Effect of interval between neoadjuvant chemoradiotherapy and surgery on oncological outcomes in poor responders with locally advanced rectal cancer: a retrospective cohort study

Background: The optimal interval from completion of neoadjuvant chemoradiotherapy (CRT) to surgery in locally advanced rectal cancer remains controversial. It seems that delayed surgery is associated with an increase in pathological complete response rates. However, the prognostic effect of delayed surgery in poor responders is unclear. Materials and methods: Patients with locally advanced mid or distal rectal cancer undergoing neoadjuvant CRT followed by total mesorectal excision at a university teaching cancer center between June 2010 and December 2018 were retrospectively reviewed in this study. According to the tumor regression grade, poor responders (tumor regression grade 2–3) to neoadjuvant CRT were selected for analyses. Patients were divided into the longer interval group (greater than 8 weeks) and the shorter interval group (8 weeks or less) based on the wait time from completion of neoadjuvant CRT therapy to surgery. Results: among 916 eligible patients, 522 patients had a poor tumor response. There were 217 patients in the shorter interval group and 305 patients in the longer interval group. At the baseline, patients in the longer interval group were more likely to have a T3 stage and positive vascular invasion. Compared with patients in the shorter interval group, patients in the longer interval group had significantly worse overall survival and disease-free survival (DFS) (log-rank test, overall survival: P=0.004, DFS: P<0.001). The 3-year DFS rates were 75.6 and 63.1% in the shorter interval group and the longer interval group, respectively. In the multivariate analysis, delayed surgery was associated with an increased risk of mortality (hazard ratio: 2.003, 95% CI: 1.233–3.253, P=0.005) and recurrence (hazard ratio: 1.555, 95% CI: 1.121–2.156, P=0.008). Conclusion: Patients who had a poor tumor response should be identified by restaging MRI and receive radical surgery in time, without a prolonged interval.


Introduction
Neoadjuvant chemoradiotherapy (CRT) followed by total mesorectal excision (TME) and adjuvant chemotherapy has been recommended as the standard therapy for locally advanced rectal cancer (LARC). Compared with postoperative CRT, neoadjuvant CRT had the advantages of improved resectability and local control, as well as less therapeutic toxicity [1,2] . Our previous study demonstrated that preoperative radiotherapy was superior to postoperative radiotherapy in terms of oncological outcomes in metastatic rectal cancer patients who received definitive surgical resection of the primary tumor [3] .
After receiving neoadjuvant CRT, patients showed varying degrees of tumor downsizing and downstaging. Approximately 20% of patients achieved a pathological complete response (pCR), with excellent long-term outcomes [4][5][6] . This subgroup has the potential of receiving a 'watch-and-wait' approach without radical surgery [7] .
Several tumor regression grading (TRG) systems have been developed to stratify tumor pathological response to neoadjuvant therapy [8,9] . To date, there is no uniform consensus on whether TRG is a prognostic factor in LARC [10][11][12][13] .
Coordination of neoadjuvant CRT and surgery is also important because delayed surgery is associated with higher pCR rates [14][15][16] . Recently, a pooled analysis of 3085 patients from seven randomized trials suggested that the best time to achieve pCR in LARC is at 10 weeks and lengthening the surgical interval had no impact on oncological outcomes [17] . However, several studies have demonstrated that a longer interval is associated with higher rates of positive margins and lower rates of sphincter preservation [18,19] . It is rational to prolong the surgical interval to achieve maximum tumor regression in patients who eventually have a complete or good response confirmed by pathology. Nevertheless, it is unclear whether delayed surgery has a clinical benefit in poor responders.
The aim of this study was to determine the prognostic implications of the surgical interval after neoadjuvant CRT in LARC patients who had a poor tumor response based on the TRG score.

Study population
The records of rectal cancer patients who received surgery at a university teaching cancer center between June 2010 and December 2018 were retrospectively reviewed. A biopsy was performed on all patients to obtain histological confirmation. Patients were staged before neoadjuvant CRT using computed tomography (CT) scans as well as MRI. The inclusion criteria were as follows: patients were diagnosed with LARC, patients received neoadjuvant CRT and curative surgery, mid to low rectal adenocarcinoma was diagnosed by endoscopy (defined as the distal edge of the tumor within 10 cm from the anal verge), pathologically confirmed TRG2-3. The exclusion criteria were as follows: patients received neoadjuvant chemotherapy alone, patients were diagnosed with stage IV disease, and patients received palliative surgery. The following characteristics were extracted: TRG score; age at diagnosis; sex; histologic type; differentiation; yield pathological T stage; ypN stage; perineural invasion; vascular invasion; surgical procedures; and survival data. The study protocol was approved by the ethics committee of the university teaching cancer center. This work was reported in line with the strengthening the reporting of cohort studies in surgery (STROCSS) criteria [20] , (Supplemental Digital Content 1, http://links.lww.com/JS9/A489).

