The combination of prostate imaging reporting and data system version 2 (PI-RADS v2) and periprostatic fat thickness on multi-parametric MRI to predict the presence of prostate cancer

Navigating the gray zone: Machine learning can differentiate malignancy in PI-RADS 3 lesions

INTRODUCTION

The objective of this study is to predict the probability of prostate cancer in PI-RADS 3 lesions using machine learning methods that incorporate clinical and mpMRI parameters.

METHODS

The study included patients who had PI-RADS 3 lesions detected on mpMRI and underwent fusion biopsy between January 2020 and January 2024. Radiological parameters (Apparent diffusion coefficient (ADC), tumour ADC/contralateral ADC ratio, Ktrans value, periprostatic adipose tissue thickness, lesion size, prostate volume) and clinical parameters (age, body mass index, total prostate specific antigen, free PSA, PSA density, systemic inflammatory index, neutrophil-lymphocyte ratio [NLR], platelet lymphocyte ratio, lymphocyte monocyte ratio) were documented. The probability of prostate cancer prediction in PI-RADS 3 lesions was calculated using 6 different machine-learning models, with the input parameters being the aforementioned variables.

RESULTS

Of the 235 participants in the trial, 61 had malignant fusion biopsy pathology and 174 had benign pathology. Among 6 different machine learning algorithms, the random forest model had the highest accuracy (0.86±0.04; 95% CI 0.85-0.87), F1 score (0.91±0.03; 95% CI 0.91-0.92) and AUC value (0.92±0.06; 95% CI 0.88-0.90). In SHAP analysis based on random forest model, tumour ADC, tumour ADC/contralateral ADC ratio and PSA density were the 3 most successful parameters in predicting malignancy. On the other hand, systemic inflammatory index and neutrophil lymphocyte ratio showed higher accuracy in predicting malignancy than total PSA, age, free PSA/total PSA and lesion size in SHAP analysis.

CONCLUSION

Among the machine learning models we developed, especially the random forest model can predict malignancy in PI-RADS 3 lesions and prevent unnecessary biopsy. This model can be used in clinical practice with multicentre studies including more patients.

A modified U-Net convolutional neural network for segmenting periprostatic adipose tissue based on contour feature learning

Objective

This study trains a U-shaped fully convolutional neural network (U-Net) model based on peripheral contour measures to achieve rapid, accurate, automated identification and segmentation of periprostatic adipose tissue (PPAT).

Methods

Currently, no studies are using deep learning methods to discriminate and segment periprostatic adipose tissue. This paper proposes a novel and modified, U-shaped convolutional neural network contour control points on a small number of datasets of MRI T2W images of PPAT combined with its gradient images as a feature learning method to reduce feature ambiguity caused by the differences in PPAT contours of different patients. This paper adopts a supervised learning method on the labeled dataset, combining the probability and spatial distribution of control points, and proposes a weighted loss function to optimize the neural network's convergence speed and detection performance. Based on high-precision detection of control points, this paper uses a convex curve fitting to obtain the final PPAT contour. The imaging segmentation results were compared with those of a fully convolutional network (FCN), U-Net, and semantic segmentation convolutional network (SegNet) on three evaluation metrics: Dice similarity coefficient (DSC), Hausdorff distance (HD), and intersection over union ratio (IoU).

Results

Cropped images with a 270 × 270-pixel matrix had DSC, HD, and IoU values of 70.1%, 27 mm, and 56.1%, respectively; downscaled images with a 256 × 256-pixel matrix had 68.7%, 26.7 mm, and 54.1%. A U-Net network based on peripheral contour characteristics predicted the complete periprostatic adipose tissue contours on T2W images at different levels. FCN, U-Net, and SegNet could not completely predict them.

Conclusion

This U-Net convolutional neural network based on peripheral contour features can identify and segment periprostatic adipose tissue quite well. Cropped images with a 270 × 270-pixel matrix are more appropriate for use with the U-Net convolutional neural network based on contour features; reducing the resolution of the original image will lower the accuracy of the U-Net convolutional neural network. FCN and SegNet are not appropriate for identifying PPAT on T2 sequence MR images. Our method can automatically segment PPAT rapidly and accurately, laying a foundation for PPAT image analysis.

