24 Rectal Adenocarcinoma
Abstract
Modern care of the patient with rectal carcinoma is multidisciplinary and requires close cooperation of surgery, medical and radiation oncology, radiology, and pathology specialties. Accurate locoregional staging will influence the decision to administer neoadjuvant therapies. Preoperative chemoradiation will decrease the local recurrence rates for locally advanced rectal cancer by 50%. High-quality surgery according to total mesorectal excision principles remains essential for curative resection, as is specialized pathologic evaluation of the surgical specimen. Newer organ-preserving modalities further contribute to the treatment armamentarium. Adjuvant therapy is controversial but remains the standard of care. Patients should further be followed after definitive management according to established surveillance protocols to identify local and distant recurrences. Optimization of outcomes for patients with rectal carcinoma is dependent on adherence to these principles of high-quality evidence-based care.
24.1 Introduction
Cancers arising in the rectum (i.e., the distal 15 cm of the large bowel) share many of the same genetic and pathologic characteristics of cancer originating in the more proximal colon, but the anatomic constraints of the bony pelvis along with the proximity to other urogenital structures, anal sphincters, and autonomic nerves create significant issues for surgical access. As a result, rectal cancer outcomes have been historically worse than those of colon cancers. Rectal cancer surgery has evolved from perineal (Jacques Lisfranc) 1 and posterior (Paul Kraske) 2 approaches in the 1800s to the introduction of the abdominoperineal excision by Sir Ernest Miles in 1908 3 and the sphincter-preserving anterior resection by Dixon in 1948. 4 However, local recurrence rates and survival remained poor until the introduction of the total mesorectal excision (TME) concept, which emphasizes dissection along the avascular “holy plane” and en bloc removal of the rectum and its encompassing mesorectum, by R.J. Heald in 1982. 5 Local recurrence rates further improved when the benefits of neoadjuvant (chemo)radiation were demonstrated by the German Rectal Cancer Study 6 and the Swedish Rectal Cancer Trial. 7 The introduction of minimally invasive techniques and organ preservation has further added to the treatment armamentarium. The management of rectal cancer has become truly multidisciplinary with optimal management dependent on medical and radiation oncologists, radiologists, pathologists, and surgeons.
24.2 Epidemiology and Presentation
Colorectal cancer is common, with an estimated 1.4 million new cases worldwide including 693,900 deaths. 8 Incidence rates are highest in North America, Europe, Oceania, and Japan, but are increasing in many countries in which rates are historically low. 9 The reasons for this increase in countries in Latin America, Asia, and Eastern Europe are thought to be changing dietary and activity patterns, and increase in smoking prevalence. 10 , 11 , 12 In the United States, it is estimated that 134,500 new colorectal cancer cases are diagnosed annually, including 39,220 rectal cancers. 13 While still among the highest in the world, the incidence rate in the United States and Western Europe is starting to decrease, likely as a result of screening and removal of precancerous lesions. 14 , 15 However, the incidence of colorectal cancer in patients younger than 50 years is increasing at a rate of 2.1% per year from 1992 to 2012, for whom average-risk screening is not recommended. Young-onset colorectal cancer is more likely to be left sided or rectal, poorly differentiated, have mucinous or signet-ring cell histology, and present at advanced stages. 16 , 17 The large majority of these patients were symptomatic with rectal bleeding (50–60%), abdominal pain (30–60%), or change in bowel habits (20–70%), 16 , 17 , 18 which likely prompted further evaluation given that screening of asymptomatic average-risk individuals in this age group is not standard of care. However, stage-specific survival in young-onset colorectal cancer is similar to cancers occurring in older individuals. 19 The molecular genetics of colorectal cancer are covered in detail elsewhere and will not be discussed in this chapter.
Patients with colorectal cancer can present in three ways: asymptomatic individuals in whom the tumor is detected on routine screening; evaluation after worrisome symptoms; or emergently with perforation, obstruction, or life-threatening bleeding. However, 80 to 90% of colorectal cancers are detected after evaluation for symptoms—most commonly rectal bleeding, abdominal pain, and change in bowel habits. 20 , 21 In contrast, emergent presentation occurs in approximately 5 to 10% of patients, usually as a result of obstruction. 22 Despite the increase in screening programs, 20 to 30% of patients with newly diagnosed colorectal cancer will not have previously undergone a screening colonoscopy. 20 , 21 , 22 Furthermore, approximately 20% of patients will have stage IV disease at presentation, most commonly to the liver, lungs, or peritoneum as a result of hematogenous spread. The liver is usually the initial site of metastasis due to the portal venous drainage of the large bowel. However, tumors in the distal rectum often develop pulmonary metastases first because the inferior rectal vein drains directly into the vena cava instead of the portal vein.
24.3 Preoperative Evaluation
Once the diagnosis of rectal cancer is made, the patient must undergo complete evaluation prior to initiation of any treatment (▶ Table 24.1). 23 A comprehensive history and physical examination should be performed to assess for associated signs and symptoms. The history should include any current comorbidities that might affect perioperative management, previous surgical history, and complete family history. Concerning cancer-specific history may include significant weight loss, extensive bleeding per rectum, tenesmus, pain with defecation, and obstructive symptoms. The presence of any of these symptoms should alert the clinician to more advanced disease. In particular, tenesmus, the sensation of incomplete evacuation or constant feeling of needing to evacuate, is suggestive of a large and potentially fixed tumor. Pain with defecation may suggest lower tumors with possible sphincter involvement, as rectal cancer above the anorectal junction is typically painless unless there is invasion of surrounding structures. Pretreatment history should also include sphincter and urinary function, as well as sexual function in males, as these may be affected by neoadjuvant and/or surgical therapy. Perioperative cardiac risk assessment should be performed according to the American College of Cardiology/American Heart Association guidelines. 24 The patient should also be specifically questioned about functional capacity by determining if they can climb two flights of stairs without shortness of breath. This activity represents a functional capacity of at least four metabolic equivalents (METs). Physical examination should assess for any previous abdominal incisions that might affect potential surgical approach. A digital rectal examination (DRE) must be performed to assess for sphincter involvement. Palpable low rectal tumors should also be evaluated if they are mobile and soft, or hard and fixed. For lesions that are not palpable, the surgeon should also perform a rigid proctoscopy to determine the distal extent of the tumor. Patients should undergo a full colonoscopy if not already done to rule out synchronous lesions, which are found in approximately 3 to 5% of patients from population-based studies. 25 , 26 A carcinoembryonic antigen (CEA) level should be obtained for every patient before surgery. It must be noted that the purpose of the CEA is not for screening or diagnosis. A meta-analysis of the diagnostic characteristics of CEA for colorectal cancer reported a pooled sensitivity of 46% (95% confidence interval [CI]: 0.45–0.47) and specificity of 89% (95% CI: 0.88–0.92). 27 , 28 Other etiologies for an elevated CEA include gastritis, peptic ulcer disease, diverticulitis, liver disease, chronic obstructive pulmonary disease, diabetes, and other acute inflammatory states. CEA may also be falsely elevated in smokers. 29 Several surgical and medical oncology societies have recommended against the use of tumor markers, including CEA, as a screening or diagnostic test. 30 , 31 CEA does have utility as a prognostic tool, as well as for surgical planning and posttreatment surveillance, and therefore a baseline measurement should be obtained prior to any treatment. Patients with preoperative serum CEA ≥ 5 ng/mL have worse prognosis, stage for stage, than patients with serum CEA < 5 ng/mL. 31 , 32 , 33 , 34 Furthermore, normalization of CEA in patients with an elevated serum CEA who are undergoing neoadjuvant therapy is a strong predictor of complete pathologic response. 35 , 36 , 37 Elevated CEA levels that do not normalize after curative resection should raise the suspicion of residual disease and should prompt further evaluation. Likewise, patients with rising CEA levels after curative resection should undergo systemic imaging to rule out local and/or distant recurrence. Routine laboratory tests other than a CEA are not necessary. Transaminases are neither sensitive nor specific for liver metastases.
Baseline |
|
Imaging |
|
Preoperative planning |
|
Staging is performed according to the American Joint Commission on Cancer 7th edition TNM classification (▶ Table 24.2 and ▶ Table 24.3). Computed tomography (CT) scans of the chest, abdomen, and pelvis are required to rule out concurrent distant metastases. CT scans of the abdomen and pelvis may show tumor-related complications such as perforation, obstruction, or clear invasion into surrounding organs, or distant metastases or lymphatic involvement. It must be noted that CT scan has limited sensitivity for low-volume lesions (especially under 0.5–1.0 cm). According to one study, only 11% of nodules less than 0.5 cm were seen on CT scan, and only 37% of lesions between 0.5 and 1.0 cm were seen. 38 , 39 An extremely elevated CEA in the absence of metastatic disease apparent on CT imaging should alert the clinician about the possibility of peritoneal carcinomatosis. In these cases, a positron emission tomography (PET) scan may be useful. 27 , 28 Even if the PET scan is normal, visual and/or manual inspection of the peritoneal cavity at the beginning of a curative-intent procedure is imperative to rule out synchronous metastatic disease prior to undergoing proctectomy. Routine CT imaging of the chest is controversial, as the incidence of indeterminate lesions is high, but these nodules are often benign; however, numerous groups include this as the optimal imaging choice; for example, in the United States both National Comprehensive Cancer Network (NCCN) and the American College of Surgeons Commission on Cancer (ACS CoC) National Accreditation program for Rectal Cancer (NAPRC) recommend CT rather than plain chest X-ray.
