Treatment and prognosis of colorectal cancer with synchronous peritoneal metastases: 11-year single institute experience

Introduction

Colorectal cancer (CRC) is the third most common malignant neoplasm worldwide, including in China.1–3 Nearly 25% of patients with CRC have metastatic disease at diagnosis.4 The typical locations for distant metastases are the liver, peritoneum and lung, among which synchronous peritoneal metastases (SPM) have worse progression-free survival and overall survival (OS).5 6 Early studies have shown that the median OS of patients with colorectal peritoneal metastases was ≤8 months; however, with the improvement of treatment strategies, the median OS has increased to 42 months.7

SPM are defined as peritoneal metastasis diagnosed at the same time as the primary tumour and are observed in 5%–10% of patients with CRC.8 9 The main treatments for SPM include cytoreductive surgery (CRS), systemic chemotherapy and hyperthermic intraperitoneal chemotherapy (HIPEC). However, the treatment of patients with colorectal SPM remains controversial in terms of the necessity of primary tumour resection, indications for CRS and HIPEC and regimens for perioperative chemotherapy.10–12

Initially, CRC with SPM was viewed as a terminal-stage disease for which palliative treatment was the only choice.13 In recent years, promising outcomes have been observed for patients undergoing selective CRS plus HIPEC.14 15 However, the National Comprehensive Cancer Network guideline seems conservative regarding CRS plus HIPEC and states that ‘The goal of treatment for most peritoneal metastases is palliative, rather than curative, and primarily consists of systemic therapy with palliative surgery or stenting if needed for obstruction or impending obstruction. CRS and/or intraperitoneal chemotherapy can be considered in experienced centres for selected patients with limited peritoneal metastases for whom R0 resection can be achieved.’ This guideline has limited indications for the clinical recommendation of CRS plus HIPEC, and clinicians still tend to favour palliative care for patients with CRC with peritoneal metastases. Hence, the optimal treatment approach for patients with colorectal SPM should be investigated further.

In addition, most studies on the treatment and prognosis of patients with colorectal SPM were from Europe and North America4 16–18 and a study with a large sample size from China is yet to be conducted. Hence, this study aimed to summarise the treatment strategies of patients with colorectal SPM and assess their clinical outcomes, as well as to develop a prognostic model to predict their OS.

Patients and methods

Results

Patient characteristics

A total of 8631 patients with CRC were identified. Of these, 378 had SPM (4.4%; figure 1). Seven patients were excluded due to loss to follow-up or non-tumor-related death; finally, 371 patients were found eligible for this study.

Request permission

Figure 1

Study flowchart. Colorectal cancer with SPM in The Sixth Affiliated Hospital of Sun Yat-sen University from June 2007 to June 2018. CRC, colorectal cancer; PM, peritoneal metastasis; SPM, synchronous peritoneal metastases.

Of the 371 patients, 95 (26%) were aged >65 years. The median age of the entire cohort was 58 (range: 18–92) years. The proportion of men was 63% (235/371). The preoperative diagnostic rate of peritoneal metastases by imaging (CT or MRI) was 53%. Furthermore, 133 (36%) patients had extraperitoneal metastasis, 226 (61%) had a complete or incomplete digestive obstruction and 240 (65%) had high or moderate histological differentiation (table 1).

Table 1

View inline|View popup

Characteristics of 371 patients with colorectal SPM

Survival information

The median follow-up time was 44.0 months (95% CI, 33.6 to 54.4) for the entire population, and 282 patients (76.0%) died during follow-up. The median OS was 15.0 months (95% CI, 13.1.0 to 16.9) in the entire population, and the 3-year and 5-year OS rates were 23.7% and 16.9%, respectively. Kaplan-Meier analysis showed that undergoing chemotherapy (p<0.001), CRS (p<0.001) and HIPEC (p<0.001) were significantly associated with improved OS. Stratified by chemotherapy, the median OS was 18 months in the chemotherapy group and 10 months in the other group. Stratified by the CRS, the median OS of patients who did not receive CRS was 6 months, while that of patients who received CRS of CC0–1 and CC2–3 was 49 and 13 months, respectively. HIPEC was significantly associated with increased OS, with a median OS of 24 months in the HIPEC group and 12 months in the non-HIPEC group (figure 2). The median OS of patients with treatment scores of 0, 1, 2 and 3 was 5, 10, 21 and 29 months, respectively, with 3-year survival rates of 0%, 12.3%, 31.0% and 41.8%, respectively.

