cases	Bleeding (%)	Fistula or leak (%)	Intraperitoneal abscess (%)	Obstruction or stenosis (%)	Operative mortality (%)
Chassin et al 21	181	0	1.1	1.7	1.1	0.7
Fortin et al 22	118	0	5.0	0	0.8	2.5
Brodman and Brodman 23	88	0	0	2.3	a	0
Reuter 24	69	0	9.0	a	a	2.9
Scher et al 25	35		2.9	a		8.6
Steichen and Ravitch 26	264	0.4	3.4	a	0.8
Tuchmann et al 27	51	2.0	6.0	a	a	0.4
Kyzer and Gordon 20	223	2.2	0.9	0.4	0	1.5
Kracht et al 28	106	a	2.8	1.9	a	1.9
a Not addressed.

A variety of commercial instruments are available for the construction of stapled anastomoses (▶ Fig. 22.14).

When a hand-sutured anastomosis is elected, the bowel edges are transected obliquely to ensure adequate blood supply at the bowel edges to be anastomosed. Today most surgeons prefer end-to-end anastomoses. Even where there is a disparity in size of the bowel lumina, as occurs in anastomoses of the ileum to the transverse colon, the discrepancy can be readily overcome by division of the antimesenteric border of the ileum. Considerable controversy has been engendered about whether to use a one-layer or a two-layer anastomosis and the type of suture material to be used. The two-layer technique, which has been used successfully in the past, consists of a posterior row of 4–0 silk or polyglactin (Vicryl, Ethicon) placed on a fine atraumatic needle into the seromuscular layer. An inner layer of 4–0 chromic catgut or polyglactin is placed through the full thickness of the bowel wall, begun on one edge, and continued on the posterior wall in a simple running over-and-over suture but with a change to the Connell suture on the anterior half. The anterior seromuscular layer is then completed with 4–0 silk, polyester, or polyglactin sutures (▶ Fig. 22.15).

A growing number of surgeons have favored a single-layer inverting interrupted technique. A posterior interrupted single layer of a 3–0 or 4–0 absorbable suture has been used. The suture is then continued on the anterior wall (▶ Fig. 22.16). Care must be taken not to invert excessive amounts of tissue, thereby causing narrowing of the lumen, but this is true for any type of hand-sutured anastomosis. Some surgeons prefer to use a Gambee suture for the single-layer anastomosis (▶ Fig. 22.17). Other suture materials such as polypropylene have been used.

Saline irrigation of the abdominal cavity is performed to remove blood, bacteria, and debris. Drains are not necessary. Wounds are closed with continuous absorbable sutures for the fascia, with staples or subcuticular continuous absorbable material used for the skin.

Resection of Transverse Colon

The appropriate operation for a carcinoma of the transverse colon has been a controversial matter. The reason is the desire to fulfill the criteria for resection of the regional lymphatic drainage. Depending on portion of the transverse colon that is involved, drainage may occur through the middle and/or right colic branches and possibly the left colic branches. For a lesion that is located in the midtransverse colon, a transverse colectomy would be in order.

The procedure might begin with division of the greater omentum from the greater curvature of the stomach, either above or below the gastroepiploic arterial arcade, with care taken not to injure the wall of the stomach (▶ Fig. 22.18). In the event of a very redundant transverse colon, the omentum may be divided vertically on either side at the proposed proximal and distal lines of resection of the colon. For a short transverse colon, the entire omentum may be included in the resected specimen. To avoid tension on the anastomosis, either one or both of the hepatic and splenic flexures will require mobilization. It is often technically easier to resect the right and transverse colon rather than attempt to mobilize both flexures. The technique of mobilization of the hepatic flexure has been described in the discussion of right hemicolectomy.

Mobilization of the splenic flexure is facilitated by incising the lateral peritoneal attachment along the descending colon (▶ Fig. 22.19). As the splenic flexure is approached, great care must be exercised to avoid injury to the spleen. The lienocolic ligament can be accentuated by passage of a finger along the colonic wall from the descending colon side toward the splenic flexure. The ligament then can be clamped and divided, or, alternatively, it can be divided with the use of cautery or other energy sources. Great caution should be exercised in this maneuver since there are frequently numerous adhesions to the splenic capsule. The peritoneum is incised on the mesocolon, and in the process the splenic flexure is mobilized downward and to the right, exposing the retroperitoneum. If the greater omentum becomes a limiting factor, division of the omentum is begun. Varying other posterior attachments may require division, with care taken not to incite bleeding in this location. The trunk of the middle colic vessel and smaller vessels are secured (▶ Fig. 22.20). It should be noted that the origin of the middle colic vessels is quite proximal on the superior mesenteric vessels and must be pursued with extreme caution to prevent injury to these structures. The bowel is divided and an anastomosis is created as previously described. In reconstituting the mesenteric defect between the ileum and the descending colon, care must be exercised to avoid narrowing of the duodenojejunal junction.

For lesions at or near the hepatic flexure or the ascending colon, a right hemicolectomy is performed. For lesions near the splenic flexure, a partial left colectomy with anastomosis of the transverse colon to the proximal sigmoid is performed (▶ Fig. 22.21). Resection of this type may necessitate division of the left branch of the middle colic and left colic vessels.

Sigmoid Resection

Some controversy exists as to the most appropriate procedure for removal of a sigmoid carcinoma. One school of thought supports the necessity for a radical left hemicolectomy with anastomosis of the transverse colon to the rectum. However, there are a growing number of surgeons who realize that the extended resection has not resulted in increased survival rates. When patients who have lymphatics involved to the root of the inferior mesenteric artery have these resected, no increased survival rate is noted in comparison with patients who have a less radical procedure. ⁷ It would, therefore, seem that the extra mobilization, with its potential risks and prolonged operating time, is not justified. The extent of the resection depends on the portion of the sigmoid colon involved. Lesions of the proximal sigmoid would require an anastomosis performed between the descending colon and the distal sigmoid, those of the distal sigmoid would involve an anastomosis between the proximal sigmoid and the upper rectum, and those of the midportion of the sigmoid, depending on the redundancy of the colon, would require an anastomosis between the sigmoid-descending junction and the rectosigmoid. The splenic flexure is not routinely mobilized, but, depending on the location of the lesion and the redundancy of the colon, it may require mobilization to avoid tension on the anastomosis.

The patient may be placed in the supine position, but for more distal lesions it is preferable to have the patient in the modified lithotomy position so that simultaneous access can be obtained through the abdomen and the rectum. This access is necessary to allow use of the circular stapling device or inspection of the anastomosis by proctosigmoidoscopy.

The procedure is initiated by incising the peritoneum along the white line of Toldt in the left paracolic gutter, freeing the distal descending colon and the sigmoid from their developmental attachments from the splenic flexure to the pelvic brim (▶ Fig. 22.22). In the midportion of the sigmoid mesocolon is the intersigmoid fossa, a small depression in the peritoneum that acts as a guide to the underlying ureter (▶ Fig. 22.23). As the sigmoid mesentery is further mobilized, care is taken to displace the mesosigmoid from the left ureter, which is seen coursing over the iliac vessels (▶ Fig. 22.24). The gonadal vessels should be protected in a similar way because injury will result in troublesome bleeding. After lateral mobilization and determination of the proximal line of resection, the peritoneum over the medial aspect of the mesosigmoid is incised toward the root of the inferior mesenteric artery to the level of the proposed ligation and then downward toward the pelvis. The inferior mesenteric artery, with its left colic and sigmoidal branches, will be identified (▶ Fig. 22.25). The inferior mesenteric artery distal to the left colic branch is then divided and ligated. Smaller vessels leading toward the planned lines of resection are secured, and the bowel is transected proximally and distally. As with the technique for right hemicolectomy, some surgeons advocate ligation and division of the blood supply prior to any other manipulation, but the same general principles pertain. In such a situation, depending on the extent of resection, the inferior mesenteric artery at its origin from the aorta (or distal to the left colic branch) and the inferior mesenteric vein at the level of the duodenum (or more distally for a lesser resection) require ligation and division. ^{21 , 22 , 23 , 24 , 25 , 26 , 27 , 28 , 29} Abcarian and Pearl ³⁰ have described a simple technique for high ligation of the inferior mesenteric artery and vein. After completion of abdominal exploration, the small bowel is packed away toward the right side of the abdominal cavity to expose the duodenojejunal flexure. The peritoneum overlying the lateral border of the fourth portion of the duodenum is incised, exposing the inferior mesenteric vein. The vein is mobilized for 2 to 3 cm, ligated in continuity with nonabsorbable suture, and divided. This incision is extended diagonally 5 to 6 cm medially to expose the infrarenal aorta proximal to its bifurcation. The inferi- or mesenteric artery is easily identified, ligated in continuity, and divided at its origin. The lymph nodes surrounding the takeoff of the inferior mesenteric artery are dissected sharply in a proximal-to-distal manner to allow for their complete excision (▶ Fig. 22.26). Heald ³¹ recommends division of the inferior mesenteric artery approximately 2 cm from the aorta to preserve the autonomic nerves, which split around its origin.

The anastomosis is then created according to the surgeon’s method of choice. However, if the anastomosis is low, the authors’ preference is to use the circular stapler as described in detail in this chapter. For surgeons who deem it necessary to perform a radical left hemicolectomy, the operation is conducted in a similar way by combining the mobilization of the sigmoid, the splenic flexure, and the distal transverse colon. The notable difference is the level at which the vessels are secured. To accomplish the radical left hemicolectomy, the posterior parietal peritoneum is incised to expose the inferior mesenteric vessels. The artery is tied flush with the aorta, and the vein is ligated separately at the level of the duodenum (▶ Fig. 22.27).

During the last several years, a variety of individuals including Hohenberger, West, and Quirke have touted the advantages of complete mesocolic excision for the attempted curative treatment of colorectal carcinoma. ^{32 , 33 , 34} While numerous studies have found the technique to confer oncologic benefit, other publications have failed to identify such benefits. Part of the problem may be a confusion between the terms “complete mesocolic excision” and “central venous ligation.” Bertelson et al, ³⁵ representing the Dutch Colorectal Cancer Group, were able to echo the results of Hohenberger, West, and Quirke by assessing over 1,000 patients who underwent standard traditional surgery as compared to 364 patients who underwent complete mesocolic excision (CME). The authors identified significant differences in the numbers of lymph nodes retrieved at 10 versus 36, respectively, and also found that 12 nodes were harvested in 89% versus 99%, respectively. Very important was the four-year disease-free survival of 75.9% versus 85.8%, respectively. Other similar differences were identified between the two groups of patients in favor of complete mesocolic excision. Although it is important to perform complete mesocolic excision, the performance of central venous ligation is more contentious and potentially introduces additional morbidity.

Bilateral Salpingo-Oophorectomy

In a review of their experience and the surgical literature, Birnkrant et al ³⁶ found the incidence of ovarian metastasis from colorectal carcinoma to be approximately 6% with a range of 1.5 to 13.6%. In a prospective controlled study, Graffner et al ³⁷ detected ovarian metastases in 10.3% of patients undergoing operations on all segments of the large bowel. Since bilateral involvement occurs between 50 and 70% of the time, a bilateral oophorectomy is recommended, especially for postmenopausal women.

However, controversy exists about the role of prophylactic oophorectomy during resection of primary colorectal carcinoma. Sielezneff et al ³⁸ attempted to prospectively assess the prognostic impact of simultaneous bilateral oophorectomy in postmenopausal women undergoing curative resection for colorectal carcinoma. Ovarian metastases were detected in 2.4% of the operative specimens. Local recurrence or liver metastases rates were not affected by oophorectomy. Five-year actuarial survival rates were not significantly different whether patients had oophorectomy (81.6%) or not (87.9%). Their results suggested that microscopic synchronous ovarian metastasis is rare at the time of curative resection of a colorectal carcinoma in postmenopausal women and does not modify prognosis. Young-Fadok et al ³⁹ conducted a prospective randomized trial of 152 patients to evaluate the influence of oophorectomy on recurrence and survival in patients with Dukes’ B and C stage colorectal carcinomas. In 76 patients randomized to oophorectomy, no incidence of gross or microscopic metastatic disease to the ovary was found. Preliminary survival curves suggest a survival benefit for oophorectomy of 2 to 3 years after operation, but this benefit does not appear to persist at 5 years (▶ Fig. 22.28). There has been no incidence of colorectal carcinoma metastatic to the ovaries in this series of Dukes’ B and C stage carcinomas, unlike other nonrandomized studies of all stages, which have reported a 4 to 10% incidence. These authors concluded that the possibility of a survival advantage emphasizes the need to continue this preliminary work.

Concomitant oophorectomy is controversial, and its efficacy in prolonging survival has been questioned since few patients survive 5 years after operation. However, it generally adds little to the operation and will prevent the subsequent development of ovarian carcinoma. ⁴⁰ Removal of the ovaries at the time of bowel resection will eliminate the need for repeat laparotomy to resect an ovarian mass in approximately 2% of women with large bowel carcinoma. ³⁶ Oophorectomy is often recommended more strongly for patients who have carcinoma of the rectum, but no site in the colon results in a greater proportion of ovarian metastases. The apparent greater proportion from the left colon is probably due to the fact that this portion of the large bowel harbors the largest number of malignancies.

Oophorectomy should be performed in premenopausal women if any gross abnormality of the ovary is detected. ³⁶ Indeed, ovarian metastases of colorectal origin have been reported to occur more commonly in premenopausal women, with rates ranging from 3.8 to 28%. ^{36 , 41} This finding might support the recommendation for prophylactic oophorectomy regardless of patient age. ⁴² Certainly, contiguous involvement necessitates en bloc resection. Ovarian involvement carries with it a poor prognosis. ⁴³

22.2.5 Postoperative Care

The postoperative care of the patient is discussed in Chapter 5. It should be noted that it is not necessary to use nasogastric suctioning on a routine basis.

