Japanese multicenter prospective study investigating laparoscopic surgery for locally advanced rectal cancer with evaluation of CRM and TME quality (PRODUCT trial)

Abstract Aim In Japan, we have not been able to validate the results of laparoscopic surgery for locally advanced rectal cancer using the universal index “circumferential resection margin (CRM).” Previously, we established a semi‐opened circular specimen processing method and validated its feasibility. In the PRODUCT trial, we aimed to assess CRM in patients with locally advanced rectal cancer who underwent laparoscopic rectal resection. Methods This was a multicenter, prospective, observational study. Eligible patients had histologically confirmed rectal adenocarcinoma located at or below 12 cm above the anal verge with clinical stage II or III and were scheduled for laparoscopic or robotic surgery. The primary endpoint was pathological CRM. CRM ≤1 mm was defined as positive. Results A total of 303 patients operated on between August 2018 and January 2020 were included in the primary analysis. The number of patients with clinical stage II and III was 139 and 164, respectively. Upfront surgery was performed for 213 patients and neoadjuvant therapy for 90 patients. The median CRM was 4.0 mm (IQR, 2.1‐8.0 mm), and CRM was positive in 26 cases (8.6%). Univariate and multivariate analyses demonstrated that a predicted CRM from the mesorectal fascia of ≤1 mm on MRI was the significant factor for positive CRM (P = .0012 and P = .0045, respectively). Conclusion This study showed the quality of laparoscopic rectal resection based on the CRM in Japan. Preoperative MRI is recommended for locally advanced rectal cancer to prevent CRM positivity.


| INTRODUC TI ON
Rectal resection still remains the mainstay in multidisciplinary treatment for locally advanced rectal cancer (LARC), and the quality of surgery is directly associated with postoperative local recurrence. [1][2][3] Although the local recurrence rate for Dukes stage C was as high as 40% two decades ago, the establishment of total mesorectal excision (TME) since 1982 and adoption of neoadjuvant chemoradiotherapy as standard treatment have decreased the rate to approximately 5%. [4][5][6] Different from Western countries, however, Japanese surgeons have adopted a unique strategy for LARC in which rectal resection with lateral lymph node dissection (LLND) is performed first regardless of preoperative staging, followed by postoperative chemotherapy. 1 Despite this discrepancy, the long-term results are comparable between Japan and Western countries. 7 Although the treatment strategies evolved individually in Japan and in Western countries, the addition of LLND after neoadjuvant CRT can further prevent recurrence at the lateral pelvic cavity for patients with swollen nodes at baseline. 8 This indicates that not all metastasized lateral lymph nodes can be eradicated with CRT alone, urging Western surgeons to indicate LLND for high-risk patients.
Japanese surgeons started to use neoadjuvant CRT to prevent local recurrence that cannot be prevented by LLND. The optimization of combination treatment with neoadjuvant CRT and LLND would be a key factor in completely preventing local recurrence after surgery for LARC. 9 The indication for the combination of neoadjuvant CRT and/or LLND has to be discussed on the premise that the quality of rectal resection is appropriately ensured. However, the quality of Japanese surgery cannot have been evaluated in a pathological manner that is comparable to the method used in Western countries. According to the Japanese guidelines, the mesorectum is dissected off completely to harvest the perirectal lymph nodes and the rectum longitudinally opened to assess the macroscopic features of the tumor, which have long been emphasized for the evaluation of malignant potential, 10,11 leaving an inappropriate specimen for measuring the circumferential resection margin (CRM). Due to these differences, we have not been able to assess the pathological CRM to date, and the results of LARC treatment including LLND in Japan have not been accurately interpreted in Western countries. To resolve the inability to measure CRM in Japanese practice, we established a semi-opened circular specimen processing method and validated its feasibility 12,13 to successfully measure pathological CRM. Based on this background, we aimed to prospectively assess the quality of laparoscopic surgery for LARC in Japan in a multicenter study by assessing the universal standard, CRM.

| Study design
This was a multicenter, prospective, observational study conducted in Japan. A total of 18 institutions from the Japan Society of Laparoscopic Colorectal Surgery participated in the study. Eligible patients were ≥20 years old, had histologically confirmed rectal CRM in patients with locally advanced rectal cancer who underwent laparoscopic rectal resection.
Methods: This was a multicenter, prospective, observational study. Eligible patients had histologically confirmed rectal adenocarcinoma located at or below 12 cm above the anal verge with clinical stage II or III and were scheduled for laparoscopic or robotic surgery. The primary endpoint was pathological CRM. CRM ≤1 mm was defined as positive.
Results: A total of 303 patients operated on between August 2018 and January 2020 were included in the primary analysis. The number of patients with clinical stage II and III was 139 and 164, respectively. Upfront surgery was performed for 213 patients and neoadjuvant therapy for 90 patients. The median CRM was 4.0 mm (IQR, 2.1-8.0 mm), and CRM was positive in 26 cases (8.6%). Univariate and multivariate analyses demonstrated that a predicted CRM from the mesorectal fascia of ≤1 mm on MRI was the significant factor for positive CRM (P = .0012 and P = .0045, respectively).