Periprostatic Adipose Tissue: A New Perspective for Diagnosing and Treating Prostate Cancer

Prostate cancer (PCa) is the most common tumor of the male genitourinary system. It will eventually progress to fatal metastatic castration-resistant prostate cancer, for which treatment options are limited. Adipose tissues are distributed in various parts of the body. They have different morphological structures and functional characteristics and are associated with the development of various tumors. Periprostatic adipose tissue (PPAT) is the closest white visceral adipose tissue to the prostate and is part of the PCa tumor microenvironment. Studies have shown that PPAT is involved in PCa development, progression, invasion, and metastasis through the secretion of multiple active molecules. Factors such as obesity, diet, exercise, and organochlorine pesticides can affect the development of PCa indirectly or directly through PPAT. Based on the mechanism of PPAT's involvement in regulating PCa, this review summarized various diagnostic and therapeutic approaches for PCa with potential applications to assess the progression of patients' disease and improve clinical outcomes.

Management Strategy for Prostate Imaging Reporting and Data System Category 3 Lesions

PURPOSE OF REVIEW

Prostate Imaging Reporting and Data System (PI-RADS) category 3 lesions present a clinical dilemma due to their uncertain nature, which complicates the development of a definitive management strategy. These lesions have an incidence rate of approximately 22-32%, with clinically significant prostate cancer (csPCa) accounting for about 10-30%. Therefore, a thorough evaluation is warranted.

RECENT FINDINGS

This review highlights the need for radiology peer review, including the confirmation of dynamic contrast-enhanced (DCE) compliance, as the initial step. Additional MRI models such as VERDICT or Tofts need to be verified. Current evidence shows that imaging and clinical indicators can be used for risk stratification of PI-RADS 3 lesions. For low-risk lesions, a safety net monitoring approach involving annual repeat MRI can be employed. In contrast, lesions deemed potentially risky based on prostate-specific antigen density (PSAD), 68 Ga-PSMA PET/CT, MPS, Proclarix, or AI/machine learning models should undergo biopsy. It is recommended to establish a multidisciplinary team that takes into account factors such as age, PSAD, prostate, and lesion size, as well as previous biopsy pathological findings. Combining expert opinions, clinical-imaging indicators, and emerging methods will contribute to the development of management strategies for PI-RADS 3 lesions.

Obesity-Related Cross-Talk between Prostate Cancer and Peripheral Fat: Potential Role of Exosomes

The molecular mechanisms of obesity-induced cancer progression have been extensively explored because of the significant increase in obesity and obesity-related diseases worldwide. Studies have shown that obesity is associated with certain features of prostate cancer. In particular, bioactive factors released from periprostatic adipose tissues mediate the bidirectional communication between periprostatic adipose tissue and prostate cancer. Moreover, recent studies have shown that extracellular vesicles have a role in the relationship between tumor peripheral adipose tissue and cancer progression. Therefore, it is necessary to investigate the feedback mechanisms between prostate cancer and periglandular adipose and the role of exosomes as mediators of signal exchange to understand obesity as a risk factor for prostate cancer. This review summarizes the two-way communication between prostate cancer and periglandular adipose and discusses the potential role of exosomes as a cross-talk and the prospect of using adipose tissue as a means to obtain exosomes in vitro. Therefore, this review may provide new directions for the treatment of obesity to suppress prostate cancer.

Mechanistic Insights into the Link between Obesity and Prostate Cancer

Obesity is a pandemic of increasing worldwide prevalence. There is evidence of an association between obesity and the risk of prostate cancer from observational studies, and different biologic mechanisms have been proposed. The chronic low-level inflammation within the adipose tissue in obesity results in oxidative stress, activation of inflammatory cytokines, deregulation of adipokines signaling, and increased circulating levels of insulin and insulin-like growth factors (IGF). These mechanisms may be involved in epithelial to mesenchymal transformation into a malignant phenotype that promotes invasiveness, aggressiveness, and metastatic potential of prostate cancer. A thorough understanding of these mechanisms may be valuable in the development of effective prostate cancer prevention strategies and treatments. This review provides an overview of these mechanisms.

The Association of Periprostatic Fat and Grade Group Progression in Men with Localized Prostate Cancer on Active Surveillance

PURPOSE

Evidence suggests that visceral fat quantity may be associated with post-prostatectomy outcomes and risk of prostate cancer related death. We evaluated whether increased fat volume, normalized to prostate size, is associated with decreased risk of disease progression.

