Introduction

Colorectal cancer (CRC) is the third most common cause of death in Japan, and its prevalence is increasing. Synchronous peritoneal metastases (PM) are found in 4–10 % of patients with primary CRC, and are a well-known poor prognostic factor [1, 2]. The TNM Classification of Malignant Tumors (TNM) classifies malignant CRC with PM as stage IVB, and includes patients with multiple organ metastases [3]. The degrees of PM are organized according to focal number, range, and size of PM in most proposed classifications for prognostic prediction and decisions about treatment strategies [4–8]. Previous reports have identified histological type, the presence of lymph node metastases, and R0 surgery as prognostic variables for CRC with PM [9, 10], but there have been no detailed reports of clinical outcomes of CRC with PM after R0 surgery. Although the prognostic factors and risk factors for recurrence are unclear, this information can be important for postoperative follow-up and decisions about adjuvant chemotherapy administration.

The aim of this study was to investigate the factors affecting recurrence and prognosis in CRC with synchronous PM after R0 resection as an aid in determining appropriate postoperative follow-up and treatment strategies.

Methods

Patients

Data were collected from 1217 consecutive patients with CRC and synchronous PM who underwent initial surgery between January 1997 and December 2007, using the databases of 16 referral hospitals of the Japanese Society for Cancer of the Colon and Rectum (JSCCR). Among the 1217 patients, 224 underwent R0 resection, and 172 of these had sufficient background records. From this group, 72 patients with PM localized to the adjacent peritoneum (P1 according to the classification of JSCCR [6]) and no distant metastases were enrolled in the study for final analysis. No patients in our series underwent hyperthermic intraperitoneal chemotherapy (HIPEC). The study protocol was approved by the ethics committee of the JSCCR and the local institutional review board.

Parameters

The parameters used in this study were as follows: age, gender, location of primary tumor, size of primary tumor, lymph node dissection, number of dissected lymph nodes, histology, depth of tumor invasion, lymph node metastases, postoperative chemotherapy, lymphatic invasion, venous invasion, grade of PM according to JSCCR classification, maximum size of PM, number of PM, extent of PM, and preoperative serum levels of the carcinoembryonic antigen (CEA) and carbohydrate antigen 19-9 (CA19-9). Clinical and pathological data were recorded according to the JSCCR classification system [6]. Depth of tumor invasion and lymph node (LN) metastases were classified according to the seventh edition of TNM classification system [3]. The cutoff values for CEA and CA19-9 were 5.0 ng/mL and 37.0 U/mL, respectively. Serial univariate analyses were conducted for each number of dissected lymph nodes, 39 times in total, to determine the cut-off values for the number of dissected lymph nodes in the analyses of factors affecting recurrence and prognosis, and the numbers of dissected lymph nodes with the lowest p value were defined as the cut-off values. Surgical charts were thoroughly reviewed and scrutinized with respect to the narrative descriptions, figures, and comments. Patients were classified according to the size of the largest disseminated lesion, number of disseminated lesions, and numbers of regional dissemination lesions. The numbers and sizes of PM nodules were converted to numerical data based on the original description provided in the surgical charts. For example, “a few” disseminated lesions were recorded as “≤3”, “miliary” and “rice-sized” lesions were recorded as “≤5 mm,” and “azuki bean-sized” and “thumb-sized” lesions were recorded as “5–20 mm.” The site of the PM was classified into nine areas, and the extent of the PM was evaluated as the number of sites, according to a previous study [11]. PM grade was classified using the Japanese classification of PM defined by JSCCR as follows: P0, no PM; P1, metastasis localized to the adjacent peritoneum; P2, metastases limited to the distant peritoneum; and P3, diffuse metastases in the distant peritoneum [6]. Intensive chemotherapy was defined as a regimen containing fluorouracil (5FU) or 5FU derivatives and camptothecin (CPT-11) or oxaliplatin. The median follow-up period was 30.7 months (range 0.5–209.0 months).

Statistical analysis

Differences between continuous and categorical variables were identified using the Mann–Whitney U test and Chi-squared or Fisher’s exact test, respectively. Factors affecting recurrence were evaluated using binomial logistic regression analyses for factors that were compared by Ch-squared or Fisher’s exact test to a level of p < 0.1 in order to determine factors independently associated with recurrence. All data were expressed as medians and ranges, and survival rates were calculated using the Kaplan–Meier method and compared using the log-rank test to examine associations between survival and the clinical and pathological variables. Survival analyses were performed using Cox regression models for factors identified as influencing patient survival in Kaplan–Meier analyses, and the differences were identified using the log-rank test to a level of p < 0.1. Multiple comparisons were performed by Bonferroni adjustment. The differences between groups and associations with survival of CRC patients with PM were considered significant at p < 0.05. All analyses were performed using JMP 11 statistical software (SAS Institute Inc., Cary, NC, USA).