Conclusion:
This study showed the quality of laparoscopic rectal resection based on the CRM in Japan. Preoperative MRI is recommended for locally advanced rectal cancer to prevent CRM positivity.

K E Y W O R D S
laparoscopy, magnetic resonance imaging, margins of excision, rectal neoplasms, total mesorectal excision adenocarcinoma located ≤12 cm above the anal verge with clinical   stage II or III (T3N0M0, T1-4aN1-2M0), and scheduled for laparoscopic or robotic surgery. The deepest part of the tumor could be   diagnosed as T4a, a tumor extending above the peritoneal reflection, and be regarded as eligible on the condition that the adjacent organ was not invaded. Rectal magnetic resonance imaging (MRI) was used for the assessment of T and N staging, and the clinical stage was assessed based on the images before neoadjuvant therapy. The indication for neoadjuvant therapy, surgical approaches, or LLND were at the discretion of each hospital. Exclusion criteria were as follows: a history of active double cancer (synchronous cancer, or metachronous cancer with disease-free interval <5 years), cancer invading the adjacent organs, or psychiatric or addictive disorders that affected compliance with the protocol. The target sample size was 300, considering that previous randomized controlled trials comparing laparoscopic surgery to open surgery included approximately 200 patients for laparoscopic surgery, [14][15][16] and that the es-

| Outcomes
The primary outcome was a pathological CRM measured using the semi-opened circular specimen processing method. CRM was defined as negative if the distance between the closest tumor invasion and dissected plane was more than 1 mm. The secondary outcomes included the quality of TME, surgical and pathological findings, disease-free survival, overall survival, and local recurrence rate. Our original plan was to report short-term results, including CRM, which is the primary endpoint of this study, first and then demonstrate long-term results after completing a follow-up of all patients according to the previous randomized controlled trials, including ALaCaRT or Z6051. 5,6 The quality of the mesorectal excision was categorized as complete, nearly complete, or incomplete according to the Dutch TME trial: complete, smooth surface of mesorectal fascia with all fat contained in the enveloping fascia; nearly complete, the mesorectal envelope was intact except for defects no more than 5 mm deep, with no loss of mesorectal fat; incomplete, low bulk mesorectum with defects down onto the muscularis propria and/or a very irregular circumferential resection margin. 17 The site of the lower border of the primary tumor (upper or lower) was categorized according to the subclassification of the 12 cm of rectum into equal halves. In assessing the clinical and pathological stage, the lateral lymph node was regarded as regional nodes according to the Japanese guidelines, and tumors with isolated metastasis to lateral lymph nodes were categorized as stage III.

| Pathological assessment
The resected specimen was photographed from four directions to confirm the quality of the dissected mesorectal fascia: anteriorly, from right, from left, and posteriorly. The procedure for semi-opened circular specimen processing was described in detail in our previous reports. 12 Briefly, the area of the rectum between 2 cm above and below the borders of the rectal cancer is not incised and the corresponding mesorectum is left attached to the rectum in order to measure the CRM. In assessing the CRM, if the tumor invasion in lymph node metastasis, extramural vascular invasion (EMVI), or tumor deposit is closer to the dissected plane than the main tumor invasion, the closer distance is recorded as the CRM. The feasibility of this procedure for pathological assessment was previously verified in a multicenter, prospective, observational study, thereby confirming that the quality of semi-opened circular specimen processing could be maintained in our study group. 13

| Statistical analysis
Statistical analyses were performed using JMP pro 15.1.0 software (SAS Institute, Cary, NC, USA). The results are expressed as the number of cases evaluated for categorical data, or as the median and interquartile range (IQR) for quantitative data. Univariate analyses were performed using Fisher's exact test or the Mann-Whitney U test as appropriate. Factors associated with CRM positivity were assessed using a multivariate logistic regression model. p<0.05 was considered to be statistically significant.

| Patient background and tumor characteristics
Between August 2018 and January 2020, a total of 308 patients were enrolled in this study ( Figure 1). Five patients were excluded after enrollment: two patients were ineligible because the tumor invaded the adjacent organs at baseline diagnosis, one patient withdrew consent, and two patients were ineligible for other reasons.