Neoadjuvant CRT and surgical resection
In general, patients received pelvic radiation at a dose of 50 Gy/25 fractions with concurrent capecitabine at 825 mg/m 2 twice daily for 5 days/week. Consolidation chemotherapy with one or more cycles of capecitabine plus oxaliplatin was performed 2 weeks after the end of CRT (oxaliplatin 130 mg/m 2 on day 1 and capecitabine 1000 mg/m 2 twice daily on days [1][2][3][4][5][6][7][8][9][10][11][12][13][14]. The detailed cycles of consolidation chemotherapy depended on the surgical interval. The surgical interval was defined as the waiting time from the completion of CRT to surgery. Surgical resection was performed at least four weeks after the completion of neoadjuvant CRT. All patients received TME. In general, a total of six cycles of chemotherapy (preoperative and postoperative) was administered.

TRG score
The TRG was scored by two independent specialist pathologists who were not aware of the MRI diagnosis according to the seventh edition of the American Joint Committee on Cancer Staging Manual and the College of American Pathologists Guidelines as modified by Ryan et al. [21] . The 4-point TRG system was graded on a scale of 0-3. TRG0: complete regression with no residual cancer cells; TRG1: almost complete regression with only one single residual cancer cell or a cluster of cancer cells; TRG2: moderate regression with many residual cancer cells; TRG3: minimal regression with nearly no cancer cells killed. TRG scoring was evaluated in the pathological report. In the present study, patients with TRG2-3 were considered poor responders.

Follow-up
Clinical examination, tumor marker assessment, colonoscopy, chest and abdominal CT scan, and pelvic CT or MRI were included in the follow-up protocol. Medical records review, telephonic follow-ups, and death registry data linkage were used in combination for collecting survival data. The last follow-up was on 30 September 2020. The primary endpoints were overall survival (OS) and disease-free survival (DFS). OS was defined as the time from the initiation of therapy to the final follow-up or until death from any cause. DFS was defined as the time from the initiation of therapy to the date of local or distant relapse, death from any cause or the date of last follow-up.

Statistical analysis
All statistical analyses were conducted using SPSS version 25.0 (SPSS). χ 2 -test was used to compare categorical variables between two groups. Survival analyses were undertaken using Kaplan-Meier analysis. The log-rank test was used for univariate comparisons. Multivariate Cox regression analysis was performed on variables that were significant in univariate analysis to identify the independent prognostic factors. A P < 0.05 was considered as statistically significant.

Results
A total of 7632 rectal cancer patients were retrospectively reviewed. Of these, 916 patients with locally advanced (clinical stage II or III) mid or distal rectal cancer underwent neoadjuvant CRT followed by TME. Finally, a total of 522 patients who had a poor tumor response were included in the analysis. The detailed patient selection process is shown in Supplementary Figure 1.  Table 1. Patients in the longer interval group were more likely to have a T3 stage and positive vascular invasion. The other characteristics were comparable between the two groups.
The median follow-up was 30.8 months. During the follow-up, 170 patients (32.6%) experienced a recurrence and 86 patients (16.5%) died. Patients in the longer interval group had significantly worse OS (log-rank test, P = 0.004) and DFS (log-rank test, P < 0.001) as compared to patients in the shorter interval group (Fig. 1). The 3-year DFS rates were 75.6 and 63.1% in the shorter interval group and the longer interval group, respectively. Multivariate Cox analysis revealed that delayed surgery was associated with an increased risk of mortality (hazard ratio: 2.003,  Besides that, tumor differentiation, ypN stage, vascular invasion, and surgical procedures were also found to be independently associated with OS and DFS in poor responders. Furthermore, patients had a trend of lower survival as waiting time was prolonged, especially for more than 16 weeks (Fig. 2).

