Clinical factors of post-chemoradiotherapy as valuable indicators for pathological complete response in locally advanced rectal cancer

OBJECTIVES: Pathological complete response has shown a better prognosis for patients with locally advanced rectal cancer after preoperative chemoradiotherapy. However, correlations between post-chemoradiotherapy clinical factors and pathologic complete response are not well confirmed. The aim of the current study was to identify post-chemoradiotherapy clinical factors that could serve as indicators of pathologic complete response in locally advanced rectal cancer. METHODS: This study retrospectively analyzed 544 consecutive patients with locally advanced rectal cancer treated at Sun Yat-sen University Cancer Center from December 2003 to June 2014. All patients received preoperative chemoradiotherapy followed by surgery. Univariate and multivariate regression analyses were performed to identify post-chemoradiotherapy clinical factors that are significant indicators of pathologic complete response. RESULTS: In this study, 126 of 544 patients (23.2%) achieved pathological complete response. In multivariate analyses, increased pathological complete response rate was significantly associated with the following factors: post-chemoradiotherapy clinical T stage 0-2 (odds ratio=2.098, 95% confidence interval=1.023-4.304, p=0.043), post-chemoradiotherapy clinical N stage 0 (odds ratio=2.011, 95% confidence interval=1.264-3.201, p=0.003), interval from completion of preoperative chemoradiotherapy to surgery of >7 weeks (odds ratio=1.795, 95% confidence interval=1.151-2.801, p=0.010) and post-chemoradiotherapy carcinoembryonic antigen ≤2 ng/ml (odds ratio=1.579, 95% confidence interval=1.026-2.432, p=0.038). CONCLUSIONS: Post-chemoradiotherapy clinical T stage 0-2, post-chemoradiotherapy clinical N stage 0, interval from completion of chemoradiotherapy to surgery of >7 weeks and post-chemoradiotherapy carcinoembryonic antigen ≤2 ng/ml were independent clinical indicators for pathological complete response. These findings demonstrate that post-chemoradiotherapy clinical factors could be valuable for post-operative assessment of pathological complete response.


' INTRODUCTION
Currently, preoperative chemoradiotherapy (CRT) followed by total mesorectal excision (TME) is the standard treatment for patients with locally advanced rectal cancer (LARC) to improve local disease control and sphincter preservation rate (1,2). Following preoperative chemotherapy, variable responses have been observed among individuals, with a range of 12% to 40% of cases achieving pathological complete response (pCR) (3)(4)(5). Patients achieving pCR after CRT have a lower incidence of recurrence and more favorable long-term survival than those without pCR (6,7). Additionally, due to the associated relatively long-term survival, tumor complete response allows the adoption of a ''wait-and-see'' policy for patients undergoing neoadjuvant CRT instead of necessitating immediate surgery (8,9). Therefore, it is of considerable clinical importance to identify significant factors that can successfully predict pCR after preoperative CRT. Although many techniques have been applied towards this goal, correlations between post-CRT clinical factors and pCR are seldom reported and remain poorly defined (10). To identify significant post-CRT clinical factors associated with pCR, we conducted the current study using a large cohort of patients undergoing CRT and surgery at a single center.

Patient selection
This study included a total of 577 consecutive patients with rectal cancer who underwent preoperative CRT followed by surgery from December 2003 through June 2014 at Sun Yat-sen University Cancer Center, China. The following inclusion criteria were applied: (1) histologically confirmed rectal adenocarcinoma; (2) cT3-4 or N+ disease before CRT; (3) no other anti-tumor therapy was received before CRT; and (4) no distant metastatic disease. Thirty-one patients were excluded from this study due to distant metastatic disease before CRT and an additional two patients who were confirmed to have adenosquamous carcinoma were also eliminated. This left a total of 544 individuals who were included for analysis. The detailed clinical data of the eligible patients were reviewed using our electronic medical record system. Prior to CRT and surgery, informed consent was obtained from each patient. The study was undertaken in accordance with the ethical standards of the World Medical Association Declaration of Helsinki and study approval was obtained from independent ethics committees at Sun Yat-Sen University Cancer Center.

