Identification of Precise 3D CT Radiomics for Habitat Computation by Machine Learning in Cancer.
Olivia PriorCarlos MacarroVíctor NavarroCamilo MonrealMarta LigeroAlonso Garcia-RuizGarazi SernaSara SimonettiIrene BrañaMaria VieitoManuel EscobarJaume CapdevilaAnnette T ByrneRodrigo DienstmannRodrigo De Almeida ToledoPaolo G NuciforoElena GarraldaFrancesco GrussuKinga BernatowiczRaquel Perez-LopezPublished in: Radiology. Artificial intelligence (2024)
"Just Accepted" papers have undergone full peer review and have been accepted for publication in Radiology: Artificial Intelligence . This article will undergo copyediting, layout, and proof review before it is published in its final version. Please note that during production of the final copyedited article, errors may be discovered which could affect the content. Purpose To identify precise 3D radiomics features in CT images that enable computation of stable and biologically meaningful habitats with machine learning for cancer heterogeneity assessment. Materials and Methods This retrospective study included 2436 liver or lung lesions from 605 CT scans (November 2010-December 2021) in 318 patients with cancer (mean age, 64.5 years ± 10.1 [SD]; 185 male patients). 3D radiomics were computed from original and perturbed (simulated retest) images with different combinations of feature computation kernel radius (R) and bin size (B). The lower 95% confidence limit (LCL) of the intraclass correlation coefficient was used to measure repeatability and reproducibility. Precise features were identified by combining repeatability and reproducibility results (LCL ≥ 0.50). Habitats were obtained with Gaussian Mixture Models in original and perturbed data using precise radiomics features and compared with habitats obtained using all features. The Dice similarity coefficient (DSC) was used to assess habitat stability. Biologic correlates of CT habitats were explored in a case study, with a cohort of 13 patients with CT, multiparametric MRI (mpMRI) and tumor biopsies. Results 3D radiomics showed poor repeatability (median LCL[IQR] of 0.442 [0.312-0.516]) and poor reproducibility against R (0.44[0.33-0.526]) but excellent reproducibility against B (0.929[0.853-0.988]). Twentysix radiomics features were precise, differing in lung and liver lesions. Habitats obtained with precise features (DSC, 0.601 (0.494-0.712) and 0.651 (0.52-0.784) for lung and liver lesions, respectively) were more stable than those obtained with all features (DSC, 0.532 (0.424-0.637) and 0.587 (0.465-0.703), respectively; P < .001). In the case study, CT habitats correlated quantitatively and qualitatively, with heterogeneity observed in mpMRI habitats and histology. Conclusion Precise 3D radiomics features were identified on CT that enabled tumor heterogeneity assessment through stable tumor habitat computation. ©RSNA, 2024.
Keyphrases
- contrast enhanced
- machine learning
- artificial intelligence
- magnetic resonance imaging
- diffusion weighted imaging
- computed tomography
- dual energy
- image quality
- magnetic resonance
- deep learning
- lymph node metastasis
- big data
- climate change
- positron emission tomography
- single cell
- papillary thyroid
- squamous cell carcinoma
- end stage renal disease
- chronic kidney disease
- ejection fraction
- rheumatoid arthritis
- patient safety
- squamous cell