Scenario 1: Patients Unfit for/Unwilling to Undergo Major Surgery

This group includes patients who decline surgery or are not candidates for operation. Most of these patients are not surgical candidates due to high operative risk related to frailty and excess comorbidity or advanced primary or metastatic disease, precluding curative resection. The primary aim in these cases is most often local control rather than long-term survival. The aim of treatment is palliation. To facilitate these goals, many centers prefer to give a higher radiotherapy dose than is delivered in the neoadjuvant setting. The assessment of response, usually by clinical examination and/or CT imaging, is imprecise because it is of no therapeutic consequence. Thus, retrospective analysis of the rate of complete response is not highly accurate, and comparing these reports is unreliable.

The largest series in the literature is from the Princess Margaret Hospital in Toronto, reporting retrospectively on 271 patients, treated from 1978 to 1997 with external beam radiation therapy doses ranging from 40 to 60 Gy. A complete clinical response, defined as no gross tumor on endoscopy or digital rectal examination, was obtained in 30% of patients after a median time of 4 months. The local relapse rate was 78% after a median time of 18 months. Washington University in St. Louis reported on 199 patients treated from 1981 to 1995 with endocavitary contact radiotherapy, with or without external beam radiation, as primary therapy. These were patients of advanced age, with excessive comorbidity, or who refused to proceed with any resection requiring colostomy. The local control rate in this series was 71% after a median follow-up of 70 months. A series from Lyon, France, reports on 63 patients with intermediate-size tumors, treated with radiotherapy alone from 1986 to 1998. A high localized dose of radiation was obtained using a combination of endoluminal contact therapy and external beam and interstitial brachytherapy. The cCR rate was an impressive 92% at 2 months after treatment, with a local control rate of 63% at 54 months. A Polish series, which used high doses (55–70 Gy) of conformal radiotherapy, with or without chemotherapy, in primary inoperable and recurrent rectal cancer, noted a cCR of 39%. In addition to these series, there are several smaller reports that mostly include patients with large and inoperable tumors treated with standard external beam radiotherapy, with or without chemotherapy. The rate of complete clinical response in these series varied from 10% to 50%, indicating a heterogeneous group of tumors.

Although it is not easy to draw conclusions from these series, some observations can be made. First, the definition and assessment of clinical response in these patient cohorts is highly variable. Second, maximal response to radiation, with or without chemotherapy, can take up to 4 to 6 months. Finally, these series confirm the basic rule of radiotherapy: that cCR and long-term local control are more likely in the setting of higher doses and smaller tumors.

Scenario 2: Patients with Early Rectal Cancer

This group includes patients with clinically small and superficial tumors who are attempting to avoid radical surgery. Common approaches include local excision and/or radiation. Failure rates after local excision, transanal excision, and transanal endoscopic microsurgery (TEM) without radiation range from 5% to 20%. Undetected and untreated locoregional metastasis is hypothesized to result in a significant proportion of local recurrences. Pelvic radiation (external beam and contact) is an alternative approach with a potential to treat the primary tumor and locoregional lymph nodes. Described by Papillon in France in the 1950s, endocavitary contact therapy delivers repetitive high doses to the luminal surface of a small distal rectal tumor. The result in early tumors is impressive, with a recent review reporting long-term local control rates of 85% to 90% in more than 1000 patients with clinical T1N0 tumors. The success has led investigators to use this approach to avoid major surgery in more advanced rectal cancers. In a 2008 review, Borschitz and colleagues reported a long-term local recurrence rate of 7% in 237 cT2/3 tumors treated with chemoradiation and local excision; 22% of patients had a pCR, and recurrence rates were directly associated with residual disease: 0% for ypT0, 2% for ypT1, 7% for ypT2, and 21% for ypT3 tumors. A report from the United Kingdom describes 220 patients with early rectal cancer receiving a combination of internal (contact) and external radiotherapy and local excision, with curative intent; 10% of patients showed poor response and underwent immediate rescue surgery. Of the remaining 90%, only 10% had late local recurrence, indicating that rectal preservation is possible in a significant subset of patients.

The common theme of these studies is that smaller tumors sustain greater response to chemoradiation. When considering all patients, major surgery may be avoidable in up to 50%. The dilemma remains, however: Does the benefit to patients who respond to chemoradiation and avoid surgery outweigh the added toxicity and potential of overtreatment? This cohort of small and early rectal cancers would have excellent prognosis if treated with surgery alone. Maximizing response to increase the proportion of patients that achieves a cCR and avoids surgery would make chemoradiation more attractive.

