Purpose: Grading of chronic pancreatitis (CP) is a clinical and radiologic challenge. Retrograde cholangiopancreatography (ERCP) and magnetic resonance cholangiopancreatography (MRCP) use a version of the Cambridge criteria for ductal evaluation and CP staging, but interchangeability between the modalities lacks validation. This work compares ERCP and MRCP Cambridge scores and evaluates diagnostic performance of MRCP in a large cohort of patients with CP.
Methods: A large radiology database was searched for CP patients who underwent MRCP between 2003 and 2013. Next, patients who also had an ERCP within 90 days of their MRCP were selected. These were categorized into mild, moderate, and severe CP using the standardized Cambridge classification for ERCP. Radiologists blinded to ERCP findings then rated MRCP with modified Cambridge scores.
Results: The cohort comprised 325 patients (mean age 51 years; 56% female). By ERCP Cambridge classification, 122 had mild CP, 109 moderate CP, and 94 severe CP. MRCP and ERCP showed total agreement of Cambridge score in only 43% of cases. With ERCP as reference, the sensitivity and specificity of MRCP in detecting Cambridge scores 4 + 5 (main-duct predominant) were 75.9% and 64.3%, and for Cambridge score 3 (side-branch predominant) it was 60.0% and 76.9%, respectively.
Conclusions: There is a lack of strong concordance between ERCP- and MRCP-based grading of CP using the Cambridge criteria. MRCP had moderate to good performance in diagnosing side-branch predominant versus main-duct predominant CP. This suggests an inherent challenge in comparing literature and calls for a revision of the standards.
Keywords: Chronic pancreatitis; Diagnosis; ERCP; Magnetic resonance cholangiopancreatography.