Citations

Background. MicroRNAs are small, single-stranded non-coding RNA molecules that play major role in gene regulation. miRNAs have been reported as both tumor suppressive and oncogenic and are promising candidates for cancer prognosis and treatment. miR-494 is known to promote cell migration and invasion in CRC. Our study aimed to further explore the role of miR-494 modulation (inhibition and overexpression) in CRC. Methods. We evaluated the impact of tumor gene expression of MIR494 on outcomes (overall survival: OS; progression-free survival: PFS) in data for 433 metastatic CRC (mCRC) patients (pts) (bevacizumab [bev] (N = 226) vs cetuximab [cet] (N = 207)) enrolled in randomized phase III CALGB/SWOG 80405 trial obtained from Alliance for Clinical Trials in Oncology. RNA extracted from FFPE tumor samples was sequenced using the HiSeq 2500 (Illumina). Gene expression was divided into high, medium and low. Associations were assessed by log-rank test and multivariable Cox proportional hazards regression model. Treatment-gene interactions were analyzed using the likelihood ratio test. CRC cell lines HCT116 and HT29 were transfected with miR-494 inhibitors and mimics for 48 hours and total RNA was collected, followed by RT-qPCR (miR-494 and target genes BMAL1, PAK1, CDH1, CDH2, SNAI1, TWIST) and clonogenic assays. Results. High miR-494 was significantly associated with worse PFS in overall (median survival: high vs low: 9.3 vs 11.3 months, P<0.025) and cet-treated (median survival high vs low: 9.2 vs 9.6 months, P<0.025) mCRC cohorts, while no significance was observed in bev-treated mCRC. Multivariable Cox model showed significant association between high miR-494 and worse PFS and OS in the cet-treated cohort; however, treatment-gene interaction was not statistically significant. Expression analysis revealed miR-494 was significantly reduced and increased following inhibitor and mimic transfection, respectively. Differential gene expression of the target genes analyzed was observed between the miR-494 inhibitor and mimic-treated HCT116 and HT29 samples. In HCT116 cells, inhibition of miR-494 significantly (P < 0.01) reduced expression of PAK1, while increased the expression of CDH1. The number of colonies formed relatively increased after transfection with miR-494 inhibitor compared to miR-494 mimic. Conclusion. Our findings showed that high miR-494 expression correlated with worse PFS overall and in cet-treated mCRC pts. It would be interesting to check for potential associations with EGFR signaling (direct or indirect). Further studies are warranted for efficient targeting and to gain better understanding of the associated signaling involved in miR-494 modulation in CRC. Support: U10CA180821, U10CA180882; U10CA180888 (SWOG): Genentech. https://acknowledgments.alliancefound.org; P30CA014089, UG1CA180830 [HJL]. Clinicaltrials.gov Identifier: NCT00265850 Citation Format: Pooja Mittal, Joshua Millstein, Shivani Soni, Francesca Battaglin, Karam Ashouri, Sandra Algaze, Priya Jayachandran, Yan Yang, Lesly T. Gonzalez, Unnati H. Shah, Wu Zhang, Fang-Shu Ou, Alan P. Venook, Indrakant K. Singh, Heinz J. Lenz. Effect of miR-494 modulation in colorectal cancer (CRC): Clinical outcome, target expression and cell characteristics [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 2787.

