A statistical framework for radiation dose estimation with uncertainty quantification from the γ-H2AX assay

Abstract

Over the last decade, the γ–H2AX focus assay, which exploits the phosphorylation of the H2AX histone following DNA double–strand–breaks, has made considerable progress towards acceptance as a reliable biomarker for exposure to ionizing radiation. While the existing literature has convincingly demonstrated a dose–response effect, and also presented approaches to dose estimation based on appropriately defined calibration curves, a more widespread practical use is still hampered by a certain lack of discussion and agreement on the specific dose–response modelling and uncertainty quantification strategies, as well as by the unavailability of implementations. This manuscript intends to fill these gaps, by stating explicitly the statistical models and techniques required for calibration curve estimation and subsequent dose estimation. Accompanying this article, a web applet has been produced which implements the discussed methods.

The effect of data aggregation on dispersion estimates in count data models

“…From this one can deduce some sort of empirical dose-response relationship, which appears roughly linear over a considerable dose range, noting a saturation effect [17] for higher doses. It has been observed in the literature [12,18], and can also be hinted at from Figure 2, that overdispersion is present in H2AX foci data so that, for instance, quasi-Poisson or negative binomial models appear adequate. The quasi-Poisson model is essentially a Poisson model which allows for variance/mean ratios different from one.…”

“…Poisson regression models can be easily adapted to allow for situation (2.6), since the dispersion cancels out from the score equations and so the estimates of regression parameters are unaffected. One speaks then of quasi-Poisson regression models [19], which have gained interest specifically in the field of biodosimetry [12]. Under such a framework, standard errors can be conveniently computed in a post-hoc manner by multiplying the standard errors from the Poisson model with the square root of the (estimated) dispersion parameter [20].…”

“…The suitability of this histone as a biomarker for DSBs [4,5], and by extension, ionising radiation exposure [6][7][8][9], has long been established in the literature. However, statistical work to quantify this relationship and facilitate the actual dose estimation has only been carried out quite recently [10][11][12]. It should be noted that -H2AX foci data is not only used for biological dosimetry but much more prominently for several research questions in radiation biology [13][14][15][16].…”

The 2019 Bo Lindell Laureate Lecture: On the use of interdisciplinary, stakeholder-driven, radiation protection research in support of medical uses of ionising radiation

“…transplantation status) (Harbron et al, 2018a), together with improved dose assessment (Harbron et al, 2016) and overall risk of cancers (Journy et al, 2016(Journy et al, , 2017Harbron et al, 2017aHarbron et al, , 2018a for low-dose medically exposed populations. In the area of radiobiology, key recent outputs include development and validation of cytogenetic and genetic biomarkers of radiation exposure in medically exposed populations to underpin dose assessment (Cruz-Garcia et al, 2018;Einbeck et al, 2018;Moquet et al, 2018;O'Brien et al, 2018;Tichy et al, 2018), development of a new method of premature chromosome condensation to increase the speed of biological assessment of higher doses (Sun et al, 2019(Sun et al, , 2020, development of a new protocol for rapid gene-expression-based dose estimation (Polozov et al, 2019), and identification of further new genes suitable for biodosimetric purposes using rapid long-read DNA sequencing methods (Cruz-Garcia et al, 2020). Other major contributions include publication of peer-reviewed papers focused on the dose to the lens of the eye following CT scan exposures (Harbron et al, 2019), and the limited impact of iodinated contrast media on doses to haematopoietic stem cells (Harbron et al, 2017b(Harbron et al, , 2018b.…”

DeepFoci: Deep Learning-Based Algorithm for Fast Automatic Analysis of DNA Double Strand Break Ionizing Radiation-Induced Foci

“…Unless all data are analyzed by a single observer, which is practically impossible, the results may suffer from dramatic variations [28]. Hence, the results obtained by different observers and/or labs can only be compared with extreme caution [30,62,63]. This unsatisfactory situation means that, without suitable software, the evaluation of large image datasets, as generated for instance in the case of mass radiation accidents, remains unrealistic.…”