Scenario 3: Patients with Locally Advanced Rectal Cancer Treated with Neoadjuvant Chemoradiation and Planned Surgery

The largest series of patients in this category is reported in several articles by Habr-Gama and colleagues, and was summarized in a recent overview. These studies involve generally fit patients with mid- to distal rectal cancers, for which the standard treatment is radiation with concurrent fluoropyrimidine, followed by total mesorectal excision. If patients had a cCR that was sustained for a year, they were entered into the study and managed nonoperatively. The proportion of sustained cCRs was reportedly approximately 30%, suggesting that a substantial number of less advanced tumors were included. The largest study from the Sao Paulo group was reported in 2006 and includes 361 patients treated in two hospitals. Of 122 patients who had an initial cCR, 23 underwent rescue surgery within a year for early recurrence. Of the remaining 99 patients with sustained cCRs, 5 developed isolated local recurrence and 1 developed local and distant recurrence. With a mean follow-up of 60 months, 5-year overall survival was 93% and disease-free survival was 85%. The most recent update in 2011 reported on 173 patients treated at one of these two Sao Paulo centers, and confirms the excellent results: 96% overall survival and 75% disease-free survival at 5 years, with a mean follow-up of 65 months. Of the 67 patients (39%) who achieved an initial cCR, 8 (11%) recurred locally; all recurrences were endoluminal and amenable to curative salvage surgery. Nine underwent diagnostic local excision.

Other, smaller studies have reported their experiences with observation after chemoradiation for rectal cancer. A retrospective series from Memorial Sloan–Kettering Cancer Center reported on 32 patients who achieved a cCR and were managed nonoperatively. All patients were initially diagnosed with stage III disease or had a distal T2 tumor. Patients who declined surgery were older, had comorbidity, and/or had a strong desire to avoid the long-term morbidity associated with major surgery. After a median follow-up of 28 months, 6 patients suffered a local recurrence (5 endoluminal and 1 nodal), all within 14 months. The 3 patients with isolated local recurrence remain disease-free after salvage surgery. The overall survival rate was 96%, and similar to a comparison group of 57 patients with a pCR undergoing surgery.

A prospective series from Maastricht in the Netherlands described 21 patients with a cCR who were managed nonoperatively. With a median follow-up of 25 months, there was only 1 patient with an isolated local recurrence (after 22 months), which was amenable to salvage surgery. This compared well with a group of patients who had pCR after major surgery. As expected, there were fewer stomas and better bowel function after organ-preserving treatment. In this series, modern MRI techniques played an important role in patient selection and follow-up.

Another retrospective study from India reported on 33 patients who achieved a cCR, of a cohort of 291 (11%) patients with locally advanced or very distal rectal cancer treated with neoadjuvant therapy. Response was assessed by digital rectal examination and sigmoidoscopy at 4 to 6 weeks, and the definition of cCR allowed for minimal induration or mucosal irregularity; 23 of the 33 patients with cCR chose a watch-and-wait policy, the majority because of a wish to avoid a permanent stoma. Seven patients (30%) had local recurrence, 5 of these within the first year. A smaller series from the United Kingdom evaluated 49 patients with locally advanced rectal cancer treated with neoadjuvant therapy. Of the 6 patients with a cCR who opted to avoid surgery, none has recurred.

In contrast to these studies reporting good outcomes, two older studies from another center in Sao Paulo noted high rates of local recurrence after nonoperative management in a small number of patients with cCR after chemoradiation (with and without brachytherapy) for locally advanced distal rectal cancer. The reason for these divergent results is unknown, but may be a result of patient selection. Similar to other series, however, a majority of local recurrences were identified within the first year.

The use of transanal excision of the primary site after neoadjuvant therapy, in order to help identify cCR, was examined in a study of 174 patients from Israel. All patients had locally advanced cancers of the mid- and lower rectum that responded well to neoadjuvant therapy. Digital examination, proctoscopy, and endorectal ultrasound identified 68 patients (39%) with possible cCR; 31 underwent local excision, yielding 23 with ypT0NX. All other patients underwent a major resection. With a long follow-up of 87 months, there were no local recurrences. Although this may constitute a compelling argument for using local excision to help identify cCR after chemoradiation, the approach is not without complications, including delayed healing and extended perioperative discomfort.

Although the studies are heterogenous, with varying inclusion criteria, definitions of cCR, and modes of follow-up, some common themes exist. Recurrence after cCR most often presents within the first year. Salvage appears possible in the majority of cases. At this juncture, however, it must still be assumed that improvement in quality of life via nonoperative strategies is a trade-off, associated with some degree of increased oncologic risk. Many experts have recently written reviews and commentaries advising caution, noting that it is too early to propose nonoperative management as a standard treatment.