Introduction

For patients with locally advanced rectal cancer, long-course chemoradiation therapy (LCCRT) and short-course radiation therapy (SCRT) both represent valid neoadjuvant approaches based on multiple studies supporting these regimens over the past several decades. LCCRT generally consists of treatment to 45–54 Gy administered over 25–30 fractions with concurrent 5-fluorouracil (5-FU)-based chemotherapy, followed by surgery several weeks later. SCRT generally consists of treatment to 25 Gy administered over five fractions, without concurrent chemotherapy, with surgery either within 1 week or delayed until several weeks later.

Despite evidence supporting both approaches, the use of SCRT in the USA has been limited. Indeed, a recent analysis of the National Cancer Data Base (NCDB) indicated that while use of neoadjuvant radiation for rectal cancer significantly increased from 2004 to 2012, the frequency of administration of short-course radiation remained < 1% over this period [1]. Survey data of practicing radiation oncologists confirm a general hesitation towards SCRT for rectal cancer, with radiation oncologists citing concerns of insufficient downstaging, greater acute and late toxicities, and difficulty coordinating with surgical teams when administering SCRT, along with a desire for longer follow-up from SCRT trials before routine incorporation into practice [2].

Nevertheless, each regimen has its relative merits, and it is therefore critical that radiation oncologists who routinely treat rectal cancer have fluency in both approaches, particularly given emerging data supporting neoadjuvant strategies that incorporate short-course radiation. Herein, we review the key clinical literature supporting short-course radiation, and ongoing research in this field.

Historical Use of SCRT

SCRT was used in three landmark clinical trials that helped establish the role of neoadjuvant radiation for rectal cancer. From 1987 to 1990, the Swedish Rectal Cancer Trial randomized 1168 patients with operable rectal adenocarcinoma to surgery alone or neoadjuvant SCRT followed by surgery within 1 week [3]. At the time, it was unclear if the risk reduction in recurrence rates from use of neoadjuvant radiation would outweigh its potential adverse effects. Radiation in this study was delivered to 25 Gy in five fractions using a three-field or four-field technique, with the L4/L5 interface serving as the superior border, although protocol guidelines were violated in 9% of patients who were treated with a two-field approach. After following patients for a minimum of 5 years, the group reported significant improvements in local recurrence, cancer-specific survival, and overall survival, which persisted after a median follow-up of 13 years [4, 5]. While post-operative mortality within 30 days of surgery was no different between the two arms (3 vs. 4%, p = 0.31), there was a significant increase in post-operative mortality for patients treated with a two-field technique as compared to patients treated per protocol (15 vs. 3%, p < 0.01) [3]. Furthermore, subsequent analyses of hospital admissions data showed a higher rate of late small bowel obstruction amongst patients receiving neoadjuvant radiation (13.9 vs. 5.5%, p < 0.01) [6, 7]. Additionally, survey data of patients who lived at least 5 years indicated that those randomized to the pre-operative radiation arm experienced worse long-term bowel function, including urgency, incontinence, and emptying difficulty, which was associated with impairment of social functioning [8]. Taken together, this trial showed clear oncologic benefit with neoadjuvant short-course radiation, with improvement in local control translating into a survival advantage, albeit with appreciable late toxicities relative to surgery alone. Notably, large treatment volumes with limited treatment field number appeared to be correlated to worse normal tissue toxicity in other early studies of SCRT as well [9].