22.3.1 Radiotherapy

Although radiotherapy has been used extensively in various settings for the treatment of rectal carcinoma, it plays a limited role with colon cancer. ⁴⁵ Exceptions relate to the presence of a carcinoma in a portion of the colorectum fixed to the retroperitoneum (cecum, ascending colon, and descending colon) or pelvis (rectosigmoid or rectum). Indications that have been considered appropriate for postoperative radiotherapy include the following: (1) involvement of lymph nodes; (2) known inadequate margins of resection; (3) adherence to the retroperitoneum, sacrum, or pelvic side walls; (4) transmural penetration to a macroscopic degree; and (5) extensive microscopic penetration with the presence of positive lymph nodes. ⁴⁶

A few retrospective, single institution studies have shown that adjuvant radiotherapy improves local control for colon cancer patients at high risk of recurrence after surgery. ^{47 , 48 , 49 , 50 , 51 , 52} Unfortunately, the single randomized prospective trial comparing chemotherapy alone with combined chemotherapy and radiotherapy lacks sufficient power to draw valid conclusions. ⁵³ Current National Comprehensive Cancer Network (NCCN) guidelines recommend that radiotherapy for colon cancer be considered in patients with T4 tumors with penetration to a fixed structure. ⁵⁴ The radiation field should include the tumor bed as defined by preoperative imaging and the placement of surgical clips at the time of operation. A dose of 45 to 50 Gy in 25 to 28 fractions is recommended and should be delivered with concomitant 5-fluorouracil (5-FU) chemotherapy. Thus, the colorectal surgeon should always be ready to place clips in and around the tumor bed during operations involving the resection of a fixed T4 colon tumor in order to help direct postoperative radiotherapy. Neoadjuvant chemoradiotherapy can be considered for select patients with bulky tumors invading other structures.

22.3.2 Chemotherapy

The use of adjuvant chemotherapy is attractive because it may offer the possibility of identifying patients who are likely to have occult, residual, or disseminated disease at the time of operation. Chemotherapy is most effective when the burden of carcinoma is smallest and the fraction of malignant cells in growth phase is the highest. ⁵⁵ Currently adjuvant chemotherapy is playing an increasingly important role. ⁴⁵ For patients with stage III colon cancer, adjuvant chemotherapy has been recommended since 1990. ⁵⁶ More recently, the National Quality Forum has endorsed metrics related to the administration of chemotherapy in stage III colon cancer patients in order to ensure that patients with stage III colon cancer not only are considered for chemotherapy, but also are given chemotherapy in a timely fashion. ⁵⁷ For patients with stage I colon cancer, surgery alone is highly successful, and thus no adjuvant therapy is currently recommended. On the other hand, selected patients with stage II colon cancer may benefit from adjuvant treatment and this remains the focus of clinical trials. Finally, stage IV colon cancer patients are usually primarily treated with chemotherapy.

Adjuvant chemotherapy is recommended for all stage III colon cancer patients because it decreases recurrence and increases survival when compared to surgery alone. ^{58 , 59} After surgery alone for stage III colon cancer, overall 5-year survival is 40 to 60%. ^{60 , 61 , 62 , 63 , 64 , 65} Current chemotherapeutic regimens improve overall survival to 70 to 80%. ⁶⁶ Thus, 5-year overall survival of stage III colon cancer patients improves by an absolute 20 to 25% with adjuvant chemotherapy. ▶ Table 22.2 summarizes the results of key clinical trials establishing the efficacy of adjuvant chemotherapy for nonmetastatic colon cancer. ^{66 , 67 , 68 , 69 , 70 , 71} If all patients with stage III colon cancer receive adjuvant chemotherapy, roughly one-third to one-half of disease recurrences would be prevented.

Given the significant survival benefit of adjuvant chemotherapy, colon and rectal surgeons need to ensure that their stage III colon cancer patients are evaluated for chemotherapy after surgery. The National Quality Forum has endorsed two metrics regarding the treatment of stage III colon cancer patients. ⁵⁷ The first metric estimates how many stage III patients are referred or treated with chemotherapy, whereas the second metric looks at the timeliness of the administration of chemotherapy. Specifically, the first metric (measure 0385) determines the percentage of patients ≥ 18 years old who are either referred for adjuvant chemotherapy, prescribed adjuvant chemotherapy, or have previously received adjuvant chemotherapy in the last 12 months. The other metric (measure 0223) determines the percentage of patients younger than 80 years for whom adjuvant chemotherapy is considered or administered within 4 months of the diagnosis. Thus, it is important for colon and rectal surgeons to promptly refer all stage III colon cancer patients for adjuvant chemotherapy.

For patients with stage III colon cancer, the NCCN guidelines recommend adjuvant treatment with folinic acid, fluorouracil and oxaliplatin (FOLFOX) or CapeOx for 6 months. ⁵⁴ FOLFOX has been found to be superior to 5-FU/leucovorin (LV), ^{66 , 72} and CapeOx is superior to bolus 5-FU/LV. ^{72 , 73 , 74} While used frequently in patients with metastatic disease, biologic therapy with antibodies directed at vascular endothelial growth factor A (VEGF-A; bevacizumab) and epidermal growth factor receptor (EGFR) antibody (panitumumab, cetuximab) is not recommended for adjuvant therapy of stage III disease. ^{75 , 76 , 77 , 78} The current FOLFOX regimen, mFOLFOX6, and the CapeOx regimen are outlined in ▶ Table 22.3. These agents act in different ways on colon cancer cells. 5-Fluorouracil is a pyrimidine analog that incorporates into deoxyribonucleic acid (DNA) to stop DNA synthesis. Capecitabine is an oral 5-FU prolog and thus works in the same way as 5-FU. Folinic acid (LV) is a vitamin B derivative that increases the cytotoxicity of 5-FU. Oxaliplatin inhibits DNA synthesis by forming inter- and intrastrand cross-links in DNA preventing replication and transcription. Using FOLFOX, the survival benefit of adding oxaliplatin to 5-FU does come at a price, the added side effect of peripheral sensory neuropathy (PSN). While 40 to 50% of patients given oxaliplatin will develop PSN, only 10 to 20% of patients will have grade 3 PSN, which is defined as severe symptoms limiting activities of daily living. ⁷⁹ Fortunately, only 1% of patients will have grade 3 PSN at 12 months after treatment. ⁶⁶ Since the benefit of the addition of oxaliplatin to 5-FU/LV is unproven in patients over the age of 70 years, capecitabine alone or 5-FU/LV should be considered in elderly patients with stage III colon cancer. ⁵⁴ Capecitabine-based regimens can be particularly complicated by palmar–plantar erythrodyskinesia (hand–foot syndrome), but this side effect can be limited by symptomatic treatment and resolves after treatment is concluded. ⁸⁰

The 5-year overall survival of patients with stage II colon cancer is 65 to 85% with surgery alone. ⁸¹ Unlike stage III disease, the role of adjuvant chemotherapy in stage II disease remains controversial, with some studies showing a benefit ⁷⁰ and others showing no benefit. ⁸² If there is a benefit to adjuvant chemotherapy in stage II colon cancer patients, the benefit does not improve survival by more than 5% unlike the 25 to 30% improvement for stage III patients receiving adjuvant chemotherapy. ⁵⁴

Following surgery for stage II colon cancer, the current NCCN guidelines (February 2015) recommend observation (surgery alone), enrollment in a clinical trial, or adjuvant chemotherapy. ⁵⁴ To sort out these options, a detailed discussion with the patient is recommended to highlight the potential benefits and risks of chemotherapy. Any high-risk features should be identified and discussed (▶ Table 22.4). Patients with or without high-risk features should consider observation, clinical trial, or chemotherapy with capecitabine or 5-FU/LV. Only those patients with high-risk features should be considered candidates for FOLFOX or CapeOx. It is important to remember that the addition of oxaliplatin has not been shown to improve survival in stage II colon cancer patients. ⁶⁶ Finally, decision making regarding the use of adjuvant chemotherapy for stage II disease may be aided by performing genetic testing of the tumor after surgical resection. Genetic testing of stage II tumors has been shown to be independently predictive of prognosis. High microsatellite instability (MSI-H) or defective mismatch repair (dMMR) status has been shown to be associated with a lower recurrence rate (11 vs. 26%) after surgical resection alone. ⁸³ In addition, MSI-H tumors do not benefit from 5-FU adjuvant therapy. ⁷⁵ Thus, MSI/MMR testing is recommended in all patients with stage II disease in order to avoid giving adjuvant chemotherapy in patients who will derive no benefit from it. In addition to MSI/MMR testing, multigene colon cancer assays such as Oncotype Dx, ColoPrint, and ColDx are now available that can also predict prognosis and risk of recurrence. All three of these multigene assays predict recurrence independent from other factors such as TNM stage, MMR status, tumor grade, and nodes. ^{84 , 85 , 86 , 87 , 88 , 89 , 90} While these assays provide additional information regarding prognosis and recurrence risk, they are not predictive of the potential benefit of chemotherapy, and consequently are, to date, of limited clinical value.

Bevacizumab (Avastin; Genentech Inc., South San Francisco, CA) is a recombinant humanized anti-VEGF monoclonal antibody that inhibits neoplastic angiogenesis, and has demonstrated survival benefit in patients with previously untreated metastatic colorectal carcinoma when combined with irinotecan/fluorouracil/LV (IFL). Kabbinavar et al ⁹¹ combined analysis of data from three randomized clinical studies evaluating bevacizumab in combination with FU/LV alone. The median duration of survival was 17.9 months in 5-FU/LV bevacizumab group compared with 14.6 months in the combined control group, corresponding to a hazard ratio for death of 0.74. The median duration of progression-free survival was 8.8 months in the FU/LV bevacizumab group, compared with 5.6 months in the combined control group, corresponding to a hazard ratio for disease progression of 0.63. The addition of bevacizumab also improved the response rate (34.1 vs. 24.5%).

In a phase III trial, combining bevacizumab with irinotecan, bolus fluorouracil, and LV (IFL) increased survival compared with IFL alone in first-line treatment of patients with metastatic colorectal carcinoma. Hurwitz et al ⁹² described the efficacy and safety results of the patient cohort who received bevacizumab combined with fluorouracil LV and compared them with results of concurrently enrolled patients who received IFL. Median overall survivals were 18.3 and 15.1 months with fluorouracil LV bevacizumab (n = 110) and IFL/placebo (n = 100), respectively. Median progression-free survivals were 8.8 and 6.8 months, respectively. Overall response rates were 40 and 37% and median response durations were 8.5 and 7.2 months, respectively. Adverse events consistent with those expected from the fluorouracil LV or IFL-based regimens were seen, as were modest increases in hypertension and bleeding in the bevacizumab arm, which were generally easily managed. They concluded the fluorouracil LV bevacizumab regimen seems as effective as IFL and has an acceptable safety profile. They further concluded that fluorouracil LV bevacizumab is an active alternative treatment regimen for patients with previously untreated metastatic colorectal carcinoma.

While significant progress has been made in defining optimal cytotoxic regimens in the adjuvant treatment of colorectal cancer, several questions remain regarding the optimal duration of chemotherapy treatment, the role of radiotherapy in rectal cancer, the possibility of nonsurgical interventions for rectal cancer, and the emerging role of immunotherapy.

Prior studies have shown no benefit from extending adjuvant therapy beyond 6 months in patients with stage III colon cancer. ⁹³ However, a shorter duration of chemotherapy has not been adequately investigated. CALGB 80702 is currently investigating 6 cycles (3 months) versus 12 cycles (6 months) of FOLFOX chemotherapy in patients with completely resected stage III colon cancer (NCT01150045). This will be one of six ongoing clinical trials evaluating 3 versus 6 months of adjuvant oxaliplatin-based chemotherapy. A meta-analysis of these studies (International Duration Evaluation in Adjuvant Chemotherapy [IDEA] collaboration) will test the noninferiority of a 3-month strategy to a 6-month strategy. In addition to the investigation of the duration of adjuvant treatment in colon cancer, efforts are ongoing to define the role of cyclooxygenase (COX) inhibition on disease recurrence. Analysis of the Nurses’ Health Study (NHS) and Health Professional Follow-Up Study (HPFS) has shown a decreased recurrence rate in patients with a diagnosis of colon cancer with regular aspirin intake. ⁹⁴ The benefit appeared to be limited to patients with COX-2 overexpressing tumors. ⁹⁵

These analyses were limited by their retrospective nature and require further support from prospectively conducted trials. CALGB 80702 randomizes all enrolled subjects to celecoxib versus placebo in order to investigate the role of COX-2 inhibition in the adjuvant treatment of colon cancer. Similarly, the A SCOLT clinical trial (NCT00565708) is randomizing patients with stage II or III disease to 3 years of aspirin versus placebo to address the role of aspirin in preventing colorectal cancer recurrence. Finally, several studies are investigating immunotherapy as an adjuvant form of treatment in colon cancer. An ongoing phase III clinical trial is evaluating the role of cytokine-induced killer cell immunotherapy for stage III colon cancer following surgery and completion of adjuvant therapy (NCT02280278).

