Gastrointestinal Cancer: Surgical Oncology

Monica M. Bertagnolli

Hippocrates had this to say about intestinal cancer: “It is better not to treat deep cancers, for those so treated die rapidly whereas the nontreated live for a longer time” (1). This statement held true until the advent of safe anesthesia and surgery, advancements that were not fully developed until the mid-20th century. The earliest successful surgical treatment of a gastrointestinal cancer was probably accomplished by Pillore of Rouen, who performed a cecostomy for an obstructing colon cancer in 1776. Unfortunately, in a disastrous attempt to relieve the obstruction, the patient ingested 2 lb of mercury before surgery. The patient died 28 days after surgery, and an autopsy showed that the cause of death was obstruction of the distal small intestine by the mercury (2). In 1844, Reybard reported a survival after resection and anastomosis for cancer of the colon. This is a remarkable accomplishment, considering that it antedates general anesthesia. The most significant of the many problems encountered during early intestinal surgery was the extremely high rate of fatal infection in the era before wide acceptance of aseptic technique and availability of antibiotics. In 1879, Christian Billroth resected a sigmoid cancer and exteriorized the proximal bowel as a permanent colostomy. This operation had a much lower incidence of infection and mortality than earlier attempts at resection and anastomosis. Billroth’s technique was later modified by Bloch (1892), Paul (1895), and von Mikulicz-Radecki (1895–1905) to what became known as an “obstructive” resection for colon cancer. During these procedures, the loop of bowel containing the tumor was brought outside the abdominal cavity, and the incision was closed. Approximately 1 week after the initial surgery, the bowel was divided with a cautery, removing the segment containing the tumor and creating a loop colostomy. Because the abdominal cavity was closed several days before the bowel was opened, peritoneal contamination was less, and mortality from infection was significantly reduced. This technique was widely used by American surgeons as recently as 1940 (2).

The first successful surgery for gastric carcinoma is attributed to Billroth in 1881. During this procedure, the antrum and pylorus were removed and the duodenum was sutured to the gastric cardia, an operation that later became known as a Billroth I gastrectomy. W.C. Roentgen discovered x-rays in 1895, and by 1910, contrast-enhanced radiography was used to diagnose intestinal cancers, which allowed somewhat earlier diagnosis and better preoperative planning. The first cancer operation to meet present-day standards of adequate primary tumor resection and complete lymphadenectomy was performed by Miles, who developed the combined abdominoperineal resection in 1926. In 1938, Whipple successfully resected a tumor in the region of the pancreatic head, a complex procedure that involved removal of the distal stomach, duodenum, and pancreatic head, and required reanastomosis of the pancreatic duct, distal common bile duct, and stomach to the proximal jejunum. After World War II, advances in blood transfusion, antibiotics, metabolic support, and anesthesia substantially reduced the mortality from radical cancer resections such as these.

In 1958, Hirschowitz developed the first flexible fiberoptic instrument for endoscopy, and instruments allowing routine evaluation of the upper and lower gastrointestinal tract became widely available in the 1970s. These advances allowed the earlier detection of neoplastic lesions and endoscopic removal of premalignant adenomas, the first intervention shown to decrease colon cancer incidence. During the 1970s, the distinct natural histories of different tumors were recognized. For most solid tissue tumors, clinicians developed the view that tumor involvement of regional lymph nodes was an indicator of systemic disease unlikely to respond to more aggressive local surgery. In the late 20th century, contributions from radiation and chemotherapy produced modest improvements in survival when used as adjuvant therapy for patients with stage III colorectal cancer. Comparative gains are still lacking for the other primary gastrointestinal cancers.

This book contains sections that provide a comprehensive review of surgery for specific gastrointestinal cancers. This introductory chapter focuses on the general approach to surgical management of gastrointestinal cancer patients and also describes some of the less understood therapeutic approaches, such as intraoperative radiation therapy and intraperitoneal chemotherapy. In addition, this section outlines several new areas of investigation in management of gastrointestinal cancers that may improve patient care in the future.

