Objectives HGSOC is the deadliest gynecologic cancer. Molecular analysis to samples available at The Cancer Genome Atlas (TCGA) allowed the identification of 4 molecular subtypes with different biology/prognosis. Recently, our group demonstrated by bioinformatic analysis that patients expressing high levels of genes related to obesity and lipid metabolism (e.g. CD36/TGFß) experienced poorer outcomes in two international cohorts (TCGA and AOCS). Here, our goal was to establish if there was correlation between such patterns and the prevalence and prognosis of different molecular subtypes of HGSOC.

Methods We retrieved matching data of 455 cases of stage II-IV HGSOC from TCGA and categorized into 4 clusters based first on molecular subtypes (differentiated, immunoreactive, mesenchymal and proliferative) and then subdivided based on obesity and lipid metabolism gene pattern expression (CD36/TGFß high vs low). Proportion and survival analyses were carried out. Chi square, Kaplan Meier were used to assess statistical significance.

Results Mesenchymal subtype was significantly more prevalent among women expressing CD36/TGFß high (51,2% vs 6,9%, p<0.0001). Proliferative and differentiated were more prevalent in women expressing CD36/TGFß low (71,3% vs 27,7%, p<0.0001). The 5-year overall survival was significantly different between subgroups with the immunoreactive subtype expressing CD36/TFGß low experiencing the best outcome compared to the rest (median 67,7 vs 44,1 months, log-rank p<0.01). Cox analysis (including clustering, age, cytoreduction and therapeutic response) confirmed the immunoreactive CD36/TGß/low as independent prognostic factor.

Conclusions Our results confirm the importance of immune response and the negative impact that obesity and lipid metabolism alterations have in defining prognosis in HGSOC (supported by Fondecyt 1160800).