Classify it by site (intrahepatic, perihilar, distal) and growth pattern, then map what the surgeon needs: the longitudinal biliary extent (Bismuth–Corlette), hepatic artery and portal vein involvement, lobar atrophy and nodes.
Orient first
- Intrahepatic mass-forming tumours show peripheral arterial rim enhancement with progressive central (fibrous) enhancement on delayed images, capsular retraction and upstream duct dilatation.
- Perihilar (Klatskin) tumours are often small, periductal-infiltrating strictures — the dilated ducts point to them.
- Resectability is decided by bilateral second-order duct extension, bilateral or main vascular involvement, and a future liver remnant — the radiologist maps each.
Acquire the study
- Late arterial, portal venous and 3–5 min delayed phases, 1 mm with coronal and curved reformats along the ducts.
The manoeuvre
- Delayed phase: progressive enhancement of the fibrous tumour — defines its extent better than the portal phase.
- Coronal curved reformat: the upper and lower limits of the ductal stricture; classify Bismuth–Corlette I–IV.
- Arterial phase: right and left hepatic arteries — encasement vs abutment, and variants.
- Portal venous phase: main and branch portal veins; lobar atrophy with crowded dilated ducts (atrophy–hypertrophy complex).
- Nodes: hepatoduodenal, retropancreatic, and distant (para-aortic = M1 — verify the edition).
What confirms it
- A mass or stricture with delayed enhancement, upstream duct dilatation and no benign explanation (stone, IgG4, PSC stricture) — histology confirms.
What licenses you to exclude it
- Smooth tapering strictures without a mass, with elevated IgG4 or a stone, favour a benign cause — say "no imaging feature of cholangiocarcinoma" rather than excluding it.
The classic misread
- Reading the extent only on the portal venous phase and understaging the tumour.