Dicentric chromosome assay using a deep learning-based automated system

Jeong, Soo Kyung; Kim, Song-Hyun; Jang, S. D.; Kang, Yeong-Rok; Kim, HyoJin; Kye, Yong-Uk; Lee, Seong Hun; Lee, Chang Geun; Park, Moon‐Taek; Kim, Joong Sun; Jeong, Min Ho; Jo, Wol Soon

doi:10.1038/s41598-022-25856-1

Cited by 9 publications

(12 citation statements)

References 46 publications

(60 reference statements)

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The distribution patterns for DC and MN were analyzed for compliance of Poisson distribution. Our ndings indicate that the distribution of DC detected by both, manual and automated scoring follows a Poisson distribution, as demonstrated in Tables 1 and 2, consistent with previous studies [7,11,42,43]. Moreover, in contrary to lower doses (0-3 Gy), the frequencies of DC observed with the automated scoring method were lower for higher doses (4 and 5 Gy) compared to those obtained with the manual scoring method.…”

Section: Discussionsupporting

confidence: 91%

“…A dose-dependent increase in the number of cells exhibiting more than one DC formations was observed, with the notable nding that no instances of cells with two or more DC formations were observed up to 0.5 Gy, consistent with manual scoring observations. Statistical analysis con rmed a Poisson distribution for DC occurrences across all doses, consistent with earlier reports [10,11,23,33,34]. The dispersion index (u) and relative variance (σ²/Y) were calculated for each dose points, following the guidelines provided by the IAEA (Table 2) [24].…”

Section: Establishment Of Dose Response Curve For 60 Co-γ-ray Induced...supporting

confidence: 66%

“…To address this issue, many researchers opt for manual con rmation of dicentric chromosomes (DC) identi ed by automated methods. These manual techniques concentrate on reevaluating and validating only the metaphases that were previously agged for dicentric chromosomes [7,11,42,43]. The time required to validate DC is minimal, especially considering the typically low frequency of DC.…”

Section: Discussionmentioning

confidence: 99%

“…Over the past six decades, signi cant efforts have been dedicated for standardizing and re ning the DC chromosome assay (DCA), making it a robust and widely accepted method in biological dosimetry [11]. DC, primarily induced by ionizing radiation, offer speci city and sensitivity, particularly in detecting chronic or low-dose radiation exposure [12].…”

Section: Introductionmentioning

confidence: 99%

See 3 more Smart Citations

Lab to Emergency: Establishment and validation of automated method for rapid biodosimetry

Vijayalakshmi,

Chaurasia,

Nair

et al. 2024

Preprint

View full text Add to dashboard Cite

In inadvertent radiation exposure scenarios, accurate measurement of absorbed dose by biodosimetric techniques is crucial, especially when physical dosimetry is unavailable or in dispute. Chromosomal aberrations like dicentric (DC) and micronuclei (MN) are widely used for biodosimetry, with the former being the gold standard for individual dose assessment. However, both methods are labour-intensive requiring skilled analysis. This study aims to assess and compare the frequency and accuracy of DC and MN scoring through manual and automated methods. It investigates the establishment and validation of dose-response curves, in the dose range of 0-5 Gy, for 60Co-γ-ray-induced DC and MN using peripheral blood samples from three human volunteers. Statistical analysis confirmed dose-dependent increases in aberration frequencies, with Poisson distribution validation for all dose points. The manual and automated scoring methods yielded notable differences in linear and quadratic coefficients. Correlation analyses demonstrated substantial agreement between manual and automated scoring methods, particularly for MN (R2= 0.98). Towards validation, 16 dose-blinded samples were analysed for both endpoints using manual and automated scoring methods. Results yielded a close match between estimated and delivered doses, particularly evident with automated scoring (of both DC and MN) displaying superior accuracy for most of the doses. Notably, the variation was minimal, with -1.96% for DC and 2.85% for MN at 0.5 and 4 Gy, respectively. Automated scoring was approximately 15-20 times faster than manual scoring, suggestive of a faster practical option, especially in emergency scenarios such as criticality accidents and radiological emergencies, thus facilitating speedy triage and medical decisions.

show abstract

Section: Discussionsupporting

confidence: 91%

Section: Establishment Of Dose Response Curve For 60 Co-γ-ray Induced...supporting

confidence: 66%

Section: Discussionmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

Lab to Emergency: Establishment and validation of automated method for rapid biodosimetry

Vijayalakshmi,

Chaurasia,

Nair

et al. 2024

Preprint

View full text Add to dashboard Cite

show abstract

“…In [16], the authors demonstrate that Faster R-CNN can be used to detect DCs in metaphases. This model is thoroughly evaluated in [17].…”

Section: Related Workmentioning

confidence: 99%

Ensembling Unets, sparse representation and low dimensional visualization for rare chromosomal aberration detection in light microscopy images

Deschemps,

Grégoire,

Martinez

et al. 2023

Preprint

View full text Add to dashboard Cite

In biological dosimetry, a radiation dose is estimated using the average number of chromosomal aberration per peripheral blood lymphocytes. To achieve an adequate precision in the estimation of this average, hundreds of cells must be analyzed in 2D microscopy images. Currently, this analysis is performed manually, as conventional computer vision techniques struggle with the wide variety of shapes showcased by chromosomes. The false discovery rate of current automated detection systems is high and variable, depending on small variations in data quality (chromosome spread, illumination variations ...), which makes using it in a fully automated fashion impossible. Automating chromosomal aberration is needed to reduce diagnosis time. Furthermore, an automated system can process more images, which improves confidence intervals around the estimated radiation dose. We build an object detection model to automate chromosomal aberration detection using recent advances in deep convolutional neural networks and statistical learning. We formulated the problem of rare aberration detection as a heatmap regression problem requiring the minimization of a sparsity-promoting loss to reduce the false alarm rate. Our Unet-based approach is analoguous to a one-stage object detector, and keeps the number of hyperparameters to a minimum. Finally, we demonstrate large performance improvements using an ensemble of checkpoints collected during a single run of training.A PCA-based strategy is used to provide cues for interpretation of our deep neural network-based model. The methodology is demonstrated on real, large, and challenging datasets depicting rare chromosomal aberrations and is favorably compared to a reference dosimetry technique.

show abstract

Adaptive Segmentation of DAPI-stained, C-banded, Aggregated and Overlapping Chromosomes

Platkov,

Gardos,

Gurevich

et al. 2024

Cell Biochem Biophys

View full text Add to dashboard Cite

Dicentric chromosome assay using a deep learning-based automated system

Cited by 9 publications

References 46 publications

Lab to Emergency: Establishment and validation of automated method for rapid biodosimetry

Lab to Emergency: Establishment and validation of automated method for rapid biodosimetry

Ensembling Unets, sparse representation and low dimensional visualization for rare chromosomal aberration detection in light microscopy images

Adaptive Segmentation of DAPI-stained, C-banded, Aggregated and Overlapping Chromosomes

Contact Info

Product

Resources

About