Comparison of Computed Tomography Ventilation Imaging and Hyperpolarized ³He Magnetic Resonance Imaging in Patients With Lung Cancer.

Abstract

Computed tomography ventilation imaging (CTVI) has great potential for clinical translation and has been validated in numerous studies. However, previous studies focused on image comparisons, and little is known about the dosimetric implications of "good" or "poor" image correlations or agreements in radiation therapy (RT). This study assessed the agreement of dose-function metrics between CTVI and hyperpolarized ³He magnetic resonance imaging (MRI) as a reference. This study included 27 patients with non-small cell lung cancer who were randomized in a phase 2 trial examining ³He MRI-guided functional avoidance RT. All patients underwent 4-dimensional computed tomography and ³He MRI. CTVI was retrospectively created. Pearson correlation and Bland-Altman analyses were performed to assess the agreements between ³He MRI- and CTVI-based well-ventilated lung dose-function metrics, including functional V_20Gy (fV_20Gy) and functional mean lung dose, of RT plans. To assess the associations between dose-function metrics and clinical outcomes, we dichotomized patients by a clinically meaningful decline in Functional Assessment of Cancer Therapy-Lung: Lung Cancer Subscale (≥3-point) and the occurrence of grade ≥2 radiation pneumonitis and compared the dose-function metrics between the 2 groups. Correlation coefficients of dose-function metrics were 0.80 to 0.82 for both fV_20Gy and functional mean lung dose. Bland-Altman analyses showed a mean difference < 1% for fV_20Gy, with 95% limits of agreement of 8% to 10%, indicating that agreements varied among patients. Among Lung Cancer Subscale-evaluable patients (n = 20), those with a ≥3-point decline (n = 8) had higher fV_20Gy than those without (CTVI: 28.2% vs 21.9%, P = .05; ³He MRI: 26.2% vs 22.3%, P = 0.24). Patients with grade ≥2 radiation pneumonitis (n = 4) also showed higher values than those without (CTVI: 28.2% vs 24.2%, P = .34; ³He MRI: 27.0% vs 23.5%, P = .25). CTVI demonstrated strong correlations and small bias in dose-function metrics with ³He MRI, including fV_20Gy and functional mean lung dose. These findings support CTVI as a ventilation surrogate for functional avoidance RT.