Clinical significance of subtypes of Gleason pattern 4 prostate cancer

Lower urinary tract symptoms in elderly men: Considerations for prostate cancer testing

PURPOSE

Both lower urinary tract symptoms (LUTS) and prostate cancer (PCa) are common in elderly men. While LUTS are generally due to a benign etiology, they may provoke an evaluation with prostate-specific antigen (PSA), which can lead to a cascade of further testing and possible overdiagnosis in patients with competing risks. There is limited patient and provider understanding of the relationship between LUTS and PCa risk, and a lack of clarity in how to evaluate these men to balance appropriate diagnosis of aggressive PCa with avoidance of overdiagnosis.

METHODS

A literature review was performed using keywords to query the electronic database PubMed. All articles published before November 2023 were screened by title and abstract for articles relevant to our subject.

RESULTS

Epidemiological studies suggest that LUTS and PCa are largely independent in elderly men. The best available tools to assess PCa risk include PSA permutations, novel biomarkers, and imaging, but there are limitations in older men based on lack of validation in the elderly and unclear applicability of traditional definitions of "clinically significant" disease. We present a three-tiered approach to evaluating these patients.

CONCLUSION

Elderly men commonly have LUTS as well as a high likelihood of indolent PCa. A systematic and shared decision-making-based approach can help to balance objectives of appropriate detection of phenotypically dangerous disease and avoidance of over-testing and overdiagnosis.

Contemporary Diagnostic Reporting for Prostatic Adenocarcinoma: Morphologic Aspects, Molecular Correlates, and Management Perspectives

The diagnosis and reporting of prostatic adenocarcinoma have evolved from the classic framework promulgated by Dr Donald Gleason in the 1960s into a complex and nuanced system of grading and reporting that nonetheless retains the essence of his remarkable observations. The criteria for the "Gleason patterns" originally proposed have been continually refined by consensuses in the field, and Gleason scores have been stratified into a patient-friendly set of prognostically validated and widely adopted Grade Groups. One product of this successful grading approach has been the opportunity for pathologists to report diagnoses that signal carefully personalized management, placing the surgical pathologist's interpretation at the center of patient care. At one end of the continuum of disease aggressiveness, personalized diagnostic care means to sub-stratify patients with more indolent disease for active surveillance, while at the other end of the continuum, reporting histologic markers signaling aggression allows sub-stratification of clinically significant disease. Whether contemporary reporting parameters represent deeper nuances of more established ones (eg, new criteria and/or quantitation of Gleason patterns 4 and 5) or represent additional features reported alongside grade (intraductal carcinoma, cribriform patterns of carcinoma), assessment and grading have become more complex and demanding. Herein, we explore these newer reporting parameters, highlighting the state of knowledge regarding morphologic, molecular, and management aspects. Emphasis is made on the increasing value and stakes of histopathologists' interpretations and reporting into current clinical risk stratification and treatment guidelines.

Transcriptomic Heterogeneity of Expansile Cribriform and Other Gleason Pattern 4 Prostate Cancer Subtypes

BACKGROUND

Prostate cancers featuring an expansile cribriform (EC) pattern are associated with worse clinical outcomes following radical prostatectomy (RP). However, studies of the genomic characteristics of Gleason pattern 4 subtypes are limited.

OBJECTIVE

To explore transcriptomic characteristics and heterogeneity within Gleason pattern 4 subtypes (fused/poorly formed, glomeruloid, small cribriform, EC/intraductal carcinoma [IDC]) and the association with biochemical recurrence (BCR)-free survival.

DESIGN, SETTING, AND PARTICIPANTS

This was a retrospective cohort study including 165 men with grade group 2-4 prostate cancer who underwent RP at a single academic institution (2016-2020) and Decipher testing of the RP specimen. Patients with Gleason pattern 5 were excluded. IDC and EC patterns were grouped. Median follow-up was 2.5 yr after RP for patients without BCR.

OUTCOMES MEASUREMENTS AND STATISTICAL ANALYSIS

Prompted by heterogeneity within pattern 4 subtypes identified via exploratory analyses, we investigated transcriptomic consensus clusters using partitioning around medoids and hallmark gene set scores. The primary clinical outcome was BCR, defined as two consecutive prostate-specific antigen measurements >0.2 ng/ml at least 8 wk after RP, or any additional treatment. Multivariable Cox proportional-hazards models were used to determine factors associated with BCR-free survival.

