202 Chapter 15 Abstract Background The definitive identification of the invasive component of mixed pattern non-mucinous adenocarcinoma of the lung is crucial, as only the invasive size is used for staging. The separation of true invasion from iatrogenic collapse is difficult. This study aims to investigate whether incorporating additional histopathological criteria for collapse enhances the identification of genuinely invasive regions in non-mucinous adenocarcinomas, and whether this impacts invasion measurements and subtype classification. Design Cases were obtained by combining two retrospective cohorts of pulmonary nonmucinous adenocarcinoma from institutes in different countries. Tumors were resected between the 1st January 2011 and the 31th December 2016, had a tumor diameter ≤3 cm, no metastases, and were linked to follow up data. Freshly cut sections of 70 cases were stained for H&E, elastin and CK7, scanned to digital image files, and uploaded to the Pathogate website. The cases were reviewed twice by 42 pulmonary pathologists to determine the presence of invasion: firstly, according to the 2021 WHO classification, and secondly using an alternative approach recognizing collapsed AIS. In between the 2 rounds an on-line tutorial was presented, together with written materials. The parameters recorded were presence of invasion, growth pattern percentages, total tumor size, location invasive area, invasive size, and an uncertainty score about decision of invasion. The heatmap analysis revealed areas more or less commonly scored as invasive in each case. Based on feedback from 41 peers in the first two rounds, which included scores and heatmaps highlighting the underlying histology of hotspots of invasion, a third round was conducted with the expectation that an updated classification system would result in improved consensus. Results A total of 42 pathologists from 13 countries scored all 70 cases in the first 2 rounds, while 36 pathologists scored 41 non-unanimous cases in the 3rd round. The kappa values for rounds 1, 2 and 3 were, 0.27, 0.45 and 0.62, respectively. In the first round, more than 21 pathologists scored “No Invasion” in 3 cases, and this increased to 10 cases in the 2nd and 3rd rounds; crucially none of these cases recurred. In measurements of total tumor size, the mean coefficient of variation was 6% in two rounds (range 1st round 1-26%; 2nd 1-29%). In contrast, when measuring invasive tumor size, the mean coefficient of variation rose from 34% in the first round to 40% in the second round, and these changes were for each case visualized by heatmap analyses. Growth pattern subtyping showed least variation for “solid” pattern: kappa 0.65 and for the others <0.3.
RkJQdWJsaXNoZXIy MTk4NDMw