22.3.3 Immunotherapy

Immunotherapy was believed to have some effect on colon carcinoma, but there is no conclusive evidence to indicate significant improvement in survival. ⁹⁶ A review of prospective randomized trials by Lise et al, ⁹⁷ which included an immunotherapy arm, failed to demonstrate any benefit. A report of a controlled randomized trial consisting of a 2-year program of vaccination with bacille Calmette-Guérin (BCG) and neuraminidase-treated autologous carcinoma cells at 5-year follow-up failed to alter either the disease-free interval or the survival of patients. ⁹⁸ A controlled clinical trial of interferon-a as postoperative surgical adjuvant therapy for patients with colon carcinoma demonstrated significant enhancement of nonspecific immune function but no significant difference in patient survival. ⁹⁹ A study in which 189 patients with Dukes’ C colorectal carcinoma who underwent resection for cure were randomized to observation or postoperative treatment with 17–1A antibody. After a median follow-up of 5 years, antibody treatment was reported to have reduced the overall death rate by 30%. ¹⁰⁰ In the future, genetic engineering techniques may allow generation of substances during the immune response, and these may have therapeutic value by modifying the biologic response to malignancy. ¹⁰¹

22.4.1 Obstruction

When complete obstruction of the colon arises as a result of a carcinoma, the recommended treatment depends on the level of the colon that is obstructed as well as the beliefs and experience of the treating surgeon. ^{104 , 105} In their review of 115 obstructing carcinomas, Sjödahl et al ¹⁰⁶ found that 37% were right sided (proximal to splenic flexure) and 63% were left sided. Only 4% were Dukes’ A, while 15% already had distant metastases.

Interestingly, a study by Nozoe et al ¹⁰⁷ found the mean size of the obstructing carcinoma was 3.7 cm, which was significantly smaller than that of nonobstructing carcinomas (5.4 cm). The proportion of lymph node metastases in obstructing carcinomas was 66.9%, which was significantly higher than that in nonobstructing carcinomas (42.4%). The proportion of carcinomas classified into Dukes’ C or D in obstructing carcinomas was 84.6% and was significantly higher than that in nonobstructing carcinomas (52.5%).

If the patient’s condition can be stabilized and there is evidence of resolution of the occlusion, bowel preparation and elective resection is the ideal solution. This clinical course is unusual, and therefore decisions on how to proceed must be made. For right-sided colonic obstructions, it is generally accepted that the treatment of choice is a resection and primary anastomosis with removal of the right and proximal transverse colon. ¹⁰⁸ Even though the bowel is not prepared, the resection usually can be readily accomplished.

When the obstruction is located in the distal transverse colon, the matter of how to proceed is controversial. Some surgeons believe that the patient should have a proximal diversion, followed by a definitive resection. However, under these circumstances a growing number of surgeons have adopted the procedure of an extended right hemicolectomy, followed by a primary ileo-descending colon anastomosis.

Lee et al ¹⁰⁹ compared the operative results of 243 patients who had emergency operations for right- and left-sided obstructions from primary colorectal carcinomas. One hundred and seven patients had obstruction at or proximal to the splenic flexure (right-sided lesions) and 136 had lesions distal to the splenic flexure (left-sided lesions). The primary resection rate was 91.8%. Of the 223 patients with primary resection, primary anastomosis was possible in 88% of patients. Among the 101 primary anastomosis patients with left-sided obstruction, segmental resection with on-table lavage was performed in 75 patients and subtotal colectomy was performed in 26 patients. The overall operative mortality rate was 9.4%, although that of the patients with primary resection and anastomosis was 8.1%. The anastomotic leakage rate for those with primary resection and anastomosis was 6.1%. There were no differences in the mortality or leakage rates between patients with right- and left-sided lesions (mortality 7.3 vs. 8.9% and leakage 5.3 vs. 6.9%). Colocolonic anastomosis did not show a significant difference in leakage rate when compared with ileocolonic anastomosis (6.1 vs. 6%).

On-Table Lavage

Still others have recommended resection of the primary disease combined with on-table lavage and primary anastomosis. A major perceived disadvantage of on-table lavage is that it is time consuming. The operative technique consists of the mobilization of the appropriate segment of colon according to oncologic principles. In most circumstances, both the hepatic and splenic flexures require mobilization. The bowel at an appropriate distance distal to the carcinoma is divided, as is the proximal bowel 5 to 10 cm distal to the proximal site of the anastomosis, thus removing the carcinoma-bearing portion of colon. A no. 22 or 24 bladder catheter is inserted into the cecum through the freshly amputated appendicular stump or through the terminal ileum if the patient has had an appendectomy (▶ Fig. 22.29). The catheter balloon is inflated and held in place by a pursestring suture. A noncrushing clamp is placed across the terminal ileum to prevent reflux of the irrigation fluid. A standard intravenous infusion set is connected to the bladder catheter. The bowel, having been adequately mobilized, permits the distal portion to be placed in a kidney basin and hard fecal scybala can be “milked” into the kidney basin. A sterile corrugated anesthetic scavenger tube, 22 mm in diameter, is then inserted into the distal bowel and secured in place with strong tapes. The free end of this tube is draped over the side of the patient and secured in an appropriate collecting system. The colon is then lavaged with warm isotonic saline solution until the effluent in the corrugated tube is clear. The volume of lavage solution required is determined by the extent of fecal loading but is usually 3 to 6 L. Lavage time may vary from 20 to 45 minutes. Once the effluent is clear, the bladder catheter is removed, and the appendiceal opening or ileum is closed. A short segment of bowel containing the irrigated tube is resected and an anastomosis created according to the surgeon’s preference.

Most series quote an operative mortality rate of approximately 10%. ¹¹¹ Anastomotic leakage rates following primary anastomosis are low. Tan and Nambiar ¹¹⁷ performed 36 primary resections and anastomoses following intraoperative antegrade colonic lavage for left-sided obstructing carcinoma. There were two deaths (one from anastomotic leak). Other complications included chest infection in 11% and wound infection in 19.4%. Others support this form of management. ^{114 , 118 , 119} Leakage rates are approximately 4%. ¹¹¹ Wound infection rates remain a problem, with rates of 25 to 60%, and hospital stays of around 20 days. ¹¹¹

The Scotia Study Group ¹²⁰ conducted the first multicenter prospective randomized trial comparing subtotal colectomy with segmental resection and primary anastomosis following intraoperative irrigation for the management of malignant left-sided colonic obstruction. Of the 91 eligible patients recruited by 12 centers, 47 were randomized to subtotal colectomy and 44 to on-table irrigation and segmental colectomy. Hospital mortality and complication rates did not differ significantly, but 4 months after operation, increased bowel frequency (three or more bowel movements per day) was significantly more common in the subtotal colectomy group (14 of 35 vs. 4 of 35). More patients in the subtotal colectomy group reported that they had consulted their general practitioner with bowel problems than those in the segmental resection group (15 of 37 vs. 3 of 35). The authors believe that segmental resection following intraoperative irrigation is the preferred option except when there is cecal perforation or if synchronous neoplasms are present in the colon, when subtotal colectomy is more appropriate.

Chiappa et al ¹²¹ reported 39 patients who were treated with intraoperative decompression, on-table lavage, resection, and primary anastomosis. The primary anastomosis was intraperitoneal in 74% and below the peritoneal reflection of the rectum in 26% of patients. Operative mortality was 3% and anastomotic leakage was observed in 6% of patients. Complications included intra-abdominal abscess (3%) and wound infections (8%). Ohman ¹²² also found a higher operative mortality rate for primary resection (14%) compared with staged resection (5%) and, although there was an early apparent superior survival rate with the staged procedures, it did not persist into the fourth and fifth years. Umpleby and Williamson ¹²³ reported a better 5-year survival rate following resection and anastomosis (48%) than after staged procedures (18%).

22.4.2 Stenting

Intestinal stenting is a procedure that is becoming more widespread. It was first introduced by Dohmoto in 1991 ¹²⁹ as definitive palliative treatment for patients with obstructive disease where resection for cure was not appropriate due to very advanced local disease, metastatic disease, or because of an unacceptably high operative risk. In 1994, Tejero et al ¹³⁰ proposed stent placement as a “bridge-to-surgery” for emergency relief of colonic obstruction with an aim to subsequent elective resection. The technique can be applied in patients who refused operative treatment. Colostomy can be avoided with an improved quality of life especially in the palliative setting.

Suitable lesions for endoluminal colorectal stenting include obstructing both primary left-sided colorectal carcinomas and extracolonic malignancies such as prostate, bladder, ovarian, or pancreatic. It is also not appropriate for lesions less than 5 cm from the anal verge. The actual length of the lesion is not a theoretical limitation. It is contraindicated in the presence of colonic perforation with peritonitis and would not prove effective with multiple sites of obstruction.

Although it is not mandatory, it is probably best that stents be placed under endoscopic guidance with the aid of fluoroscopy. The administration of prophylactic antibiotics is probably wise. The procedure is conducted under conscious sedation. A catheter over a guide wire is advanced through the lesion. Contrast is injected into the proximal lumen. Once deployed, the stents expand and become incorporated into the surrounding tissue by pressure necrosis, thus anchoring the stent.

Dauphine et al ¹³¹ reviewed their experience with 26 self-expanding metal stents as the initial interventional approach in the management of acute malignant large bowel obstruction. In 14 patients, the stents were placed for palliation, whereas in 12, they were placed as a bridge to surgery. In 85%, stent placement was successful on the first occasion. In the remaining four individuals, one was successfully stented at the second occasion, and three required emergency operation. Nine of the 12 patients (75%) in the bridge-to-surgery group underwent elective colon resection. In the palliative group, 29% had reobstruction of the stents and in 9% the stent migrated. In the remaining 62%, the stent was patent until the patient died or until the time of last follow-up. Colonic stents achieved immediate nonoperative decompression and proved to be both safe and effective.

Since first described, there have been numerous publications on the subject. Khot et al ¹³² conducted a systematic review of the published data on stenting for the treatment of colorectal obstruction. A total of 58 publications were found, of which 29 case series were included in the analysis. Technical and clinical success, complications, and reobstruction, both in palliation and as a “bridge to surgery” were assessed. Pooled results showed that stent insertion was attempted in 598 instances. Technical success was achieved in 92% and clinical success in 88%. Palliation was achieved in 90% of 336 cases, while 85% of 262 insertions succeeded as a “bridge to surgery” (95% had a one-stage operative procedure with a mean time to the operating room of 8.9 days). Technical reasons for failure included inability to place the guide wire, malposition, or perforation. Clinical failures included perforation, persistent obstructive symptoms, or adhesion of colonic wall to the stent. There were three deaths (1%). Perforation occurred 22 times (4%), 1% in balloon dilatation versus 2% in non–balloon dilatation. Stent migration was reported in 10% of 551 technically successful cases. Management included stent removal, stent reinsertion, operation, and no immediate intervention but proceed to planned operation. The rate of stent reobstruction was 10% of the 525, mainly in the palliative group. Reason for obstruction included ingrowth of malignancy, stent migration, and fecal impaction. Bleeding occurred in 5% of patients, the majority requiring no treatment, but three patients received transfusions. Another 5% of patients experienced pain, either abdominal or rectal, and this was controlled with oral analgesics. They concluded that the evidence suggests that colorectal stents offer good palliation and are safe and effective as a “bridge to surgery.” Stent usage can avoid the need for a stoma and is associated with low rates of mortality and morbidity. Dilatation of malignant strictures at the time of stent placement appears to be dangerous and should be avoided.

Law et al ¹³³ evaluated the outcomes of self-expanding metallic stents as a palliative treatment for malignant obstruction of the colon and rectum. The insertion of self-expanding metallic stents was attempted for palliation in 52 patients. Successful insertion of the stent was achieved in 50 patients. The median survival of patients was 88 (range, 3–450) days. Complications occurred in 13 patients (25%). These included perforation of the colon (n = 1), migration or dislodgement of the stents (n = 8), severe tenesmus (n = 1), colovesical fistula (n = 1), and ingrowth of malignancy (n = 2). Insertion of a second stent was required in eight patients. Subsequent operations were performed in nine patients, and stoma creation was required in seven patients.

Saida et al ¹³⁴ evaluated the long-term prognosis of expandable metallic stent insertion compared with emergency operation without expandable metallic stent. Forty emergency operations and 44 expandable metallic stent insertions were retrospectively compared. Postoperative complications were significantly less frequent in the expandable metallic stent group: wound infection was 14 versus 2%; leakage following anastomosis was 11 versus 3%; 3-year overall survival rate was 50 versus 48%; 5-year survival rate was 44 versus 40% in the emergency operation and expandable metallic stent groups, respectively. They concluded that because preoperative expandable metallic stent insertion for obstructive colorectal carcinoma had good postoperative results and no disadvantages in long-term prognosis, this procedure should be used in preoperative treatments of obstructive colorectal carcinoma.

Martinez-Santos et al ¹³⁵ evaluated primary anastomosis and morbidity rates obtained with self-expandable stents in comparison with the results of emergency surgical treatment. Patients with left-sided malignant colorectal obstruction were enrolled. Forty-three patients were assigned to preoperative stent and elective operation or palliative stent (emergency surgical treatment). In the stent group, the obstruction was relieved in 95% after the stent placement. Of 26 patients who underwent operative treatment, a primary anastomosis was possible in 84.6 versus 41.4% in the immediate operative group, with lower need for a colostomy (15.4 vs. 58.6%) in the immediate operative group. The anastomotic failure rate was similar and the reintervention rate was lower (0 vs. 17%). The total stay (14.2 vs. 18.5 days), the intensive care unit stay (0.3 vs. 2.9 days), and the number of patients with severe complications (11.6 vs. 41.2%) were significantly lower in the stent group.

Johnson et al ¹³⁶ studied 36 patients, of whom 18 had obstructing left-sided colon carcinomas relieved by placement of endoluminal stents. These were compared with 18 historical controls with similar clinicopathological features that were treated more conventionally with palliative stoma formation. Both groups of patients gained relief of obstructive symptoms. There were no differences in survival or in-hospital mortality. The median length of palliation was 92 days for stenting and 121 days for palliative stoma formation. Formation of a stoma required a significantly longer stay in the intensive care unit, but hospital stay was similar. They concluded as an alternative to palliative operation, selected patients benefit from colonic endoluminal stenting with relief of obstructive symptoms and no adverse effect on survival. Patients may be spared the potential problems associated with palliative stoma formation and the morbidity of operation. Stenting can be offered to the very frail patient who would otherwise be managed conservatively.