Results

Characteristics of the patients (Table 1)

The median patient age was 66.0 years (age range 34–91 years). Among the 72 patients, 21 patients had a primary tumor in the right colon, including the cecum, ascending colon, and transverse colon; 36 patients had a tumor in the left colon, including the descending colon, sigmoid colon, and rectosigmoid colon; and 27 patients had a tumor in the rectum. Forty-five patients (62.5 %) had PM measuring 5 mm or smaller. Fifty-nine patients (81.9 %) had < 4 PM. One regional dissemination lesion was observed in 66 (91.7 %) of the 72 patients. Adjuvant chemotherapy was administered after R0 surgery in 44 patients (61.1 %), among whom 8 patients (11.1 %) received intensive chemotherapy.

Table 1 Univariate and multivariate analyses of factors affecting recurrence in colorectal cancer patients with peritoneal metastases after R0 resection
Full size table

Factors affecting recurrence after R0 surgery

Recurrence after R0 surgery was observed in 51 patients (70.8 %). An association between lymph node metastases (p = 0.06), venous invasion (p = 0.06), and number of dissected lymph nodes (p = 0.09) was revealed using the Chi-squared test (Table 1). Among these factors, logistic regression analyses revealed that lymph node metastases independently affected recurrence (Table 1). Hematogenous recurrence was noted in 41.7 % of all patients who experienced recurrence; 29.2 % of these patients experienced peritoneal recurrence (Table 2).

Table 2 Recurrence site after R0 surgery
Full size table

Among patients who received postoperative chemotherapy, no significant differences in rates of recurrence were found between those receiving intensive adjuvant chemotherapy and non-intensive chemotherapy. There was also no significant difference in recurrence rates between patients who underwent surgery after 2005 (17 of 24 cases; 70.8 %), when oxaliplatin was approved for use in CRC in Japan, and those before 2005 (34 of 48 cases; 70.8 %). In addition, no significant difference in post-surgical follow-up duration was found between the period after 2005 (median 34.4 months, range 13.9–70.3) and the period until 2005 (median 27.3 months, range 0.5–209.0) (p = 0.30).

The median disease-free interval was 11.5 months (range: 0.9–55.7) in 47 patients whose recurrence date was known. Twenty-three (48.9 %) of 47 patients experienced recurrence within 12 months after R0 surgery, and 33 patients (70.2 %) within 18 months (Fig. 1). Recurrence within 18 months was observed in two (40.0 %) of five patients who underwent intensive postoperative chemotherapy, in 19 (70.4 %) of 27 patients who underwent non-intensive postoperative chemotherapy, and in 12 (80.0 %) of 15 patients who underwent no postoperative chemotherapy. Surgical treatment for recurrent lesions after R0 surgery was performed in nine patients (20.0 %). There was no significant difference between patients who experienced recurrence after 18 months (3 of 15 cases; 20.0 %) and those in whom recurrence was observed within 18 months (6 of 33 cases; 18.2 %, p = 0.88) (Fig. 1).

Fig. 1
figure 1

Disease-free interval after R0 surgery in patients with peritoneal metastases of colorectal cancer, and rates of surgery for recurrent lesions according to disease-free interval. Intensive ChT: postoperative intensive chemotherapy

Full size image

Factors affecting prognosis after R0 surgery

The overall 3- and 5-year survival rates of patients with PM after R0 surgery were 46.3 and 36.4 %, respectively. Having eight or fewer dissected lymph nodes (p = 0.01) and non-intensive or no postoperative chemotherapy (p = 0.03) were associated with poor prognosis using the log-rank test (Table 3), and both were found to be independent poor prognostic factors using the Cox proportional hazard model (Table 3).

Table 3 Univariate and multivariate analyses of factors affecting survival in colorectal cancer patients with peritoneal metastases
Full size table

The number of lymph node metastases was significantly higher in patients with more than eight dissected lymph nodes (median 3.5, range 0–42) than in those with eight or fewer (median 0, range 0–5) (p = 0.003). There were also significantly more patients with lymph node metastases among those with more than eight dissected lymph nodes (54 of 64 patients, 84.4 %) than eight or fewer (3 of 8 patients, 37.5.0 %) (p = 0.008).

The prognosis was significantly better in patients who underwent surgery after 2005 than in those who underwent surgery before 2005 (Fig. 2a). Furthermore, among patients who underwent postoperative chemotherapy, the prognosis was significantly better among those receiving intensive chemotherapy using CPT-11 or oxaliplatin than in those receiving non-intensive chemotherapy (Fig. 2b). Among patients who experienced recurrence, the 3- and 5-year survival rates after recurrence were 10.1 and 5.0 %, respectively. Prognosis after recurrence within 18 months was significantly better than that after 18 months (Fig. 3). There was no significant difference between patients who underwent surgery for recurrent lesions and those who underwent other treatments (p = 0.92).