Stage | T | N | M |
0 | Tis | N0 | M0 |
I | T1 T2 | N0 N0 | M0 M0 |
IIa | T3 | N0 | M0 |
IIb | T4a | N0 | M0 |
IIc | T4b | N0 | M0 |
IIIa | T1–T2 T1 | N1/N1c N2a | M0 M0 |
IIIb | T3–T4a T2–T3 T1–T2 | N1/N1c N2a N2b | M0 M0 M0 |
IIIc | T4a T3–T4a T4b | N2a N2b N1-N2 | M0 M0 M0 |
IVa | Any T | Any N | M1a |
IVb | Any T | Any N | M1b |
Source: © 2010 American Joint Committee on Cancer | |||
In theory, venous drainage of lower rectal cancers directly enters into the vena cava circulation, and bypassing the liver, through the hemorrhoidal veins. In one systematic review of 12 studies including 5,873 patients, 9% had indeterminate pulmonary nodules on preoperative chest CT, but only 11% of these lesions (~1% of the total population) declared themselves to be colorectal metastases during surveillance. 40 Given the low incidence of malignancy in these indeterminate pulmonary nodules, further preoperative evaluation can be avoided, and these lesions can be followed during surveillance. If equivocal suspicious findings on CT scan are found, the patient should undergo more specific testing, such as liver magnetic resonance imaging (MRI). The additional imaging techniques should be not routinely performed in the absence of other findings on CT scan as they are not generally useful and add unnecessary costs and burden for the patient.
A PET scan need not be routinely ordered if the patient has no evidence of metastatic disease on staging CT scans. A PET scan may be useful in the setting of resectable liver metastases to detect any extrahepatic disease that may render the patient not a candidate for curative resection, or in the situations with an elevated CEA but no metastatic disease on CT imaging.
24.3.1 Local Staging
The two imaging modalities that are the most useful for local staging of rectal adenocarcinoma are transrectal ultrasound (TRUS) and MRI. Both modalities are more accurate than CT for assessing the depth of tumor invasion, nodal staging, and assessment of the circumferential resection margin (CRM). CT of the pelvis can define tumor invasion into surrounding structures for larger tumors, but does not adequately assess depth of invasion for smaller tumors. 41 , 42 , 43 , 44 CT also does not accurately diagnose malignant perirectal adenopathy, with a sensitivity of only 49%. 45 The utility of preoperative CT is to detect tumor-related complications such as perforation, malignant fistula, or obstruction.
Transrectal Ultrasound
TRUS can differentiate between the specific layers of the rectal wall and other surrounding anatomic structures, especially those situated anteriorly, based on their different acoustic impedances (▶ Fig. 24.1). The 10-mHz crystal is most commonly used for rectal cancer staging, and has a resolution of 0.4 mm and focal range of 1.5 to 4 cm. The anal sphincter complex is especially well visualized on TRUS, but the distal most length of rectum at the anorectal junction is difficult to assess by TRUS. The optimal image requires that the transducer probe be situated in the middle of the rectum with good opposition of the rectal wall with the water-filled balloon. Tumors appear as extensions of the first hypoechoic wall. Ultrasound T-staging is based on tumor disruption of the different echographic layers of the rectal wall. The entire length of the tumor needs to be imaged as the depth of invasion is not uniform. This may cause the patient significant discomfort for tumors that are large, bulky, or near obstructing. Metastatic lymph nodes appear as hypoechoic deposits usually at the level of or proximal to the tumor, are round rather than oval, and have ill-defined borders. However, small tumor deposits may not sufficiently alter the echogenic profile of the lymph node to be detected by TRUS. As such, TRUS is technically difficult and operator dependent.
Prior to the widespread use of MRI, TRUS was considered the gold standard for locoregional staging of rectal cancer. Early studies found high diagnostic accuracy for TRUS. In a systematic review of studies published before 1993, the overall accuracy of endorectal ultrasound (ERUS) for T-staging was 84%, with 97% sensitivity and 87% specificity. 46 In particular, ERUS was highly accurate in differentiating between T1/T2 and T3 and greater, which had important management implications as T3 and above tumors should undergo neoadjuvant chemotherapy, whereas T1/T2 tumors can undergo upfront resection (in the absence of nodal metastases). However, later studies have shown TRUS to be less accurate than previously reported. In one of the largest studies on the use of TRUS for locoregional rectal cancer staging, 7,096 patients underwent TRUS followed by surgical resection without neoadjuvant therapy. 47 The overall concordance between clinical stage as reported by TRUS and final pathologic staging was only 64.7% (95% CI: 63.6–65.8). Tumors were overstaged in 17.3%, and understaged in 18%. The accuracy of TRUS varied significantly based on T-stage, with highest clinical–pathological concordance for TRUS-staged T1 tumors (kappa = 0.591) and lowest for T4 tumors (kappa = 0.321). Hospitals that performed over 30 TRUS per year had the highest diagnostic accuracy (73.1%; 95% CI: 69.4–76.5), whereas hospitals with less than 10 TRUS per year reported a diagnostic accuracy of only 63.2% (95% CI: 61.5–64.9). Another study of 545 patients reported the overall accuracy of TRUS for T-staging to be 69%, and 64% for N-staging. 48 Tumors were more often overstaged (18%) than understaged (13%). Other more recent studies report that TRUS is less reliable in differentiating between early T3 with minimal extramural spread and more advanced T3 tumors, as well as differentiating between T2 and early T3. 48 , 49 This disadvantage is particularly relevant as it may over- or undertreat a significant proportion of patients with rectal cancer under evaluation for neoadjuvant therapies. These results suggest that TRUS may be an accurate diagnostic modality to diagnose earlier stage tumors if local expertise is available.
The diagnostic accuracy of TRUS for nodal staging (70–75%) is similar to that of CT (55–65%) and MRI (60–65%). 50 , 51 A meta-analysis of 35 studies including 2,732 patients reported the pooled sensitivity and specificity of TRUS for nodal staging to be 73.2% (95% CI: 70.6–75.6) and 75.8% (95% CI: 73.5–78.0), respectively. 52 The additional value of fine-needle aspiration (FNA) to TRUS for the evaluation of nodal status is equivocal. 53 Perirectal nodes are typically too small to be visualized by TRUS unless they contain metastatic deposits, at which point they can be detected by visual morphology alone and do not require FNA biopsy confirmation. However, assessment of lymph nodes outside of the focal range of the TRUS probe, such as the lateral pelvic nodes, is lacking.
TRUS is also limited for the assessment of the mesorectal fascia, which has therapeutic and prognostic importance. Anteriorly, TRUS can detect CRM involvement at the level of the seminal vesicles, prostate, and vagina based on the close proximity of these structures to the rectal wall. Posteriorly, TRUS cannot assess the CRM based on the lack of adjacent structures. In these cases, MRI offers a clear advantage over TRUS.
Magnetic Resonance Imaging
Pelvic or rectal MRI has replaced EUS as the local staging modality. MRI offers excellent resolution and assessment of the tissues and anatomic structures of the pelvis and has replaced ERUS as the imaging modality of choice for most expert groups and authorities (▶ Fig. 24.2). Its specific advantage over other imaging options includes the ability to predict involvement of the CRM and to visualize extramural vascular invasion (EMVI).
Rectal MRI is typically performed on a 1.5- or 3-T MRI scanner. Endorectal surface coils can be used, but are very uncomfortable for the patient, and the development of high-resolution surface coils has also led to the reduction in their use. Specific rectal cancer protocols with multiplanar T2-weighted images including axial, coronal, sagittal, oblique axial, and oblique coronal view with or without multiaxial diffusion-weighted imaging should be used. The examination is performed with the patient supine. The rectum should be cleansed with enemas to decrease image misinterpretation caused by stool. Several different rectal solutions can be administered to provide a negative contrast against the rectal wall. The radiologist should be provided with tumor-specific information such as the estimated height and dimensions of the tumor, as well as any previous pelvic surgeries.
MRI may have difficulty in differentiating between the mucosa and the submucosa, as they appear as a single inner hyperintense layer on T2-weighted imaging. The muscularis propria is an intermediate hypointense layer, and the mesorectal fat appears hyperintense. The mesorectal fascia can also be clearly seen as a thin hypointense layer surrounding the mesorectal fat. Rectal tumors appear as intermediate signal intensity lesions between the high signal intensity of fat and low intensity of muscle on T2-weighted imaging and demonstrate enhancement after contrast administration. MRI has difficulty in distinguishing between early T1/T2 tumors and they are often combined together as tumors confined to the bowel wall. Invasion into the perirectal fat can be clearly seen on thin section MRIs and is crucial in differentiating between T2 and T3 tumors. The extent of extramural spread measured by MRI is highly accurate when measured against final pathology, and can be used to determine the need for neoadjuvant therapies for select T3 tumors. 54 Metastatic nodes appear as nodular structures in the mesorectum at the level of or slightly proximal to the tumor. They can also be seen along the internal iliac or superior rectal vessels. Nodes smaller than 3 mm are not well visualized on MRI, but can still harbor malignancy. 55 Lymph nodes are considered metastatic if they are larger than 5 mm in shortest axis, those with spiculated or irregular borders, or have a mottled appearance. 56 , 57 Enlarged nodes in the mesorectum in the context of rectal cancer are usually considered malignant, but nodes can be reactive to rectal manipulation such as biopsy and tattooing during endoscopy. Administration of newer contrast agents 58 , 59 or diffusion weighted imaging 60 can also be used to diagnose metastatic nodes, but these techniques are still considered experimental and are not yet widely used. MRI can also detect extramural vascular invasion, which appears as tumor deposits spreading directly off the tumor along lateral rectal vessels, which also has prognostic implications. 61 , 62
The accuracy of MRI for locoregional staging has been addressed in a meta-analysis of 21 studies that used histopathologic analysis as the reference. 63 In this study, the sensitivity and specificity of MRI to detect T3/T4 over T1/T2 was 87% (95% CI: 81–92) and 75% (95% CI: 65–80), nodal metastases 77% (95% CI: 69–84) and 71% (95% CI: 59–81), and involved CRM 77% (95% CI: 57–90) and 94% (95% CI: 88–97), respectively. However, the authors noted that there was significant heterogeneity among the studies in terms of the definition of a positive lymph node or an involved CRM.
The role of MRI in the locoregional staging of rectal cancer has evolved significantly over the past decade. The Magnetic Resonance Imaging in Rectal Cancer European Equivalence (MERCURY) study group has contributed important data that have shifted the modality of choice from TRUS to MRI. There are several major advantages of MRI over TRUS that make it the preferred imaging modality for the locoregional staging of rectal cancer, especially when neoadjuvant therapy is under consideration. While sensitivity and specificity of the two modalities for T- and nodal-staging are roughly equal, MRI can accurately predict involvement of the CRM as well as the depth of extramural invasion for cT3 tumors. These two findings will subsequently dictate further neoadjuvant management. Improvements in technology will allow for better resolution and enhanced accuracy of MRI staging.
