Abstract 764: Molecular diagnosis using spatial analysis of lung tumors with a history of pancreatic cancer

Abstract Background: In pancreatic cancer (PC) patients, the lung is the third most common site of recurrence after the liver and peritoneum. Accurate histological diagnosis of lung tumors is essential for determining treatment strategies. However, pancreatic cancer lung metastases (PCLM) and invasive mucinous adenocarcinoma (IMA), a subtype of primary lung adenocarcinoma, share histopathological features, such as mucin-rich tall columnar epithelial cells growing in glandular or papillary structures with a lepidic growth pattern. Even with immunohistochemical markers like TTF-1 or CK20, differentiation remains challenging. This study aimed to identify molecular markers unique to PCLM and IMA using spatial multi-omics analysis. Materials and Methods: We performed spatially resolved transcriptomics/proteomics on surgically resected lung tumor specimens obtained from patients with a history of PC. A total of nine patients were included: five diagnosed as possibly PCLM based on clinicopathological features and four with tumors difficult to distinguish between PCLM and primary lung IMA. Regions of interest (ROIs) were selected in each tumor component and the tumor microenvironment (TME). Gene and protein expression profiles were comprehensively analyzed using the GeoMx digital spatial profiler in each ROI. Additionally, primary PC specimens were analyzed in five cases where there were available, and three patients of primary lung IMA without a history of PC were included as controls. Results: The median age of the nine patients with PC were 67 years (range: 54-80 years), with four males and five females. Histologically, all patients exhibited tall columnar epithelial cells with glandular or papillary structures. Among them, five patients had limited or sparse mucin, leading to a diagnosis of possibly PCLM, while the remaining four were challenging to distinguish. Spatial multi-omics analyses stratified the nine patients into subtypes, revealing distinct molecular signatures. Comparison with IMA-specific signatures identified molecular markers to aid in differentiating PCLM from primary lung IMA. Additionally, differences in gene/protein expression profiles, including within the TME, were observed between primary pancreatic tumors and lung metastases. Conclusion: This study identified molecular markers that could distinguish PCLM from primary lung IMA despite their similar imaging and histological features. These findings suggest that spatial multi-omics analysis could be a valuable tool for improving the differential diagnosis of PCLM and IMA. Citation Format: Kosei Ishimura, Kazuhiko Shien, Shuta Tomida, Hidetaka Yamamoto, Hirofumi Inoue, Kazuya Hisamitsu, Ryota Fujiwara, Atsushi Matsuoka, Ryunosuke Fujii, Ryo Yoshichika, Tomohiro Higashihara, Naohiro Hayashi, Yuma Fukumoto, Kazuhiro Okada, Fumiaki Mukohara, Mao Yoshikawa, Ken Suzawa, Shinichi Toyooka. Molecular diagnosis using spatial analysis of lung tumors with a history of pancreatic cancer [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2025; Part 1 (Regular Abstracts); 2025 Apr 25-30; Chicago, IL. Philadelphia (PA): AACR; Cancer Res 2025;85(8_Suppl_1):Abstract nr 764.