Request permission

Figure 2

Overall survival of patients with colorectal synchronous peritoneal metastases. (A) Overall survival stratified by chemotherapy (p<0.001). (B) Overall survival stratified by CRS (p<0.001). (C) Overall survival stratified by HIPEC (p<0.001). (D) Overall survival stratified by treatment scores (p<0.001). CC0, no macroscopic peritoneal tumour remains following cytoreduction; CC1, tumor nodules less than 2.5 mm persist following cytoreduction; CC2, residual disease measuring 2.5 mm to 2.5 cm; CC3, presence of tumor nodules greater than 2.5 cm, or a confluence of unresectable tumor nodules at any site within the abdomen or pelvis; CRS, cytoreductive surgery; HIPEC, hypothermic intraperitoneal chemotherapy.

Establishment and assessment of the nomogram for prognosis prediction

Variables with p<0.05 from the multivariate analysis were used to establish a nomogram for predicting prognosis (figure 3). This model could predict the 1-year, 2-year and 3-year OS rates of patients with SPM. Then, the performance of the nomogram was verified using the c-index, calibration curve and receiver operating characteristic (ROC) curve. The c-index of the model was 0.747 (95% CI, 0.474 to 1.020). The c-index, calibration and ROC curves indicated satisfactory prediction performance (figures 4 and 5). In order to increase practicality, we further constructed a dynamic nomogram, which can be found at https://qxssysu.shinyapps.io/CRCSPM_OS_DynNomapp.

Request permission

Figure 3

A nomogram model for predicting overall survival in patients with colorectal synchronous peritoneal metastases (a dynamic nomogram for patients with colorectal cancer with synchronous peritoneal metastases at https://qxssysu.shinyapps.io/CRCSPM_OS_DynNomapp). CA12-5, cancer antigen 12-5; CRS, cytoreductive surgery; HIPEC, hypothermic intraperitoneal chemotherapy; PCI, peritoneal carcinomatosis index.

Request permission

Figure 4

The calibration curves for validation of the nomogram. Predicted 1-year, 2-year and 3-year overall survivals for patients with colorectal synchronous peritoneal metastases are shown on the horizontal axis, and actual 1-year, 2-year and 3-year overall survivals are shown on the vertical axis. The c-index is 0.747 (95% CI, 0.474 to 1.020).

Request permission

Figure 5

Receiver operating characteristic (ROC) curves of the nomogram for predicting 1-year, 2-year and 3-year overall survival of patients with colorectal synchronous peritoneal metastases. AUC, area under the curve.

Discussion

Patients with colorectal SPM have a poor prognosis, and the eighth edition of the American Joint Committee on Cancer (AJCC) staging system classifies peritoneal1 metastases as M1c. In this single-centre study, we identified the prognostic factors associated with OS. Our findings suggest that patients with colorectal SPM who received comprehensive treatment achieved better prognoses and complete CRS should be considered for long-term survival benefits in patients with SPM of colorectal origin.

Previous studies from China have investigated the association between treatment strategies and OS in patients with colorectal SPM.20–23 However, their clinical significance was limited due to the small number of patients investigated,20–23 involving metachronous peritoneal metastases,20 involving appendiceal tumours22 or patients being treated only CRS plus HIPEC.23 Although numerous international multicentre cohort studies have been conducted on patients with colorectal SPM, they mainly focused on the benefits of CRS plus HIPEC or primary tumour resection.10 11 24 25 Compared with these studies, the conclusions of the present study are more general.