Several studies have shown that depth of extramural invasion has important prognostic implications. Patients with invasion greater than 5 mm have worse cancer-specific survival compared to patients with less than 5 mm invasion (54 vs. 85% at 5 years; p < 0.001). 64 , 65 Increased depth of invasion is also associated with an increased risk of nodal metastases. 66 These findings suggest that the importance of distinguishing between T1/T2 and T3 tumors is less essential than identifying T3 tumors with greater than 5-mm invasion into the perirectal fat. The MERCURY study group showed that thin-slice MRI can reliably identify the depth of invasion for patients with T3 tumors within 0.5 mm of the final histopathological result. 54 MRI can also identify patients with high risk for positive CRM in which a part of the tumor of a metastatic deposit or lymph node is less than 1 mm from mesorectal fascia before neoadjuvant therapy 67 after surgical resection with 94% accuracy. 68 MRI-involved CRM also predicted local recurrence, disease-free, and overall survival. 69 The utility of MRI assessment of extramural depth of invasion, CRM status, and extramural vascular invasion was demonstrated in a study in which patients were divided into “good prognosis” and “bad prognosis.” Patients with “good prognosis” tumors had T1/T2 or T3 disease with safe CRM (tumor > 1 mm from mesorectal fascia), no evidence of extramural vascular invasion, and extramural depth of invasion less than 5 mm on preoperative MRI. None of these patients received pre- or postoperative radiotherapy and all underwent TME proctectomy. Local recurrence rate for these patients was 3% with 85% disease-free and 68% overall survival at 5 years. These findings suggest that MRI can be used to detect tumors that would traditionally undergo neoadjuvant chemoradiation that could undergo upfront resection with good outcomes. However, this approach is not yet standard of care.
Imaging after Neoadjuvant Chemoradiation
Certain patients may benefit from restaging after completion of neoadjuvant chemoradiation. MRI and TRUS can be used to assess response to neoadjuvant therapy, but their accuracy is poor as both have difficulty in differentiating between residual tumor and radiation-induced edema, inflammation, or fibrosis. Restaging will be discussed later in this chapter. A systematic review of 30 studies reporting on the accuracy of restaging MRI and TRUS after neoadjuvant chemoradiation published up to 2011 demonstrated that both modalities have poor accuracy for T-stage (65%, 95% CI: 56–72 for TRUS; 52%, 95% CI: 44–59% for MRI) and N-stage (both modalities had 72% accuracy). 70 MRI was able to accurately predict CRM involvement after neoadjuvant therapy. Given these data, restaging with MRI and TRUS remains difficult, and neither of these studies should be used as the sole method to assess tumor (complete) response. Newer MRI technologies such as diffusion-weighted imaging or dynamic contrast-enhanced MRI may potentially increase the ability to accurately assess tumor response by directly measuring local microcirculation and cellular environment. 71 These techniques are still limited to large specialty centers and have not been disseminated into smaller practices. PET scan may also be able to detect pathologic response, although the data are conflicting. 72 , 73
24.3.2 Multidisciplinary Tumor Board
All patients with rectal cancer should be discussed upon presentation to the institution at a multidisciplinary tumor board conference. 74 These conferences should consist of a panel of surgeons, radiologists, pathologists, medical oncologists, and radiation oncologists that are specialized in the care of patients with rectal cancer. Management decisions to proceed with neoadjuvant therapy, upfront surgery, or other strategies are decided on using a consensus multidisciplinary team approach. Previous studies have demonstrated that the multidisciplinary approach improves decision making, clinical, outcomes, and patient experience for patients with a variety of cancers. 75 , 76 , 77 , 78 Specifically for rectal cancer, the implementation of multidisciplinary tumor boards was associated with lower incidences of permanent stoma and local recurrence, improved delivery of evidence-based care, and, most importantly, better overall survival. 79 , 80 The importance of the multidisciplinary tumor board in the management of patients with rectal cancer is such that it is included among the core components of the American College of Surgeons (ACS) Commission on Cancer (CoC) National Accreditation Program for Rectal Cancer (NAPRC). 81 , 82
It is the goal of the NAPRC to provide standards for multidisciplinary management to improve the quality and delivery of care for patients with rectal adenocarcinoma. 81 Outcomes in the United States are highly variable. While it is difficult to study variability in more exact measures of outcome such as local recurrence rates and survival, colostomy rates are felt to be an appropriate metric of decreased surgical quality in rectal cancer. 83 As of 2007, 60% of patients undergoing proctectomy for rectal cancer received a permanent colostomy. 84 Two studies by Ricciardi et al 84 , 85 demonstrate that rates of permanent colostomy are highly variable and excessive. County-level data from 21 U.S. states revealed that 40% of surgeons performed only abdominoperineal resection (APR) for their patients with rectal cancer, 85 while a study of greater than 7,500 proctectomies from 11 states found that half of all patients received a permanent colostomy. 86 A recent study of over 47,000 patients using data from the Nationwide Inpatient Sample found that colostomy rates had decreased in the United States from 65% in 1988 to 40% in 2006. However, there was marked variability between high- and low-volume hospitals (colostomy rates 44 vs. 57%, respectively) and even the lowest colostomy rate reported was still far in excess of what would be considered acceptable in European countries where quality improvement programs have been successfully implemented. 83
This variability between high- and low-volume centers also extends to delivery of neoadjuvant care, as well as other surgical quality measures. Analysis of the National Cancer Database (NCDB) demonstrated that a higher proportion of rectal cancer patients received evidence-based neoadjuvant therapy in high-volume (more than 30 rectal resections per year) compared to low-volume centers. 87 The inverse relationship was also true, in that patients were more likely to receive adjuvant chemoradiotherapy at low-volume centers. Hodgson et al 88 also reported that low-volume hospitals had higher permanent colostomy and 30-day mortality and lower 2-year survival compared to hospitals in the higher quartiles of volume, using the California Cancer Registry. These and other studies were included in a Cochrane meta-analysis, which demonstrated that higher hospital volume was significantly associated with improved 5-year survival. 89
24.4 Neoadjuvant Therapy
24.4.1 Indications
The use of neoadjuvant therapy has led to a significant reduction in local recurrence rates for patients with resectable adenocarcinoma of the rectum. Traditionally, patients underwent upfront surgical resection, followed by chemoradiotherapy in the adjuvant setting for patients with advanced tumors, positive margins, and/or node-positive disease. The treatment paradigm has drastically changed since the late 1990s and early 2000s when several seminal trials demonstrated significantly improved outcomes for preoperative treatment. At present, the current standard of care for neoadjuvant therapy includes long-course chemoradiation, usually with 50.4 Gy delivered in 28 fractions with concurrent fluoropyrimidine-based radio-sensitizing chemotherapy and surgery after 6 to 8 weeks, or short-course preoperative radiotherapy delivered as 25 Gy in 5 fractions over 1 week, followed by surgery within a week of completion of therapy. The current accepted indications for neoadjuvant chemoradiotherapy are provided in ▶ Table 24.4.
24.4.2 Overview of Neoadjuvant Strategies
The best regimen for neoadjuvant therapy has not been determined, with significant variability internationally. In the United States, long-course chemoradiation is the favored approach for any patient with an indication for neoadjuvant therapy, whereas in Europe, short-course radiotherapy is preferred. What is clear from the data is that both neoadjuvant long-course chemoradiation and short-course radiotherapy reduce the risk of local recurrence by approximately 50% without major differences in disease-free and overall survival, especially in patients undergoing proctectomy according to TME principles. One potential benefit of long-course chemoradiation is the potential for significant tumor regression with 10 to 15% incidence of complete pathologic response, which may favorably impact prognosis. An overview of the seminal randomized controlled trials that have influenced current neoadjuvant management will be reviewed below.
Preoperative versus Postoperative Chemoradiotherapy
Several randomized trials have compared the use of chemoradiation in the preoperative versus postoperative setting and have generally demonstrated a 50% risk reduction in local recurrence in favor of preoperative therapy (▶ Table 24.5). Long-course chemoradiation has been largely based on a study by the German Rectal Cancer Study Group in which 823 patients with clinical T3/T4 or node-positive tumors were randomized to either neoadjuvant or adjuvant chemoradiation. 6 The preoperative group received 50.4 Gy delivered in 28 fractions over 5 weeks with a 5-day continuous infusion of fluorouracil (5-FU) during the first and fifth weeks, followed by surgery with TME technique 6 weeks after completion of chemoradiation. The postoperative group received the same regimen in the adjuvant setting with an additional 5.4-Gy boost to the tumor bed. Both groups received additional four cycles of bolus 5-FU (4 weeks after surgery in the preoperative group, and 4 weeks after completion of chemoradiation in the postoperative group). The 5-year local recurrence rate was 6% in the preoperative group versus 13% in the postoperative group (p = 0.006), but no difference in 5-year disease-free (68 vs. 65%) or overall survival (76 vs. 74%). Long-term follow-up data from this trial demonstrated a persistent benefit in the reduction of local recurrence (7 vs. 10%; p = 0.048) but again with no differences in disease-free (70 vs. 70%) or overall (60 vs. 60%) survival. 90 Preoperative therapy did improve sphincter preservation in the subset of patients thought to require APR (n = 194) before treatment (39 vs. 20%; p = 0.004). Complete clinical response was seen in 8% overall in the preoperative group. There were no differences in postoperative morbidity between the two groups.
The National Surgical Adjuvant Breast and Bowel Project (NSABP) R-03 trial randomly assigned 267 patients with clinical T3/T4 or node-positive tumors to a similar protocol as the German Rectal Cancer Study Group, but without mandatory TME surgery. 91 In this study, there were no differences in 5-year local recurrence rates (11% in both arms) and overall survival (75 vs. 66%; p = 0.065) between preoperative and postoperative groups despite a 16.5% incidence of complete pathologic response in patients receiving preoperative chemoradiation. However, an improvement in disease-free survival was seen in the preoperative group (65 vs. 53%; p = 0.011). The discrepancy in results between the NSABP R-03 and the German Rectal Cancer Study Group trials may be the lack of uniform TME surgery in the NSABP R-03 study.