RESULTS AND LIMITATIONS

In this cohort, 99/165 patients (60%) had EC and 67 experienced BCR. Exploratory analyses and clustering demonstrated transcriptomic heterogeneity within each Gleason pattern 4 subtype. In the multivariable model controlled for pattern 4 subtype, margin status, Cancer of the Prostate Risk Assessment Post-Surgical score, and Decipher score, a newly identified steroid hormone-driven cluster (hazard ratio 2.35 95% confidence interval 1.01-5.47) was associated with worse BCR-free survival. The study is limited by intermediate follow-up, no validation cohort, and lack of accounting for intratumoral and intraprostatic heterogeneity.

CONCLUSIONS

Transcriptomic heterogeneity was present within and across each Gleason pattern 4 subtype, demonstrating there is additional biologic diversity not captured by histologic subtypes. This heterogeneity can be used to develop novel signatures and to classify transcriptomic subtypes, which may help in refining risk stratification following RP to further guide decision-making on adjuvant and salvage treatments.

PATIENT SUMMARY

We studied prostatectomy specimens and found that tumors with similar microscopic appearance can have genetic differences that may help to predict outcomes after prostatectomy for prostate cancer. Our results demonstrate that further gene expression analysis of prostate cancer subtypes may improve risk stratification after prostatectomy. Future studies are needed to develop novel gene expression signatures and validate these findings in independent sets of patients.

External Validation of a Digital Pathology-based Multimodal Artificial Intelligence Architecture in the NRG/RTOG 9902 Phase 3 Trial

BACKGROUND

Accurate risk stratification is critical to guide management decisions in localized prostate cancer (PCa). Previously, we had developed and validated a multimodal artificial intelligence (MMAI) model generated from digital histopathology and clinical features. Here, we externally validate this model on men with high-risk or locally advanced PCa treated and followed as part of a phase 3 randomized control trial.

OBJECTIVE

To externally validate the MMAI model on men with high-risk or locally advanced PCa treated and followed as part of a phase 3 randomized control trial.

DESIGN, SETTING, AND PARTICIPANTS

Our validation cohort included 318 localized high-risk PCa patients from NRG/RTOG 9902 with available histopathology (337 [85%] of the 397 patients enrolled into the trial had available slides, of which 19 [5.6%] failed due to poor image quality).

OUTCOME MEASUREMENTS AND STATISTICAL ANALYSIS

Two previously locked prognostic MMAI models were validated for their intended endpoint: distant metastasis (DM) and PCa-specific mortality (PCSM). Individual clinical factors and the number of National Comprehensive Cancer Network (NCCN) high-risk features served as comparators. Subdistribution hazard ratio (sHR) was reported per standard deviation increase of the score with corresponding 95% confidence interval (CI) using Fine-Gray or Cox proportional hazards models.

RESULTS AND LIMITATIONS

The DM and PCSM MMAI algorithms were significantly and independently associated with the risk of DM (sHR [95% CI] = 2.33 [1.60-3.38], p < 0.001) and PCSM, respectively (sHR [95% CI] = 3.54 [2.38-5.28], p < 0.001) when compared against other prognostic clinical factors and NCCN high-risk features. The lower 75% of patients by DM MMAI had estimated 5- and 10-yr DM rates of 4% and 7%, and the highest quartile had average 5- and 10-yr DM rates of 19% and 32%, respectively (p < 0.001). Similar results were observed for the PCSM MMAI algorithm.

CONCLUSIONS

We externally validated the prognostic ability of MMAI models previously developed among men with localized high-risk disease. MMAI prognostic models further risk stratify beyond the clinical and pathological variables for DM and PCSM in a population of men already at a high risk for disease progression. This study provides evidence for consistent validation of our deep learning MMAI models to improve prognostication and enable more informed decision-making for patient care.

PATIENT SUMMARY

This paper presents a novel approach using images from pathology slides along with clinical variables to validate artificial intelligence (computer-generated) prognostic models. When implemented, clinicians can offer a more personalized and tailored prognostic discussion for men with localized prostate cancer.

