Reproducibility of Histopathological Subtypes in Pulmonary Adenocarcinoma. An International Interobserver Study
E Thunnissen, M Beasly, A Borczuk, E Brambilla, LR Chirieac, D Flieder, A Gazdar, K Keisinger, P Hasleton, Y Ishikawa, K Kerr, Y Matsuno, Y Minami, A Moreira, N Motoi, A Nicholson, M Noguchi, D Nonaka, G Pelosi, I Petersen, N Rekhtman, V Roggli, B Travis, M Tsao, I Wistuba, H Xu, Y Yatabe, J Kuik. Amsterdam, Netherlands; Winston Salem; New York; Grenoble, France; Boston; Philadelphia; Dallas; Manchester, United Kingdom; Tokyo, Japan; Aberdeen, United Kingdom; Tsukuba-shi, Japan; Nagoya, Japan; London, United Kingdom; Milan, Italy; Jena, Germany; Toronto, Canada; Durham; Rochester; Sapporo, Japan
Background: Histological subtyping of pulmonary adenocarcinoma has been associated with predictive and/or prognostic features, but inconsistencies may originate from difficulties in the proper subclassification. This study aimed to establish the reproducibility of these subtypes amongst pulmonary pathologists worldwide.
Design: In a ring study 26 pathologists involved in the IASLC reviewed 115 cases submitted by 20 pathologists, of POWERPoint images consisting of 5 typical histologic patterns: acinar (n=20), non-mucinous BAC (lepidic, n=19), micropapillary (n=16), papillary (n=19) and solid (n=20), in addition to a 6th problem case (n=21). All cases were randomized (JK) and reviewed in 3 subsequent rounds. For each case the reviewer provided a diagnosis of one dominant pattern and also additional pattern(s), if any. Kappa score was separately calculated for the 5 typical patterns and the difficult cases.
Results: In total number of diagnoses for the typical patterns combined and the difficult cases were 2444 and 546, respectively. The mean kappa score (±SD) for the 5 typical patterns combined and for difficult cases were 0.77 ± 0.07 and 0.38 ± 0.14, respectively. More than one pattern was reported for the typical and difficult cases in 35% and 65% of the cases, here the overlap was between papillary and micropapillary (22%); acinar and solid (15%); acinar and papillary (14%); BAC and papillar (10%); acinar and BAC (9%); acinar and micropapillary (6%); BAC and micropapillary (6%) and among several other combinations(19%).
Conclusions: Substantial reproducibility was found for typical patterns of pulmonary adenocarcinoma subtypes. When multiple patterns are present, reproducibility is fair.
Monday, March 22, 2010 1:15 PM
Platform Session: Section E, Monday Afternoon