Multiparametric MRI and Radiomics in Prostate Cancer: A Review of the Current Literature

Can PSMA PET detect intratumour heterogeneity in histological PSMA expression of primary prostate cancer? Analysis of [68Ga]Ga-PSMA-11 and [18F]PSMA-1007

PURPOSE

Prostate-specific membrane-antigen positron emission tomography (PSMA PET) is a promising candidate for non-invasive characterization of prostate cancer (PCa). This study evaluated whether PET with tracers [68Ga]Ga-PSMA-11 or [18F]PSMA-1007 is capable to depict intratumour heterogeneity of histological PSMA expression.

METHODS

Thirty-five patients with biopsy-proven primary PCa without evidence of metastatic disease nor prior interventions were prospectively enrolled. All patients underwent PSMA PET combined with computer tomography (CT) with either [68Ga]Ga-PSMA-11 (cohort I, 20 patients) or [18F]PSMA-1007 (cohort II, 15 patients) followed by radical prostatectomy. Specimens were scanned by ex-vivo CT and histologically prepared. On digitized whole-mount prostate sections, PCa areas with different morphologies were manually defined and H-Score of immunohistochemical PSMA expression was calculated with assistance by artificial intelligence (AI). PCa areas with similar H-Score were unified in segmentation on ex-vivo CT. After co-registration on PSMA PET-CT, Spearman's coefficients of PSMA expression to mean and maximum standardized uptake value (SUVmean and SUVmax) were calculated. Furthermore, the agreement of the co-registered tumour areas to gross tumour volume (GTV) in PSMA PET was analysed.

RESULTS

Thirty-two patients were included in the final analysis. For histological PCa areas, immunohistochemical PSMA expression correlated significantly to SUVmean and SUVmax (p < 0.001, p = 0.001). An approximate linear correlation between H-Score and SUVmean / SUVmax was found for tumour areas larger than 400 μm² in histology (p < 0.001). Tumour areas with strong PSMA expression showed a significantly larger overlap to GTV in PSMA PET after co-registration than tumour areas with very low PSMA expression (p < 0.01). No significant differences were found between the two tracer cohorts (p = 0.72).

CONCLUSION

PSMA PET with both [68Ga]Ga-PSMA-11 or [18F]PSMA-1007 is able to detect changes in histological PSMA expression within PCa lesions allowing biologically targeted radiotherapy.

3D-AttenNet model can predict clinically significant prostate cancer in PI-RADS category 3 patients: a retrospective multicenter study

PURPOSES

The presence of clinically significant prostate cancer (csPCa) is equivocal for patients with prostate imaging reporting and data system (PI-RADS) category 3. We aim to develop deep learning models for re-stratify risks in PI-RADS category 3 patients.

METHODS

This retrospective study included a bi-parametric MRI of 1567 consecutive male patients from six centers (Centers 1-6) between Jan 2015 and Dec 2020. Deep learning models with double channel attention modules based on MRI (AttenNet) for predicting PCa and csPCa were constructed separately. Each model was first pretrained using 1144 PI-RADS 1-2 and 4-5 images and then retrained using 238 PI-RADS 3 images from three training centers (centers 1-3), and tested using 185 PI-RADS 3 images from the other three testing centers (centers 4-6).

RESULTS

Our AttenNet models achieved excellent prediction performances in testing cohort of center 4-6 with the area under the receiver operating characteristic curves (AUC) of 0.795 (95% CI: [0.700, 0.891]), 0.963 (95% CI: [0.915, 1]) and 0.922 (95% CI: [0.810, 1]) in predicting PCa, and the corresponding AUCs were 0.827 (95% CI: [0.703, 0.952]) and 0.926 (95% CI: [0.846, 1]) in predicting csPCa in testing cohort of center 4 and center 5. In particular, 71.1% to 92.2% of non-csPCa patients were identified by our model in three testing cohorts, who can spare from invasive biopsy or RP procedure.

CONCLUSIONS

Our model offers a noninvasive screening clinical tool to re-stratify risks in PI-RADS 3 patients, thereby reducing unnecessary invasive biopsies and improving the effectiveness of biopsies.

CRITICAL RELEVANCE STATEMENT

The deep learning model with MRI can help to screen out csPCa in PI-RADS category 3.