Abstract A hallmark of GBM is its highly-immunosuppressive tumor microenvironment (TME). GBM stem-like cells (GSCs) can evade immune surveillance to allow for continued tumor growth and propagation. GSCs regulate the immune TME through recruitment and production of immunosuppressive cells and factors via tumor cell-intrinsic mechanisms. Blocking these cell-intrinsic immunosuppressive mechanisms is widely viewed as a way to enhance the effects of current immune therapies. We recently showed that TGFβ type II receptor (TGFBR2) promotes a mesenchymal transition in GBM cells characterized by high expression levels of immunosuppressive genes NT5E (CD73), ENTPD1 (CD73), LGALS1 (galectin-1), CD274 (PD-L1), PDCD1LG2 (PD-L2), and TGFB1. miRNA network analysis identified miR-16-5p and miR-124-3p as high-confidence candidates to block the expression of these immunosuppressive signaling axis in GBM cells. Transgenic miR-16-5p and miR-124-3p simultaneously reduced expression of CD274, ENTPD1, NT5E, TGFB1, and LGALS1 and increased tumor cell killing of PBMC-derived CD4+ and CD8+ T cells. We also measured an increase in drivers of the CD8+ effector state and a decrease in expression of exhaustion marker PD1 in PBMC-derived CD8+ T cells. In addition to the effects on CD4+ and CD8+ T cell activation and function, miRNA-mediated knock-down of this immunosuppressive gene signature in tumor cells also resulted in a decrease in CD206 expression and reduced release of anti-inflammatory chemokines in THP1 cells. These findings demonstrate that miRNA-based gene inhibition can block cell-intrinsic immunosuppressive pathways to enhance anti-tumor immune cell function in vitro and predicts that delivering these miRNAs using advanced nanocarriers will cooperate with current immunotherapies in vivo. Citation Format: TC Taylor, Amanda Johnson, Harmon Khela, Jack Korleski, Sophie Sall, Bachchu Lal, Karen Smith-Connor, Hernando Lopez-Bertoni. miR-16-5p and miR-124-3p block the immunosuppressive phenotype of GBM cells in vitro. [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 1357.

Abstract Background: Replication checkpoint modulators (RCMs), including inhibitors of CHK1, ATR, and WEE1, have been widely studied for their ability to inhibit cell cycle checkpoints and enhance the toxicity of agents that stall replication forks. In preclinical studies each of these inhibitors also demonstrates notable activity as a single agent against relapsed/refractory acute myeloid leukemia (R/R AML). In particular, our prior study (PMID: 33414171) demonstrated that several myeloid cell lines underwent cell death at nanomolar concentrations, with a median IC50 of 9 nM irrespective of the TP53 status. Examination of pathways involved in RCM-induced leukemia cell killing demonstrated that CHK1 inhibition induces CDC25A upregulation accompanied by activation of a death program that involves a CDK2-initiated 10- to 50-fold upregulation of TNFα and downstream engagement of the extrinsic apoptotic pathway. Here we have comprehensively characterized the processes that occur between the interruption of replication checkpoints and leukemic cell death in these sensitive cells versus cells selected for RCM resistance. Methods: Using the CHK1 inhibitors MK8776 and prexasertib, ATR inhibitors berzosertib and ceralasertib, and WEE1 inhibitor adavosertib as paradigm drugs, we examined signaling in two AML cell lines, U937 and THP.1, that were selected for CHK1 inhibitor resistance, examined cross-resistance patterns, and assessed the biochemical basis for leukemic cell death. Changes in gene expression and signaling induced by RCMs were analyzed in isogenic pairs using RNA sequencing, flow cytometry and immunoblotting. Results: In contrast to parental U937 and THP.1 cells, CHK1 inhibitor-selected U937 and THP.1 cell exhibited a lack of TNFα mRNA induction and did not undergo cell death in response to multiple CHK1 inhibitors. RNA sequencing of CHK1 inhibitor-selected AML lines revealed an almost complete absence of induction of CDK2-activated genes following prexasertib treatment. Interestingly, unlike solid tumors, AML cell lines that were resistant to CHK1 inhibitors exhibited hypersensitivity to WEE1 or ATR inhibition. Further studies demonstrated that CHK1 inhibitor response was varied in primary AML samples, with 50% of primary AML specimens resistant to CHK1i monotherapy, again without any relationship to TP53 status. The common features among sensitive vs. resistant leukemias was the successful inhibition of CDK2 phosphorylation and the associated induction of TNFα. Conclusions: Our findings reveals critical new insights into the factors that influence the response to CHK1 inhibitors in AML. Importantly, we found no evidence of cross-resistance to inhibitors of other replication checkpoint proteins, enabling alternative options. Citation Format: Annapoorna Venkatachalam, Husheng Ding, Nicole Vincelette, Cordelia McGehee, Kevin Peterson, Paula Schneider, Cristina Correia, Hu Li, Scott Kaufmann. CDK2 activation mediates response of acute myeloid leukemia to CHK1, ATR and WEE1 inhibitors [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 423.

