Instead of the traditional lab-to-clinic pipeline, a "reverse translation" approach uses AI to analyze data from patients who fail standard-of-care treatments. This identifies the specific unmet need and biological target first, guiding subsequent lab research for higher success rates.

To overcome the industry bottleneck of few validated solid tumor targets (15-20), Memo analyzes tumor-infiltrating B-cells from patients with superior outcomes. This approach aims to identify unique antibody-target pairs, unlocking new biological pathways for next-generation therapies like ADCs and CAR-Ts.

Unlike most trials that avoid patients who failed other therapies, Corvus intentionally included them, considering it a 'stacking deck against yourself'. This high-risk bet, based on their drug's unique mechanism, paid off by showing efficacy in a tough-to-treat population and demonstrating a lack of cross-resistance.

Contrary to the belief that AI needs massive datasets, Dr. Joseph Juraji's approach with NetraAI focuses on finding small, specific patient subpopulations within small trials. This allows the identification of a drug's 'superpower' without the need for big data, transforming trial economics.

By first targeting T-cell lymphoma, Corvus gathers crucial safety and biologic effect data in humans. This knowledge about the drug's impact on T-cells directly informs and de-risks subsequent trials in autoimmune diseases like atopic dermatitis, creating a capital-efficient development path.

The standard approach to reducing cancer drug toxicity is narrowing the target to specific mutations (e.g., HER2, KRAS). While this improves safety, it drastically shrinks the addressable patient population for each new therapy. This puts immense pressure on the pharmaceutical business model, where development costs average $2.5 billion per drug.

The company's core IP stems from a proprietary biobank of AML patient samples collected over 20 years at Oxford University. This historical dataset, containing samples from elite responders to stem cell transplants, is described as "very hard to replicate," creating a significant and durable competitive advantage in target discovery.

The company not only identifies targets from its elite patient cohort but also isolates the corresponding T-cell receptors (TCRs). Because these TCRs have been circulating safely in patients for years, they offer a strong starting point for safety. They are also naturally "highly selected," providing significant initial affinity for their targets, which can accelerate development.

Researchers analyzed the Oxford biobank samples in an "unbiased way," without preconceived notions of what targets they would find. This approach surprisingly revealed that all 22 initial tumor-specific targets were HLA Class II-restricted, a category previously overlooked in favor of HLA Class I. This highlights the power of agnostic discovery to challenge existing scientific dogma.