KEY POINTS

AttenNet models included channel attention and soft attention modules. 71.1-92.2% of non-csPCa patients were identified by the AttenNet model. The AttenNet models can be a screen clinical tool to re-stratify risks in PI-RADS 3 patients.

Evaluating the feasibility of AI-predicted bpMRI image features for predicting prostate cancer aggressiveness: a multi-center study

OBJECTIVE

To evaluate the feasibility of utilizing artificial intelligence (AI)-predicted biparametric MRI (bpMRI) image features for predicting the aggressiveness of prostate cancer (PCa).

MATERIALS AND METHODS

A total of 878 PCa patients from 4 hospitals were retrospectively collected, all of whom had pathological results after radical prostatectomy (RP). A pre-trained AI algorithm was used to select suspected PCa lesions and extract lesion features for model development. The study evaluated five prediction methods, including (1) A clinical-imaging model of clinical features and image features of suspected PCa lesions selected by AI algorithm, (2) the PIRADS category, (3) a conventional radiomics model, (4) a deep-learning bases radiomics model, and (5) biopsy pathology.

RESULTS

In the externally validated dataset, the deep learning-based radiomics model showed the highest area under the curve (AUC 0.700 to 0.791). It exceeded the clinical-imaging model (AUC 0.597 to 0.718), conventional radiomic model (AUC 0.566 to 0.632), PIRADS score (AUC 0.554 to 0.613), and biopsy pathology (AUC 0.537 to 0.578). The AUC predicted by the model did not show a statistically significant difference among the three externally verified hospitals (p > 0.05).

CONCLUSION

Deep-learning radiomics models utilizing AI-extracted image features from bpMRI images can potentially be used to predict PCa aggressiveness, demonstrating a generalized ability for external validation.

CRITICAL RELEVANCE STATEMENT

Predicting the aggressiveness of prostate cancer (PCa) is important for formulating the best treatment plan for patients. The radiomic model based on deep learning is expected to provide an objective and non-invasive method for evaluating the aggressiveness of PCa.

KEY POINTS

Predicting the aggressiveness of PCa is important for patients to obtain the best treatment options. The deep learning-based radiomics model can predict the aggressiveness of PCa with high accuracy. The model has good universality when tested on multiple external datasets.

Utility of quantitative measurement of T2 using restriction spectrum imaging for detection of clinically significant prostate cancer

The Restriction Spectrum Imaging restriction score (RSIrs) has been shown to improve the accuracy for diagnosis of clinically significant prostate cancer (csPCa) compared to standard DWI. Both diffusion and T2 properties of prostate tissue contribute to the signal measured in DWI, and studies have demonstrated that each may be valuable for distinguishing csPCa from benign tissue. The purpose of this retrospective study was to (1) determine whether prostate T2 varies across RSI compartments and in the presence of csPCa, and (2) evaluate whether csPCa detection with RSIrs is improved by acquiring multiple scans at different TEs to measure compartmental T2 (cT2). Data includes two cohorts scanned for csPCa with 3T multi-b-value diffusion-weighted sequences acquired at multiple TEs. cT2 values were computed from multi-TE RSI data and compared by compartment. CsPCa detection was compared between RSIrs and a logistic regression model (LRM) to predict the probability of csPCa using cT2 in combination with RSI measurements. Two-sample t-tests (α = 0.05) and the area under the receiver operating characteristic curve (AUC) were used for the statistical analyses. In both cohorts, T2 was different (p < 0.05) across the four RSI compartments (C1, C2, C3, C4). Voxel-level, cohort 1: T2 was different in csPCa for C1, C2, C3 (p < 0.001). Patient-level, cohort 1: T2 was different in csPCa patients in C3 (p = 0.02); cohort 2: T2 differed in csPCa patients in C1 (p = 0.01), C3 (p = 0.01) and C4 (p < 0.01). Voxel-level csPCa detection: cT2 did not improve discrimination over RSIrs alone (p = 0.9). Patient-level: RSIrs and the LRM performed better than diffusion alone (p < 0.001), but the difference in AUCs between RSIrs and the LRM was not significantly different (p = 0.54). In conclusion, significant differences in cT2 were observed between normal and cancerous prostatic tissue. With our data, however, consideration of cT2 in addition to diffusion did not significantly improve cancer detection performance.

