Peritoneal carcinomatosis index predicts survival in colorectal patients undergoing HIPEC using oxaliplatin: a retrospective single-arm cohort study

Peritoneal carcinomatosis (PC) affects up to 30% of colorectal cancer patients and remains the second leading cause of death after liver metastasis [1, 2]. PC treated with palliative chemotherapy alone is associated with a median survival of 5.2–12.6 months [3]. However, in the early 1990s, complete cytoreductive surgery (CRS) combined with hyperthermic intra-peritoneal chemotherapy (HIPEC) was introduced by Dr. Sugarbaker as a new treatment for isolated colorectal and appendiceal PC [4]. Few randomized trials are available studying CRS/HIPEC; however, Verwaal et al. showed in their landmark randomized trial from the Netherlands that colorectal PC treated with CRS/HIPEC using mitomycin C had a median survival of 22.2 months compared to 12.6 months when treated with chemotherapy alone [5, 6]. Multiple retrospective studies, such as Elias et al., subsequently demonstrated an improved 5-year overall survival of 51% when PC was treated with CRS/HIPEC compared with 13% when these patients were treated with chemotherapy alone [3]. As a result, a curative intent complete cytoreduction combined with HIPEC is now considered the standard of care for PC of colorectal origin with an expected long-term survival of 40% at 5 years [7].

The cytoreductive surgery has been increasingly standardized over the years with complete removal of macroscopic carcinomatosis achieved through radical organ resection, omentectomy, peritoneal stripping, and ablation of any disease that cannot be resected [8]. Contraindications to proceeding with cytoreduction include extensive and unresectable mesenteric, small bowel, or hepatoduodenal involvement [9]. The peritoneal carcinomatosis index score is also now universally used to evaluate extent of disease, communicate between clinicians and investigators, and has been found to be prognostic regarding patient survival.

Unfortunately, despite the consensus on the importance of CRS/HIPEC in the management of colorectal PC, there has not been consensus on the optimal HIPEC approach, standardization of the protocol, nor on the best chemotherapeutic agent to use during the procedure. HIPEC can be delivered with the abdomen open [5, 6] or with the skin temporarily closed using large bore drains and a pump for chemotherapy circulation [9, 10]. Intra-peritoneal chemotherapy can be given at room temperature, post-operatively, using intra-abdominal drains (EPIC) [11, 12], instead of the heated (42–43 °C) intra-operative chemotherapy given during HIPEC [11, 13]. Some protocols call for IV chemotherapy to be infused at the time of HIPEC, [14] while others do not [5, 6]. Several different chemotherapeutic agents have been used for intraperitoneal chemotherapy including mitomycin C [5, 6, 10, 15], cisplatin [16‐18], oxaliplatin [14].

Given the wide heterogeneity of methodology and approaches to CRS/HIPEC in both clinical practice and the literature, there is great benefit from conducting a study to assess the survival benefit in colorectal PC patients in the setting of a homogeneous protocol with a homogeneous chemotherapeutic agent. Therefore, we present here perioperative and oncologic outcomes from a cohort of colorectal PC patients presenting to a large Canadian referral center. The primary outcomes of overall survival and disease-free survival compare favorably with the literature, and secondary endpoints including complications were also acceptably low. On multivariate analysis, we confirmed the well-known prognostic value of PCI in predicting long term OS and DFS.

Of the 113 PC patients who underwent laparotomy at our center, 91 were resectable and underwent CRS/HIPEC. Ninety patients achieved a CCR score of 0, while one patient had a CCR score of 1. Twenty-two patients were considered unresectable during exploration at our center, were closed without cytoreduction, and went on to receive more systemic chemotherapy. Five of those patients were brought back for a second-look laparotomy, but all remained unresectable and were not cytoreduced (Fig. 1, schematic; Table 1).

