Poster #78: Kasonde Chewe

Integrative Analyses of Biological Data on Multiple Scales

Timothy Khumpan 1,2, *, Kasonde Chewe 1,4, *, Namrata Vijayvergia 3, *, #, Eric Eleam 1, Juan Nicolas Quiñones-Romero 1, Suraj Peri 3,5, Gong Yulan 6, Paul F. Engstrom 3, Cai Qi 6, Jianming Pei 7, Gulnaz Alekbaeva 1, Kerry Campbell 8, Johnathan Whetstine 1, Igor Astsaturov 3, Hayan Lee 1,9,10 # 1 Cancer Epigenetics Institute, Nuclear Dynamics and Cancer Program, Fox Chase Cancer Center, Philadelphia, USA. 2 Department of Computer and Information Sciences, College of Science and Technology, Temple University, Philadelphia, USA.​ 3 Department of Hematology/Oncology, Fox Chase Cancer Center​, Philadelphia, USA. 4 School of Biomedical Engineering, Science, Health Systems, Drexel University, Philadelphia, USA. 5 Bioinformatics Oncology Target Discovery, Merck, Boston, USA 6 Histopathology Facility, Fox Chase Cancer Center, Philadelphia, USA.​ 7 Clinical Genomics, Fox Chase Cancer Center​, Philadelphia, USA. 8 Cancer Signaling and Microenvironment Program, Fox Chase Cancer Center, Philadelphia, USA. ​ 9 Department of Cancer and Cellular Biology, Lewis Katz School of Medicine, Temple University, Philadelphia, USA. 10 Department of Medical Genetics and Molecular Biochemistry, Lewis Katz School of Medicine, Temple University, Philadelphia, USA.

Neuroendocrine neoplasms (NENs) are a biologically heterogeneous group of epithelial malignancies arising across multiple organs. Although rare compared with other solid tumors, their incidence has increased more than six-fold in recent decades. A major clinical challenge is distinguishing aggressive, poorly differentiated neuroendocrine carcinomas (PD-NECs) from slower-growing, well-differentiated neuroendocrine tumors (WD-NETs), which differ substantially in prognosis and treatment response. Conventional histopathology may be insufficient for reliable classification or determination of tumor origin, motivating molecularly informed diagnostic approaches. Methods We performed transcriptomic profiling of 36 formalin-fixed, paraffin-embedded NENs, including 21 WD-NETs and 15 PD-NECs from 11 anatomical sites. We developed an interpretable AI-based histology classifier that achieved an area under the receiver operating characteristic curve of 0.87–0.93. Results PD-NECs exhibited upregulation of genes involved in cell-cycle regulation and DNA damage response, including SFN, CHEK1, E2F1, CDC6, and TTK. In contrast, neuroendocrine lineage-associated genes, including CAMK2B and PAK3, were enriched in WD-NETs. AI-derived transcriptomic biomarkers distinguishing WD-NETs and PD-NECs, PAK3 and EZH2, respectively, were independently validated by immunohistochemistry. Findings were further supported by an independent, publicly available lung NEN DNA methylation dataset. DNA methylation-based mitotic age analysis also demonstrated greater stemness and proliferative potential in lung NECs than in lung NETs, consistent with the transcriptomic findings. Conclusions This study establishes an integrative multi-omics framework for molecular classification of NENs that complements conventional histopathological assessment. By combining transcriptomic profiling, interpretable machine learning, and epigenetic validation, this approach may improve tumor stratification, identify mixed or potentially misclassified cases, and support biomarker-driven diagnostic and therapeutic strategies.