Proposal for a new classification for perihilar cholangiocarcinoma based on tumour depth

Background The T system for distal cholangiocarcinoma has been revised from a layer‐based to a depth‐based approach in the current American Joint Committee on Cancer (AJCC) classification. In perihilar cholangiocarcinoma, tumour depth in the staging scheme has not yet been addressed. The aim of this study was to propose a new T system using measured tumour depth in perihilar cholangiocarcinoma. Methods Patients who underwent hepatectomy for perihilar cholangiocarcinoma between 2001 and 2014 were reviewed retrospectively. The vertical distance between the top of the tumour and deepest invasive cells was measured as invasive tumour thickness (ITT) by two independent pathologists. Log rank statistics were used to determine cut‐off points, and the concordance (C) index was used to assess survival discrimination of each T system. Results ITT was measurable in all 440 patients, with a median value of 6·0 (range 0–45) mm. The median difference in ITT between observers was 0·6 (range 0–20) mm. Cut‐off points for prognosis were 1, 5 and 8 mm. Five‐year survival decreased with increasing ITT (P < 0·001): 67 per cent for ITT less than 1 mm (25 patients), 54·9 per cent for ITT 1 mm and over to less than 5 mm (138 patients), 43·4 per cent for ITT 5 mm and over to less than 8 mm (118 patients), and 32·2 per cent for ITT 8 mm and over (159 patients). The C‐index of this classification was comparable to that of the current AJCC T classification (0·598 versus 0·589). Conclusion ITT is a reliable approach for making a depth assessment in perihilar cholangiocarcinoma. A four‐tier ITT classification with cut‐off points of 1, 5 and 8 mm is an adequate alternative to the current layer‐based T classification. This study proposed a new T classification using tumour depth, which was measured in 440 patients with perihilar cholangiocarcinoma. A four‐tier T classification using cut‐off points of 1, 5 and 8 mm showed prognostic discrimination comparable to that of the current layer‐based AJCC T classification. Might be more logical