Oncologic results in rectal cancer patients with a subcentimeter distal margin after laparoscopic-assisted sphincter-preserving surgery: a propensity-score matched cohort analysis


 Background: Distal resection margin (DRM) is closely associated with sphincter-preserving surgery and oncologic safety for patients with mid-low rectal cancers. However, the optimal DRM has not been determined. The purpose of this study to assess the impact of a DRM of ≤ 1 cm on oncologic safetyMethods: Data of 378 rectal cancer patients who underwent laparoscopic-assisted sphincter-preserving surgery from 2009 to 2015 were retrospectively analyzed. Patients were divided into two groups based on DRM: ≤ 1 cm (n=74) and >1 cm (n=304). To minimize the differences between the two groups, propensity-score matching on baseline features was performed. Stratified analysis was performed according to neoadjuvant chemoradiotherapy.Results: Before propensity-score matching, no significant differences in 5-year disease-free survival (DFS) (92.8 vs. 81.3%; P=0.128) and 5-year overall survival (OS) (83.7 vs. 82.2%; p=0.892) were observed in patients with DRMs of ≤1(n=74) and >1cm (n=304), respectively. After propensity-score matching (1:1), there were also no significant differences in DFS (88.1 vs. 78.2%; P=0.162) and OS (84.5vs. 84.9%; P=0.420) between the DRM of ≤1 cm group (n=65) and >1 cm group (n=65), respectively. A total of 44 patients received preoperative chemoradiotherapy. In this cohort, the 5-year local recurrence (LR) rates ( p=0.118) and the 5-year DFS rates ( p=0.298) were not significantly different between two groups. A total of 334 patients received surgery without neoadjuvant chemoradiotherapy. There were also no significant differences in the 5-year LR rates ( p=0.150) and 5-year DFS rates ( p=0.172) between two groups.Conclusions: No matter whether patients with rectal cancers receiving neoadjuvant therapy or not, sphincter-preserving surgery with a DRM of ≤1 cm may be acceptable in mid-low rectal cancer without jeopardizing oncologic safety.

DRM and its oncological implications during sphincter-preserving surgery for rectal carcinoma patients with or without NCRT need further investigation. In addition, most previous studies on this issue did not perform strati ed analyses based on NCRT and were limited by small sample size.
The purpose of the present study was to evaluate the impact of a DRM of ≤ 1 cm on long-term oncologic outcomes after sphincter-sparing resection for patients with mid-low rectal cancers by strati ed analysis based on NCRT.

Patients
Clinicopathological data of 378 patients with rectal cancers who were treated by laparoscopy-assisted anterior resection with standard total mesorectal excision (TME) were collected from January 2009 through December 2015. Patients with a history of other malignant tumors, distant metastasis, emergency surgery, palliative surgery, Hartmann procedure, abdominoperineal resection, or a pathologically-proven positive circumferential resection margin (CRM) or DRM were excluded. Patients treated with and without NCRT were analyzed separately.
Neoadjuvant chemoradiotherapy and surgical procedure NCRT was recommended for participants with cT3 or T4 tumor and/or lymph nodes metastasis. However, not all eligible patients received neoadjuvant therapy due to various reasons. Three-dimensional conformal intensity-modulated radiotherapy was used. Neoadjuvant chemoradiotherapy protocol comprised a total irradiation dose of 50.4 Gy, delivered in 2-Gy fractions, 5 days per week for 5 weeks.
Concurrent chemotherapy consisted of continuous infusion of 5-uorouracil and leucovorin or oral capecitabine. Surgery occurred at least 6-8 weeks after completion of long-course preoperative radiotherapy.
Laparoscopy-assisted low anterior resection was performed by experienced surgeons. The TME principle was followed for patients with mid-low rectal lesions, and a partial mesorectal excision was conducted for patients with upper rectal tumors. 5 cm distal margin was recommended for upper rectal cancer, 1-2cm for mid-low rectal tumor. For lesions close to the anorectal junction, a microscopically negative DRM was acceptable. The DRM of the fresh resected specimen was inspected by the surgeon. If it was suspiciously positive or microscopically involved at a frozen-section examination, an additional part of the distal rectum was removed, or abdominoperineal resection (APR) was adopted. Digestive tract reconstruction was performed using double stapled anastomosis or hand-sewn. Whether to perform preventive ileostomy is determined by the surgeon based on the intraoperative situation.

Histopathology
All resected specimens were inspected by surgeons rstly and then sent to pathologists for further examination. The DRM was measured following the microscopic examination of the formalin-xed specimen. It was de ned as the closest distance from the lowest border of the lesion ( or the scar tissue after NCRT) to the distal mucosal resection margin. The doughnut was examined microscopically, but not included in this measurement.

