Authors: Erqi Pollom (1Department of Radiation Oncology, Stanford School of Medicine), Vipul R. Sheth (2Body MRI Division, Department of Radiology, Stanford School of Medicine), Aaron J. Dawes (3Section of Colon & Rectal Surgery, Department of Surgery, Stanford University School of Medicine; 4Stanford-Surgical Policy Improvement Research and Education Center, Department of Surgery, Stanford University School of Medicine), Thomas Holden (5Department of Medicine, Stanford School of Medicine)
Categories: Article, total neoadjuvant therapy, watch and wait, non-operative management
Source: Cancer journal (Sudbury, Mass.)
Authors: Erqi Pollom, Vipul R. Sheth, Aaron J. Dawes, Thomas Holden
The treatment paradigm for rectal cancer has been shifting towards de-escalated approaches to preserve patient quality of life. Historically, the standard treatment in the United States for locally advanced rectal cancer has standardly comprised preoperative chemoradiotherapy coupled with total mesorectal excision. Recent data challenges this “one-size-fits-all” strategy, supporting the possibility of omitting surgery for certain patients who achieve a clinical complete response to neoadjuvant therapy. Consequently, patients and their physicians must navigate diverse neoadjuvant options, often in the context of pursuing organ preservation. Total neoadjuvant therapy, involving the administration of all chemotherapy and radiation before total mesorectal excision, is associated with the highest rates of clinical complete response. However, questions persist regarding the optimal sequencing of radiation and chemotherapy and the choice between short-course and long-course radiation. Additionally, meticulous response assessment and surveillance is critical for selecting patients for non-operative management without compromising the excellent cure rates associated with tri-modality therapy. As non-operative management becomes increasingly recognized as a standard of care treatment option for patients with rectal cancer, ongoing research in patient selection and monitoring as well as patient-reported outcomes is critical to guide personalized rectal cancer management within a patient-centered framework.
In the United States, the conventional treatment for locally advanced rectal cancer has historically involved preoperative chemoradiotherapy (CRT) followed by total mesorectal excision (TME) and adjuvant chemotherapy^1,2^. Given the excellent locoregional control of this tri-modality treatment approach, there has been growing interest in reducing treatment-associated morbidity by de-escalating locoregional therapy. For example, the recent PROSPECT trial showed that among patients with favorable risk rectal cancer, neoadjuvant FOLFOX, compared to neoadjuvant CRT, was noninferior and had less detrimental impact on sexual function^3,4^.
However, TME itself can be associated with potential complications, including sexual and urinary dysfunction, anastomotic leak or strictures, and issues with perineal wound healing. Patients undergoing abdominoperineal resection have reported worse body image, micturition symptoms and sexual enjoyment at 1 year post surgery compared to those who had sphincter-sparing surgery^5^. Even with sphincter-sparing surgery, 50 to 90% of patients experience bowel dysfunction including low anterior resection syndrome^6^. This syndrome involves frequent and urgent bowel movements, an inability to completely empty the bowels, and even accidental leakage of stool, significantly impacting patient function and overall quality of life.
Patients who achieve a pathologic complete response (pCR) after CRT have a better prognosis and potentially could avoid TME^7^. Habr-Gama et al. published one of the first organ preservation experiences for rectal cancer, reporting 5-year overall and disease-free survival of 100% and 92%, respectively, without surgery, for those patients achieving a clinical complete response (cCR)^8^. Since then, numerous series have demonstrated acceptable rates of sustained local control and excellent survival with an organ preservation approach. A meta-analysis of 23 of these studies reported a 2 year local regrowth rate of 16.7%, with similar survival compared with patients with pCR who underwent surgery^9^. Another large international dataset of 800 patients managed with watch-and-wait at 47 centers in 15 countries showed favorable 5 year overall and disease-free survival of 94 and 85%, respectively^10^. Finally, the recently published OPRA trial demonstrated impressive 3-year colostomy-free survival of 53% with their watch-and-wait strategy^11^.
As we become more comfortable offering organ preservation to patients with rectal cancer, crucial questions arise concerning how to tailor neoadjuvant interventions to enhance the likelihood of cCR. These include determining the optimal sequencing of treatments, the duration and intensity of radiation, and the specific chemotherapy regimens employed. Identifying the individuals most likely to benefit from organ preservation approaches is another complex task that involves considering tumor characteristics, overall health, and patient preferences. The goal is to minimize treatment-related morbidity while not jeopardizing the excellent cure rates achieved with Tri-modality therapy.
Many of the watch-and-wait experiences are heterogeneous in the neoadjuvant therapies used. Nonetheless, a recurrent finding “total neoadjuvant therapy (TNT),” wherein the chemotherapy traditionally administered post-TME is instead delivered prior to TME, is consistently associated with the highest rates of cCR.
