Treatment of synchronous peritoneal carcinomatosis and liver metastases from colorectal cancer
Introduction
In patients with potentially resectable liver metastasis (LM), the discovery of peritoneal carcinomatosis (PC) is usually a contraindication to hepatectomy.1–3 Similarly, if curative treatment of PC is a goal, the discovery of LM is also considered a contraindication.45 In contrast to this conservative approach, recently a few series reported cure after complete resection of LM associated with extrahepatic disease, notably PC,67 and selected cases of PC have been cured after complete cytoreductive surgery followed by intraperitoneal chemohyperthermia.8–11 As a result of these observations, the rationale and role of surgical cytoreduction in multimodality treatment for cancer have become greatly changed.12 An attractive and more frequently used treatment pattern consists of induction chemotherapy followed by optimum cytoreductive resection of residual tumor and then additional chemotherapy. Concomitantly, progress in anesthetic and surgical techniques has sufficiently reduced the risk of the most radical surgical procedures to rival current risk of systemic chemotherapy. The development of radiofrequency ablation of liver metastases provided an additional surgical tool for safe hepatic cytoreduction.#
Is it beneficial for selected patients presenting with synchronous LM and PC to be treated with curative intent? The aim of this prospective study was to evaluate tolerance and the oncologic results of this new therapeutic approach.#
Material and methods
Population
From January 1993 to November 2003, 24 patients were treated with a curative intent for synchronous PC and LM originating from colorectal adenocarcinomas. They represented 1.4% of the 679 hepatectomies performed for malignant tumors in our Institute and 8.9% of the 267 PC patients treated with a curative intent during the same period. Their mean age was 44.8years (range: 26–64).#
LM had been diagnosed before laparotomy in all cases. Relatively extensive PC had been diagnosed before laparotomy in 14 patients; it was discovered only at laparotomy in the other 10 cases who had a minimal volume of carcinomatosis. The characteristics of the liver disease and of the peritoneal disease are reported for these two different groups of patients in Table 1.#
Selection criteria
The selection criteria for hepatectomy plus peritonectomy in the 14 patients in whom LM and PC were detected preoperatively, were: (1) a young age, (2) a good performance status (WHO 0 or 1), (3) moderate operative risk of liver resections avoiding patients with invasion of the hepatic hilum, the inferior vena cava or the hepatic veins, (4) a moderate volume of PC and absent symptoms, and (5) an objective response or stabilization of LM after 3months of systemic chemotherapy. These patients signed an informed consent to participate in a clinical trial prior to treatment. For 10 patients in whom PC was discovered at laparotomy, the decision was taken during the operation after considering the resectability of PC and the difficulty of the hepatectomy, the performance status and the response to the adjuvant chemotherapy. These patients could not give beforehand a signed informed consent to participate to the trial on intraperitoneal chemotherapy. No patients were operated on for isolated PC with LM detected only at laparotomy.#
Phases of treatment
Treatment included two phases. A complete resection (R0) of all detectable liver lesions was performed by partial hepatectomy with intermittent vascular exclusion.13 Radiofrequency ablation was utilized for small-sized central LM. The resection of PC by peritonectomy with removal of digestive tissue when necessary was done. PC was scored with Sugarbaker's peritoneal cancer index, which ranges from 1 to 39.5 A score of 0 to 3, according to the size of the peritoneal deposits, was determined for 13 areas within the abdominal and pelvic cavity.#
Microscopic residual PC was treated perioperatively with intraperitoneal chemotherapy, before cancer cells were trapped in fibrin. Intraperitoneal chemotherapy was performed early and continuously over the five postoperative days,14 or with intraoperative chemohyperthermia.15 The modalities for intraperitoneal chemotherapy (Table 2) were dependent upon the ongoing prospective trials at the time of the treatment and on the modality of discovery of the PC. If the carcinomatosis was not diagnosed prior to surgery, consent for intraperitoneal chemohyperthermia could not be obtained so that early postoperative intraperitoneal chemotherapy was utilized.#
Statistics
Patients were recorded prospectively in a specific database. The exact status of each patient was known on the date of the analysis (November 2004). Univariate study of the following parameters was performed: age (>40), origin (colon or rectum), presentation PC before or after laparotomy, number of LMs, peritoneal cancer index, mean duration of the surgery, and blood loss during surgery. No multivariate study was done, due to the small number of patients. The chi-square test, Fisher's exact test or Wilcoxon test, where appropriate, were used for univariate analysis. Differences were considered significant at p=0.05. Survival curves were plotted using the Kaplan–Meier method. The patient who died during the postoperative course was included in the survival curves.#
