An exploratory study of body composition as a predictor of dose-limiting toxicity in metastatic pancreatic cancer treated with gemcitabine plus nab-paclitaxel.

BACKGROUND: Body composition is increasingly being studied as a method of predicting chemotherapy toxicity. Our study aimed to evaluate associations of body composition with treatment toxicity in a group of pancreatic cancer patients treated with gemcitabine plus nab-paclitaxel. METHODS: A retrospective review was performed for all patients who received first-line gemcitabine plus nab-paclitaxel for metastatic pancreatic cancer at a northern Alberta cancer institute (Canada) from 2014 to 2017. Total lean body mass (LBM) was derived from measurements of muscle surface area at L3 on baseline computed tomography (CT) scans. Optimal stratification, or minimal p-value analysis, was used to assess for a threshold of nab-paclitaxel dose per LBM (mg/kg) associated with a higher risk of dose-limiting toxicity (DLT). RESULTS: 152 patients were included in the study, of whom 62 (40.8%) experienced DLT. nab-Paclitaxel dose/LBM ranged from 0.98 to 8.76 mg/kg. A threshold for nab-paclitaxel dose/LBM that optimally predicted risk of DLT was identified at 5.83 mg/kg. Above this cut-off, 18/31 (58.1%) patients experienced DLT, compared to 44/121 (36.4%) patients below (p = 0.028). Patients above this cut-off had a higher incidence of peripheral neuropathy compared to those below, though this was not statistically significant based on an adjusted p-value threshold (48.4 vs. 29.8% respectively, p = 0.050). Body mass index, body surface area, and absolute initial doses of nab-paclitaxel or gemcitabine did not significantly impact likelihood of DLT. CONCLUSIONS: nab-Paclitaxel dose normalized to LBM, based on CT-derived measures of skeletal muscle, has potential to predict risk of chemotherapy toxicity. Chemotherapy dosing based on body composition, rather than conventional anthropometric measures, may be effective in reducing treatment toxicity.