The prevalence of SPM was 4.4% in this study, similar to that reported in most previous studies.4 24 26 Our centre is one of the largest gastrointestinal treatment centres in South China. Thus, our data may be representative to some extent. In this study, HIPEC has been increasingly used, reflecting the acceptance of the CRS plus HIPEC treatment pattern by surgeons. Of note, between 2015 and 2018, 41% of patients with SPM received only one type of treatment, mainly treated by ostomy or systemic chemotherapy. In these patients, the volume of SPM might be too large to perform CRS. Another possible reason is that CRS was not widely accepted in the past years. Currently, most surgeons still resort to conservative treatment for SPM.

Our study showed that CRS, HIPEC and systemic chemotherapy are independently associated with OS. Patients who received all three treatments had a median OS of 29 months, which was longer than those receiving any one or two modalities of treatments. Previous studies showed that patients who received CRS plus HIPEC had a median OS of 32.4–62.7 months,16 27–30 which was longer than that observed in the present study. This might be because most patients who received CRS in our study did not undergo complete cytoreduction. However, patients who received CRS of CC0–1 had a median OS of 49 months, which was consistent with the findings reported in the PRODIGE 7 trail.7 Patients with SPM generally have an inferior status,31 and the morbidity and mortality rates of CRS plus HIPEC are 20%–40% and 3%, respectively.32 Therefore, the management of these patients requires the collaborative efforts of radiologists, medical oncologists, radiation oncologists and surgeons from several disciplines.

The prognostic nomogram showed moderate predictive ability, which can aid clinical decision making. Previous studies have proposed models for predicting the prognosis of CRC with peritoneal metastases. The most commonly mentioned scores are the Peritoneal Surface Disease Severity Score (PSDSS) and Colorectal Peritoneal Metastases Prognostic Surgical Score (COMPASS).33 34 PSDSS is used to predict the prognosis of patients with peritoneal metastases after CRS plus HIPEC in combination with clinical symptoms, CT-PCI and histopathological characteristics of the primary tumour. However, this scoring system is mainly used to select patients suitable for CRS plus HIPEC, and its clinical application is relatively limited. COMPASS further optimised the indices of PSDSS and visualised the prediction model without changing the application premise, and the c-index of the model was 0.720. Compared with PSDSS and COMPASS, the nomogram established in this study can be more widely used and has a better prediction ability, with a c-index of 0.747.

The poor prognosis of SPM emphasises the importance of early diagnosis. Unfortunately, imaging examinations are limited in detecting minor lesions.35 In our study, only 52.8% of patients with SPM were diagnosed using preoperative imaging examinations. Although positron emission tomography (PET)-CT has a higher detection rate, clinicians do not use it routinely because of its high cost. Therefore, it is important to accurately identify these high-risk patients and perform routine PET-CT. In addition, new techniques can also be considered for the early diagnosis of PM. Dong et al developed a radiomic nomogram to identify occult PM in patients with advanced gastric cancer, and the area under the curve (AUC) was 0.920 in the external validation cohort.36 We recently reported an image-based deep learning algorithm to identify SPM of colorectal origin based on CT images, and it has shown great potential in the prediction of SPM, showing an accuracy of 94.11% with an AUC 0.922.37 Furthermore, studies in recent years have indicated that circulating tumour DNA is of great value in detecting minimal residual disease and monitoring the recurrence and treatment response of CRC.38–40 Circulating tumour DNA from peritoneal lavage fluids may be helpful to identify patients at high risk of metachronous peritoneal metastases.

This study has some limitations. First, as this was a retrospective study, we relied on the accuracy of pre-input data. To be as accurate as possible, two doctors checked all the data. Second, in this study, chemotherapy was defined as receiving at least two consecutive courses of systemic chemotherapy during perioperative or palliative care, and a specific chemotherapy regimen; the application of targeted or immunological drugs and the course of chemotherapy were not subdivided, which could lead to some inherent potential bias. Finally, this study lacks validation, and the Surveillance, Epidemiology, and End Results database should be considered for validation in the future.

Conclusions

This observational study indicated that patients with colorectal SPM who received integrated treatment had better prognoses and that complete CRS should be considered for long-term survival benefits. The developed nomogram demonstrated good predictive power, which could be used for the prognostic assessment of patients with colorectal SPM and for selecting patients suitable for CRS plus HIPEC. Certainly, more refined stratification and randomised clinical trials are needed to address potential confounding factors and bias.