Patient demographics are described in Table 1 for the entire cohort (N = 113). Median age was 55 years (range 33–71), with 59 males (52%) and 55 females. Median ASA score was II. Median follow-up was 24 months (range 0–137). Most tumors were cancers of the colon (N = 102, 90%), with fewer rectal tumors (N = 11, 10%). Histological distribution was classical intestinal adenocarcinoma (N = 92, 81.4%) and mucinous adenocarcinoma (N = 21, 18.6%) (Table 1).

Fifty-three (46.9%) patients from our entire cohort presented with synchronous stage IV peritoneal disease. Most of these patients had been previously explored at other centers, given systemic chemotherapy, and referred to our center. Regarding their initial exploration, 17 patients had only the primary tumor resected, 28 had the primary resected with an incomplete cytoreduction, 4 patients were explored without any resection, and 4 patients had a complete cytoreduction and HIPEC at initial exploration at our center (Table 1).

The remaining 60 (53.1%) patients presented with metachronous peritoneal disease. Four patients (3.5%) presented initially with stage IV disease in other metastatic locations, 38 (33.6%) with stage III disease, 17 (15.0%) with stage II disease, and 1 (0.9%) patient with stage I disease (Table 1).

For the patients who underwent CRS/HIPEC, 81 were colon cancer as the primary site and 10 were of rectal origin (Table 2). The median PCI was 6 (range 0–23) with 5 patients having PCI > 20. A complete cytoreduction (CCR0) was achieved in all but one case (CCR1). A median of 2 organs were resected (range 0–6), and a median of 1 (range 0–3) anastomosis was performed in each case. Median operative time was 330 min (range 105–690 min), and blood loss was 375 cc (100–5000 ml). Most patients were admitted post-operatively to the ICU (N = 76, 83.5%) and required parenteral nutrition (N = 68, 74.7%). Median hospitalization was 16 days (range 6–67). Sixty-day post-operative morbidities and mortalities are listed in Table 2. There were a total of 10 patients (11.0%) with severe grade III/IV complications, and 1 mortality (1.1%) within the post-operative period (Table 2).

Peritoneal carcinomatosis arising from colorectal origin was once considered a terminal stage IV condition but now is represented by a spectrum of disease severity amenable to curative resection in the context of systemic and peritoneal chemotherapy. While CRS and HIPEC have come to be regarded as the standard of care, significant progress still needs to be made in standardizing the approach and chemotherapeutic agents used. One of the reasons for this lack of consensus is the lack of high-quality studies with homogeneous disease, techniques, and chemotherapeutic agents. Consequently, our study fills an important need, providing a relatively homogeneous population of colorectal carcinomatosis patients, all managed with systemic chemotherapy, CRS, and HIPEC with oxaliplatin. Additionally, patients were managed by the same surgical group at a tertiary referral center, using the same inclusion criteria developed in the Canadian HIPEC guidelines [19]. Our study follows up and expands on the patient population originally reported on by our group in 2013 [14]. This earlier study reported 40 patients who underwent CRS and oxaliplatin HIPEC with a 5-year survival of 33%. Our present study has added to this overall population for a total of 91 patients who underwent CRS and oxaliplatin HIPEC with a 5-year survival of 55%, grade III–IV complication rate of 11.0%. Finally, our study confirms that oxaliplatin-based HIPEC retains PCI as a relevant prognostic factor that can be used as a continuous variable in evaluating patients for surgery.

