Incidence, time course and independent risk factors for metachronous peritoneal carcinomatosis of gastric origin – a longitudinal experience from a prospectively collected database of 1108 patients

Data from our cohort, which is to our knowledge the largest ever studied, show that neither the incidence nor the time course and prognosis of metachronous peritoneal carcinomatosis in gastric cancer patients treated with curative intent, have changed over the last three decades. Our data are in line with previous studies reporting an advanced tumour stage in 45.9 percent of the patient cohort and with synPC PC in 14.6 percent at primary diagnosis respectively [4]. Our findings show that tumour stage at diagnosis has not changed and, especially, the rates of synPC has not decreased within the last two decades, which is consistent with data reported by a recent European population based study [4]. This could firstly be explained by the fact that screening for gastric cancer is only recommended for a small proportion of people with well-established risk factors [19,24,25], and secondly, by the overall poor perception of any cancer screening programs [26].

For further analysis we excluded patients with both, a metastatic state and/or positive margins after D2 gastrectomy, as they are associated with a very poor prognosis, [27,28] and revisional surgery and/or radio chemotherapy is recommended under these circumstances [19,20]. Despite Ro D2 gastrectomy 50% of the patients were diagnosed with tumour recurrence, 16% developed metaPC after a median time of 17.7 (15.1-20.3) months from surgery resulting in a tumour related mortality of 100% with a median survival of 3 months. The data on the incidence of metaPC from this study, however, vary considerably [29-33] with that previously reported, as in our cohort peritoneal tumour progression occurred in up to 60% of patients. The differences in the results of our study cohort could be explained by several reasons. Firstly, baseline demographic and tumour characteristics of our patient cohort were different to those of other studies. Secondly, the vast majority of the previous findings were obtained from an Asian population being reported to develop gastric cancer with a more malignant tumour biology [33]. Thirdly, the patient cohorts of those studies may have had a more structured follow up postoperatively, more extensively exploiting the prevalence reserve of metachronous PC. Fourthly, older studies were published more than 13 years ago and therefore did not incorporate the currently available perioperative chemotherapeutic regimes [5,10]. For this precise reason we performed a subgroup analysis using three time periods based on the different perioperative chemotherapeutic regimes available over the last few decades (Table 2).

In concordance with the literature we observed an improved overall survival in patients treated with modern combination chemotherapies during time period III [5,34]. Survival after R0 D2 gastrectomy did, however, not differ in patient among the different time periods. It appears that patients diagnosed with metaPC also did not experience survival benefits as the subgroup analysis did not show significant survival differences among the three different time periods. Interestingly, patients in our cohort were more often diagnosed with metaPC and had an increased risk for metaPC during the most recent time period compared to the patients treated within the earlier time periods. However, multivariate analysis did not show that the time period of treatment itself was an independent risk factor for the incidence and prognosis of metaPC (Table 5, Figure 3).

The higher incidence of metaPC during the most recent time period may, on one hand, reflect the usage of an improved and, therefore, more sensitive imaging technology [35]. We also assumed that patients underwent imaging studies more frequently as potent second and third line chemotherapeutic regimes became available and were applied more frequently to these patients. Although pathological staging characteristics (measured by UICC VII) of the different cohorts were not different at the time point of surgery, a locally advanced tumour was found more frequently in patients in time period III. Multivariate analysis revealed that a serosa positive tumour category was an independent risk factor for the development of metaPC (Table 5, Figure 3), which is in line to previous reports [36,37]. Further, it is probable that the use of perioperative chemotherapeutic regimen during the last decade may have downsized and downstaged the tumour enabling curative surgery with benefits diminishing in the long term [10]. Therefore, it could be argued that the perioperative therapy may have enabled a curative treatment in a substantial number of patients who most likely would have been treated using palliative intentions in the former time periods. It has firstly been hypothesized that preoperative chemotherapy may decrease the biological activity of the tumour and therefore reduce the likelihood of malignant biological active cell spread into the peritoneal cavity during surgery [34]. In addition, tumour dissemination during surgery through the opening of lymphatic channels lymph node dissection and spread of viable cancer cells into the peritoneal cavity during could be reduced [10]. However, we did not observe a clear benefit in our cohort.

Multivariate analysis did not identify perioperative chemotherapy, age and gender as independent risk factors for the development of metaPC (Table 5, Figure 2b). Other histopathological characteristics such as undifferentiated ‘gradings’ (G3/4), nodal positive category (N+), signet ring cell (SRC), and serosa positve tumour category (T3/4) were identified to be independent risk factors (Table 5, Figure 3), which is consistent to the literature [38,39]. Notably, median survival of patients diagnosed with metaPC was consistently shorter compared to patients diagnosed with synPC with no trend in further subgroup analysis. This may indicate that a systematic follow up of patients at high risk of tumour progression may be beneficial if strategies to treat PC become available. Despite improvements in overall survival, the time course and occurrence rate of metaPC as well as prognosis did not improve over time. Considering these disappointing results two conclusions need to be drawn: Firstly, other strategies aiming to prevent or at least relevantly delay metachronous dissemination may be a reasonable approach. A combined strategy of cytoreductive surgery and intraperitoneal chemotherapy showed favorable results in highly selected patients [16]. This multimodal treatment is currently regarded as the only therapeutic option for selected patients with PC from gastric cancer, reporting improved 5-year survival rates ranging from 13 to 28% [14,15]. Secondly, patients treated with curative intention (D2 gastrectomy plus state of the art chemotherapy) at high risk for metaPC could be tailored to other therapeutic approaches such as the extensive intraoperative peritoneal lavage (EIPL). This straightforward adjuvant surgical technique has been advocated as a useful tool for those gastric cancer patients who are likely to suffer from peritoneal recurrence [17,18]. In a randomized controlled study the effect of EIPL therapy on prevention of peritoneal recurrence on patients with peritoneal free cancer cells without overt peritoneal metastasis was verified [17,18]. Another strategy analogous to that proposed by Elias et al. could potentially apply for patients diagnosed with gastric cancer. Elias et a. scheduled patients diagnosed with colorectal cancer at high risk for metaPC to a second look laparotomy routinely even though imaging studies did not reveal any signs of recurrence at this time point [40].