PHARMACOGENETICS Improvement of a predictive model in ovarian cancer patients submitted to platinum-based chemotherapy: implications of a GST activity profile Deolinda Pereira 1,2 & Joana Assis 3,4 & Mónica Gomes 2,3,5 & Augusto Nogueira 3,4,5 & Rui Medeiros 2,3,5,6 Received: 27 October 2015 /Accepted: 14 January 2016 # Springer-Verlag Berlin Heidelberg 2016 Abstract Purpose The success of chemotherapy in ovarian cancer (OC) is directly associated with the broad variability in platinum response, with implications in patients survival. This hetero- geneous response might result from inter-individual variations in the platinum-detoxification pathway due to the expression of glutathione-S-transferase (GST) enzymes. We hypothe- sized that GSTM1 and GSTT1 polymorphisms might have an impact as prognostic and predictive determinants for OC. Methods We conducted a hospital-based study in a cohort of OC patients submitted to platinum-based chemotherapy. GSTM1 and GSTT1 genotypes were determined by multiplex PCR. Results GSTM1-null genotype patients presented a signifi- cantly longer 5-year survival and an improved time to progres- sion when compared with GSTM1-wt genotype patients (log- rank test, P = 0.001 and P = 0.013, respectively). Multivariate Cox regression analysis indicates that the inclusion of genetic information regarding GSTM1 polymorphism increased the predictive ability of risk of death after OC platinum-based chemotherapy (c-index from 0.712 to 0.833). Namely, resid- ual disease (HR, 4.90; P = 0.016) and GSTM1-wt genotype emerged as more important predictors of risk of death (HR, 2.29; P = 0.039; P = 0.036 after bootstrap). No similar effect on survival was observed regarding GSTT1 polymorphism, and there were no statistically significant differences between GSTM1 and GSTT1 genotypes and the assessed patients’ clinical-pathological characteristics. Conclusion GSTM1 polymorphism seems to have an impact in OC prognosis as it predicts a better response to platinum- based chemotherapy and hence an improved survival. The characterization of the GSTM1 genetic profile might be a useful molecular tool and a putative genetic marker for OC clinical outcome. Keywords Epithelial ovarian cancer . GST . Polymorphism . Pharmacogenetic . Platinum-based chemotherapy . Predictive value Introduction Ovarian cancer (OC) is the third most common gynecological cancer among women worldwide, with an estimate of 239,000 new cases and 152,000 deaths each year. In Europe, OC is the fifth most incident cancer in women but represents the leading cause of death among gynecological tumors [1]. The high mortality rate is, in part, attributable to the usual absence of early symptoms of the disease and the lack of screening methods for early OC diagnosis. Consequently, the diagnosis is frequent at an advanced stage, with 75 % of all cases being diagnosed at stages III and IV where the disease has spread throughout the abdominal cavity [2–4]. Deolinda Pereira and Joana Assis contributed equally to this work. * Rui Medeiros ruimedei@ipoporto.min-saude.pt 1 Oncology Department, Portuguese Institute of Oncology, Porto, Portugal 2 ICBAS, Abel Salazar Institute for the Biomedical Sciences, Porto, Portugal 3 Molecular Oncology and Viral Pathology Group-Research Center, Portuguese Institute of Oncology, Edifício Laboratórios. 4° piso, Rua Dr. António Bernardino de Almeida, 4200-072 Porto, Portugal 4 FMUP, Faculty of Medicine, Porto University, Porto, Portugal 5 Research Department, Portuguese League Against Cancer (NRNorte), Porto, Portugal 6 CEBIMED, Faculty of Health Sciences, Fernando Pessoa University, Porto, Portugal Eur J Clin Pharmacol DOI 10.1007/s00228-016-2015-3