Inter-scan and inter-observer tumour volume delineation variability on cone beam computed tomography in patients treated with stereotactic body radiation therapy for early-stage non-small cell lung cancer

Abstract

Introduction: Quantification of volume changes on cone beam computed tomography (CBCT) during lung stereotactic body radiation therapy (SBRT) for non-small cell lung cancer (NSCLC) may provide a useful radiological marker for radiation response and for adaptive treatment planning. This study quantifies inter-scan and inter-observer variability in tumour volume delineation on CBCT.

Methods: Three clinicians independently contoured the primary gross tumour volume (GTV) manually on CBCTs taken immediately before SBRT treatment (pre) and after the same SBRT treatment (post) for 19 NSCLC patients. Relative volume differences (RVD) were calculated between the pre- and post-CBCTs for a given treatment and between any two of three observers for a given CBCT. Coefficient of variation (CV) was used to quantitatively measure and compare the extent of variability.

Results: Inter-observer variability had a significantly higher CV of 0.15 ± 0.13 compared to inter-scan CV of 0.03 ± 0.04 with P < 0.0001. The greatest variability was observed with tumours (<2 cm in diameter) versus larger tumours with 95% limit of agreement (LOA) (Mean ± Standard Deviation) of 1.90% ± 19.55% vs. -0.97% ± 12.26% for inter-scan RVD and 29.99% ± 73.84% vs. 9.37% ± 29.95% for inter-observer RVD respectively.

Conclusions: Inter-observer variability was greater than inter-scan variability for tumour volume delineation on CBCT with greatest variability for small tumours (<2 cm in diameter). LOA for inter-scan variability (~12%) helps defines a threshold for clinically meaningful tumour volume change during SBRT treatment for tumours with diameter greater than 2 cm, with larger thresholds needed for smaller tumours.

Keywords: CBCT; NSCLC; SBRT; inter-observer variability; inter-scan variability.