Surgical Evaluation

Preoperative Risk Assessment

Although minimal access surgery and early postoperative feeding have decreased the physiological impact of intestinal surgery, such surgery still requires general anesthesia and is frequently performed on patients with comorbid conditions and nutritional compromise. It is worthwhile, therefore, to consider a few issues in the preoperative evaluation and preparation of these patients. Decision making in major cancer surgery requires balancing of surgical risks and benefits, and the first step in this process involves assessing the patient’s physiological reserves. This is particularly important in cancer patients of advanced age because age alone is not an accurate predictor of surgical risk. Physical classification systems to identify patients with increased risk of adverse outcomes after major
surgery include those developed by the American Society of Anesthesiologists and by Goldman et al. (3) (Table 5.1).

Table 5.1 Functional classification of physical status

Risk category	American Society of Anesthesiologists classification	Goldman classification
Class I	No physiological, biochemical, or psychiatric disturbance; the pathological process for which surgery is to be performed is localized and not conducive to systemic disturbance	Ordinary physical activity, such as walking and climbing stairs, does not cause angina. Angina occurs with strenuous, rapid, or prolonged exertion at work or recreation.
Class II	Mild to moderate systemic disturbance caused either by the condition to be treated surgically or by other pathophysiological processes	Slight limitation of ordinary activity. Angina occurs with walking or climbing stairs rapidly; walking uphill; walking or climbing stairs after meals or in cold, in wind, or under emotional stress; or during the first few hours after awakening. Angina occurs when walking more than two blocks on the level or climbing more than one flight of stairs at a normal pace and in normal conditions.
Class III	Severe systemic disturbance or pathology	Marked limitation of ordinary physical activity. Angina occurs when walking one to two blocks on the level or climbing one flight of stairs at a normal pace and in normal conditions.
Class IV	Severe systemic disorder that is immediately life threatening and not always correctable by the operative procedure	Inability to carry on any physical activity without discomfort. Angina may be present at rest.
Class V	Moribund condition; little chance of surviving surgery
Emergency (E)	Any patient in one of the classes listed in this table operated on in an emergency situation

It is particularly important to identify significant cardiac risk factors present before surgery because more than 50% of all major perioperative complications or deaths are related to cardiovascular disease. In addition, optimal perioperative physiological support can decrease surgical morbidity and mortality in patients with cardiac disease. The most commonly used method of assigning cardiac risk for patients undergoing major surgery was developed by Goldman et al. (3) (Tables 5.2 and 5.3). Based on this classification, patients whose preoperative state places them in risk category III or IV should be considered for invasive perioperative monitoring to allow optimal physiological support. Patients with evidence of potentially treatable cardiovascular disease, such as unstable angina or transient ischemic attacks, should undergo complete evaluation and treatment of these conditions before tumor resection. For patients with good functional status and no signs of cardiovascular disease, the risk and cost of noninvasive preoperative cardiac evaluation are not warranted because these individuals have a low risk of perioperative myocardial infarction or death from cardiac causes.

Nutrition

Most patients undergoing cancer surgery can withstand the associated brief period of nutritional deficit and catabolism without difficulty. It is not unusual, however, to encounter a patient whose nutrition has been compromised by his or her underlying condition. The degree of compromise can range from mild, with no adverse effect on treatment outcome, to cancer cachexia, a paraneoplastic syndrome characterized by anorexia, weight loss, and progression to multiple-organ dysfunction. For many patients with intestinal tumors, anorexia and the resulting malnutrition are exacerbated by tumor-associated changes in gastrointestinal function and by the surgery, chemotherapy, and radiation therapy used to treat the tumor. It is essential, therefore, to assess the nutritional status of cancer patients before therapy and provide optimal nutritional support during treatment and recovery.

Nutritional assessment for patients with cancer includes taking a dietary history, with documentation of recent weight change, anorexia, early satiety, or dysphagia. Findings on physical examination that suggest malnutrition include evidence of muscle wasting; dry, flaky skin texture; brittle hair or unusual hair loss; and ridging or spooning of the nails. Important laboratory tests for nutritional evaluation include serum albumin and transferrin levels. These assessments allow patients to be classified into clinically relevant categories that indicate the extent of their nutritional reserves (Table 5.4).