A third trial conducted in Korea compared pre- and postoperative chemoradiation with capecitabine as the chemosensitizing agent with pelvic radiotherapy along with surgery according to TME principles for patients with clinical T3/T4 or node-positive tumors. 92 This trial could not demonstrate the benefit of preoperative chemoradiation as the 5-year rates of local recurrence, disease-free, and overall survival were similar between the two groups. Pathologic complete response was seen in 17%. One benefit of neoadjuvant chemoradiation in this study was a higher incidence of sphincter preservation in low-lying tumors (68 vs. 42%).
Preoperative Radiotherapy with or without Concurrent Chemotherapy
Several randomized trials have shown that the addition of concurrent chemotherapy as a radiosensitizing agent decreases local recurrence rates by 50% compared to preoperative radiotherapy alone. The largest study, the European Organization for Research and Treatment of Cancer (EORTC) 22921 trial, used a 2 × 2 factorial design to randomly assign patients with clinical T3/T4 tumors into four groups: preoperative radiotherapy alone (45 Gy over 5 weeks), preoperative chemoradiation (bolus 5-FU/leucovorin [LV] on weeks 1 and 5 of radiotherapy), preoperative radiotherapy and postoperative chemotherapy, and preoperative chemoradiation and postoperative chemotherapy. TME surgery was performed in only 37% of all procedures. Local recurrence was lower in all groups receiving chemotherapy regardless of the timing of administration compared to preoperative radiotherapy alone (17% for preoperative radiotherapy alone, 8.7% for preoperative chemoradiation, 9.6% for preoperative radiotherapy and postoperative chemotherapy, and 7.6% for preoperative chemoradiation and postoperative chemotherapy; p = 0.002). However, no survival benefit at 5 years was found between any of the four groups. These results persisted at 10-year follow-up. 93
The French Fédération Francophone de Canérologie Digestive (FFCD) 9203 trial randomly assigned 742 patients with clinical T3/T4 to a preoperative radiotherapy regimen 45 Gy in 25 fractions with 5-FU/LV during the first and fifth weeks versus radiotherapy alone. 94 Surgery was performed within 3 and 10 weeks after completion of radiotherapy. TME was recommended but not required. Both arms received adjuvant chemotherapy. Local recurrence at 5 years was lower in patients who received chemoradiation (8.1 vs. 16.5%; p =0.004).
Preoperative Short-Course Radiotherapy versus Surgery Alone
The role of preoperative short-course radiotherapy versus surgery alone has been investigated in three randomized trials (▶ Table 24.6). The three trials are consistent in reporting a decrease in local recurrence for patients receiving preoperative radiotherapy versus surgery alone, with a significant effect even in patients undergoing TME surgery. The Swedish Rectal Cancer Trial randomized 1,168 patients with resectable rectal adenocarcinoma to preoperative short-course radiotherapy regimen consisting of 25 Gy delivered in 5 fractions over a 1-week period followed by surgery within 1 week versus surgery alone with no additional radiotherapy. 7 Importantly, surgery was not performed according to TME principles in this trial, and approximately one-third of the included patients were stage I tumors. Patients receiving preoperative radiotherapy experienced a significantly lower local recurrence (11 vs. 27%; p < 0.001) and higher overall survival (58 vs. 48%; p = 0.002) at 5 years compared to patients in the surgery-only arm. Long-term follow-up data reported the persistent benefits in local control and overall survival after a median of 13-year follow-up. 95 However, this trial must be interpreted with caution, as the surgical technique did not adhere to TME principles, which may explain the 27% local recurrence rate in the surgery-only arm, which is more than double that of the surgery-only arms in the Dutch Colorectal Cancer Group and MRC-CR07 trials that included TME-guided surgery. The 16% difference in local recurrence between the two arms may also explain the survival advantage in the preoperative therapy arm, which again was not seen in the two other trials.
The Dutch Colorectal Cancer Group trial investigated the role of short-course radiotherapy in patients undergoing proctectomy according to TME principles. This trial enrolled 1,861 patients to Swedish-style preoperative short-course 5 × 5 Gy radiotherapy and then surgery versus surgery alone. Postoperative therapy was administered to patients with intraoperative tumor spillage or positive margins. Five-year local recurrence was lower in the preoperative therapy arm (5.6 vs. 10.9%; p < 0.001), but no differences in overall survival (64.2 vs. 63.5%; p = 0.902). Long-term follow-up of this trial demonstrated the improvement in local recurrence rates for the preoperative radiotherapy arm persisted at 10 years (5 vs. 11%; p < 0.001). Overall survival was similar between the two arms (48 vs. 49%; p = 0.20). Importantly, local recurrence rates were still lower after preoperative radiotherapy in patients with a negative circumferential margin (3 vs. 9%; p < 0.001).
The Medical Research Council CR-07/National Cancer Institute of Canada (NCIC) Clinical Trials Group C016 randomly assigned 1,350 patients with operable rectal cancer to preoperative 5 × 5 Gy radiotherapy followed by surgery versus surgery and selective postoperative chemoradiation (45 Gy in 25 fractions with concurrent infusional 5-FU) for positive CRM on pathology. Both groups were administered adjuvant chemotherapy if they had positive CRM and/or positive nodes. There were no differences in the incidence of CRM positivity after surgery between the two arms, but 5-year local recurrence rates were lower in the preoperative radiotherapy arm (4.7 vs. 11.5%; p < 0.001). The preoperative radiotherapy arm also had improved disease-free (73.6 vs. 66.7%; p = 0.013), but not overall survival (70.3 vs. 67.9%; p = 0.91).
Preoperative Short-Course Radiotherapy versus Long-Course Chemoradiation
Several trials have directly compared preoperative short-course radiotherapy to long-course chemoradiation and have not demonstrated an improvement in local recurrence or sphincter preservation for either approach. The Polish Colorectal Study Group randomly assigned 312 patients with clinically accessible T3/T4 tumors to conventional chemoradiation with 50.4 Gy in 28 fractions with bolus 5-FU/LV or short-course radiotherapy with 5 × 5 Gy and surgery within 1 week. 96 All patients underwent TME surgery. This study was powered to detect a 15% difference in sphincter preservation between the two arms, with the decision to perform sphincter preservation at the time of surgery. Compliance was higher and there was less early radiation toxicity in the short-course radiotherapy arm. The incidence of complete response was 16% in the chemoradiation arm compared to 1% in the short-course radiotherapy arm (p < 0.001). There were also higher rates of positive CRM in the short-course radiotherapy arm (12.9 vs. 4.4%; p = 0.017), but these findings did not translate into higher sphincter preservation (58 vs. 61%; p = 0.57), 4-year local recurrence (9 vs. 14%; p = 0.170), disease-free survival (55.6 vs. 58.4%; p = 0.820), or overall survival (66.2 vs. 67.2%; p = 0.960) in the chemoradiation arm.
The Trans-Tasman Radiation Oncology Group 01.04 trial also compared conventional long-course chemoradiation with 50.4 Gy with infusional 5-FU versus short-course radiotherapy with 5 × 5 Gy in a randomized trial of 326 patients with cT3N0–2M0 rectal adenocarcinoma. 97 Similarly to the Polish Colorectal Study Group trial, complete pathologic response was higher in the chemoradiation arm (15 vs. 1%), but there was no difference in the proportion of patients undergoing APR for tumors less than 5 cm from the anal verge (short-course 79 vs. long-course 77%) or margin positivity (5 vs. 4%). At 5-year follow-up, local recurrence rates were similar (5.7 vs. 7.5%; p = 0.51), as were disease-free and overall survival, and late complications.
Short-course radiotherapy is generally not recommended for patients with clinical T4 or large bulky tumors since it is not used to induce downsizing or downstaging. However, there have been some data to support the use of consolidation chemotherapy after short-course radiotherapy. One trial randomly assigned 541 patients with clinical T4 or fixed T3 tumors to preoperative short-course radiotherapy followed by 6 cycles of FOLFOX4 (folinic acid, fluorouracil, and oxaliplatin) followed by surgery versus long-course chemoradiation but with the addition of oxaliplatin to bolus 5-FU/LV. 98 In this study, toxicity was lower in the short-course radiotherapy arm but there were similar rates of R0 resection and complete pathologic response. At 3-year follow-up, overall survival was higher in the short-course group. While these results appear promising, the control arm of this study used oxaliplatin along with 5-FU/LV, which has been shown to increase treatment toxicity without any improvement in outcomes compared to conventional 5-FU-based chemotherapy.
Timing of Surgery after Neoadjuvant Therapy
The optimal timing of surgery after completion of neoadjuvant therapy has not yet been determined. Traditionally, this interval was 6 weeks based on the protocol used by the German Rectal Cancer Study Group. 6 However, tumor response to radiotherapy takes time—one study reported that a tumor size of 54 cm3 would require an interval of 20 weeks to regress to less than 0.1 cm3. 99 Pooled results of 13 observational studies including 3,584 patients show that delaying surgery beyond 8 weeks increased the incidence of pathologic complete response by 6% without increasing complications, but with no difference in R0 resections or disease-free or overall survival. 100 Similarly, an analysis of the National Cancer Database reported that an interval greater than 8 weeks was associated with increased pCR (odds ratio [OR]: 1.12; 95% CI: 1.01–1.25) and tumor downstaging (OR: 1.11; 95% CI: 1.02–1.25) compared to 6- to 8-week interval. 101 There have been two randomized trials investigating the timing of surgery after neoadjuvant radiation. The Lyon R90–01 trial compared 2 weeks versus 6 to 8 weeks after long-course radiotherapy and reported a higher pathologic complete response in the longer duration (26 vs. 10%), but with no differences in overall survival. 102 However, this trial used preoperative long-course radiotherapy with 39 Gy in 13 fractions without concurrent chemotherapy rather than conventional chemoradiation, thus limiting the generalizability of these results to standard practice.
