Objectives: To investigate whether radiomics features extracted from iliac bone marrow on prostate MRI can differentiate patients with monoclonal gammopathy of undetermined significance (MGUS) alone from those with concomitant MGUS and prostate cancer (PCa). Methods: This retrospective study included 28 patients with hematologically confirmed MGUS who underwent multiparametric prostate MRI and MRI/ultrasound fusion-guided biopsy, stratified into MGUS + PCa (n = 16) and MGUS-only (n = 12). The right iliac bone marrow was manually segmented on axial T2-weighted images. A total of 107 radiomics features were extracted using PyRadiomics. A hybrid descriptive-inferential approach was applied for feature selection, followed by discriminant analysis with 5-fold cross-validation. Diagnostic performance was evaluated using sensitivity, specificity, accuracy, PPV, NPV, and AUROC with 95% bootstrap confidence intervals. Results: ANOVA identified two features with significant group differences. After hybrid feature selection, original_glcm_ClusterShade was the most discriminative feature. The model achieved sensitivity of 87.5%, specificity of 41.7%, accuracy of 67.9%, PPV of 66.7%, NPV of 71.4%, and AUROC of 0.745 (95% CI: 0.54–0.91; p = 0.013). Conclusions: Radiomics analysis of the iliac bone marrow on routine prostate MRI may reveal microstructural differences between patients with isolated MGUS and those with concomitant MGUS and PCa. These hypothesis-generating findings warrant prospective validation in larger multicenter cohorts before clinical application.
Radiomics Assessment of Iliac Bone Marrow on Prostate MRI Discriminates Monoclonal Gammopathy of Undetermined Significance with and Without Concomitant Prostate Cancer: A Proof-of-Concept Study
Girometti R.;
2026-01-01
Abstract
Objectives: To investigate whether radiomics features extracted from iliac bone marrow on prostate MRI can differentiate patients with monoclonal gammopathy of undetermined significance (MGUS) alone from those with concomitant MGUS and prostate cancer (PCa). Methods: This retrospective study included 28 patients with hematologically confirmed MGUS who underwent multiparametric prostate MRI and MRI/ultrasound fusion-guided biopsy, stratified into MGUS + PCa (n = 16) and MGUS-only (n = 12). The right iliac bone marrow was manually segmented on axial T2-weighted images. A total of 107 radiomics features were extracted using PyRadiomics. A hybrid descriptive-inferential approach was applied for feature selection, followed by discriminant analysis with 5-fold cross-validation. Diagnostic performance was evaluated using sensitivity, specificity, accuracy, PPV, NPV, and AUROC with 95% bootstrap confidence intervals. Results: ANOVA identified two features with significant group differences. After hybrid feature selection, original_glcm_ClusterShade was the most discriminative feature. The model achieved sensitivity of 87.5%, specificity of 41.7%, accuracy of 67.9%, PPV of 66.7%, NPV of 71.4%, and AUROC of 0.745 (95% CI: 0.54–0.91; p = 0.013). Conclusions: Radiomics analysis of the iliac bone marrow on routine prostate MRI may reveal microstructural differences between patients with isolated MGUS and those with concomitant MGUS and PCa. These hypothesis-generating findings warrant prospective validation in larger multicenter cohorts before clinical application.| File | Dimensione | Formato | |
|---|---|---|---|
|
applsci-16-06814.pdf
accesso aperto
Tipologia:
Versione Editoriale (PDF)
Licenza:
Creative commons
Dimensione
5.52 MB
Formato
Adobe PDF
|
5.52 MB | Adobe PDF | Visualizza/Apri |
I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.


