CLINICIAN-DESIGNED MULTIMODAL AI
A clearer view of changing prostate cancer risk.
OncoVera is developing VioPRO, an investigational multimodal AI system designed to bring quantitative prostate biopsy pathology together with MRI, PSA and clinical variables, and longitudinal follow-up to estimate the risk of clinically significant adverse reclassification during active surveillance.
Investigational · In DevelopmentPATHOLOGY
MRI
PSA + CLINICAL
LONGITUDINAL
VioPRO
Individualized risk estimate + contributing signals + change over time
THE CLINICAL DECISION
Active surveillance depends on recognizing when risk changes.
For appropriately selected men with localized prostate cancer, active surveillance can delay or avoid unnecessary treatment while preserving the option for curative treatment if the disease shows clinically important change.
The challenge is that evolving risk is reflected across multiple sources of information: biopsy morphology, PSA and PSA density, MRI findings, repeat biopsy results, and the patient's longitudinal clinical history.
VioPRO is being developed to bring these signals together around one clinically defined question: when does a patient who initially appears suitable for surveillance warrant closer reassessment?
OUR FIRST PRODUCT
VioPRO
Multimodal risk stratification for prostate cancer active surveillance.
VioPRO is OncoVera's first investigational product. It is being developed to support clinicians managing patients with apparently low-risk localized prostate cancer by estimating the probability of clinically significant adverse reclassification/progression over a defined time horizon.
The system combines quantitative pathology, imaging, structured clinical information, and longitudinal change in a single modeling framework.
Discover VioPROOncoVera
VioPRO
PATHOLOGY
Quantitative tissue phenotype
MRI
Imaging context
PSA + CLINICAL
Clinical context
LONGITUDINAL
Risk changes over time
Investigational · In Development
Four perspectives on one evolving disease.
PATHOLOGY
Quantitative tissue phenotype
Whole-slide prostate biopsy images are used to quantify clinically meaningful morphology selected with GU pathology expertise. The pathology representation can combine expert-defined quantitative features, conventional pathology variables, and complementary learned representations when they add reproducible value.
MRI
Imaging context
Structured multiparametric MRI information can include PI-RADS assessment, lesion characteristics, prostate volume, suspicion of extraprostatic extension, and longitudinal imaging change where available.
PSA + CLINICAL
Clinical context
PSA, PSA density, age, stage, biopsy history, prostate volume, and other routinely available clinical factors selected with clinical collaborators.
LONGITUDINAL
Risk changes over time
Serial PSA, repeat imaging, repeat biopsy findings, and relevant management events allow the model to incorporate change rather than rely on a single baseline snapshot.
VioPRO
Individualized risk estimate + contributing signals + change over time
Built around clinical meaning, not modality alone.
EXPERT-GUIDED PATHOLOGY
Pathology expertise defines the signal.
The pathology layer is not limited to generic image embeddings. Clinically meaningful tissue phenotypes are selected with GU pathology expertise and quantified at scale using whole-slide imaging AI.
MULTIMODAL BY DESIGN
The clinical picture is larger than the slide.
Pathology is combined with imaging, PSA and clinical variables, and longitudinal follow-up rather than treated as the entire prediction problem.
LONGITUDINAL RISK
Designed to evolve with the patient.
The architecture is intended to update risk as new PSA, MRI, and biopsy information becomes available.
INTERPRETABLE DEVELOPMENT
Designed to show what contributes.
Structured clinical variables and expert-defined quantitative pathology provide an interpretable foundation, while learned representations can be evaluated for incremental predictive value rather than treated as an opaque replacement for domain expertise.
Built for rigorous validation from the beginning.
Development is structured so that the pathology and multimodal workstreams can progress in parallel. Once a reproducible pathology representation meets predefined stability criteria, it can be versioned for multimodal modeling while pathology development continues underneath it.
01
PATHOLOGY REPRESENTATION
Define and quantify clinically meaningful prostate biopsy morphology.
02
MULTIMODAL MODELING
Combine pathology with clinical, MRI, and longitudinal data and measure the incremental contribution of each source.
03
VALIDATION
Perform patient-level temporal validation, external validation, calibration analysis, and component-level comparison.
04
TRANSLATION
Evaluate clinical workflow integration, regulatory strategy, reimbursement and economic value, and commercialization pathways.
Building VioPRO requires multiple clinical perspectives.
OncoVera is developing VioPRO through collaboration spanning GU pathology, urology, AI and engineering, biomedical informatics, and commercialization, with planned expansion of the platform into additional prostate cancer decision points.
We welcome conversations with clinicians, researchers, health systems, translational partners, and investors interested in advancing multimodal prostate cancer risk stratification.
Collaborate With Us