RE: The BOIN clinical trial design21 Jan 2026 21:09
AI probably quite good for this sort of thing:
‘ The starting dose in a Phase 1 drug trial is a critically calculated low level, derived from preclinical (animal) toxicity data (like the No Observed Adverse Effect Level - NOAEL), scaled to humans, and then reduced by a significant safety factor (often 10x) to ensure patient safety, aiming to avoid toxicity while being high enough for potential biological activity, balancing risk and efficiency. Regulatory guidance (FDA, EMA) provides frameworks, but the exact dose is unique, using methods like the "Maximum Recommended Starting Dose" (MRSD) calculation, often involving human equivalent dose (HED) conversion and safety margins.
Key Factors in Determining the Starting Dose:
Preclinical Data: The NOAEL from the most sensitive animal species is a primary input.
Species Scaling: The NOAEL is converted to a Human Equivalent Dose (HED) using specific scaling factors, often based on body surface area.
Safety Factor: A safety margin (e.g., 10x) is applied to the HED to get the MRSD, ensuring the initial dose is far below predicted toxicity.
Drug Type: Different rules apply to small molecules (often escalating by ~100% or half-log) versus biologics (often half-log or 3x increments).
Trial Goals: While safety is paramount, the dose shouldn't be so low as to be sub-therapeutic, minimizing unnecessary exposure to ineffective doses. ’