Follow-up
Clinical follow-up was performed every 3months. The median follow-up was 73.4months (range: 12–143) after this combined treatment. No patient was lost to follow-up.#
Results
Operative parameters
A major hepatectomy (resecting at least three segments according to Couinaud's classification) was performed in 12 cases. Radiofrequency ablation was also used in five patients. One to 15 LMs were resected during the above procedures (mean: 4.4, median: 3.6). Only 8 patients had a single LM. The mean number of LMs was not significantly different between the two groups of patients (Table 1). The mean peritoneal cancer index of resected PC was 8.6 (range: 2–25). The mean peritoneal cancer index was smaller (p<0.001) in the group in which PC was discovered at laparotomy (3.7 versus 11.5 in the group with PC detected preoperatively) (Table 1). Circular digestive anastomoses were necessary in 18 patients (mean per patient: 1.8, range: 0–5). The mean operative time was 357±112min (range: 165–720), and mean blood loss was 925ml (median: 719, range: 200–4500), but only 7 patients received blood transfusion. Microscopic residual PC was treated with early postoperative intraperitoneal chemotherapy in 14 patients and with intraoperative chemohyperthermia in 10 patients.#
Mortality and morbidity
One postoperative death occurred at day 14 due to peritonitis that had been undiagnosed; in this patient there was also aplasia grade 3. Morbidity was 58% (14 patients), with 5 grade 3–4 aplasia, 5 transient liver failures, 4 lung infections, 3 bilioma, and 2 digestive fistula. Severe aplasia was often observed after extended hepatectomy. The mean hospital stay was 21.4±4.2days (range: 15–49).#
Survival rates and recurrences
Median follow-up was 6.1years (range: 12–143months). Overall and disease-free survival curves are shown in Fig. 1. At 2years, they attained 61.3% (CI: 42–80) and 37.7% (CI: 21–58), at 3years 41.5% (CI: 23–63) and 23.6% (CI:11–45), and at 5years 26.5% (CI: 11–52) and 15.7% (CI: 5–39), respectively. In four patients, disease did not recur; 19 patients developed recurrence (median delay: 22.5months, range: 4–46), but 3 of them are currently free of disease because they underwent repeated hepatectomy resecting one (n=2) or two (n=1) recurrent LMs. To date seven patients are free of disease and their mean follow-up has been 27.8months since their last surgery (range: 5–99).#
First recurrences occurred in the following sites: (a) isolated sites: liver 6, lungs 2, brain 2, bone 1, distal lymph nodes 1; (b) associated sites: liver–lung 2, liver–PC 3, liver–bone 1, liver–lymph nodes 1. Finally, only three patients had recurrence in the peritoneum (all in association with LM) and 13 had recurrence in the liver.#
Prognostic factors
Age (>40), origin (colon/rectum), duration of surgery and mean blood loss did not show any prognostic impact. Results concerning the type of presentation (PC known before laparotomy or not), the peritoneal score, and the number of LMs, are shown in Table 3. Only the number of LMs (>2) had a significant (p<0.01) negative impact on survival.#
Discussion
This prospective study shows that complete resection of the concomitant cancer dissemination to the liver and to the peritoneum is feasible in selected patients, and allows, in association with systemic chemotherapy, a cure in a few. Traditionally, treatment with curative intent in patients with colorectal cancer presenting simultaneously with LM and PC is considered beyond the standard of practice. However, the same was thought about hepatectomy for LM before 1960, and about cytoreductive surgery with intraperitoneal chemotherapy for PC before 2000.#
The surgical tradition that would limit cancer resection to patients with regional disease in the absence of systemic involvement is being questioned. Several discoveries that invalidate this belief have appeared. It is now known that micro-metastases and circulating cancer cells in blood, bone marrow and lymph nodes are almost always present in advanced-stage cancers.16 Most cells derived from the primary tumor are not filtered out by the liver and enter the systemic circulation.17 It is also well known that cancer cells are often found in the blood accumulating in the surgical field.1819 Thus, a complete cancer resection is in reality only a cytoreductive surgery in that treatment does not completely eradicate all tumor cells, even if it entirely resects all visible and detectable disease. However, cure can be achieved with the assistance of the patients' natural immunologic defenses and/or with chemotherapy. Surgical tumor burden reduction may provide an immunologic benefit, as tumor cells produce immunosuppressive cytokines, precipitate immune complexes, and produce peptidoglycans that interfere with normal defenses.2021 Furthermore, according to the log-kill hypothesis, each dose of chemotherapeutic agent kills a constant fraction of cells, rather than a specific number of them.22 Therefore, by reducing the initial tumor volume by surgical cytoreduction, one increases the likelihood of chemotherapy to reduce the number of viable tumor cells towards the desired endpoint of zero.#
