Automatic detection of micronuclei by cell microscopic image processing

An efficient cell micronucleus classification network based on multi-layer perception attention mechanism

Cellular micronucleus detection plays an important role in pathological toxicology detection and early cancer diagnosis. To address the challenges of tiny targets, high inter-class similarity, limited sample data and class imbalance in the field of cellular micronucleus image detection, this paper proposes a lightweight network called MobileViT-MN (Micronucleus), which integrates a multilayer perceptual attention mechanism. Considering that limited data and class imbalance may lead to overfitting of the model, we employ data augmentation to mitigate this problem. Additionally, based on domain adaptation, we innovatively introduce transfer learning. Furthermore, a novel Deep Separation-Decentralization module is designed to implement the reconstruction of the network, which employs attention mechanisms and an alternative strategy of deep separable convolution. Numerous ablation experiments are performed to validate the effectiveness of our method. The experimental results show that MobileViT-MN obtains outstanding performance on the augmented cellular micronucleus dataset. Avg_Acc reaches 0.933, F1 scores 0.971, and ROC scores 0.965. Compared with other classical algorithms, MobileViT-MN is more superior in classification performance.

Chromosomal Instability in Various Generations of Human Mesenchymal Stem Cells Following the Therapeutic Radiation

Background

Radiotherapy is a crucial treatment for most malignancies. However, it can cause several side effects, including the development of secondary malignancies due to radiation-induced genomic instability (RIGI). The aim of this study was to evaluate genomic instability in human mesenchymal stem cells (hMSCs) at different X-ray radiation doses. Additionally, the study aimed to examine the relative expression of certain genes involved in DNA repair, proto-oncogenes, and tumor suppressor genes.

Methods

After extracting, characterizing, and expanding hMSCs, they were exposed to X-ray beams at doses of 0, 0.5, 2, and 6 Gy. Nuclear alterations were evaluated through the cytokinesis-block micronucleus (CBMN) assay at 2, 10, and 15 days postirradiation. The expressions of BRCA1, BRCA2, TP53, Bax, Bcl2, and KRAS genes were analyzed 48 hr after irradiation to evaluate genomic responses to different radiation doses.

Results

The mean incidence of micronuclei, nucleoplasmic bridges, and nuclear buds was 4.8 ± 1.6, 47.6 ± 6, and 18 ± 2.6, respectively, in the nonirradiated group 48 hr after the fourth passage, per 1,000 binucleated cells. The incidence of micronuclei in groups exposed to 0.5, 2, and 6 Gy of radiation was 14.3 ± 4.9, 32.3 ± 6.5, and 55 ± 9.1, respectively, 48 hr after irradiation. The expression levels of the BRCA2, Bax, TP53, and KRAS genes significantly increased after exposure to 6 Gy radiation compared to the control groups. However, there was no significant increase in BRCA1 and Bcl2 gene expression in our study.

Conclusion

This study demonstrated significant nuclear alterations in the 10 days postirradiation due to the RIGIs that they inherited from their irradiated ancestral cells. While chromosomal instability is a prevalent event in malignant cells, so it seems necessary to optimize radiotherapy treatment protocols for tissues that contain stem cells, especially with IMRT, which delivers a low dose to a larger volume of tissues.

A deep learning workflow for quantification of micronuclei in DNA damage studies in cultured cancer cell lines: A proof of principle investigation

The cytokinesis block micronucleus assay is widely used for measuring/scoring/counting micronuclei, a marker of genome instability in cultured and primary cells. Though a gold standard method, this is a laborious and time-consuming process with person-to-person variation observed in quantification of micronuclei. We report in this study the utilisation of a new deep learning workflow for detection of micronuclei in DAPI stained nuclear images. The proposed deep learning framework achieved an average precision of >90% in detection of micronuclei. This proof of principle investigation in a DNA damage studies laboratory supports the idea of deploying AI powered tools in a cost-effective manner for repetitive and laborious tasks with relevant computational expertise. These systems will also help improving the quality of data and wellbeing of researchers.