Deep Learning Features Can Improve Radiomics-Based Prostate Cancer Aggressiveness Prediction

PURPOSE

Emerging evidence suggests that the use of artificial intelligence can assist in the timely detection and optimization of therapeutic approach in patients with prostate cancer. The conventional perspective on radiomics encompassing segmentation and the extraction of radiomic features considers it as an independent and sequential process. However, it is not necessary to adhere to this viewpoint. In this study, we show that besides generating masks from which radiomic features can be extracted, prostate segmentation and reconstruction models provide valuable information in their feature space, which can improve the quality of radiomic signatures models for disease aggressiveness classification.

MATERIALS AND METHODS

We perform 2,244 experiments with deep learning features extracted from 13 different models trained using different anatomic zones and characterize how modeling decisions, such as deep feature aggregation and dimensionality reduction, affect performance.

RESULTS

While models using deep features from full gland and radiomic features consistently lead to improved disease aggressiveness prediction performance, others are detrimental. Our results suggest that the use of deep features can be beneficial, but an appropriate and comprehensive assessment is necessary to ensure that their inclusion does not harm predictive performance.

CONCLUSION

The study findings reveal that incorporating deep features derived from autoencoder models trained to reconstruct the full prostate gland (both zonal models show worse performance than radiomics only models), combined with radiomic features, often lead to a statistically significant increase in model performance for disease aggressiveness classification. Additionally, the results also demonstrate that the choice of feature selection is key to achieving good performance, with principal component analysis (PCA) and PCA + relief being the best approaches and that there is no clear difference between the three proposed latent representation extraction techniques.

Application of Advanced Imaging to Prostate Cancer Diagnosis and Management: A Narrative Review of Current Practice and Unanswered Questions

Major advances in prostate cancer diagnosis, staging, and management have occurred over the past decade, largely due to our improved understanding of the technical aspects and clinical applications of advanced imaging, specifically magnetic resonance imaging (MRI) and prostate-cancer-specific positron emission tomography (PET). Herein, we review the established utility of these important and exciting technologies, as well as areas of controversy and uncertainty that remain important areas for future study. There is strong evidence supporting the utility of MRI in guiding initial biopsy and assessing local disease. There is debate, however, regarding how to best use the imaging modality in risk stratification, treatment planning, and assessment of biochemical failure. Prostate-cancer-specific PET is a relatively new technology that provides great value to the evaluation of newly diagnosed, treated, and recurrent prostate cancer. However, its ideal use in treatment decision making, staging, recurrence detection, and surveillance necessitates further research. Continued study of both imaging modalities will allow for an improved understanding of their best utilization in improving cancer care.

Pan-cancer Analysis Identifies AIMP2 as a Potential Biomarker for Breast Cancer

Introduction

Aminoacyl tRNA synthetase complex interacting with multifunctional protein 2 (AIMP2) is a significant regulator of cell proliferation and apoptosis. Despite its abnormal expression in various tumor types, the specific functions and effects of AIMP2 on tumor immune cell infiltration, proliferation, and migration remain unclear.

Materials and Methods

To assess AIMP2's role in tumor immunity, we conducted a pan-cancer multi-database analysis using the Cancer Genome Atlas (TCGA), Genotype-Tissue Expression (GTEx), and Cancer Cell Lines Encyclopedia (CCLE) datasets, examining expression levels, prognosis, tumor progression, and immune microenvironment. Additionally, we investigated AIMP2's impact on breast cancer (BRCA) proliferation and migration using cell counting kit 8 (CCK-8) assay, transwell assays, and western blot analysis.

Results

Our findings revealed that AIMP2 was overexpressed in 24 tumor tissue types compared to normal tissue and was associated with four tumor stages. Survival analysis indicated that AIMP2 expression was strongly correlated with overall survival (OS) in certain cancer patients, with high AIMP2 expression linked to poorer prognosis in five cancer types.

Conclusion

Finally, siRNA-mediated AIMP2 knockdown inhibited BRCA cell proliferation and migration in vitro. In conclusion, our pan-cancer analysis suggests that AIMP2 may play a crucial role in tumor immunity and could serve as a potential prognostic marker, particularly in BRCA.