Meisner et al ¹³⁷ reported on 104 procedures with self-expanding metal stents performed in 96 patients. The goals of the procedure were either postponement of emergency operation or definitive palliative treatment. Technical success was achieved in 92% and clinical success in 82%. Procedure-related complications included perforation in three patients during stenting and in one instance 6 to 7 hours after. Other technical problems could mainly be overcome by introducing an additional stent. They believe complications seen in the group treated with self-expanding metal stents and subsequent resection (mortality 18% and anastomotic leakage 18%) do not differ from the number of complications usually seen in patients who undergo colorectal resection.

Suzuki et al ¹³⁸ reviewed 36 patients with malignant obstruction, and 6 patients with benign obstructive disease who underwent placement of self-expandable stents using a combined endoscopic and fluoroscopic technique. Stent placement was successful in 86%. Complication occurred in 44%: migration (n = 7), reobstruction (n = 5), perforation (n = 2), fistula formation (n = 1), and stent fracture (n = 1). Stent placement was successful in 100% of patients with benign strictures, but post–stent migration was frequent (2/6).

Tomiki et al ¹³⁹ compared the clinical outcome of 18 patients who had stent placement and 17 patients who underwent only colostomy. The postoperative hospital stay was 22.3 days for stent placement compared with 47.4 days for colostomy. The duration to readmission was 129.2 days for stent placement and 188.4 days for colostomy. The estimated duration of primary stent patency was 106 days. Mean survival period was 134 days in patients with stent placement and 191 days in patients with colostomy. They concluded that stent placement increases the option of palliative treatment and is an effective treatment contributing to improving quality of life.

Sebastian et al ¹⁴⁰ systematically reviewed the efficacy and safety of self-expanding metal stents in the setting of malignant colorectal obstruction. Fifty-four studies reported the use of stents in a total of 1,198 patients. The median technical and clinical success rates were 94 and 91%, respectively. The clinical success when used as a bridge to surgery was 71.7%. Major complications related to stent placement included perforation (3.8%), stent migration (11.8%), and reobstruction (7.3%). Stent-related mortality was 0.58%.

Carne et al ¹⁴¹ compared the use of expandable metallic stents as a palliative measure to traditional open surgical management. Patients with left-sided (splenic flexure and distal) colorectal carcinoma and nonresectable metastatic disease (stage IV) were treated with expandable metal stents or open resection or stoma. Twenty-two of 25 patients had colonic stents successfully inserted and 19 patients underwent open operation. The malignancies were primary in 22 stent procedures and 18 open operations. The open operations were laparotomy only (n = 2), bypass (n = 1), stoma (n = 7), resection with anastomosis (n = 4), and resection without anastomosis (n = 5). The complications after open operation were urinary (n = 2), stroke (n = 1), cardiac (n = 2), respiratory (n = 2), deep venous thrombosis (n = 1), and anastomotic leak (n = 1). There were no stent-related complications. The mean length of stay was significantly shorter in the stent group (4 vs. 10.4 days). There was no difference in survival between the two groups (median survival: stent group, 7.5 months; open operation, 3.9 months). They concluded that patients treated with stents are discharged earlier than after open operation. Stents do not affect survival.

Although stents are expensive, the procedure appears to be cost-effective since emergency operation can be avoided with acute bowel obstruction, and in those with advanced disease no resection of the colon is necessary.

22.4.3 Perforation

Perforation has been reported to occur in 3 to 9% of patients with colorectal carcinomas. ¹⁴² Patients who develop a free perforation of the colon associated with a carcinoma present with signs and symptoms of generalized peritonitis. The carcinoma itself may be perforated, or there may be a left-sided carcinoma associated with a right-sided perforation. Each situation is handled differently. In the clinical setting, for treatment of a perforated carcinoma, older reports recommended that the perforation be managed by diversion, with a proximal colostomy or cecostomy performed in association with repair of the perforation. However, this treatment does not relieve the septic process, and the aim of therapy should be to remove the diseased segment. Otherwise, contamination will continue from the level of the stoma to the level of the perforation. On completion of the resection, the question arises as to how to handle the bowel ends. If the patient already has generalized peritonitis, it seems inappropriate to perform a primary anastomosis. In this event, the proximal bowel is brought out as a stoma, and the distal bowel is drawn out as a mucous fistula or closed as a Hartmann pouch. For a right-sided perforation, a similar procedure can be performed. Another option is to resect the perforated diseased bowel and perform a primary anastomosis with a proximal diversionary stoma, either a proximal colostomy or a loop ileostomy. If technically feasible, the two ends of bowel should be brought out adjacent to each other as described for the end loop stoma. The advantage of this technique is that bowel continuity can be established at a later date without the need for a formal laparotomy.

When there is an obstructing lesion of the left colon and a perforation of the right colon, a viable option is a subtotal colectomy encompassing removal of the perforated colon and the malignancy in one operation. Saegesser and Sandblom ¹⁴³ stressed the fact that simple suture repair of an ischemic colon will not hold and that a temporary colostomy placed in an ischemic or inflamed bowel will pull through. The authors believe that the practice of closure of the perforation and relief of obstruction by colostomy or by exteriorization of the perforated cecum is illogical and inadequate. The surgeon should proceed with resection of the carcinoma and the entire distended part of the ischemic and perforated colon. A subtotal colectomy might even be considered if only a left-sided perforation is present, since this operation would fulfill the criteria of removing the diseased and unprepared bowel. Another option for management of the patient with a perforation remote from the diseased segment is to bring out the perforated segment as a stoma, either by colostomy or cecostomy.

For the patient who presents with localized peritonitis on the right side, the diagnosis may be confused with that of appendicitis. If the diagnosis is definite at the time of laparotomy, it is reasonable to proceed with a right hemicolectomy and primary anastomosis. If the localized peritonitis occurs on the left side, the differential diagnosis will include diverticulitis. Resection of the diseased segment is indicated, and management of the ends involves the same considerations as with the obstructed unprepared bowel.

22.4.4 Bleeding

Massive bleeding from a carcinoma is an unusual complication, but when it arises, it offers the built-in advantage of being a colonic cathartic. Therefore, if bleeding is so profuse that urgent operation is required, a mechanical cleansing is automatically present, and the affected portion of bowel can be resected with a primary anastomosis.

22.4.5 Obstructive Colitis

Obstructive colitis is an ulceroinflammatory condition that occurs in a dilated segment of the colon proximal to an obstructing or partially obstructing lesion. The entity is rarely reported in the literature and the following information was drawn from the review by Tsai et al. ¹⁴⁴ Obstructive colitis is only encountered in 0.3 to 3.1% of all colorectal carcinomas and affects both men and women over 50 years of age. Minor degrees of obstructive colitis may be overlooked and its prevalence may be as high as 7% when specifically sought. The left side of the colon, especially the sigmoid colon, is usually involved in obstructive colitis. Patients with obstructive colitis usually complain of bleeding per rectum and abdominal pain as well as nausea and vomiting, all of which are indistinguishable from the symptoms of colorectal carcinoma. Regardless of severity and distribution pattern, a diagnostic feature of obstructive colitis is the presence of an intact mucosal segment of about 2 to 6 cm long between the carcinoma and the colitis. The area of colitis is usually a single confluent area, often with regular geographic margins, which is well demarcated from the surrounding normal mucosa.

Microscopically, focal areas of colitis associated with obstructive colitis show replacement of mucosa by active granulation tissue. Acute and chronic inflammatory cells are moderate in amount and seldom extend beyond sites of granulation. Pseudopolyps of granular tissue or edematous mucosa may occur, and crypt abscesses may involve the mucosa at the ulcer margin. The mucosa in the intervening segment and distal to the obstructing lesion is usually normal. It is differentiated from ulcerative colitis, which is characterized histologically by an intense inflammation of the mucosa and submucosa in addition to the presence of multiple crypt abscesses. The rectum is always involved and the disease extends proximally for varying distances but always with continuity of involvement to the proximal extent of the disease process.

One suggested pathogenic mechanism of obstructive colitis is that of secondary ischemia caused by hypoperfusion. Additional factors, such as preexisting atheroma, anemia, or a past history of pelvic irradiation, may play a role in precipitating the colitis.

Obstructive colitis can cause both diagnostic and therapeutic problems. The signs and symptoms arising from obstructive colitis may be attributed to the primary obstructive lesion, which is usually most obvious on radiological and endoscopic studies. Areas of colitis may be a source of septicemia or may perforate and lead to peritonitis. Anastomoses in the unrecognized area of colitis may break down. Up to 25% of cases of obstructive colitis have been associated with anastomotic complications. Their frequently normal appearance at operation may lead to involved segments of colon being used for anastomoses with consequent complications. Because perforation through a colonic carcinoma is a grim prognostic event with negligible 5-year survival, it is important to distinguish this from perforation through the obstructive colitis, which may have a different prognosis. Awareness of the features and incidence of obstructive colitis should help surgeons avoid these diagnostic and therapeutic problems.

22.4.6 Invasion of Adjacent Viscera

Occasionally a carcinoma becomes attached to the abdominal wall or the adjacent viscera, such as the small bowel, urinary bladder, uterus, stomach, spleen, ureter, or duodenum. It is estimated that such attachment occurs in approximately 10% of all patients with colon carcinomas with a reported range of 3.1 to 16.7%. ¹⁴⁵ The philosophy of treatment to be followed in these circumstances might best be expressed by the quote attributed to Hippocrates: “To extreme diseases, extreme remedies.” In order to perform an adequate curative operation, it sometimes becomes necessary to excise en bloc all or part of the attached viscus (▶ Fig. 22.30). Often, these adhesions are inflammatory in nature and not caused by malignant infiltration, so the prognosis frequently is better than might have been anticipated originally (▶ Table 22.5). With this knowledge, the surgeon should not hesitate to resect attached structures. ^{146 , 147 , 148 , 149 , 150 , 151 , 152 , 153 , 154 , 155}

An exception to these recommendations might be when the duodenum or bladder base is extensively involved, in which case the primary lesion is removed and the structures at risk are marked with metal clips. Under these circumstances, the morbidity and mortality rates of the radical operation involving an anterior exenteration or Whipple’s procedure would probably exceed any possible benefit derived from a very radical operation. However, Curley et al ¹⁵⁶ reported on 12 patients in whom the carcinoma involved the duodenum or pancreatic head and who underwent an en bloc extended right hemicolectomy and pancreaticoduodenectomy. There were no operative deaths, and malignant invasion was confirmed in all patients. At the time of reporting, 8 of the 12 patients were alive at a median of 42 months.

Similarly, Koea et al ¹⁵⁷ reported their experience with eight patients with bulky primary carcinomas of the right colon infiltrating the duodenum (n = 4) or pancreatic head (n = 4) surgically managed at Memorial Sloan Kettering Cancer Center. Six patients presented with anemia, and one patient each with epigastric pain and an abdominal mass. All patients had T4 lesions, whereas five had lymph node metastases at presentation. All patients were resected with clear pathologic margins by either right colectomy and en bloc duodenectomy (n = 4) or en bloc pancreaticoduodenectomy (n = 4). The 30-day mortality rate was 0. Six patients remained alive and free of disease at a median follow-up of 26 months, and there was one longterm survivor who was alive and free of disease at 84 months after resection.

Talamonti et al ¹⁵⁸ reviewed 70 patients who underwent resection of a carcinoma of the colon and rectum with en bloc total cystectomy (36 patients) or partial cystectomy (34 patients). There were three postoperative deaths in the total cystectomy group but none after partial cystectomy. The 5-year actuarial survival rate for the 64 patients with negative resection margins was 52%. In exceptionally good-risk patients, consideration may be given to a total pelvic exenteration.

In a review of 1,918 patients who underwent curative resection for colorectal carcinoma, Gall et al ¹⁴⁷ noted that 121 patients had multivisceral organ involvement. Extended multivisceral radical resections resulted in a postoperative mortality rate of 12% (compared to 6% without such resection), with a 5-year survival rate of 54% for patients with inflammatory adherence and 49% for patients with malignant infiltration. In this series, the most frequently used extensions of resection were total hysterectomy (39%), small bowel (21%), urinary bladder (16%), and abdominal wall (4%). It is worth noting that when carcinoma was inadvertently torn or transected during resection, the 5-year survival rate dramatically dropped to 17%. Despite the increased operative mortality with extended resection, the authors of this review believe that the benefit outweighs the disadvantages. Hunter et al ¹⁴⁸ reviewed their results of colorectal carcinoma in three treatment groups: standard colectomy, en bloc resection, and colectomy with separation of adherent organs. The 5-year survival rates were 55, 61, and 23%, respectively. No operative mortality occurred with en bloc resection. The 5-year survival rate, recurrence rate, and local recurrence rate for standard colectomy were 55, 33, and 11%, respectively; for en bloc resection, 61, 36, and 18%; and for separation of organs, 23, 77, and 69%. The authors concluded that colorectal carcinoma adherent to adjacent organs must be treated by en bloc resection because separation of organs results in unacceptably high local recurrence and poor 5-year survival rates. On the other hand, the results of en bloc resection were comparable to those of standard colectomy for non-adherent carcinomas.

Nakafusa et al ¹⁵⁵ evaluated the short- or long-term outcome of multivisceral resection relative to that of the standard operation. Of 323 patients, 16.4% received multivisceral resection because of adhesion to other organs. Overall, morbidity rates were 49.1% for multivisceral resection versus 17.8% for the standard operation and postoperative mortality was 0% in both groups. Only multivisceral resection (OR, 2.7) was an independent factor for overall postoperative complications. The survival of patients with multivisceral resection was similar to that after the standard operation (5-year rate, 76.6 vs. 79.5%). Lymph node metastases (hazard ratio, 2.5) and blood transfusion (hazard ratio, 2.4) were independently associated with patient survival.

