Prostate Imaging-Reporting and Data System Version 2: Beyond Prostate Cancer Detection

Proton versus photon therapy for high-risk prostate cancer with dose escalation of dominant intraprostatic lesions: a preliminary planning study

Background and purpose

This study aimed to investigate the feasibility of safe-dose escalation to dominant intraprostatic lesions (DILs) and assess the clinical impact using dose-volume (DV) and biological metrics in photon and proton therapy. Biological parameters defined as late grade ≥ 2 gastrointestinal (GI) and genitourinary (GU) derived from planned (D P) and accumulated dose (D A) were utilized.

Materials and methods

In total, 10 patients with high-risk prostate cancer with multiparametric MRI-defined DILs were investigated. Each patient had two plans with a focal boost to the DILs using intensity-modulated proton therapy (IMPT) and volumetric-modulated arc therapy (VMAT). Plans were optimized to obtain DIL coverage while respecting the mandatory organ-at-risk constraints. For the planning evaluation, DV metrics, tumor control probability (TCP) for the DILs and whole prostate excluding the DILs (prostate-DILs), and normal tissue complication probability (NTCP) for the rectum and bladder were calculated. Wilcoxon signed-rank test was used for analyzing TCP and NTCP data.

Results

IMPT achieved a higher Dmean for the DILs compared to VMAT (IMPT: 68.1 GyRBE vs. VMAT: 66.6 Gy, p < 0.05). Intermediate-high rectal and bladder doses were lower for IMPT (p < 0.05), while the high-dose region (V60 Gy) remained comparable. IMPT-TCP for prostate-DIL were higher compared to VMAT (IMPT: 86%; α/β = 3, 94.3%; α/β = 1.5 vs. VMAT: 84.7%; α/β = 3, 93.9%; α/β = 1.5, p < 0.05). Likewise, IMPT obtained a moderately higher DIL TCP (IMPT: 97%; α/β = 3, 99.3%; α/β = 1.5 vs. VMAT: 95.9%; α/β = 3, 98.9%; α/β = 1.5, p < 0.05). Rectal D A-NTCP displayed the highest GI toxicity risk at 5.6%, and IMPT has a lower GI toxicity risk compared to VMAT-predicted Quantec-NTCP (p < 0.05). Bladder D P-NTCP projected a higher GU toxicity than D A-NTCP, with VMAT having the highest risk (p < 0.05).

Conclusion

Dose escalation using IMPT is able to achieve a high TCP for the DILs, with the lowest rectal and bladder DV doses at the intermediate-high-dose range. The reduction in physical dose was translated into a lower NTCP (p < 0.05) for the bladder, although rectal toxicity remained equivalent.

The absolute tumor-capsule contact length in the diagnosis of extraprostatic extension of prostate cancer

BACKGROUND

Extraprostatic extension (EPE) of prostate cancer is associated with a poor prognosis. The broad-based capsule-tumor interface has been recognized as one of the worrisome imaging features in multiparametric prostate MRI (mpMRI). However, there was significant heterogeneity among the measurement method used in prior studies.

OBJECTIVES

This study's objectives were to investigate and compare the accuracy between the curvilinear and linear measurement, find the optimal cut-off contact surface threshold for the diagnosis of EPE, and assess the benefit of the additional contact surface measurement versus visual assessment alone.

METHODS

The status of EPE in mpMRI and the overall PI-RADS were assessed. The tumor's dimensions, the actual tumor-capsule contact length (ACTCL), and the absolute tumor-capsule contact length (ABTCL) were measured. The parameters were analyzed and correlated with the EPE status from prostatectomy specimens.

RESULTS

Ninety-five patients who underwent mpMRI followed by prostatectomy were included in the study. High Gleason score (score 8-9), radiologist's impression of EPE, and PI-RADS 5 were significantly correlated with EPE in surgical specimens (p = 0.014, p < 0.001, and p < 0.001, respectively). Both ACTCL and ABTCL of patients with EPE were significantly higher than those without EPE in all imaging sequences (p < 0.001 to p = 0.003). The ABTCL has higher accuracy than the ACTCL. Dynamic contrast enhancement (DCE) was the most accurate sequence to measure the contact interface. The recommended cut-off value of ABTCL was 15.0 mm, which had a sensitivity and specificity of 75.86% and 72.09%. Multivariable analysis revealed that the ABTCL > 15 mm and the radiologist's impression on visual assessment were the only two independent predictors for the prediction of EPE (p = 0.048 and p = 0.016, respectively). Improvement of diagnostic performance was achieved when the two factors were combined.