Incidence and Predicting Factors of Histopathological Features at Robot-Assisted Radical Prostatectomy in the mpMRI Era: Results of a Single Tertiary Referral Center

Background and Objectives: To describe the predictors of cribriform variant status and perineural invasion (PNI) in robot-assisted radical prostatectomy (RARP) histology. To define the rates of upgrading between biopsy specimens and final histology and their possible predictive factors in prostate cancer (PCa) patients undergoing RARP. Material and Methods: Within our institutional database, 265 PCa patients who underwent prostate biopsies and consecutive RARP at our center were enrolled (2018-2022). In the overall population, two independent multivariable logistic regression models (LRMs) predicting the presence of PNI or cribriform variant status at RARP were performed. In low- and intermediate-risk PCa patients according to D'Amico risk classification, three independent multivariable LRMs were fitted to predict upgrading. Results: Of all, 30.9% were low-risk, 18.9% were intermediate-risk and 50.2% were high-risk PCa patients. In the overall population, the rates of the cribriform variant and PNI at RARP were 55.8% and 71.1%, respectively. After multivariable LRMs predicting PNI, total tumor length in biopsy cores (>24 mm [OR: 2.37, p-value = 0.03], relative to <24 mm) was an independent predictor. After multivariable LRMs predicting cribriform variant status, PIRADS (3 [OR:15.37], 4 [OR: 13.57] or 5 [OR: 16.51] relative to PIRADS 2, all p = 0.01) and total tumor length in biopsy cores (>24 mm [OR: 2.47, p = 0.01], relative to <24 mm) were independent predicting factors. In low- and intermediate-risk PCa patients, the rate of upgrading was 74.4% and 78.0%, respectively. After multivariable LRMs predicting upgrading, PIRADS (PIRADS 3 [OR: 7.01], 4 [OR: 16.98] or 5 [OR: 20.96] relative to PIRADS 2, all p = 0.01) was an independent predicting factor. Conclusions: RARP represents a tailored and risk-adapted treatment strategy for PCa patients. The indication of RP progressively migrates to high-risk PCa after a pre-operative assessment. Specifically, the PIRADS score at mpMRI should guide the decision-making process of urologists for PCa patients.

Convolutional Neural Network Quantification of Gleason Pattern 4 and Association with Biochemical Recurrence in Intermediate Grade Prostate Tumors

Differential classification of prostate cancer (CaP) grade group (GG) 2 and 3 tumors remains challenging, likely due to the subjective quantification of percentage of Gleason pattern 4 (%GP4). Artificial intelligence assessment of %GP4 may improve its accuracy and reproducibility and provide information for prognosis prediction. To investigate this potential, a convolutional neural network (CNN) model was trained to objectively identify and quantify Gleason pattern (GP) 3 and 4 areas, estimate %GP4, and assess whether CNN-assessed %GP4 is associated with biochemical recurrence (BCR) risk in intermediate risk GG 2 and 3 tumors. The study was conducted in a radical prostatectomy cohort (1999-2012) of African American men from the Henry Ford Health System (Detroit, Michigan). A CNN model that could discriminate four tissue types (stroma, benign glands, GP3 glands, and GP4 glands) was developed using histopathologic images containing GG 1 (n=45) and 4 (n=20) tumor foci. The CNN model was applied to GG 2 (n=153) and 3 (n=62) for %GP4 estimation, and Cox proportional hazard modeling was used to assess the association of %GP4 and BCR, accounting for other clinicopathologic features including GG. The CNN model achieved an overall accuracy of 86% in distinguishing the four tissue types. Further, CNN-assessed %GP4 was significantly higher in GG 3 compared with GG 2 tumors (p=7.2*10^-11). %GP4 was associated with an increased risk of BCR (adjusted HR=1.09 per 10% increase in %GP4, p=0.010) in GG 2 and 3 tumors. Within GG 2 tumors specifically, %GP4 was more strongly associated with BCR (adjusted HR=1.12, p=0.006). Our findings demonstrate the feasibility of CNN-assessed %GP4 estimation, which is associated with BCR risk. This objective approach could be added to the standard pathological assessment for patients with GG 2 and 3 tumors and act as a surrogate for specialist genitourinary pathologist evaluation when such consultation is not available.

Impact of cribriform pattern 4 and intraductal prostatic carcinoma on National Comprehensive Cancer Network (NCCN) and Cancer of Prostate Risk Assessment (CAPRA) patient stratification

