Double duct dilatation (DDD) without an associated mass found on cross-sectional imaging warrants further investigations to exclude an occult periampullary tumour, although benign findings are common. Endoscopic ultrasound (EUS) can be performed in these patients. We aim to investigate the difference in longitudinal outcomes of patients who underwent EUS versus those who did not, and whether there were predictive factors for findings of pancreaticobiliary malignancies in these patients. 

Patients with DDD between January 2015 to December 2023 were identified from radiology electronic patient records within a tertiary centre based in the United Kingdom. A dedicated radiologist re-reviewed the index scans to confirm the presence of DDD. DDD was defined as a common bile duct of greater than 7mm and a pancreatic duct of greater than 3mm at the pancreatic head. A retrospective analysis using clinical data from electronic patient records was performed. Follow-up of each case was defined as the point of review of their medical records, date of development of a malignant pancreaticobiliary condition or the date of death. Data was analysed using chi-squared, mann-whitney u test and fisher exact test depending on the type of data. 

Initial search of the hospital medical imaging system resulted in 470 scans. The list was subsequently cut down to 153 individual cases by exclusion criteria including no DDD or associated mass found on imaging. 56% had EUS (G1) performed following their scans, whilst 44% had no EUS (G2). The median age of G1 was lower than G2 [median 68 years (58-75) vs 77 (69-85.5), p<0.00001]. G1 had 53% females and G2 had a higher proportion of 71% (p=0.03).  

Baseline characteristics revealed G1 had significantly more patients with chronic pancreatitis (19% vs 6%, p=0.03). Frequency of previous cholecystectomy (12% vs 10%, p=0.77), opioid use (35% vs 34%, p=0.85) and previous endoscopic retrograde cholangiopancreatography (4% vs 1%, p=0.63) were similar in both groups. Biliary pain and jaundice followed a similar pattern of increased frequency in G1 but with no significant differences (39% vs 28%, p=0.16) and (25% vs 18%, p=0.29) respectively. Also, there was no difference with regards to weight loss (25% vs 26%, p=0.8). 

Abnormal LFTs were similar in both groups (56% vs 60%, p=0.76). However, ALT was found to be significantly different [median 44 (17.5-154) vs 32 (12-85), p=0.03]. Bilirubin [median 10 (6-29.5) vs 9 (6-18), p=0.42], ALP [median 161 (95.5-485.5) vs 155 (91-390), p=0.58] and Ca19-9 [median 28 (10-60) vs 19 (8.5-50), p=0.42] had no significant differences. Proportions of computed tomography (52% vs 60%, p=0.29) and magnetic resonance imaging (48% vs 40%, p=0.29) scans were similar between both groups. Common bile duct [median 13 (11-17) vs 13 (11-18), p=0.51] and pancreatic duct [median 6 (5.5-7) vs 7 (6-7.5), p=0.12] sizes were also similar.

Lastly, median follow-up was significantly different with a longer time frame in G1 [40.6 months (26.5-65.9) vs 34.4 (13.4-56.1), p=0.01]. The incidence of malignancy, however, was similar (5% vs 13%, p=0.08). Patients diagnosed with malignancy during index EUS (n=14) had predictive factors of jaundice (n=11), weight loss (n=6) and biliary pain (n=6). The same factors were frequent in patients diagnosed during follow-up too (n=13) with jaundice (n=10), weight loss (n=3) and biliary pain (n=3). 

Our study has demonstrated in cases where no cause of DDD is seen on cross-sectional imaging or EUS, there are low rates of malignancy in the follow-up period. Though not statistically significant, patients who did not have EUS had an increased occurrence of malignancy. In the absence of jaundice, weight loss and biliary pain, the likelihood of malignant pathology is low. Given the risks associated with pancreaticobiliary endoscopy, especially in an aging population, this study stresses the importance of careful case selection and allowing time to counsel patients for invasive investigations, if deemed appropriate.