Purpose The fluorescent in situ hybridization (FISH) technique, which easily detects reciprocal translocations, is currently used to estimate doses in retrospective biological dosimetry, after suspected accidental overexposure to ionizing radiation (IR). This study of 42 cases aimed to verify the appropriateness of this assay for radiation dose reconstruction, compared to the dicentric assay, and to evaluate other limitations. Material and methods We labeled chromosomes 2, 4, and 12 by 3-color FISH painting to detect translocations on lymphocytes of patients with suspected past IR overexposure. Result Translocation dose estimation showed doses significantly different from 0 Gy in 25 of the 42 cases. The lowest positive dose measured was 0.3 Gy. Several months after IR exposure, the doses measured by translocation and dicentric assays are quite similar. For a year, dose estimation by translocation assay becomes more relevant as dicentric frequency starts to decrease, coming close to 0 for more than a year after the exposure. The persistence of translocations enabled us to corroborate an overexposure 44 years earlier. Interpretation of the observed translocation yield requires the knowledge of the patient’s other radiation exposures. A dose assessment by this biomarker is relevant only if the radiation exposure is confirmed. Conclusions This technique is appropriate for corroborating a former IR exposure of individuals. When the radiation dose is greater than 1 Gy, the translocations in complex exchanges must be considered. Another relevant point is the use of an appropriate background yield of translocations. The dose assessment, however, also depends on exposure to various genotoxic agents besides IR. If no evidence about the existence of radiation exposure is available, dose assessment is not useful. For this reason, report only the translocation frequency and its comparison with the background yield by age class is preferable. © 2018 Informa UK Limited, trading as Taylor and Francis Group.