Assessing robustness of radiomic features by image perturbation
This work addresses the need for reliable radiomic models in medical imaging by providing a method to identify robust features when test-retest data is unavailable, though it is incremental as it builds on existing perturbation techniques.
The study tackled the problem of ensuring radiomic feature robustness for reproducibility by investigating 18 image perturbation methods to assess feature stability, finding that a specific perturbation chain minimized false positive robust features to 3.3% in NSCLC and 10.0% in HNSCC cohorts.
Image features need to be robust against differences in positioning, acquisition and segmentation to ensure reproducibility. Radiomic models that only include robust features can be used to analyse new images, whereas models with non-robust features may fail to predict the outcome of interest accurately. Test-retest imaging is recommended to assess robustness, but may not be available for the phenotype of interest. We therefore investigated 18 methods to determine feature robustness based on image perturbations. Test-retest and perturbation robustness were compared for 4032 features that were computed from the gross tumour volume in two cohorts with computed tomography imaging: I) 31 non-small-cell lung cancer (NSCLC) patients; II): 19 head-and-neck squamous cell carcinoma (HNSCC) patients. Robustness was measured using the intraclass correlation coefficient (1,1) (ICC). Features with ICC$\geq0.90$ were considered robust. The NSCLC cohort contained more robust features for test-retest imaging than the HNSCC cohort ($73.5\%$ vs. $34.0\%$). A perturbation chain consisting of noise addition, affine translation, volume growth/shrinkage and supervoxel-based contour randomisation identified the fewest false positive robust features (NSCLC: $3.3\%$; HNSCC: $10.0\%$). Thus, this perturbation chain may be used to assess feature robustness.