Notably, total mesorectal excision (TME), which is now the standard oncologic surgery for rectal cancer, was not mandated in the Swedish Rectal Cancer Trial resulting in a high rate of local recurrences seen in the surgery-alone arm (27%). The value of radiation therapy in the context of optimal oncologic surgery was unclear. As such, a Dutch study was conducted from 1996 to 1999 in which 1805 patients with operable rectal adenocarcinoma were randomized to TME with or without neoadjuvant SCRT. Local recurrences were significantly less common in the Dutch study than in the Swedish study, resulting in no difference in overall survival between study arms. Nevertheless, a local recurrence benefit was still seen in patients treated with neoadjuvant radiation, with a roughly 50% relative risk reduction which persisted at 12 years of follow-up (5 vs. 11%, p < 0.0001) [10, 11]. While patients who received pre-operative radiation experienced slightly increased blood loss, a higher incidence of perineal wound complications, and worse late bowel and sexual function, this did not contribute significantly to overall health-related quality of life [12, 13]. Additionally, both the hospitalization rate and the incidence of bowel obstruction were not significantly different between arms [14]. Notably, unlike the Swedish study, the superior field border of the radiation field in the Dutch study was at the sacral promontory and protocol violations were less frequent.

Whether radiation could be selectively reserved for patients with a close or positive circumferential resection margin (CRM) was analyzed in the Medical Research Council (MRC) CR07 study, in which patients with operable rectal adenocarcinoma were randomized to routine SCRT followed by TME within 10 days of the first fraction of radiation or to selective post-operative chemoradiation for patients with involvement of the CRM, defined as tumor < 1 mm from the margin. At 5 years, patients randomized to routine neoadjuvant radiation experienced an improvement in both local recurrence (4.7 vs. 11.5%, p < 0.0001) and disease-free survival (73.6 vs. 66.7%, p = 0.013) [15]. Neoadjuvant SCRT was associated with impact on sexual and bowel function [16]. Similar field design was used in the MRC study as was used in the Dutch study.

The above studies demonstrated the value of SCRT for reducing local recurrence in patients with operable rectal cancer. While significant late bowel obstruction was reported in the Swedish study, less late bowel toxicity was observed in the Dutch and MRC studies. Notably, while all stages of resectable rectal cancer were included in the aforementioned trials, low absolute rates of local recurrence were seen in early-stage patients who underwent TME, suggesting restriction of neoadjuvant radiation to patients with locally advanced disease. It is now generally agreed that radiation should be limited to stage II–III patients. Ongoing efforts to further risk stratify stage II–III patients and delineate those who benefit most from neoadjuvant radiation are beyond the scope of this review [17].

In parallel with the Dutch and MRC studies, a German study randomized patients with stage II–III rectal adenocarcinoma to neoadjuvant chemoradiation to 50.4 Gy with concurrent 5-fluorouracil (5-FU) followed by TME 6 weeks later versus post-operative chemoradiation to 55.8 Gy with concurrent 5-FU [18]. In this study, patients receiving neoadjuvant therapy experienced improved local recurrence, downstaging, sphincter preservation, and acute and late toxicity, as compared to patients treated with post-operative chemoradiation. The German study set the modern standard for neoadjuvant LCCRT, but its relative merits as compared to short-course radiation remained unclear.

Short-Course Versus Long-Course Chemoradiation

The first randomized trial comparing short-course and long-course radiation was conducted by the Polish Colorectal Study Group [19,20,21]. Eligible patients included those with clinical stage T3 or T4 operable rectal adenocarcinoma. As MRI was not readily available at participating institutions, staging was primarily based on clinical exam with “circular” or “tethered” lesions defined as cT3-4 lesions. Patients with freely moveable tumors underwent endorectal ultrasound and/or CT to exclude early-stage tumors, while patients with fixed lesions or sphincter involvement were excluded. The study randomized patients to neoadjuvant SCRT (25 Gy in five fractions), with surgery performed within a week from the end of radiation, or chemoradiation to 50.4 Gy in 28 fractions, with concurrent 5-FU/leucovorin administered during the first and fifth weeks of therapy and with surgery performed 4–6 weeks thereafter. TME was mandated, and sphincter preservation was to be determined based on the tumor status at time of surgery, with a recommendation for a 1-cm distal margin. Sphincter preservation rate was the primary endpoint of the trial.