| Pathological outcomes
Pathological outcomes are shown in Table 3. The median length of pathologically assessed CRM, the primary endpoint of this study, was 4.0 mm (IQR, 2.1-8.0 mm), and CRM was ≤1 mm in 26 cases, corresponding to a positivity rate of 8.6%. In these 26 cases, the CRM was positive at the site of the main tumor in 18 cases and positive at metastasized lymph nodes in six cases, a tumor nodule in one case, and at the intra-lymphatic duct invasion in one case.
Of the eight cases in which CRM was positive at a site other than the main tumor, there were no cases in which CRM status could be diagnosed correctly based on MRI. Regarding pathological staging, pCR was confirmed in 14 cases, and two, 76, 104, and 107 cases were stage 0, I, II, and III, respectively. The pCR rate corresponded to 15.6%. Metastasis to lateral lymph nodes was confirmed in 12 cases, including three cases with isolated metastasis to lateral lymph nodes. In evaluation of the TME quality, 293 cases were complete and 10 nearly complete. The distal resection margin (DRM) was positive in one case. Table 4 summarizes the results of univariate analyses for the association between CRM positivity and the preoperative and operative variables. The predicted CRM from the mesorectal fascia being ≤1 mm on baseline MRI was found to be a significant risk factor for pathologically positive CRM (P = .0012). Among the 90 cases with neoadjuvant therapy, 84 underwent a second MRI to assess the therapeutic effect. The predicted CRM was positive on MRI in seven cases, four of which were diagnosed as pathologically positive. In contrast, among 77 cases with a negative predicted CRM on MRI, six were pathologically positive, which was significantly lower than the number predicted CRM to be positive (P = .0001).

| Analysis of the risk factors for positive CRM
Although predicted CRM on MRI was suggested to be a risk factor after neoadjuvant therapy, we used the baseline MRI findings as candidate factors in the subsequent analysis because it was difficult to universally diagnose MRI after neoadjuvant therapy, which can be affected by therapeutic modifications. 18 The multivariate analysis was carried out using the independent factors, including the predicted CRM from the mesorectal fascia on MRI and tumor stage cT4a, as well as tumor site and distance from the anal verge, which are the established risk factors for CRM positivity in the MERCURY II study. 19 The cut-off value for the distance from the anal verge was defined as 60 mm, considering the fact that low rectal cancer located ≤6 cm from the anal verge was regarded as a definite risk factor for CRM positivity in rectal cancer. 19 As shown in Table 5, a predicted CRM from the mesorectal fascia ≤1 mm on baseline MRI was demonstrated to be a significant factor, but the other factors were not shown to be related to the CRM positivity.

| DISCUSS ION
In this study, we assessed the pathological CRM for LARC operated on patients in Japan in a multicenter prospective study using the semi-opened circular specimen processing method. Until now, we have not been able to validate the oncological results of rectal cancer treatment in Japan by directly comparing it to the results in   14,16 In analyses of the long-term results of the above two trials, laparoscopic surgery was eventually demonstrated not to be a risk factor for disease-free survival, though pathological CRM was the single, poor prognostic factor. 5,6 These data suggested that, in the era of multimodal therapy for locally advanced rectal cancers, obtaining a CRM of more than 1 mm is the most crucial factor to achieving sufficient curability regardless of the surgical approach. In this regard, assessment of CRM is essential to validate the surgical quality and predict patient prognosis. The CRM in this study was shown to be 8.6%, and approximately 70% of the cases were positive at the site of the main tumor. Intriguingly, one-third of the patients were positive at a site other than the main tumor, suggesting the importance of preoperative assessment in tumor expansion in the mesorectum. The positivity rate of CRM was lower in the COREAN (2.1%) and ALaCaRT (6.7%) trials, and higher in the COLOR II (9.5%) and Z6051 (12.1%) trials, compared to the rate in the PRODUCT trial.
In interpreting the differences in CRM positivity, the differences in the enrolled patients should be considered. In contrast to the other trials, neoadjuvant therapy was administered in as few as 30% of cases in the presented study, and NAC was used in roughly half of the patients. Furthermore, more advanced cases were enrolled in this study based on the pathological tumor stage, and more than   patients is going to be carried out in a future ancillary analysis.
In conclusion, this study has demonstrated the results of Japanese CRM after laparoscopic rectal resection for LARC for the first time. CRM positivity was found in 8.6% of cases, one-third of which were positive at a site other than the main tumor. MRI should be carried out for patients with possible LARC at baseline to decide on the surgical plane. The rate of positive CRM shown in this study will play an important role as a reference value in future studies.

ACK N OWLED G EM ENT
The following doctors cooperated in this study: Hasegawa Tadashi