Mount Sinai Becomes First NYC Comprehensive Cancer Center with AI Trial Matching
Mount Sinai Tisch Cancer Center announced on January 8, 2026, the launch of PRISM, an innovative artificial intelligence platform designed to connect cancer patients across the Mount Sinai Health System with potentially life-saving clinical trials. This deployment positions Mount Sinai as the first National Cancer Institute-designated Comprehensive Cancer Center in New York City to implement an oncology-specific AI tool for systemwide clinical trial matching.
The PRISM platform, developed by AI company Triomics, represents a major advancement in expanding access to innovative cancer research, accelerating clinical trial enrollment, and ensuring patients across the entire health system are considered for cutting-edge treatment options earlier in their care journey. The initiative addresses longstanding challenges in clinical trial participation that have historically concentrated research opportunities at flagship academic medical centers while leaving patients at community hospitals underserved.
By automating the complex, time-intensive process of matching patients to appropriate clinical trials, PRISM fundamentally transforms how oncologists identify research opportunities for their patients. The technology analyzes electronic health records, medical histories, diagnoses, and clinical characteristics to determine trial eligibility in real-time, replacing manual chart reviews that previously consumed significant clinician time while frequently missing eligible candidates.
OncoLLM: Purpose-Built Large Language Model for Oncology
The PRISM platform is powered by OncoLLM, a large language model-based pipeline specifically designed for cancer care rather than adapted from general-purpose AI systems. This oncology-specific architecture represents a critical distinction from generic language models that have raised concerns about accuracy and reliability in high-stakes clinical settings.
Generic large language models frequently generate “hallucinations”—plausible-sounding but factually incorrect information—when applied to specialized medical domains. In cancer care, where treatment decisions directly impact patient survival and quality of life, such inaccuracies carry unacceptable risks. OncoLLM addresses these concerns through training specifically focused on oncology knowledge, terminology, treatment protocols, and clinical reasoning patterns.
The domain-specific approach enables OncoLLM to understand nuanced aspects of cancer care including specific biomarkers, complex treatment histories, disease staging systems, molecular subtyping, and eligibility criteria for clinical trials. This specialized knowledge base allows the system to identify appropriate trial matches with greater accuracy and reliability than general-purpose AI tools applied to oncology applications.
Industry observers note that PRISM represents a shift toward rigorous, specialized AI governance in healthcare settings. Approximately 12% of hospitals have moved toward implementing domain-specific AI systems with enhanced oversight to ensure patient safety and data accuracy, reflecting growing recognition that general-purpose AI tools may be insufficient for specialized medical applications.
Comprehensive Systemwide Implementation Strategy
Mount Sinai’s strategic decision to deploy PRISM across its entire health system addresses a fundamental inequity in clinical trial access. Historically, trial participation has been concentrated at flagship academic hospitals with dedicated research infrastructure, while patients receiving care at community hospitals faced significant barriers to trial enrollment even when potentially eligible.
With systemwide PRISM deployment, patients seen at Mount Sinai Queens, Mount Sinai Brooklyn, Mount Sinai South Nassau, Mount Sinai Morningside, Mount Sinai West, and other network facilities now have identical access to available trials as those treated at The Mount Sinai Hospital. This geographic equity directly advances Mount Sinai’s mission as an NCI-designated Comprehensive Cancer Center.
The initiative strengthens systemwide access to clinical trials by creating unified patient identification processes across all facilities, accelerating research integration across diverse care settings ranging from urban academic centers to community hospitals, supporting improved outcomes for the diverse populations Mount Sinai serves across New York City and surrounding regions, and expanding the pool of patients who can be considered for research at the right moment in their treatment journey.
By democratizing trial access across its network, Mount Sinai positions itself to meet and potentially exceed national benchmarks for clinical trial participation rates. Higher participation benefits both individual patients who gain access to novel therapies and the broader cancer research enterprise by accelerating enrollment timelines for important studies.
