Objective To identify CT-acquisition parameters accounting for radiomics variability and to develop a post-acquisition CT-image correction method to reduce variability and improve radiomics classification in both phantom and clinical applications. Methods CT-acquisition protocols were prospectively tested in a phantom. The multi-centric retrospective clinical study included CT scans of patients with colorectal/renal cancer liver metastases. Ninety-three radiomics features of first order and texture were extracted. Intraclass correlation coefficients (ICCs) between CT-acquisition protocols were evaluated to define sources of variability. Voxel size, ComBat, and singular value decomposition (SVD) compensation methods were explored for reducing the radiomics variability. The number of robust features was compared before and after correction using two-proportion z test. The radiomics classification accuracy (K-means purity) was assessed before and after ComBat- and SVD-based correction. Results Fifty-three acquisition protocols in 13 tissue densities were analyzed. Ninety-seven liver metastases from 43 patients with CT from two vendors were included. Pixel size, reconstruction slice spacing, convolution kernel, and acquisition slice thickness are relevant sources of radiomics variability with a percentage of robust features lower than 80%. Resampling to isometric voxels increased the number of robust features when images were acquired with different pixel sizes (p < 0.05). SVD-based for thickness correction and ComBat correction for thickness and combined thickness–kernel increased the number of reproducible features (p < 0.05). ComBat showed the highest improvement of radiomics-based classification in both the phantom and clinical applications (K-means purity 65.98 vs 73.20). Conclusion CT-image post-acquisition processing and radiomics normalization by means of batch effect correction allow for standardization of large-scale data analysis and improve the classification accuracy. Key Points • The voxel size (accounting for the pixel size and slice spacing), slice thickness, and convolution kernel are relevant sources of CT-radiomics variability. • Voxel size resampling increased the mean percentage of robust CT-radiomics features from 59.50 to 89.25% when comparing CT scans acquired with different pixel sizes and from 71.62 to 82.58% when the scans were acquired with different slice spacings. • ComBat batch effect correction reduced the CT-radiomics variability secondary to the slice thickness and convolution kernel, improving the capacity of CT-radiomics to differentiate tissues (in the phantom application) and the primary tumor type from liver metastases (in the clinical application).

In this trial, local control and survival rates after SBRT were very good. Treatment with SBRT had no significant impact on lung function at 36 months. These findings provide further support for the use of SBRT as a radical treatment for NSCLC. Lung toxicity is minimal, even in patients with poor pulmonary function before treatment.

BackgroundTo compare thermoplastic masks (TMP) and vacuum cushion system (VCS) to assess differences in interfraction set up accuracy in patients treated with stereotactic radiotherapy (SBRT) for oligometastatic lung cancer. Secondarily, to survey radiotherapy technologists to assess their satisfaction with the two systems.MethodsRetrospective study of patients treated with lung SBRT between 2008 to 2012 at our institution. Immobilization was performed for 73 treatment sessions (VCS = 40; TMP = 33). A total of 246 cone-beams were analysed. Patients considered ineligible for surgery with a life expectancy ≥6 months and performance status > 1 were included. Target lesion location was verified by cone beam computed tomography (CBCT) prior to each session, with displacements assessed by CBCT simulation prior to each treatment session. Couch shifts were registered prospectively in vertical, longitudinal, and latero-lateral directions to obtain Kernel coordinates (3D representation). Technologists were surveyed to assess their satisfaction with indexing, positioning, and learning curve of the two systems. Setup displacements were obtained in all patients for each treatment plan and for each session. To assess differences between the immobilization systems, a t-test (Welch) was performed.ResultsMean displacements for the TMP and VC systems, respectively, were as follows: session one, 0.64 cm vs 1.05 cm (p = 0.0002); session two, 0.49 cm vs 1.02 cm (p < 0.0001), and session three, 0.56 vs 0.97 cm (p = 0.0011). TMP resulted in significantly smaller shifts vs. VCS in all three treatment sessions. Technologists rated the learning curve, set up, and positioning more highly for TMP versus VCS.ConclusionsDue to the high doses and steep gradients in lung SBRT, accurate and reproducible inter-fraction set up is essential. We found that thermoplastic masks offers better reproducibility with significantly less interfractional set up displacement than vacuum cushions. Moreover, radiotherapy technologists rated the TMP system higher. Taken together, these two findings suggest that TMP may be preferable to VCS. However, more research is needed to determine both inter- and intrafraction error to identify the optimal immobilisation system for use in lung SBRT.

No abstract