Modest increase of KIF11 expression exposes fragilities in the mitotic spindle, causing chromosomal instability

Chromosomal instability (CIN), the process of increased chromosomal alterations, compromises genomic integrity and has profound consequences on human health. Yet, our understanding of the molecular and mechanistic basis of CIN initiation remains limited. We developed a high-throughput, single-cell, image-based pipeline employing deep-learning and spot-counting models to detect CIN by automatically counting chromosomes and micronuclei. To identify CIN-initiating conditions, we used CRISPR activation in human diploid cells to upregulate, at physiologically relevant levels, 14 genes that are functionally important in cancer. We found that upregulation of CCND1, FOXA1 and NEK2 resulted in pronounced changes in chromosome counts, and KIF11 upregulation resulted in micronuclei formation. We identified KIF11-dependent fragilities within the mitotic spindle; increased levels of KIF11 caused centrosome fragmentation, higher microtubule stability, lagging chromosomes or mitotic catastrophe. Our findings demonstrate that even modest changes in the average expression of single genes in a karyotypically stable background are sufficient for initiating CIN by exposing fragilities of the mitotic spindle, which can lead to a genomically diverse cell population.

Automatic recognition of micronucleus by combining attention mechanism and AlexNet

Background

Micronucleus (MN) is an abnormal fragment in a human cell caused by disorders in the mechanism regulating chromosome segregation. It can be used as a biomarker for genotoxicity, tumor risk, and tumor malignancy. The in vitro micronucleus assay is a commonly used method to detect micronucleus. However, it is time-consuming and the visual scoring can be inconsistent.

Methods

To alleviate this issue, we proposed a computer-aided diagnosis method combining convolutional neural networks and visual attention for micronucleus recognition. The backbone of our model is AlexNet without any dense layers and it is pretrained on the ImageNet dataset. Two attention modules are applied to extract cell image features and generate attention maps highlighting the region of interest to improve the interpretability of the network. Given the problems in the data set, we leverage data augmentation and focal loss to alleviate the impact.

Results

Experiments show that the proposed network yields better performance with fewer parameters. The AP value, F1 value and AUC value reach 0.932, 0.811 and 0.995, respectively.

Conclusion

In conclusion, the proposed network can effectively recognize micronucleus, and it can play an auxiliary role in clinical diagnosis by doctors.

Rapid and automatic detection of micronuclei in binucleated lymphocytes image

Cytokinesis block micronucleus (CBMN) assay is a widely used radiation biological dose estimation method. However, the subjectivity and the time-consuming nature of manual detection limits CBMN for rapid standard assay. The CBMN analysis is combined with a convolutional neural network to create a software for rapid standard automated detection of micronuclei in Giemsa stained binucleated lymphocytes images in this study. Cell acquisition, adhesive cell mass segmentation, cell type identification, and micronucleus counting are the four steps of the software's analysis workflow. Even when the cytoplasm is hazy, several micronuclei are joined to each other, or micronuclei are attached to the nucleus, this algorithm can swiftly and efficiently detect binucleated cells and micronuclei in a verification of 2000 images. In a test of 20 slides, the software reached a detection rate of 99.4% of manual detection in terms of binucleated cells, with a false positive rate of 14.7%. In terms of micronuclei detection, the software reached a detection rate of 115.1% of manual detection, with a 26.2% false positive rate. Each image analysis takes roughly 0.3 s, which is an order of magnitude faster than manual detection.

Radiation-Induced Chromosomal Aberrations and Immunotherapy: Micronuclei, Cytosolic DNA, and Interferon-Production Pathway

Radiation-induced chromosomal aberrations represent an early marker of late effects, including cell killing and transformation. The measurement of cytogenetic damage in tissues, generally in blood lymphocytes, from patients treated with radiotherapy has been studied for many years to predict individual sensitivity and late morbidity. Acentric fragments are lost during mitosis and create micronuclei (MN), which are well correlated to cell killing. Immunotherapy is rapidly becoming a most promising new strategy for metastatic tumors, and combination with radiotherapy is explored in several pre-clinical studies and clinical trials. Recent evidence has shown that the presence of cytosolic DNA activates immune response via the cyclic GMP-AMP synthase/stimulator of interferon genes pathway, which induces type I interferon transcription. Cytosolic DNA can be found after exposure to ionizing radiation either as MN or as small fragments leaking through nuclear envelope ruptures. The study of the dependence of cytosolic DNA and MN on dose and radiation quality can guide the optimal combination of radiotherapy and immunotherapy. The role of densely ionizing charged particles is under active investigation to define their impact on the activation of the interferon pathway.