Tag
A novel deep learning network trained on synchrotron imaging data automates accurate inner ear segmentation in clinical CT, surpassing expert performance and establishing a new standard for preoperative planning.
A deep learning model classifying usual interstitial pneumonia patterns on CT scans predicts fibrosis progression and mortality in individuals with interstitial lung abnormalities, offering a promising tool for risk stratification and clini
A novel deep learning model significantly improves diagnostic accuracy in distinguishing fibro-osseous jaw lesions (FD, COF, COD) from histology images, outperforming expert pathologists and offering a promising diagnostic support tool.
Baseline retinal vascular features quantified by automated deep learning on fundus photographs are associated with future visual field decline in primary open-angle glaucoma, offering potential prognostic biomarkers to enhance patient risk
A novel deep learning algorithm accurately detects moderate or greater aortic stenosis using AI-guided focused cardiac ultrasound acquired by novice operators, presenting a scalable solution to expand screening accessibility.
A validated deep learning model accurately detects surgical site infections from wound images, offering a promising telemedicine tool for early postoperative monitoring and timely clinical intervention.
This review synthesizes evidence on deep learning models for assessing technical skills and error detection in robotic-assisted vaginal cuff closure, emphasizing objective surgical skill evaluation, model accuracy, and implications for surg
Reflectance confocal microscopy features at week 4 post-dupilumab treatment, combined with deep learning analysis, enable non-invasive prediction of therapeutic response in facial atopic dermatitis patients.
This study presents a deep learning model that generates synthetic T-max perfusion maps from noncontrast DWI and FLAIR sequences, potentially streamlining stroke imaging and eliminating the need for contrast agents.
A novel AI model, SÉMIL, integrating vision-language and multiple instance learning, improves prognostic accuracy for stage II colorectal cancer by refining risk stratification beyond current clinical assessments, especially within NCCN hig
The MAIPI deep learning model offers a novel, accurate prognosis tool for mantle cell lymphoma, using standard H&E stained biopsy images, bypassing need for molecular tests or expert pathology, and enhancing risk stratification for indi
This review evaluates how deep learning-assisted delta radiomics on MRI, combined with AFP response, provides a non-invasive, high-accuracy prediction of pathological complete response in unresectable HCC undergoing conversion immunotherapy
A multicenter study in JAMA Ophthalmology introduces a deep learning model using domain-shift technology to accurately detect macular diseases across different OCT hardware vendors, achieving high negative predictive value and establishing
Researchers developed EEGSurvNet, a deep survival model that analyzes routine EEG to predict seizure risk over two years. Outperforming clinical models, it achieved an AUROC of 0.80 at two months and showed high efficacy in patients without
A deep learning-derived sarcopenia marker, the muscle-bone ratio (MBR), has been shown to predict the efficacy of anti-EGFR maintenance therapy in patients with RAS wild-type metastatic colorectal cancer, potentially identifying those who t
This critical synthesis of a 2020–2025 narrative review (327 DL studies) evaluates deep learning (DL) applications across diagnosis, prognosis, segmentation, and emerging intraoperative uses in otolaryngology and outlines practical steps fo
Deep learning models predicting RNFL thickness from optic disc photographs can identify ocular hypertension patients at higher risk for primary open-angle glaucoma, enhancing early risk assessment and monitoring disease progression.