Kroneman et al ¹⁵⁹ evaluated the results of 33 patients who underwent curative en bloc resection. Adherent organs excised included small bowel, urinary bladder, abdominal wall, uterus, duodenum, pancreas, stomach, and kidney. The postoperative morbidity rate was 6%, the mortality was 3%, and the 4-year survival rate was 33%. Poeze et al ¹⁶⁰ reported on 1,346 patients with colorectal carcinoma, 144 (11%) of whom underwent multivisceral resections for invasion of adjacent organs. In patients who had disease-free margins, there was no compromise of long-term survival (i.e., local invasion to adjacent organs with or without lymph node involvement was not related to survival). The overall operative mortality rate was 5%. Izbicki et al ¹⁵³ reported on 83 patients who underwent en bloc resection. Mean survival was 44 months after extended resection. The postoperative mortality, morbidity, and survival rates were comparable to those in patients who underwent nonextended resections.

Landercasper et al ¹⁶¹ reported on 54 of 1,284 patients (4%) who underwent potentially curative resections of right colon lesions found to be adherent to adjacent organs, abdominal wall, or retroperitoneum. Postoperative complications developed in 24% of patients. The mortality rate was 1.9% and the 5-year survival rate was 31%. Only one of nine patients with pancreatic or duodenal adherence treated with limited resection remained disease free. The authors recommend radical en bloc resection if no distant metastases are present. Adjuvant radiation therapy or chemotherapy did not improve survival.

To determine the perioperative mortality and morbidity and the long-term prognosis of patients undergoing extended pelvic resections for localized advanced primary adenocarcinoma of the rectum, Orkin et al ¹⁴⁹ reviewed their experience with 65 patients. Local invasion without distant metastases was present in all patients at operation and en bloc resection of all involved organs was performed with intent of cure. Average age at operation was 61 years; 23% were men and 77% were women. Operations included abdominoperineal resection in 57%, low anterior resection in 31%, and Hartmann’s procedure in 12%. Additionally, women (81%) with intact uteri underwent en bloc hysterectomy, 77% of women with intact ovaries had oophorectomy, and 50% of women had partial vaginal resection. Twenty-six percent of the 65 patients had a cystectomy, and 2 patients had a portion of small intestine resected in continuity with their carcinoma. Pathologic examination revealed lymph node involvement in 45% and histologic confirmation of adjacent organ extension in 57%. There were no perioperative deaths. Overall 5-year survival was 52% with 65% of deaths attributable to either recurrent carcinoma or a new primary lesion. The cumulative probability of recurrence at 5 years was 39%.

Carne et al ¹⁵⁴ reported on multicenter experiences of en bloc bladder resection for colorectal carcinoma adhering to the urinary bladder. Fifty-three patients were identified, of which 45 had en bloc partial cystectomy performed, 4 en bloc total cystectomy, and 4 had the adhesions disrupted and no bladder resection. All patients who did not have a bloc resection developed local recurrence and died from their disease. Mean follow-up was 62 months. The extent of bladder resection did not seem important in determining local recurrence. The decision to perform total rather than partial cystectomy should be based on the anatomic location of the carcinoma.

Rowe et al ¹⁴⁵ determined the therapeutic benefit of multivisceral resection in patients with locally advanced colorectal carcinomas. The study population was composed of 118 patients whose resection of the primary lesion included one or more adhesed adjacent secondary organs or structures. Their survival is reported in ▶ Table 22.5, but clinical relevance is that there was no statistically significant difference in the 5-year survival rates when multiple adjacent secondary organs or structures were resected and therefore they believe an aggressive operative approach is warranted. ^{160 , 161}

Yamada et al ¹⁶² reported 64 patients with locally advanced primary or recurrent rectal carcinoma with abdominoperineal resection with sacral resection performed in 9 patients, anterior pelvic exenteration in 8 patients, total pelvic exenteration in 27 patients, and total pelvic exenteration with sacral resection in 20 patients. Rates of morbidity, reoperation, and mortality were 50, 4.5, and 0% in 22 patients with primary carcinoma, and 60, 2.4, and 2.4% in 42 patients with recurrent disease, respectively. Major complications, such as sepsis, intra-abdominal abscess, and enteric fistula caused one hospital death and reoperation in two patients. In 21 patients who underwent curative resection for primary carcinoma, the overall 5-year survival rates were 74.1% for Dukes’ B and 47.4% for Dukes’ C although the difference was not statistically significant. Thirty patients with recurrent carcinoma who underwent curative resection had significantly improved survival with a 5-year survival rate of 22.9% compared with 12 patients who underwent palliative resection resulting in a survival rate of 0%.

In an excellent clinical review of the role of extended resection in the initial treatment of locally advanced colorectal carcinoma, Lopez and Monafo ¹⁶³ collated information on the results of multivisceral resection for colorectal carcinoma. In 11 publications in which 609 patients underwent extended resection for colorectal carcinoma, the operative morbidity rate was 27%, the operative mortality rate was 6%, and lymph node metastases occurred in 39%. The 5-year survival rate was 68% for node-negative status and 23% for node-positive status. If adherence to adjacent viscera was benign, the 5-year survival rate was 68%, but it declined to 40% if the attachment was malignant. Survival in locally advanced colorectal carcinoma is more dependent on lymph node status than on the extent of local invasion. ¹⁵⁰ In 23 publications in which 248 patients underwent total pelvic exenteration for rectosigmoid carcinoma, the operative morbidity rate was 60% and the operative mortality rate was 12%. The 5-year survival rate was 64% for node-negative status and 32% for node-positive status.

In the unique situation in which there is isolated invasion of the prostate by a rectal carcinoma, Campbell et al ¹⁶⁴ described the use of radical retropubic prostatectomy in conjunction with restorative proctosigmoidectomy for en bloc excision. This novel technique offers an alternative to total pelvic exenteration, thereby obviating the need for urinary and fecal diversion. The expected 5-year survival of patients subjected to en bloc resection ranged from 30 to 79% (▶ Table 22.5) and thus justifies an aggressive approach.

22.4.7 Urinary Tract Involvement by Colorectal Carcinoma

McNamara et al ¹⁶⁵ recently reviewed the literature on urinary tract involvement by colorectal carcinoma with the aim of highlighting technical and oncologic issues that should be considered when dealing with this complex problem. From the relevant literature, they identified three distinct clinical scenarios in which the urinary tract may be affected by colorectal carcinoma: involvement by primary colorectal carcinoma, involvement of recurrent carcinoma, and unexpected intraoperative findings of urinary tract involvement. The following information and guidelines draw heavily from their dissertation.

22.4.8 Unresectable Carcinoma

In the unusual circumstance in which a lesion is totally unresectable, it usually can be bypassed satisfactorily.

22.4.9 Palliative Resection

One of the most unsatisfying situations facing any surgeon who operates on patients with colon and rectal carcinoma is that of recommending a major abdominal procedure, with its potential complications, to a patient who has definite evidence of incurable disease. The decision regarding operative intervention is usually reached with some trepidation, since many of these patients are in poor physical condition and have a limited life expectancy. However, even for patients with metastatic carcinoma of the large bowel, resection performed to eliminate the symptoms of local disease has been advocated as a worthwhile procedure for avoiding the potential complications of obstruction and massive bleeding and the effects of local invasion of the primary lesion. In general, resections relieve patients of their symptoms and sometimes may even prolong life expectancy. ¹⁶⁶ The most common symptoms are pain and bleeding. ^{167 , 168}

It has been estimated that 10 to 20% of patients who are seen with primary operable colorectal carcinoma already have associated liver metastases. ¹⁶⁶ Unfortunately, not all patients will benefit from resection, and, in fact, some patients will be caused additional morbidity. This morbidity, together with the mortality of the operative procedure, may exceed the benefit of any temporary symptomatic relief. Thus, the role of palliative resection for malignant neoplasms has been questioned from time to time. This is especially true for the decision to perform a palliative abdominoperineal resection, an operation that entails not only an operative procedure of considerable magnitude, but also the establishment of a permanent colostomy in a patient with only a chance of limited survival.

In the presence of metastatic disease, survival will depend on the nature and extent of the metastases. Indeed, some metastatic lesions should be resected in addition to the primary lesion. Survival depends on the pattern of metastatic disease. For example, Joffe and Gordon ¹⁶⁷ noted survival with unilobar liver metastases to be 16.9 months, while with bilobar metastases, survival was only 8.5 months. Cady et al ¹⁶⁶ noted a survival of 13 months, Takaki et al ¹⁶⁹ 12 months, and Goslin et al ¹⁷⁰ a similar length of survival. Under such circumstances, the recommendation for resection should be tempered by a consideration of factors such as extensive hepatic replacement or jaundice, marked ascites, or massive peritoneal seeding, in which case life expectancy is very short and no benefit could be accrued from a resection. The prognosis is poor for patients with extensive liver metastases, patients older than 75 years, and patients with a previous history of cardiovascular disease. ¹⁶⁷

Mäkelä et al ¹⁷¹ reviewed 96 patients who underwent palliative operations with an 8% postoperative mortality rate (5% for resections and 17% for nonresection procedures) and a 24% postoperative morbidity rate. Median survival was 10 months (15 months for resections and 7 months for nonresection procedures) and 5% of patients survived longer than 5 years. The median relief of symptoms related to the malignancy was 4 months (4 months after resection and 1 month after nonresection procedures). Twenty-five patients underwent a second palliative operation.

Liu et al ¹⁶⁸ studied 68 patients with incurable colon carcinoma to try to identify objective criteria that might help surgeons decide which patients will benefit from palliative operations. The postoperative mortality rate was 10% and the complication rate was 10%. The mean survival after palliative resection was 10.6 months, after bypass was 3.4 months, and after diagnosis in patients not operated on was 2.0 months. Of the variables studied, the only factors affecting survival were poorly differentiated lesions and greater than 50% replacement of liver. The authors concluded that although resection carries with it a relatively high postoperative mortality rate, it is worthwhile as long as hepatic metastases occupy less than 50% of liver volume.

The macroscopic features of the primary disease must be taken into consideration because of the ever-present concern of obstruction. However, endoluminal stenting is an option in dealing with obstructive symptoms.

22.4.10 Synchronous Carcinomas

Recommendations for the appropriate treatment of synchronous carcinomas of the colon are at least in part based on the magnitude of the risk of development of metachronous adenomas and carcinomas after conventional resections. The incidence of synchronous carcinomas has been reported to be 1.5 to 7.6%. ¹⁷² In a series of 2,586 patients, an incidence of 1.8% was reported. ¹⁷² Bussey et al ¹⁷³ reported on 3,381 patients who survived conventional resections for carcinoma of the colon and rectum at St. Mark’s Hospital in London and found an overall incidence of metachronous carcinoma of 1.5%. The incidence rose to 3% in those cases followed up for at least 20 years. For those patients in whom an associated adenomatous polyp was found in the original operative specimen, the level rose to 5%. In a more recent study, synchronous carcinomas were found in 4.4% of patients. ¹⁷⁴ Passman et al ¹⁷⁵ reported on an 18-year multi-institutional database of 4,878 patients with colon carcinoma. There were 160 patients (3.3%) with 339 synchronous carcinomas. Eight percent of these patients had more than two lesions at the time of diagnosis. Based on highest stage lesion, 1% of patients were at stage 0, 28% at stage I, 33% at stage II, 25% at stage III, and 11% at stage IV. The disease-specific 5-year survival rate by highest stage was 87% for stage 0 or I, 69% for stage II, 50% for stage III, and 14% for stage IV. These “highest stage” survival rates for patients with synchronous carcinomas were not significantly different from survival of patients with same-stage solitary carcinomas in their database. In light of this, it seems reasonable that if synchronous carcinomas are located in the same anatomic region, a conventional resection should be performed. When the carcinomas are widely separated, a subtotal colectomy is the operation of choice.

22.4.11 Synchronous Polyps and Carcinoma

Recommendation for the treatment of patients with colon carcinoma and associated polyps involves the same considerations as for synchronous carcinomas. However, it also depends on the number, location, and size of these polyps. For example, if the polyps were confined to the region of the index carcinoma, the conventional operation for that portion of the bowel would be indicated. With the availability of colonoscopy, assessment and possible therapy of associated polyps can be accomplished. If the remaining bowel contains only occasional polyps that can be easily excised with the colonoscope, it would appear reasonable to have these polyps excised and to proceed with a conventional resection of the carcinoma. If one of the excised polyps should contain a carcinoma or if the polyps were of a size deemed in excess of colonoscopic polypectomy, a subtotal colectomy would be appropriate. ¹⁷⁶ Subtotal colectomy even has been recommended for colon carcinoma and synchronous polyp in good-risk patients. ¹⁷⁴ An individual who has exhibited the propensity for growth of many polyps in the colon, although not in adequate numbers to be considered familial adenomatous polyposis, would still qualify for a subtotal colectomy and ileorectal, or at least ileosigmoid, anastomosis.

22.4.12 Metachronous Carcinoma

Gervaz et al ¹⁷⁷ assessed the incidence of metachronous colorectal carcinomas in a population-based study of 500,000 residents. Of this total, 5,006 patients had sporadic carcinoma of the colon or rectum with 34% being located proximal to the splenic flexure. Occurrence of a second primary colorectal carcinoma was observed in 2.4% of this population. The risk for developing a second incidence of primary colorectal carcinoma was higher in patients whose initial carcinoma was located in the proximal colon (3.4 vs. 1.8%; OR, 1.9). The risk for each segment of large bowel was as follows: cecum, 3.4%; right colon, 3%; transverse colon, 3.8%; left colon, 2.8%; sigmoid colon, 1.7%; and rectum, 1.8%. By contrast, the risk for developing a second extracolonic carcinoma did not differ between patients with proximal and distal carcinomas (13.7 vs. 13.4%).