Background: Tumor immune microenvironment (TIME) organization fundamentally shapes cancer progression and immunotherapy outcomes. Traditional 2D spatial molecular approaches inadequately capture the complex three-dimensional architecture and cellular interactions within TIME. We developed an AI-driven platform integrating multiple spatial molecular datasets, enabling comprehensive 3D visualization for therapeutic discovery and patient selection. Methods: Our platform combines holotomography imaging, providing 3D refractive index maps, with diverse spatial molecular data: in situ molecular imaging, spatial proteomics, multiplex immunohistochemistry, and spatial transcriptomics. AI algorithms align datasets within a unified 3D coordinate system, while Poisson implicit surface reconstruction generates subcellular-resolution models. Platform validation used clinical samples from immunotherapy-treated patients. Results: The platform integrated multimodal spatial data into high-resolution 3D TIME representations, revealing novel relationships between protein expression, transcriptional patterns, and cellular architecture. 3D surface protein analysis with subcellular morphology identified patients with target protein expression outside tight junctions, indicating higher risk for adverse effects from antibody-drug conjugates and cellular therapies. This insight, unavailable through 2D analysis, enabled precise patient selection. Conclusions: This platform provides detailed subcellular-resolution TIME architecture analysis, offering actionable insights for therapeutic targeting, biomarker discovery, and patient selection. AI-driven 3D spatial multimodal TIME analysis advances precision medicine across multiple therapeutic approaches, including immunotherapy, targeted therapy, and drug delivery optimization. AI Disclosure: Original abstract content was written by the authors. Claude was used for editorial assistance and formatting without altering the scientific content or context of the work. Citation Format: Inyeop Jang, Yoonho Choi, Minji Kim, Jeonghwan Park, Jiyeon Sung, Byungsoo Ann, Clemenceau R. Jean, TaeHyun Hwang. AI-enabled 3D multimodal integration platform reveals subcellular architecture of tumor immune microenvironment for therapeutic target and biomarker discovery and patient selection [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 3648.

Abstract Background: Understanding the spatiotemporal dynamics of the tumor microenvironment (TME) during treatment is crucial for improving outcomes in pancreatic ductal adenocarcinoma (PDAC). This study presents a pre-planned spatiotemporal analysis from a phase II trial evaluating neoadjuvant mFOLFIRINOX in resectable PDAC, aimed at elucidating treatment resistance mechanisms and identifying predictive biomarkers. Methods: This multi-institutional, single-arm phase II trial enrolled treatment-naïve patients with histologically confirmed resectable PDAC (NCT05066802) who underwent six cycles of neoadjuvant mFOLFIRINOX followed by surgery. The primary endpoint was the R0 resection rate. Matched pre-treatment tumor biopsies and post-treatment surgical specimens were analyzed from selected responders (R) and non-responders (NR) using high-resolution spatial transcriptomics (10X Genomics Visium). This approach enabled comprehensive mapping of TME evolution and treatment-induced changes. Results: Between May 2020 and February 2022, 27 patients were enrolled. Of these, 24 patients underwent surgery after neoadjuvant treatment (resection rate: 88.9%), and 23 achieved R0 resection (85.2%), meeting the primary endpoint. However, 20 resected patients (83.3%) recurred. Spatiotemporal analysis was performed on matched samples from four R and four NR. In pre-treatment samples from NR group, clusters dominated by fibroblasts (enriched for ECM, EMT, and CAF phenotypes) and B cell subtypes were distinct compared to R group, which exhibited elevated antibody responses and robust T cell and macrophage activity. NR samples showed increased IgM expression but lacked corresponding Fc receptors, suggesting limited immune effector function. Post-treatment