Deep Machine Learning for Medical Diagnosis, Application to Lung Cancer Detection: A Review

BioMedInformatics, Journal Year: 2024, Volume and Issue: 4(1), P. 236 - 284

Published: Jan. 18, 2024

Deep learning has emerged as a powerful tool for medical image analysis and diagnosis, demonstrating high performance on tasks such cancer detection. This literature review synthesizes current research deep techniques applied to lung screening diagnosis. summarizes the state-of-the-art in detection, highlighting key advances, limitations, future directions. We prioritized studies utilizing major public datasets, LIDC, LUNA16, JSRT, provide comprehensive overview of field. focus architectures, including 2D 3D convolutional neural networks (CNNs), dual-path networks, Natural Language Processing (NLP) vision transformers (ViT). Across studies, models consistently outperformed traditional machine terms accuracy, sensitivity, specificity detection CT scans. is attributed ability automatically learn discriminative features from images model complex spatial relationships. However, several challenges remain be addressed before can widely deployed clinical practice. These include dependence training data, generalization across integration metadata, interpretability. Overall, demonstrates great potential precision medicine. more required rigorously validate address risks. provides insights both computer scientists clinicians, summarizing progress directions analysis.

Language: Английский

---

A Thorough Review of the Clinical Applications of Artificial Intelligence in Lung Cancer

Cancers, Journal Year: 2025, Volume and Issue: 17(5), P. 882 - 882

Published: March 4, 2025

According to data from the World Health Organization (WHO), lung cancer is becoming a global epidemic. It particularly high in list of leading causes death not only developed countries, but also worldwide; furthermore, it holds place terms cancer-related mortality. Nevertheless, many breakthroughs have been made last two decades regarding its management, with one most prominent being implementation artificial intelligence (AI) various aspects disease management. We included 473 papers this thorough review, which published during 5-10 years, order describe these breakthroughs. In screening programs, AI capable detecting suspicious nodules different imaging modalities-such as chest X-rays, computed tomography (CT), and positron emission (PET) scans-but discriminating between benign malignant well, success rates comparable or even better than those experienced radiologists. Furthermore, seems be able recognize biomarkers that appear patients who may develop cancer, years before event. Moreover, can assist pathologists cytologists recognizing type tumor, well specific histologic genetic markers play key role treating disease. Finally, treatment field, guide development personalized options for patients, possibly improving their prognosis.

Language: Английский

---

Interpretable Machine Learning Techniques

Published: March 7, 2025

---

Diagnostic performance of radiomics in prediction of Ki-67 index status in non-small cell lung cancer: A systematic review and meta-analysis

Journal of medical imaging and radiation sciences, Journal Year: 2024, Volume and Issue: 55(4), P. 101746 - 101746

Published: Sept. 13, 2024

Language: Английский

---

Integrating radiomics into predictive models for low nuclear grade DCIS using machine learning

Scientific Reports, Journal Year: 2025, Volume and Issue: 15(1)

Published: March 3, 2025

Abstract Predicting low nuclear grade DCIS before surgery can improve treatment choices and patient care, thereby reducing unnecessary treatment. Due to the high heterogeneity of limitations biopsies in fully characterizing tumors, current diagnostic methods relying on invasive face challenges. Here, we developed an ensemble machine learning model assist preoperative diagnosis DCIS. We integrated clinical data, ultrasound images, mammography Radiomic scores from 241 cases. The model, based Elastic Net, Generalized Linear Models with Boosting (glmboost), Ranger, improved ability predict preoperatively, achieving AUC 0.92 validation set, outperforming using data alone. comprehensive also demonstrated notable enhancements discrimination improvement net reclassification ( p < 0.001). Furthermore, effectively stratified patients by risk disease-free survival. Our findings emphasize importance integrating into prediction models, offering fresh perspectives for personalized management

Language: Английский

---

Diagnostic Performance of Radiomics in Prediction of Ki-67 Index Status in Non-small Cell Lung Cancer: A Systematic Review and Meta-Analysis

medRxiv (Cold Spring Harbor Laboratory), Journal Year: 2024, Volume and Issue: unknown

Published: Jan. 12, 2024

Abstract Background Lung cancer is a global health concern, in part due to its high prevalence and invasiveness. The Ki-67 index, indicating cellular proliferation, pivotal for assessing lung aggressiveness. Radiomics the inference of quantifiable data features from medical images through algorithms may offer insights into tumor behavior. Here, we perform systematic review meta-analysis assess performance radiomics predicting status Non-small Cell Cancer (NSCLC) on CT scan. Methods materials A comprehensive search current literature was conducted using relevant keywords PubMed/MEDLINE, Embase, Scopus, Web Science databases inception November 16, 2023. Original studies discussing CT-based NSCLC cohorts were included. quality assessment involved diagnostic accuracy (QUADAS-2) score (RQS). Quantitative meta-analysis, R, assessed pooled sensitivity specificity cohorts. Results We identified 10 that met inclusion criteria, involving 2279 participants, with 9 these included quantitative meta-analysis. overall moderate based QUADAS-2 RQS assessment. radiomics-based models training 0.78 (95% CI [0.73; 0.83]) 0.76 [0.70; 0.82]), respectively. validation 0.79 0.84]) 0.69 [0.61; 0.76]), Substantial heterogeneity noted (I 2 > 40%). It utilizing ITK-SNAP as segmentation software contributed significantly higher sensitivity. Conclusion This indicates promising NSCLC. study underscores radiomics’ potential personalized management, advocating prospective standardized methodologies larger samples.

Language: Английский

---

Interpretation of Diabetic Foot Ulcer Image Classification Using Layer Attribution Algorithms

Lecture notes in networks and systems, Journal Year: 2024, Volume and Issue: unknown, P. 13 - 22

Published: Jan. 1, 2024

Language: Английский

---