Maintenance of adequate preoperative oral nutrition should be a high priority for all patients undergoing major surgery. This imperative is sometimes overlooked when gastrointestinal cancer patients receive multiple preoperative endoscopic or radiologic evaluations requiring fasting or have tumor-associated anorexia or nausea. Controversy remains over the clinical benefits of nutritional intervention in cancer patients. In the most extreme case of patients with severe malnutrition, most studies support the use of perioperative total parenteral nutrition (TPN) (4). Aggressive nutritional support is also necessary for patients whose treatment results in prolonged periods (10–14 days) of inadequate nutritional intake (4). The benefits to these patients, including decreased operative morbidity and mortality, exceed the increased risk due to TPN-related infections. Patients with severe malnutrition should receive a minimum of 7 days of nutritional therapy before a major surgical procedure and should continue to receive adequate nutrition via TPN
or feeding jejunostomy, if necessary, as soon as possible after surgery. For all other patients, the risk of TPN-related complications is probably greater than the benefits of decreased recovery time and marginally improved survival. It is clear, however, that maximizing enteral nutrition by early postoperative feeding or nasogastric or jejunostomy feedings when necessary is an effective, safe, and sometimes overlooked way to speed patient recovery.

Table 5.2 Goldman’s cardiac risk index

Risk factor				Points
Jugular venous distension/S₃ gallop				11
Myocardial infarction in the previous 6 mo				10
Any rhythm other than sinus or PACs on preoperative ECG				7
More than 5 PVCs/min on preoperative ECG				7
Age older than 70 yr				5
Emergency procedure				4
Intrathoracic, intraperitoneal, or aortic operation				3
Poor general medical condition (Po₂ <60 mm Hg or Pco₂ >50 mm Hg; K⁺ <3 mEq/L or BUN >50 mg/dL; abnormal SGOT; signs of chronic liver disease)				3
Hemodynamically significant aortic stenosis				3
Class	Point total	Minor complications	Life-threatening complications	Cardiac deaths
I (n = 537)	0–5	532 (99%)	4 (0.7%)	1 (0.2%)
II (n = 316)	6–12	295 (93%)	12 (5%)	5 (2%)
III (n = 130)	13–25	112 (86%)	15 (11%)	3 (2%)
IV (n = 18)	≥26	4 (22%)	4 (22%)	10 (56%)
PAC, premature atrial contraction; ECG, electrocardiogram; PVC, premature ventricular contraction; Po₂, partial pressure of oxygen; Pco₂, partial pressure of carbon dioxide; BUN, blood urea nitrogen; SGOT, serum glutamic-oxaloacetic transaminase.

Staging

Accurate staging enhances the care of gastrointestinal cancer patients and becomes even more important as new therapeutic modalities emerge. Patients known to have localized disease can avoid extensive surgery or toxic adjuvant therapies, and identification of clinically significant micrometastatic disease selects a group of patients who are likely to benefit from adjuvant therapy. The wide range of treatment responses observed within the present staging categories for gastrointestinal cancers suggests that these categories are too large and that, particularly for stage II disease, more specific distinctions are needed. A number of different approaches to improved gastrointestinal cancer staging are under investigation. These include surgical methods, such as staging laparoscopy and sentinel node excision, as well as new techniques for examining tissue using immunohistochemical and molecular markers. A brief description of several promising staging modalities is presented here.

