Combination of Peri-Tumoral and Intra-Tumoral Radiomic Features on Bi-Parametric MRI Accurately Stratifies Prostate Cancer Risk: A Multi-Site Study

Abstract

Background: Prostate cancer (PCa) influences its surrounding habitat, which tends to manifest as different phenotypic appearances on magnetic resonance imaging (MRI). This region surrounding the PCa lesion, or the peri-tumoral region, may encode useful information that can complement intra-tumoral information to enable better risk stratification. Purpose: To evaluate the role of peri-tumoral radiomic features on bi-parametric MRI (T2-weighted and Diffusion-weighted) to distinguish PCa risk categories as defined by D’Amico Risk Classification System. Materials and Methods: We studied a retrospective, HIPAA-compliant, 4-institution cohort of 231 PCa patients (n = 301 lesions) who underwent 3T multi-parametric MRI prior to biopsy. PCa regions of interest (ROIs) were delineated on MRI by experienced radiologists following which peri-tumoral ROIs were defined. Radiomic features were extracted within the intra- and peri-tumoral ROIs. Radiomic features differentiating low-risk from: (1) high-risk (L-vs.-H), and (2) (intermediate- and high-risk (L-vs.-I + H)) lesions were identified. Using a multi-institutional training cohort of 151 lesions (D1, N = 116 patients), machine learning classifiers were trained using peri- and intra-tumoral features individually and in combination. The remaining 150 lesions (D2, N = 115 patients) were used for independent hold-out validation and were evaluated using Receiver Operating Characteristic (ROC) analysis and compared with PI-RADS v2 scores. Results: Validation on D2 using peri-tumoral radiomics alone resulted in areas under the ROC curve (AUCs) of 0.84 and 0.73 for the L-vs.-H and L-vs.-I + H classifications, respectively. The best combination of intra- and peri-tumoral features resulted in AUCs of 0.87 and 0.75 for the L-vs.-H and L-vs.-I + H classifications, respectively. This combination improved the risk stratification results by 3–6% compared to intra-tumoral features alone. Our radiomics-based model resulted in a 53% accuracy in differentiating L-vs.-H compared to PI-RADS v2 (48%), on the validation set. Conclusion: Our findings suggest that peri-tumoral radiomic features derived from prostate bi-parametric MRI add independent predictive value to intra-tumoral radiomic features for PCa risk assessment.

An integrated nomogram combining deep learning, Prostate Imaging–Reporting and Data System (PI-RADS) scoring, and clinical variables for identification of clinically significant prostate cancer on biparametric MRI: a retrospective multicentre study

“…Patch-based deep learning approaches using convolutional neural networks have been used to detect and characterise prostate cancer; 17,18 however, these studies did not make use of important information in the peritumoural region, which has been shown to improve disease characterisation and classification performance. 15,16,26 Algohary and colleagues 19 showed the benefit of extracting radiomic representations from peritumoural regions along with intratumoural regions. They highlighted the differences in concentration of epithelial cells and lymphocytes between low-risk and high-risk regions by mapping the representative peritumoural regions on biparametric MRI to the whole-mount pathology slides.…”

“…We aimed to construct an integrated nomogram (referred to as ClaD) combining deep learning-based imaging predictions, PI-RADS scoring, and clinical variables using multivariable logistic regression to identify clinically significant prostate cancer on biparametric MRI. Since previous research 19 has suggested the importance of the peritumoural region for characterisation of prostate cancer on biparametric MRI, we explored multiple input configurations of the deep learning network, with patches extracted at different scales and each subsequent scale incorporating additional information from the peritumoural region.…”

Artificial Intelligence Compared to Radiologists for the Initial Diagnosis of Prostate Cancer on Magnetic Resonance Imaging: A Systematic Review and Recommendations for Future Studies

“…The 27 studies, and by extension, the CAD systems presented or evaluated within them, were categorized as either ROI Classification (ROI-C), Lesion Localization and Classification (LL&C), or Patient Classification (PAT-C); the categories are shown diagrammatically in Figure 2. ROI-C refers to (n = 16) studies where CAD systems classified pre-defined regions of interest (ROI), e.g., manually contoured lesions [19][20][21][22][23][24][25][26][27][28][29][30][31][32]44,45], LL&C refers to (n = 10) studies where CAD systems performed simultaneous lesion localization and classification [33][34][35][36][37][38][39][40][41][42], and PAT-C refers to (n = 1) studies where CAD systems classified patients directly [43].…”

“…All 27 included studies used a retrospective study design. The median size of patient cohorts used for evaluation was 98 (range 30 to 417, n = 26) for studies where the size of the evaluation patient cohort was reported [19][20][21][22][23][24][25][26][27][28][29][31][32][33][34][35][36][37][38][39][40][41][42][43][44][45]. Most studies (n = 18) considered clinically suspected patient cohorts [20][21][22][23]27,31,[34][35][36][37][38][39][40][41][42][43][44][45], while fewer studies (n = 9) considered patient cohorts with biopsy-proven prostate cancer [19,[24][25][26]…”

Radiogenomics in brain, breast, and lung cancer: opportunities and challenges

“…8 Medical imaging enables a non-invasive analysis of the functional and physiological properties of tumors, and the different available modalities are increasingly recognized for containing high-dimensional mineable data, which in turn can be used to improve medical decision making. 9,10 Imaging can also help in characterizing peritumoral regions, which are not always surgically removed for molecular characterization 11,12 and may convey information related to the tumor microenvironment. 13,14 For example, imaging characteristics of tumors are increasingly being used to predict gene expression.…”