MATERIALS AND METHODS

Patients enrolled on a prospective active surveillance trial for at least 6 months who had magnetic resonance imaging within 2 years of enrollment were eligible. The surveillance protocol included a standardized followup regimen consisting of biennial prostate specific antigen and examination and yearly biopsy. Clinicopathological characteristics were collected at baseline. Three fat measurements were taken using prostate magnetic resonance imaging, including subcutaneous, linear periprostatic (pubic symphysis to prostate) and volumetrically defined periprostatic. Progression was defined as increase in Gleason grade group. Multivariable Cox proportional hazards models were used to evaluate fat volumes normalized by prostate size (stratified into tertiles).

RESULTS

A total of 175 patients were included in the study. Average age was 62.5 years (SD 7.4) and average prostate specific antigen was 5.4 ng/dl (SD 3.9). Median followup was 42 months (IQR 18-60) and 50 patients (28.6%) had progression. Compared to the lowest tertile, the highest tertile of volumetric periprostatic fat measurement (HR 2.63, 95% CI 1.23-5.60, p=0.01) and linear periprostatic fat measurement (HR 2.30, 95% CI 1.01-5.22, p=0.05) were associated with worsened progression-free survival, while subcutaneous fat measurement (p=0.97) was not. Importantly, the model did not substantively change when accounting for patient body mass index and other factors.

CONCLUSIONS

Increased periprostatic fat volume, normalized to prostate size, may be associated with shortened progression-free survival in men with prostate cancer on active surveillance.

Periprostatic fat thickness measured on MRI correlates with lower urinary tract symptoms, erectile function, and benign prostatic hyperplasia progression

This study investigated the correlation between periprostatic fat thickness (PPFT) measured on magnetic resonance imaging and lower urinary tract symptoms, erectile function, and benign prostatic hyperplasia (BPH) progression. A total of 286 treatment-naive men diagnosed with BPH in our department between March 2017 and February 2019 were included. Patients were divided into two groups according to the median value of PPFT: high (PPFT >4.35 mm) PPFT group and low (PPFT <4.35 mm) PPFT group. After the initial evaluation, all patients received a combination drug treatment of tamsulosin and finasteride for 12 months. Of the 286 enrolled patients, 244 completed the drug treatment course. Patients with high PPFT had larger prostate volume (PV; P = 0.013), higher International Prostate Symptom Score (IPSS; P = 0.008), and lower five-item version of the International Index of Erectile Function (IIEF-5) score (P = 0.002) than those with low PPFT. Both high and low PPFT groups showed significant improvements in PV, maximum flow rate, IPSS, and quality of life score and a decrease of IIEF-5 score after the combination drug treatment. The decrease of IIEF-5 score was more obvious in the high PPFT group than that in the low PPFT group. In addition, more patients in the high PPFT group underwent prostate surgery than those in the low PPFT group. Moreover, Pearson's correlation coefficient analysis indicated that PPFT was positively correlated with age, PV, and IPSS and negatively correlated with IIEF-5 score; however, body mass index was only negatively correlated with IIEF-5 score.

A Novel Prediction Tool Based on Multiparametric Magnetic Resonance Imaging to Determine the Biopsy Strategy for Clinically Significant Prostate Cancer in Patients with PSA Levels Less than 50 ng/ml

PURPOSE

To develop and internally validate nomograms to help choose the optimal biopsy strategy among no biopsy, targeted biopsy (TB) only, or TB plus systematic biopsy (SB).

PATIENTS AND METHODS

This retrospective study included a total of 385 patients who underwent magnetic resonance imaging (MRI)-guided TB and/or SB at our institute after undergoing multiparametric MRI (mpMRI) between 2015 and 2018. We developed models to predict clinically significant prostate cancer (csPCa) based on suspicious lesions from a TB result and based on the whole prostate gland from the results of TB plus SB or SB only. Nomograms were generated using logistic regression and evaluated using receiver-operating characteristic (ROC) curve analysis, calibration curves and decision analysis. The results were validated using ROC curve and calibration on 177 patients from 2018 to 2019 at the same institute.