Patients who underwent LCCRT experienced higher levels of acute grade 3+ toxicity as compared to patients treated with SCRT (18 vs. 3%, p < 0.001), but no significant differences in post-operative complications or late toxicities were seen [20, 21]. Patients treated with LCCRT were more likely to experience a clinical complete response (13 vs. 2%, p < 0.001) and a pathologic complete response (16 vs. 1%) as compared to patients in the SCRT arm [19]. The low pCR rate in the SCRT arm at least in part reflected the short interval from completion of radiation to surgery; times of 1 week or less are not sufficient to allow for the histological impact of radiation to manifest. Despite the difference in clinical response between the two arms, there was no difference in rate of sphincter preservation. Whether this discrepancy was due to non-uniformity in the practice of participating surgeons to base sphincter preservation off of tumor status at the time of surgery, despite stipulation in the protocol, is unclear. Importantly, patients who underwent LCCRT were less likely to have positive surgical margins (4 vs. 13%, p = 0.017). Nevertheless, at a median of 48 months of follow-up, there was no difference in overall survival, disease-free survival, or local control. Of note, there was an imbalance between study arms in the administration of adjuvant chemotherapy, with more frequent administration in the short-course radiation arm (46.4 vs. 30.1%, p = 0.005). Given the lack of difference in oncologic outcomes between the two study arms, the authors concluded that short-course radiation should be the primary consideration in Poland for locally advanced rectal cancer due to its improved acute toxicity profile, decreased cost, and increased convenience.

The Trans-Tasman Radiation Oncology Group (TROG) conducted a subsequent trial comparing SCRT with immediate surgery within 1 week with LCCRT with surgery 4–6 weeks later, with all patients receiving TME as standard surgery [22]. This study was restricted to cT3 patients as defined by endorectal ultrasound or MRI, providing a more uniform study population than the Polish study. Additionally, adjuvant chemotherapy was standardized, consisting of 4 and 6 months of 5-FU/leucovorin in the long-course and short-course arms, respectively. Similar to prior studies, a three-field or four-field technique was used to cover a clinical target volume (CTV) that included the primary tumor as well as the mesorectal, presacral, pelvic sidewall, and internal iliac nodes, with the superior border at the sacral promontory. After a minimum follow-up of 3 years and a median follow-up of 5.9 years, there was no significant difference in distant recurrence or overall survival between the two groups, and late toxicity profiles were also similar. Local recurrence rates at 3 years, the primary endpoint of the study, were similar as well (SCRT 7.5% vs. LCCRT 4.4%, p = 0.24). Although this difference did not reach statistical significance, the confidence interval (− 2.1 to 8.3%) was broad. Moreover, when examining tumors < 5 cm from the anal verge, there was a non-significant albeit numerically large difference in local recurrence risk between the two arms (SCRT 12.5% vs. LCCRT 3.2%, p = 0.21). As expected, patients in the long-course chemoradiation arm experienced a higher rate of pathologic complete response (15 vs. 1%, p < 0.001), but there was no difference in sphincter preservation for lower tumors. There was also no difference in the rates of severe late toxicities between the two arms.

The TROG study authors concluded that while no differences in outcomes were noted between study groups, LCCRT may be considered for distal and/or bulky tumors that would benefit from downstaging, but that SCRT should be otherwise considered. Both the Polish and TROG studies support increased downstaging with LCCRT and a delay to surgery, compared to SCRT with surgery within 1 week. Importantly, no significant differences in late normal tissue toxicities have been reported from these studies, an important consideration given the contribution of late toxicity concerns to the lack of adoption of short-course radiation in the USA.

SCRT With Delayed Surgery

The Stockholm III trial compared three treatment groups: SCRT with surgery within 1 week, SCRT with a delay to surgery (of 4–8 weeks), or long-course radiation therapy (without concurrent chemotherapy). The study was designed as a non-inferiority trial comparing the two regimens with delayed surgery to SCRT with early surgery [23, 24•, 25]. Following a protocol amendment, centers in Sweden could choose to randomize patients to only the two SCRT arms rather than all three arms.

