Dynamic susceptibility contrast MRI in advanced pancreatic cancer: semi-automated analysis to predict response to chemotherapy

Abstract

The purpose of this study was to assess whether dynamic susceptibility contrast magnetic resonance imaging (DSC-MRI) can predict response to chemotherapy in advanced pancreatic cancer. DSC-MRI was performed using gradient-echo echo-planar imaging after bolus injection of contrast material. Fifty-four patients with advanced pancreatic cancer who were scheduled for chemotherapy were enrolled. ΔR2* was calculated using semi-automated computer analysis capable of tracking moving lesions during DSC-MRI. Pre-treatment maximum ΔR2* and clinical factors including gender, age, tumor stage (UICC III/IV), initial tumor size, and chemotherapy regimen were compared between patients with progressive disease and patients with stable disease as was determined at 3-month follow-up, and between patients with progressive disease and patients with stable disease as was determined at 6-month follow-up. Receiver operating characteristic (ROC) analysis and the Kaplan-Meier method with log-rank test were used to assess the relationship between the pre-treatment maximum ΔR2* and early progression (i.e. at 3-month follow-up). The pre-treatment maximum ΔR2* of patients with disease progression at 3-month follow-up (10.68 ± 3.88 s(-1)) was significantly different (p < 0.01) from that of patients with stable disease at 3-month follow-up (6.94 ± 3.12 s(-1)). Pre-treatment maximum ΔR2* of patients with disease progression at 6-month follow-up was not significantly different from that of patients with stable disease at 6-month follow-up, although a trend was noted (p = 0.08). Pre-treatment clinical factors were not significantly different between progressive and stable patients at 3- and 6-month follow-up. Tumor progression rate was significantly higher in patients with a higher pre-treatment maximum ΔR2* than in those with a lower pre-treatment maximum ΔR2* (median progression time, 38 vs 138 days, p < 0.01, using a cut-off value of 8.13 s(-1) as determined by ROC analysis). In conclusion, DSC-MRI may predict early progression in patients with advanced pancreatic cancer undergoing chemotherapy.