NR surgical specimens displayed elevated expression of MUC5AC and CEACAM6 in malignant ductal cells, alongside genes promoting tumor growth. Neoadjuvant mFOLFIRINOX increased fibroblast, while reducing immune cell components. NR tumors formed distinct "resistance niches" after chemotherapy, marked by ECM production, CAF activation, and RTK signaling. Conversely, R tumors showed enhanced epithelial organization with increased type-I hemidesmosome complexes facilitating epithelial adhesion to the basement membrane. Conclusion: Three months of neoadjuvant mFOLFIRINOX was feasible and tolerable for resectable PDAC but associated with high recurrence rate, emphasizing the need for better patient selection. Spatiotemporal analysis revealed critical differences in TME dynamics according to treatment response, including the emergence of resistance niches and distinct pre-treatment spatial signatures. These findings support the integration of spatial biology into future pancreatic cancer trials to improve stratification and therapy. Citation Format: Choong-kun Lee, Changjin Hong, Ho Kyoung Hwang, Min Hwan Kim, Jean Rene Clemenceau, Seock-Jin Chung, Isabel Barnfather, Sumanth Reddy Nakkireddy, Jeong Hwan Park, Ji-Youn Sung, Minji Kim, Yooho Choi, Inyeop Jang, Sunho Park, Sung Hyun Kim, Chang Moo Kang, Hye Jin Choi, Tae Hyun Hwang. Spatiotemporal evolution of the tumor microenvironment predicts treatment response in phase II trial of neoadjuvant chemotherapy for resectable 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 771.

Abstract In the US and globally, Black men experience the highest prostate cancer burden among all racial and ethnic groups, underscoring the urgent need for research to guide the development and implementation of culturally tailored strategies aimed at reducing disparities and improving outcomes. In response, the Inclusive Cancer Care Research Equity (iCCaRE) Consortium was created to advance health equity and reduce disparities in prostate cancer. Importantly, a Survivorship Care Plan (SCP) is recommended to educate and empower patients, enabling them to be better informed about their cancer diagnosis, treatment, and survivorship care recommendations. To date, there is no SCP targeted to Black prostate cancer survivors (CaPS) who have the greatest disparities with the poorest survivorship outcomes (ACS F&F; AACR HD Report; Fullwood, 2024). Therefore, we aimed to develop a SCP tailored for Black men. Addressing the unique needs of Black men require a genuine bidirectional relationship between scientist and the community rooted in community engaged research (CER). The SCP was first informed by the literature (Bonner, 2024) and the ASCO template (canceradvocacy.org). With input from the iCCaRE leadership and the Partnership Engagement Services (PES) Core, the Whole Person SCP better responds to the informational, clinical and quality of life needs of Black Prostate Cancer survivors. Based on CER, we included 22 diverse, Black CaPs to further guide and tailor the iCCaRE. Our findings indicate that the iCCaRE Survivorship Care Plan (SCP) enhanced the ASCO template by incorporating and expanding several unique domains, including family history, care team contacts, follow-up and surveillance, symptoms, quality of life, survivor concerns, major comorbidities, health advisories, and culturally relevant resources, creating a comprehensive, whole-person SCP tailored specifically for Black men. It is crucial to continue to improve support for clinical research and advocate for Black men and their communities to reduce prostate cancer burden. Prostate cancer also needs to include incidence, mortality, and disparities for Black men to raise awareness of scarcely studied field of Black men and prostate cancer survivorship. Citation Format: Kimlin Tam Ashing, Timethia Bonner, Ewan K. Cobran, Folakemi Odedina, Aj Merriweather, Ernie Kaninjing, Floyd Willis, Cassandra Moore, Joanne Oliver, Fornati Bedell, John McCall. Advancing health equity and reducing disparities in prostate cancer in Black men: iCCaRE [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 994.

