Following the publication of the Intergroup-0116 (INT-0116) and the MAGIC studies [12, 13] which demonstrated a survival benefit from the combination of either postoperative chemoradiation or perioperative chemotherapy with surgery for resectable oesophagogastric cancer, a number of studies have suggested that the survival advantage conferred by this multimodality approach could be cautiously extended to patients with Stage IV disease. However, as this approach is potentially associated with high incidence of side effects it would be important to identify factors that predict survival to assist with patient selection and justify indication and feasibility.
Our retrospective analysis has evolved PS, liver metastasis and DNA content (DNA Index) as independent predictors of survival in our patient cohort which included patients with Stage IV gastric adenocarcinoma treated with palliative gastrectomy and systemic chemotherapy. HP infection was not identified as significant prognostic indicator by either univariate or multivariate analysis.
Although numerous studies report the detrimental effect of poor PS on survival in Stage IV gastric cancer, these mostly involve patients treated with palliative chemotherapy alone [14–16] like the one by Kim et al. which involved 304 consecutive patients with newly diagnosed metastatic or recurrent gastric cancer treated with one or more cycles of cisplatin-based chemotherapy . We could only identify one study which assessed the prognostic significance of PS in a multimodality setting which involved intraperitoneal chemotherapy to treat intra-abdominal gross residual lesions after palliative gastrectomy with maximal cytoreduction. This particular study by Jeung et al. which involved 53 patients identified PS as the only significant defining factor for progression-free survival (P = 0.009) by multivariate analysis .
Liver metastasis is also a well-established prognostic indicator in patients with advanced gastric cancer treated with either palliative chemotherapy or surgery but not the combination. For example a pool analysis of 1080 chemotherapy naïve patients with locally advanced or metastatic oesophagogastric cancer by Chau et al. identified liver metastasis as an adverse prognostic factor by multivariate analysis . Similarly, liver invasion has evolved as an independent survival indicator in a large prospective study involving 539 patients with advanced gastric cancer that had undergone surgical resection .
The prognostic value of DNA content in gastric cancer as measured by the DNA Index (DNA ploidy) is controversial as divergent DNA content analysis results have been reported by various studies. These are thought to reflect objective differences in the analytical techniques employed (image cytometry vs. flow cytometry) and intratumoural DNA ploidy heterogeneity. Image cytometry is considered superior to flow cytometry as it allows direct visualization and selection of tumour cells for inclusion in the DNA measurement. This qualitative feature appears to outbalance its lower throughput when compared with flow cytometry as was illustrated in a recent study by Belien et al. on the prognostic value of both image and flow cytometric analysis of DNA content in gastric cancer . Their patient cohort consisted of 221 cases of gastric cancer analyzed for DNA content using the guidelines of the European consensus report on standardization of diagnostic DNA image cytometry and flow cytometry [20, 21]. Although this study has demonstrated equal sensitivity for both methods in detecting DNA non-diploid gastric cancers, image cytometry DNA content analysis outperformed flow cytometry in predicting survival by multivariate analysis.
Other studies demonstrating the adverse prognostic significance of DNA analysis in advanced gastric cancer includes a study by Kimura et al. who analyzed the DNA content of 270 patients with advanced gastric cancer by flow cytometry to conclude that high DNA ploidy index was the third strongest prognostic factor for survival behind peritoneal dissemination and liver metastases (P < 0.01) . A different study by Baba et al. including 93 patients with advanced gastric cancer showed that high DNA ploidy manifests with higher incidence of vessel invasion and lymph node metastasis invasion conferring a poor 5- year survival in elderly patients (P < 0.05) . Similarly the flow cytometric analysis of gastric carcinomas performed by Danova et al. to evaluate ploidy patterns and the distribution of cells in the different cell cycle phases has demonstrated that DNA aneuploidy was a strong independent adverse prognostic factor for survival in patients with limited or advanced stage gastric cancer .