Treatment
All of the subjects underwent preoperative CRT. Radiation therapy (RT) was administered via three-dimensional conformal radiation therapy (3D-CRT) or intensity-modulated radiation therapy (IMRT). The patients were scheduled to receive a total irradiation dose of 46.0-50.4 Gy to the pelvic area, delivered in 1.8-2.0 Gy fractions daily on five consecutive days per week for 5-6 weeks. If a patient could not tolerate the full dose or suffered from severe toxic effects, the RT was stopped. Patients who were reluctant to undergo surgery were delivered a radical dose of RT.
One of the following chemotherapeutic regimens was delivered concurrently with the RT: (1) XELOX regimen consisting of 130 mg/m 2 oxaliplatin administered intravenously on day 1 and 1,000 mg/m 2 capecitabine administered orally twice daily on days 1-14 for a 3-week cycle; (2) FOLFOX regimen consisting of 85 mg/m 2 oxaliplatin and 400 mg/m 2 leucovorin administered intravenously over 2 hours on day 1, 400 mg/m 2 5-FU administered intravenously on day 1 and 1200 mg/m 2 5-FU administered intravenously for 2 days for a 2-week cycle; or (3) 825 mg/m 2 capecitabine administered orally twice daily during RT without weekend breaks.
All subjects were scheduled for radical resection 6-8 weeks after the completion of preoperative CRT. If bowel obstruction or perforation occurred after CRT, emergent surgery was performed immediately. Surgery was delayed in the event of grade 3 or 4 toxicity. The surgical procedure used depended on tumor location and invasive extent; TME was performed whenever possible.

Evaluation
Post-CRT clinical evaluation was performed at the 6 th week after completion of CRT and included a combination of physical examination, post-CRT carcinoembryonic antigen (CEA) test, colonoscopy, transrectal ultrasonography (EUS), chest computed tomography (CT) and abdominal and pelvic nuclear magnetic resonance (MRI). Disease stage was classified according to the 2010 American Joint Committee on Cancer/International Union Against Cancer (AJCC/ UICC) staging system. According to EUS images, post-CRT clinical (yc) T stage 0 was defined as no tumor signs in the bowel wall; ycT stage 1 was defined as a tumor signal confined to the submucosa; and ycT stage 2 was defined as a malignancy that had invaded the muscularis propria. Using MR images, ycN stage 0 was defined as the absence of metastatic lymph nodes; ycN stage 1 and ycN stage 2 were respectively defined as the presence of 1-3 and 43 metastatic lymph nodes, which appeared irregular in shape, showed enhancement and possessed a diameter 45 mm; ycT stage 3 was defined as a malignancy that broke through the muscularis propria; and ycT stage 4 was defined as a tumor that had invaded adjacent organs. All radiological reports were confirmed by two independent radiologists. Postoperative final tumor pathological staging was also evaluated by two independent pathologists. pCR was defined as the absence of viable tumor cells, with only fibrotic masses or acellular mucin pools present in proximity to the primary tumor and lymph nodes (11). Patients confirmed as having achieved pCR were classified as the pCR cohort, while the remaining patients formed the non-pCR cohort.

Statistical analysis
Data were analyzed using SPSS version 17.0 (Chicago, IL, USA). Continuous variables were shown as the median (range) or the mean (standard deviation) and categorical variables were presented as percentages. All continuous and polytomous variables were dichotomized by a cut-point prior to univariate and multivariate analyses. Univariate comparisons of factors were performed using the chi square test or Fisher's exact test for categorical variables and an unpaired t-test for continuous variables. Factors with statistical significance in univariate analyses were included in multivariate analyses. Multivariate analyses were performed using logistic regression to identify significant post-CRT clinical indicators and therapeutic factors for pCR. All tests were twotailed and po0.05 was considered statistically significant.