CONCLUSION

The ABTCL has better accuracy than the curvilinear measurement in the prediction of EPE. The optimum sequence for the measurement of the contact surface is the DCE. We recommended using 15.0 mm as a cut-off point.

CLINICAL IMPACT

The addition of the ABTCL measurement showed an increase in diagnostic performance. We encourage radiologists to use the capsular contact measurement in addition to their visual assessment to detect EPE in pre-operative MRI.

Diagnostic Performance of PI-RADS v2, Proposed Adjusted PI-RADS v2 and Biparametric Magnetic Resonance Imaging for Prostate Cancer Detection: A Preliminary Study

Purpose: To evaluate the diagnostic performance of PI-RADS v2, proposed adjustments to PI-RADS v2 (PA PI-RADS v2) and biparametric magnetic resonance imaging (MRI) for prostate cancer detection. Methods: A retrospective cohort of 224 patients with suspected prostate cancer was included from January 2016 to November 2018. All the patients underwent a multi-parametric MR scan before biopsy. Two radiologists independently evaluated the MR examinations using PI-RADS v2, PA PI-RADS v2, and a biparametric MRI protocol, respectively. Receiver operating characteristic (ROC) curves for the three different protocols were drawn. Results: In total, 90 out of 224 cases (40.18%) were pathologically diagnosed as prostate cancer. The area under the ROC curves (AUC) for diagnosing prostate cancers by biparametric MRI, PI-RADS v2, and PA PI-RADS v2 were 0.938, 0.935, and 0.934, respectively. For cancers in the peripheral zone (PZ), the diagnostic sensitivity was 97.1% for PI-RADS v2/PA PI-RADS v2 and 96.2% for biparametric MRI. Moreover, the specificity was 84.0% for biparametric MRI and 58.0% for PI-RADS v2/PA PI-RADS v2. For cancers in the transition zone (TZ), the diagnostic sensitivity was 93.4% for PA PI-RADS v2 and 88.2% for biparametric MRI/PI-RADS v2. Furthermore, the specificity was 95.4% for biparametric MRI/PI-RADS v2 and 78.0% for PA PI-RADS v2. Conclusions: The overall diagnostic performance of the three protocols showed minimal differences. For lesions assessed as being category 3 using the biparametric MRI protocol, PI-RADS v2, or PA PI-RADS v2, it was thought prostate cancer detection could be improved. Attention should be paid to false positive results when PI-RADS v2 or PA PI-RADS v2 are used.

MRI index lesion radiomics and machine learning for detection of extraprostatic extension of disease: a multicenter study

Objectives

To build a machine learning (ML) model to detect extraprostatic extension (EPE) of prostate cancer (PCa), based on radiomics features extracted from prostate MRI index lesions.

Methods

Consecutive MRI exams of patients undergoing radical prostatectomy for PCa were retrospectively collected from three institutions. Axial T2-weighted and apparent diffusion coefficient map images were annotated to obtain index lesion volumes of interest for radiomics feature extraction. Data from one institution was used for training, feature selection (using reproducibility, variance and pairwise correlation analyses, and a correlation-based subset evaluator), and tuning of a support vector machine (SVM) algorithm, with stratified 10-fold cross-validation. The model was tested on the two remaining institutions’ data and compared with a baseline reference and expert radiologist assessment of EPE.

Results

In total, 193 patients were included. From an initial dataset of 2436 features, 2287 were excluded due to either poor stability, low variance, or high collinearity. Among the remaining, 14 features were used to train the ML model, which reached an overall accuracy of 83% in the training set. In the two external test sets, the SVM achieved an accuracy of 79% and 74% respectively, not statistically different from that of the radiologist (81–83%, p = 0.39–1) and outperforming the baseline reference (p = 0.001–0.02).

Conclusions

A ML model solely based on radiomics features demonstrated high accuracy for EPE detection and good generalizability in a multicenter setting. Paired to qualitative EPE assessment, this approach could aid radiologists in this challenging task.

Key Points

• Predicting the presence of EPE in prostate cancer patients is a challenging task for radiologists.

• A support vector machine algorithm achieved high diagnostic accuracy for EPE detection, with good generalizability when tested on multiple external datasets.

• The performance of the algorithm was not significantly different from that of an experienced radiologist.