TNT addresses concerns related to the risk of distant disease both by enabling earlier treatment of micrometastases and by improving treatment compliance^12,13^. Delivering most or all of the chemotherapy before, rather than after surgery, also has practical benefits for the patients, including reducing the interval to diverting loop ileostomy takedown. Finally, as only up to 15–20% of patients have a cCR following standard CRT, TNT can increase extent of tumor response, with rates in some series exceeding 50%^14^, and likelihood of patient eligibility for non-operative management.
Given these advantages, TNT became a standard approach for rectal cancer at many institutions. However, there had not been prospective trials showing clear benefit until recently with the RAPIDO and PRODIGE23 trials (Table 1). The RAPIDO trial compared standard pre-operative long-course CRT with pre-operative short-course radiation (SCRT) followed by 6 cycles of capecitabine and oxaliplatin (CAPOX) or 9 cycles of folinic acid, oxaliplatin, and fluorouracil (FOLFOX) for patients with high-risk rectal cancers (those with T4a/b or N2 disease, extramural vascular invasion, involved mesorectal fascia, or enlarged lateral lymph nodes). PRODIGE23 compared standard CRT to 6 cycles of folinic acid, oxaliplatin, fluorouracil, and irinotecan (FOLFIRINOX) followed by CRT prior to surgery. Both confirmed that TNT reduces the risk of distant metastases and increases pCR rates to almost 30%. Neither trial showed an improvement in overall survival with TNT.
Thus, if patients desire non-operative management, TNT should be employed to give the best chance of cCR. However, it is noteworthy that RAPIDO and PRODIGE23 used significantly divergent TNT regimens. One sequenced radiation first, prior to chemotherapy, while the other sequenced radiation after chemotherapy. Further, SCRT and doublet chemotherapy was used in RAPIDO while PRODIGE23 used long-course CRT and triplet chemotherapy.
The OPRA trial helped address the TNT sequencing question by comparing 16–18 weeks of chemotherapy followed by long-course CRT versus CRT followed by 16–18 weeks of chemotherapy. The rationale for delivering CRT after chemotherapy has in part been due to concern that a prolonged delay between the completion of radiation and surgery may lead to increased surgical morbidity resulting from pelvic fibrosis and wound-healing complications^15^. On the other hand, extending time to surgery following radiation can allow more time for tumor downstaging and increase chance of cCR^16^.
The OPRA trial found that patients who started with CRT had a higher cCR rate (76% vs 71%), lower local regrowth rate (27% vs 40%), and higher rate of a preserved rectum at 3 years (53% vs 41%), compared to those who started with chemotherapy. This difference was sustained at the 5-year long-term analysis showing a preserved rectum in 54% compared to 39% in those who started with chemotherapy with similar disease-free survival (DFS) in both groups^14^. Importantly, the rate of distant metastases was not significantly different between patients having surgery at regrowth versus re-staging, suggesting that even patients with regrowth may not be worse off than they would have been had they chosen surgical resection at the outset^11^. Another randomized phase II trial showed that upfront CRT followed by 3 cycles of FOLFOX, compared to 3 cycles of FOLFOX followed by CRT, resulted in better compliance with CRT and improved pCR rates (25 versus 17%). Further, this trial did not show increased surgical morbidity or lower R0 resection rates with the prolonged duration between CRT and surgery^17^.
Despite these findings, a persistent concern with long-course CRT first is the resultant delay to initiating systemic therapy. SCRT, as given on the RAPIDO trial, could be a mechanism by which to deliver radiation therapy first while minimizing delay in systemic therapy. Another approach would be to deliver a few (but not all) cycles of chemotherapy first, followed by long-course CRT, and then completing the remaining chemotherapy while allowing time for response after CRT. Both approaches may allow earlier initiation of systemic therapy while ensuring sufficient time after radiation to maximize response and increase chance of nonoperative management.
While the German Rectal trial^1^ established preoperative long-course CRT as standard of care for locally advanced rectal cancer in the United States, multiple trials have demonstrated the efficacy of SCRT^18–23^. Thus, a persistent clinical question has been whether to use SCRT or long-course CRT for patients with rectal cancer.
While there were initial concerns for lack of downstaging of bulky tumors and inferior outcomes for patients with low-lying tumors who received SCRT^18,23^, the addition of a delay (during which chemotherapy can be delivered) prior to TME has resulted in comparable outcomes with SCRT as CRT^23,24^. However, concerns persist regarding the potentially lower biologically effective doses of SCRT and the impact on local control. The 5-year follow-up of the RAPIDO trial found that locoregional recurrence was higher in the short-course TNT arm compared to CRT (10 versus 6%)^25^. Risk factors for locoregional recurrence on this trial included enlarged lateral lymph nodes, positive circumferential resection margin, tumor deposits, and node positivity. The higher number of breached mesorectums in the SCRT TNT arm raises questions about whether inferior outcomes were due to SCRT or prolonged time between radiation and surgery. However, the prevailing body of evidence supporting nonoperative management predominantly relies on long-course CRT. The OPRA trial, for instance, demonstrated substantial and sustained cCR rates, with up to a 3-year TME-free survival of 53% achieved with long-course CRT.