Table 5.3 Reduction of perioperative cardiac risk

Risk	Recommended preoperative cardiac evaluation	Recommended action
Low: Class I and <12 points	None	Proceed with surgery with usual monitoring
Medium: Class II or III, 12–26 points or cannot assess by history	Exercise testing, or if patient is unable to exercise, dipyridamole thallium or stress echo testing	Consider invasive perioperative monitoring
High: Class IV or >26 points	Cardiac catheterization	Coronary artery revascularization, if possible, before surgery

Table 5.4 Preoperative nutritional assessment

	Normal nutritional reserves	Mild to moderate nutritional deficit	Severe malnutrition
Weight	Normal or recent loss of <6% of body weight	Recent loss of 6%–12% of body weight	Recent loss of ≥12% of body weight
Physical examination	Normal	Normal	Muscle wasting, skin, hair, or nail changes
Serum albumin level	≥3.5 g/dL	2.6–3.4 g/dL	≤2.5 g/dL
Serum transferrin level	≥200 mg/dL	151–199 mg/dL	≤150 mg/dL

Staging Laparoscopy

The application of minimally invasive technology to intestinal surgery has decreased staging and treatment morbidity for gastrointestinal cancer patients. The best example of this is avoidance of laparotomy by the use of staging laparoscopy, an approach that is particularly effective for cancers of the gastroesophageal junction, stomach, liver, and pancreas. Because laparoscopy is best used to examine the visible surfaces of the peritoneum and the abdominal organs, the sensitivity of laparoscopy for detection of unresectable disease is significantly increased by the application of endoscopic or laparoscopic ultrasonography.

In a study of 76 patients with upper gastrointestinal malignancies, including lower esophageal, gastric, and pancreatic cancers, laparoscopy alone provided more staging information than conventional imaging in 17 of 39 patients. Fourteen patients were upstaged through detection of peritoneal deposits and liver metastases, and three patients were downstaged and subsequently underwent successful resection. The addition of laparoscopic ultrasonography changed the clinical management of 13 of 37 patients by supplying more detailed information than that obtained with conventional imaging. These additional findings included portal vein invasion in three patients and liver metastases in three patients (5). In a prospective study comparing laparoscopy, ultrasonography, and computed tomography (CT) for staging of gastric cancer, 103 consecutively examined patients with gastric adenocarcinoma were assessed. Laparoscopy with ultrasonography was the most sensitive method for detection of hepatic, nodal, and peritoneal metastases, with an accuracy rate of 99% compared to 76% for ultrasonography and 79% for CT (6). In a study of 114 patients with pancreatic cancer and no evidence of distant disease by CT, laparoscopy confirmed intraabdominal disease extension in 27 (24%) (7). Of the remaining 87 patients without metastatic disease, 42 were shown to have vascular invasion by angiography, a determination that could have also been made by laparoscopic or endoscopic ultrasonography. Forty patients proceeded to laparotomy, and in 30 of these, the tumors were resectable (8).

The gold standard for evaluation of liver tumors is careful intraoperative palpation and intraoperative liver ultrasonography (9). Because patients with resectable hepatic metastases from colorectal cancer can achieve 25% to 30% long-term survival after hepatic resection, detection of small localized liver lesions is important (10). In one study, the combination of laparoscopy with laparoscopic ultrasonography was used to evaluate 15 patients undergoing elective laparotomy for colorectal cancer. Complete inspection and ultrasonography of the liver was possible in 13 patients, and liver metastases were identified in four patients. All patients proceeded to laparotomy with complete palpation and open ultrasonography of the liver, and one additional 0.8-cm lesion was detected; however, this was found to be benign by biopsy (11). In a study of 50 patients with liver tumors, laparoscopy with laparoscopic ultrasonography detected disease that precluded curative resection in 23 patients (46%). These included new lesions not seen by magnetic resonance imaging or CT in 14 patients (28%). Fourteen patients of this cohort were selected for resection, which was accomplished successfully in 13 patients. A historical control group had 58% resectability (12). Laparoscopy with laparoscopic ultrasonography may therefore be most useful in determining unresectability in patients with hepatic tumors without the need for open laparotomy.

