Estimating partial body ionizing radiation exposure by automated cytogenetic biodosimetry

Abstract

PurposeInhomogeneous exposures to ionizing radiation can be detected and quantified with the Dicentric Chromosome Assay (DCA) of metaphase cells. Complete automation of interpretation of the DCA for whole body irradiation has significantly improved throughput without compromising accuracy, however low levels of residual false positive dicentric chromosomes (DCs) have confounded its application for partial body exposure determination.Materials and MethodsWe describe a method of estimating and correcting for false positive DCs in digitally processed images of metaphase cells. Nearly all DCs detected in unirradiated calibration samples are introduced by digital image processing. DC frequencies of irradiated calibration samples and those exposed to unknown radiation levels are corrected subtracting this false positive fraction from each. In partial body exposures, the fraction of cells exposed, and radiation dose can be quantified after applying this modification of the contaminated Poisson method.ResultsDose estimates of three partially irradiated samples diverged 0.2 to 2.5 Gy from physical doses and irradiated cell fractions deviated by 2.3-15.8% from the known levels. Synthetic partial body samples comprised of unirradiated and 3 Gy samples from 4 laboratories were correctly discriminated as inhomogeneous by multiple criteria. Root mean squared errors of these dose estimates ranged from 0.52 to 1.14 Gy2 and from 8.1 to 33.3%2 for the fraction of cells irradiated.ConclusionsAutomated DCA can differentiate whole-from partial-body radiation exposures and provides timely quantification of estimated whole-body equivalent dose.Biographical NoteBen Shirley M.Sc. is Chief Software Architect, CytoGnomix Inc. Canada; Joan Knoll Ph.D. Dipl.ABMGG, FCCMG is Professor in Pathology and Laboratory Medicine, Schulich School of Medicine and Dentistry, University of Western Ontario, Canada and cofounder, CytoGnomix Inc.; Jayne Moquet Ph.D. is Principal Radiation Protection Scientist in the Cytogenetics Group, Public Health England; Elizabeth Ainsbury Ph.D. is Head, Cytogenetics Group and the Chromosome Dosimetry Service, Public Health England; Pham Ngoc Duy M.Sc. is deputy director of Biotechnology Center, Dalat Nuclear Research Institute, Vietnam; Farrah Norton M.Sc.is Research Scientist and Lead of the Biodosimetry emergency response and research capability at Canadian Nuclear Laboratories; Ruth Wilkins, Ph.D. is Research Scientist and Chief of the Ionizing Radiation Health Sciences Division at Health Canada, Ontario, Canada; and Peter K. Rogan Ph.D. is Professor of Biochemistry and Oncology, Schulich School of Medicine and Dentistry, University of Western Ontario, Canada, and President, CytoGnomix Inc.

Radiation Exposure Determination in a Secure, Cloud-Based Online Environment

“…Dose estimates for homogeneously irradiated HC test samples (HS01, HS04, HS05, HS07, HS08, HS10) were generated after application of 4 different image selection models (A_B, A_C, A_D, Automated178981) and matched previously published results (7). Estimates of partial-body dose and fraction of blood irradiated agreed with previously published results after application of the C_B750 image selection model (8). Note that slight variation in results are expected in every partial-body analysis in ADCI by a random sampling algorithm that removes false positive DCs in partial-body irradiated samples (8).…”

“…A set of calibration (10 samples, 50,497 images) and test (6 samples, 4,098 images) samples obtained from Health Canada (HC) were uploaded to cloud storage and processed with ADCI_Online (7). Additionally, a set of calibration and test samples obtained from Public Health England (PHE) were also uploaded (8). Calibration curves were generated, and whole-body exposures of HC samples were determined using the Dose Estimation Wizard for image section models A_B, A_C, A_D, and Automated178981 (7).…”

“…DC counts associated with the machine learning tuning parameter σ (a measure of sensitivity and specificity) are stored in an "adcisample" file generated during sample processing (18). England (PHE) were also uploaded (8). Calibration curves were generated, and whole-body exposures of HC samples were determined using the Dose Estimation Wizard for image section models A_B, A_C, A_D, and Automated178981 (7).…”

“…Calibration curves were generated, and whole-body exposures of HC samples were determined using the Dose Estimation Wizard for image section models A_B, A_C, A_D, and Automated178981 (7). Partial-body exposure levels were estimated for PHE samples, with each receiving the equivalent of 50% fractional exposures based on an associated dose calibration curve generated with image selection model C_B750 (8).…”

“…ADCI first utilizes novel image processing segmentation techniques to locate and characterize features of chromosomes (3,4), followed by tandem machine learning to determine whether each chromosome is monocentric or dicentric (5,6). ADCI autonomously determines image selection criteria used to select optimal images for interpretation (3,4), constructs dose calibration curves, and estimates whole or partial-body radiation dose in test samples with unknown exposures (7,8). The estimated time for semi-automated DC review was compared to unattended processing by ADCI in a scenario involving 1,000 potentially exposed samples (2).…”