Despite the current preference for long-course CRT for nonoperative management, SCRT warrants further investigation. There have been several SCRT nonoperative management studies from Washington University School of Medicine. One of these was the NORMAL-R trial, on which patients were treated with 25 Gy in 5 daily fractions of pelvic radiation (37% of patients received a simultaneous integrated radiation boost of 30 Gy to primary and 35 Gy to extramesorectal lymph nodes) followed by 8 cycles of FOLFOX or 5 cycles of CAPOX. Of 19 pts enrolled, 21% had T2N0 rectal cancer. They reported impressive initial cCR of 74% and 1-year cCR of 68%. The same group also reported a larger retrospective series of 86 patients treated with a similar approach. A quarter of patients again had favorable disease (stage I or II). Half of patients had a cCR, and of these, 21% ultimately experienced local regrowth^26^.
Ongoing studies are also exploring SCRT with radiation boost and triplet chemotherapy to increase cCR rates^27^. The ACO/ARO/AIO18.1, which is ongoing, seeks to compare the RAPIDO approach of SCRT to the CAO/ARO/AIO-04/−12 approach of CRT, both followed by consolidation chemotherapy with a primary outcome of organ preservation (NCT04246684). Although we do not have a long-term DFS and organ preservation data for TNT with SCRT versus long-course CRT, SCRT can be advantageous in improving treatment compliance, patient convenience and cost, and minimizing the time patients are off multi-agent systemic therapy.
There have been efforts to intensify systemic therapy to not only reduce the risk of distant metastases but also to maximize response. In metastatic colorectal cancer, a triplet regimen with FOLFOXIRI has been shown to be superior to standard doublet chemotherapy^28^. This intensified regimen has been explored in non-metastatic rectal cancer, including on the PRODIGE23 trial as discussed above^29,30^. The JANUS trial is comparing CRT followed by consolidation doublet (mFOLFOX6 or CAPOX) or triplet chemotherapy (mFOLFIRINOX) for 4 months, with a primary endpoint of cCR.
Additionally, escalating the radiation dose to gross tumor can also help to improve rates of cCR. In a prospective observational study, patients with cT2–3 cN0–1 rectal tumors within 6 cm of the anal verge were treated with CRT using concurrent tegafur-uracil chemotherapy and high radiation doses (60 Gy to primary tumor using external beam radiation followed by a 5 Gy endorectal brachytherapy boost). With this regimen, cCR was 78%^31^. The OPERA study randomized patients with early T2–3 rectal cancer to a boost of external beam radiotherapy at 9 Gy in 5 fractions versus contact x-ray brachytherapy (90 Gy in 3 fractions) following neoadjuvant chemoradiotherapy (45 Gy external beam radiotherapy in 25 fractions over 5 weeks with concurrent oral capecitabine). The 3-year organ preservation rate was significantly higher with the addition of the contact x-ray brachytherapy boost (97 vs 63%)^32^.
Intensification with novel therapies has also been explored. NRG-GI002 utilized 2 experimental and separate TNT comparator arms with veliparib or pembrolizumab added concurrently to either CRT or TNT with all arms sequencing induction chemotherapy followed by CRT^33^. The experimental arms were only compared to the control arm and not each other. Long term follow-up showed no significant improvement in short-term outcomes (pCR, cCR, organ preservation, neoadjuvant rectal score) in either experimental arm. In the pembrolizumab arm, there was an improvement in 3-year overall survival but not disease-free survival. Future and ongoing biomarker analyses must be pursued to identify who derives short- and long-term benefit from the addition of targeted modalities.
Paradoxically, while these measures are intended to give patients a better chance for treatment de-escalation, they do raise concern for overtreatment, if patients are not able to avoid surgery.
