ABSTRACT: The definition of overtreatment of rectal cancer is controversial, and thus it is difficult to accurately quantitate its prevalence. All components of rectal cancer treatment are associated with significant potential for morbidity and dysfunction that may have a negative impact on the patient’s quality of life. No one would disagree with the tenet that overtreatment should be avoided whenever possible. Despite that consensus, little attention is given in the literature to the issues of overtreatment of rectal cancer. This review article presents a variety of clinical scenarios and summarizes available data demonstrating that overtreatment of some patients with rectal cancer is occurring on a regular basis. It is hoped that this will stimulate clinicians to critically review their own practices to eliminate such overtreatment. Development of new clinical trials to determine whether current practice guidelines are promoting overtreatment of selected rectal cancer patients is proposed.
Optimal therapy of any condition implies that all components of the treatment regimen are essential to achieve the intended therapeutic effect and are not redundant or associated with excess morbidity. Both undertreatment and overtreatment are undesirable. Because we cannot accurately distinguish patients with small or micrometastatic foci of cancer from those free of disease, it is often necessary to empirically provide adjuvant therapy to at-risk patients, recognizing that doing so represents appropriate therapy for some but overtreatment for others. Thus, some overtreatment is inevitable given our current limited ability to accurately detect and stage cancer. Nonetheless, patients diagnosed with cancer rarely raise concerns about overtreatment, probably because the diagnosis triggers an emotional response favoring overtreatment rather than risking undertreatment.
It behooves physicians treating cancer patients to critically and nonemotionally determine whether overtreatment is occurring and, if so, at what cost to the patient in terms of quality of life and unnecessary morbidity and mortality. Because all components of the current treatment regimens for rectal cancer are associated with a risk of significant morbidity and/or detrimental impact on normal functions, it is appropriate to review the possibility that we could eliminate some of the overtreatment of patients with rectal cancer.[1-3]
Variety of Treatment Options
The goals of treatment of a patient with rectal cancer are to control the local disease, prevent distant spread, restore bowel continuity, and maintain normal anal continence, preserve sexual and bladder function, and minimize other treatment-associated morbidity and mortality. Today, a variety of rectal cancer treatment regimens are available. These span the spectrum from simple polypectomy to prolonged, potentially morbid, multimodality regimens involving neoadjuvant chemoradiation therapy, radical extirpative surgery, and adjuvant chemotherapy.
Results of treatment of rectal cancer are usually reported by stage as defined by the TNM system (Table 1). Choice of therapy, however, is based on a somewhat subjective analysis of many factors that must be considered before treatment is initiated. These include clinical staging; imaging studies such as endorectal ultrasonography, computed tomography (CT), magnetic resonance imaging (MRI), and positron-emission tomography (PET) scanning; histopathology of the biopsy and/or resected specimen; technical considerations; and the patient's operative risks, comorbidities, prior therapies as well as the specific desires or needs of the patient.
Given the complexity of decisionmaking in this area and the lack of a uniformly accepted standard based on irrefutable evidence, it is difficult to be dogmatic about the definitions of optimal treatment, undertreatment, and overtreatment. Nonetheless, there are generally accepted treatment guidelines based primarily on pretreatment stage of disease that can serve as a platform for considering the issues of overtreatment of rectal cancer patients.[6,7]
This review is focused on curative intent therapy of stage 0, I, II, and III rectal cancers and will not address the issues related to overtreatment of stage IV disease or overuse of palliative therapy. A discussion of whether radical resection of the rectum after a complete clinical response to neoadjuvant chemoradiation therapy constitutes overtreatment is beyond the scope of this article.
Cancer in a Polyp
Cancer in a polyp is an imprecise term that encompasses a broad spectrum of polypoid neoplasms. The risk of developing a malignancy in a polyp is related to both its size and histology. Although it is rare to find invasive cancer in a polyp smaller than 0.6 cm, the risk of carcinoma in a polyp greater than 3 cm is as high as 35% to 50% depending on its histology.[ 9,10] Malignancy is noted in 5% of tubular adenomas, 22% of tubulovillous adenomas, and up to 40% of villous adenomas.[11-13]
Accurate pathologic assessment of an excised polyp is essential to calculate the risks of subsequent development of local recurrence and lymph node or distant metastasis. To accurately assess the level of invasion of the cancer, all polypoid lesions should be completely excised, preferably in one piece, and oriented before fixation. If the dysplastic tissue is confined to the mucosa, the cancer is termed "in situ" and there is no risk of metastasis. If tumor cells penetrate through the muscularis mucosa (basement membrane) of the bowel wall, the cancer is considered "invasive" because dysplastic cells have access to lymphatics and vessels within the submucosa and can recur locally or metastasize to distant sites.