Results from this study were published after a minimum follow-up of 2 years [23]. Results were analyzed separately for 385 patients enrolled in the three-arm randomization. Additionally, 455 patients who were enrolled on the two-arm short-course radiation randomization were pooled with patients who received short-course radiation in the three-arm randomization and also analyzed separately. In the three-arm randomization, there was no difference in local recurrence, distant metastasis, or overall survival between the groups. Similarly, no difference in these endpoints was seen in the pooled analysis of the short-course radiation arms. In the three-arm randomization, no patients in the short-course radiation with immediate surgery arm experienced grade 3+ acute toxicity, whereas 6 and 5% of patients in the short-course radiation with delayed surgery arm and long-course radiation with delayed surgery arm experienced grade 3+ toxicity. In the pooled analysis of the short-course radiation arms, the rate of grade 3+ acute toxicity was 7% in patients undergoing delayed surgery versus < 1% in patients undergoing early surgery (p < 0.0001). However, in the pooled analysis, patients undergoing early surgery experienced a higher incidence of both surgical complications (36 vs. 28%, p = 0.03) and any post-operative complication (53 vs. 41%, p = 0.001). While these differences in surgical and post-operative complications were not statistically significant in the three-arm randomization, complication rates were numerically higher in the short-course radiation with immediate surgery arm. Notably, an early analysis suggested a significant difference in pathologic complete response between the two short-course radiation arms (delayed surgery 11.8% vs. early surgery 1.8%, p = 0.001) [25].

The Stockholm III trial illustrated a number of important points. First, SCRT with delayed surgery was feasible, increased downstaging as compared to SCRT with early surgery, and resulted in similar oncologic efficacy as compared to both SCRT with early surgery and LCRT with delayed surgery. Moreover, while less acute toxicity was experienced with SCRT with early surgery, the reduction in acute toxicity came at the cost of increased post-operative complications.

Short-Course Radiation and Neoadjuvant Systemic Therapy

Given the high local control rates with modern neoadjuvant radiation and TME, distant relapse predominates over local relapse for patients with locally advanced rectal cancer. Neoadjuvant systemic therapy may offer the ability to address micrometastatic disease earlier in the treatment course and reduce distant relapse while also contributing to tumor downstaging and local control. Indeed, significant local impact of systemic therapy, as measured by pathologic CR rate, has been demonstrated in the GEMCAD 0801 study as well as in a report from Memorial Sloan Kettering Cancer Center [26, 27].

The delay between neoadjuvant radiation therapy and surgery presents an opportunity for the delivery of systemic therapy. A study from Poland randomized patients with locally advanced rectal cancer to either SCRT followed by three cycles of FOLFOX and surgery or LCCRT to 50.4 Gy in 28 fractions with concurrent oxaliplatin and boluses of 5-FU/leucovorin followed by surgery [28]. Eligible patients for this study had cT4 tumors on CT or MRI or palpably fixed cT3 tumors. Administration of concurrent oxaliplatin in the long-course chemoradiation arm was based on retrospective data at the time supporting its use, prior to multiple subsequent prospective studies which showed no oncologic benefit and increased toxicity with concurrent oxaliplatin [29,30,31]. After a median follow-up of 35 months, no differences were reported in local recurrence or distant recurrence. Despite this, a difference in overall survival was seen in favor of the short-course radiation arm. No differences were noted between arms in post-operative complications or late toxicity. Overall acute toxicity was slightly improved in the short-course radiation arm, although grade 3+ acute toxicity was no different. Washington University investigators initiated a phase II study of short-course radiation followed by four cycles of FOLFOX for cT3 or cT4 patients, with T-stage based on endorectal ultrasound, MRI, or clinical exam findings of a tethered lesion [32•]. As opposed to prior SCRT protocols, intensity-modulated radiation therapy (IMRT) was used in an attempt to minimize toxicity. Additionally, a simultaneous integrated boost (SIB) technique was employed in which 25 Gy was administered to the primary tumor, involved nodes, and entire mesorectum 1 cm above and below gross tumor involvement and 20 Gy was administered to elective nodes, including the full mesorectum, presacral nodes, and internal iliac nodes. At a median follow-up of 26 months, local control and freedom from relapse were 95 and 87%, respectively, for the 76 enrolled patients. Pathologic complete response for the overall cohort was 25%. Amongst favorable risk locally advanced patients who met eligibility criteria for the PROSPECT trial, namely patients with tumors > 5 cm from the anal verge, <cT4 disease, <cN2 disease, and no disease within 3 mm of the mesorectal fascia, the pathologic complete response rate was 39% [33].

