Adenocarcinoma In Situ, Minimally Invasive Adenocarcinoma and Invasive Pulmonary Adenocarcinoma – Evaluation of Interobserver Agreement in 294 Nodules with Survival Analysis
Jennifer M Boland, Jason A Wampfler, Ping Yang, Marie Christine Aubry, Mariza de Andrade, Eunhee S Yi. Mayo Clinic, Rochester, MN
Background: Recent studies have shown patients with pulmonary adenocarcinoma (PADC) showing entirely lepidic growth (adenocarcinoma in situ, AIS) or invasion ≤5 mm (minimally invasive adenocarcinoma, MIA) have superior outcome compared to patients with conventional invasive adenocarcinoma (IA). This has led to the new IASLC/ATS/ERS classification of lung adenocarcinoma proposed in early 2011. Data on interobserver variability within this new classification system is limited. Further validation of the superior survival of patients with AIS and MIA is also needed, since these tumors are rare and often make up a small minority of tumors in large studies.
Design: Patients were selected from the Mayo Clinic Epidemiology and Genetics of Lung Cancer Study database who underwent surgical resection of PADC at Mayo Clinic from 1997-2010. Cases were enriched for AIS and MIA by giving priority to cases diagnosed as bronchioloalveolar carcinoma (BAC) or adenocarcinoma with BAC features. 294 nodules were reviewed from 254 patients. Existing pathology slides were reviewed by two independent pathologists, who measured sizes of maximum invasion and central scar, if present.
Results: 237 of 294 nodules (81%) were classified into the same invasive category by both observers: 11 AIS, 71 MIA and 155 IA (kappa= 0.62, 95% CI 0.54-0.71). In 9 cases (3%) there was a disagreement between AIS and MIA. In 48 cases (16%) there was a disagreement between MIA and IA. The average difference in invasion measurement between observers was 3.4 mm. The average difference in scar measurement (when present) was 2.6 mm. Overall 5 year survival was significantly different among categories as determined by both observers: for observer 1,100% for AIS, 78% for MIA (relative risk over AIS (RR)=3.3) and 63% for IA (RR=7.6) (p=0.0007); for observer 2, 83% for AIS, 79% for MIA (RR=1.8) and 60% for IA (RR=4.5) (p=0.0001). The 5 year survival for the 48 patients where there was disagreement in classification between MIA and IA was 77%.
Conclusions: Moderate to good agreement was seen between observers (kappa 0.62), with concordance in the determination of AIS, MIA and IA in 81% of cases, and correlation between the combined AIS/MIA group versus IA in 84% of cases. While patients with MIA had better survival than IA, outcome was not as good as patients with AIS. In cases where there was a disagreement between MIA versus IA, survival was similar to the MIA group.
Monday, March 19, 2012 1:00 PM
Platform Session: Section D, Monday Afternoon