RESULTS

In the multivariate analyses, prostate-specific antigen level, prostate volume, and the Prostate Imaging Reporting and Data System score were predictors of csPCa in both nomograms. Age was also included in the model for suspicious lesions, while obesity was included in the model for the whole gland. The area under the curve (AUC) in the ROC analyses of the prediction models was 0.755 for suspicious lesions and 0.887 for the whole gland. Both models performed well in the calibration and decision analyses. In the validation cohort, the ROC curve described the AUCs of 0.723 and 0.917 for the nomogram of suspicious lesions and nomogram of the whole gland, respectively. Also, the calibration curve detected low error rates for both models.

CONCLUSION

Nomograms with excellent discriminative ability were developed and validated. These nomograms can be used to select the optimal biopsy strategy for individual patients in the future.

Pre-treatment ratio of periprostatic to subcutaneous fat thickness on MRI is an independent survival predictor in hormone-naïve men with advanced prostate cancer

BACKGROUND

Epidemiological studies have shown an association between obesity and prostate cancer (PCa) aggressiveness. However, little is known about periprostatic fat (PPF) and its relationship with overall fat deposition in PCa. PPF is thought to contribute to PCa growth and migration via secreted factors and induction of chronic inflammation. We investigated if pre-treatment PPF thickness correlates with overall survival (OS).

METHODS

We reviewed 85 hormone-naïve men with advanced PCa who had received androgen deprivation therapy (ADT). PPF thickness was measured by magnetic resonance imaging (MRI) and compared with subcutaneous fat (SCF) thickness as an internal control. Visceral fat (VF) area measured by computed tomography served as an additional control. We evaluated the relationship between laboratory data, pathology results, and obesity parameters and OS.

RESULTS

Median follow-up was 50.6 months. Thirty-six patients died during follow-up. Univariate analysis revealed that nadir PSA titer, Gleason score, N stage, M stage, extent of disease by bone scan grade, hemoglobin, lactate dehydrogenase, alkaline phosphatase, and PPF/SCF ratio were associated with OS. Multivariate analysis revealed that nadir PSA titer, N stage, and PPF/SCF ratio were independent prognostic factors for survival. The 5-year OS in the patients with higher PPF/SCF ratio (≥ 1) and lower PPF/SCF ratio (< 1) was 49.5% and 66.5%, respectively (P = 0.039).

CONCLUSIONS

Pre-treatment ratio of PPF-to-SCF thickness on MRI is an independent predictor of survival in hormone-naïve men with advanced PCa. This could be useful for predicting which patients are more likely to develop castration-resistant PCa.

Periprostatic Fat Thickness on MRI is an Independent Predictor of Time to Castration-resistant Prostate Cancer in Chinese Patients With Newly Diagnosed Prostate Cancer Treated With Androgen Deprivation Therapy

BACKGROUND

The aim of this study was to evaluate the association between periprostatic fat thickness (PPFT) and time to castration-resistant prostate cancer (CRPC) in newly diagnosed patients with prostate cancer (PCa) treated with androgen deprivation therapy (ADT).

PATIENTS AND METHODS

We retrospectively reviewed the medical records of 150 patients with PCa treated with ADT at our hospital between June 2011 and June 2017. PPFT measured on magnetic resonance imaging (MRI) and PPFT/periprostatic fat volume (PPFV) measured on computed tomography (CT) were evaluated. Kaplan-Meier curves and log-rank tests were used to assess significant differences in time to CRPC between the 2 groups (high PPFT vs. low PPFT, determined by PPFT > or < the median value, respectively). Univariable and multivariable Cox regression analyses were employed to identify the potential prognostic factors for survival.

RESULTS

The median value of PPFT measured on MRI was 0.555 cm. PPFT was significantly associated with PPFV measured on CT images (with a correlation coefficient of 0.825; P < .001). A total of 66 patients (44%) progressed to CRPC during the follow-up period. Patients with high PPFT (measured on MRI) showed a significantly shorter PFS than patients with low PPFT. Multivariable Cox analysis demonstrated that T stage, presence of distant metastasis, shorter time to prostate-specific antigen nadir, higher prostate-specific antigen nadir, Gleason score (greater than 4 + 4), and high PPFT were significantly associated with shorter PFS.