Abstract Most disease-associated single-nucleotide polymorphisms (SNPs) identified by genome-wide association studies (GWASs) reside in noncoding regions. There is increasing evidence that they play regulatory roles through epigenetic mechanisms by altering local chromatin state, 3D chromatin structure, and looping interactions. We and others have identified 42 CLL risk loci from GWASs. At 12 risk loci located outside of gene promoters, we previously demonstrated that functional SNPs are highly enriched in enhancers and super-enhancers that mostly target BCL2 family genes and transcriptional regulators. However, the study was limited by using only public epigenomic data generated in lymphoblastoid cell lines (LCL) and normal immune related cell types. To gain a better understanding of how the functional SNPs alter local chromatin state and modulate target gene expression in CLL, we generated RNA-seq data (n=588) and transposase-accessible chromatin with sequencing (ATAC-seq) data (n=69) in CLL tumor B cells. We performed molecular quantitative trait loci (QTL) analysis with index and correlated SNPs (R2 0.5) at 41 risk loci (excluding the HLA locus). The analyses identified cis expression QTLs (eQTLs), within 1Mb of genes’ transcription start site, at 25 risk loci and chromatin accessibility QTLs (caQTLs, within peak) at 11 risk loci, with both eQTLs and caQTLs at 8 risk loci. We identified a total of 54 eGenes, for which the expression is affected by nearby SNPs within 1Mb, including 40 novel eGenes at 16 CLL risk loci not previously reported. Notably, based on publicly available chromosome interaction data from CLL, LCL, and normal B cells, for over 70% of the novel eGenes, their promoters showed looping interactions with the regions carrying the eQTLs. Further, 13 SNPs were identified as both caQTL and eQTL, with the associated novel eGenes including CD81 at 11p15.5, IPCEF1 at 6q25.2, as well as CXXC1 and MBD1 at 18q21.1. The result suggests that these SNPs play a role in altering chromatin accessibility contributed to gene expression variation. At 11p15.5 risk locus, SNP rs2651823 (A/G) represents a novel caQTL and eQTL. The alternative G allele is associated with increased chromatin accessibility and CD81 expression. Hi-C data from GM12878 and CD20+ B cells supported interaction between the open chromatin region carrying rs2651823 and CD81 promoter. CD81 facilitates the trafficking of CD19 to the surface of activated B cells, and CD81-null mice showed impaired immune response. In a mouse model, CD81 contributes to tumor progression and metastasis in Lewis lung carcinoma. In summary, our molecular QTL analysis revealed that functional SNPs, located in CLL risk loci, modulate chromatin accessibility and expression of downstream target genes that could impact CLL risk. Citation Format: Huihuang Yan, Zhiquan Wang, Mingma Hoel, Cristine Allmer, Chantal McCabe, Daniel O’Brien, Dennis Robinson, James Cerhan, Sameer Parikh, Esteban Braggio, Susan Slager. Molecular QTL analyses identify functional variants at inherited risk loci in chronic lymphocytic leukemia [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 2763.