Discussion
The optimal time interval between neoadjuvant CRT and surgery in LARC patients remains unknown. Special attention was paid to the correlation between a longer waiting period and an improvement in the rate of pCR since pCR is considered as a surrogate marker for estimating long-term survival benefits. Notably, a watch-and-wait strategy is a valuable alternative for rectal resection in patients who were evaluated with a clinical complete response. Patients could undergo salvage surgical treatment for local regrowth with adequate pelvic control [22] . To achieve enhanced tumor regression, total neoadjuvant therapy (TNT), defined as delivering all planned chemotherapy before surgery, was developed. A higher pCR rate in a TNT approach might result in more patients adopting a watch-and-wait  [23] , which can be considered as the goal toward delaying surgery indefinitely until the tumor regrows. To date, multiple trials have compared different TNT strategies with standard CRT therapy. Yet, the oncological safety of TNT remains controversial. Particularly, the updated RAPIDO results demonstrated that a TNT strategy (short-course radiotherapy, chemotherapy, and surgery) was associated with an increased risk of locoregional recurrence and the reduction in disease-related treatment failure and distant metastases compared with the standard-care group [24] . Notably, a meta-analysis conducted sensitivity analyses and concluded that the decision to make adjuvant chemotherapy optional in the standard-care group may have biased the results in favor of the above TNT strategy [25] . Evidence indicated that a longer waiting period after neoadjuvant CRT does not improve the oncological outcomes of LARC [26] , which raises the question whether the survival benefit of patients achieving a satisfactory pathological response is impaired or masked by poor responders. Patients with a pCR are likely to have better outcomes, but improved DFS or OS is not observed in the entire population. Therefore, some patients with intensification may have a worse outcome, and resistant tumors may become even more resistant and aggressive. Moreover, delayed surgery was associated with increased complications and a worse quality of mesorectal resection [27] . To the best of our knowledge, this is the first study to investigate the prognostic implication of the surgical interval after neoadjuvant CRT in LARC patients who had a poor tumor response based on the TRG score. This study showed that poor responders in the longer interval group had significantly worse OS and DFS compared with those in the shorter interval group. In this study, the therapy was not suspended in the longer interval group. Consolidation chemotherapy was performed 2 weeks after the end of CRT until the surgery was scheduled. Therefore, the rationality of adopting the TNT strategy in patients who had a poor tumor response remains unclear. For good responders, a longer waiting time may be suitable to achieve a clinical response. A prospective observational study with weekly MRI found that tumor shrinkage is the fastest in the beginning of treatment (26%/week), slows to 7%/week in the last 2 weeks of CRT, and finally to 1.3%/week in the last 5 weeks before surgery [28] . It is difficult to achieve a pCR for some patients with TRG2/TRG3 who received delayed surgery. Therefore, poor responders could be identified by restaging MRI in the early stages and receive radical surgery timely. TNT seems more appropriate for good responders during the CRT course itself.
Recently, a multicenter retrospective cohort study reported that a longer interval before surgery after completing neoadjuvant CRT was associated with worse oncological outcomes in tumors with a poor pathological response to neoadjuvant CRT [29] . Our results are consistent with this conclusion. Nevertheless, they defined a ypT stage of 2 to 3 or a ypN positive as a minor or poor response from the point of yp stage instead of the TRG score. Downstaging cannot accurately reflect the tumor response to neoadjuvant therapy. In particular, if patients achieve a good response with only microscopic foci of tumor cells in the subserosa, they would be staged as ypT3. Besides that, there were more deaths (37.7%) than recurrences (26.9%) in this study, indicating that several patients may die from noncancer-related causes.
Notably, patients with a poor tumor response need to be identified in time and receive radical surgery without delay. How to accurately identify the poor responders in time? Preoperative response evaluation with neoadjuvant CRT remains a challenge in LARC. MRI is accurate and sensitive to evaluate the tumor response to neoadjuvant CRT. In previous studies, the MRI-TRG system predicted survival outcomes for the favorable and poor responders [30,31] . Sclafani et al. [32] explored the correlation between MRI tumor regression grade (mrTRG) and pathological tumor regression grade (pTRG) in rectal cancer, and concluded that the agreement between mrTRG and pTRG is low and mrTRG cannot be used as a surrogate of pTRG. Pang et al. proposed a new fourcategory MRI-TRG system based on the volumetric analysis of the residual tumor and CRT-induced anorectal fibrosis, which could be a surrogate for the TRG classification scheme [33] . Furthermore, radiomics of pretreatment MRIs and a deep learning model based on diffusion kurtosis MRI could show good performance for predicting tumor response [34,35] . This study had some limitations. First, this retrospective study may have introduced potential bias and confounding variables. For example, there were more patients with mucinous and Signet ring cell histology in the long interval group. Second, data on several important clinicopathological characteristics, such as detailed clinical stages, cycles of preoperative chemotherapy, MSI status, et al., were not recorded. Third, circumferential resection margin (CRM) status is a very strong independent predictor of survival; however, the effects of CRM status on oncological outcomes were not evaluated due to the limited number of patients with positive CRM. Fourth, it is more difficult to assess clinical response with MRI or CT in large, circumferential or mucinous/signet ring tumors, which requires time to determine whether or not the tumors respond to treatment.

Conclusion
In summary, poor responders should be identified by restaging MRI and receive radical surgery in time, without a delay.

Ethics approval and consent to participate
This study was approved by the Institutional Review Board of Fudan University Shanghai Cancer Center. The written informed consent was obtained from all patients or legally authorized representative of dead participants.

Data availability statement
The FUSCC dataset used during the current study are available from the corresponding author on reasonable request.

Provenance and peer review
Not commissioned, externally peer-reviewed.