Laparoscopy also provides the opportunity to perform peritoneal lavage for cytologic study. The use of peritoneal cytologic analysis for staging of gastrointestinal malignancies, however, is controversial. In patients with pancreatic cancer, the presence of positive peritoneal cytologic results is an ominous sign. In a report of 32 consecutively presenting pancreatic patients with positive results on peritoneal cytologic analysis, only two had disease amenable to resection, and the median survival in those with and without visible intraabdominal metastases was 7.8 and 8.6 months, respectively (13). At Academic Medical Center in Amsterdam, laparoscopic staging with peritoneal cytologic analysis was performed on 449 patients from 1992 to 1997. The indication was a variety of gastrointestinal tumors, including 87 esophageal cancers, 72 proximal bile duct tumors, 236 periampullary tumors, 17 tumors of the pancreatic body or tail, and 7 primary and 32 metastatic liver tumors. Lavage changed the assessment of stage and accurately predicted unresectable disease in only 6 of 449 patients (1.3%). In addition, of the 28 patients with positive lavage findings, 19 (68%) also had metastatic disease identifiable at laparoscopy, and three had false-positive results on peritoneal cytologic analysis and were found to have resectable disease on exploration. The authors conclude that the technique is not effective (14).

In summary, therefore, staging laparoscopy is best used to document unresectability in patients who would otherwise require laparotomy to reach this conclusion. This technique is most often useful for upper gastrointestinal malignancies that tend to exhibit peritoneal spread missed by noninvasive imaging techniques.

Significance of Micrometastatic Disease

One population for which staging could be improved includes patients with node-negative (N0) disease by conventional histopathology. For example, in patients undergoing potentially curative surgery for colon cancer, 35% to 45% will
have N0 nodal status (15). Approximately 25% of these stage II patients will experience posttreatment disease recurrence, suggesting that for one in four patients with N0 disease, current histopathological staging methods fail to identify those destined to manifest tumor progression. The application of better techniques to detect these high-risk cases would identify patients who may benefit from adjuvant chemotherapy. These patients are also an important population for evaluating new treatments to target minimal residual disease.

“Micrometastasis” is a term used to describe evidence of tumor cell spread beyond a primary tumor that does not meet histologic criteria for N1 or M1 disease. Examples include tumor deposits in regional lymph nodes of <200 μm in diameter, or single cells identified only following immunohistochemical stains for tumor-associated proteins such as cytokeratins or carcinoembryonic antigen. For solid tissue malignancies, the clinical significance of tumor cells discovered in lymph nodes, bone marrow, or circulating blood is unclear. Tumors readily shed individual cells into the circulation, and even individual tumor cells can be detected in tissue samples by immunohistochemical stains or polymerase chain reaction–based methods. The clinical importance of circulating tumor cells resides in their ability to lodge in host tissues and form an independent metastatic colony. Such micrometastases likely represent a small fraction of the cells shed from a tumor. To date, numerous studies have failed to correlate the presence of small numbers of tumor cells in lymph nodes, bone marrow, or circulating blood with clinical behavior of tumors (16,17,18,19). This distinction is obviously most important for tumors that are responsive to adjuvant therapies, such as colorectal or gastric cancers.

At present, the most important obstacles to understanding the clinical significance of micrometastatic disease are technical in nature. Techniques for detecting micrometastatic disease are difficult to standardize due to differences in antibodies, staining techniques, and scoring systems. The presence of tumor cells in the circulation and in lymphatics may also be affected by surgical manipulation of a tumor during resection (20,21). Perhaps for these reasons, most of the available retrospective studies failed to find prognostic significance for metastases detected by immunohistochemistry only (16,17,18,19).

Although micrometastatic disease has yet to be assigned clinical significance in gastrointestinal malignancies, research designed to investigate this issue is proceeding in parallel with work to improve early disease detection. In addition to benefit achieved by improvements in clinical staging, understanding the differences in character between micrometastatic disease and nonmetastasizing precursors would provide important targets for improved diagnostic and therapeutic agents. In the future, it is possible that imaging techniques, such as preoperative magnetic resonance or intravenously administered near-infrared fluorescence probes, will accurately detect small volume metastases from gastrointestinal solid tumors (22).

