The Prostate Cancer Foundation Challenge Awards are specifically designed to provide financial support to bold, cross-disciplinary research efforts that otherwise might struggle to secure funding. Dr. Khurana’s team, which notably includes Dr. Iman Hajirasouliha—associate professor at Weill Cornell Medicine—and physician-scientist Dr. Yu Chen alongside pathologist Dr. Anuradha Gopalan from Memorial Sloan Kettering Cancer Center, exemplifies the power of integrating diverse scientific disciplines to tackle complex biomedical challenges. Their joint mission focuses on leveraging advanced computational methodologies to combat the growing challenge of treatment-resistant prostate cancer.

Prostate cancer remains one of the most commonly diagnosed malignancies among men in the United States, with approximately 300,000 new cases identified annually. The lifetime risk of developing this disease stands at nearly 12%. Tradition dictates that prostate tumor proliferation is largely driven by androgen receptor signaling pathways, which can be disrupted therapeutically through androgen deprivation therapy or direct androgen receptor inhibitors. Despite the initial efficacy of such treatments, a significant subset of tumors evolve into aggressive subtypes that bypass these signaling dependencies, resulting in treatment resistance. Unfortunately, current clinical diagnostic tools lack the precision necessary to detect these resistant tumor phenotypes during their early emergence.

Building on foundational work that identified key treatment-resistant prostate tumor subtypes, Dr. Khurana and her collaborators aim to address this critical diagnostic gap through AI-driven innovation. Their strategy centers on training machine learning models using extensive datasets comprising digitized pathology slides, tumor gene expression profiles, and associated clinical treatment outcomes. By assimilating these multidimensional data, the AI system aspires to classify individual tumors accurately, revealing their subtype composition and predicting therapeutic responsiveness before conventional methods would allow.

The technical challenge lies in designing algorithms that not only achieve high sensitivity—minimizing false negatives—but also maintain specificity to avoid false positives, thereby ensuring clinical reliability. The trained models must interpret complex histopathological features, integrate transcriptomic signatures, and correlate these with treatment trajectories to generate predictive insights. Such a sophisticated AI platform could revolutionize patient stratification, enabling clinicians to identify candidates for experimental therapies tailored to resistant subtypes and simultaneously avoid ineffective standard treatments.

If successful, the AI classifiers developed through this project will fill a significant unmet need in urologic oncology by enabling early, non-invasive detection of tumor heterogeneity and adaptive resistance mechanisms. This capability has profound implications for personalized medicine, allowing therapeutic interventions to be precisely targeted at subpopulations of cancer cells poised to evade current treatments. This would herald a shift from a one-size-fits-all approach toward highly customized clinical management in prostate cancer care.

Looking beyond algorithm development, Dr. Khurana’s team plans to advance the AI model toward clinical validation via prospective trials. This stage will test the practical application of their technology in real-world diagnostic workflows, evaluating performance in diverse patient cohorts to ensure robustness and reproducibility at scale. Successful translation from computational model to bedside tool could accelerate drug development by refining patient selection for clinical trials of novel therapeutics.

This research initiative exemplifies the convergence of cutting-edge computational science and clinical oncology, illustrating how data-driven AI tools can empower clinicians with unprecedented diagnostic precision. The integration of pathology imaging, genomics, and AI encapsulates a modern multidisciplinary approach crucial for overcoming the biological complexity inherent in prostate cancer progression and drug resistance.

In summary, Dr. Ekta Khurana’s groundbreaking project, supported by the Prostate Cancer Foundation’s substantial award, promises to transform the landscape of prostate cancer diagnosis and treatment. By harnessing artificial intelligence to detect treatment-resistant tumor subtypes early, this work could significantly improve clinical outcomes for thousands of men facing advanced prostate malignancies. The collaborative effort from leading institutions underscores a new era where computational innovation and biomedical expertise unite to tackle some of the most formidable challenges in cancer care.

Subject of Research: Prostate cancer; AI-based early detection of treatment-resistant tumor subtypes
Article Title: AI-Powered Early Detection of Treatment-Resistant Prostate Tumors: A Paradigm Shift in Cancer Care
News Publication Date: Not specified
Web References:

• Prostate Cancer Foundation
• Weill Cornell Medicine – Dr. Ekta Khurana
• Weill Cornell Medicine – Dr. Iman Hajirasouliha
• Memorial Sloan Kettering Research – Dr. Yu Chen
• Memorial Sloan Kettering – Dr. Anuradha Gopalan
  Image Credits: Weill Cornell Medicine
  Keywords: Prostate cancer, Prostate tumors, Artificial intelligence, Treatment resistance, Computational biomedicine, Machine learning, Pathology, Genomics, Clinical trials

Synovial sarcoma’s hallmark is a unique chromosomal translocation that fuses two genes, SS18 and SSX, generating the SS18::SSX fusion oncoprotein. This aberrant protein acts as a molecular driver of cancer, orchestrating epigenetic and transcriptional reprogramming that sustains the malignant identity and proliferative capacity of these cells. The exact mechanisms through which the fusion oncoprotein hijacks cellular processes have remained elusive, complicating efforts to develop targeted therapies.