The French Research Group of Rectal Cancer Surgery 6 (GREC-CAR-6) trial randomly assigned 265 patients with clinical T3/T4 or node-positive tumors of the mid and low rectum who received conventional neoadjuvant long-course chemoradiation to surgery 7 versus 11 weeks after completion of neoadjuvant therapy. 103 The longer 11-week interval to surgery did not increase the incidence of pathologic complete response (17 vs. 15%; p = 0.60), which was the primary endpoint of the study, or sphincter preservation (90 vs. 89%). The 11-week arm also experienced higher postoperative morbidity (44.5 vs. 32%; p = 0.040) through higher medical complications (32.8 vs. 19.2%; p = 0.014), but similar rates of anastomotic leakage and a trend toward more perineal wound problems after APR and conversion to open surgery. The quality of mesorectal excision was also worse in the 11-week arm (complete mesorectum 78.7 vs. 90%; p = 0.016). The authors concluded that waiting 11 weeks for surgical resection after completion of chemoradiation did not confer any benefit over a 7-week interval and increased postoperative morbidity. Many surgeons now wait 10 to 12 weeks following the completion of neoadjuvant chemoradiotherapy.
Alternative Chemotherapeutic Agents during Neoadjuvant Radiotherapy
At the present time, current standard of care for patients undergoing long-course chemoradiation is to deliver infusional 5-FU during neoadjuvant radiotherapy, with capecitabine as an oral alternative for 5-FU. Two trials have compared capecitabine to infusional 5-FU as the radiosensitizing agent during radiotherapy. 104 , 105 In both trials, locoregional control and overall survival were similar between the two agents, but with different toxicity profiles. These data suggest the equivalency of oral capecitabine and infusional 5-FU during radiotherapy for neoadjuvant therapy.
Oxaliplatin has also been investigated as an addition to 5-FU-based chemotherapy during neoadjuvant radiotherapy in multiple randomized trials with mixed results. 105 , 106 , 107 , 108 , 109 In all trials, the toxicity profile of oxaliplatin was higher than that of traditional 5-FU-based chemotherapy. Only two trials demonstrated benefit in terms of improved pathologic complete response rates 108 , 109 and disease-free survival. 108 Given its greater toxicity profile and unclear effectiveness, oxaliplatin should not be routinely administered as part of neoadjuvant chemotherapy during radiotherapy. Irinotecan has also been investigated in the neoadjuvant setting, but its efficacy was not demonstrated in a small trial of 106 patients, 110 although some benefit has been reported in nonrandomized trials. 111 , 112 , 113 The addition of bevacizumab or epidermal growth factor inhibitors to traditional 5-FU-based chemoradiotherapy does not have any level I evidence for or against the inclusion of these agents.
Initial Neoadjuvant Chemotherapy Instead of Chemoradiation
While the previously mentioned studies have investigated the addition of different agents to chemotherapy during neoadjuvant radiotherapy, the use of neoadjuvant chemotherapy rather than chemoradiotherapy is under active investigation. As the MERCURY study group has shown, select tumors with favorable characteristics that would otherwise undergo neoadjuvant (chemo)radiotherapy can undergo upfront resection without increased risk of local recurrence or worse survival. 114 Furthermore, modern chemotherapy regimens are effective against rectal cancer and are better tolerated in the preoperative versus postoperative setting. This approach may allow for more selective use of radiotherapy, which would spare the patient from the short- and long-term effects of radiation. A pilot study of 32 patients from Memorial Sloan Kettering Cancer Center prospectively enrolled patients with stage II or III rectal cancer to a preoperative regimen of six cycles of FOLFOX (including four cycles of bevacizumab) followed by restaging. Patients with T4 lesions, CRM involvement on preoperative MRI, and/or bulky nodal disease were excluded. Patients who demonstrated stable or responsive disease underwent surgery, while those demonstrating progression underwent conventional long-course chemoradiation. All patients in this study were able to achieve R0 resection, including 25% with complete pathologic response. After a median follow-up of 54 months, there were no locoregional recurrences and 4-year disease-free and overall survival were 92 (95% CI: 82.1–100) and 91.6% (84.0–100), respectively. This study formed the basis of the ongoing multi-institutional PROSPECT trial that should further define management for patients with favorable stage II or III tumors. Initial results of the FOWARC trial, which randomly assigned 495 patients with clinical stage II or III rectal cancer to conventional preoperative chemoradiation, preoperative long-course radiotherapy with concurrent FOLFOX, or FOLFOX alone, show that the chemotherapy arm had equal tumor downsizing to the conventional chemoradiation arm but with less toxicity and postoperative complications. 109 Long-term results are still pending. Despite these promising data, patients with T3N0 or T1–3N1 disease should still undergo conventional neoadjuvant chemoradiation unless they cannot or are unwilling to receive pelvic radiation, unless they are part of a clinical trial.
Induction chemotherapy prior to chemoradiation should be considered for patients with locally invasive tumors (T4) or those with bulky nodal disease, At present, current standard of care indicates that patients undergoing neoadjuvant chemoradiation with curative intent should receive adjuvant chemotherapy, but a multicenter study of specialty cancer centers reported that a sizable minority of these patients do not receive adjuvant chemotherapy. 115 Administering systemic therapy in the neoadjuvant setting may lead to higher rates of resectability and pathologic complete response. In addition, neoadjuvant chemotherapy addresses potential systemic disease earlier and in a more effective manner. FOLFOX is generally better tolerated in the neoadjuvant setting and more patients complete the proposed therapy. While there are no randomized trials, several phase II trials report promising results. A study from the United Kingdom prospectively enrolled patients with high-risk criteria on MRI (tumors within 1 mm of the mesorectal fascia, at or below the levators, extending ≥ 5 mm into the perirectal fat, T4, or T1–2N2) into a treatment regimen consisting of 12 weeks of neoadjuvant capecitabine and oxaliplatin, followed by conventional long-course chemoradiation and TME surgery. 116 Radiologic response was seen in 74% after completion of chemotherapy and 89% after chemotherapy and chemoradiation, while 96% of patients who underwent surgery had R0 resection. Disease-free and overall survival at 3 years were 68 and 83%, respectively. A Spanish study randomly assigned 108 patients to induction versus adjuvant chemotherapy with capecitabine/oxaliplatin for patients undergoing preoperative chemoradiation and TME surgery for T3/T4 or node-positive tumors. 117 Patients in the induction arm were better able to tolerate systemic chemotherapy, but no differences in short-term outcomes were found. Despite the lack of randomized data supporting its effectiveness, induction chemotherapy is recognized as an acceptable option according to the NCCN guidelines for patients with an indication for neoadjuvant (chemo)radiation, although it is generally reserved for patients with high-risk tumors in which there is concern of a positive margin, or with bulky nodal disease.
24.4.3 Morbidity of Neoadjuvant Therapy
The benefits of an approximate 50% reduction in local recurrence for both neoadjuvant long-course chemoradiation or short-course radiotherapy and the 10 to 15% pathologic complete response is associated with significant adverse events that result from these therapies.
The EORTC 22921 study demonstrated that the addition of concurrent chemotherapy to a long-course dose of radiotherapy increases the incidence of grade 3 or higher toxicity during neoadjuvant treatment by almost 50% (7.4% for radiotherapy alone vs. 13.9% for chemoradiation; p < 0.001). 118 In the Polish Colorectal Study Group trial, patients undergoing long-course chemoradiation had a much higher incidence of grade 3/4 toxicity compared to the short-course radiotherapy arm (18.2 vs. 3.2%; p < 0.001), 96 although this was not replicated in the Trans-Tasman trial. 97 The German Rectal Cancer Study Group trial showed no difference in toxicity between pre- and postoperative long-course chemoradiation, although the incidence of anastomotic stenosis was higher in patients receiving postoperative chemoradiation. 6 Because of the lower toxicity profile, compliance is higher for short-course radiotherapy.
However, short-course radiotherapy may have a significant negative impact on postoperative morbidity. In all three trials comparing preoperative short-course radiotherapy versus upfront surgery, the preoperative radiotherapy arm had higher perineal wound complications. 7 , 119 , 120 The Stockholm III randomized trial compared preoperative 5 × 5 short-course radiotherapy with a short (1-week) and long (4- to 8-week) interval to surgery and long-course radiotherapy with a long interval to surgery. 121 The incidence of postoperative complications was highest in patients who underwent preoperative 5 × 5 radiotherapy but had surgery 11 to 17 days after radiotherapy (65%). There were no complications between short- and long-course radiotherapy, although the long-course arm did not receive concurrent chemotherapy. The Polish Colorectal Study Group also reported no differences in quality of life as measured by the EORTC QLQ-C30 or sexual dysfunction between long-course chemoradiation and short-course radiotherapy. 122 Preoperative radiotherapy is also associated with significant impairments in anorectal function. Analyses of longterm data from the Swedish Rectal Cancer trial and the Dutch trial reported higher rates of fecal and urinary incontinence and small bowel obstructions in patients who received preoperative radiotherapy versus upfront surgery. 123 , 124 , 125 These impairments are likely a result of radiotherapy rather than the mode of delivery (i.e., same long-term effects regardless or short- or long-course regimens). 126 Preoperative radiotherapy does not appear to increase the risk of secondary malignancy. 127 These data underline the importance of accurate staging to avoid overtreatment.
Assessment of Tumor Response after Neoadjuvant Therapy
The benefit of repeating the staging investigations after completion of neoadjuvant therapy and prior to surgery is not clear. Several studies have reported a change in management in up to 15% of patients based on repeat CT, PET, or MRI, mostly based on the interval development of metastatic disease. 128 , 129 , 130 , 131 , 132 Restaging may also demonstrate significant tumor regression, which has important prognostic implications. 133 The MERCURY study reported that tumor regression grade was significantly associated with both disease-free and overall survival. Those with a good tumor regression grade had better 5-year overall survival compared to those with poor response (72 vs. 27%; p = 0.001). 134 The German CAO/ARO/AIO-94 trial also showed that patients with a complete response had a 10-year disease-free survival of 89.5% compared to 73.6% for moderate regression and 63% for poor regression. 135 A retrospective review of 725 patients reported similar findings. 136 While the GRECCAR-6 study did not demonstrate a difference in pathologic complete response after 11-versus 6-week interval between completion of chemoradiation and surgery, other studies have shown a higher incidence of tumor regression or pathologic complete response with longer interval to surgery. 137 , 138 , 139 , 140 Given these data, it is reasonable to perform repeat DRE, CEA, endoscopy, and MRI to assess tumor response between 4 and 6 weeks after completion of neoadjuvant chemoradiation. Patients who demonstrate evidence of significant tumor regression on repeat staging and who are asymptomatic should be considered for a longer interval to surgery to induce further response, although this should be balanced against the potentially increased risk of postoperative complications. 103 , 137 , 140 Patients with minimal to no response should undergo surgery within the usual 6- to 8-week intervals.
