Résumé : In the phase III MPACT trial, nab-paclitaxel plus gemcitabine (nab-P + Gem) demonstrated superior efficacyversus Gem alone for patients with metastatic pancreatic cancer. We sought to examine the feasibility of positron emissiontomography (PET) and to compare metabolic response rates and associated correlations with efficacy in the MPACT trial.Patients and methods: Patients with previously untreated metastatic adenocarcinoma of the pancreas were randomized1:1 to receive nab-P + Gem or Gem alone. Treatment continued until disease progression by RECIST or unacceptable toxicity.Results: PET scans were carried out on the first 257 patients enrolled at PET-equipped centers (PET cohort). Most patients(252 of 257) had ≥ 2 PET-avid lesions, and median maximum standardized uptake values at baseline were 4.6 and 4.5 in thenab-P + Gem and Gem-alone arms, respectively. In a pooled treatment arm analysis, a metabolic response by PET (best responseat any time during study) was associated with longer overall survival (OS) (median 11.3 versus 6.9 months; HR, 0.56;P < 0.001). Efficacy results within each treatment arm appeared better for patients with a metabolic response. The metabolicresponse rate (best response and week 8 response) was higher for nab-P + Gem (best response: 72% versus 53%,P = 0.002; week 8: 67% versus 51%; P = 0.014). Efficacy in the PET cohort was greater for nab-P + Gem versus Gem alone,including for OS (median 10.5 versus 8.4 months; hazard ratio [HR], 0.71; P = 0.009) and ORR by RECIST (31% versus 11%;P < 0.001).Conclusion: Pancreatic lesions were PET avid at baseline, and the rate of metabolic response was significantly higher fornab-P + Gem versus Gem alone at week 8 and for best response during study. Having a metabolic response was associatedwith longer survival, and more patients experienced a metabolic response than a RECIST-defined response.ClinicalTrials.gov: NCT00844649.