Jacobian-Based Biomechanical Texture Radiomics from Inspiratory–Expiratory CT for Predicting Target-Lobe Volume Reduction and Treatment Response In Endobronchial Valve Therapy
Abstract
Purpose
Endobronchial Valves (EBV) are one of the few treatment options for patients with moderate to severe emphysema. Eligibility is typically assessed from CT image data analysis including Emphysema Score (RA950) and Fissure Integrity Score (FIS). However not all EBV candidates are responders. The purpose of this work was to develop an approach to quantify the CT texture of biomechanical lung expansion and contraction during lung inspiration and expiration to improve prediction of candidates who will benefit from EBV treatment.
Methods
We performed a retrospective study using data from 50 patients with moderate to severe emphysema who underwent EBV treatment. Each patient underwent baseline (pre-treatment) CT scanning at both Total Lung Capacity (TLC) and Residual Volume (RV) breath-holds. Lung lobe segmentations were performed on CT image data to automatically identify lobar regions. RA950 was extracted for each lobe. Jacobian determinant was extracted from registered in RV-TLC images to quantify lung expansion and contraction. 90 features from 4 texture-radiomic feature sets were extracted from lobar regions. Pearson correlations between volume collapse of the target lobe post implantation and the radiomic features were captured.
Results
First-order minimum Jacobian intensity within the ipsilateral lung was positively correlated with change in target lobe volume reduction (TLVR) (r = 0.321, p = 0.023). Texture features reflecting spatial organization and ventilation heterogeneity were negatively correlated with TLVR: GLCM Autocorrelation (−0.368, p = 0.009), GLCM Joint Average (−0.316, p = 0.025), GLRLM SRHGLE (−0.363, p = 0.009). Within the target lobe, GLSZM features—LAE (−0.341, p = 0.015), LAHGLE (−0.371, p = 0.008), Zone Variance ( −0.342, p = 0.015)—were significantly associated with TLVR.
Conclusion
Texture-based radiomics derived from Jacobian ventilation maps capture clinically relevant patterns of regional ventilation heterogeneity associated with EBV-induced lobar deflation and may provide complementary biomechanical imaging biomarkers beyond conventional CT metrics.