More recently, the same group of investigators compared outcomes from the aforementioned phase II trial to a cohort of matched historical controls from their institution as well as Stanford University who underwent long-course chemoradiation followed by surgery and adjuvant chemotherapy [34•]. Despite being well-matched for key prognostic factors, patients who underwent SCRT followed by neoadjuvant systemic therapy experienced higher rates of pathologic response and higher rates of distant metastasis-free survival and disease-free survival.

Ultimately, randomized controlled trial data will be needed to definitively compare SCRT followed by neoadjuvant chemotherapy versus LCCRT. The RAPIDO study is specifically looking at this comparison in high-risk locally advanced patients, with eligible patients having cT4 disease, cN2 disease, extramural vascular invasion, mesorectal fascial involvement, or positive lateral lymph nodes on MRI [35]. Treatment on the standard arm of the RAPIDO trial consists of 50.4 Gy in 28 fractions or 50 Gy in 25 fractions, administered with concurrent capecitabine, followed by surgery with TME 6–8 weeks after completion of radiation, with a recommendation for eight cycles of adjuvant capecitabine and oxaliplatin (CAPOX). Treatment on the experimental arm consists of short-course radiation to 25 Gy in five fractions, followed by six cycles of CAPOX, without administration of adjuvant chemotherapy. Long-term oncologic and toxicity outcomes from this study are awaited. The STELLAR trial is an additional phase III study being conducted in Asia in which all stage II/III patients are eligible and are randomized to either long-course chemoradiation to 50 Gy in 25 fractions with concurrent capecitabine, followed by surgery with TME 6–8 weeks later, with six cycles of adjuvant CAPOX, or short-course radiation to 25 Gy in five fractions, followed by four cycles of CAPOX, surgery 4 weeks later, and two additional cycles of adjuvant CAPOX. Early results from STELLAR show no difference between the two groups with respect to surgical complications, with an improved pathologic CR rate in the experimental group [36].

Candidacy for SCRT

LCCRT has been the standard treatment for rectal tumors with the most adverse prognostic signs on imaging, including patients with involved radial margins, involvement of unresectable structures, and/or lateral lymph node involvement [37]. SCRT with early surgery, as we have noted, is not associated with significant tumor downstaging and this may have influenced decisions about its use in this group of tumors in spite of the long-term oncologic results [38]. However, SCRT with delay to surgery and SCRT with intervening systemic therapy prior to surgery are both associated with meaningful pathologic CR rates and overall tumor downstaging. Along these same lines, SCRT may also be adapted to the growing interest in “watch and wait” approaches for patients with rectal cancer [39]. The results of SCRT in patients with the highest-risk features will need to be further studied.

Conclusion

SCCRT has not been widely adopted in the USA based on concerns about toxicity, downstaging effects, and lack of long-term follow-up. However, randomized clinical trials have demonstrated its efficacy in reducing the risk of local-regional recurrence relative to surgery alone and shown similar oncologic outcomes to LCCRT, without major differences in long-term toxicities. While high levels of late bowel toxicity were seen with short-course radiation using older techniques and larger volumes, more modern studies have not shown significant differences increased late toxicity between short-course radiation, either with immediate or delayed surgery, as compared to LCCRT. Delaying surgery after short-course radiation increases downstaging, and incorporation of neoadjuvant systemic therapy after short-course radiation may further increase downstaging while also addressing micrometastatic disease. While longer follow-up and results from ongoing trials will continue to refine recommendations, sufficient data exist for routine incorporation of SCRT into the clinical management of many patients with rectal cancer.