Abstract Background: Cancer-associated fibroblasts (CAFs) are a highly diverse cell population that plays a crucial role in shaping the tumor microenvironment (TME). Their spatial interactions within the TME remain to be systematically characterized, and the factors driving their phenotypic heterogeneity are not yet fully understood. Recent advancements in single-cell spatial transcriptomics technologies and cell segmentation approaches enable the precise measurement of thousands of transcripts at subcellular resolution. Combined with innovative bioinformatics, these advancements have unlocked unprecedented opportunities for in-depth characterization of CAFs and their interactions within the TME. Methods: We conducted a comprehensive spatial transcriptomic analysis of CAFs across 8 cancer types, using the CosMx and MERSCOPE platforms to profile nearly 6 million cells as part of our discovery cohorts. We developed a novel computational method to spatially cluster CAF subpopulations and defined four distinct spatial CAF subtypes. We further examined their phenotypic differences, spatial niche organization, and interactions with neighboring cells. Additionally, we investigated the influence of CAF subtypes on immune cell phenotypes and infiltration levels. To validate our findings, we employed multiple spatial platforms, including Visium, Xenium, COMET, CODEX, and IMC, to ensure reproducibility across diverse technologies. Moreover, we performed correlative studies in large-cohorts with rich clinical data to evaluate the clinical significance of these CAF subtypes. Results: Our approach provided an unprecedented high-resolution examination of the spatial phenotypes of CAFs in large, complex tissues, uncovering novel insights. We comprehensively characterized CAFs across various solid tumor types and identified four conserved spatial CAF subtypes. These subtypes are characterized by distinct cell organizational patterns, neighboring cell compositions, interaction networks, and transcriptomic features. Moreover, they demonstrate significant correlations with the abundance, distribution, and phenotypes of tumor-infiltrating immune cells, as well as associations with tumor histopathological features and clinical outcomes. Conclusions: Our study provides a pan-cancer, high-resolution characterization of CAF spatial organization and interactions in solid tumors and identified conserved spatial CAF subtypes linked to distinct organizational patterns. Our findings reveal that the transcriptomic states and phenotypes of CAFs are significantly shaped by their interactions with neighboring cancer and immune cells. Beyond offering novel insights into the spatial heterogeneity of CAFs, this work paves the way for developing strategies to effectively target and modulate CAFs for therapeutic benefit. Citation Format: Yunhe Liu, Ansam Sinjab, Jimin Min, Guangchun Han, Francesca Paradiso, Yuanyuan Zhang, Yang Liu, Ruiping Wang, Guangsheng Pei, Kyung Serk Cho, Enyu Dai, Akshay Basi, Jared K. Burks, Kimal I. Rajapakshe, Yanshuo Chu, Jiahui Jiang, Daiwei Zhang, Xinmiao Yan, Paola A. Guerrero, Alejandra Serrano, Mingyao Li, Tae Hyun Hwang, Andrew Futreal, Jaffer A. Ajani, Luisa M Solis Soto, Amir A. Jazaeri, Humam Kadara, Anirban Maitra, Linghua Wang. Spatial subtypes and cellular interactions of cancer-associated fibroblasts revealed by single-cell spatial omics [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 166.

Abstract Background: Chromosomal instability (CIN) is frequent in patients with metastatic castration-resistant prostate cancer (mCRPC) and determines taxane resistance. A circulating tumor cell (CTC) assay which morphologically identifies CTCs with CIN (CTC-CIN) from a single tube of whole blood could predict clinical outcomes in this population. We tested this hypothesis as a pre-planned, secondary outcome of the CARD trial, where patients with mCRPC progressing within a year of treatment with an androgen receptor pathway inhibitor (ARPI; enzalutamide or abiraterone acetate + prednisolone/prednisone) were randomized to treatment with cabazitaxel or the alternative ARPI. Methods: CTCs were isolated from blood samples obtained at baseline; cycle two, and end of treatment. The associations of baseline CTC and counts with imaging-based progression free survival (ibPFS), overall survival (OS), time to prostate-specific antigen (PSA) progression, RECIST 1.1 objective response rate (ORR), and PSA50 response rate (PRR) were assessed. Results: Of 255 patients randomized, CTC counts were acquired from 237 patients at screening, 213 patients