Shitoh et al ¹⁷⁸ reported that microsatellite instability can be regarded as an independent marker for predicting the development of metachronous colorectal carcinoma after operation. In a study of 328 colorectal carcinoma patients surveyed by periodic colonoscopy for at least 3 years after operation, 17 metachronous colorectal carcinomas were detected during the follow-up period. The percentage of microsatellite instability–positive cases was 26.4%. Incidences of metachronous colorectal carcinomas in microsatellite instability-positive and microsatellite instability–negative cases were 15.3 and 3%, respectively. The cumulative 5-year incidence of metachronous colorectal carcinomas was significantly higher in microsatellite instability–positive cases than in microsatellite instability–negative cases (12.5 vs. 2.5%).

22.4.13 Treatment of Metastatic Disease

Liver

Metastases to the liver from carcinoma of the colon or rectum are frequent occurrences. Indeed, the liver is the dominant site of treatment failure and the major cause of death in patients with colorectal carcinoma. Studies have demonstrated that up to 30% of patients undergoing apparently curative operation already have hepatic metastases that are not evident to the surgeon at the time of laparotomy. ^{179 , 180 , 181} Furthermore, another 50% have recurrent disease develop within the liver. ¹⁸² Some 90% of patients who die from colorectal carcinoma have liver metastases. ¹⁸³ In a study of doubling times, Finlay et al ¹⁸⁴ determined that the mean doubling time for overt metastases was 155 ± 34 days (± standard error of mean [SEM]) compared with 86 ± 12 days for occult metastases. The mean age of the metastases at the time of operation was estimated by extrapolation of the observed growth curve, assuming Gompertzian kinetics, to be 3.7 ± 0.9 years (± SEM) for overt metastases and 2.3 ± 0.4 years for occult metastases.

There is a perception that streamline flow of blood in the portal vein may influence the anatomic distribution of liver metastases, depending on the site of the primary lesion. It has previously been reported that carcinomas arising in the right colon are distributed to the right lobe of the liver 10 times more commonly than to the left lobe, whereas liver metastases from carcinomas arising from the left colon and rectum are believed to be distributed homogenously. Wigmore et al ¹⁸⁵ collected data prospectively on the anatomic site of hepatic metastases in 207 patients with colorectal metastases. This study could not find any evidence to support a differential pattern of metastasis within the liver dependent on the location of the primary colorectal carcinoma.

In an effort to accurately detect liver metastases, van Ooijen et al ¹⁸⁶ prospectively compared continuous CT angiography to preoperative ultrasonography and conventional CT in 60 patients with primary or secondary colorectal carcinoma. The standard references were palpation of the liver and intraoperative ultrasonography. Continuous CT angiography had a high sensitivity of 94% in contrast to ultrasonography (48%) and conventional CT (52%). However, there was a higher false-positive rate because of variations in the perfusion of normal liver parenchyma. Overall, continuous CT angiography had the highest accuracy (74%) compared with ultrasonography (57%) and CT (57%). The low specificity will hamper its routine application.

Strasberg et al ¹⁸⁷ reviewed 43 patients with metastatic colorectal carcinoma referred for hepatic resection after conventional staging with CT. PET scanning was performed on all patients. PET identified additional carcinoma not seen on CT in 10 patients. Operation was contraindicated in 6 of these patients because of the findings on PET. Laparotomy was performed in 37 patients. In all but 2, liver resection was performed. The Kaplan–Meier estimate of overall survival at 3 years was 77%. This figure is higher than the 3-year estimate of survival found in previously published series. They concluded that preoperative PET scan lessens the recurrence rate in patients undergoing hepatic resection for colorectal metastases to the liver by detection of disease not found on conventional imaging. Some patients who will not benefit from operation can thus be spared a laparotomy and major resection.

Liver surgeons usually recommend against biopsy of colorectal liver metastases because of the risk of local dissemination. Rodgers et al ¹⁸⁸ conducted a multicenter retrospective review of cases of colorectal liver metastases presenting for operation that had undergone a preoperative biopsy. Of 231 cases of colorectal metastases, 18.6% had undergone a preoperative biopsy. Evidence of dissemination related to the biopsy was 16%. Within the operative period (median 21 months), 3 of the 7 cases with evidence of dissemination and 11 of the 35 without dissemination were alive without disease. They concluded there is a significant risk of local dissemination with biopsy of colorectal liver metastases.

The value of intraoperative hepatic ultrasonography was discussed previously. Fuhrman et al ¹⁸⁹ reported on the use of intraoperative ultrasonography in the assessment of porta hepatis lymph nodes and the evaluation of resection margins to determine whether this modality would improve the selection of patients likely to benefit from operation. Of 151 patients undergoing exploration, 30 patients were considered unresectable and 14 (9.2%) demonstrated by intraoperative ultrasonography. The authors concluded that intraoperative ultrasonography did, indeed, improve the selection process.

The question of what to do for patients with these metastases has been a matter of controversy. At one point, any suggestion of an operative approach to metastatic disease was deemed foolish by some. The natural history of untreated hepatic metastases confirms a median survival of 6 to 12 months and of 4.5 months if metastases are synchronous. ^{190 , 191 , 192} If not resected, 3-year survival rates ranged from 3 to 7% and only 1 to 2% of patients will survive for 5 years. ¹⁸² Six studies of the natural history of such metastases in a total of 1,151 patients described a 5-year survival rate of 3%. In a study of 484 untreated patients, 6 independent determinants of survival were identified in the following order: (1) percent liver volume replaced by carcinoma, (2) grade of malignancy of the primary lesion, (3) presence of extrahepatic disease, (4) mesenteric lymph node involvement, (5) serum carcinoembryonic antigen (CEA), and (6) patient age. ¹⁹³ The prognosis is closely related to the extent of liver replacement.

A variety of chemotherapeutic regimens, including systemic chemotherapy and direct intraportal and intra-arterial modes of administration, have been attempted, all with limited and short (if any) benefit, but with the cost of considerable toxicity and anxiety. Systemic chemotherapy has resulted in response rate ranges of 18 to 28%, ¹⁹⁴ and the median survival rate ranges of 8 to 14 months. ¹⁸² Other efforts have been directed at hepatic artery embolization, hepatic artery ligation, and even irradiation, all without significant worthwhile benefit.

The lack of effective therapeutic alternatives has made hepatic resection the primary treatment consideration. Indeed, worthwhile survival rates in selected patients have been reported (▶ Table 22.6). ^{195 , 196 , 197 , 198 , 199 , 200 , 201 , 202 , 203 , 204 , 205 , 206 , 207 , 208 , 209 , 210 , 211 , 212 , 213 , 214} The timing and extent of operation varies. In the patient who presents with a synchronous lesion, which is amenable to operation, it appears appropriate to excise the lesion at the time of operation. If the lesion requires a major hepatic resection, the combination of partial hepatectomy and colectomy appears to be too great a task for one operation. After the colonic resection, if there is no other evidence of metastases and if a thorough evaluation, including a CT scan, has demonstrated removable disease, proceeding with resection is the treatment of choice. If the patient presents with metastatic disease at a later date, evaluation is necessary to ensure that the metastatic disease is confined; at the same time, evaluation should be performed to rule out the presence of recurrent disease at the area of the index carcinoma. The resection of hepatic metastases in patients with intra-abdominal extrahepatic disease is of no proven benefit. ²¹⁵ Even with preoperative staging, as many as 26% of patients will have intra-abdominal extrahepatic metastases, most commonly in portal and celiac lymph nodes. ²¹⁵ It is necessary to rule out evidence of other metastatic disease. Unfortunately, investigation of patients rarely unveils a solitary lesion. Only approximately 10% of patients develop metastases suitable for operation. In their study of the natural history of hepatic metastases from colorectal carcinoma, Wagner et al ²¹⁶ found that the median survival rate for unresected solitary and multiple unilobar metastases was 21 and 15 months, respectively. Earlier series reported untreated patients to have a median survival of 6 to 12 months. It is understandable why hepatic resection became an attractive option.

Table 22.6 Survival following resection of hepatic metastasis

Author(s)	No. of patients	Survival (%)		Operative mortality (%)	Complication rate (%)
		3 y	5 y
Hughes et al 195 , a	800		32
Schlag et al 196	122	40	30	4	34
Petrelli et al 197	62		26	8	30
Doci et al 198	100		30	5	41
Rosen et al 199	280	47	25	4
Nakamura et al 200	31	45	45	3	16
Van Ooijen et al 201	118		21	8	35
Gayowski et al 202	204	43	32	1
Scheele et al 203	434	45	33	4	22
Fuhrman et al 189 , b	107		44	3
Hananel et al 204	26		31	0	66
Rougier et al 205	123	35	21
Wade et al 206	133		26	4
Wanebo et al 207	74		24	7	35
Ohlsson et al 208	111		25	4
Fong et al 209	1,001	57	37	3	31
Buell et al 210	110	54	40	2	21
Elias et al 211	111	38	20	4	28
Kato et al 212 , c	585		33
Teh and Ooi 213	96	71		0	7
Weber et al 214	62	45	22	0	36
a Tumor registry of 24 institutions (24%, 5-year disease-free survival). b Ultimate patient selection with intraoperative ultrasonography. c Postoperative hepatic artery chemotherapy in 33% but no difference in survival noted in those with or without chemotherapy.

Surgical resection of primary colorectal carcinoma in patients with stage IV disease at initial presentation remains controversial. Although bowel resection to manage symptoms such as bleeding, perforation, or obstruction has been advocated, management of asymptomatic patients has not been well defined. Patient-dependent factors (performance status, comorbid disease) and extent of distant metastases are among the considerations that have an impact on the decision to proceed with operative management in asymptomatic stage IV colorectal carcinoma. To ascertain the natural history of a group of untreated patients and to evaluate simultaneously in another group whether or not the administration of systemic chemotherapy modifies this natural history, Luna-Perez et al ²¹⁸ followed up 77 patients with liver metastases from colorectal carcinoma. Untreated patients consisted of 45 patients; 41 developed extrahepatic metastatic disease and their median survival rate was 13 months. The group who received chemotherapy included 32 patients; 29 developed extrahepatic metastatic disease and their median survival was 15 months. There were no differences in overall survival in both groups. The administration of systemic chemotherapy did not modify the natural course of the disease. Dismal results of this nature mandate a better form of therapy, namely, operative. Ruo et al ²¹⁹ reviewed 127 patients who underwent elective resection of their asymptomatic primary colorectal carcinoma. Over the same time period, 103 stage IV patients who did not undergo resection were identified. The resected group could be easily distinguished from the nonresected group by a higher frequency of right colon carcinomas and metastatic disease restricted to the liver or one other site apart from the primary carcinoma. Resected patients had prolonged median (16 vs. 9 months) and 2-year (25 vs. 6%) survival compared with patients who were never resected. Univariate analysis identified three significant prognostic variables (number of distant sites involved, metastases to liver only, and volume of hepatic replacement by malignancy) in the resected group. Volume of hepatic replacement was also a significant predictor of survival. Subsequent to resection of asymptomatic primary colorectal carcinoma, 20% developed postoperative complications. Median hospital stay was 6 days. Two patients (1.6%) died within 30 days of operation. They concluded stage IV patients selected for elective palliative resection of asymptomatic primary colorectal carcinomas had substantial postoperative survival that was significantly better than those never having resection.

A review by Blumgart and Fong ¹⁸² revealed an operative mortality of less than 5% in most series, but complications arose in excess of 20% in most series. Myocardial complications were seen in 1%, pleural effusion requiring thoracotomy in 5 to 10%, pneumonia in 5 to 22%, and pulmonary embolism in 1%. Complications specifically related to liver resection included liver failure (3–8%), bile leak and biliary fistula (4%), perihepatic abscess (2–10%), and significant hemorrhage (1–3%). The most common sites for failure were the liver and lung with the liver involved as a site of recurrence in 45 to 75% of patients having liver resection. In light of this, adjuvant systemic chemotherapy seems to be an attractive option, but to date its role is unproven. Because the liver is the most common site of recurrence and may be the sole site in up to 40% of patients, regional hepatic chemotherapy is theoretically attractive, but studies in this arena have also failed to prove the benefit of that therapy. However, some studies are encouraging.

The prognosis of metastatic carcinoma is grave. Kuo et al ²²⁰ collected data from 74 patients with stage IV colorectal carcinoma to identify prognostic factors for predicting selection criteria for operative treatment in patients with metastatic disease. Overall survival time was 16.1 months. Survival in the curative resection group was significantly longer than in the noncurative groups (31.9 vs. 12.7 months). The operative mortality and morbidity rates were 5.6 and 21%, respectively. The two most common complications were leakage at the site of anastomosis and urinary tract infection. Based on these results, they concluded that patients older than 65 years with metastases at multiple sites, intestinal obstruction, preoperative CEA level 2,500 ng/mL, lactate dehydrogenase 2,350 units/L, hemoglobin less than 10 mg/dL, or hepatic parenchymal replacement by metastatic disease greater than 25% have poor prognosis for operative intervention. They noted the more aggressively they performed radical resection and metastasectomy in selected patients, the more survival benefits the patients obtained.