Recently, a multidisciplinary group of researchers from Sanford Burnham Prebys Medical Discovery Institute, alongside collaborators at UCLA, UC San Diego, and the University of Edinburgh, published groundbreaking findings that illuminate a novel vulnerability in synovial sarcoma’s molecular armor. By integrating publicly available genomic datasets with their own experimental screenings in cell-based and animal models, this team identified the SUMO2 gene as a critical dependency selectively essential for synovial sarcoma cell growth.

SUMO2 encodes a small ubiquitin-like modifier protein that participates in post-translational modifications known as SUMOylation. This cellular process modulates protein function, localization, and interactions, thereby influencing epigenetic landscapes and gene expression patterns. Their data suggest that SS18::SSX fusion oncoprotein activates SUMO2, facilitating the cancer cells’ aberrant epigenetic programs and promoting sarcomagenesis.

To explore the therapeutic potential of targeting SUMO2, the researchers employed TAK-981, a small molecule inhibitor that impedes the SUMOylation pathway by blocking SUMO2 conjugation. Treatment with TAK-981 significantly impaired synovial sarcoma cell viability in vitro, accompanied by downregulation of gene networks under the control of the SS18::SSX fusion oncoprotein. The inhibitor not only disrupted the proliferation of cancerous cells but also lowered cellular levels of the fusion oncoprotein itself, underscoring a feedback mechanism that may enhance treatment efficacy.

Complementing cellular studies, in vivo experiments in mouse models demonstrated that SUMO2 inhibition curtailed tumor growth, reinforcing the notion that targeting this pathway can effectively attenuate sarcomagenesis. These findings also imply that TAK-981 might sensitize synovial sarcoma cells to standard chemotherapeutic regimens, suggesting a combinatorial strategy could yield synergistic effects in the clinical setting.

The significance of these results lies in bridging the gap between genomic data and actionable therapeutic interventions. By leveraging public cancer dependency maps and validating hits in biologically relevant models, the investigators exemplify the power of precision medicine approaches in uncovering cancer-specific vulnerabilities. Their work exemplifies how data-driven methodologies guide innovative drug discovery, particularly for rare cancers lacking effective targeted therapies.

Despite advancements, synovial sarcoma remains a formidable disease with roughly a 50-60% five-year survival rate for patients with metastatic progression. The ability of this malignancy to metastasize predominantly to the lungs, combined with the absence of tailored treatments, underscores the urgent need for new modalities. The discovery of SUMO2’s central role offers promise not only as a monotherapy target but as a gateway to understanding cancer epigenetics in fusion-driven sarcomas.

According to Dr. Rema Iyer, lead author and recent graduate from Sanford Burnham Prebys Graduate School of Biomedical Sciences, the complexity of synovial sarcoma’s epigenetic rewiring has hindered targeted drug development. The study’s insights into SUMO2 highlight a viable node for therapeutic intervention that had previously escaped attention because of the intricate interplay of oncoproteins and cellular epigenomic states.

Senior author Dr. Ani Deshpande, professor at Sanford Burnham Prebys and leader of the Cancer Genome and Epigenetics Program, emphasizes that SUMO2 inhibitors like TAK-981 carry strong potential for clinical translation. Given prior evidence of TAK-981’s efficacy in preclinical models of acute myeloid leukemia and pancreatic cancer, these findings strengthen the rationale for advancing this inhibitor into clinical trials for synovial sarcoma patients.

The methodology underpinning this research involved rigorous comparative screening across various platforms: analyses of DepMap’s expansive genomic datasets, cell culture model systems, and live animal experiments. This multi-layered approach allowed for a robust identification of genes essential to synovial sarcoma growth, out of which SUMO2 emerged as a consistent and druggable target.

While the immediate therapeutic implications center on SUMO2 inhibition, the broader impact resides in the conceptual framework that fusion oncoproteins like SS18::SSX impose epigenetic dependencies exploitable by precision drugs. Researchers worldwide now may consider SUMOylation pathways as fertile ground in the fight against other fusion-driven sarcomas and potentially beyond.

This study marks a critical advance in synovial sarcoma research, paving the way for targeted, mechanism-based therapies. It is a testament to the synergy between cutting-edge genomic technology and translational science, promising a future where even the rarest and most aggressive cancers can be tackled with tailored, effective interventions.

—

Subject of Research: Animals

Article Title: Targeting SUMO2 reverses aberrant epigenetic rewiring driven by SS18::SSX fusion oncoproteins and impairs sarcomagenesis

News Publication Date: 13-Aug-2025

Web References:

• The EMBO Journal article
• DepMap Consortium

References: DOI 10.1038/s44318-025-00526-w

Keywords: Cancer, Metastasis, Sarcoma, Oncoproteins