Objectives: This study aims to evaluate the impact of therapy assessment PET/computed tomography (CT) scan on the management of locally advanced pancreatic adenocarcinoma (LAPC), and the value of qualitative versus quantitative PET/CT interpretation for patient outcome prediction.
Materials and methods: Forty-two LAPC patients were retrospectively included. PET/CT was performed at a median of 4.6 weeks after completion of chemo ± radiotherapy to assess the primary treatment response. PET was interpreted visually using a qualitative five-point scale (Hopkins criteria for therapy assessment). Quantitative PET parameters including maximum and peak standardized uptake value (SUV max and SUV peak), total lesion glycolysis, and metabolic tumor volume (MTV) were also measured using the gradient segmentation method. Kaplan-Meier and Cox regression analyses were performed.
Results: Thirty-five patients were followed up until death. Therapy assessment PET/CT led to a change in the overall management of 22 (52.4%) patients, prompting surgical resection (eight patients), adding radiation therapy (eight patients), or starting palliative chemotherapy (six patients). The median survival in patients with a negative or a positive PET scan, according to the Hopkins criteria, was 14.6 and 8.7 months, respectively (P=0.06). The median quantitative thresholds of SUV peak 2.64 [hazard ratio (HR)=2.67, P=0.03], total lesion glycolysis 44.0 g (HR=2.64, P=0.005), and MTV 24.7 ml (HR=2.57, P=0.008) were significant predictors of overall survival. Using combined quantitative scoring, patients with high SUV peak and high MTV (>median cut point) had a 5.45-fold (95% confidence interval: 1.76-16.87) increased risk for death compared with those with both low SUV peak and MTV (the reference group).
Conclusion: PET-based volumetric parameters can predict survival outcomes of patients with LAPC. A combined quantitative PET/CT scoring system provides significantly improved prognostication.