Towards the Definition of Radiomic Features and Clinical Indices to Enhance the Diagnosis of Clinically Significant Cancers in PI-RADS 4 and 5 Lesions

Prostate cancer (PC) is the most frequently diagnosed cancer among adult men, and its incidence is increasing worldwide [...].

MRI-derived radiomics models for diagnosis, aggressiveness, and prognosis evaluation in prostate cancer

Prostate cancer (PCa) is a pernicious tumor with high heterogeneity, which creates a conundrum for making a precise diagnosis and choosing an optimal treatment approach. Multiparametric magnetic resonance imaging (mp-MRI) with anatomical and functional sequences has evolved as a routine and significant paradigm for the detection and characterization of PCa. Moreover, using radiomics to extract quantitative data has emerged as a promising field due to the rapid growth of artificial intelligence (AI) and image data processing. Radiomics acquires novel imaging biomarkers by extracting imaging signatures and establishes models for precise evaluation. Radiomics models provide a reliable and noninvasive alternative to aid in precision medicine, demonstrating advantages over traditional models based on clinicopathological parameters. The purpose of this review is to provide an overview of related studies of radiomics in PCa, specifically around the development and validation of radiomics models using MRI-derived image features. The current landscape of the literature, focusing mainly on PCa detection, aggressiveness, and prognosis evaluation, is reviewed and summarized. Rather than studies that exclusively focus on image biomarker identification and method optimization, models with high potential for universal clinical implementation are identified. Furthermore, we delve deeper into the critical concerns that can be addressed by different models and the obstacles that may arise in a clinical scenario. This review will encourage researchers to design models based on actual clinical needs, as well as assist urologists in gaining a better understanding of the promising results yielded by radiomics.

Novel Multiparametric Magnetic Resonance Imaging-Based Deep Learning and Clinical Parameter Integration for the Prediction of Long-Term Biochemical Recurrence-Free Survival in Prostate Cancer after Radical Prostatectomy

Radical prostatectomy (RP) is the main treatment of prostate cancer (PCa). Biochemical recurrence (BCR) following RP remains the first sign of aggressive disease; hence, better assessment of potential long-term post-RP BCR-free survival is crucial. Our study aimed to evaluate a combined clinical-deep learning (DL) model using multiparametric magnetic resonance imaging (mpMRI) for predicting long-term post-RP BCR-free survival in PCa. A total of 437 patients with PCa who underwent mpMRI followed by RP between 2008 and 2009 were enrolled; radiomics features were extracted from T2-weighted imaging, apparent diffusion coefficient maps, and contrast-enhanced sequences by manually delineating the index tumors. Deep features from the same set of imaging were extracted using a deep neural network based on pretrained EfficentNet-B0. Here, we present a clinical model (six clinical variables), radiomics model, DL model (DLM-Deep feature), combined clinical-radiomics model (CRM-Multi), and combined clinical-DL model (CDLM-Deep feature) that were built using Cox models regularized with the least absolute shrinkage and selection operator. We compared their prognostic performances using stratified fivefold cross-validation. In a median follow-up of 61 months, 110/437 patients experienced BCR. CDLM-Deep feature achieved the best performance (hazard ratio [HR] = 7.72), followed by DLM-Deep feature (HR = 4.37) or RM-Multi (HR = 2.67). CRM-Multi performed moderately. Our results confirm the superior performance of our mpMRI-derived DL algorithm over conventional radiomics.

Value of handcrafted and deep radiomic features towards training robust machine learning classifiers for prediction of prostate cancer disease aggressiveness

There is a growing piece of evidence that artificial intelligence may be helpful in the entire prostate cancer disease continuum. However, building machine learning algorithms robust to inter- and intra-radiologist segmentation variability is still a challenge. With this goal in mind, several model training approaches were compared: removing unstable features according to the intraclass correlation coefficient (ICC); training independently with features extracted from each radiologist's mask; training with the feature average between both radiologists; extracting radiomic features from the intersection or union of masks; and creating a heterogeneous dataset by randomly selecting one of the radiologists' masks for each patient. The classifier trained with this last resampled dataset presented with the lowest generalization error, suggesting that training with heterogeneous data leads to the development of the most robust classifiers. On the contrary, removing features with low ICC resulted in the highest generalization error. The selected radiomics dataset, with the randomly chosen radiologists, was concatenated with deep features extracted from neural networks trained to segment the whole prostate. This new hybrid dataset was then used to train a classifier. The results revealed that, even though the hybrid classifier was less overfitted than the one trained with deep features, it still was unable to outperform the radiomics model.