Supplementary Information

The online version contains supplementary material available at 10.1007/s00330-021-07856-3.

Risk Factors for Pathologically Confirmed Lymph Nodes Metastasis in Patients With Clinical T2N0M0 Stage Prostate Cancer

Objective

To explore the risk factors for postoperatively pathological lymph node metastasis in patients with clinical T2N0M0 stage prostate cancer (PCa).

Methods

We retrospectively analyzed clinicopathological data of 316 patients with clinical T2 stage PCa and preoperative negative lymph nodes [LN(−)] indicated by imaging (cT2N0M0) between January 2014 and May 2019. Multivariate logistic regression analysis was performed to determine risk factors for postoperatively pathological pLN(+) in patients with cT2N0M0 stage PCa. Spearman correlation analysis was used to explore the relationship between tumor burden and Prostate Imaging Reporting and Data System version 2 (PI-RADS v2) score.

Results

A total of 45 patients (14.2%) were confirmed by postoperative pathology to have LN metastasis. Univariate analysis indicated that total prostate-specific antigen (tPSA), PI-RADS v2 score, postoperative Gleason grade group (GGG), intraductal carcinoma of the prostate (IDC-P), clinical T2 substaging, and postoperative pathological tumor burden were risk factors for pLN(+) in all patients. Multivariate analysis showed that tPSA and postoperative GGG were risk factors for pLN(+) in all patients. Univariate analysis revealed that tPSA, PIRADS v2 score, clinical T2 substaging, IDC-P, postoperative pathological tumor burden, and postoperative GGG were risk factors for pLN(+) in patients with GGG ≥ 3. Multivariate analysis suggested that tPSA, PI-RADS v2 score, clinical T2 substaging, postoperative pathological tumor burden, and GGG were risk factors for pLN (+) in patients with GGG ≥ 3. Spearman correlation analysis showed that PI-RADS v2 score was positively correlated with clinical T2 substaging and postoperative pathological tumor burden.

Conclusion

There was a high risk of LN metastasis in patients with cT2 PCa if they had high preoperative tPSA or high postoperative GGG. Patients with cT2 PCa and GGG ≥ 3 had a high risk of LN metastasis if they had high PI-RADS v2 score, high preoperative clinical stage or high postoperative pathological tumor burden. PI-RADS v2 score predicted tumor burden well in patients with GGG ≥ 3.

Effect of observation size and apparent diffusion coefficient (ADC) value in PI-RADS v2.1 assessment category 4 and 5 observations compared to adverse pathological outcomes

OBJECTIVE

To compare observation size and apparent diffusion coefficient (ADC) values in Prostate Imaging Reporting and Data System (PI-RADS) v2.1 category 4 and 5 observations to adverse pathological features.

MATERIALS AND METHODS

With institutional review board approval, 267 consecutive men with 3-T MRI before radical prostatectomy (RP) between 2012 and 2018 were evaluated by two blinded radiologists who assigned PI-RADS v2.1 scores. Discrepancies were resolved by consensus. A third blinded radiologist measured observation size and ADC (ADC.mean, ADC.min [lowest ADC within an observation], ADC.ratio [ADC.mean/ADC.peripheral zone {PZ}]). Size and ADC were compared to pathological stage and Gleason score (GS) using t tests, ANOVA, Pearson correlation, and receiver operating characteristic (ROC) analysis.

RESULTS

Consensus review identified 267 true positive category 4 and 5 observations representing 83.1% (222/267) PZ and 16.9% (45/267) transition zone (TZ) tumors. Inter-observer agreement for PI-RADS v2.1 scoring was moderate (K = 0.45). Size was associated with extra-prostatic extension (EPE) (19 ± 8 versus 14 ± 6 mm, p < 0.001) and seminal vesicle invasion (SVI) (24 ± 9 versus 16 ± 7 mm, p < 0.001). Size ≥ 15 mm optimized the accuracy for EPE with area under the ROC curve (AUC) and sensitivity/specificity of 0.68 (CI 0.62-0.75) and 63.2%/65.6%. Size ≥ 19 mm optimized the accuracy for SVI with AUC/sensitivity/specificity of 0.75 (CI 0.66-0.83)/69.4%/70.6%. ADC metrics were not associated with pathological stage. Larger observation size (p = 0.032), lower ADC.min (p = 0.010), and lower ADC.ratio (p = 0.010) were associated with higher GS. Size correlated better to higher Gleason scores (p = 0.002) compared to ADC metrics (p = 0.09-0.11).