Sentinel Nodes in Gastrointestinal Cancer

Sentinel lymph node sampling (SLNS) identifies a small number of regional lymph nodes that accurately predict the status of all regional nodes in a patient undergoing cancer surgery. SLNS is based on the assumption that lymphatic flow drains sequentially from peripheral to central tissue locations, with limited functional collaterals outside the dominant vascular supply. Consequently, tracer substances injected at a tumor site must follow the same pathway by which metastatic tumor cells traverse lymphatic channels. If these conditions are met, the first lymph node encountered, termed the sentinel node, is a reliable indicator of the tumor status of the entire nodal basin. To a high degree, these principles hold true for the integumentary system because SNLS is a clinically validated indicator of nodal status for both breast cancer and melanoma (23,24).

SLNS has been evaluated for colon, gastric, and esophageal malignancies. The staging issues addressed by SLNS differ significantly among these three diseases. For esophageal and gastric cancer, assessment of nodal status prior to resection of the primary tumor can aid surgical decision making. For example, by identifying patients with stage III disease, SLNS is a minimally invasive method to select patients with advanced esophageal or gastric cancer that may benefit from preoperative chemoradiation. At the other end of the disease spectrum, studies are investigating the efficacy of minimally invasive surgery for resection of superficial gastric cancer. This procedure requires confirmation of node-negative status prior to definitive surgery, and single-institution studies indicate that the SLNS technique may achieve this goal (25). The development of minimally invasive surgery for early gastric cancer is therefore proceeding in close association with studies to confirm the accuracy of SLNS for this disease.

Unlike surgery for breast cancer or melanoma, the anatomy of the colon permits wide lymphadenectomy without significantly increasing the difficulty or morbidity of tumor resection. SLNS for colon cancer would be useful if it is therapeutic or if it improved the accuracy of pathological staging (26,27,28). Although proponents of SNLS for colon cancer report that this technique is highly accurate, with values ranging from 89% to 97% for the detection of node-positive disease (27,29,30,31), most of the studies that reached this conclusion applied different diagnostic criteria to the sentinel node than those used for nonsentinel nodes. For example, it was common for sentinel nodes but not nonsentinel nodes to be scored positive based on immunohistochemical results rather than standard histopathological criteria. In a recent prospective study performed by the Cancer and Leukemia Group B, sentinel nodes were sampled in 66 patients with respectable colon cancer (32). Examination of these nodes using standard histopathology failed to predict the presence of nodal disease in 13/24 (54%) of node-positive cases. Immunostains were then performed on both sentinel and nonsentinel nodes for cases whose lymph nodes were negative by standard histopathology. Depending on the immunohistochemical criteria used to assign lymph node positivity, sentinel node exam resulted in either an unacceptably high false-positive rate (20%), or a low sensitivity for detection of micrometastatic disease (40%). By examining both sentinel and nonsentinel nodes, this multiinstitutional study showed that sentinel nodes did not accurately predict the presence of either conventionally defined nodal metastases or micrometastatic disease (32). At the present time, therefore, SLNS is not used for staging of resectable colon cancer.

Radioimmunoguided Surgery

Radioimmunoguided surgery (RIGS) is a technique for intraoperative localization of tumor in tissues that would not ordinarily be removed as part of a standard cancer resection. This is done so additional tumor sites can be treated by extended resection, intraoperative radiation therapy, or adjuvant postoperative chemotherapy or radiation therapy. A radiolabeled monoclonal antibody (MAb) is injected into the patient 3 to 4 weeks before surgery. The antibody used most often for defining the extent of gastrointestinal cancers by RIGS is CC49, an MAb against the epithelial tumor antigen, tumor-associated glycoprotein-72 (TAG-72). This MAb is generally labeled with iodine 125 and therefore must be administered with a thyroid-blocking agent such as potassium iodide. The MAb is cleared from the bloodstream and concentrated in the tumor, and the patient is taken to surgery when a 2-second measurement of the precordial area with a gamma probe yields counts of <30 (33). During exploratory laparotomy, in addition to the usual
inspection and palpation, the abdomen is scanned with a handheld gamma probe, taking as background the count corresponding to that of circulating blood, obtained at the bifurcation of the aorta. The local tumor field and associated nodal basins are then scanned for increased signal, which indicates residual disease. Complete survey of the abdomen with the gamma probe also includes the liver, stomach, duodenum, posterior retroperitoneum including kidneys, and pelvis. In some studies of recurrent colorectal cancer, use of this technique altered surgical decisions as much as 30% of the time, generally by extending the resection to adjacent tissues to obtain a RIGS-negative field. Problems with the technique include low specificity in lymph node tissue, possibly due to clearance of the antibody by the reticuloendothelial system.