Developing a predictive model for clinically significant prostate cancer by combining age, PSA density, and mpMRI

PURPOSE

The study aimed to construct a predictive model for clinically significant prostate cancer (csPCa) and investigate its clinical efficacy to reduce unnecessary prostate biopsies.

METHODS

A total of 847 patients from institute 1 were included in cohort 1 for model development. Cohort 2 included a total of 208 patients from institute 2 for external validation of the model. The data obtained were used for retrospective analysis. The results of magnetic resonance imaging were obtained using Prostate Imaging Reporting and Data System version 2.1 (PI-RADS v2.1). Univariate and multivariate analyses were performed to determine significant predictors of csPCa. The diagnostic performances were compared using the receiver operating characteristic (ROC) curve and decision curve analyses.

RESULTS

Age, prostate-specific antigen density (PSAD), and PI-RADS v2.1 scores were used as predictors of the model. In the development cohort, the areas under the ROC curve (AUC) for csPCa about age, PSAD, PI-RADS v2.1 scores, and the model were 0.675, 0.823, 0.875, and 0.938, respectively. In the external validation cohort, the AUC values predicted by the four were 0.619, 0.811, 0.863, and 0.914, respectively. Decision curve analysis revealed that the clear net benefit of the model was higher than PI-RADS v2.1 scores and PSAD. The model significantly reduced unnecessary prostate biopsies within the risk threshold of > 10%.

CONCLUSIONS

In both internal and external validation, the model constructed by combining age, PSAD, and PI-RADS v2.1 scores exhibited excellent clinical efficacy and can be utilized to reduce unnecessary prostate biopsies.

Predicting Local Failure after Partial Prostate Re-Irradiation Using a Dosiomic-Based Machine Learning Model

The aim of this study is to predict local failure after partial prostate re-irradiation for the treatment of isolated locally recurrent prostate cancer by using a machine learning classifier based on radiomic features from pre-treatment computed tomography (CT), positron-emission tomography (PET) and biological effective dose distribution (BED) of the radiotherapy plan. The analysis was conducted on a monocentric dataset of 43 patients with evidence of isolated intraprostatic recurrence of prostate cancer after primary external beam radiotherapy. All patients received partial prostate re-irradiation delivered by volumetric modulated arc therapy. The gross tumor volume (GTV) of each patient was manually contoured from planning CT, choline-PET and dose maps. An ensemble machine learning pipeline including unbalanced data correction and feature selection was trained using the radiomic and dosiomic features as input for predicting occurrence of local failure. The model performance was assessed using sensitivity, specificity, accuracy and area under receiver operating characteristic curves of the score function in 10-fold cross validation repeated 100 times. Local failure was observed in 13 patients (30%), with a median time to recurrence of 36.7 months (range = 6.1-102.4 months). A four variables ensemble machine learning model resulted in accuracy of 0.62 and AUC 0.65. According to our results, a dosiomic machine learning classifier can predict local failure after partial prostate re-irradiation.

Quality of Multicenter Studies Using MRI Radiomics for Diagnosing Clinically Significant Prostate Cancer: A Systematic Review

Background: Reproducibility and generalization are major challenges for clinically significant prostate cancer modeling using MRI radiomics. Multicenter data seem indispensable to deal with these challenges, but the quality of such studies is currently unknown. The aim of this study was to systematically review the quality of multicenter studies on MRI radiomics for diagnosing clinically significant PCa. Methods: This systematic review followed the 2020 Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) checklist. Multicenter studies investigating the value of MRI radiomics for the diagnosis of clinically significant prostate cancer were included. Quality was assessed using the checklist for artificial intelligence in medical imaging (CLAIM) and the radiomics quality score (RQS). CLAIM consisted of 42 equally important items referencing different elements of good practice AI in medical imaging. RQS consisted of 36 points awarded over 16 items related to good practice radiomics. Final CLAIM and RQS scores were percentage-based, allowing for a total quality score consisting of the average of CLAIM and RQS. Results: Four studies were included. The average total CLAIM score was 74.6% and the average RQS was 52.8%. The corresponding average total quality score (CLAIM + RQS) was 63.7%. Conclusions: A very small number of multicenter radiomics PCa classification studies have been performed with the existing studies being of bad or average quality. Good multicenter studies might increase by encouraging preferably prospective data sharing and paying extra care to documentation in regards to reproducibility and clinical utility.