Post-CRT clinical indicators for pCR
The cut-off values for post-CRT CEA, DAV and interval from the completion of CRT to surgery were set according to the median values. In the univariate analyses, the potential post-CRT clinical indicators included ycT stage, ycN stage, post-CRT DAV, interval from completion of CRT to surgery,    (Table 3).  Unlike the majority of previous studies, which have focused on pre-CRT factors that can predict pCR (12,13), the present study evaluated the value of certain post-CRT parameters and treatment variables for predicting pCR. Using univariate and multivariate analyses, we showed that ycT stage 0-2, yc N stage 0, interval from completion of CRT to surgery of 47 weeks and post-CRT CEA p2 ng/ml could be applied to assess pCR post-operatively.
In our study, a favorable pCR rate of 23.2% was achieved, which was comparable to previous reports in which CRT was delivered with fluoropyrimidine-based chemotherapy combined with oxaliplatin (4,14,15). To improve pCR rates, several randomized trials have been conducted to integrate oxaliplatin into the currently widely used fluoropyrimidinebased preoperative chemotherapy regimen. The majority of these trials have achieved a favorable pCR rate, ranging from 28% to 40% (5,16). However, another phase III study showed that the addition of oxaliplatin into CRT did not increase pCR: the pCR rate in this study was 13.9% in the capecitabine group compared to 19.2% in the capecitabine plus oxaliplatin group (p=0.09) (17). To the best of our knowledge, drug dose, patient tolerance during treatment and administration schedule all influence the achievement of pCR. A better understanding of chemotherapy schedule as well as the features of specific cohorts of patients might contribute to increase pCR after CRT.
CEA is not only a prognostic factor but also a predictor for response to preoperative CRT in LARC (18). Pretreatment CEA level has been demonstrated as a useful marker for pCR following preoperative CRT (13,19). However, few studies have evaluated the value of post-CRT CEA level in assessing pCR following preoperative CRT. Yang (21). In our current study, a post-CRT CEA level p2 ng/ml was significantly associated with a higher pCR rate (28.5% vs. 20.3%, p=0.037). Although the exact mechanism is unclear, we suppose that the lower post-CRT CEA level implies a lower tumor burden and, subsequently, less residual tumor in the rectum after CRT.
Several retrospective studies have emphasized the role of a longer interval between CRT and surgery in achieving a histological tumor response (22,23). Our study showed that an interval of 47 weeks between completion of CRT and surgical resection significantly increased pCR rate (26.9% vs. 18.3%, p=0.02). A previous study also reported that an interval 47 weeks could improve pCR after assessing 132 patients with LARC (35% vs. 17%, p=0.03) (24). Recently, a large retrospective study evaluating 17,255 patients from the National Cancer Data Base in America indicated that an interval greater than 8 weeks was associated with higher odds of pCR (OR=1.12, 95% CI=1.01-1.25, p=0.040) (25). However, the optimal timing for surgery after neoadjuvant CRT in LARC remains controversial. Stein et al. suggested that there was no incremental benefit from delaying surgery after CRT for tumor regression (26). To settle this issue, a prospective randomized trial named GRECCAR-6 was designed and the final result aims to identify the optimal interval for delivering pCR (27).
It is well known that TNM system category is closely associated with the prognosis of colorectal cancer. Some previous studies have found that pretreatment clinical N stage is an independent clinical predictor for achieving pCR (12,28). To date, few studies have assessed whether post-CRT TNM stage could be used to predict tumor response to CRT (13,29). To identify useful predictors for pCR, a recent study screened yc T stage (po0.001) and yc N stage (po0.001) as predictors of pCR after CRT (30). After assessing post-CRT clinical parameters, we found similar results, namely, that ycT stage 0-2 (OR=2.098, 95% CI=1.023-4.304, p=0.043) and ycN stage 0 (OR=2.011, 95% CI=1.264-3.201, p=0.003) are both independent indicators for pCR. To the best of our knowledge, lower clinical T and N stages are associated with less tumor burden in primary organs and lymph nodes. Further studies should be conducted to confirm the authentic values of ycT stage and ycN stage as indicators of pCR.
We would like to acknowledge several limitations in the present study. The first limitation is the retrospective methodology from a single-institution experience. Additionally, some post-CRT variables, including tumor gross change, circumferential extent of tumor and clinical stage shift after CRT, could not be fully evaluated. Furthermore, the patients included in our study were treated over a course of 11 years. This is a long time and CRT regimens and methods of pathologic assessment have changed in this period. However, even when considering the above limitations, our study identified several post-CRT indicators for pCR. External validation using another large database or prospective clinical studies would be of value. Based on our results, individualized therapy can be tailored for patients after CRT. For instance, patients with lower rectal cancer who are confirmed as complete responders to preoperative CRT may choose a wait-and-see policy with avoidance of abdominal perineal resection, as such patients may not achieve sufficient benefit from surgery. On the contrary, patients who fail to achieve a good response tend to have a poor prognosis and consequently need more intensive treatments or the application of novel therapies to minimize the possibility for local and distant recurrences.
This large, retrospective study identified post-CRT clinical parameters of considerable value as indicators for pCR. Post-CRT clinical (yc) T stage 0-2, ycN stage 0, an interval from the completion of CRT to surgery of 47 weeks and post-CRT CEA p2 ng/ml could be useful for predicting pCR. These indicators may help clinicians evaluate the efficiency of preoperative CRT and formulate individualized treatment strategies before surgery.

' ACKNOWLEDGMENTS
We thank all the staff engaged in this study, especially those at the Department of Colorectal Surgery at the Sun Yat-sen University Cancer Center.