Dose-response of X-ray induced dicentrics determined by fully- & semi-automated DCScore evaluation after 3 h or 24 h colcemid treatment.

Purpose: Optimization of automated dicentric evaluation in the BIR laboratory is necessary to improve and accelerate individual biological dosimetry in radiation accident scenarios. Therefore, two different DCScore classifiers were analyzed for their suitability for use with laboratory-specific protocols, including two different lymphocyte culture conditions, 3-hour or 24-hour colcemid treatment.

Materials and methods: Dicentric formation was compared in 3 h and 24 h colcemid-treated cultures by fully- and semi-automated dicentric scoring using two different classifiers. Various calibration curves were constructed and absorbed doses of blinded X-irradiated blood samples were estimated after 24 h of colcemid treatment using both classifiers and scoring modes.

Results: 24 h colcemid treatment results in twice as many metaphases as 3 h colcemid treatment and the courses of dicentric frequencies after short- and long-term colcemid treatment differ, especially > 1 Gy. The "short-term colcemid classifier" detects more dicentric candidates and true positive dicentrics, respectively, especially > 2 Gy than the "long-term classifier" on the same slides.

Conclusion: Neither classifier was significantly better suited for the lab-specific MP preparations with regard to triage dose estimates for blinded samples by fully- as well as semi-automated analysis. For accurate dose assessment, it is recommended to adapt an available classifier to laboratory-specific conditions and protocols to optimize the identification of true dicentrics by DCScore.