Teaching Case

79-year-old male with comorbidities including diabetes, hypertension, ischaemic heart disease, peptic gastritis and HCV positivity, suffering from upper abdominal pain, nausea and malaise.
At admission, physical examination revealed jaundice (corresponding to 7.4 mg/dl total bilirubin, mostly direct). Laboratory abnormalities included increased serum creatinine (1.6 mg/dl), liver transaminases, gamma-glutamyltranspeptidase and alkaline-phosphatase.

Sonography (Fig. 1) detected a 3-cm hypoechoic pancreatic head mass causing mild dilatation of common bile duct (CBD) and left lobe intrahepatic ducts.
CT (Fig. 2) confirmed circumscribed, solid lesion with loss of normal pancreatic lobulations and parenchymal enhancement, atrophied upstream body and tail with dilated main pancreatic duct (MPD), without calcifications.
MR-cholangiopancreatography (Fig. 3) depicted T1-hypointense, mildly T2-hyperintense pancreatic mass with restricted diffusion, biliary obstruction, dilated MPD and saccular dilatations of side branches, strongly suspicious for carcinoma.
Endoscopic ultrasound (not shown) confirmed hypoechoic T2N1 mass with small-sized regional adenopathies. Despite inconclusive biopsies, considering elevated serum CA19-9 (122.5 U/ml after adjustment for bilirubin), Whipple pancreaticoduodenectomy was performed. Pathology diagnosed mass-forming chronic pancreatitis (MF-CP) with mixed fibrosis and inflammation, which could be suggested by low normalised apparent diffusion coefficient (Fig. 3g) and smoothly tapering MPD entering the mass ("duct-penetrating sign", Fig. 3i).
The table (Fig. 4) summarizes some useful features to discriminate between pancreatic head cancer and MF-CP.

Most commonly alcohol-related, chronic pancreatitis (CP) is characterised by progressive inflammatory damage with pancreatic fibrosis, resulting in irreversible exocrine and endocrine functional impairment. Anatomic changes may include atrophy, focal or diffuse enlargement, calcifications, dilatation of main pancreatic duct (MPD) and side branches. Clinical manifestations such as abdominal pain, anorexia and weight loss commonly suggest tumour, particularly in case of jaundice, elevated CA19-9 marker, or lacking history of pancreatitis [1, 2].
Furthermore, CP is associated with an increased risk of pancreatic carcinoma, which coexists in 1-6% of patients. Unfortunately, despite state-of-the art cross-sectional techniques, differentiation between CP and cancer is challenging when the former presents as focal mass without calcifications. As a result, 5-10% of patients having pancreatico-duodenectomy for suspected malignancy are ultimately found to have benign disease at pathology, and unnecessarily undergo high iatrogenic morbidity and non-negligible mortality. Therefore, correct preoperative diagnosis is crucial, and generally requires endoscopic ultrasound-guided biopsy, which however has limited (75-80%) yield [3, 4].
Often misinterpreted as tumour, mass-forming CP (MF-CP) shows hypo-to-isointense T1-weighted signal intensity reflecting chronic inflammation and fibrosis, is mostly T2-iso-to-hyperintense, shows variable dynamic patterns with loss of normal early homogeneous pancreatic enhancement, and may cause dilatation of MPD, common bile duct (CBD) or both. Albeit commonly regarded as characteristic of pancreatic head cancer, the “double-duct sign” (dilated MPD plus CBD) has 46-62% sensitivity and 78-83% specificity. With close inspection, a non-obstructed or smoothly tapering MPD entering through the mass (“duct penetrating sign”) favouring a MF-CP diagnosis may be differentiated from abrupt ductal cut-off suggesting tumour [3-6].
Current MRI protocols routinely include diffusion-weighted (DWI) acquisitions, which have been investigated for differentiating pancreatic carcinoma versus MFCP, reaching an overall 82% specificity in a meta-analysis; however, DWI results are questionable due to limited samples and technical inconsistencies. Pancreatic malignancies show restricted diffusion and significantly lower apparent diffusion coefficients (ADC) values than normal parenchyma and benign tumours, but their ADC increases when necrosis is present. Since MF-CP may contain variable proportions of fibrosis, the major DWI limitation is the inability to distinguish between inflammation and carcinoma due to overlapping ADC values. As in this case, normalisation of ADC (calculated as ratio of pancreatic lesion to apparently normal adjacent pancreas) has been reported to improve discrimination of MF-CP (range 0.708-0.890x10-3 mm2/sec) from pancreatic tumours (0.895-0.985x10-3 mm2/sec) [3, 7-10].