Introduction
EASL guidelines recommend surveillance of patients with primary sclerosing cholangitis (PSC) with magnetic resonance cholangiography (MRC). Relevant stricture (RS), defined as presence of high-grade stricture (HGS) in MRC and signs or symptoms of cholestasis, is an indication for endoscopic retrograde cholangiography (ERC)1. Dominant stricture (DS) is the only established definition in ERC and a risk factor for cholangiocarcinoma and worse prognosis2-4. Whether HGS and RS correlate with presence of DS in ERC or biliary neoplasia is unknown.
Aims & Methods
To evaluate the predictive value of HGS and RS for the presence of DS in ERC and prevalence of biliary dysplasia in PSC patients with HGS, RS and DS.
We recruited all PSC patients with MRC done ≤4 months before ERC in two tertiary centers 10/2021-1/2024. Two expert radiologists evaluated MRC images in consensus, noting presence and location of HGS. RS was defined as HGS with cholestatic symptoms or elevated bilirubin before ERC. At ERC, disease severity was assessed with Helsinki score5, presence and location of DS, biliary brush cytology (Pap stain: 1. benign 2. atypia suspicious of dysplasia or malignancy 3. malignant), lab values (ALP, Bil) and cholestatic symptoms (simple cholestatic complaints score, SCCS6) were noted. Response to endoscopic treatment was defined as 20% decrease in ALP and/or Bil after 4-16 weeks.
Results
Of 128 patients 81 (63%) were male, median age was 42 years (range 20-73), median PSC duration 6 years (range 0-35), and 99 (77%) had inflammatory bowel disease.
HGS was seen in 94 (73%) patients and 57 (45%) had DS at ERC. Of 123 patients with complete cholestatic symptoms data, 44 (36%) had RS.
Positive predictive value (PPV) of HGS identifying DS in the same bile duct segment was 40% (95% CI 22-61%) in common bile duct, 31% (95% CI 19-48%) in common hepatic duct, 32% (95% CI 22-45%) in right hepatic duct and 50% (95% CI 35-65%) in left hepatic duct. HGS associated with higher cholestatic lab values, worse bile duct disease, cholestatic symptoms, and need for endoscopic dilatation (Table). 6/57 (11%) patients with DS lacked HGS in MRC.
| Variable | HGS, n = 94 | No HGS, n= 34 | p-value |
| Helsinki score > 7, n (%) | 70 (74) | 14 (44) | 0.01 |
| Helsinki score (0-16), median (range) | 8 (2-12) | 6 (2-14) | 0.002 |
| Dilatation in ERC, n (%) | 70 (74) | 15 (44) | 0.003 |
| ALP (IU/l), median (IQR) | 199 (114-349) | 119 (94-202) | 0.009 |
| Bil (µmol/l), median (IQR) | 14 (9-19) | 11 (7-18) | 0.05 |
| Symptoms (SCCS > 0), n (%) | 29 (31) | 7 (21) | 0.02 |
Sensitivity (Sn), specificity (Sp), PPV, and negative predictive value (NPV) of HGS predicting DS in ERC were 89% (95% CI 78-98%), 39% (95% CI 29-51%), 54% (95% CI 44-64%) and 82% (95% CI 66-92%).
Sn, Sp, PPV, and NPV of RS predicting DS in ERC were 45% (95% CI 33-67%), 72% (95% CI 60-81%), 57% (95% CI 42-70%) and 62% (95% CI 51-72%).
71 patients with elevated ALP and/or Bil received endoscopic treatment. Treatment response did not correlate with stricture findings: in patients with HGS, RS, and DS, 19/51 (37%), 13/33 (39%), and 16/33 (48%) showed response, compared to 3/6 (50%), 8/21 (38%) and 6/24 (25%) in patients without these strictures. 8/128 (6%) patients had atypia suspicious of malignancy or dysplasia in brush cytology, with no correlation to HGS (5%), RS (5%) or DS (7%) .
Conclusion
Prevalence of DS was similar in patients with HGS (54%) and RS (57%). Absence of HGS had good rule-out power for DS (NPV 82%). HGS associated with more advanced bile duct disease and need for endoscopic treatment. Prevalence of biliary cytology suspicious for dysplasia or malignancy was similar in patients with and without HGS, RS, or DS.
References
- Chazouilleres et al. Journal of Hepatology 2022; 77: 761-806.
- Rudolph G, et al. J Hepatol 2009; 51: 149-155.
- Tenca A, et al. United European Gastroenterology Journal 2024; 12: 717-725.
- Chapman MH, et al. Eur J Gastroenterol Hepatol 2012; 24: 1051-1058.
- Boyd S, et al. Endoscopy 2016; 48: 432-439.
- Munster KN, et al. Liver International 2020; 40: 2758-2766.
