Mathematical methods in biological dosimetry: the 1996 Iranian accident

To report 18 months of cytogenetic follow-up for an Iranian worker accidentally overexposed to 192Ir, the mathematical extrapolation and comparison with clinical data.

Unstable chromosome aberrations were measured using conventional cytogenetic tests by French and Iranian biological dosimetry laboratories on five occasions after the exposure. The decrease in dicentrics over time was analysed mathematically. In addition, Dolphin and Qdr extrapolations were applied to the data to check the exposure estimates. FISH determination of translocation yields was performed twice by the French laboratory and the results compared with the Dolphin and Qdr corrected values.

Dose estimates based on dicentrics decreased from 3.1 +/- 0.4 Gy at 5 days after the accident to 0.8 +/- 0.2 Gy at 529 days. This could be fitted by double-exponential regression with an inflexion point between rapid and slow decrease of dicentrics after about 40 days. Dose estimates of 3.4 +/- 0.4 Gy for the Qdr model and 3.6 +/- 0.5 Gy for the Dolphin model were calculated during the post-exposure period and were remarkably stable. FISH translocation data at 26 and 61 days appeared consistent with the Dolphin and Qdr estimates.

Dose correction by the Qdr and Dolphin models and translocation scoring appeared consistent with the clinical data and provided better information about the radiation injury than did crude estimates from dicentric scoring alone. Estimation by the Dolphin model of the irradiated fraction of the body seemed unreliable: it correlated better with the fraction of originally irradiated lymphocytes.