PhaseV Trials

PhaseV is a Boston-based technology company that leverages advanced causal machine learning, generative AI, and an expansive data lake of real-world data (RWD) to empower biopharma sponsors and CROs with fast, accurate, and efficient clinical development decisions. Headquartered in Cambridge, MA, PhaseV serves 45+ biopharma and CRO customers worldwide, maintaining over 2 million patient-level records in its data lake and more than 50 disease predictive models, backed by 10+ scientific publications.

The PhaseV platform enables sponsors and CROs to rapidly design and implement adaptive and Bayesian clinical trials, monitor progress, and make real-time adjustments to optimize outcomes. It also supports analysis of heterogeneous treatment effects, patient stratification, early-derived endpoint identification, site selection, and indication expansion — all with the goal of bringing new treatments to more patients in a more precise and efficient way.

Core Products and Modules

• Trial Optimizer: Runs millions of simulations within minutes, benchmarks Bayesian designs, and enables real-time, data-driven trial decisions with AI-powered optimization. Used to redesign Phase 2 and Phase 3 studies, it has demonstrated sample size reductions of 10–15% while maintaining efficacy.
• Response Optimizer: Identifies optimal patient subgroups and biomarkers by detecting heterogeneity in complex biological signals and multi-variable clinical data. Applied in Type II Diabetes, it helped reinitiate a previously failed trial by identifying a responsive patient subset, with findings validated against multiple Phase 3 trials and subjected to rigorous sensitivity and false discovery analysis.
• Portfolio Optimizer: Uncovers key biological factors driving disease progression and assesses their impact on clinical outcomes using AI-driven causal graphs. Used in oncology to closely recapitulate patient-level variability observed in the Atezolizumab trial, and in metabolic disease to uncover and validate new indications for lifecycle management strategy.
• ClinOps Optimizer: Enhances clinical operations through Causal-ML driven site selection and real-time dashboards for trial and site performance analytics, enabling identification of top-performing recruitment sites while achieving geographic and demographic diversity.
• ML Early-Derived Endpoint Identification: Predicts and validates early-derived endpoints to accelerate decision-making. Proven case studies include neurology and immunology, with applications across various diseases. In a Systemic Lupus Erythematosus study, PhaseV demonstrated that BILAG data at week 8 could reliably predict improvement at week 24, enabling adaptive trial designs with significantly reduced sample sizes.

Proven Outcomes and Key Metrics

• 40% reduced enrollment in optimized trials
• 50% lower trial costs achieved through leaner designs
• 40% shorter trial duration through streamlined protocols
• 30%+ higher probability of success with AI-powered optimization tools
• These values are based on outcomes from real-world collaborations where PhaseV supported biopharma companies in optimizing trial design and operations.

Therapeutic Areas and Case Studies

• Type II Diabetes (Oramed): Response Optimizer identified a responsive patient subgroup, enabling reinitiation of a previously failed trial, with results validated by external Phase 3 data.
• Infectious Disease: Trial Optimizer was used to redesign a Phase 3 study, reducing sample size by 10–15% while meeting trial objectives.
• Rheumatic Disease: Phase 2 simulations guided a biotech client toward an optimal trial design through extensive multi-configuration analysis.
• Systemic Lupus Erythematosus: Combined ML Early-Derived Endpoint Identification and Trial Optimizer to demonstrate that an early endpoint at week 8 could predict 24-week outcomes, significantly improving trial efficiency.
• Oncology: Portfolio Optimizer modeled PD-L1 blocker patient-level variability consistent with the Atezolizumab trial (NCT02108652).
• Rheumatology / Osteoarthritis (Enlivex): Identified a higher-responder subgroup for the drug Allocetra in knee osteoarthritis patients.
• Immunology (EMD Serono): Response Optimizer identified a clinically meaningful subpopulation that would not otherwise have been detected.
• Kidney Disease: Portfolio Optimizer established data-driven connections between biomarkers and clinical outcomes, informing development strategy and future combination arm studies.

Technology and Team

• PhaseV combines expertise in machine learning, software engineering, statistics, data science, and clinical and pharmaceutical development to address complex optimization challenges in drug development.
• The platform is built on causal machine learning, generative AI, and cutting-edge software for clinical data analysis, trial design, and execution.
• Leadership includes co-founders Raviv Pryluk, PhD (CEO) and Elad Berkman, MSc (CTO), along with Dan Goldstaub, PhD (Scientific Co-Founder) and Brad Carlin, PhD (Sr. Director, Data Science and Statistics).
• A distinguished Scientific Advisory Board includes experts such as Sofia Villar, PhD; Jennifer Hill, PhD; David Perry, MD; Murray Urowitz, MD; Howard Trachtman, MD; Marcia Levenstein, ScD, MBE; and Jerald S. Schindler, DrPH.
• PhaseV has been recognized in industry media as a leading company applying AI to de-risk clinical trials, and recently launched an AI-powered Enrollment Lab to model clinical trial feasibility before protocol lock.

PhaseV is advancing paradigm shifts in clinical development by combining a rare multidisciplinary team with a powerful AI/ML platform, enabling biopharma sponsors and CROs to make smarter, faster, and more data-driven decisions across every phase of the clinical trial process.