A proposed new Japanese classification of synchronous peritoneal metastases from colorectal cancer: A multi‐institutional, prospective, observational study conducted by the Japanese Society for Cancer of the Colon and Rectum

Abstract Aim To establish a new Japanese classification of synchronous peritoneal metastases from colorectal cancer. Methods This multi‐institutional, prospective, observational study enrolled patients who underwent surgery for colorectal cancer with synchronous peritoneal metastases. Overall survival rates were compared according to the various models using objective indicators. Each model was evaluated by Akaike's information criterion (AIC). The region of peritoneal metastases was evaluated by the peritoneal cancer index (PCI). Results Between October 2012 and December 2016, 150 patients were enrolled. The AIC of the present Japanese classification was 1020.7. P1 metastasis was defined as confined to two regions. The minimum AIC was obtained with the cutoff number of 10 or less for P2 metastasis and 11 or more for P3 metastasis. As for size, the best discrimination ability between P2 and P3 metastasis was obtained with a cutoff value of 3 cm. The AIC of the proposed classification was 1014.7. The classification was as follows: P0, no peritoneal metastases; P1, metastases localized to adjacent peritoneum (within two regions of PCI); P2, metastases to distant peritoneum, number ≤10 and size ≤3 cm; P3, metastases to distant peritoneum, number ≥11 or size >3 cm; P3a, metastases to distant peritoneum, number ≥11 and size ≤3 cm, or number ≤10 and size >3 cm; P3b, metastases to distant peritoneum, number ≥11 and size >3 cm. Conclusion This objective classification could improve the ability to discriminate prognosis in patients with synchronous peritoneal metastases from colorectal cancer.


| INTRODUC TI ON
Colorectal cancer is the second most common cause of cancer death in the United States and Japan. 1,2 Furthermore, the incidence of colorectal cancer has been increasing in Japan. 3,4 Peritoneal metastasis is one of the factors associated with a poor prognosis in patients with colorectal cancer and is found in approximately 5% of primary colorectal cancer cases. 5 Colorectal cancer with synchronous peritoneal metastases is classified as Stage IVC in the current AJCC Cancer Staging Manual. 6 The Japanese classification of colorectal carcinoma is published by the Japanese Society for Cancer of the Colon and Rectum (JSCCR). 7 In the present Japanese classification, synchronous peritoneal metastasis was classified as follows: P0, no peritoneal metastasis;
On the other hand, in specialized centers for peritoneal diseases, the peritoneal cancer index (PCI) has been used worldwide. 8 The PCI classifies the abdominal cavity into 13 regions. In each region, 0 to 3 points are given according to the size of peritoneal metastases. Therefore, the maximum PCI score is 39. The PCI has superior objectivity, but it seems cumbersome for general surgeons.
The aim of this study was to establish a new objective classification of peritoneal metastases from colorectal cancer in a multiinstitutional, prospective, observational study.

| Study design
The 28 member hospitals of the JSCCR were involved in this multiinstitutional, prospective, observational study. These hospitals joined a committee of the JSCCR, named "Grading of peritoneal metastasis from colorectal cancer." Patients who underwent surgery for colorectal cancer with synchronous peritoneal metastases between October 2012 and December 2016 were enrolled. Clinical and pathological information was registered within 3 months after surgery. Prognostic information was collected 3 years after surgery.
Written, informed consent was obtained from all patients before enrollment. Since synchronous peritoneal metastases from colorectal cancer are often found accidentally during surgery, written informed consent could be obtained after surgery in such cases. The ethics committees of the Japanese Society of the Colon and Rectum and each institution approved this study. In this study, tumor location was classified into right colon (velmiformis, cecum, ascending colon, transverse colon), left colon (descending colon, sigmoid colon), and rectum (RS, Ra, Rb).

| Surgical procedure
The surgical procedure was not determined by the protocol of this study, because there are a variety of conditions in patients with synchronous peritoneal metastases. Each surgeon made a decision about primary tumor resection and peritoneal metastasis resection.

| Data collection
All data were collected prospectively. The preoperative data included physical information, blood tests, and preoperative diagnosis. The information regarding peritoneal metastases included both the Japanese classification and the PCI. The size and number of peritoneal metastases in each region were recorded. Surgical procedures and other pathological information were sent within 3 months after surgery. Information regarding postoperative chemotherapy and outcomes was collected at least 3 years after surgery.

