Cervical cancer segmentation based on medical images: a literature review.

Background and Objective: Cervical cancer clinical target volume (CTV) outlining and organs at risk segmentation are crucial steps in the diagnosis and treatment of cervical cancer. Manual segmentation is inefficient and subjective, leading to the development of automated or semi-automated methods. However, limitation of image quality, organ motion, and individual differences still pose significant challenges. Apart from numbers of studies on the medical images' segmentation, a comprehensive review within the field is lacking. The purpose of this paper is to comprehensively review the literatures on different types of medical image segmentation regarding cervical cancer and discuss the current level and challenges in segmentation process. Methods: As of May 31, 2023, we conducted a comprehensive literature search on Google Scholar, PubMed, and Web of Science using the following term combinations: "cervical cancer images", "segmentation", and "outline". The included studies focused on the segmentation of cervical cancer utilizing computed tomography (CT), magnetic resonance (MR), and positron emission tomography (PET) images, with screening for eligibility by two independent investigators. Key Content and Findings: This paper reviews representative papers on CTV and organs at risk segmentation in cervical cancer and classifies the methods into three categories based on image modalities. The traditional or deep learning methods are comprehensively described. The similarities and differences of related methods are analyzed, and their advantages and limitations are discussed in-depth. We have also included experimental results by using our private datasets to verify the performance of selected methods. The results indicate that the residual module and squeeze-and-excitation blocks module can significantly improve the performance of the model. Additionally, the segmentation method based on improved level set demonstrates better segmentation accuracy than other methods. Conclusions: The paper provides valuable insights into the current state-of-the-art in cervical cancer CTV outlining and organs at risk segmentation, highlighting areas for future research.

Segmentation of Clinical Target Volume From CT Images for Cervical Cancer Using Deep Learning.

Introduction: Segmentation of clinical target volume (CTV) from CT images is critical for cervical cancer brachytherapy, but this task is time-consuming, laborious, and not reproducible. In this work, we aim to propose an end-to-end model to segment CTV for cervical cancer brachytherapy accurately. Methods: In this paper, an improved M-Net model (Mnet_IM) is proposed to segment CTV of cervical cancer from CT images. An input and an output branch are both proposed to attach to the bottom layer to deal with CTV locating challenges due to its lower contrast than surrounding organs and tissues. A progressive fusion approach is then proposed to recover the prediction results layer by layer to enhance the smoothness of segmentation results. A loss function is defined on each of the multiscale outputs to form a deep supervision mechanism. Numbers of feature map channels that are directly connected to inputs are finally homogenized for each image resolution to reduce feature redundancy and computational burden. Result: Experimental results of the proposed model and some representative models on 5438 image slices from 53 cervical cancer patients demonstrate advantages of the proposed model in terms of segmentation accuracy, such as average surface distance, 95% Hausdorff distance, surface overlap, surface dice, and volumetric dice. Conclusion: A better agreement between the predicted CTV from the proposed model Mnet_IM and manually labeled ground truth is obtained compared to some representative state-of-the-art models.

GSEnet: feature extraction of gene expression data and its application to Leukemia classification.

Gene expression data is highly dimensional. As disease-related genes account for only a tiny fraction, a deep learning model, namely GSEnet, is proposed to extract instructive features from gene expression data. This model consists of three modules, namely the pre-conv module, the SE-Resnet module, and the SE-conv module. Effectiveness of the proposed model on the performance improvement of 9 representative classifiers is evaluated. Seven evaluation metrics are used for this assessment on the GSE99095 dataset. Robustness and advantages of the proposed model compared with representative feature selection methods are also discussed. Results show superiority of the proposed model on the improvement of the classification precision and accuracy.

Frangi based multi-scale level sets for retinal vascular segmentation.

Retinal vascular disease has always been the focus of medical attention. However, segmentation of the retinal vessels from fundus images is still an open problem due to intensity inhomogeneity in the image and thickness diversity of the retinal vessels. In this paper, we propose Frangi based multi-scale level sets to segment retinal vessels from fundus images. Vascular structures are first enhanced by the Frangi filter with local optimal scales being obtained at the same time. The enhanced image and local optimal scales are taken considered as inputs of the proposed level set models. Effectiveness of the proposed multi-scale level sets to their scale fixed versions has been evaluated using DRIVE and STARE image repositories. In addition, the proposed level set models have been tested on the DRIVE and STARE images. Experiments show that the proposed models produce segmentation accuracy at the same level with state-of-the-art methods.

The effect of vascular risk factor burden on the severity of COVID-19 illness, a retrospective cohort study.

BACKGROUND: Patients with cardiovascular comorbidities are at high risk of poor outcome from COVID-19. However, how the burden (number) of vascular risk factors influences the risk of severe COVID-19 disease remains unresolved. Our aim was to investigate the association of severe COVID-19 illness with vascular risk factor burden. METHODS: We included 164 (61.8 ± 13.6 years) patients with COVID-19 in this retrospective study. We compared the difference in clinical characteristics, laboratory findings and chest computed tomography (CT) findings between patients with severe and non-severe COVID-19 illness. We evaluated the association between the number of vascular risk factors and the development of severe COVID-19 disease, using a Cox regression model. RESULTS: Sixteen (9.8%) patients had no vascular risk factors; 38 (23.2%) had 1; 58 (35.4%) had 2; 34 (20.7%) had 3; and 18 (10.9%) had ≥4 risk factors. Twenty-nine patients (17.7%) experienced severe COVID-19 disease with a median (14 [7-27] days) duration between onset to developing severe COVID-19 disease, an event rate of 4.47 per 1000-patient days (95%CI 3.10-6.43). Kaplan-Meier curves showed a gradual increase in the risk of severe COVID-19 illness (log-rank P < 0.001) stratified by the number of vascular risk factors. After adjustment for age, sex, and comorbidities as potential confounders, vascular risk factor burden remained associated with an increasing risk of severe COVID-19 illness. CONCLUSIONS: Patients with increasing vascular risk factor burden have an increasing risk of severe COVID-19 disease, and this population might benefit from specific COVID-19 prevention (e.g., self-isolation) and early hospital treatment measures.