P1.09-29 In Situ Growth Pattern in Lung Adenocarcinoma Is Divisible into Distinct Categories with Divergent Biological and Survival Implications

Abstract

The morphological stepwise progression of lung adenocarcinomas is well established, but little is known about the molecular events that underlie this. In particular, in situ patterns of growth are frequently seen in adenocarcinomas, and often it is not clear if this truly always represents early-stage preinvasive disease. Therefore we set out to better characterise in situ tumour growth, and to identify molecular and biological correlates of tumour invasion. We constructed a retrospective cohort of 964 locally held adenocarcinomas, with patient data and tissue microarrays (TMAs). Immunohistochemistry for Ki67 and epithelial-mesenchymal transition markers was applied to TMAs and quantified. In situ and adjacent invasive areas of 23 early tumours were subjected to further in situ assays and laser capture microdissection. Genomic DNA was extracted and driver genes were panel sequenced. We morphologically identified two distinct types of in situ tumour growth in early mixed pattern tumours: a low-grade precursor ('C1'- after the classic Noguchi classification) associated with higher-grade invasive disease, and a high-grade lepidic outgrowth ('C2') associated with invasive growth of similar cytological grade. C1 and C2-type in situ tumour growth are sharply separated by their proliferation rate (P=0.005) and their propensity for nodal metastasis (P=0.03), suggesting that this distinction is likely to be important in future grading/growth pattern classification. Furthermore, molecular analysis supports the classification; invasive areas of C1 tumours show driver mutations which are absent from neighbouring in situ disease (4/18 cases), indicating molecular progression, while in 5 sequenced C2 cases no evidence of molecular progression was seen. The difference between low-grade precursor and high-grade in situ patterns was further investigated in our full set of 964 tumours. We find that the prognostic power of proliferation rate (Ki67) is driven almost entirely by its effects in in situ areas of tumour growth. Proliferation rate in invasive tumour areas at most weakly predictive of patient outcome. We make several key findings: i) In situ disease in lung adenocaricnoma is divided into two biologically and prognostically distinct groups, with implications for our understanding of stepwise progression in lung cancer ii) These two groups can easily be separated on the basis of cellular proliferation rate ii) We identify mutations in key driver genes that are explicitly asociated with the transition from in situ to invasive growth iii) Proliferation rate is a potentially valuable prognostic marker, but this may be restricted to its ability to separate these two key biologically distinct growth patterns