Ex Vivo Fluorescence Confocal Microscopy (FCM) of Prostate Biopsies Rethought: Opportunities of Intraoperative Examinations of MRI-Guided Targeted Biopsies in Routine Diagnostics

Background: The diagnosis of prostate carcinoma (PCa) requires time- and material-consuming histopathological examinations. Ex vivo fluorescence confocal microscopy (FCM) can detect carcinoma foci in diagnostic biopsies intraoperatively. Methods: MRI-guided and systematic biopsies were identified in a dataset of our previously published study cohort. Detection rates of clinically relevant tumors were determined in both groups. A retrospective blinded trial was performed to determine how many tumors requiring intervention were detectable via FCM analysis of MRI-guided targeted biopsies alone. Results: MRI-guided targeted biopsies revealed tumors more frequently than systematic biopsies. Carcinomas in need of intervention were reliably represented in the MRI-guided biopsies and were identified in intraoperative FCM microscopy. Combined with serum PSA levels and clinical presentation, 91% of the carcinomas in need of intervention were identified. Conclusions: Intraoperative FCM analysis of MRI-guided biopsies is a promising approach for the efficient diagnosis of PCa. The method allows a timely assessment of whether a tumor disease requiring intervention is present and can reduce the psychological stress of the patient in the waiting period of the histological finding. Furthermore, this technique can lead to reduction of the total number of biopsies needed for the diagnosis of PCa.

Radiomics Analysis on [68Ga]Ga-PSMA-11 PET and MRI-ADC for the Prediction of Prostate Cancer ISUP Grades: Preliminary Results of the BIOPSTAGE Trial

Simple Summary

Radiomics analysis is used on magnetic resonance imaging – apparent diffusion coefficient (MRI-ADC) maps and [68Ga]Ga-PSMA-11 PET uptake maps to assess unique tumor traits not visible to the naked eye and predict histology-proven ISUP grades in a cohort of 28 patients. Our study’s main goal is to report imaging features that can distinguish patients with low ISUP grades from those with higher grades (ISUP one+) by employing logistic regression statistical models based on MRI-ADC and 68Ga-PSMA data, as well as assess the features’ stability under small contouring variations. Our findings reveal that MRI-ADC and [⁶⁸Ga]Ga-PSMA-11 PET imaging features-based models are equivalent and complementary for predicting low ISUP grade patients. These models can be employed in broader studies to confirm their ISUP grade prediction ability and eventually impact clinical workflow by reducing overdiagnosis of indolent, early-stage PCa.

Abstract

Prostate cancer (PCa) risk categorization based on clinical/PSA testing results in a substantial number of men being overdiagnosed with indolent, early-stage PCa. Clinically non-significant PCa is characterized as the presence of ISUP grade one, where PCa is found in no more than two prostate biopsy cores.MRI-ADC and [⁶⁸Ga]Ga-PSMA-11 PET have been proposed as tools to predict ISUP grade one patients and consequently reduce overdiagnosis. In this study, Radiomics analysis is applied to MRI-ADC and [⁶⁸Ga]Ga-PSMA-11 PET maps to quantify tumor characteristics and predict histology-proven ISUP grades. ICC was applied with a threshold of 0.6 to assess the features’ stability with variations in contouring. Logistic regression predictive models based on imaging features were trained on 31 lesions to differentiate ISUP grade one patients from ISUP two+ patients. The best model based on [⁶⁸Ga]Ga-PSMA-11 PET returned a prediction efficiency of 95% in the training phase and 100% in the test phase whereas the best model based on MRI-ADC had an efficiency of 100% in both phases. Employing both imaging modalities, prediction efficiency was 100% in the training phase and 93% in the test phase. Although our patient cohort was small, it was possible to assess that both imaging modalities add information to the prediction models and show promising results for further investigations.