Key points

Introduction

The incidence of rectal cancer is estimated to be 39,220 in the United States in 2016, with an estimated 49,190 deaths from combined colon and rectal cancer. Although surgery remains the primary definitive treatment of colorectal cancers, additional local treatment with radiotherapy is indicated in many patients because of the anatomy of the rectum and pelvis. The limited space within the pelvis can make complete surgical resection more difficult for rectal tumors while at the same time providing a fixed target for radiation treatments. The role of radiotherapy as adjunct to surgery has evolved over the decades with changes in the timing (preoperative vs postoperative), length (short course vs long course), intent (neoadjuvant vs definitive), and delivery (3-dimensional conformal radiation therapy [3D-CRT] vs intensity-modulated radiation therapy [IMRT]), significantly improving the outcomes of patients. This review summarizes the evolution in radiation therapy (RT) for the management of rectal cancer and addresses some of the current questions regarding its role in the future.

Introduction

The incidence of rectal cancer is estimated to be 39,220 in the United States in 2016, with an estimated 49,190 deaths from combined colon and rectal cancer. Although surgery remains the primary definitive treatment of colorectal cancers, additional local treatment with radiotherapy is indicated in many patients because of the anatomy of the rectum and pelvis. The limited space within the pelvis can make complete surgical resection more difficult for rectal tumors while at the same time providing a fixed target for radiation treatments. The role of radiotherapy as adjunct to surgery has evolved over the decades with changes in the timing (preoperative vs postoperative), length (short course vs long course), intent (neoadjuvant vs definitive), and delivery (3-dimensional conformal radiation therapy [3D-CRT] vs intensity-modulated radiation therapy [IMRT]), significantly improving the outcomes of patients. This review summarizes the evolution in radiation therapy (RT) for the management of rectal cancer and addresses some of the current questions regarding its role in the future.

From surgery alone to combined modality therapy

Historically, the primary management of rectal carcinoma was surgical resection, with rates of local failure of 25% or higher with older surgical techniques. It was also observed that positive resection margin resulted in significantly worse local failure rates. The advent of total mesorectal excision (TME), which uses sharp dissection of mesorectal contents, further reduced the rate of local recurrence to as low as 8%. For patients with higher risk disease (such as T3/T4 disease and node positivity), the risk of local and distant failure remained elevated, so adjuvant therapies including radiation and chemotherapy were administered in an attempt to reduce rates of local and distant failure. Multiple studies examined this adjuvant role of pelvic radiation and collectively found that adjuvant radiation alone decreased the risk of local recurrence but did not significantly improve overall survival. The addition of adjuvant chemotherapy was noted to reduce the rates of distant metastases while improving disease-free survival. Since that time, more modern trials have examined the appropriate sequencing of therapies (surgery, radiation, chemotherapy) and assessed which patients may be candidates for omission of certain modalities to reduce the late effects of treatment.

Preoperative radiation therapy regimens

In the 1980s and 1990s, the standard management of rectal cancer was surgical resection with low anterior resection (LAR) or abdomino-perineal resection for proximal and distal tumors, respectively, followed by adjuvant radiation (with or without chemotherapy) and further chemotherapy. For some patients, it was difficult to complete the entire postoperative course of adjuvant treatment, which led to the emergence of several trials to address the question of whether neoadjuvant radiation therapy (RT) with or without chemotherapy could be safely administered without sacrificing disease control. The German Rectal Cancer Study CAO/ARO/AIO-94 randomly assigned more than 800 patients to treatment with either neoadjuvant RT to 5040 cGy with continuous infusion 5-fluorouracil (5-FU) or adjuvant RT to 5580 cGy with 5-FU. Both groups underwent TME resection and 4 cycles of adjuvant 5-FU. With the long-term update, neoadjuvant chemoradiation (chemo-RT) was found to decrease local recurrence (7.1% vs 10.1%; P = .048); however, there were no differences in overall survival [OS] rate (59.6% vs 59.9%; P = .85) or distant metastases (29.8% vs 29.6%; P = .9). Pathologic complete response (pCR) rate in the preoperative group was 9% as of the long-term update. In the initial publication, acute grade 3 or higher toxicity, an important endpoint, was significantly less in the preoperative group (27% vs 40%; P = .001), and late grade 3 or higher toxicity was also decreased with preoperative therapy (14% vs 24%; P = .01).

Two additional trials, the European Organisation for Research and Treatment of Cancer (EORTC) 22,921 and the Federation Francophone de Cancerologie (FFCD) 9203, examined the effect of neoadjuvant chemo-RT over RT alone (with or without adjuvant chemotherapy) on outcomes for patients with cT3-4 resectable disease. Pooled analysis of the data found that chemo-RT significantly improved 3-year local control (92.3% vs 84.7%; P <.0001) and pCR rates (11.2% vs 3.7%; P <.0001). However, there were no differences in 5-year OS (66.3% vs 65.9%; P = .66) or 3-year distant progression-free rates (71.3% vs 70.7%; P = .5). Based on the totality of the above data, neoadjuvant chemo-RT has become the current standard of care in the United States and many regions of Europe.