RIGS findings may have prognostic value. In a multicenter phase III trial for recurrent or metastatic intraabdominal cancer, surgical decision making was altered by RIGS analysis 20% of the time (34). In the patients with liver metastasis, RIGS identified occult metastasis in the periportal lymph nodes in 28.5% and identified those patients who were not likely to be cured by liver resection (35). Studies of the use of RIGS have not yet demonstrated an impact on patient treatment morbidity or survival, and this technique is costly and fairly cumbersome. Further study is required to determine both how accurate RIGS staging is and whether this method can be applied effectively in routine clinical care.

Molecular Characterization of Tumors

Research since the mid-1990s has provided an improved understanding of the molecular nature of gastrointestinal carcinogenesis. These observations are beginning to be translated into clinically useful predictors of tumor behavior. Tumor-associated genotypic or phenotypic markers add to conventional assessment of histologic grade or degree of invasion by measuring cell cycle control, angiogenic potential, or genomic stability, to name a few functional categories by which various markers are classified (Table 5.5).

One example of how a tumor-associated marker can provide useful clinical information is provided by thymidylate synthase (TS) measurement. TS is an enzyme required for DNA synthesis, as it converts 2′-deoxyuridine 5′-monophosphate to thymidine 5′-monophosphate. TS is a critical target for 5-fluorouracil (5-FU), the most common chemotherapeutic agent for gastrointestinal malignancies. DNA synthesis is inhibited by 5-FU through formation of a ternary complex between TS, the 5-FU metabolite fluorodeoxyuridylate, and CH₂FH₄, a folic acid derivative (36,37). Overexpression of TS protein in colorectal and gastric tumors is associated with resistance to 5-FU and may also indicate loss of p53 function (38,39,40). Increased expression of TS in tumors predicts a poor response to chemotherapy regimens using 5-FU in multiple clinical studies (41,42).

Table 5.5 Putative markers of prognosis or treatment response

Carcinoembryonic antigen	PCNA	p53
Thymidylate synthase	Ki67	p21
Thymidylate phosphorylase	Cyclin D1	p27
Matrix metalloproteinases	Ploidy	Myc
Cathepsin D	Apoptotic index	Bcl-2/Bax
Sialyl Lewis A/sialyl Lewis X	Microvascular density	17pLOH
CD44 v6, v8–10	VEGF	18qLOH
Plasminogen activator	Sucrase-isomaltase	DCC
uPA receptor	Prolactin receptor	Ki-ras
HER-2/neu	Vitamin D receptor	Microsatellite instability
PCNA, proliferating cell nuclear antigen; LOH, loss of heterozygosity; VEGF, vascular endothelial growth factor; DCC, deleted in colon cancer; uPA, urokinase plasminogen activator; HER-2, human epidermal growth receptor 2.

Most reports linking genotypic or phenotypic characteristics of tumors to clinical outcome are single-marker studies, performed either retrospectively or prospectively with small numbers of patients. Although some of these markers, like TS, appear to have important independent prognostic value, their clinical utility is still in question because of variability in laboratory methods, differences in treatment within the study cohort, and insufficient clinical follow-up. At the present time, however, several promising clinical trials are underway within the cancer cooperative groups to determine the association between multiple tumor markers and clinical outcome for colorectal, gastric, and pancreatic malignancies.

Advancing Surgical Techniques

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