Key result
Automated cardiothoracic ratio calculation using a neural network model was more accurate than manual calculation by nurse practitioners (mean difference from experts 0.83% vs 1.52%; p<0.001).
Why the study?
Assessing cardiomegaly via cardiothoracic ratio monitoring from chest X-rays is subjective and varies between operators.
Does an automated neural network model improve the accuracy and speed of cardiothoracic ratio calculation from chest X-rays in hemodialysis patients compared to manual calculation?
Cross-Sectional
No
Does an automated neural network model improve the accuracy and speed of cardiothoracic ratio calculation from chest X-rays in hemodialysis patients compared to manual calculation?
Absolute Event Rate: 0.83% vs 1.52%
p-value: p=< 0.001
An automated neural network model accurately and rapidly calculates the cardiothoracic ratio from chest X-rays in hemodialysis patients, outperforming manual assessment.
May support automated CTR assessment in hemodialysis patients; leaves open prospective validation for clinical workflows or outcomes.
Cardiomegaly is associated with poor clinical outcomes and is assessed by routine monitoring of the cardiothoracic ratio (CTR) from chest X-rays (CXRs). Judgment of the margins of the heart and lungs is subjective and may vary between different operators. Methods: Patients aged > 19 years in our hemodialysis unit from March 2021 to October 2021 were enrolled. The borders of the lungs and heart on CXRs were labeled by two nephrologists as the ground truth (nephrologist-defined mask). We implemented AlbuNet-34, a U-Net variant, to predict the heart and lung margins from CXR images and to automatically calculate the CTRs. Results: The coefficient of determination (R2) obtained using the neural network model was 0.96, compared with an R2 of 0.90 obtained by nurse practitioners. The mean difference between the CTRs calculated by the nurse practitioners and senior nephrologists was 1.52 ± 1.46%, and that between the neural network model and the nephrologists was 0.83 ± 0.87% (p < 0.001). The mean CTR calculation duration was 85 s using the manual method and less than 2 s using the automated method (p < 0.001). Conclusions: Our study confirmed the validity of automated CTR calculations. By achieving high accuracy and saving time, our model can be implemented in clinical practice.
No takes yet. Share an insight, caveat, or question.
Chou et al. (2023) conducted a cross-sectional in Hemodialysis. Automated cardiothoracic ratio calculation (AlbuNet-34) vs. Manual calculation by nurse practitioners was evaluated on Mean difference in cardiothoracic ratio compared to nephrologist-defined ground truth (p=< 0.001). Automated cardiothoracic ratio calculation using a neural network model was more accurate than manual calculation by nurse practitioners (mean difference from experts 0.83% vs 1.52%; p<0.001).
Synapse has enriched 5 closely related papers on similar clinical questions. Consider them for comparative context: