🤖 AI Summary
This study addresses the instability of conventional radiomics features in anisotropic CT/MRI images, which arises from conflating voxel geometry with signal alterations introduced by interpolation. To resolve this issue, the authors propose a voxel-spacing-aware radiomics framework that explicitly models the physical geometry of vox日晚间 through metadata, thereby decoupling intrinsic geometric structure from interpolation-induced distortions—all without image resampling or alteration of the original signal. Implemented within PyRadiomics, four configurations (NR, RS, VS, FK) were evaluated for robustness using intraclass correlation coefficients (ICC), Friedman tests, feature selection, and multilayer perceptron performance. Results demonstrate that the proposed VS method achieves near-perfect agreement with non-resampled baseline features in both CT and MRI (ICC > 0.997), delivers comparable predictive performance, and effectively circumvents resampling-related biases.
📝 Abstract
Objective: Radiomic texture features are usually computed in voxel-index neighborhoods, implicitly assuming isotropic spatial relationships. In anisotropic images, this can confound voxel geometry with interpolation-induced signal changes. We developed a voxel-spacing-aware radiomic framework that incorporates physical geometry into texture computation without resampling.
Approach: We modified PyRadiomics to account for voxel spacing while preserving the native image signal. Four configurations were compared: native non-resampled extraction (NR), isotropic resampling (RS), voxel-spacing-aware extraction (VS), and fake-isotropic preprocessing (FK), in which spacing metadata were overwritten without altering the image array. Experiments included 685 LIDC-IDRI pulmonary nodules and 209 I-SPY2 breast MRI cases, with 196 radiomic descriptors. Robustness was assessed using ICC, within-subject variability, Friedman testing, feature selection, machine learning, a multilayer perceptron, and external validation.
Main results: VS showed near-native agreement with NR: median ICC(A,1) was 0.9976 in CT and 0.9984 in MRI. RS produced lower agreement and larger deviations, while FK showed intermediate behavior, confirming that spacing metadata alone can affect radiomic features. Gradient-derived and neighborhood-sensitive descriptors were most affected by preprocessing. VS preserved predictive performance comparable to NR in external CT validation, whereas MRI showed greater variability across preprocessing strategies and classifiers.
Significance: Voxel-spacing-aware extraction separates geometric modeling from interpolation-induced signal modification while preserving the native image signal, offering a coherent alternative to isotropic resampling for radiomic analysis of anisotropic CT and MRI.