Our OS of 55% exceeds the rates quoted in the literature of 32–47% [23] and several factors could account for this. First, our center has a strict patient selection policy based on the Canadian HIPEC guidelines created jointly by our institution and others in Canada. This provides us with a more selected patient population. Secondly, the completeness of cytoreduction (CCR) score is known to be a powerful prognostic indicator in carcinomatosis patients [13, 24]. In our patients, however, we were able to achieve a CCR score of 0 in 99% of patients, which would naturally improve the OS of our entire cohort compared with studies where the CCR scores were more heterogeneous [11]. Third, PCI is known to be a significant prognostic factor, and our guidelines recommend against performing HIPEC in PCI > 20. Therefore, our patient population is biased toward lower PCI (with a median of 6) and this could contribute to the improved OS. Lastly, most HIPECs performed in the USA use mitomycin C as the chemotherapeutic agent [11, 24]; therefore, one hypothesis is that oxaliplatin is simply a more effective HIPEC agent than mitomycin C. However, this hypothesis would need to be tested in a head-to-head randomized trial comparing the two agents to be validated.

Although mitomycyin C is quite widely used as the agent of choice in the USA, our center has been using oxaliplatin as the primary agent for our protocol for several years due to a number of reasons [3]. First, there is a significant amount of moderate-quality evidence ranging from pharmacokinetics, animal studies, single institution retrospective, and few randomized trials that multiple agents including mitomycin C, cisplatin, oxaliplatin, and even heated water are effective during HIPEC. Secondly, oxaliplatin has a long proven track record of being active against colorectal tumors as an IV drug and the backbone of FOLFOX-based regimens. Lastly, pharmacokinetics studies in animals and humans indicate that oxaliplatin has a favorable tissue penetration profile (392 ng oxaliplatin/mg peritoneal tissue) [25‐27].

The peritoneal carcinomatosis index (PCI) has been well-documented as an indicator of prognosis in patients receiving CRS + HIPEC in the setting of several agents including oxaliplatin [7, 8, 11, 13, 14]. However, there has been controversy on the most appropriate PCI cutoff value to use as an exclusion criteria when evaluating patients for CRS + HIPEC. Several studies, including the current Canadian HIPEC guidelines, recommend excluding colorectal carcinomatosis patients with a PCI > 20 [11, 19]; however, PCI values of 14 [13] and 16 [7] have been proposed by others. By expanding the original patient population reported by our group, we were able to stratify the population by PCI as a continuous variable. Our multivariate analysis was able to show a statistically significant reduction in both OS and DFS with each incremental increase in the PCI. The practical implications of this are that even when a particular PCI value is chosen as a clinical exclusion criterion, this should be considered only as a relative contraindication for young, healthy patients with favorable tumor biology who may still benefit from aggressive surgical intervention. Therefore, using the example of the five patients with PCI > 20 in our series, they achieved a median survival of 19 months which was inferior to the 62 months achieved in patients with PCI < 20 (Fig. 3a). However, these patients exceeded the historical expectation of 6–12-month overall survival for all comers treated with palliative chemotherapy and therefore may have been provided a benefit by undergoing CRS and HIPEC. This is of course acknowledging that our CRS HIPEC patient population has already been selected to be healthy and young enough to undergo the procedure and therefore might be more likely to have a longer (12 months) survival with palliative chemotherapy than the average stage IV patient with carcinomatosis.

Our study has limitations relating to the retrospective design, potential for clinician bias without blinding, referral bias, and the lack of a control arm for comparison. Nevertheless, we have attempted to control for these factors in a number of ways. First, the data was collected prospectively and all patients that presented were included in the study. We cannot mitigate the referral bias; however, we would expect this to worsen our survival outcomes due to receiving more complex patients. Lastly, we referred to historical controls as our reference population both for patients treated with palliative chemotherapy and patients receiving comparable CRS/HIPEC under different regimens which we felt were acceptable given the large body of literature describing both of these populations.

Taken together, we feel that our study advances the understanding of the precise effectiveness and survival advantages of an oxaliplatin-based CRS and HIPEC in a homogeneous population of colorectal carcinomatosis patients. Our study also conclusively re-emphasizes the importance of PCI in prognosis and patient selection even in a cohort of patients achieving excellent CCR0 disease control. Finally, given the favorable survival outcomes and low morbidity profile we have reported, our protocol can serve as a template for oxaliplatin-based HIPEC programs.