Enhancing Ovarian Tumor Diagnosis: Performance of Convolutional Neural Networks in Classifying Ovarian Masses Using Ultrasound Images

Journal of Clinical Medicine, Год журнала: 2024, Номер 13(14), С. 4123 - 4123

Опубликована: Июль 15, 2024

Background/Objectives: This study aims to create a strong binary classifier and evaluate the performance of pre-trained convolutional neural networks (CNNs) effectively distinguish between benign malignant ovarian tumors from still ultrasound images. Methods: The dataset consisted 3510 images 585 women with tumors, 390 195 malignant, that were classified by experts verified histopathology. A 20% to80% split for training validation was applied within k-fold cross-validation framework, ensuring comprehensive utilization dataset. final an aggregate three CNNs (VGG16, ResNet50, InceptionNet), experimentation focusing on aggregation weights decision threshold probability classification each mass. Results: model outperformed all individual models, achieving average sensitivity 96.5% specificity 88.1% compared subjective assessment’s (SA) 95.9% 93.9% specificity. All above results calculated at 0.2. Notably, misclassifications made similar those SA. Conclusions: AI-assisted image analysis can enhance diagnosis aid ultrasonographers less experience minimizing errors. Further research is needed fine-tune validate their in diverse clinical settings, potentially leading even higher overall accuracy.

Язык: Английский

---

Large language models in methodological quality evaluation of radiomics research based on METRICS: ChatGPT vs NotebookLM vs radiologist

European Journal of Radiology, Год журнала: 2025, Номер unknown, С. 111960 - 111960

Опубликована: Янв. 1, 2025

Язык: Английский

---

Clinical value of the nomogram model based on endoscopic ultrasonography radiomics and clinical indicators in identifying benign and malignant lesions of the pancreas

Frontiers in Oncology, Год журнала: 2025, Номер 15

Опубликована: Фев. 19, 2025

Based on endoscopic ultrasonography (EUS) radiomics and clinical data, we constructed a model nomogram for identifying benign malignant pancreatic lesions, explored the diagnostic performance of these two prediction models. Images data 151 patients with lesions detected by EUS from January 2018 to September 2023 were retrospectively collected. The randomly divided into training set validation at ratio 7:3. Through feature extraction screening images, calculated score (rad-score) realize construction model. Collecting laboratory test results, rad-scores patients, univariate multivariate logistic regression analyses used screen statistically significant influencing factors that could help identify pancreas, was constructed. utility models evaluated using receiver operating characteristic (ROC) curves, calibration decision curve analysis (DCA). screening, eight non-zero coefficient features finally selected calculate rad-score. Multivariate showed rad-score, age, CA199 in predicting lesions. A based three factors. In set, exhibited superior an AUC = 0.865 (95% CI 0.761-0.968) compared DCA depicted demonstrated accuracy yielded higher net benefit decision-making indicators, promising accurately

Язык: Английский

---

Evaluation of a novel ensemble model for preoperative ovarian cancer diagnosis: Clinical factors, O-RADS, and deep learning radiomics

Translational Oncology, Год журнала: 2025, Номер 54, С. 102335 - 102335

Опубликована: Март 5, 2025

Язык: Английский

---

Evolving and Novel Applications of Artificial Intelligence in Cancer Imaging

Cancers, Год журнала: 2025, Номер 17(9), С. 1510 - 1510

Опубликована: Апрель 30, 2025

Artificial intelligence (AI) is revolutionizing cancer imaging, enhancing screening, diagnosis, and treatment options for clinicians. AI-driven applications, particularly deep learning machine learning, excel in risk assessment, tumor detection, classification, predictive prognosis. Machine algorithms, especially frameworks, improve lesion characterization automated segmentation, leading to enhanced radiomic feature extraction delineation. Radiomics, which quantifies imaging features, offers personalized response predictions across various modalities. AI models also facilitate technological improvements non-diagnostic tasks, such as image optimization medical reporting. Despite advancements, challenges persist integrating into healthcare, tracking accurate data, ensuring patient privacy. Validation through clinician input multi-institutional studies essential safety model generalizability. This requires support from radiologists worldwide consideration of complex regulatory processes. Future directions include elaborating on existing optimizations, advanced techniques, improving patient-centric medicine, expanding healthcare accessibility. can enhance optimizing precision medicine outcomes. Ongoing multidisciplinary collaboration between radiologists, oncologists, software developers, bodies crucial AI's growing role clinical oncology. review aims provide an overview the applications oncologic while discussing their limitations.

Язык: Английский

---

Multimodal ultrasound-based radiomics and deep learning for differential diagnosis of O-RADS 4–5 adnexal masses

Cancer Imaging, Год журнала: 2025, Номер 25(1)

Опубликована: Май 23, 2025

Язык: Английский

---

O impacto da inteligência artificial na interpretação de exames de imagem em diagnóstico médico

Brazilian Journal of Health Review, Год журнала: 2024, Номер 7(3), С. e69808 - e69808

Опубликована: Май 20, 2024

Introdução: A inteligência artificial (IA) refere-se à capacidade de sistemas computacionais processarem extensas quantidades dados para simular comportamentos inteligentes, incluindo aprendizado autônomo. Objetivo: Investigar os impactos do uso da IA na interpretação exames imagem médicos. Metodologia: Revisão integrativa nas bases científicas: LILACS, SCIELO e PUBMED. Resultados Discussões: 17 artigos foram selecionados. está transformando o diagnóstico por imagem, melhorando a precisão, agilizando processamento dos fornecendo análises mais objetivas reprodutíveis. Essas tecnologias também estão contribuindo medicina permitindo diagnósticos tratamentos personalizados. Além disso, modificam configuração das estações trabalho radiologistas, que agora incluem um terceiro monitor assistidas IA. Conclusão: implementação em clínicas tem demonstrado potencial significativo personalizar otimizar tratamentos, além agilizar precisar diagnósticos, é crucial avanços uma centrada no paciente. Essa tecnologia promete transformações maneira como médicos pacientes abordam tratamento acompanhamento diversas doenças.

---

Exploratory study on the enhancement of O-RADS application effectiveness for novice ultrasonographers via deep learning

Archives of Gynecology and Obstetrics, Год журнала: 2024, Номер 310(6), С. 3111 - 3120

Опубликована: Ноя. 23, 2024

Язык: Английский

---

A Study on Automatic O-RADS Classification of Sonograms of Ovarian Adnexal Lesions Based on Deep Convolutional Neural Networks

Ultrasound in Medicine & Biology, Год журнала: 2024, Номер 51(2), С. 387 - 395

Опубликована: Ноя. 26, 2024

Язык: Английский

---