Beyond early discovery, LLMs deliver significant value in clinical trials. They accelerate timelines by automating months of post-trial documentation work. More strategically, they can improve trial success rates by analyzing genomic data to identify patient populations with a higher likelihood of responding to a treatment.

To raise capital, biotechs need specific clinical data. Raj Devraj specifies the three essential components investors look for: 1) confirmation of good drug exposure in humans, 2) a favorable early safety profile, and 3) biomarker data that provides proof of the drug's biological mechanism. Lacking any of these makes fundraising significantly harder.

Trials like TaylorX and MINDACT use genomic scores to identify patients with early-stage, HR+/HER2- breast cancer who won't benefit from adjuvant chemotherapy. This avoids significant toxicity for two-thirds to over 80% of patients who would have received it under older guidelines, without compromising their outcomes.

Establishing a multi-disciplinary molecular tumour board helps operationalize biomarker strategies. This collaborative body, including oncologists and surgeons, not only interprets complex molecular data for trial matching but also collectively advocates for health insurance reimbursement for necessary tests, addressing a key practical barrier.

Unlike rare biomarkers that necessitate a 'test-and-wait' approach, IB6 is expressed in over 80-90% of NSCLC tumors. This ubiquity could make pre-screening unnecessary for drugs like Sigvotatug Vedotin, allowing clinicians to initiate targeted therapy much faster and for a broader patient population.

With over 5,000 oncology drugs in development and a 9-out-of-10 failure rate, the current model of running large, sequential clinical trials is not viable. New diagnostic platforms are essential to select drugs and patient populations more intelligently and much earlier in the process.

In rare diseases with small patient pools, recruiting for clinical trials is a major challenge. Effion Health's highly sensitive digital biomarkers can detect therapeutic efficacy with fewer participants, potentially reducing the required number of patients by 30%, which saves significant time and money for pharmaceutical companies.

The panel suggests AKT inhibitor trials in prostate cancer have been disappointing due to suboptimal biomarker selection (e.g., PTEN IHC). A similar drug in breast cancer showed significant survival benefit when using a more precise NGS-based strategy, indicating a potential path forward if the right patient population is identified genetically.

Fibrogen uses its PET imaging agent in Phase 2 not to pre-select patients, but to correlate target expression with treatment response. This data will allow them to enrich their Phase 3 trial with patients most likely to respond, significantly increasing the probability of success.