Overcoming Traditional Clinical Trial Participation Barriers
Dr. Karyn Goodman, Professor and Vice Chair of Clinical Research in the Department of Radiation Oncology at the Icahn School of Medicine at Mount Sinai and Associate Director of Clinical Research at Tisch Cancer Center, articulated the core problem PRISM addresses: “Clinical trials are essential to advancing cancer care, but too often patients and their treating physicians are not aware of studies that may be appropriate for them.”
This awareness gap creates a paradoxical situation where potentially eligible patients miss trial opportunities despite multiple ongoing studies that could benefit them. Busy oncologists treating high patient volumes often lack time to manually review dozens or hundreds of active trial protocols, cross-referencing complex eligibility criteria against individual patient characteristics.
Traditional barriers to clinical trial participation include limited clinician awareness of available trials due to information overload, time-intensive manual chart review processes consuming hours per patient, fragmented medical records dispersed across multiple systems requiring extensive data gathering, complex eligibility criteria spanning dozens of parameters difficult to assess manually, and geographic concentration of trial opportunities at select academic centers.
These barriers disproportionately affect patients from underserved communities who may receive care at facilities with limited research infrastructure. By automating trial identification, PRISM helps level the playing field, ensuring that eligibility depends on medical appropriateness rather than geographic location or provider familiarity with specific studies.
Automated Real-Time Electronic Health Record Analysis
The clinical trial matching process involves determining which research studies a patient may be eligible for based on comprehensive review of their medical information. PRISM automates this traditionally manual process by analyzing electronic health records in real-time as patients move through the healthcare system.
The platform evaluates multiple data dimensions simultaneously including cancer type and subtype based on pathology reports, disease stage and extent of spread, molecular characteristics and biomarker profiles, prior treatment history and responses, current performance status and comorbidities, and demographic characteristics when relevant to specific trials.
This automated analysis happens continuously as new clinical information becomes available, enabling the system to identify trial opportunities at optimal moments during a patient’s treatment journey. A patient might become eligible for a specific trial following disease progression on first-line therapy, after specific biomarker testing results become available, or when entering maintenance phases following initial treatment.
Dr. Goodman emphasized this benefit: “By deploying an AI platform trained specifically for oncology, we can identify trial opportunities earlier, more consistently, and more equitably, allowing clinicians to focus on meaningful conversations with patients rather than manual chart review.”
The shift from manual chart review to automated matching frees oncologists to invest time in substantive discussions with patients about trial participation rather than administrative tasks. This enables more thorough informed consent processes, better alignment between patient preferences and research opportunities, and ultimately higher quality decisions about trial enrollment.
Advancing Research Access Equity
Mount Sinai’s PRISM deployment directly addresses health equity concerns in clinical research participation. Numerous studies have documented disparities in clinical trial enrollment, with patients from minority communities, lower socioeconomic backgrounds, and rural areas significantly underrepresented in cancer research despite often experiencing higher disease burden.
Multiple factors contribute to these disparities, including geographic barriers when trials concentrate at distant academic centers, systemic mistrust of medical research in communities with histories of research exploitation, language and literacy barriers affecting informed consent processes, socioeconomic constraints limiting travel to trial sites, and clinician bias in which patients are offered trial participation.
By extending trial matching capabilities across Mount Sinai’s diverse network of hospitals serving varied communities throughout New York City and surrounding areas, PRISM addresses several equity barriers simultaneously. Patients no longer need to travel to Manhattan academic centers to access trial opportunities, community hospital oncologists receive the same trial identification support as academic specialists, and systematic automated screening reduces potential for unconscious bias in patient selection.
The diverse communities Mount Sinai serves include populations with high cancer incidence and mortality rates who have been historically underrepresented in research. Improving their trial access not only benefits individual patients but also enhances the generalizability of research findings by including more representative participant populations.
Technical Superiority of Domain-Specific AI Architecture
The oncology-specific architecture of OncoLLM provides several technical advantages over generic large language models applied to clinical trial matching. General-purpose language models, while impressive in broad knowledge domains, often struggle with the precise technical requirements of medical applications.
Generic language models trained on broad internet corpora may misunderstand medical terminology, incorrectly interpret complex eligibility criteria, fail to recognize nuanced clinical scenarios, and generate confident-sounding but medically inaccurate recommendations. In 2026, these “AI hallucination” concerns have led many healthcare systems to avoid deploying general-purpose models for high-stakes clinical decisions.