CONCLUSIONS

PPFT is significantly associated with PPFV measured on CT images. PPFT measured on MRI is a readily available and significant predictor of time to CRPC in patients with PCa receiving ADT as the primary treatment.

Current status and progress of focal therapy in Asia

PURPOSE OF REVIEW

With the increasing incidence of low-to-intermediate risk of prostate cancer (PCa) by the introduction of prostate-specific antigen (PSA) screening, focal therapy has become one of the promising treatment options in the world. In Asia, same movement are occurring using several technologies including cryoablation, high-intensity focused ultrasound, brachytherapy and irreversible electroporation. However, these are still not common strategies to treat organ-confined PCa. The purpose of this review is to summarize the most updated experience and future direction of focal therapy in Asian countries.

RECENT FINDINGS

The prevalence and diagnosis of PCa are increasing in Asian countries. This increase is related to various factors including the widespread implementation of PSA testing and lifestyle changes to more Westernized diets. With the increasing detection rate of early stage PCa, overdetection and overtreatment are recognized even in Asia. In this setting, accumulating data on multiparametric MRI and MRI-targeted biopsy as well as MRI-transrectal ultrasound (TRUS) fusion biopsy suggest the potential in improving the detection of clinically significant PCa in Asia. Furthermore, targeted focal therapy has emerged as a promising treatment strategy aiming for both providing oncological outcome and maintaining functional preservation in many Asian countries.

SUMMARY

At present, focal therapy is not a current standard choice for the treatment of localized PCa in Asian countries. However, with the increase of localized PCa and patient's preference for less invasive treatment with preservation of organ-function, focal therapy should become a definite treatment option for localized PCa in Asia.

Prostate cancer susceptibility and growth linked to Y chromosome genes

The role of Y chromosome in prostate cancer progression and incidence is not well known. Among the 46 chromosomes, Y chromosome determines the male gender. The Y chromosome is smaller than the X chromosome and contains only 458 genes compared to over 2000 genes found in the X chromosome. The Y chromosome is prone to high mutation rates, created exclusively in sperm cells due to the highly oxidative environment of the testis. Y chromosome harbors epigenetic information, which affects the expression of genes associated with the incidence and progression of prostate cancer. In this review, we focus on Y chromosome related genetic abnormalities, likely to be involved in the development and progression of prostate cancer.

PI-RADS v2 and periprostatic fat measured on multiparametric magnetic resonance imaging can predict upgrading in radical prostatectomy pathology amongst patients with biopsy Gleason score 3 + 3 prostate cancer

PURPOSE

An underestimated biopsy Gleason score 3 + 3 can result in unfounded optimism amongst patients and cause physicians to miss the window for prostate cancer (PCa) cure. This study aims to evaluate the effectiveness of Prostate Imaging Reporting and Data System (PI-RADS) version 2 as well as periprostatic fat (PPF) measured on multiparametric magnetic resonance imaging (mp-MRI) at predicting pathological upgrading amongst patients with biopsy Gleason score 3 + 3 disease.

PATIENTS AND METHODS

A retrospective analysis of 56 patients with biopsy Gleason score 6 PCa who underwent prebiopsy mp-MRI and radical prostatectomy (RP) between November 2013 and March 2018 was conducted. Two radiologists performed PI-RADS v2 score evaluation and different fat measurements on mp-MRI. The associations amongst clinical information, PI-RADS v2 score, different fat parameters and pathologic findings were analyzed. A nomogram predicting upgrading was established based on the results of logistic regression analysis.

RESULTS

A total of 38 (67.9%) patients were upgraded to Gleason ≥7 disease on RP specimens. Prostate-specific antigen density (PSAD) (p < .001), positive core (p < .001), single-core positivity (p = .039), PI-RADS score (p < .001), front PPF area (p = .007) and front-to-total ratio (the ratio of front PPF area to total contour area) (p < .001) were risk factors for upgrading. On multivariate analysis, Epstein criteria (p = .02), PI-RADS score >3 (p = .024), and front-to-total ratio (p = .006) were independent risk factors for pathologic upgrading. The AUC value of the nomogram was 0.893 (95% CI, 0.787-0.999).

CONCLUSION

The combination of PI-RADS v2 and periprostatic fat measured on mp-MRI can help predict pathologic upgrading amongst patients with biopsy Gleason score 3 + 3 PCa.