24.4.4 Summary of the Decision-Making Process for Preoperative Therapy
Patients with clinical stage I tumors (cT1–2, cN0) have a very low rate of local recurrence if high-quality TME is performed; thus, these patients do not usually require neoadjuvant chemoradiation. Subgroup analyses of the Dutch trial reported a 1.7% 5-year local recurrence for stage I tumors, with no statistically significant reduction in local recurrence with the addition of preoperative short-course radiotherapy. In comparison, 5-year local recurrence for stage I tumors decreased from 15% for surgery alone to 5% for the short-course radiotherapy arm in the original Swedish trial, but this may be reflective of the radiotherapy making up for suboptimal surgery (no TME performed) in this study. However, neoadjuvant therapy should be considered for stage I tumors if the preoperative MRI demonstrates a threatened CRM. This is mainly relevant for low rectal tumors as the mesorectum is significantly thinner at this level. The Norwegian Colorectal Cancer Group demonstrated that T2 tumors with a CRM ≤ 2 mm after resection were at much higher risk of local recurrence versus those with a CRM greater than 2 mm (hazard ratio [HR[: 2.76; 95% CI: 1.05–7.38). 141 Neoadjuvant therapy should also be considered in cases where the tumor encroaches on the sphincter complex. In these cases, neoadjuvant chemoradiation may result in tumor downsizing or also downstaging and allow for a sphincter-preserving procedure where an APR would have been otherwise required.
Neoadjuvant chemoradiation has been traditionally recommended for patients with clinical stage II tumors (cT3–4, N0). Pooled analysis from rectal cancer adjuvant trials (before the TME era) demonstrates an 84% 5-year overall survival for T3N0 tumors, with a local recurrence rate ranging from 5 to 11% based on the postoperative regimen used. 142 The addition of preoperative radiotherapy was shown to decrease local recurrence in the Swedish 7 , 95 and CR07 120 trials, but not in the Dutch trial. 119 , 143 While the results of the Swedish trial may not be generalizable in the TME era, the difference in local recurrence rates in the Dutch and CR07 trials may be attributable to the inclusion of adjuvant chemotherapy in the CR07 trial. However, there is emerging evidence to suggest that not all cT3 tumors require preoperative treatment. Patients with T3N0 tumors with less than 5 mm of extramural spread or with a clear CRM greater than 2 mm have acceptable rates of local recurrence (in the 5–10% range depending on the plane of dissection). 64 , 144 One of the issues has been inaccurate local staging techniques, but the MERCURY study group has shown that high-resolution MRI can predict a pathologically negative CRM as well as accurately assess extramural spread within 0.5 mm of final histopathological measurement. 54 , 68 , 69 Furthermore, high-resolution MRI can identify good prognosis cT3 tumors with a clear CRM greater than 1 mm, extramural extension less than 5 mm, and absence of lymphovascular invasion that can undergo proctectomy and TME without need for neoadjuvant therapy with local recurrence rates of less than 5%. 114 However, this approach is not yet standard of care in North America, and these patients should be discussed at a multidisciplinary rectal cancer tumor board to discuss the possible management strategies. Patients with cT4 tumors are at high risk of local recurrence and distant metastasis up to 20 and 60%, respectively. 145 Therefore, these patients should undergo long-course chemoradiation to improve local control and minimize distant recurrence.
Patients with stage III disease (i.e., any cT, cN1–2) should be considered for neoadjuvant therapy with long-course chemoradiation or short-course preoperative radiotherapy. Long-course chemoradiation is standard of care in North America for patients with stage III disease compared to the European approach where short-course Swedish-style radiotherapy is more common. Regardless, it is clear from the available data that patients with N + disease have a significantly higher risk of local recurrence and potentially worse survival outcomes in the absence of neoadjuvant treatment. Long-term follow-up of the Swedish trial showed 23% local recurrence rate for patients treated with preoperative short-course radiotherapy followed by surgery compared to 46% for patients treated with surgery alone (p < 0.001) at a median follow-up of 13 years. No differences in cancer-specific or overall survival were found between the two groups. Again it must be noted that the higher rates of local recurrence in the Swedish trial are reflective of the non-TME surgical technique. Similarly, 12-year follow-up of the Dutch rectal cancer trial showed a significant decrease in local recurrence from 19% in the surgery-alone group compared to 9% in the preoperative short-course radiotherapy + surgery group (p < 0.001). Just like in the Swedish trial, no difference in overall survival was found between the two groups. The MRC CR07 trial also demonstrated a significantly lower local recurrence rate at 3 years for stage III rectal cancer patients treated with preoperative radiotherapy (HR: 0.46; 95% CI: 0.28–0.76).
Finally, patients who present with synchronous stage IV distant metastases in the setting of rectal cancer (i.e., stage IV disease) should be managed based on burden of disease and symptoms arising from the primary rectal tumor. Preoperative systemic therapy should be strongly considered based on the EORTC Intergroup trial 40983 in patients with resectable hepatic disease, which is randomizing patients to six cycles of FOLFOX4 before and after liver resection vs. surgery alone, of which 46% were rectal cancers. 146 The 3-year progression-free survival increased from 28.1% in the surgery-only arm to 36.2% in the perioperative chemotherapy arm (p = 0.041). A liver-first strategy should be adopted as cancer survival is related to burden of systemic disease rather than the primary tumor. 147 The optimal strategy with regard to the timing and type of radiotherapy has not yet been determined. One potential strategy is the administration of perioperative oxaliplatin-based systemic therapy with liver surgery, followed by short-course radiotherapy and then proctectomy. Short-course radiotherapy may be more tolerable in this setting compared to long-course chemoradiation, and shortens time to proctectomy if patients are symptomatic. While preoperative chemoradiation does not affect overall survival, decreasing the risk of local recurrence in patients with resectable disease is still a valid goal given the morbidity of a pelvic recurrence. In cases with unresectable distant metastases, systemic therapy should be initiated with interval restaging to assess for conversion to resectability. The management of the primary tumor is controversial. Patients with obstructive symptoms should be considered for a diverting ostomy prior to starting systemic therapy. However, even patients with a near-obstructing lesion are likely to be able to avoid surgery with palliative radiotherapy and chemotherapy. In one study, only 23% of patients with a near-obstructing lesion required palliative stoma creation with 5 × 5 Gy irradiation, followed by oxaliplatin-based chemotherapy. 148 In patients with limited remaining lifespan, or those who are unable to tolerate surgery, systemic therapy, and/or radiation, colonic stenting should be considered.
24.5 Principles of Rectal Cancer Surgery
24.5.1 Local Excision
Select patients with early rectal cancer can be managed by local excision instead of radical surgery. Early rectal cancer is defined as well to moderately differentiated clinical T1 tumors with absence of lymphovascular and perineural invasion. These patients are at the lowest risk of lymph node metastasis and local recurrence and therefore are amenable for local excision with curative intent. The main controversy surrounding management of early rectal cancer is trade-off between the excellent oncologic outcomes associated with radical surgery for T1 tumors (with 5-year survival approaching 90%) 149 , 150 , 151 and the lower perioperative and long-term morbidity associated with local excision, as radical surgery is associated with significant perioperative complications and long-term functional impairments. 152 , 153 The indications according to the most recent NCCN and American Society of Colon and Rectal Surgeons (ASCRS) guidelines for local excision are reported in ▶ Table 24.7 . The main limitation of local excision is the inability to pathologically assess the draining nodal basins; therefore, careful selection of patients is necessary. Locoregional recurrence after local excision for T1 tumors can be as high as 20%, 154 although it is more commonly quoted in the 10 to 15% range. 150 , 155 , 156 , 157 Furthermore, tumor biology of the rectum is different than that of the colon. 158 The incidence of nodal involvement for T1 tumors of the rectum can be high as 18%, whereas it is 3 to 8% in the colon. 149 , 158 , 159 Kikuchi et al 160 showed that the risk of lymph node involvement was associated with the depth of invasion into the submucosa. None of the patients with tumors that invaded into the first third of the submucosa (sm1) had nodal involvement or local recurrence compared to 8% of patients with invasion into the middle third (sm2) and 20% of patients with invasion into the deepest third (sm3). An analysis of T1 tumors undergoing radical excision from the Surveillance, Epidemiology, and End Results database reported that the overall incidence of nodal involvement was 16.3%. 159 Tumors that were sized over 1.5 cm and poorly differentiated were at significantly higher risk. Similarly, a meta-analysis of 23 studies including 4,510 patients reported that T1 tumors with greater than 1 mm invasion into the submucosa (OR: 3.87; 95% CI: 1.50–10.00), lymphovascular invasion (OR: 4.81; 95% CI: 3.14–7.37), and poor differentiation (OR: 5.60; 95% CI: 2.90–10.82) were independent risk factors for lymph node metastasis. 161 Similarly, larger tumors, depth of invasion beyond sm1, and lymphovascular invasion were found to be independent predictors of local recurrence after local excision of rectal cancer in an Association of Coloproctology of Great Britain and Ireland Transanal Endoscopic Microsurgery Collaborative study. 155 Patients with any of these risk factors should not undergo curative local excision, or if these features are found on final pathology after local excision, radical surgery should be recommended.