Invasive carcinoma in a polyp is a common clinical scenario occurring in 1.5% to 12% of polyps removed by colonoscopic polypectomy.[9,10,14-18] In general, the deeper the level of invasion, the greater the risk of recurrence and/or metastasis. Several schemas have been proposed to classify early-stage cancers arising in polyps based on their level of invasiveness.
Haggitt et al correlated the level of cancer invasion with the risk of recurrence and/or metastasis and proposed a classification to predict prognosis and guide decision-making regarding the optimal treatment of polyps with cancer (Figure 1). A noninvasive cancer confined to the mucosa is classified as level 0, while invasive carcinomas invading into the head, neck, and stalk of a pedunculated polyp are classified as level 1, 2, and 3, respectively. A cancer invading into the submucosa at the base of the stalk of a pedunculated polyp or into the submucosa of a sessile polyp is classified as Haggitt level 4.
Kikuchi et al proposed a different classification system of T1 cancers based on extent of submucosal invasion by separating them into upper one-third (Sm1), middle one-third (Sm2), and lower one-third (Sm3) invasion. In this classification system, Haggitt level 1, 2, and 3 cancers are all Sm1 cancers, whereas Haggitt level 4 cancers, whether arising in sessile or pedunculated polyps, could be Sm1, Sm2, or Sm3 depending on the depth of invasion into the submucosa. Lesions extending into the muscularis propria are classified as T2 cancers, and neither the Haggitt nor the Kikuchi system of predicting prognosis and risk of recurrence and/or metastasis are applicable.
The choices of therapy following removal of a rectal polyp with cancer include observation, local excision of the polypectomy site, endocavitary radiation to the polypectomy site, radiation of the pelvis plus sensitizing chemotherapy with or without systemic chemotherapy, radical surgery, or a combination of these modalities. The level of invasion of the cancer in the polyp, the adequacy of the resection margins, the size of the lesion, and other prognostic features discussed below-coupled with the clinician's knowledge of the patient's risk factors and desires-are all important variables used to select the optimal therapy for a given patient.
Carcinoma In Situ
An in situ noninvasive carcinoma confined to the mucosa is classified as Haggitt level 0. Because such lesions have no potential for recurrence or metastasis, sessile and pedunculated polyps containing carcinoma in situ (also called high-grade dysplasia, severe dysplasia, mucosal carcinoma, or noninvasive carcinoma) can be managed solely by polypectomy or local excision, as long as margins are free of dysplasia. If a polyp is too large to remove in one or two pieces by the snare excision polypectomy technique, transanal excision is advised. Rarely, as in the case of a large circumferential villous lesion or a large, proximal sessile rectal polyp, radical resection is necessary.
For the most part, radical surgery for in situ carcinoma of the rectum should be considered overtreatment. Chemoradiation is similarly unnecessary and inappropriate. The incidence of such overtreatment is unknown, but there are three scenarios that suggest overtreatment of such lesions is occurring:
• Not infrequently, gastrointestinal pathologists and colorectal surgeons in referral centers are asked to review cases in which a major resection is being contemplated because of the finding of "cancer" in a rectal polyp. Careful review of the pathology slides sometimes shows that the "rectal cancer" is not invasive but only in situ.
• A second scenario is that the pathology report correctly labels the cancer as carcinoma in situ, but the consulting surgeon interprets this as a condition that justifies a radical resection. Indeed, some gastrointestinal pathologists now avoid the terms "in situ carcinoma" and "intramucosal carcinoma" to minimize the risk that a well-intentioned surgeon may perform an unnecessary, radical operation.
• In the third scenario, review of the pathology slides reveals that the specimen is so disoriented or that the polyp was removed in so many pieces that the pathologist cannot accurately assess the depth of invasion of a particular focus of cancer. Such confusion can create a dilemma for the clinician and patient in choosing the appropriate therapy. In this setting, there is often a tendency to opt for aggressive therapy rather than risking undertreatment.
Much of the overtreatment that does occur in this setting could be easily avoided if such lesions were routinely removed in one piece and oriented properly before fixation and if pathologists routinely classified in situ carcinomas as "high-grade dysplasia (Tis, Nx, Mx)" in their reports, adding a simple explanatory statement that this is a noninvasive lesion with no capacity to metastasize. Although the number of patients who undergo radical surgery for an in situ carcinoma is unknown, anecdotal reports and personal experience suggest this is not a rare problem. The consequences of overtreatment by radical surgery or chemoradiation for in situ carcinoma are significant.
Invasive Carcinoma in a Pedunculated Polyp
The choice of appropriate therapy for a pedunculated polyp with invasive carcinoma is based primarily on a determination of the margin of resection, the presence of unfavorable histologic features associated with an increased risk of metastases, and the level of invasion of the dysplastic lesion. Polypectomy alone is curative for the vast majority of such lesions, providing there is no evidence of lymphovascular invasion, poorly differentiated histology, or tumor within 2 mm of the resection margin. The presence of such features mandates more aggressive treatment because of the greatly increased risk of local recurrence and metastatic spread.