The relative consistency of reports on DNA Index analysis measurements in the few studies publishing results on advanced gastric cancer, irrespectively of the method applied, strengthens the validity of our data. This may be explained by the observation that aneuploidy appears to be more frequent in advanced gastric cancer and the majority of these aneuploid tumors are not DNA ploidy heterogeneous. That was depicted in a study by Osterheld et al. who performed DNA cytophotometry on multiple samples collected from 16 advanced gastric carcinomas and found 15 DNA-aneuploid tumours (94%) and one diploid tumour; multiple DNA-stemlines were found in 4 cases (26%). Furthermore, analysis of proliferative activity performed on the same samples revealed higher proliferation rate in DNA aneuploid homogeneous tumours than in aneuploid heterogeneous tumours and heterogeneous tumours did not overexpress p53. The authors suggested that the higher proliferative activity in homogeneous-aneuploid carcinomas and their more frequent overexpression of p53 support the hypothesis that in gastric cancer tumour progression implies the development of a dominant and more aggressive (higher proliferative activity, p53 overexpression) aneuploid cell clone . This is particularly important as it has been shown that it is the DNA index of the subpopulation that is most widely distributed within gastric tumour is significantly associated with lymph node metastases (P < 0.001) and histologic grade (P < 0.001) .
Based on our analysis, the high DNA content (DNA Index) along with the poor PS and the presence of liver metastasis have evolved as independent predictors of survival in patients with Stage IV gastric adenocarcinoma treated with palliative gastrectomy and systemic chemotherapy. Since all these three variables can be assessed in preoperative biopsies, they may serve as a prognostic tool in order to preoperatively select patients eligible for palliative gastrectomy; however, prospective studies are needed to confirm this hypothesis.
There is considerable evidence to support causality between HP and sporadic noncardia gastric cancer . For example, an individual-subject meta-analysis of 12 prospective serological studies including 1228 gastric cancer cases, in whom HP status was assessed by anti HP IgG antibodies with ELISA, reported that the relative risk of non cardia gastric cancer associated with HP infection is 5.9 . It is generally accepted that HP infection leads to gastric cancer by inducing sequential alterations of the gastric mucosa, including chronic inflammation, atrophy and intestinal metaplasia, the latter considered as precancerous lesions. The extension of the preneoplastic lesions in the gastric mucosa although increases the risk for cancer development, creates an unfavorable environment for HP colonization and may account for the under-detection of HP infection in blood collected after the diagnosis of gastric cancer in case controlled studies . In addition, it has been shown that conventional ELISAs used to assess serological presence of HP IgG antibodies in most epidemiological studies are likely to produce false-negative results for gastric cancer patients, as compared to population controls, which may further underestimate the risk [29, 30]. In view of the above considerations, results from studies investigating the association of HP infection and prognosis in gastric cancer should be viewed with caution.
Meimarakis et al. used bacterial culture, histological analysis and serology (HP IgA and IgG ELISA) to assess HP status in 166 patients with gastric cancer and reported HP infection to be an independent prognostic factor for relapse free survival (HR 2.16, 95% CI 1.33-3.49) and OS (HR 2.00, 95%CI 1.22-3.57). Yet their patient cohort was different from ours as all their patients underwent curative resection (i.e. R0) followed by adjuvant treatment in 12 cases only (10 patients received intraoperative radiotherapy and 2 had postoperative chemotherapy with Cisplatin 5-FU and folinic acid); HP positivity was 75.3% . Equivalent results were reported by Marrelli et al in a cohort of 297 patients with similar characteristics . Furthermore, HP negative status (as assessed by histology) was reported to be the most significant independent prognostic factor of poor OS (HR 3.45, 95% CI: 2.43-4.89, p < 0.0001) in a different study which included only patients with locally advanced gastric cancer who underwent adjuvant chemotherapy after curative resection (≥D2 dissection) . An attempt to explain the association between HP negativity and adverse survival implies a likely contribution of HP infection in augmenting anti-tumour immunity, especially in early stage gastric cancer , but this needs further validation.
However, in our study which included patients with advanced gastric cancer no association between HP infection status and survival was observed. HP seroprevalance was observed in only 34.9% of our patients as compared to approximately 65% expected for this age group (55–64 years) in Greece . This may be related to the inferior methodology (ELISA) applied to assess HP positivity or it may simply depict mucosal changes and progressive pH alkalization in advanced gastric cancer that create a less favorable environment for HP colonization, as we mentioned above and has also been suggested by others . Moreover, our findings are in agreement with a recent report by Qiu et al who used real-time PCR for HP detection in 157 gastric cancer patients and found no significant association between HP infection and OS or relapse-free survival in patients who underwent curative surgery .
The limitations of our study evolve around its retrospective nature and the objectivity of the methodologies used to assess key indicators such as HP status and DNA Index. Despite these limitations it has clinical relevance as we have validated a number of factors that could potentially be used to assess the likelihood of clinical benefit of a multimodal therapeutic approach with the combination of palliative gastrectomy and postoperative chemotherapy in gastric cancer patients with stage IV disease.