Using this hypothesis, in the last few years the rationale and role of surgical cytoreduction in multimodality treatment for cancer has greatly expanded.12 An attractive and frequently used treatment scheme consists of induction chemotherapy, followed by optimum cytoreductive resection of residual tumor and then additional systemic chemotherapy. With 26.5% global survival rate at 5years for patients presenting with LM plus CP, our study appears as a good example of the validity of this new therapeutic concept.#
This hypothesis has been facilitated by improved technical aspects of liver resection and peritonectomy. Progress in anesthetic and surgical techniques has sufficiently reduced the risk of the most radical surgical procedures to compare favorably with the current risk of systemic chemotherapy. The development of radiofrequency ablation provided an additional surgical tool for safe cytoreduction. In our study, the technical feasibility of hepatectomy, peritonectomy, intestinal resection, and intraoperative chemohyperthermia was demonstrated.#
To our knowledge, this is the first series of patients that shows the results of a combined treatment for patients presenting synchronous LM and CP. A recent confirmation of our data has appeared.23 Also, in an earlier report focusing on the resection of LM associated with the resection of miscellaneous extrahepatic tumor disease, the 5-year survival rate was 29% in 77 patients who underwent a R0 resection.6#
Our study concerns only selected cases, with good performance status. Patients showed chemotherapy response or stable disease under induction systemic chemotherapy. All had a moderate PC with 16 of these 24 patients (66.6%) having a peritoneal cancer index of less than 10, and only one had a score greater than 20. This moderate extension of the PC explains why in this small series the extent of PC did not appear as a significant prognostic factor. Moreover, we could only detect three patients who recurred on the peritoneum all in association with a liver recurrence.#
In contrast, the number of LMs had a significant prognostic impact in this study. Two LMs seemed to be a limiting factor and it is important to emphasize that 6 of the 7 disease-free patients are in this group. Also, in our series the liver appears by far as the most frequent site of recurrence, leading us to consider the addition of postoperative intra-arterial hepatic chemotherapy with oxaliplatin in the future.24#
Postoperative complications in this group of patients amounted to greater than 50% and aplasia was more common than expected. We observed more severe aplasia when an extended hepatectomy was performed with at least two thirds of the liver resected. The induced transient and relative liver insufficiency seemed to modify the pharmacokinetics of the peritoneal chemotherapy and to increase its toxicity. From these results we suggest the doses of intraperitoneal chemotherapy be decreased when a major hepatectomy is performed.#
Finally, after complete liver or peritoneal cytoreduction or complete resection of all detectable tumors whatever their sites, it is impossible, in the absence of a randomized study, to determine if perioperative intraperitoneal chemotherapy is essential. It is possible that systemic chemotherapy delivered after the same cytoreductive surgery may result in similar survival statistics. This prospective study remains to be promoted.#
In conclusion, using new multimodality treatments in oncology, it is feasible and beneficial to treat with a curative intent selected patients presenting synchronous LM and CP. Data suggest selection of patients presenting no more than two (possibly three) LMs, a moderate volume of PC, and known as responders to chemotherapy.#
Figures and Tables
Table 1
| Group | LM+PC diagnosed before laparotomy | PC not diagnosed before laparotomy | p value |
| No. of patients | 14 | 10 | |
| No. of LM | |||
| Mean | 3.4 | 5.8 | NS |
| Range | 1–11 | 1–15 | |
| Peritoneal cancer index | |||
| Mean | 12.4 | 3.7 | <0.001 |
| Range | 6–25 | 2–6 | |
| Hepatectomy | |||
| Major | 5 | 6 | NS |
| Minor | 9 | 4 | |
| LM, liver metastases; PC, peritoneal carcinomatosis; NS, not significant. |
Table 2
| Protocol | Patients | Drug schedules |
| Early postoperative | n=14 | Day 1: mitomycin 10mg/m2 |
| Days 2–5: 5-fluorouracil 15mg/m2 | ||
| Volume of dialysate: 1L/m2. Drains were clamped 23h/24h | ||
| Intraoperative | n=2 | Mitomycin (20mg/m2), Cisplatin (200mg/m2) |
| Chemohyperthermia | n=8 | Oxaliplatin: 460mg/m2 |
| Volume of dialysate: 2L/m2, intraperitoneal hyperthermia from 42°C to 44°C. Duration: 30min after reaching 42°C |
Table 3
| No. of patients | Patients free of recurrence | p value | |
| Presentation | |||
| LM+PC | 14 | 5 | NS |
| LM alone | 10 | 2 | |
| Peritoneal cancer index | |||
| <10 | 15 | 5 | NS |
| ≥10 | 9 | 2 | |
| No. of LM | |||
| ≤2 | 12 | 6 | <0.01 |
| ≥3 | 12 | 1 |
References
21. D.Y.KavanaughD.P.CarboneImmunologic dysfunction in cancerHematol Oncol Clin North Am101996927951
22. H.B.MussChemotherapy of gynecologic cancer2nd ed.1990Wiley-LissNew York