Cross-view discrepancy-dependency network for volumetric medical image segmentation

Medical Image Analysis, Journal Year: 2024, Volume and Issue: 99, P. 103329 - 103329

Published: Aug. 30, 2024

Language: Английский

---

Geometric and dosimetric evaluation of deep learning based auto‐segmentation for clinical target volume on breast cancer

Journal of Applied Clinical Medical Physics, Journal Year: 2023, Volume and Issue: 24(7)

Published: March 15, 2023

Recently, target auto-segmentation techniques based on deep learning (DL) have shown promising results. However, inaccurate delineation will directly affect the treatment planning dose distribution and effect of subsequent radiotherapy work. Evaluation geometric metrics alone may not be sufficient for accuracy assessment. The purpose this paper is to validate performance automatic segmentation with dosimetric try construct new evaluation comprehensively understand dose-response relationship from perspective clinical application.A DL-based model was developed by using 186 manual modified radical mastectomy breast cancer cases. resulting DL were used generate alternative contours in a set 48 patients. Auto-plan reoptimized ensure same optimized parameters as reference Manual-plan. To assess impact auto-segmentation, only common but also spatial distance relative volume ( RV${R}_V$ ) used. Correlations performed Spearman's correlation between changes.Only strong (|R2 | > 0.6, p < 0.01) or moderate 0.4, Pearson established traditional metric three indices (conformity index, homogeneity mean dose). For organs at risk (OARs), inferior no significant found differences. Furthermore, we that OARs affected boundary error instead target.Current could reflect certain degree variation. find contour variations do lead dosimetry changes, clinically oriented more accurately how quality affects should constructed.

Language: Английский

---

An MRI radiomics model for predicting a prostate-specific antigen response following abiraterone treatment in patients with metastatic castration-resistant prostate cancer

Frontiers in Oncology, Journal Year: 2025, Volume and Issue: 15

Published: Jan. 27, 2025

Objective To establish a combined radiomics-clinical model for the early prediction of prostate-specific antigen(PSA) response in patients with metastatic castration-resistant prostate cancer(mCRPC) after treatment abiraterone acetate(AA). Methods The data total 60 mCRPC from two hospitals were retrospectively analyzed and randomized into training group(n=48) or validation group(n=12). By extracting features biparametric MRI, including T2-weighted imaging(T2WI), diffusion-weighted imaging(DWI), apparent diffusion coefficient(ADC) maps, radiomics dataset selected using least absolute shrinkage selection operator(LASSO) regression. Four predictive models developed to assess efficacy treating mCRPC. primary outcome variable was PSA following AA treatment. performance each evaluated area under receiver operating characteristic curve(AUC). Univariate multivariate analyses performed Cox regression identify significant predictors Results integrated constructed seven extracted T2WI, DWI, ADC sequence images data. This demonstrated highest AUC both cohorts, values 0.889 (95% CI, 0.764-0.961) 0.875 0.564-0.991). Rad-score served as an independent predictor (HR: 2.21, 95% CI: 1.01-4.44). Conclusion MRI-based has potential predict Clinical relevance statement could be used noninvasively patients, which is helpful clinical decision-making.

Language: Английский

---

Diagnostic of fatty liver using radiomics and deep learning models on non-contrast abdominal CT

PLoS ONE, Journal Year: 2025, Volume and Issue: 20(2), P. e0310938 - e0310938

Published: Feb. 13, 2025

Purpose This study aims to explore the potential of non-contrast abdominal CT radiomics and deep learning models in accurately diagnosing fatty liver. Materials methods The retrospectively enrolled 840 individuals who underwent quantitative (QCT) examinations at First Affiliated Hospital Zhengzhou University from July 2022 May 2023. Subsequently, these participants were divided into a training set (n = 539) testing 301) 9:5 ratio. liver fat content measured by experienced radiologists using QCT technology served as reference standard. images scans then segmented regions interest (ROI) which features extracted. Two-dimensional (2D) three-dimensional (3D) models, well 2D 3D developed, machine based on clinical data constructed for four-category diagnosis characteristic curves each model plotted, area under receiver operating curve (AUC) calculated assess their efficacy classification Results A total included (mean age 49.1 years ± 11.5 [SD]; 581 males), whom 610 (73%) had Among patients with liver, there 302 mild (CT fraction 5%–14%), 155 moderate 14%–28%), 153 severe >28%). all used random forest algorithm achieved highest AUC (0.973), while Bagging decision tree showed sensitivity (0.873), specificity (0.939), accuracy (0.864), precision (0.880), F1 score (0.876). Conclusion systematic comparison was conducted performance comprehensive provides broader perspective determining optimal diagnosis. It found that algorithms show high consistency QCT-based radiologists.

Language: Английский

---

MRI quantified enlarged perivascular space volumes as imaging biomarkers correlating with severity of anxiety depression in young adults with long-time mobile phone use

Frontiers in Psychiatry, Journal Year: 2025, Volume and Issue: 16

Published: Feb. 20, 2025

Long-time mobile phone use (LTMPU) has been linked to emotional issues such as anxiety and depression while the enlarged perivascular spaces (EPVS), marker of neuroinflammation, is closely related with mental disorders. In current study, we aim develop a predictive model utilizing MRI-quantified EPVS metrics machine learning algorithms assess severity symptoms in patients LTMPU. Eighty-two participants LTMPU were included, 37 suffering from 44 depression. Deep used segment lesions extract quantitative metrics. Comparison correlation analyses performed investigate relationship between self-reported mood states. Training testing datasets randomly assigned ratio 8:2 perform radiomics analysis, where combined sex age select most valuable features construct models for predicting Several significantly different two comparisons. For classifying status, eight selected logistic regression model, an AUC 0.819 (95%CI 0.573-1.000) dataset. K nearest neighbors value 0.931 0.814-1.000) The utilization machine-learning presents promising method evaluating LTMPU, which might introduce non-invasive, objective, approach enhance diagnostic efficiency guide personalized treatment strategies.

Language: Английский

---