at C2D1, and 166 patients at EOT. High baseline CTC counts were significantly associated with worse OS (median OS, 14.9 vs 9.8 months; univariate HR, 1.81; 95% CI, 1.30 - 2.51; p < 0.001). High baseline CTC-CIN counts were significantly associated with worse OS after adjustment for confounding variables (median OS, 15.3 vs 8.9 months; univariate HR, 2.16; 95% CI, 1.52 - 3.06; p < 0.001; multivariate HR, 1.56; 95% CI, 1.01 - 2.43; p = 0.047). Low CTC-CIN at baseline predicted ibPFS, ORR, and PRR benefit with cabazitaxel; conversely, high at baseline predicted no different ibPFS or OS outcomes with cabazitaxel or a second ARPI. Conclusions: This preplanned biomarker analysis of CARD confirmed that CTC-CIN counts are prognostic for OS and may be predictive of cabazitaxel benefit. (Funded by Sanofi and Epic Sciences; CARD ClinicalTrials.gov number, NCT02485691.) Citation Format: Ossian Longoria, Jan Rekowski, Santosh Gupta, Nick Beije, Klaus Pantel, Eleni Efstathiou, Cora Sternberg, Daniel Castellano, Karim Fizazi, Bertrand Tombal, Adam Sharp, Oliver Sartor, Sandrine Mace, Christine Geffriaud-Ricouard, Richard Wenstrup, Ronald de Wit, Johann de Bono. Morphological evaluation of chromosomal instability in circulating tumor cells and prediction of taxane resistance in metastatic prostate cancer: a prespecified sub-analysis of the CARD trial [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 1145

Background: Cholangiocarcinoma is a molecularly heterogenous cancer arising from the biliary epithelium. Individual omic techniques have identified clinically relevant, targetable subtypes but most tumors have no targetable alterations based on exome or transcriptomic evaluation. Integrated multiomics approaches incorporating transcriptomic, proteomic, and phosphoproteomic characterization provide deeper understanding and identify non-mutated, activated pathways that can be targeted. To optimize the potential clinical utility of these approaches, minimal classifier features must be defined. Methods: Approximately 60,000 RNA, protein, and phosphoprotein features were extracted from 35 cholangiocarcinomas treated at the primary US study site. 206 patients treated and profiled at an institution in China were also included. Using the Multiomics Factor Analysis clustering method, patients were sorted into 3 distinct molecular profiles. Important molecular features were selected using competing feature selection (FS) methods: Boruta, RreliefF, variable selection using random forests (RF), recursive factor elimination (RFE), median decrease accuracy in RF, and Concrete Autoencoders. Machine learning classification prediction models including Extreme Gradient Boosting, gradient-boosted trees, multinomial logit, and support vector machine (SVM) were tested on each FS. Each combination of FS and predictive model was tested with 10 replicates of 10-fold cross-validation of a randomly-selected balanced test set and an oversampled, balanced training subset. Accuracy, F1 score, AUC, precision, and recall were assessed for each FS and predictive model combination. A custom function of physical assay cost and classification accuracy was optimized to find the best assay. KEGG pathway overrepresentation analysis was used to analyze feature subsets. Results: A consistent set of 50 important RNA and protein features was identified, including previously reported prognostic factors, like SpryD4 protein expression, as well as previously unreported pathway surrogates, such as Slc2A2 or Pipox protein. Pathway overrepresentation analysis found insulin signaling and resistance and amino acid biosynthesis/metabolism pathways were overrepresented in the discriminative feature set. The gradient-boosted trees algorithm combined with the RFE FS method had the best performance, with a multiclass macro average (±sd) cross-validation accuracy 89% (±6%), F1 score 89% (±8%), AUC 0.98 (±0.02), precision 90% (±10%), recall 89% (±6%). Conclusions: Feature selection methods and predictive machine learning models work synchronously to develop accurate molecular subtype assays based on just 50 protein and RNA levels. This strategy has the potential to make molecular subtyping for cholangiocarcinoma and other molecularly complex cancers fast and widely available. Citation Format: Ellen Larson, Erik Jessen, Dong-Gi Mun, Jennifer Tomlinson, Amro Abdelrahman, Danielle Carlson, Hojjat Salehinejad, Rory Smoot. Feature selection and machine learning strategies optimize an affordable molecular assay for cholangiocarcinoma subtype [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 1107.