Are We Overtreating Some Patients With Rectal Cancer?

Adjuvant therapy, almost by definition, overtreats patients. It is the holy grail of those of us involved in adjuvant therapy to define the patients who are going to fail so that we can decrease the incidence of tumor recurrence and avoid giving additional therapy to patients who have been cured by their primary treatment. Improved Patient Selection
What is obviously needed is a much-improved determination of which patients will fail with singlemodality therapy. The frustration with much of the work done to date is that we have developed a plethora of prognostic factors (pathologic, surgical, biologic), but the improvement generated by each of these factors is small. Conventional TNM staging criteria are enormously useful in determining recurrence risk and prognosis, but these parameters are crude. There is a vast emerging literature on the use of molecular markers of all types,[1-3] and many studies show a statistical impact from a single marker. However, just the existence of such a plethora of markers is a good indication that no single marker is very useful. Given the great variation of the known contributing factors to tumor initiation, progression, and dissemination, it is not surprising that a single molecular marker, or even a set of markers, will not provide the complete answer. What can we do to improve the situation? First, we must acknowledge that failures of therapy can always occur. It is usually far easier to find a reason to deliver an adjuvant therapy than it is to prove that it should not be used. Patients like to hear that as physicians, we are "doing everything" for them. Clinical trials that withhold potentially beneficial therapy-however scientifically justified-are notoriously difficult to accomplish from both a practical and ethical standpoint. Our responsibility as physicians is to make the difficult decision, ie, to justify the reasons for not treating a patient in order to reduce the morbidity and cost of overtreatment. Over the past several decades, we have been effective in reducing operative morbidity/ mortality and the complication rates associated with chemoradiation. The article by Rothenberger et al addresses a critical issue but unfortunately one without clear answers. As most clinicians are acutely aware, decisions may be evidence-based but must also be individualized; the recommendations for a healthy young patient may be different than those for an older patient with significant comorbidities. Disease Process and Treatment Failure
The first task of an accurate assessment is for the physician to understand the disease process. If the physician does not understand the difference between noninvasive and invasive cancer, it is impossible to have a meaningful discussion of failure risks. Patients with noninvasive tumors need a primary resection with negative margins, however that can best be accomplished, and no further therapy. When a colorectal tumor becomes invasive, the situation is more complicated. Small tumors in polyps on a stalk with minimal invasion are almost certainly handled well by polypectomy. However, as pointed out by Rothenberger and colleagues, for sessile T1 and (especially) T2 tumors, local recurrence rates are higher than initially expected after local excision alone; additional therapy, either radical surgery or chemoradiation, is appropriate.[ 4] We anxiously await molecular profiles that may help to define those patients most likely to develop a recurrence. We completely agree with Rothenberger et al that adjuvant radiation therapy and chemotherapy is not usually necessary for tumors that are neither through the bowel wall (T3) nor node-positive. The problem is identifying these patients reliably, especially as evidence converges that radiochemotherapy given preoperatively is more beneficial than when given postoperatively. Conventional computed tomography (CT) and magnetic resonance imaging (MRI) as preoperative staging modalities are only 50% to 70% accurate in determining T and N stage of a rectal tumor.[5,6] MRI with endorectal coils has an accuracy nearing 80% in experienced hands, but is not generally available.[7,8] Currently, the best tool for determining T and N stage for rectal malignancies remains endorectal ultrasonography (EUS). Using EUS as an extension of the physical examination and rigid proctoscopy, experienced clinicians have reported up to 80% to 90% accuracy in the preoperative assessment of the T stage and 70% to 80% for N stage [9-11]; however, the technique is highly operatordependent with a significant learning curve. Overall, studies suggest an overstaging rate of more than 20%.[12,13] Stage II/III Patients
Given the difficulty in accurately assessing the preoperative stage of rectal cancers, the challenge remains how to apply the current evidence to avoid overtreating patients with presumed stage II and III cancers. Risk stratification from several large Intergroup trials leads to a better understanding of those patients at a relatively low risk of locoregional failure. These include patients who meet the following criteria: (1) stage T1/2, N1 and T3, N0 tumors,[14,15] (2) tumors located high in the rectum, (3) tumors operated on by experienced surgeons trained in total mesorectal excision (TME), (4) those with wide circumferentially negative surgical margins,[ 16] and (5) adequate nodal evaluation by the pathologist (at least 12 to 14 nodes examined).[17] It is important to understand that an experienced surgeon is an independent variable for local recurrence. As discussed by Rothenberger et al, no one has been as steadfast as Heald in asserting that improving surgical techniques will have an impact on outcomes in rectal cancer.[18] This assertion is supported by the low recurrence rates (8.2% in all patients, including those with stage 0 and I tumors) after TME alone in the Dutch TME trial[19]; the Dutch TME trial succeeded in standardizing operative techniques. However, the unfortunate truth is that many ultrasonographers and surgeons lack experience and training, many pathologists do not fully evaluate the specimen either for nodes or margins, and many treating clinicians do not carefully evaluate all of these factors. The rising costs of specialized care and multidisciplinary approaches will continue to fuel the debate as to whether such complex diseases as rectal cancer should be limited to tertiary referral centers. Translational Research
Part of the future of adjuvant therapy will be to define the selection criteria for patients who will most likely benefit-those at highest risk for recurrent disease. This is the emphasis and the critical message of the article by Rothenberger and coauthors. Translational scientific strategies will be critical to the development of a better understanding of the biology of disease. However, we must be careful not to go so far as to avoid necessary adjuvant therapy. The data from the Dutch TME trial demonstrate that the local recurrence rate with TME alone is approximately 20% at 4 years in patients with node-positive disease[ 19]; these patients clearly need additional therapy. The recent report by Sauer et al of a German trial comparing pre- to postoperative radiochemotherapy strongly suggests that this adjuvant therapy should be given preoperatively.[12] There will never be a clear dividing line, because we will likely never find a means of defining patients who have (or are likely to have) a single clonogen remaining after surgery; we can only determine risk of recurrence. Conclusions
Although this article by Rothenberger and others like it focus on local recurrence and overall survival rates as end points, we tend to underestimate the impact of local recurrence on the patient. In addition to the pain and dysfunction associated with a locally recurrent rectal tumor, it is difficult to quantitate the socioeconomic and psychological effects on the patient; what clinicians may deem as over/undertreatment may not be viewed in a similar light by patients. Clearly, it is all about patient selection and informed communication. However, the considerations expressed by Rothenberger and colleagues in this commentary will help decrease the number of rectal cancer patients who are overtreated while assuring that those at high risk for failure receive the treatments that will give them a better chance of longterm cure.