New Approaches to AI Methods for Screening Cardiomegaly on Chest Radiographs DOI Creative Commons
Patrycja S. Matusik, Zbisław Tabor, Iwona Kucybała

et al.

Applied Sciences, Journal Year: 2024, Volume and Issue: 14(24), P. 11605 - 11605

Published: Dec. 12, 2024

Background: Cardiothoracic ratio (CTR) and transverse cardiac diameter (TCD) are parameters that used to assess size on chest radiographs (CXRs). We aimed investigate the performance efficiency of artificial intelligence (AI) in screening for cardiomegaly CXRs. Methods: The U-net architecture was designed lung heart segmentation. CTR TCD were then calculated using these labels a mathematical algorithm. For training set, we retrospectively included 65 randomly selected patients who underwent CXRs, while testing chose 50 magnetic resonance (CMR) imaging had available CXRs medical documentation. Results: Using Dice coefficient 0.984 ± 0.003 (min. 0.977), it 0.983 0.004 0.972). 0.970 0.012 0.926), 0.950 0.021 0.871). mean measurements slightly greater when from either manual or automated segmentation than manually read. Receiver operating characteristic analyses showed both segmentation, read, good predictors diagnosed CMR. However, McNemar tests have shown diagnoses made with TCD, rather CTR, more consistent CMR diagnoses. According different definition based imaging, accuracy ranged 62.0 74.0% automatic (for 64.0 72.0%). Conclusion: use AI may optimize process Future studies should focus improving algorithms assessing usefulness cardiomegaly.

Language: Английский

Introduction to Artificial Intelligence for General Surgeons: A Narrative Review DOI Open Access

Blanche Lee,

Nikhil Narsey

Cureus, Journal Year: 2025, Volume and Issue: unknown

Published: March 1, 2025

Artificial intelligence (AI) has rapidly progressed in the last decade and will inevitably become incorporated into trauma surgical systems. In such settings, surgeons often need to make high-stakes, time-sensitive, complex decisions with limited or uncertain information. AI great potential augment pre-operative, intra-operative, post-operative phases of care. Despite expeditious advancement AI, many lack a foundational understanding terminology, its processes, applications clinical practice. This narrative review aims educate general about basics highlight thoraco-abdominal trauma, discuss implications incorporating use Australian health care system. found that studies have predominantly focused on machine learning deep applied diagnostics, risk prediction, decision-making. Other subfields include natural language processing computer vision. While tools care, current is limited. Future prospective, locally validated research required prior

Language: Английский

Citations

0
The use of artificial intelligence in blunt chest trauma DOI

Sagar Galwankar,

Łukasz Szarpak, Başar Cander

et al.

The American Journal of Emergency Medicine, Journal Year: 2024, Volume and Issue: 87, P. 157 - 158

Published: Sept. 3, 2024

Language: Английский

Citations

0
New Approaches to AI Methods for Screening Cardiomegaly on Chest Radiographs DOI Creative Commons
Patrycja S. Matusik, Zbisław Tabor, Iwona Kucybała

et al.

Applied Sciences, Journal Year: 2024, Volume and Issue: 14(24), P. 11605 - 11605

Published: Dec. 12, 2024

Background: Cardiothoracic ratio (CTR) and transverse cardiac diameter (TCD) are parameters that used to assess size on chest radiographs (CXRs). We aimed investigate the performance efficiency of artificial intelligence (AI) in screening for cardiomegaly CXRs. Methods: The U-net architecture was designed lung heart segmentation. CTR TCD were then calculated using these labels a mathematical algorithm. For training set, we retrospectively included 65 randomly selected patients who underwent CXRs, while testing chose 50 magnetic resonance (CMR) imaging had available CXRs medical documentation. Results: Using Dice coefficient 0.984 ± 0.003 (min. 0.977), it 0.983 0.004 0.972). 0.970 0.012 0.926), 0.950 0.021 0.871). mean measurements slightly greater when from either manual or automated segmentation than manually read. Receiver operating characteristic analyses showed both segmentation, read, good predictors diagnosed CMR. However, McNemar tests have shown diagnoses made with TCD, rather CTR, more consistent CMR diagnoses. According different definition based imaging, accuracy ranged 62.0 74.0% automatic (for 64.0 72.0%). Conclusion: use AI may optimize process Future studies should focus improving algorithms assessing usefulness cardiomegaly.

Language: Английский

Citations

0