A deep learning model for translating CT to ventilation imaging: analysis of accuracy and impact on functional avoidance radiotherapy planning

• Published in: Japanese journal of radiology 2024-07, Vol.42 (7), p.765-776
• Main Authors: Hou, Zhen, Kong, Youyong, Wu, Junxian, Gu, Jiabing, Liu, Juan, Gao, Shanbao, Yin, Yicai, Zhang, Ling, Han, Yongchao, Zhu, Jian, Li, Shuangshuang
• Format: Article
• Language: English
• Subjects: Aged; Algorithms; Avoidance; Avoidance learning; Carcinoma, Non-Small-Cell Lung - diagnostic imaging; Carcinoma, Non-Small-Cell Lung - radiotherapy; Clinical trials; Computed tomography; Deep Learning; Female; Four-Dimensional Computed Tomography - methods; Humans; Imaging; Lung - diagnostic imaging; Lung Neoplasms - diagnostic imaging; Lung Neoplasms - radiotherapy; Lungs; Male; Medical imaging; Medicine; Medicine & Public Health; Middle Aged; Nuclear Medicine; Original Article; Planning; Pneumonitis; Radiation therapy; Radiology; Radiotherapy; Radiotherapy Planning, Computer-Assisted - methods; Respiratory function; Risk; Tomography, X-Ray Computed - methods; Toxicity; Ventilation
• ISSN: 1867-1071; 1867-108X; 1867-108X
• DOI: 10.1007/s11604-024-01550-2

Purpose Radiotherapy planning incorporating functional lung images has the potential to reduce pulmonary toxicity. Free-breathing 4DCT-derived ventilation image (CTVI) may help quantify lung function. This study introduces a novel deep-learning model directly translating planning CT images into CTVI. We investigated the accuracy of generated images and the impact on functional avoidance planning. Materials and methods Paired planning CT and 4DCT scans from 48 patients with NSCLC were randomized to training ( n  = 41) and testing ( n  = 7) data sets. The ventilation maps were generated from 4DCT using a Jacobian-based algorithm to provide ground truth labels (CTVI 4DCT ). A 3D U-Net-based model was trained to map CT to synthetic CTVI (CTVI Syn ) and validated using fivefold cross-validation. The highest-performing model was applied to the testing set. Spearman's correlation ( r s ) and Dice similarity coefficient (DSC) determined voxel-wise and functional-wise concordance between CTVI 4DCT and CTVI Syn . Three plans were designed per patient in the testing set: one clinical plan without CTVI and two functional avoidance plans combined with CTVI 4DCT or CTVI Syn , aimed at sparing high-functional lungs defined as the top 50% of the percentile ventilation ranges. Dose–volume (DVH) parameters regarding the planning target volume (PTV) and organs at risk (OARs) were recorded. Radiation pneumonitis (RP) risk was estimated using a dose–function (DFH)-based normal tissue complication probability (NTCP) model. Results CTVI Syn showed a mean rs value of 0.65 ± 0.04 compared to CTVI 4DCT . Mean DSC values over the top 50% and 60% of ventilation ranges were 0.41 ± 0.07 and 0.52 ± 0.10, respectively. In the test set ( n  = 7), all patients’ RP-risk benefited from CTVI 4DCT -guided plans (Risk mean_4DCT_vs_Clinical : 29.24% vs. 49.12%, P  = 0.016), and six patients benefited from CTVI Syn -guided plans (Risk mean_Syn_vs_Clinical : 31.13% vs. 49.12%, P  = 0.022). There were no significant differences in DVH and DFH metrics between CTVI Syn and CTVI 4DCT -guided plan ( P  > 0.05). Conclusion Using deep-learning techniques, CTVI Syn generated from planning CT exhibited a moderate-to-high correlation with CTVI 4DCT . The CTVI Syn -guided plans were comparable to the CTVI 4DCT -guided plans, effectively reducing pulmonary toxicity in patients while maintaining acceptable plan quality. Further prospective trials are needed to validate these findings.