Fig. 2
figure 2

Overall survival curves according to a era (p = 0.047) and b regimen of adjuvant chemotherapy (p = 0.04) after R0 surgery

Full size image
Fig. 3
figure 3

Overall survival curves after recurrence according to disease-free interval (p = 0.008)

Full size image

Discussion

Synchronous PM are a poor prognostic factor for patients with primary CRC [12]. The majority of CRC patients with synchronous PM have hematogenous metastases, and most patients with PM are managed exclusively with palliative care for symptom relief, with or without chemotherapy [9]. In the present study, among all patients with synchronous PM, R0 surgery was performed in 18.4 %; this figure was similar to those of previous reports on synchronous liver metastases without neoadjuvant chemotherapy [13–15], although an increase in the resection rate has been reported with recent preoperative chemotherapy for liver metastases [16–18]. Aggressive surgery can be performed for prognostic improvement in synchronous PM of CRC if it is performed as R0 surgery. Aggressive cytoreductive surgery (CRS) combined with hyperthermic intraperitoneal chemotherapy (HIPEC) has been attempted as a way to improve treatment outcomes for CRC patients with PM [7, 19–21], and significant prognostic improvement was shown with this approach in a randomized controlled trial [22]. It has not been used as standard treatment, however, because of frequent and significant morbidity and mortality. Furthermore, a recent retrospective international registry study showed that complete cytoreduction was associated with improved overall survival, while HIPEC was not [23]. Recently, intensive systemic chemotherapy has been expected to improve survival.

Some retrospective multi-center analyses have reported that R0 resection is an independent favorable prognostic factor in patients with PM caused by CRC [9, 10]. In the present study, however, the recurrence rate was 70.8 % even in localized PM cases, and this high rate of recurrence was similar to that found in previous cases of metastases to other organs, even when R0 surgery was performed [24, 25]. We found that lymph node metastasis was an independent factor affecting recurrence after R0 resection in patients with PM caused by CRC. Postoperative treatment other than intensive chemotherapy and a total of eight or fewer dissected lymph nodes were independent poor prognostic factors. Focal number, range, and size of PM have been used to classify the progression of PM [4–8], and the JSCCR classification according to focal number and range of PM has been reported as one of the factors affecting R0 resection and prognosis [9]. However, the number of patients eligible for R0 surgery is limited, and all of those in the present study belonged to P1. As a result, the focal number, range, and size of PM, which are used by many classifications, were not associated with recurrence or prognosis in patients after R0 surgery. In the present study, a total of eight or fewer dissected lymph nodes was one of the factors of poor prognosis. Furthermore, among these patients, the number of node metastases and rates of occurrence of node metastasis were significantly lower than in other patients. A minimum of 12 lymph nodes should be examined to prevent stage migration and to establish the N stage, particularly in stage II disease [26]. In PM cases in which R0 surgery was possible, our outcomes suggest that the harvesting of more than eight lymph nodes through suitable lymph node dissection for a cure was important for the prediction of prognosis.

Postoperative intensive chemotherapy was an independent favorable prognostic factor, and prognosis was significantly better among patients who underwent surgery after 2005, the year in which oxaliplatin was approved in Japan for use in colorectal cancer, than in those who underwent surgery before 2005. Furthermore, the prognosis was significantly better in patients who received postoperative intensive chemotherapy versus non-intensive or no chemotherapy, although there was no significant difference in recurrence rate between the two groups. However, the number of patients with recurrence after 18 months was greater among those who received postoperative intensive chemotherapy than those who did not. These outcomes suggest that intensive chemotherapy contributed to prognostic improvement by extending the disease-free period after R0 surgery. A pooled analysis of two large prospective randomized trials (N9741 and N9841) investigating the efficacy of chemotherapy for metastatic CRC showed significantly shorter median survival among patients with PM than those without PM [27]. The efficacy of chemotherapy, even intensive chemotherapy such as FOLFOX/FOLFIRI ± bevacizumab/cetuximab, in treating PM has not yet been demonstrated [27–29]. The results of analysis in the present study may reflect the more frequent hematogenous than peritoneal recurrence after R0 surgery for PM of CRC, and thus may explain why adjuvant chemotherapy after R0 surgery was effective for reducing recurrence and improving prognosis. Recurrence was observed within 12 months after R0 surgery in almost half of the patients with PM, and within 18 months in approximately 70 % of patients. Recurrence, particularly hematogenous recurrence, should thus be a significant concern for the first 18 months after R0 surgery for CRC with PM. The disease-free interval has been associated with prognosis after recurrence, and prognosis was poor in patients who had recurrence within 18 months. Complete surgical removal of cancer, followed by adjuvant therapy, may extend the disease-free period and improve prognosis.

There were several limitations with this retrospective design. Some cases had to be excluded from further analysis because of a lack of sufficient background records. Nonetheless, our observations warrant further consideration and validation in a larger series of patients with CRC.

In conclusion, lymph node metastasis was an independent factor for recurrence after R0 resection in patients with PM of CRC. In addition, the number of dissected lymph nodes and postoperative intensive chemotherapy were independent prognostic factors. In cases of CRC with PM, the harvesting of more than eight lymph nodes through suitable lymph node dissection for cure and the administration of intensive adjuvant chemotherapy after R0 surgery are recommended for greater prediction accuracy and improved prognosis.