Purpose: Four-dimensional-computed-tomography (4DCT) is routinely utilized in clinics for radiotherapy treatment planning and has been previously investigated as a method to generate specific-ventilation-maps (Guerrero et al. Int J Radiat Oncol Biol Phys, 2005). Currently, 4DCT specific-ventilation has been implemented in multiple clinical-trials to generate functional-lung-avoidance-plans for image-guided radiotherapy (NCT02308709, NCT02528942, NCT02843568). In a similar fashion to 4DCT specific-ventilation generation, free-breathing-pulmonary proton MRI (FBMRI) using clinically available MRI systems and pulse-sequences provides a non-contrast-enhanced method to generate specific-ventilation-maps (Capaldi et al. Radiology, 2018). Here, our objective was to perform a â€œhead-to-headâ€? comparison between free-breathing MRI versus CT to generate specific-ventilation in non-small-cell-lung-cancer (NSCLC) patients.
Methods: Stage IIIA/IIIB NSCLC patients (n=12, 67Â±6yrs) provided written-informed-consent to a randomized controlled clinical-trial (NCT02002052) that aimed to compare outcomes related to MRI-guided versus conventional radiotherapy planning. Free-breathing proton MRI, as well as inhaled hyperpolarized noble-gas MRI were acquired (Capaldi et al. Acad Radiol, 2015). 4DCT was acquired within <7 days of MRI. Non-rigid image registration was performed on both free-breathing MR and CT images using the modality-independent-neighbourhood-descriptor (MIND) deformable approach (Heinrich et al. Med Image Anal, 2012). FBMRI and 4DCT specific-ventilation-maps were generated using a density-based-method (Castillo et al. Phys Med Biol, 2010). Spearman correlation coefficients (Ï?) were used to determine the relationships between whole-lung MR and 4DCT specific-ventilation-measurements. Bland-Altman analysis was implemented to determine the agreement between specific-ventilation-measurements.
Results: FBMRI specific-ventilation was correlated with 4DCT specific-ventilation (Ï?=.61, p=.03) and hyperpolarized noble-gas MRI ventilation-defect-percent (Ï?=-.70, p=.01), although Bland-Altman analysis indicated that there were biases of 0.07Â±0.06 between FBMRI and 4DCT specific-ventilation.
Conclusion: In this proof-of-concept demonstration in NSCLC-patients, free-breathing MR and CT derived specific-ventilation-maps were qualitatively and quantitatively related. These results suggest that dynamic imaging methods may interrogate similar airway abnormalities that influence the time-constants for lung filling or ventilation as compared to static breath-hold methods.