Simultaneous Colorectal and Hepatic Resection

Weber et al ²¹⁴ reported that in selected patients, simultaneous resection of the colorectal primary carcinoma and liver metastases does not increase mortality or morbidity rates compared with delayed resection, even if a left colectomy and/or a major hepatic resection are required. de Santibañes et al ²²¹ reviewed the results of liver resection performed simultaneously with colorectal resection in 71 cases. The median hospital stay was 8 days. Morbidity was 21% and included nine pleural effusions, seven wound abscesses, four instances of hepatic failure, three systemic infections, three intra-abdominal abscesses, and one colonic anastomotic leakage. Operative mortality was 0%. Recurrence rate was 57.7% and progression of disease was detected in 33.8%. Overall and disease-free survivals at 1, 3, and 5 years were 88, 45, and 38% and 67, 17, and 9%, respectively. Prognostic factors with notable influence on patient outcomes were nodal stage as per TNM classification, number of liver metastases, diameter (smaller or larger than 5 cm), liver resection specimen weight (lighter or heavier than 90 g), and liver resection margin (smaller or larger than 1 cm).

Chua et al ²²² retrospectively analyzed 96 consecutive patients with synchronously recognized primary carcinoma and hepatic metastases who underwent concurrent (64 patients) or staged (32 patients) colonic and hepatic resections. No significant differences were observed between concurrent and staged in type of colon resection or hepatic resection, overall operative duration, blood loss, volume of blood products transfused, perioperative morbidity (53 vs. 41%), disease-free survival from date of hepatectomy (median 13 vs. 13 months), or overall survival from date of hepatectomy (median 27 vs. 34 months). There was no operative mortality. Overall duration of hospitalization was significantly shorter for concurrent than for staged resection (mean 11 vs. 22 days). They concluded that concurrent colectomy and hepatectomy is safe and more efficient than staged resection and should be the procedure of choice for selected patients in medical centers with appropriate capacity and experience.

Tocchi et al ²²³ reviewed the results of 78 patients who underwent resection of primary colorectal carcinoma and hepatic metastases with curative intent. Adverse predictors of the longterm outcome included the number of metastases (> 3), preoperative CEA value greater than 100 ng/mL, resection margin less than 10 mm, and portal nodal status.

Tanaka et al ²²⁴ reported on 39 consecutive patients with synchronous colorectal carcinoma metastases to the liver who underwent curative simultaneous “one-stage” hepatectomy and resection of the colorectal primary. Only the volume of the resected liver was selected as a risk factor for postoperative complications (350 g mean resected liver volume in patients with postoperative complications vs. 150 g in those without complications). Patient age of 70 years or older and poorly differentiated mucinous adenocarcinoma as the primary lesion predicted decreased overall survival. They concluded that a one-stage procedure appears desirable for synchronous colorectal hepatic metastases except for patients requiring resection of more than one hepatic section, patients aged 70 years or older, and those with poorly differentiated or mucinous adenocarcinomas as primary lesions.

Currently, there is no consensus as to which factors are important in selecting patients for operation and which factors are important in determining the patient’s prognosis. For example, Attiyeh and Wichern ²²⁵ found no significant difference in the survival rates of patients with a solitary metastasis and in those with multiple lesions, nor was survival influenced by the size of the metastasis. The survival rate was better in patients whose primary colorectal lesion was Dukes’ B compared with those whose lesion was Dukes’ C. Adson ²²⁶ listed several determinants for a favorable prognosis, including (1) primary colorectal carcinoma of limited locoregional extent (Dukes’ A or B), (2) presence of fewer than four liver lesions, (3) metastases that appear a long time after the primary lesion was removed, (4) lesions that can be removed with wide margins, and (5) lack of extrahepatic metastases. Combining their own experience with reports in the literature, Bozzetti et al ²²⁷ found that sites of failure after liver resection were hepatic in 16%, extrahepatic in 15%. and both in 14%. Patterns of recurrence in our patients were hepatic in 31%, hepatic and an extrahepatic site in 15%, and lung in 15%. ²⁰⁴ Nagorney ¹⁹¹ reported that the only characteristic associated with prolonged survival was the stage of the primary lesion, with Dukes’ B patient survival being greater than Dukes’ C. In Nagorney’s review, site of origin and degree of differentiation of the primary carcinoma did not correlate with survival rate. Characteristics of metastatic disease that influenced survival included the number of hepatic metastases (one to three are better than four or more), the interval between diagnosis of the primary lesion and hepatic metastases, the resection margin (a margin > 1 cm is better), and the presence of extrahepatic disease. The size and distribution of lesions within the liver had no association with survival. We were also interested in variables related to survival and reviewed 26 selected patients with liver colorectal metastases who underwent hepatic resection. ²⁰⁴ The patient’s age, sex, site of primary lesion, histologic grade, lymph node involvement, location, size, and number of hepatic metastases, type of hepatic resection, and preoperative CEA blood levels were documented. Complete removal with histologically negative resection margins were accomplished in 24 patients. The extent of resection performed was hepatic lobectomy in 12 patients, segmentectomy in 8 patients, and wedge resection in 4 patients. The 5-year survival rate was 30.5%. Patients with metachronous metastases had a better survival rate than those with synchronous lesions (46.6 vs. 13.6%, respectively). None of the other factors studied showed a significant effect on survival. During a median follow-up of 30.9 months, 20 patients developed recurrence of their disease (60% in the liver). There was no perioperative mortality. Morbidity arose in 66% of patients, with a majority of the complications minor. Wanebo et al ²⁰⁷ reported a significant relationship with survival and the number of metastases (three or fewer vs. four or more), the presence of bilobar versus unilobar metastases, and the extent of liver resection (wedge and segmental vs. lobectomy and trisegmentectomy). They believe that resection of bilobar disease or extended resection should generally be avoided, especially in medically compromised patients. Nakamura et al ²⁰⁰ adopted a very aggressive approach for patients suffering from liver metastases. Of 31 patients, 22 underwent lymph node dissection of the hepatic hilus, in the minds of most surgeons, a current contraindication to hepatic resection.

Six of the 22 patients who underwent lymph node dissection had nodes positive for carcinoma. Ten patients underwent removal of recurrent lesions in the liver, lung, adrenal glands, and brain after initial hepatic resection. Based on an overall 5-year survival rate of 45%, the authors concluded that repeat hepatectomy and dissection of hilar lymph nodes improves prognosis in selected patients with hepatic metastases of colorectal carcinoma.

In an analysis of risk factors, Gayowski et al ²⁰² found that gender, Dukes’ classification, site of primary colorectal carcinoma, histologic differentiation, size of metastatic lesion, and intraoperative transfusion requirement were not statistically significant prognostic factors. In patients 60 years of age or older, an interval of 24 months or less between colorectal and hepatic resection, four or more metastatic lesions, bilobar involvement, positive resection margins, lymph node involvement, and the direct invasion of adjacent organs were significant poor prognostic factors.

Hughes et al ¹⁹⁵ collated information from a registry of 24 participating institutions. Factors that they found to affect prognosis detrimentally were (1) more than four metastatic lesions, (2) a short disease-free interval from initial resection to appearance of metastases (< 1 year), (3) a pathologic margin of less than 1 cm on the liver specimen, and (4) the presence of lymph node metastases at the time of initial resection.

Using a multivariate regression analysis, Scheele et al ²⁰³ found that survival was dependent on the presence of satellite metastases, grade of the primary carcinoma, time of the diagnosis of metastases (synchronous vs. metachronous), diameter of the largest metastases (> 5 cm), anatomic versus nonanatomic approach, year of resection, and mesenteric lymph node involvement. Rougier et al ²⁰⁵ studied 544 patients with resected hepatic metastases from colorectal carcinoma to determine prognostic factors. Among the 20 variables assessed, 8 items were singled out. In decreasing order of relative risk, they included performance status (2–4 vs. 0–1), alkaline phosphatase level (greater than normal vs. normal), number of involved segments (24 vs. 3), chemotherapy (no vs. yes), extrahepatic metastases (yes vs. no), primary location (right vs. other), prothrombin time (< 75 vs. > 75%), and resection of the primary carcinoma (no vs. yes). Specific criteria for the selection process are constantly evolving. Adson ²²⁶ has offered a thoughtful set of guidelines. Patients whose primary colorectal lesions are well confined, who have one to three evident unilobar hepatic metastases that likely can be removed with wide margins, and who have no evidence of extrahepatic metastases should undergo resection. Patients with extrahepatic metastases, numerous hepatic metastases involving more than one-half of the liver, large lesions that encroach on major hepatic veins, or contralateral hilar ducts or veins or lesions sited so as to preclude resection with free margins have an unfavorable prognosis and should not undergo resection. Unfortunately, many patients do not fall neatly into one of these categories, and the surgeon must exercise considerable judgment in making a definitive recommendation. ²²⁷

The role of neoadjuvant chemotherapy for patients with multiple (five or more) bilobar hepatic metastases irrespective of initial resectability is being considered with increased frequency. Tanaka et al ²²⁸ compared the outcome of hepatectomy alone with that of hepatectomy after neoadjuvant chemotherapy for multiple bilobar hepatic metastases from colorectal carcinoma. The outcome of 48 patients treated with neoadjuvant chemotherapy, followed by hepatectomy was compared with that of 23 patients treated by hepatectomy alone. Patients who received neoadjuvant chemotherapy had better 3- and 5-year survival rates from the time of diagnosis than those who did not (67.0 and 38.9% vs. 51.8 and 20.7%, respectively) and required few extended hepatectomies (four segments or more; 39 of 48 vs. 23 of 23). In patients with bilateral multiple colorectal liver metastases, neoadjuvant chemotherapy before hepatectomy was associated with improved survival and enabled complete resection with fewer extended hepatectomies.

Allen et al ²²⁹ compared the treatment and outcome in patients referred for staged resection of synchronous colorectal liver metastases between patients who did not receive neoadjuvant chemotherapy and had exploratory operations after recovery from colon resection and patients who did receive chemotherapy before liver resection. Neoadjuvant chemotherapy was given to 52 patients; in 29 of them, the disease did not progress, but in 17 the disease progressed while they were receiving treatment. Median follow-up was 30 months. Five-year survival was statistically similar between patients who received and did not receive neoadjuvant therapy (43 vs. 35%). Patients within the neoadjuvant group whose disease did not progress while they were receiving chemotherapy experienced significantly improved survival as compared to patients who did not receive chemotherapy (85 vs. 35%). In the setting of synchronous colorectal metastases, the response to neoadjuvant chemotherapy may be a prognostic indicator of survival and may assist in the selection of patients for conventional or experimental adjuvant therapies.

Fong et al ²⁰⁹ reported on 1,001 consecutive patients undergoing liver resection for metastatic colorectal carcinoma. These resections included 237 trisegmentectomies, 394 lobectomies, and 370 resections encompassing less than a lobe. The operative mortality rate was 2.8%. The 5-year survival rate was 37% and the 10-year survival rate was 22%. Seven factors were found to be significant and independent predictors of poor long-term outcome: positive margin, extrahepatic disease, node-positive primary, disease-free interval from primary to metastases less than 12 months, number of hepatic lesions greater than 1 cm, largest hepatic lesion greater than 5 cm, and CEA level greater than 200 ng/mL. When the last five of these criteria were used in a preoperative scoring system, assigning 1 point for each criterion, the total score was highly predictive of outcome. The 5-year actuarial survival for patients with 0 points was 60%, 1 point was 44%, 2 points was 40%, 3 points was 20%, 4 points was 25%, and 5 points was 14%. In fact, no patient with 5 points survived 5 years. Patients with up to two criteria can have a favorable outcome. Patients with three, four, or five criteria should be considered for experimental adjuvant trials.

Iwatsuki et al ²³⁰ examined various clinical and pathologic risk factors in 305 consecutive patients who underwent primary hepatic resection for metastatic colorectal carcinoma. Preliminary multivariate analysis revealed that independently significant negative prognosticators were (1) positive surgical margins, (2) extrahepatic carcinoma involvement including the lymph nodes, (3) three or more metastatic lesions, (4) bilobar metastases, and (5) time from treatment of the carcinoma to hepatic recurrence of 30 months or less. Because the survival rates of the 62 patients with positive margins or extrahepatic metastases were uniformly very poor, multivariate analysis was repeated in the remaining 243 patients who did not have these lethal risk factors. The reanalysis revealed that independently significantly poor prognosticators were (1) three or more metastases, (2) metastases size greater than 8 cm, (3) time to hepatic recurrence of 30 months or less, and (4) bilobar metastases. Risk scores (R) for recurrence were divided into five groups: grade 1, no risk factors; grade 2, one risk factor; grade 3, two risk factors; grade 4, three risk factors; and grade 5, four risk factors. Grade 6 consisted of the 62 culled patients with positive margins or extrahepatic metastases. Estimated 5-year survival rates of grade 1 to 6 patients were 48.3, 36.6, 19.9, 11.9, 0, and 0%, respectively. The proposed risk-score grading predicted the survival differences.

Smith et al ²³¹ found that in patients who are undergoing curative resection of hepatic colorectal metastases, an elevated expression of the biomarkers hTERT and Ki-67 are better predictors of poor long-term survival than is a score based on clinical features. Kato et al ²¹² reported on 585 patients who underwent hepatectomy at 18 institutions. The 5-year survival rate for those treated by hepatectomy was significantly higher (32.9%) than for those not undergoing hepatectomy (3.4%). After hepatectomy for hepatic metastases, the most prevalent form of recurrence was in the remnant liver (41.4%), followed by recurrence of pulmonary metastases (19.2%), and other (7.2%). Factors of the primary carcinoma that adversely affect prognosis included poorly differentiated adenocarcinoma or mucinous carcinoma, depth of invasion, lymph node metastases of stage n3 and n4 by the Japanese classification of colorectal carcinoma, number of metastatic lymph nodes of more than four, and Dukes’ stage D. Factors at the time of hepatectomy adversely affecting prognosis after operation for hepatic metastases included residual carcinoma, extrahepatic metastases, hepatic metastases of degree H3 stipulated by the Japanese classification of colorectal carcinoma, number of metastases of four or more, pathology of hepatic metastases of poorly differentiated carcinoma, resection margin of less than 10 mm, and CEA value higher than normal preoperative and 1-month postoperative.