| Statistical analysis
Various classification models were constructed by combinations of region, number, and size of peritoneal metastases and compared.
The discrimination ability of each model was evaluated by Akaike's information criterion. Actuarial survival after surgery was depicted by Kaplan-Meier curves. The log-rank test was used to compare overall survivals. Differences in continuous variables were compared using the Kruskal-Wallis test. JMP 13 software (SAS Institute Japan, Ltd.) was used for data analysis. The data are expressed as medians and range or numbers of patients and percentages (%). A p value less than 0.05 was taken to indicate significance in this study. Table 1 shows the patients' characteristics for the entire cohort. A total of 150 patients were enrolled in this study; their median age was 66 years. The location of the primary tumor was the right colon in approximately half of the patients. The primary tumor was not resected in 24 patients in whom pathological T and N factors were not available. Cytoreductive surgery with hyperthermic intraperitoneal chemotherapy (HIPEC) was performed in only two patients.

| Clinical question 1 (range of P1 metastases)
One of the clinical questions was the range of P1 metastases, i.e., localized peritoneal metastases. In this study, 30 patients were classified as having P1 metastases. Of the 30 P1 metastases, 23 (77%) were confined to one area, and seven (23%) were confined to two areas. Therefore, all P1 metastases were confined to one or two areas.

| Clinical question 2 (cutoff number between P2 and P3 metastases)
Another clinical question was the cutoff number between P2 and P3 metastases. The AIC was evaluated according to the number of peritoneal metastases. The AIC according to the number of P2 ≤9, ≤10, ≤19, and ≤20 was 1023.3, 1020.7, 1021.2, and 1022.2, respectively. Therefore, the best cutoff number of P2 metastases was 10. Figure 2 shows the survival curves when P2 was defined as metastases to distant peritoneum (number of peritoneal metastases ≤10, Figure 2).

| Clinical question 3 (size of peritoneal metastases)
The present Japanese classification does not include the concept of size of peritoneal metastases. As a first step, the AIC was investigated P3, metastasis to distant peritoneum, number ≥11 or size >3 cm; P3a, metastasis to distant peritoneum, number ≥11 and size ≤3 cm, or number ≤10 and size >3 cm; and.
The proposed classification reflected the prognosis of patients with and without distant metastasis (p = 0.027 and 0.0017).

| DISCUSS ION
It has been reported that the present Japanese classification of peritoneal metastases from colorectal cancer is easy to use and useful in predicting prognosis. 5,[9][10][11][12] In addition, the present Japanese classification is also useful in determining whether synchronous peritoneal metastases should be resected during surgery. The present Japanese classification predicted the prognosis of patients with synchronous peritoneal metastases from colorectal cancer very well. On the other hand, it has been pointed out that the present Japanese classification seems less objective. Therefore, a major aim of this study was to add objectivity to the present Japanese classifi-  9 reported that it was appropriate that P3 metastases be defined as >10 peritoneal metastases in their multi-institutional, retrospective study. Their cutoff number between P2 and P3 metastases was consistent with that in the present study.
The third clinical question related to whether the concept of size of peritoneal metastases could improve the ability of the Japanese classification to discriminate prognosis. The PCI uses the cutoff sizes of 5 and 50 mm. However, in the current study, there was no difference in survival according to the sizes of 5 and 50 mm. The subclassification using the cutoff size of 30 mm most effectively improved the ability to discriminate prognosis in the current study.
The present Japanese classification has been used as an indicator for R0 resection in patients with synchronous peritoneal metastases. In the Japanese guidelines for the treatment of colorectal cancer, complete resection is strongly recommended for P1 metastases. 14 Complete resection is recommended for P2 metastases when they are easily resectable. The R0 resection rate for P1 and P2 metastases using the proposed classification in the current study was comparable to that in the present Japanese classification.
There are some limitations in this study. First, although the data were collected prospectively using the same case report form, the surgical procedure was dependent on each surgeon. Most syn- In conclusion, this study succeeded in establishing a new objective classification of synchronous peritoneal metastases from colorectal cancer while following the present Japanese classification. In addition, the concept of the size of peritoneal metastases improved the ability to discriminate prognosis.

ACK N OWLED G M ENTS
This study was performed as part of a project study of the Japanese Society for Cancer of the Colon and Rectum.

E TH I C S S TATEM ENT
Approval of the research protocol: The study protocol followed the ethical guidelines of the 2008 Declaration of Seoul and was approved by the Institutional Review Board at the JSCCR and each institution.
Informed Consent: A written informed consent was obtained from each patient.
Registry and the Registration No. of the study/Trial: N/A.
Animal Studies: N/A.