In contrast to the neoadjuvant chemo-RT approach with 5 to 6 weeks of daily RT (long course), other countries have assessed whether preoperative short-course RT alone is sufficient to reduce the local recurrence risk in stage II and III rectal cancers. The Swedish Rectal Cancer Trial investigated whether preoperative short-course RT of 25 Gy in 5 fractions followed by surgery 1 week later improved outcomes over surgery alone. TME was not mandated in this trial. On long-term follow-up, short-course RT improved OS (38% vs 30%; P = .008) and reduced local recurrence (9% vs 26%; P <.001). One of the criticisms of this trial is that TME was not mandated, which may have resulted in worse outcomes in the control group as demonstrated by the high rate of local recurrence. The Dutch TME trial therefore sought to address the benefit of short-course RT in patients undergoing TME by comparing preoperative RT, 25 Gy/5 fractions, with surgery alone (although patients with positive margins received mandatory adjuvant RT). Long-term follow-up found that short-course RT reduced the 10-year local recurrence rate (5% vs 11%; P <.0001) but had no effect on 10-year OS (48% vs 49%; P = .86). However, nearly two-thirds of patients on this study had stage I/II disease for which radiation may not provide a benefit, so subset analysis was performed and discovered that patients with stage III and negative circumferential margins had increased OS with short-course RT (50% vs 40%; P = .032).

Two additional trials addressed whether preoperative short-course RT results in inferior outcomes to those of standard long-course chemo-RT. Bujko and colleagues randomly assigned more than 300 patients with cT3/T4 disease to preoperative short-course RT, 25 Gy/5 fractions, followed by surgery within 7 days or to preoperative long-course chemo-RT, 50.4 Gy/28 fractions, with 5-FU/leucovorin followed by surgery 4 to 6 weeks later. No differences were detected in 4-year local recurrence rates (9% vs 14.2%; P = .17), OS (67.2% vs 66.2%; P = .96), disease-free survival (58.4% vs 55.6%; P = .82), or late toxicity (10.1% vs 7.1%; P = .36) for short- versus long-course RT, respectively. Notably, short-course RT showed lower rates of pCR (0.7% vs 16.1%; P <.001) and increased rates of positive circumferential margins (12.9% vs 4.4%; P = .017) but no difference in sphincter preservation (61.2% vs 58%; P = .57). In a similar trial design, the Trans-Tasman Radiation Oncology Group (TROG) 01.04 randomly assigned more than 300 patients with cT3N0-2 disease less than 12 cm from anal verge to short-course RT, 25 Gy/5 fractions, followed by surgery and 6 cycles of adjuvant chemotherapy versus long-course RT, 50.4 Gy/28 fractions, with 5-FU and 4 additional cycles of adjuvant chemotherapy. Although patients were randomly assigned, there were more patients with distal tumors (<5 cm from anal verge) in the short-course RT arm (30% vs 19%) and fewer patients with proximal tumors (10–12 cm) in the short-course arm (16% vs 26%). No differences were detected in 3-year local recurrence (7.5% vs 4.4%; P = .24), 5-year OS (74% vs 70%; P = .62), 5-year distant recurrence (27% vs 30%; P = .92), or late toxicity (5.8% vs 8.2%; P = .53). Similar to the Polish trial, pCR rates were lower in the short-course group (1% vs 15%; P value not reported), and there were no differences in sphincter preservation rates (63% vs 69%; P = .22). Although not statistically significant, the actuarial rates of local failure were higher in patients with distal tumors who received short-course RT (12.5% vs 0.3%; P = .21), which the authors concluded may favor long course in patients with distal tumors.

Short-course RT has not been routinely adopted in the United States because of concern over the late gastrointestinal (GI) morbidity from hypofractionated RT. Although neither the Polish nor Australian trials observed a difference in late GI toxicity between short-course and long-course arms, there is concern that a median follow-up of 4 and 6 years, respectively, may not be sufficient time to observe the late effects of the hypofractionated treatment. Current National Comprehensive Cancer Network guidelines (v2.2016), however, do allow for the use of short-course RT in cT3 or cN1-2 patients but recommends against its use in T4 tumors or in patients for whom down-staging is desired.

Alternative neoadjuvant treatment strategies

Although the current standard of care for locally advanced rectal cancer is neoadjuvant RT (with or without chemotherapy) and surgical resection followed by adjuvant chemotherapy, several recent and ongoing clinical trials have addressed the sequencing of the adjuvant therapies. These studies were conducted with the intent to ensure completion of all therapy, improve pCR rates, and potentially reduce the late effects of treatment. Because it has been observed that pCR rates are associated with superior survival and outcomes, it is important to define treatment strategies that best accomplish this goal ( Table 1 ).

