Quantitative Measures of Pure Ground-Glass Nodules from an Artificial Intelligence Software for Predicting Invasiveness of Pulmonary Adenocarcinoma on Low-Dose CT: A Multicenter Study
These videos have been matched automatically. Contact us if they are not relevant.
These videos have been matched automatically. Contact us if they are not relevant.
View abstract on PubMed
Summary
This summary is machine-generated.Artificial intelligence (AI) software accurately measures pure ground-glass nodule (pGGN) diameter, comparable to radiologists. AI-derived mass is the best predictor of pulmonary adenocarcinoma invasiveness on low-dose CT scans.
Area Of Science
- Radiology and Medical Imaging
- Artificial Intelligence in Medicine
- Oncology
Background
- Deep learning-based artificial intelligence (AI) tools are increasingly used for pulmonary nodule detection and segmentation.
- Accurate prediction of pulmonary adenocarcinoma invasiveness is crucial for effective lung cancer screening.
- Pure ground-glass nodules (pGGNs) on low-dose CT (LDCT) require careful characterization.
Purpose Of The Study
- To evaluate the diagnostic performance of AI-derived quantitative measures for predicting the invasiveness of pulmonary adenocarcinomas presenting as pGGNs on LDCT.
- To compare the accuracy of AI-generated measurements with radiologist measurements for pGGN diameter.
Main Methods
- A total of 388 pGGNs were analyzed from three cohorts (training, testing, external validation).
- AI software extracted quantitative measures: diameter, volume, attenuation, and mass.
- Radiologists manually measured nodule diameter; agreement assessed using intra-class correlation coefficient (ICC) and Bland-Altman analysis. Diagnostic performance evaluated using receiver operating characteristic (ROC) curves and area under the curve (AUC).
Main Results
- AI-derived diameter measurements showed high agreement with radiologist measurements (ICC 0.972-0.981, low bias).
- The 'mass' quantitative measure demonstrated the highest AUCs (0.893-0.915) across all cohorts for predicting invasiveness.
- Mass outperformed diameter (AI and radiologist), volume, and attenuation in diagnostic performance (all p < 0.05).
Conclusions
- AI software provides automated pGGN diameter measurements with accuracy comparable to radiologists on LDCT.
- AI-derived 'mass' is the most effective quantitative predictor of pulmonary adenocarcinoma invasiveness in pGGNs.
- This AI-driven quantitative analysis can aid clinical decision-making for pGGNs in lung cancer screening.
These videos have been matched automatically. Contact us if they are not relevant.
These videos have been matched automatically. Contact us if they are not relevant.