Adjuvant therapy and follow-up
Single-agent capecitabine was given to stage patients with unfavorable prognostic factors. CAPOX (capecitabine plus oxaliplatin) was used to patients with positive lymph nodes. Traditionally, patients were followed-up every 3 months for the rst two years after operation, every 6 months for the next 3 years, and every year thereafter. Examinations included physical examinations, carcinoembryonic antigen (CEA), X-ray/CT of chest, CT/ultrasonography of abdomen and pelvis, and endoscopy, but the adoption of each modality was at the discretion of physicians. PET/CT or bone scan was conducted when appropriate. Survival details were obtained from the follow-up o ce in our hospital or directly from patients or relatives by telephone or WeChat.

De nition of recurrence
Local recurrence (LR), including anastomotic and pelvic lymph nodes recurrence, was de ned as any clinically or histopathologically con rmed carcinoma recurrence within the pelvis after primary operation. Local recurrence rate was regarded as the sum of all local recurrences, regardless of distant metastases.
Distant metastasis is de ned as the metastasis of cancer cells to distant organs (e.g., lung, liver, bone) or lymph nodes via lymphatic or blood pathways following primary surgery. The rates of distant metastases were regarded as the sum of metastases regardless of local recurrence.

Statistical analysis
Categorical variables are presented as frequencies and percentages. Continuous variables are described as mean (s.d.) values. The Student ' s t-test and c 2 test and Mann-Whitney U test were used to analyze continuous and categorical variables, respectively. The Kaplan-Meier method with the log-rank test was used for survival analysis. P-value 0.05 was considered statistically signi cant.
Given the differences in the baseline features between patients in the two groups (Table 1), propensityscore matching was performed to minimize baseline differences. Nearest-neighbor matching without replacement was used. Matching was conducted with the use of a 1:1 matching protocol, with a caliper width equal to 0.2 of the standard deviation of the logit of the propensity score. Confounding variables used to calculate the propensity score were age, sex, tumor distance from anal verge, pathologic T staging, pathologic TNM staging and perineural invasion. The OS and DFS analyses were performed using the overall patients and the matched patients by the DRM, respectively. The survival analyses of patients with or without neoadjuvant therapy were limited to the overall patients rather than the matched patients, due to the small size of patients treated by neoadjuvant therapy. All analyses were carried out with SPSS version 23.0 (SPSS, Inc., Chicago, IL, USA).

Clinicopathological characteristics of patients with mid-low rectal cancer
Detailed clinicopathological data of the 378 rectal cancer patients who underwent anterior resection are shown in Table 1. In the present study, 74 cases with a DRM of ≤1cm were compared to 304 cases with a DRM of >1cm. The mean DRM length was 0.8±0.3 cm in the DRM of ≤1 cm group and 2.6±1.0 cm in the DRM of >1 cm group. The mean distance from the anal verge in the DRM of ≤1 cm group was signi cantly different from that in the DRM of >1 cm group (8.3±3.3 vs. 9.3±2.8cm, p=0.017). The groups were comparable in gender, age, preoperative CRT, tumor differentiation, and lymphovascular invasion.
After applying propensity-score matching strategy (1:1), 65 patients with a DRM of ≤1cm were matched to 65 patients with a DRM of >1cm. Form Table 1, there were no signi cant differences in bassline clinicopathological data between two groups (p>0.05).  The patterns of local recurrence and distant metastasis in the two groups are presented in Table 3. With regard to local recurrence, pelvic lymph nodes recurrence was more common than anastomotic recurrence in both groups. As for distant metastasis, the lung was the most common metastatic organ in two groups. Four patients had lung metastases, 2 patients had simultaneous lung and liver metastases, and another patient had paraaortic lymph nodes relapse in the DRM of ≤1 cm group. Similarly, there were 16 lung metastases, 12 liver metastases, 8 simultaneous lung and liver metastases, 2 paraaortic lymph nodes relapse, and 2 ovary metastases in the DRM of >1 cm group.  1).

Subgroup analysis of overall survival strati ed by DRM and the use of NCRT
Considering the in uence of NCRT on the distance of tumor invasion, we analyzed the effect of DRM on overall survival of rectal cancer patients according to whether preoperative chemoradiotherapy was adopted.
A total of 44 patients received preoperative chemoradiotherapy. In this cohort, the 5-year local recurrence rates were similar between the DRM of ≤1 cm group and the DRM of >1 cm group ( 8.  Table 4 and Fig. 3).

Discussion
The optimal length of DRM has not been determined during sphincter-sparing surgery for patients with mid-low rectal cancers, due to the lack of evidence from high-level randomized controlled studies. In this situation, it is challenging to choose a reasonable surgical type (e.g., low anterior resection or APR) for both surgeons and patients. Therefore, the investigation of the effect of the DRM on the oncological results of the low-lying rectal cancer patients is necessary and of important clinical signi cance. In the current study, patients with positive CRM and/or DRM were excluded in order to avoid interference with the oncological results. Several studies [17,18] have con rmed that positive CRM was signi cantly associated with worse clinical outcomes. Patients with microscopically positive DRMs were related to a signi cantly higher local recurrence rate [3,19].
The length of the DRM is mainly depended on the tumor location in the rectum. In this study, the mean distance from the anal verge in the DRM of ≤1 cm group was shorter than that in the DRM of >1 cm group, similar to previous studies [13,20,21]. Meanwhile, we found that the proportion of patients with T1 and T2 stages in the DRM of ≤1 cm group was higher than that in the control group. Similarly, the proportion of the patients with stage and in the DRM of ≤1 cm group was higher than that in the DRM of >1 cm group. This condition is mainly caused by case selection. Compared to low-lying rectal cancer patients with advanced T/TNM stage, patients with earlier stage are more likely to be received sphincterpreserving resection rather than APR.
Overall, there were no signi cant differences in the 5-year LR rate, 5-year DFS, and 5-year OS between the two groups in this study. Similarly, A systemic review of 5574 rectal cancer patients conducted by Bujko et al. [22] was to evaluate whether a DRM of <1cm jeopardizes oncologic safety. They concluded that a DRM of <1cm did not compromise oncologic outcomes. Dong Woo Kang et al. [21] reported that the 5year LR rate was 8.8% in the DRM of ≤1 cm group and 8.5% in the DRM of >1 cm group (p = 0.630). The 5-year DFS rate was 75.1 and 76.3% (p = 0.895), and the 5-year OS rate was 82.6 and 85.9% (p=0.401), respectively. In the study, about 41.4% of enrolled patients were treated by NCRT according to risk factors, such as cT4, positive CRM.
Many previous studies [13-16, 19, 20, 23] attempted to de ne the narrowest su cient DRM ( 5mm, 8mm, 1cm, and 2cm) in patients with sphincter-preserving surgery. However, most of these studies, the analysis of the impact of DRM on survival results was not well strati ed by NCRT.
A subgroup analysis of this study showed that there were no signi cant differences in the 5-year LR and 5-year DFS between the two groups in 44 patients who received NCRT. In agreement with our result, Philipp et al. analyzed 88 patients with mid-low rectal cancer receiving NCRT. They found that the 5-year LR rate was 6.7% in patients with DRM of <1cm and 5.5% in patients with DRM of ≧1cm [13]. Preoperative chemotherapy may increase the rate of sphincter-preserving surgery due to the regression of primary tumors [24,25]. It was reported that neoadjuvant chemoradiation of tumors was associated with shorter distal spread compared to tumors without preoperative radiation [26,27].
In addition, many previous studies showed that distal intramural spread greater than 1 cm was only in 0-5% of patients [26,28,29]. However, Akihiro Kondo et al.[16] analyzed 71 patients with low rectal carcinoma who received preoperative chemotherapy. They found that 42 (59%) patients had distal spread. Distal spreads of 10-19mm, and ≧2cm were observed in 11 (15%), and 4 (6%) patients, respectively. Besides, they revealed that the presence of different therapeutic effect between the mucosal and deeper layers and poorly differentiated adenocarcinoma were independent risk factors for DRM ≧ 1cm after preoperative chemotherapy. Thus, these ndings suggest that for carefully selected patients with neoadjuvant chemoradiotherapy, a subcentimeter DRM is acceptable and not associated with worse oncologic results. It should be noted that for patients with high-risk factors who underwent preoperative chemotherapy, a DRM of 1 cm may not be su cient.
By subgroup analysis in the present study, a DRM of ≤1 cm did not compromise the oncologic outcomes of patients who received TME surgery alone. The introduction and application of TME principle for advanced rectal cancer have signi cantly reduced local recurrence and improved a safe DRM [7,30,31]. A study including 152 mid-low rectal cancer patients with TME surgery alone reported that there were no signi cant differences in 10-year recurrence rates between the DRM ≤1 cm group and the DRM >1 cm group (0.0 vs. 3.6%, p=0.27). The authors believed that sphincter-saving TME with a subcentimeter DRM was a preferable option to preserve anal function without compromising oncological safety. Although patients were not treated by NCRT, postoperative adjuvant chemotherapy and/or radiotherapy may compensate for the negative impact of near margin on survival.
The present study has several drawbacks. First, some statistical bias is inevitable due to the retrospective nature of the research design, although the propensity-score matching was performed. Prospective cohort studies will be conducted to reduce such errors. Second, although the total sample size is large, the number of patients with a DRM of ≤1 cm group is small. Third, the proportion of patients receiving neoadjuvant therapy was low, and adjuvant therapy was not uniform, which may affect oncologic results.
In addition, not all specimens in this study were pinned. Availability of data and material The datasets used in the present study are available from the corresponding author on reasonable request.

Competing interest
The authors declare that they have no con ict of interest.