Local excision can be performed via traditional transanal excision (TAE) or through an advanced transanal endoscopic surgery (TES) operating platform such as transanal endoscopic microsurgery (TEMS), transanal endoscopic operation (TEO), or transanal minimally invasive surgery (TAMIS). While traditional local excision was limited to tumors situated in the low rectum due to the lack of access of mid- and high-rectal lesions using Parks retractors, newer platforms have been developed which offer better visualization of and access to the lesion, allowing for an improved resection quality. TEMS was first reported by Dr. Gerard Buess in 1984, 163 which used a rigid operating platform with stereoscopic views (resectoscope) to gain endoluminal access. TEO was introduced as an alternative to TEMS. Subsequently, TAMIS was described by Atallah et al in 2010 as a less expensive alternative to TES. 164 Endoluminal access is obtained using a soft single-incision operating port and standard laparoscopic equipment. The SILSPort (Covidien, New Haven, CT) was first used, but since then specialized operating ports developed specifically for TAMIS have been developed, such as the GelPOINT Path (Applied Medical, Rancho Santa Margarita, CA). There are no in vivo comparative data between the different operating platforms, and the choice will be dictated by surgeon preference and equipment availability. 165 However, TEMS has been compared to traditional TAE using Parks retractors and has been shown to be superior in terms of resection quality. A meta-analysis showed that TES was associated with a lower incidence of lesion fragmentation (OR: 0.10; 95% CI: 0.04–0.21), positive margins (OR: 0.19; 95% CI: 0.11–0.31), and local recurrence (OR: 0.25; 95% CI: 0.15–0.40) with no difference in complications (OR: 1.02; 95% CI: 0.66–1.58). 166 It must be noted that tumors excised by TAE were located almost solely in the distal rectum, which have a worse prognosis compared to lesions in the upper two-thirds, and these results may be due to selection bias. Nevertheless, the poor outcomes associated with TAE are even more reason to use an advanced operating platform (TES) to approach these lesions.
In cases with diagnostic uncertainty, local excision can act as an “excisional biopsy” and further management dictated by final pathologic results. Given the limitations of the current locoregional staging modalities, there will be a not insignificant proportion of patients who will be understaged or will have risk factors for lymph node metastasis or local recurrences that were not apparent on the preoperative biopsy. These patients should undergo salvage radical excision within 30 days of the initial local excision. Perioperative outcomes appear to be similar between salvage TME after local excision and upfront TME 167 , 168 ; however, one study reported a higher APR rate. 168 The authors hypothesized that the inflammatory process post-TES made a low colorectal or coloanal anastomosis (CAA) not feasible. Several studies have shown that salvage surgery in these cases does not compromise oncologic outcomes. 169 , 170 Patients who refuse radical surgery or are too medically unfit can be considered for adjuvant chemoradiotherapy, although level I data to support such an approach are lacking. Case series from MD Anderson and Memorial Sloan Kettering Cancer Centers show that adjuvant radiotherapy without concomitant chemotherapy may provide adequate local control, 171 , 172 although recurrence rate are still high and survival may still be inferior compared to radical resection. Long-term follow-up of the Cancer and Leukemia Group B (CALGB) 8984 trial demonstrated that 10-year local recurrence rates for T2 lesions treated with local excision and postoperative chemoradiation was 18% compared to 8% for T1 lesions treated with local excision alone. 173 Disease-free and overall survival was also lower in the T2 lesions despite chemoradiotherapy.
TES requires establishing pneumorectum through the transanal operating port when rigid platforms are employed. For TEMS, the dedicated system using a continuous flow model that prevents billowing is a key component of the technology. Standard laparoscopic insufflators and equipment can be used when flexible platforms are utilized, although newer insufflation technologies such as the high-flow CO2 (AirSeal, ConMed, Utica, NY) reduce the amount of luminal bellowing and greatly facilitate the procedure. Regardless of the operating platform that is used, a full-thickness excision into the mesorectal fat with at least 1 cm surrounding should be performed for malignant lesions (▶ Fig. 24.3), as a submucosal resection is at high risk of residual disease (OR: 6.47; 95% CI: 3.00–13.97 compared to full-thickness excision). 155 Lesions as high as 15 cm from the anal verge can be resected via TEO, although the risk of intraperitoneal perforation is highest for anterior lesions located 7 cm and above from the anal verge. The resulting full-thickness defect can either be closed or left open, as the data are equivocal. There may be fewer complications if the defect is closed, 174 but this theory has not been definitively proven. 175 Clearly if peritoneal perforation occurs, the resulting defect must be closed. This can be performed via the transanal approach, or a transabdominal approach if the resulting loss of pneumorectum precludes endoluminal closure. There are no data to suggest that oncologic outcomes are compromised if peritoneal perforation occurs. These cases are usually done as outpatient surgery, unless there is concern about peritoneal violation or bleeding. The most common immediate postoperative complications include pain, urinary retention, and bleeding. Anorectal function impairment can occur, but usually resolves within 6 months. 176 , 177 , 178
Several studies have shown less postoperative morbidity and comparable long-term outcomes between local excision and radical resection for T1 rectal adenocarcinoma. Winde et al 179 randomly assigned 52 patients with well to moderately differentiated T1 tumors to TAE versus anterior resection. There were fewer early complications and equal survival outcomes, although this study was underpowered to detect any real differences in these outcomes. Meta-analyses have reported significantly lower perioperative complications (8.2 vs. 47.2%; p = 0.01) and mortality (0 vs. 3.7%; p = 0.01) for local excision by TES compared to radical surgery. 180 However, TES local excision was associated with a higher incidence of local recurrence compared to radical resection, but without any differences in disease-free or overall survival. 180 , 181 If only “low-risk” T1 cancers are considered (well to moderate differentiation, absence of lymphovascular invasion), recurrence rate after TES was similar to that of radical surgery (4 vs. 3%), whereas for “high-risk” T1 tumors (poor differentiation or presence of lymphovascular invasion), TES had significantly higher rates of local recurrence (33 vs. 18%). Local excision is also associated with improved quality of life compared to radical surgery for early rectal cancer. Lezoche et al 182 demonstrated that quality-of-life impairments as measured by the EORTC QLQ–C30 and QLQ–CR38 lasted only 1 month after TES local excision, whereas these impairments persisted up to 6 months after laparoscopic TME. At 1 year, neither TES nor the laparoscopic TME had any changes from baseline. Long-term data show that local excision by TES and laparoscopic TME has similar quality of life (EORTC QLQ–C30 and EQ-5D), but there was a higher incidence of defecation problems in patients undergoing radical surgery. 183
The cost of rigid TES platforms is approximately US$80,000, whereas the main equipment costs of TAMIS are related to the disposable transanal port, which costs approximately US$600 to 800. 184 However, a cost analysis comparing local excision by TES versus open surgery for early rectal cancer reported that only 12 TES cases were required to recoup the capital costs of the equipment. 185 Furthermore, the improved resection quality offered over traditional TAE should offset the increased equipment costs by minimizing recurrences and associated high treatment costs.
Surveillance after local excision of early rectal cancer should follow NCCN guidelines and include history, physical examination, and CEA every 3 to 6 months for the first 2 years, then every 6 months thereafter for a total of 5 years. Given the risk of local recurrence, close endoscopic and radiologic follow-up should also be considered as part of the surveillance strategy. Flexible sigmoidoscopy (to detect mucosal recurrence) and MRI of the rectum (to detect mural recurrences that may not be apparent by endoscopy) every 3 to 6 months for the first 2 years is reasonable.
If recurrence is detected, a complete staging workup should be performed to assess whether the recurrence is locoregional or if there are distant metastases. Early studies that reported outcomes of local recurrence after TAE for T1 and T2 tumors reported that locoregional recurrences were often advanced and required multivisceral resection in a significant proportion of patients. 186 , 187 , 188 An R0 resection was achieved in 79 to 94% of patients undergoing surgical salvage. Survival was poor in these patients and was not equivalent to those undergoing upfront radical resection. More recent studies reporting on outcomes for local recurrences after TEMS report that 61 to 88% of patients with recurrent disease were eligible for curative salvage surgery, and that this may not negatively impact survival as compared to upfront surgery. 189 , 190 However, these data are heterogeneous in terms of the surveillance strategy and available technologies over the study periods. Better present imaging modalities allow for more accurate clinical staging, which may improve patient selection for local excision. Nevertheless, these data underscore the importance of proper patient selection for these procedures.
Local Excision after Chemoradiation
Unlike T1 tumors, T2 lesions have a much higher failure rate after local excision alone. Locoregional recurrence for T2 tumors is at least double that of T1 tumors, ranging from 13 to 30%, 156 , 162 , 191 , 192 which may be in part due to the 30 to 40% incidence of occult nodal involvement. 192 Despite this, a significant proportion of patients with T2 tumors are treated with local excision alone. 193 This has led to the application of neoadjuvant chemoradiation prior to local excision in order to reduce local recurrence rate and avoid the morbidity of radical oncologic surgery. There are some data to support this approach. Lezoche et al 182 randomly assigned 100 patients with T2N0M0 tumors less than 3 cm within 6 cm of the anal verge to local excision by TES versus laparoscopic TME after long-course neoadjuvant chemoradiation with 50.4 Gy and concomitant infusional 5-FU. Tumors were downstaged in 51% of cases, with 28% in the TEMS and 26% in the surgery arm achieving ypT0. No patients in the surgery arm had positive lymph nodes on final pathology. After a median follow-up of 9.6 years, the local recurrence rate was similar for both arms (TES 12 vs. surgery 10%; p = 0.686), as was cancer-related (89 vs. 94%; p = 0.687) and overall (72 vs. 80%; p = 0.609) survival. Similarly, the ACOSOG Z6041 phase II trial investigated a preoperative chemoradiation regimen consisting of capecitabine, oxaliplatin, and 54-Gy radiotherapy, followed by local excision for patients with clinical T2N0 tumors. 194 Of the 77 patients who completed the preoperative regimen and underwent local excision, 64% experienced tumor downstaging with 44% overall achieving a pathologic complete response. 195 Three-year disease-free and overall survival was 88.2 and 94.8%, respectively.
