Clinical Trial Representation and Extrapolating Data to the Patient in Front of You

William asks the question many patients raise explicitly or implicitly: “Do these data really apply to me?” Dr. Wu answers honestly and in two parts.

First, upstream disparities matter. Comprehensive molecular diagnostics in advanced NSCLC are not uniformly delivered across racial groups in the United States, and Black patients can face structural barriers to receiving next-generation sequencing (NGS) in the first place. William's access to comprehensive NGS, which ultimately identified his HER2 (ERBB2) A775_G776insYVMA mutation via liquid biopsy, is itself a point of equity worth naming with the patient.

Second, the pivotal HER2-directed trial population is not fully representative of William. In the Beamion LUNG study, the racial distribution was approximately 53% Asian, 32% white, and 15% not reported. No Black patients were included in the published cohort. That gap is real and should not be minimized.

What Dr. Wu tells patients, and how he frames it as still useful for shared decision-making, is that extrapolation is a universal challenge in oncology, not a HER2-specific one. Clinical trial patients are rarely representative of the patient in front of you: An 85-year-old, for example, is almost never the trial population, yet the biology of the mutation and the mechanism of action of the drug are not expected to differ by race or age. In William's case, the mechanism of HER2 tyrosine kinase inhibition is not predicted to behave differently in a Black man than in the trial population. Dr. Wu pairs this honest acknowledgment with a commitment to monitor closely and adjust based on William's individual response and tolerability, the same approach he uses for any patient whose demographics fall outside the trial population.

In the next episode, “Real-World Gaps, Financial Toxicity, and the Evolving HER2 Treatment Landscape,” Dr. Wu closes with where community practice diverges from guidelines and what the next 1–2 years are likely to bring.