We have developed and validated a practical approach to identifying the location on the skin surface that will receive a prespecified biopsy dose (ranging down to 1 cGy) in support of in vivo biological dosimetry in humans. This represents a significant technical challenge since the sites lie on the patient's surface outside the radiation fields. The PEREGRINE Monte Carlo simulation system was used to model radiation dose delivery, and TLDs were used for validation on phantoms and for confirmation during patient treatment. In the developmental studies, the Monte Carlo simulations consistently underestimated the dose at the biopsy site by approximately 15% (of the local dose) for a realistic treatment configuration, most likely due to lack of detail in the simulation of the linear accelerator outside the main beam line. Using a single, thickness-independent correction factor for the clinical calculations, the average of 36 measurements for the predicted 1-cGy point was 0.985 cGy (standard deviation: 0.110 cGy) despite patient breathing motion and other real-world challenges. Since the 10-cGy point is situated in the region of high-dose gradient at the edge of the field, patient motion had a greater effect, and the six measured points averaged 5.90 cGy (standard deviation: 1.01 cGy), a difference that is equivalent to approximately a 6-mm shift on the patient's surface.