The definition of an adequate margin is controversial, but most authors report that recurrences and regional lymph node metastases are rare or do not develop at all if the tumor-free margin is ≥ 2 mm.[21,22] Morson et al classified 60 malignant polyps removed by snare polypectomy at St. Mark's Hospital in London as having complete, doubtful, or incomplete margins of resection. Surgical resection was performed in 14 patients because of an incomplete or doubtful clearance. Two patients had residual tumor in the resected specimen; one patient died of metastatic disease and the other survived long-term. They found no recurrences in the 46 patients treated by polypectomy alone.
Unfavorable histologic features that predict a high rate of residual tumor and/or metastasis are poorly differentiated histology and lymphatic or vascular invasion.[24,25] When such histologic features were present and the margin of clearance was incomplete, Cooper et al reported cancer-specific failure in 20% of 71 patients. By comparison, no failures occurred in 46 patients when these factors were not present. Similar data come from Volk et al, who reported recurrence in 10 of 30 patients after local excision of poorly differentiated cancers but no recurrences in the 16 patients with margins ≥ 2 mm and tumor that was not poorly differentiated.
Carcinomas invading the head, neck, or stalk of a pedunculated polyp are classified as Haggitt levels 1, 2, and 3, respectively. The incidence of residual cancer in the bowel wall and the risk of metastasis to regional lymph nodes or to distant sites on long-term follow-up after polypectomy of such lesions is low. A Mayo Clinic study of 151 patients who underwent resection for cancer in a polyp found a 0% incidence of lymph node metastasis in Haggitt levels 1, 2, and 3 polyps in the absence of other histologic risk factors.[ 27] If other histologic risk factors are not included in the analysis, residual cancer, lymph node metastasis, or recurrent cancer is reported to be 0.7% for level 1, 5% for level 2, and 12.9% for level 3.[19,28,29]
In most instances, radical surgery or chemoradiation is not needed for Haggitt level 1, 2, and 3 invasive carcinomas in a pedunculated polyp with clear margin of resection and no unfavorable histologic features. Polypectomy alone is adequate, and more aggressive treatment such as radical resection or chemoradiation should be considered overtreatment. It is difficult to quantitate this occurrence, but personal experience suggests it is not uncommon.
Many patients request a second opinion after they have been advised to undergo radical surgery by a wellintentioned gastroenterologist or surgeon because the pathologist has identified invasive carcinoma in a pedunculated polyp. Usually, the pathology report does not mention the precise level of invasion, and usually there is no evidence that the surgeon proposing the resection has reviewed the pathology slides to determine the level of invasion and presence of adverse features. The surgeon is obligated to review the specimen and educate the patient regarding the nature of the cancer in the polyp, the inferred risk of residual cancer or metastasis based on histologic review, and the realistic morbidity and mortality of the proposed treatment. The authors' general approach is summarized in Tables 2 and 3.
Haggitt Level 4 Invasive Carcinoma in a Polyp
Haggitt level 4 invasive carcinoma in a sessile or pedunculated polyp is predictive of a 10% risk of regional lymph node metastasis. If combined with an unfavorable histology, the risk increases to 25%. In the TNM nomenclature, such cancers are classified as T1. Their optimal treatment is controversial. In Europe, many centers treat all curable rectal cancers (including T1 cancers) with a combination of preoperative radiation and radical surgery. As noted below, this clearly seems to be an overtreatment of stage I rectal cancers. In the United States, the generally accepted treatment options for patients with T1, Haggitt level 4 invasive carcinoma in a polyp are standard radical resection of the rectum or local excision with or without chemoradiation.[ 7] Neither of these options is currently considered overtreatment.
However, a different perspective is gained by considering that 75% to 90% of such patients do not have residual cancer in the adjacent wall or lymph nodes and would not develop recurrent cancer if a less radical treatment option had been used. In this view, current practice overtreats a significant number of patients. The dilemma is to accurately predict which patients need the radical treatment and which can be cured with a less morbid approach.
The Kikuchi classification based on extent of submucosal invasion was developed specifically to provide guidance for the optimal treatment of T1, Haggitt level 4 invasive cancers arising in a polyp. In their review of 182 such lesions treated by polypectomy, the Kikuchi classification was Sm1 in 64, Sm2 in 82, and Sm3 in 36. Local recurrence developed in 0, 4 (5%), and 0 patients and lymph node metastasis in 0, 4 (5%), and 9 (25%) patients, respectively. None of the 64 patients with Sm1 invasion developed a local recurrence or lymph node metastasis despite the presence of lymphovascular invasion in 30% and poorly differentiated histology in 12.5%. If other studies show similar data, Haggitt level 4, Sm1 invasive cancers arising in a polyp may be optimally treated by polypectomy alone.