Kim, Min‐Ji

Abstract Traditional cancer research methodologies predominantly focus on 2D tumor modeling, including spatial profiling and AI-driven analysis at the 2D level. To advance beyond these limitations, we implemented Holotomography (HT), a 3D, label-free optical imaging method. This technique, leveraging refractive index contrasts, allows for enhanced visualization of tumor tissues, organoids, and cells in three dimensions. Using serial H&E sections and unstained slides, we created detailed 3D tumor models in gastric cancer, providing insights into tumor, immune, and stromal composition, as well as cell-cell interactions and extracellular matrix characteristics. AI algorithms further enhanced our 3D tumor modeling, offering novel perspectives on the initiation and progression of tumor, immune, and stromal components at the subcellular level. Analyzing gastric tumors treated with immune checkpoint inhibitors before, during, and after treatment revealed dynamic changes in spatial composition over time, offering unique insights into the evolution of the 3D tumor ecosystem. Additionally, by integrating spatial transcriptomic and in situ molecular imaging data, we developed a novel Multimodal 3D Tumor Model. This model not only uncovers subcellular-level molecular characteristics but also elucidates cellular interactions during treatment, suggesting potential biomarkers and therapeutic approaches. Lastly, by constructing 3D cellular interaction networks incorporating true 3D interactions, we provided an in-depth view of cellular communication within the tumor microenvironment. This pioneering approach, representing the first Multimodal 3D Tumor Model at the subcellular level, offers novel insights into gastric cancer's spatial evolution, potentially revolutionizing our understanding and treatment of gastric cancer. Citation Format: Yoon Ho Choi, Minji Kim, Katya Redekop, Inyeop Jang, Jean R. Clemenceau, Isabel Barnfather, Sumanth Reddy Nakkireddy, Changjin Hong, Sunho Park, Ji-Youn Sung, Jeong Hwan Park, Shitara Kohei, Aoki Yu, Sung Hak Lee, Sam Wang, Jaeho Cheong, Tae Hyun Hwang. AI-driven multimodal 3D tumor modeling with spatial molecular insights for predicting ICI response in gastric cancer [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2024; Part 1 (Regular Abstracts); 2024 Apr 5-10; San Diego, CA. Philadelphia (PA): AACR; Cancer Res 2024;84(6_Suppl):Abstract nr 7400.

Abstract Background: Recent studies support that high levels of tumor-infiltrating lymphocytes (TILs) were associated with better prognosis in HPV(+) oropharyngeal squamous cell carcinoma (OPSCC). While these studies provide a potential role for TILs as a prognostic biomarker, the analyses were relied on manual quantification performed by pathologists, resulting in inter-observer variability. Deep Learning (DL)-based whole slide hematoxylin and eosin (H&E) image analyses may overcome these challenges. However, this analysis only provides the presence of cell types without consideration for the functional roles of each cell within the tumor immune microenvironment (TME). In this work, we develop a computational pipeline to integrate paired H&E and immunohistochemistry (IHC) images to functionally characterize TILs and investigate their prognostic utility. Methods: We analyzed 88 patients: 42 stage I, 39 stage II, and 7 stage III. Our data contain both H&E and IHCs examining FoxP3, CD3, PD-L1, CD20, etc. on serial sections of the tissue. In-house DL-based H&E analysis used to identify TILs, tumors, and stroma in each tumor, then performed registration between adjacent H&E and IHC images from the same tissue. The patients were then classified into three basic immune phenotypes: immune inflamed (IN; high TILs in the tumor region), immune excluded (EX; TILs are mostly localized in stroma), and immune desert (ID; few/no TILs) based on TIL enrichment in the TME. To functionally characterize TILs, we quantified protein expression from the adjacent IHCs. For example, we further classified each patient into different subtypes based on enriched protein expression (e.g., FoxP3 high IN, FoxP3 low IN). We used the Kaplan-Meier method and multivariate Cox analysis to evaluate the prognostic value of different subtypes enriched with different proteins to predict disease-free survival (DFS). Results: The IN group with 43 patients was significantly associated with good prognosis. Interestingly, further stratification of the IN subgroup based on Foxp3 quantification on TIL regions (i.e., high FoxP3 IN and low FoxP3 IN) showed that high protein expression of FoxP3 in TILs in the IN subgroup is significantly associated with a better prognosis compared to other immune subgroups (HR, 0.16; p-value, 0.003). Multivariate analysis, including other clinical covariates showed that the immune subtypes associated with high FoxP3 are independently associated with DFS. These results demonstrate that DL-based integrative IHC and H&E image analysis could be used to identify subgroups with distinct clinical outcomes. Furthermore, our results reveal unknown roles for Foxp3 expression in the TILs in HPV(+) OPSCC as a prognostic biomarker, a finding which should be evaluated in a larger cohort. Citation Format: Sumanth Reddy Nakkireddy, Inyeop Jang, Minji Kim, Linda X. Yin, Michael Rivera, Joaquin J. Garcia MD, Kathleen R. Bartemes, David M. Routman, Eric J. Moore, Daniel J. Ma, Kathryn M. Van Abel, Tae Hyun Hwang. Integrative spatial analysis of paired IHC and H&E images identifies Foxp3 enriched tumor-infiltrating lymphocytes associated with disease-free survival in human papillomavirus (HPV)-related oropharyngeal squamous cell carcinoma. [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2023; Part 1 (Regular and Invited Abstracts); 2023 Apr 14-19; Orlando, FL. Philadelphia (PA): AACR; Cancer Res 2023;83(7_Suppl):Abstract nr 4638.