Personal Visceral Fat Index (PVF Index), a New Predictive Factor for Chemotherapy Efficacy in Patients with Advanced Gastric Cancer.

Xiaodong Zhu

DOI: https://doi.org/10.1200/jco.2018.36.15_suppl.e16049

IF: 45.3

2018-01-01

Journal of Clinical Oncology

Abstract:e16049 Background: It has been reported that skeletal muscle depletion is a poor prognostic factor for a variety of cancers. Whether the changes of visceral fat or skeletal muscle during chemotherapy are related with the efficacy of treatment is still unclear. This study aimed to investigate the relationship between clinical efficacy and the changes of visceral fat or skeletal muscle during chemotherapy. Methods: This is a retrospective analysis based on a phase II clinical trial (NCT00767377) conducted during June 2007 to July 2012 in Fudan University Shanghai Cancer Center, in which all the AGC patients received EOF (epirubicin, oxaliplatin and fluorouracil) regimen as first line therapy. The computed tomography(CT) scans at baseline and after two cycles of treatment were used, and total abdominal fat area(TFA), visceral fat area(VFA), subcutaneous fat area(SFA) and skeletal muscle area(SMA) were measured with the software of GE Healthcare-Centricity RIS CE V3.0. Results: Because our PACS CT system only retains the data after 2010, only 53 patients with available data were included in the analysis. In terms of viscera fat area, we identified a variate as personal visceral fat index (PVF index), which stands for the difference between VFA/TFA after 2 cycle chemotherapy and baseline VFA/TFA. Receiver operating characteristic (ROC) curve showed that PVF index < 0.02 offered the best overall sensitivity (72.41%) and specificity (70.83%) for prediction of longer PFS time (AUC 0.7155). So patients were classified into low( < 0.02) and high(≥0.02) PVF index group. Compared with high PVF index group, patients of low PVF index group had significantly higher response rate (59.3% vs 30.8%, p= 0.037), longer PFS time (7.1months vs 5.1months, p= 0.036) and a tendency for a longer OS time (19.0months vs 9.6months, p= 0.190). Multivariate Cox proportional hazards model showed that PVF index < 0.02 was an independent predictive factor for PFS (HR, 0.377; 95%CI 0.146-0.978; P = 0.045). In terms of skeletal muscle area, no relationship between SMA and PFS, OS or response rate was found. Conclusions: PVF index can be used to predict for both favorable PFS and response in AGC patients treated with EOF regimen.