Indications for hepatectomy in patients with four or more hepatic colorectal metastases remain controversial. Imamura et al ²³² reviewed data from 131 patients who underwent a total of 198 hepatectomies. Patients were grouped according to the number of metastases. The 5-year survival rate of patients with 1 to 3, 4 to 9, and 10 or more metastases were 51, 46, and 25%, respectively. They concluded hepatic resection for patients with four to nine metastases clearly is warranted. On the other hand, in high volume centers at which the operative mortality rate is nearly zero, the presence of 10 or more nodules may not be an absolute contraindication to surgical therapy.

In the review by Jaeck ²³³ the 5-year survival rate for resection of colorectal liver metastases ranged from 20 to 54%. However, the resectability rate of colorectal liver metastases is reported to be less than 20%. This limitation is mainly due to insufficient remnant liver and to extrahepatic disease. Among extrahepatic locations, lymph node metastases are often considered indications of a very poor prognosis and a contraindication to resection. He found that the presence of hepatic pedicle lymph node metastases ranged from 10 to 20%. When located near the hilum and along the hepatic pedicle, they should not be considered an absolute contraindication to resection, and extended lymphadenectomy should be performed. However, when they reach the celiac trunk, there is no survival benefit after resection of colorectal liver metastases.

Elias et al ²³⁴ reported the long-term outcome and prognostic factors of 75 patients who underwent a complete R0 resection of extrahepatic disease simultaneously with hepatectomy for colorectal liver metastases. Extrahepatic disease localization included peritoneal carcinomatosis (limited), hilar lymph nodes, local recurrences, retroperitoneal nodes, lung, ovary, and abdominal wall. The mortality rate was 2.7% and morbidity was 25%. After a median follow-up of 4.9 years, the overall 3- and 5-year survival rates were 45 and 28%, respectively. They concluded extrahepatic disease in colorectal carcinoma patients with liver metastases should no longer be considered as a contraindication to hepatectomy. However, there must be an intended R0 resection, and it is inappropriate for patients with multiple extrahepatic disease sites or more than five liver metastases.

The optimal operative strategy for the treatment of synchronous resectable colorectal liver metastases has not been defined. Martin et al ²³⁵ reviewed their experience with 240 patients who were treated surgically for primary adenocarcinoma of the large bowel and synchronous hepatic metastasis. One hundred thirty-four patients underwent simultaneous resection of a colorectal primary and hepatic metastases in a single operation (group 1), and 106 patients underwent staged operations (group 2). Simultaneous resections tend to be performed for right colon primaries, smaller, and fewer liver metastases, and less extensive liver resection. Complications were less common in the simultaneous resection group, with 49% sustaining 142 complications compared with 67% sustaining 197 complications for both hospitalizations in the staged resection group. Patients having simultaneous resection required fewer days in hospital (median 10 vs. 18 days). Perioperative mortality was similar (3 each in simultaneous and staged). They believe simultaneous resection should be considered a safe option in selected patients with resectable synchronous colorectal metastases.

Nelson and Freels ²³⁶ assessed the effect of posthepatic resection with hepatic artery chemotherapy on overall survival. Trials were sought in Medline, the Cochrane Controlled Trial Register, the Cochrane Hepatobiliary Group Trials Register, and through contact of trial authors and reference lists using key words. Overall survival at 5 years in the hepatic artery group was 45 and 40% in the control group. No significant advantage was found in the meta-analysis for hepatic artery and chemotherapy measuring overall survival. Adverse events related to hepatic artery therapy were common including five therapy-related deaths. They concluded that this added intervention for the treatment of metastatic colorectal carcinoma cannot be recommended at this time.

Clancy et al ²³⁷ conducted a meta-analysis of prospective clinical trials to determine if adjuvant hepatic arterial infusion confers a survival benefit to treat residual microscopic disease after curative hepatic resection for colorectal carcinoma metastases. Prospective clinical trials comparing hepatic arterial chemotherapy after curative hepatic resection for colorectal carcinoma metastases against a control arm were included. The outcome measure was survival difference at 1 and 2 years after operation. Seven studies met the inclusion criteria, and all except one were randomized trials. The survival difference in months was not statistically significant at 2 years. Based on these findings, they concluded routine adjuvant hepatic artery infusion after curative resection for colorectal carcinoma of the liver cannot be recommended.

Bines et al ²³⁸ reported on a review of 131 patients who underwent hepatic resection for metastatic colorectal carcinoma. There were 31 recurrences and, of those, 13 underwent re-resection with a morbidity rate of 23%, a mortality rate of 8%, and a 5-year survival rate of 23%. The authors concluded that in properly selected patients, repeat resection yields results similar to those after initial resection. Wanebo et al ²⁰⁷ reported that 12% of their patients had repeated resection of metastases, with an overall 5-year survival rate of 43% after the first resection and 22% after the second resection. In their review of 10 reports, Blumgart and Fong ¹⁸² noted that between 15 and 40% of patients who undergo resection for hepatic metastases have the liver as the sole site of recurrence, and approximately one-third will be candidates for further resection. In the 146 patients collated, the operative mortality rate was 3%, and the complications encountered were similar to those that developed after initial resection. These results were in highly selected patients. The median survival was greater than 30 months when calculated from the time of second liver resection and greater than 47 months when calculated from the time of the first resection. However, there were only four 5-year survivors. Although resection is feasible, only approximately 5% of all patients undergoing further resection will come to a second resection. ¹⁸² Wanebo et al ²³⁹ reported recurrence rates in 65 to 85% of patients after initial hepatectomy for metastases for colorectal carcinoma. Approximately one-half of these have liver metastases and in 20 to 30% only the liver is involved. The opportunity for resection is frequently limited because of diffuse liver disease or extrahepatic extension, and only approximately 10 to 25% of these patients have conditions amenable to resection. The authors’ comprehensive review of the 28 series showed that the mean interval between the first and second liver resections varied from 9 to 33 months and was approximately 17.5 months in the two largest series. The median survival in the series reporting 10 or more patients was 19 months (mean, 24 months), which is comparable to data in the single resection series. In the large French Association series containing 1,626 patients with single resections and 144 patients with two resections, the 5-year survival rates were 25 and 16%, respectively. The recurrence rate after repeat resection was high (> 60%), and half of the recurrences were in the liver. The prognostic factors favoring repeat resection are variable, but they include absence of an extrahepatic extension of carcinoma and a complete resection of liver metastases. The authors concluded that repeat hepatic liver resection for metastatic colorectal carcinoma in carefully selected patients appears warranted. From their own experience and review of the literature, Pinson et al ²⁴⁰ came to the same conclusion. For patients collected from the literature, the authors constructed a survival curve (▶ Fig. 22.31). Fernández-Trigo et al ²⁴¹ were also encouraged to perform repeat hepatic resections for colorectal metastases because it remains the only curative treatment. Others concur with this course of management. ²⁴²

Takahashi et al ²⁴³ reviewed clinical data of patients undergoing repeat hepatectomy for metastatic colorectal carcinoma compared with those of initial hepatectomy to determine criteria for repeat hepatectomy. For 22 patients who underwent repeat hepatectomy, no mortality and an 18% morbidity rate were observed. The 3-year survival rate after hepatectomy was 49%. The only poor prognostic factor after repeat hepatectomy was a serum CEA level greater than 50 ng/mL before initial hepatectomy. Suzuki et al ²⁴⁴ assessed the risks and clinical benefits of repeat hepatectomy for those patients who underwent hepatectomy for colorectal metastases. There was no operative mortality after repeat hepatectomy in 26 patients. Operative bleeding was significantly increased in the second hepatectomy, but operating time, duration of hospital stay, and performance status after the second hepatectomy were comparable with those of the initial hepatectomy. The median survival time from the second hepatectomy was 31 months and the 3- and 5-year survival rates were 62 and 32%, respectively. A short disease-free interval (6 months or less) between the initial hepatectomy and diagnosis of hepatic recurrence in the remnant liver was significantly associated with poor survival after the second hepatectomy. Oshowo et al ²⁴⁵ conducted a study aimed to compare outcome in patients with solitary colorectal liver metastases treated by operation (n = 20) or by radiofrequency ablation. Most patients in both groups also received systemic chemotherapy. Median survival after liver resection was 41 months with a 3-year survival of 55.4%. There was one postoperative death and morbidity was minimal. Median survival after radiofrequency ablation was 37 months with a 3-year survival of 52.6%. In this study, survival after resection or radiofrequency ablation of solitary colorectal liver metastases was comparable. The latter is less invasive and requires either an overnight stay or day-case facilities only.

Berber et al ²⁴⁶ determined the predictors of survival at the time of laparoscopic radiofrequency thermal ablation in 135 patients with colorectal liver metastases who were not candidates for resection. The median survival for all patients was 28.9 months. Patients with a CEA less than 200 ng/mL had improved survival compared to those with a CEA more than 200 ng/mL (34 vs. 16 months). Patients with the dominant lesion less than 3 cm in diameter had a median survival of 38 versus 34 months for lesions 3 to 5 cm, and 21 months for lesions greater than 5 cm. Survival approached significance for patients with one to three metastases versus more than three metastases (29 vs. 22 months). The presence of extrahepatic disease did not affect survival. Only the largest liver metastases more than 5 cm was found to be a significant predictor of mortality with a 2.5-fold increased risk of death versus the largest liver metastases less than 3 cm in size.

Ueno et al ²⁴⁷ collected data from 68 patients who underwent resection of colorectal liver metastases who might benefit from prophylactic regional chemotherapy. The extrahepatic recurrence rate at 3 years after hepatectomy was 57.8%. Three variables were independently associated with extrahepatic recurrence including raised serum level of CEA after hepatectomy (relative risk 5.4), venous invasion of the primary carcinoma (relative risk 4.0), and high-grade budding of the primary carcinoma (relative risk 3.1). Patients with none of these risk factors had a 3-year extrahepatic recurrence rate of 7.1% compared with 61.6% for those with one risk factor and 100% for those with two or three risk factors. This system might be used on an individual basis to select patients with colorectal liver metastases for regional chemotherapy or systemic chemotherapy after operative intervention. The value of postoperative chemotherapy following resection of hepatic metastases was reviewed by Cohen and Kemeny. ²⁴⁸ Two studies compared hepatic artery infusion with no treatment and no overall survival benefit was reported. In one study, there was a modest 2-year survival improvement from 72 to 86%. Drugs used included floxuridine (FUDR), and 5-FU/LV, and in light of newer drugs used for systemic chemotherapy, this improvement may not be relevant. For patients with unresectable liver metastases, Gray et al ²⁴⁹ reported on the use of embolization of yttrium-90-containing microspheres into the liver via a catheter inserted into the hepatic artery at laparotomy. In 29 patients, the CEA levels fell in the 26 patients in whom this therapy was tested, and there was CT evidence of reduction in 48% of the 22 patients reexamined. Some patients also received continuous chemotherapy infusion to potentiate the radiation effect. Although this is an important first step, there is no evidence that reduction of disease is translated to either improved survival or quality of life. Stubbs et al ²⁵⁰ treated 50 patients with advanced nonresectable colorectal liver metastases with selective internal radiation therapy. Estimated liver involvement was less than 25% in 30 patients, 25 to 50% in 13 patients, and greater than 50% in 7 patients. A single dose of between 2.0 and 3.0 GBq of 90-yttrium microspheres was injected into the hepatic artery via a subcutaneous port and followed at 4-week intervals by regional chemotherapy with 5-FU. Treatment-related morbidity did occur including a 12% incidence of duodenal ulceration. Median survival for patients with extrahepatic disease was 6.9 months. For patients with no extrahepatic disease, median survival was 17.5 months. Substantial destruction of liver metastases can be achieved in more than 90% of patients with a single treatment. Lang and Brown ²⁵¹ recommended the selective embolization of doxorubicin and ethiodized oil for unresectable hepatic metastases. In his review, Stuart ²⁵² found that chemoembolization for patients with metastatic colorectal carcinoma appears to be a reasonable alternative for many who are not operative candidates. Response rates of approximately 50% have been reported, with survival longer than would be expected in studies of systemic therapy among patients who had failed standard chemotherapy. Survival may be especially enhanced in treated patients who have no extrahepatic metastases.

Cryoablation has also been used in this clinical setting but has not yet been proven to improve outcome. ¹⁸² For patients with unresectable hepatic metastases, Weaver et al ²⁵³ reported the use of hepatic cryosurgery with or without resection in 140 patients, 119 of whom had carcinomas that were colorectal in origin. The median number of lesions treated was three. The operative mortality rate was 4%, and complications included coagulopathy, hypothermia, myoglobinuria, pleural effusion, acute tubular necrosis, and infection. The median survival rate was 27 months. Ruers et al ²⁵⁴ reported on the long-term efficacy of cryosurgery as an adjunct to hepatic resection in patients with colorectal liver metastases not amenable to resection alone. Thirty patients met the following inclusion criteria: metastases confined to the liver and judged unresectable, 10 or fewer metastases, cryosurgery alone or in combination with hepatic resection allowed disease clearance. Median follow-up was 26 months. Overall, 1- and 2-year survival rates were 76 and 71%, respectively. Median survival was 32 months. Disease-free survival rates at 1 and 2 years were 35% and 7%, respectively. Six patients developed recurrence at the site of cryosurgery; given that the total number of cryosurgery-treated lesions was 69, the local recurrence rate was 9%.

Only gold members can continue reading. Log In or Register to continue

Share this:

• Click to share on Twitter (Opens in new window)
• Click to share on Facebook (Opens in new window)

Related posts:

18 Rectal Prolapse 19 Benign Neoplasms of the Colon and Rectum 27 Diverticular Disease of the Colon 30 Mesenteric Vascular Diseases 33 Constipation 32 Intestinal Stomas

Stay updated, free articles. Join our Telegram channel

Join