Response following neoadjuvant therapy should be a multidisciplinary assessment that includes flexible sigmoidoscopy, MRI pelvis, and if applicable, digital rectal exam. MRI is an essential restaging modality to identify patients with cCR. However, it can be difficult to differentiate residual tumor from fibrosis following CRT. There are several response assessment systems for MRI. The oldest is the 5-point MRI tumor regression grading (mrTRG) system, which is based on T2-weighted imaging assessment of the primary tumor site, and has been used in ongoing trials^34,35^ (Table 2). It has been found to be reliable and reproducible between multiple independent radiologists and validated against both pathology and survival outcomes^36–39^. Alternative regression assessment methods include the modified 3-point mrTRG, the ESGAR 2018 consensus guidelines, and the Memorial Sloan Kettering (MSKCC) Regression Schema^40–42^, all of which incorporate diffusion-weighted MRI in addition to T2-weighted MRI assessment and grade response on a 3-point scale instead of 5-point scale. Comparisons have shown that incorporation of diffusion weighted MRI can increase reader agreement and accuracy. However, diffusion weighted MRI is prone to artifacts, which can introduce pitfalls for inexperienced readers^43,44^. Unlike the mrTRG system, these alternative regression assessment methods have not been proven to predict survival outcomes. Flexible sigmoidoscopy can further supplement MRI findings with direct visualization of the rectal mucosa; however, it cannot assess deeper layers of the rectum.
With current tools, up to 30% of complete responders are not identified at response assessment^45,46^. Thus, in selected patients, extending the waiting interval could be considered for patients who show a “near CR” at first response assessment to allow for further regression and help determine whether or not there is a cCR^47^. On the OPRA trial, responses were classified into three tiers, based on the MSKCC schema (Table 3): cCR, near CR, and incomplete response^41^. Those who achieved cCR and near CR were offered watch-and-wait. Of the patients with a near CR, the 3-year TME rate was 48% compared with 21% in the cCR group^14,48^, thus validating the approach to deferring surgery even for those patients who technically do not have cCR as over 50% of patients with a near CR were able to ultimately avoid surgery. When compared to the mrTRG system, cCR in the MSKCC schema is equivalent to mrTRG 1 or 2; near CR is equivalent to mrTRG 3; and incomplete response is equivalent to mrTRG 4 and 5. Some trials have shown a difference in disease-free survival and overall survival between those with cCR (mrTRG 1 and 2), near CR (mrTRG 3), and incomplete response (mrTRG 4 and 5), while others have grouped mrTRG 1–3 together like the watch and wait group in the OPRA trial and shown that mrTRG1–3 is strongly predictive of pCR^49^. Figure 1 shows an example of response assessment integrating mrTRG and flexible sigmoidoscopy results.
Careful surveillance following organ preservation is critical. Although up to 30% of patients managed with an organ preservation approach will develop local regrowth, the majority of local regrowths are surgically salvageable^9,10,50^. A decision-analytic model using published data to compare organ preservation versus upfront TME for locally advanced rectal found that organ preservation had lower costs and higher quality-adjusted life-years if surgical salvage rates were high (>73%), emphasizing the importance of careful patient selection and surveillance^51^. Patients should be followed with a combination of digital rectal exam, flexible endoscopy, MRI pelvis and cross-sectional imaging of chest, abdomen, and pelvis for at least 5 years, with more frequent surveillance in the first 2 years^52^. On the OPRA trial, nearly all tumor regrowth in watch-and-wait patients occurred during the first 2 years after restaging (94% of tumor regrowth occurred within 2 years and 99% occurred within 3 years after staging)^14^. Similarly, analysis of the International Watch & Wait Database found that the probability of remaining free from local regrowth if a patient had a sustained clinical complete response for 1 year was 88.1% and for 3 years was 97.3%^53^.
Additional tools such as PET/MRI are being investigated with potential to integrate into and improve response assessment^39^. PET/MRI may detect more residual disease after TNT than MRI alone^54^. While CT of the chest and abdomen is typically used in assessment of metastatic disease, gadoxetic acid enhanced liver MRI has been shown to outperform CT in detection of metastatic colorectal cancer^55^. Circulating tumor DNA (ctDNA) from the peripheral blood can be also used as a noninvasive method for tumor interrogation and monitoring. ctDNA is detectable in over 75% of patients with localized colorectal cancers^56^ and has been shown in the post-operative setting to be an indicator of minimal residual disease and predict radiologic recurrence with better sensitivity than CEA^57,58^. Biomarker development is needed to both select patients for treatment de-escalation as well as monitor for recurrence after treatment.
In conclusion, the landscape of rectal cancer treatment is evolving towards a patient-centered approach with an emphasis on treatment de-escalation for appropriate patients. Non-operative management is now increasingly recognized as a standard of care treatment option for rectal cancer that patients and physicians can consider, based on factors such as side effects, lifestyle impact, and personal preferences. Survey studies have suggested that rectal cancer survivors were willing to accept a 20% increase in risk of local regrowth and 20% decrease in survival if it meant avoiding surgery.^59^ Long-term quality of life data comparing these different neoadjuvant approaches is needed to guide treatment decisions effectively. The goal is to minimize treatment-related morbidity while maintaining the excellent cure rates achieved with standard therapy. As we navigate this evolving paradigm, ongoing research in patient selection and monitoring as well as patient reported-outcomes will be essential to refine strategies for personalized rectal cancer management.