CONCLUSION

Among PI-RADS v2.1 category 4 and 5 observations, size was associated with higher pathological stage whereas ADC metrics were not. Size, ADC.minimum, and ADC.ratio differed in tumors stratified by Gleason score.

KEY POINTS

• Among PI-RADS category 4 and 5 observations, size but not ADC can differentiate between tumors by pathological stage. • An observation size threshold of 15 mm and 19 mm optimized the accuracy for diagnosis of extra-prostatic extension and seminal vesicle invasion. • Among PI-RADS category 4 and 5 observations, size, ADC.minimum, and ADC.ratio differed comparing tumors by Gleason score.

Preoperative PI-RADS Version 2 scores helps improve accuracy of clinical nomograms for predicting pelvic lymph node metastasis at radical prostatectomy

BACKGROUND

Lymph node invasion (LNI) is a strong adverse prognostic factor in prostate cancer (PCa). The purpose of this study was to evaluate the role of Prostate Imaging Reporting and Data System version 2 (PI-RADSv2) scores for estimating the risk of LN metastasis. The study also aimed to investigate the additional value of PI-RADSv2 scores when used in combination with clinical nomograms for the prediction of LNI in patients with PCa.

METHODS

We retrospectively identified 308 patients who underwent multiparametric magnetic resonance imaging (mpMRI) and RP with pelvic lymph node dissection (PLND). Clinicopathological parameters and PI-RADSv2 scores were assessed. Univariate and multivariate logistic analyses were performed. The area under the receiver operating characteristic curves (AUCs) and decision curve analysis (DCA) were generated for assessing the incremental value of PI-RADSv2 scores combined with the Briganti and Memorial Sloan Kettering Cancer Center (MSKCC) nomograms.

RESULTS

Overall, 20 (6.5%) patients had LNI. At univariate analysis, all clinicopathological characteristics and PI-RADSv2 scores were significantly associated to LNI (p < 0.04). However, multivariate analysis revealed that only PI-RADSv2 scores and percentage of positive cores were independently significant (p ≤ 0.006). The PI-RADSv2 score was the most accurate predictor (AUC, 80.2%). The threshold of PI-RADSv2 score was 5, which provided high sensitivity (18/20, 90.0%) and negative predictive value (203/205, 99.0%). When PI-RADSv2 scores were combined with Briganti and MSKCC nomograms, the AUC value increased from 75.1 to 86.3% and from 79.2 to 87.9%, respectively (p ≤ 0.001). The DCA also demonstrated that the two nomograms plus PI-RADSv2 scores improved clinical risk prediction of LNI.

CONCLUSIONS

The patients with a PI-RADSv2 score <5 were associated with a very low risk of LNI in PCa. Preoperative PI-RADSv2 scores could help improve the accuracy of clinical nomograms for predicting pelvic LN metastasis at radical prostatectomy.

Prostate Imaging Reporting and Data System in prostate cancer staging and planning for radical prostatectomy

Introduction

The Prostate Imaging Reporting and Data System (PI-RADS) was mainly developed for the purposes of prostate cancer (PCa) detection. However, its widespread use suggests that it may play a role in a preoperative workup prior to endoscopic radical prostatectomy (ERP).

Aim

To evaluate the prognostic value of PI-RADS in predicting extraprostatic extension (EPE) and its influence on surgical planning of ERP.

Material and methods

The analysis involved data of 154 consecutive prostate cancer patients, in whom multiparametric 3.0T magnetic resonance imaging (mpMRI) was performed before ERP. Standard descriptive assessment of mpMRI images was compared with the PI-RADS system with respect to prostate cancer staging and subsequent potential surgical template adjustment.

Results

PI-RADS significantly outperformed the standard way of mpMRI reporting in staging (AUC = 0.615 vs. 0.552, p = 0.036) with PI-RADS 5 established as the best threshold. After reevaluation of imaging, the initial surgical plan was modified based on mpMRI in terms of feasibility and extent of neurovascular bundle preservation during ERP on 96 (31.2%) sides, while on the remaining 212 (68.8%) sides the templates were left unchanged. Decisions based on mpMRI were not associated with increased risk of a positive surgical margin (PSM).

Conclusions

The PI-RADS outperforms the standard staging method using mpMRI and may assist the decision-making process regarding the extent of resection during ERP without increasing the risk of PSM.