Pathological factors affecting gastric adenocarcinoma survival in a Caribbean population from 2000-2010

To investigate pathological factors related to long term patient survival post surgical management of gastric adenocarcinoma in a Caribbean population.

This is a retrospective, observational study of all patients treated surgically for gastric adenocarcinoma from January 1(st) 2000 to December 31(st) 2010 at The University Hospital of the West Indies, an urban Jamaican hospital. Pathological reports of all gastrectomy specimens post gastric cancer resection during the specified interval were accessed. Patients with a final diagnosis other than adenocarcinoma, as well as patients having undergone surgery at an external institution were excluded. The clinical records of the selected cohort were reviewed. The following variables were analysed; patient gender, patient age, the number of gastrectomies previous performed by the lead surgeon, the gross anatomical location and appearance of the tumour, the histological appearance of the tumour, infiltration of the tumour into stomach wall and surrounding structures, presence of Helicobacter pylori and the presence of gastritis. Patient status as dead vs alive was documented for the end of the interval. The effect of the aforementioned factors on patient survival were analysed using Logrank tests, Cox regression models, Ranksum tests, Kruskal-Wallis tests and Kaplan-Meier curves.

A total of 79 patients, 36 males and 43 females, were included. Their median age was 67 years (range 36-86 years). Median survival time from surgery was 70 mo with 40.5% of patients dying before the termination date of the study. Tumours ranged from 0.8 cm in size to encompassing the entire stomach specimen, with a median tumour size of 6 cm. The median number of nodes removed at surgery was 8 with a maximum of 28. The median number of positive lymph nodes found was 2, with a range of 0 to 22. Patients' median survival time was approximately 70 mo, with 40.5% of the patients in this cohort dying before the terminal date. An increase in the incidence of cardiac tumours was noted compared to the previous 10 year interval (7.9% to 9.1%). Patients who had serosal involvement of the tumour did have a significantly shorter survival than those who did not (P = 0.017). A significant increase in the hazard ratio (HR), 2.424, for patients with circumferential tumours was found (P = 0.044). Via Kaplan-Meier estimates, the presence of venous infiltration as well as involvement of the circumferential resection margin were found to be poor prognostic markers, decreasing survival at 50 mo by 46.2% and 36.3% respectively. The increased HR for venous infiltration, 2.424, trended toward significant (P = 0.055) Age, size of tumour, number of positive nodes found and total number of lymph nodes removed were not useful predictors of survival. It is noted that the results were mostly negative, that is many tumour characteristics did not indicate any evidence of affecting patient survival. The current sample, with 30 observed events (deaths), would have about 30% power to detect a HR of 2.5.

This study mirrors pathological factors used for gastric cancer prognostication in other populations. As evaluation continues, a larger cohort will strengthen the significance of observed trends.