PET/MRI in pancreatic and periampullary cancer: correlating diffusion-weighted imaging, MR spectroscopy and glucose metabolic activity with clinical stage and prognosis.

Purpose: To correlate the clinical stage and prognosis of pancreatic or periampullary cancer with the imaging biomarkers on diffusion-weighted imaging, magnetic resonance spectroscopy and glucose metabolic activity derived from integrated PET/MRI.

Methods: This prospective study was approved by the institutional review board and informed consent was obtained. The study group comprised 60 consecutive patients with pancreatic or periampullary cancer who underwent PET/MRI before treatment. The imaging biomarkers were the minimal apparent diffusion coefficient (ADCmin), choline levels, standardized uptake values, metabolic tumour volume (MTV), and total lesion glycolysis (TLG) of the tumours. The relationships between these biomarkers and clinical TNM stage were evaluated using the Pearson test and the Mann-Whitney U test. The area under the receiver operating characteristic curve (AUROC) was used to evaluate accuracy. The correlation between the imaging biomarker and progression-free survival (PFS) was investigated using the Cox proportional hazards model.

Results: ADCmin was significantly lower in N1 and TNM stage 3+ tumours. Choline levels significantly higher in T4 tumours. TLG was significantly higher in T4, N1 and TNM stage 3+ tumours. MTV was significantly higher in T4, N1, M1, and TNM stage 3+ tumours (all P < 0.05). The MTV/ADCmin ratio exhibited the highest AUROC for predicting T4, N1, M1, and advanced TNM stages tumours, and was an independent predictor of PFS (P = 0.018) after adjustment for age, sex, tumour size and stage.

Conclusion: The imaging biomarkers from integrated PET/MRI may predict clinical stage and PFS in patients with pancreatic or periampullary cancer.