Abstract 6091: Evolutionary characterisation of lung adenocarcinoma pathological subtypes in TRACERx

Abstract Background: Lung adenocarcinoma (LUAD) is a morphologically and genetically diverse disease. The prognostic impact of LUAD histological patterns have been described, such as solid growth pattern and poor outcomes, though their underlying biology is poorly understood. Furthermore, the genomic characteristics and evolutionary constraints in relation to the inter- and intra- tumoral variance of histological patterns in primary and metastatic disease are unknown. Methods: Pathological classification of 246 patients with LUAD from the TRACERx 421 cohort was performed at the whole tumor (diagnostic samples) and multi-regional sample level (matched for tumor whole exome sequencing and RNA sequencing). Circulating tumor DNA (ctDNA) data was also integrated to determine the relationship between pathological subtypes and ctDNA detection. Results: Chromosomal instability, characterized by fraction of the genome affected by subclonal copy number alterations was significantly correlated with proportion of high-grade patterns, namely solid, cribriform and micropapillary (Spearman’s Rho 0.27, p<0.001). Analysis of somatic copy number alterations (SCNAs) and driver mutation profiles showed that specific SCNAs were associated with a predominant growth pattern, such as 3q arm gains in predominantly cribriform and solid pattern tumors. Multiregional analysis of tumors with mixed patterns showed higher grade regions to be associated with a higher frequency of LOH and expression of proliferation-related pathway genes, suggesting intra-tumoral sequential evolution from low to high grade growth patterns. No recurrent subclonal mutations or SCNAs were found to associate with progression from low to high grade patterns. The growth pattern in metastatic tumors tended to show similar or a higher-grade pattern compared with primary tumor regions harboring metastasizing clones (seeding regions). The growth pattern of the seeding regions in the primary tumor was not necessarily higher grade compared with their non-seeding counterparts. Finally, the proportion of solid pattern in the primary tumor and the presence of necrosis were found to be strongly associated with pre-operative ctDNA detection, while histological ‘spread through air spaces’ (STAS) was identified in 92% (12/13) of pre-operative ctDNA-negative tumors that subsequently were associated with recurrence. Patients with both pre-operative detectable ctDNA and STAS had a particularly poor prognosis. Conclusion: These data reveal insights into the association between morphological and molecular heterogeneity in LUAD, describe key features of tumor evolutionary tendencies and demonstrate the utility of detailed tumor morphological assessment integrated with molecular characterization and ctDNA detection. Citation Format: Takahiro Karasaki, David A. Moore, Selvaraju Veeriah, Cristina Naceur-Lombardelli, Antonia Toncheva, Maise Al Bakir, Thomas B. Watkins, Oriol Pich, Alexander M. Frankell, Emilia Lim, Mark S. Hill, Kristiana Grigoriadis, Carlos Martinez-Ruiz, James R. Black, Clare Puttick, Dhruva Biswas, Ariana Huebner, Michelle Dietzen, Emma Colliver, Claudia Lee, Nnenna Kanu, Sadegh Mohammad Saghafinia, Francisco Gimeno Valiente, Christopher Abbosh, Crispin T. Hiley, Simone Zaccaria, Nicolai J. Birkbak, Allan Hackshaw, TRACERx Consortium, Teresa Marafioti, Roberto Salgado, John Le Quesne, Andrew G. Nicholson, Nicholas McGranahan, Charles Swanton, Mariam Jamal-Hanjani. Evolutionary characterisation of lung adenocarcinoma pathological subtypes in TRACERx [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2022; 2022 Apr 8-13. Philadelphia (PA): AACR; Cancer Res 2022;82(12_Suppl):Abstract nr 6091.