Despite these promising data, the success of this approach appears to be dependent on the tumor response to neoadjuvant therapy. In a systematic review of 20 studies including 1,068 patients, Hallam et al demonstrated that the rate of local recurrence was high if a pathologic complete response was not obtained. 196 After a median follow-up of 54 months, local recurrence occurred in 4.0% (95% CI: 1.9–6.9) of patients with ypT0, 12.1% (95% CI: 6.3–19.4) for ypT1, 23.6% (95% CI: 13.0–36.1) for ypT2, and 59.6% (95% CI: 32.6–83.8) for ypT3. The pooled local recurrence rate for tumors ≥ ypT1 was 21.9% (95% CI: 15.9–28.5). Similarly, the rate of distant metastasis was 2.8% (95% CI: 0.8–6.1) for ypT0 and 20.9% (95% CI: 14.7–27.9) for tumors ≥ ypT1. The significantly higher failure rate for tumors ≥ ypT1 may be explained by the fact that more than 20% of ypT1/T2 tumors had residual lymph node metastases after radical surgery, as was shown in the German CAO/ARO/AIO-94 trial. 197 Perez et al 198 also showed that patients with cT2–4N0M0 that do not demonstrate complete clinical response (cCR) after chemoradiation have a high incidence of unfavorable histology (ypT2 or 3 in at least 66%). These data suggest that local excision alone after neoadjuvant chemotherapy in patients without apparent complete clinical or pathologic response would result in understaging and undertreatment in a significant proportion of patients.
Another consideration is that perioperative morbidity after local excision post-neoadjuvant chemoradiation is high. Several studies have reported that a significant proportion of patients experienced a rectal wound complication after local excision. Marks et al 199 showed that 33% of patients receiving neoadjuvant therapy had a complication versus 5% in patients who did not (p < 0.05). Rectal wound complications were also higher (25 vs. 0%; p = 0.015). Perez et al 200 also showed that patients undergoing local excision after neoadjuvant therapy had significantly more rectal wound dehiscence (70 vs. 23%; p = 0.03), and pain requiring readmission (43 vs. 7%; p = 0.02). 200 Anorectal function also appears to be significantly impaired in these patients. Gornicki et al 201 showed that anorectal function and quality of life was similar between local excision and radical surgery after neoadjuvant chemoradiation, suggesting that the benefits of local excision in terms of preserved function were lost with neoadjuvant treatment. Similarly, Habr-Gama et al 202 demonstrated that local excision after chemoradiation was associated with significantly worse anorectal function, as measured by manometry and the Fecal Incontinence Index and Quality of Life assessment, compared to patients who underwent observation. Given the equivocal results from the current data, local excision after neoadjuvant chemoradiation for cT2N0 tumors should not be routinely offered unless patients are unwilling or unable to undergo radical resection. Two multicenter randomized trials are currently underway to investigate this approach: ChemorAdiation therapy for rectal cancer in the distal Rectum followed by organ-sparing Transanal endoscopic microSurgery (CARTS) and Transanal endoscopic microsurgery and Radiotherapy in Early rectal Cancer (TREC). 203 , 204 Until the results of these trials are available, this approach should only be performed as part of a clinical trial.
24.5.2 Watch and Wait
Oncologic TME surgery for patients who are medically operable remains the standard of care for patients regardless of clinical response to neoadjuvant therapy. The incidence of complete pathologic response (pCR) after long-course chemoradiation ranges from 10 to 44%, with improved oncologic outcomes in these patients compared to those without pCR. 205 , 206 , 207 Maas et al performed an individual patient data meta-analysis including 3,105 patients who received neoadjuvant chemoradiation and TME. 207 The overall pCR rate was 16%, with patients achieving pCR having improved local control (HR: 0.41; 95% CI: 0.21–0.81) disease-free (HR: 0.54; 95% CI: 0.40–0.73) and overall (HR: 0.65; 95% CI: 0.47–0.89) survival compared to patients with residual disease. Given the improved outcomes with pCR and the significant short- and long-term morbidity and mortality associated with radical surgery, there has been interest in identifying patients for which surgery can be avoided (“watch-and-wait” approach). 208
Habr-Gama et al 209 were the first to report outcomes for the watch-and-wait approach. In their study, 29% of patients who underwent long-course chemoradiation were deemed to have cCR at 8 weeks using a combination of endoscopic and radiologic (CT with or without TRUS) findings. Patients with apparent cCR underwent intensive surveillance for possible recurrence with monthly follow-ups for repeat DRE, proctoscopy, and serum CEA levels, and CT scan every 6 months. These 79 patients were compared to 22 patients who had pCR on surgical pathology. After a median follow-up of 57 months, 2 patients in the observation group experienced a local recurrence that was successfully salvaged and 3 developed metastatic disease. There were no differences in disease-specific (92 vs. 83%) or overall survival (100 vs. 88%) between the observation and surgery groups. A later follow-up study on 183 patients, of which 90 (49%) had cCR (the increase in cCR was due to a change in the definition of a cCR), reported that local recurrence developed in 28 patients (31%). 210 Of these, 17 had recurred within 12 months of observation and 11 afterward. Salvage was possible in 93% with R0 resections in 89%. Overall survival at 5 years was 91%.
A Dutch study showed that only 1 patient out of 21 with cCR had a local recurrence after a mean follow-up of 25 months, which was managed by local excision. 211 No patients undergoing observation in this study had distant metastases. The endoscopic, radiologic MRI, and clinical criteria to identify patients with cCR were more stringent than Brazilian study, as was the surveillance protocol. There was no difference in disease-free or overall survival when compared to patients who underwent resection and had pCR, with better bowel function in the observation group. Similarly, the experience from Memorial Sloan Kettering Cancer Center reported 6 recurrences out of 32 patients managed by observation alone after median followup of 28 months. 212 All of the local recurrences were salvageable. There were no differences in 2-year disease-free (88 vs. 98%; p = 0.27) and overall (96 vs. 100%; p = 0.56) survival between observation and patients who were resected with a pCR.
The Oncologic Outcomes after Clinical Complete Response in Patients with Rectal Cancer (OnCoRe) study 213 performed a propensity-score-matched analysis of patients undergoing induction chemoradiotherapy managed by watch-and-wait or surgical resection from four United Kingdom cancer centers. Primary outcome in this study was non-regrowth disease-free survival. A total of 129 patients were observed after cCR, of which 34% had experienced local regrowth. Salvage therapy was feasible in 88% of these patients. After one-to-one matching with patients undergoing surgical resection, there were no differences in 3-year non-regrowth disease-free (88 vs. 78%) or overall (96 vs. 87%) survival between watch-and-wait and resected patients. These combined data suggest that observation after cCR may not need surgery, although prospective randomized trials are lacking. At the present time, surgical resection is recommended for patients who are medically operable and willing to undergo surgery.
One of the significant barriers to the watch-and-wait approach is the ascertainment of cCR of the primary tumor as well as nodal status. Lymph nodes may contain residual tumor in up to 10% of patients with ypT0. 214 , 215 A combination of physical examination with endoscopic and radiologic evaluation should be used, as the diagnostic accuracy of each of these modalities is low. Assessment by DRE and proctoscopy was only able to correctly predict pathologic complete response in 25% of cases. 216 Endoscopic findings of complete response include whitening of the mucosa, telangiectasia without mucosal ulcerations, and subtle loss of pliability of the rectal wall (▶ Fig. 24.4). Residual disease should be highly suspected in the presence of a palpable nodule, ulceration, or irregularity. However, up to 61 to 74% of patients with a pCR may still have mucosal abnormalities suggestive of an incomplete clinical response. 217 , 218 Similarly, 27% of patients with endoscopic findings suggestive of cCR had residual disease. 217 Complete response is similarly problematic to accurately predict on radiologic imaging. MRI has difficulty distinguishing between residual tumor and radiation-induced edema or fibrosis. In cases of complete response, a scar replaces the site of the tumor, which is represented as an area of low signal intensity on T2-weighted imaging. 219 Newer MRI techniques such as diffusion-weight imaging may improve diagnostic accuracy for assessment of tumor response. 71 PET/CT shows promise since complete response should eliminate metabolic activity of the tumor. A systematic review of 34 studies including 1,526 patients reported a 71% sensitivity of PET/CT for the detection of complete response. 72 A prospective study by Perez et al 220 reported that clinical evaluation alone accurately detected residual cancer in 91% of cases, and PET/CT increased overall accuracy to 96%. Normalization of CEA after neoadjuvant therapy is also reflective of tumor response. 35 , 221 Finally, local excision of the tumor scar may confirm mural sterility, but is associated with significant pain and wound complications. 199 , 222
It is not clear when to perform this assessment, as studies have ranged between 4 and 10 weeks. 209 , 211 , 212 Perez et al 223 performed PET/CT at 6 and 12 weeks after completion of neoadjuvant chemoradiation and reported that patients in whom the SUVmax increased during this interval were unlikely to develop cCR or significant tumor regression. However, not all patients should wait 12 weeks for definitive management, be it resection or observation. The GRECCAR-6 trial did not demonstrate any difference in pCR between patients randomized to 7-versus 11-week interval to surgery (15.0 vs. 17.4%; p = 0.598), but the patients in the 11-week arm experienced increased postoperative morbidity (44.5 vs. 32.0%; p = 0.040). 103 It is reasonable to assess tumor response at a minimum of 6 weeks after completion of neoadjuvant chemotherapy. Patients who demonstrate significant or complete mucosal response based on DRE/proctoscopy may be considered for further evaluation of complete response, whereas patients with moderate to poor response should undergo resection within 6 to 8 weeks after completion of neoadjuvant therapy as per current guidelines.
Patients who exhibit evidence of cCR should be willing and able to undergo a strict surveillance protocol, especially during the first year as this is when the majority of recurrences occur. Based on the approach of Habr-Gama et al, 209 patients should undergo monthly follow-up with DRE or proctoscopy for the first 3 months, then every 2 to 3 months for the remainder of the first year. CEA is checked every 2 months. Radiologic evaluation using CT or MRI should be done at the time of initial tumor assessment, and then every 6 months. Follow-up visits should continue every 3 months after the first year. Suspicious findings on clinical assessment or imaging should prompt further evaluation or radical surgery.
Related posts:
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
