Recurrence-free and overall survival in gastric cancer was highest among patients with the Epstein-Barr virus molecular subtype, investigators reported online July 27 in Clinical Cancer Research.
In contrast, the genomically stable subtype had the worst prognosis and was least likely to benefit from adjuvant chemotherapy, Bo Hwa Sohn, PhD, of the University of Texas MD Anderson Cancer Center, Houston, reported with her associates. Their prediction model “successfully stratified patients by survival and adjuvant chemotherapy outcomes,” they concluded.
wildpixel/Thinkstock
The MSI subtype was tied to a moderate prognosis. Prognosis for the CIN subtype also was moderate, but was worse in the cohort from South Korea than in the MD Anderson cohort. These results might reflect more heterogeneity within the CIN subtype, the researchers said.
Importantly, adjuvant chemotherapy most benefited CIN patients – 59% were alive without recurrence at 3 years, compared with only 34% of those who received no adjuvant therapy (hazard ratio, 0.39; 95% confidence interval, 0.16-0.94; P = .03). In contrast, adjuvant chemotherapy produced only modest benefits in MSI (HR, 0.55; P = .18), and did not benefit GS patients (P = .66).
Analyses of 24 gastric cancer cell lines from the Genomics of Drug Sensitivity in Cancer project supported that finding, said the researchers. Lines of GS cells had the highest 5-fluorouracil IC50 values, indicating they are resistant to 5-fluorouracil. The activated transcription regulator NUPR1 has been linked to chemoresistance in other cancers and is often altered in the GS subtype, they noted. They only studied adjuvant chemotherapy in the MD Anderson cohort because most South Korean patients did not receive it. Also, all EBV patients received adjuvant chemotherapy, so the researchers could not examine outcomes for this subtype.
The researchers also created a predictive model by pooling probabilities of recurrence. On a scale of 0 to 100, patients who scored under 20 had a 67% chance of recurrence-free survival at 5 years, compared with 52% for patients scoring between 20 and 30 and 38% for patients scoring above 30. Differences between these probabilities were highly statistically significant, and results for overall survival were similar. Score independently predicted recurrence-free survival after other clinicopathologic variables were controlled for in the pooled cohorts and in patients with more heterogeneous stage II cancers.
This prediction model will need to be whittled down to a few genes that adequately represent each subtype before it is useful in the clinic, the researchers acknowledged. “Nevertheless, the validation of our prediction model in two independent patient cohorts and the fact that the model reflects the biological characteristics associated with each subtype indicate that this prediction model could be used to develop rational therapy recommendations,” they concluded.
Funders included the National Institutes of Health, The University of Texas MD Anderson Cancer Center, the Korea National Research Foundation, the Scientific Research Center Program, and the Korean Research Institute of Bioscience and Biotechnology. No investigator declared potential conflicts of interest.