OncoLLM’s oncology-specific training addresses these concerns through focused learning on cancer-specific knowledge including FDA-approved oncology drugs and their mechanisms, standard treatment protocols for various cancer types, biomarker testing methodologies and interpretation, disease staging systems and prognostic factors, and clinical trial design principles and eligibility criteria patterns.
This specialized knowledge enables OncoLLM to reason about cancer care with greater accuracy and nuance than general language models. The system understands, for example, that a patient with HER2-positive breast cancer who has progressed on trastuzumab-based therapy might be appropriate for trials of novel HER2-targeted agents but not trials requiring HER2-negative disease.
Transformative Impact on Clinical Workflow Integration
Traditional clinical trial matching relied on labor-intensive manual processes that created significant bottlenecks in research enrollment. Research coordinators or physicians needed to regularly review active trial protocols, manually search patient databases for potentially eligible candidates, conduct detailed chart reviews extracting relevant clinical information, and cross-reference patient characteristics against complex eligibility criteria.
This process could require hours per potential candidate, with no guarantee that reviewed patients would ultimately prove eligible or interested in participation. The time investment often limited trial consideration to highly motivated patients who specifically requested information about research options, missing many eligible candidates who might have benefited from participation.
PRISM fundamentally transforms this workflow by automating the resource-intensive screening phase. The platform continuously monitors electronic health records, identifying potentially eligible patients as soon as relevant clinical data becomes available. Research teams receive prioritized lists of candidates most likely to meet specific trial criteria, enabling them to focus outreach efforts efficiently.
Oncologists receive alerts when their patients match specific trials, allowing timely discussions about research participation integrated naturally into treatment planning conversations. This integration ensures trial consideration happens at clinically appropriate moments rather than as an afterthought following standard treatment decisions.
Strategic Collaboration Model Between Industry and Academia
The successful PRISM deployment reflects close collaboration among Triomics (the AI platform developer), Mount Sinai Tisch Cancer Center (the clinical implementation site), and Mount Sinai Research IT (providing technical infrastructure and integration support). This multi-stakeholder partnership model combines technical innovation, clinical expertise, and operational infrastructure essential for deploying complex AI systems in healthcare settings.
Triomics brought specialized AI development capabilities focused on oncology applications, Mount Sinai provided clinical domain expertise ensuring the platform addresses real practitioner needs, and Research IT contributed healthcare IT infrastructure knowledge enabling smooth integration with existing electronic health record systems.
The partnership demonstrates that successful healthcare AI deployment requires more than just sophisticated algorithms. Clinical workflow understanding, change management strategies, technical infrastructure compatibility, data governance frameworks, and ongoing support mechanisms all prove critical for translating promising AI capabilities into sustained operational value.
Mount Sinai’s experience as an early adopter of PRISM will provide valuable insights for other cancer centers considering similar deployments. The health system plans to evaluate outcomes from this implementation, assessing metrics such as trial enrollment rates, time from patient identification to enrollment, geographic distribution of enrolled patients, and clinician satisfaction with the platform.
Research Publication and Knowledge Dissemination Plans
Mount Sinai has committed to sharing findings from the PRISM deployment through peer-reviewed publications and presentations at national scientific meetings. This knowledge dissemination will benefit the broader cancer research community by providing evidence about AI-assisted clinical trial matching effectiveness.
Key research questions the evaluation may address include quantitative improvements in trial enrollment rates following PRISM implementation, impact on time intervals from patient eligibility to trial enrollment, changes in demographic and geographic diversity of trial participants, clinician time savings from automated versus manual matching processes, and patient satisfaction with trial identification and enrollment processes.
By rigorously evaluating and publishing implementation outcomes, Mount Sinai contributes to the evidence base guiding other institutions considering similar AI deployments. The commitment to transparent outcome reporting reflects best practices in healthcare innovation, ensuring that promising technologies undergo objective assessment rather than relying solely on vendor claims or anecdotal experiences.