In the Spanish GCR-3 phase II randomized trial, patients with locally advanced cT3-T4 or cN+ disease were randomly assigned to chemo-RT using 5040 cGy with capecitabine/oxaliplatin (CAPOX) followed by TME and 4 additional adjuvant cycles of CAPOX or to the experimental arm of induction CAPOX followed by chemo-RT and then TME. There were no differences in 5-year local recurrence (2% vs 5%; P = .61), distant metastases (21% vs 23%; P = .79), OS (78% vs 75%; P = .64), or pCR rates (13% vs 14%; P = .94). This finding shows that induction chemotherapy (with a delay in chemo-RT and surgery) may be safely administered in lieu of postoperative adjuvant chemotherapy treatment.

Garcia-Aguilar and colleagues conducted a multi-institutional, nonrandomized phase II trial examining the effect of increasing preoperative chemotherapy cycles on pCR rates in patients with locally advanced rectal cancer. Patients in all arms received chemo-RT to 5040 to 5400 cGy with 5-FU followed by either (1) TME; (2) 2 cycles of modified leucovorin, 5-FU, and oxaliplatin (mFOLFOX6) then TME; (3) 4 cycles mFOLFOX6 then TME; or (4) 6 cycles mFOLFOX6 and TME. They observed that increasing cycles of neoadjuvant mFOLFOX6 resulted in higher pCR rates (groups 1–4: 18% vs 25% vs 30% vs 38%; P = .0036). Although lengthening the interval between completion of chemo-RT and TME did result in increased pelvic fibrosis at the time of TME, it did not adversely affect the ability to achieved R0 resections. Long-term follow-up with recurrence rates and survival outcomes have not been published at this time.

Preliminary results from the Chinese Neoadjuvant FOLFOX6 Chemotherapy with or without Radiation in Rectal Cancer (FOWARC) study have recently been published providing information on the role of neoadjuvant radiation in achieving pCR. Nearly 500 patients were randomly assigned to (1) standard chemo-RT with 5-FU and postoperative 5-FU, (2) chemo-RT with mFOLFOX6 and postoperative mFOLFOX6, or (3) neoadjuvant mFOLFOX6 alone (RT is omitted) and postoperative mFOLFOX6. pCR rates varied significantly between arms (groups 1–3: 14% vs 27.5% vs 6.6%; P = .005 for group 1 vs group 2). This trial concluded that radiotherapy provides a clear improvement in pCR rates that is further enhanced by intensified preoperative chemotherapy. There were no differences in R0 resection or sphincter preservation rates between groups. The addition of mFOLFOX6 to RT over 5-FU resulted in worse acute grade 3 to 4 GI toxicity (14.5% vs 7.7% diarrhea; P value not reported), but it did not impair the ability of patients to receive the full course of radiation (86.4% vs 90.5%; P value not reported). Recurrence and survival outcomes are pending.

The ongoing Alliance PROSPECT trial is investigating whether RT may be safely omitted in select LAR candidates with good response to induction FOLFOX6. Patients are randomly assigned to induction FOLFOX6 and assessed for response to the primary tumor; those with regression greater than 20% will undergo LAR, while those with poor response or regression will receive standard neoadjuvant chemo-RT with 5-FU followed by resection. The outcomes of this arm will be compared against the standard of care chemo-RT with 5-FU arm; both arms will receive adjuvant chemotherapy. This study population has more favorable disease than those in the Chinese FOWARC trial in that only T2N1 and T3N0-1 who will undergo LAR are eligible, so it will be interesting to determine whether the omission of neoadjuvant RT significantly decreases the ability to achieve pCR in this population.

Movement toward a total neoadjuvant therapy approach allows for novel clinical trial designs that may incorporate targeted systemic agents or radiation sensitizers with focus on improving pCR rates and reducing distant metastatic disease. NRG Oncology GI-002 ( clinicaltrials.gov ID 02921256) is a randomized phase II trial opened in October 2016 that compares a standard arm (mFOLFOX6 x 8 cycles followed by preoperative RT plus capecitabine and surgery) versus additional experimental arms that incorporate new systemic agents. The first planned experimental arm will add the polyADP-ribose inhibitor, veliparib, to RT plus capecitabine. The study populations are patients with locally advanced rectal cancer at high risk for distant disease meeting one of the following criteria: (1) cT3-4 and distal location (< 5 cm from anal verge), (2) bulky cT4 tumor within 3 mm of the mesorectal fascia, (3) cN2 disease, or (4) not a candidate for sphincter preservation before preoperative therapy. The primary endpoint is an improvement in pathologic Neoadjuvant Rectal Cancer score (total downstaging) compared with the control arm. Other agents under investigation for addition to neoadjuvant therapy include COX-2 inhibitors and statins after early preclinical and clinical work demonstrated poorer outcomes in tumors overexpressing COX-2 and improved outcomes in patients being treated with statins, respectively. Strategies such as this may allow further refinement of neoadjuvant therapies to improve the outcomes of patients with more aggressive disease.