Pretreatment classification tools are used in prostate cancer to inform patient management. The effect of cribriform pattern 4 (CC) and intraductal carcinoma (IDC) on such nomograms is still underexplored. We analyzed the Cancer of Prostate Risk Assessment (CAPRA) and National Comprehensive Cancer Network (NCCN) risk scores in cases with and without CC/IDC to assess impact on biochemical recurrence (BCR) and metastases/death of prostate cancer (event free survival-EFS) after prostatectomy. A matched biopsy- prostatectomy cohort (2010-2017) was reviewed for CC/IDC. CAPRA and NCCN scores were calculated. CAPRA score 0-2 were deemed "low", 3-5 "intermediate" and 6-10 "high". NCCN scores 1-2 "very low/low", 3 "favorable intermediate", 4 "unfavorable intermediate", 5-6 "high/very high". Cases were stratified by presence of CC/IDC. BCR and EFS probabilities were estimated using the Kaplan-Meier method. Prognostic performance was evaluated using log-rank tests and Harrell's concordance index. 612 patients with mean age 63.1 years were included with mean follow up of 5.3 (range 0-10.8) years. CC/IDC was noted in 159/612 (26%) biopsies. There were 101 (17%) BCR and 36 (6%) events. CAPRA discriminated three distinct risk categories for BCR (p < 0.001) while only high risk separated significantly for EFS (p < 0.001). NCCN distinguished two prognostic groups for BCR (p < 0.0001) and three for EFS (p < 0.0001). Addition of CC/IDC to CAPRA impacted scores 3-5 for BCR and scores 3-5 and 6-10 for EFS and improved the overall concordance index (BCR: 0.66 vs. 0.71; EFS: 0.74 vs. 0.80). Addition of CC/IDC to NCCN impacted scores 4 and 5-6 and also improved the concordance index for BCR (0.62 vs. 0.68). Regarding EFS, NCCN scores 4 and 5-6 demonstrated markedly different outcomes with the addition of CC/IDC. The CAPRA nomogram allows better outcome stratification than NCCN. Addition of CC/IDC status particularly improves patient stratification for CAPRA scores 3-5, 6-10, and for NCCN scores 4 and 5-6.

Diffusion Restriction Comparison between Gleason 4 Fused Glands and Cribriform Glands within Patient Using Whole-Mount Prostate Pathology as Ground Truth

The presence and extent of cribriform patterned Gleason 4 (G4) glands are associated with poor prognosis following radical prostatectomy. This study used whole-mount prostate histology and multiparametric magnetic resonance imaging (MP-MRI) to evaluate diffusion differences in G4 gland morphology. Fourty-eight patients underwent MP-MRI prior to prostatectomy, of whom 22 patients had regions of both G4 cribriform glands and G4 fused glands (G4CG and G4FG, respectively). After surgery, the prostate was sliced using custom, patient-specific 3D-printed slicing jigs modeled according to the T2-weighted MR image, processed, and embedded in paraffin. Whole-mount hematoxylin and eosin-stained slides were annotated by our urologic pathologist and digitally contoured to differentiate the lumen, epithelium, and stroma. Digitized slides were co-registered to the T2-weighted MRI scan. Linear mixed models were fitted to the MP-MRI data to consider the different hierarchical structures at the patient and slide level. We found that Gleason 4 cribriform glands were more diffusion-restricted than fused glands.

The impact of multiparametric MRI features to identify the presence of prevalent cribriform pattern in the peripheral zone tumors

PURPOSE

To assess the role of the multiparametric Magnetic Resonance Imaging (mpMRI) in predicting the cribriform pattern in both the peripheral and transition zones (PZ and TZ) clinically significant prostate cancers (csPCas).

MATERIAL AND METHODS

We retrospectively evaluated 150 patients who underwent radical prostatectomy for csPCa and preoperative mpMRI. Patients with negative (n = 25) and positive (n = 125) mpMRI, stratified according to the presence of prevalent cribriform pattern (PCP, ≥ 50%) and non-PCP (< 50%) at specimen, were included. Difference between the two groups were evaluated. Multivariate logistic regression was used to identify predictors of PCP among mpMRI parameters. The receiver operating characteristic (ROC) analysis was performed to evaluate the area under the curve (AUC) of apparent diffusion coefficient (ADC) and ADC ratio in detecting lesions harboring PCP.

RESULTS

Considering 135 positive lesions at the mpMRI, 30 (22.2%) and 105 (77.8%) harbored PCP and non-PCP PCa. The PCP lesions had more frequently nodular morphology (83.3% vs 62.9%; p = 0.04) and significantly lower mean ADC value (0.87 ± 0.16 vs 0.95 ± 0.18; p = 0.03) and ADC ratio (0.52 ± 0.09 vs 0.60 ± 0.14; p = 0.003) when compared with non-PCP lesions. At univariate and multivariate analyses, mean ADC and ADC ratio resulted as independent predictors of the presence of the PCP of the PZ tumors(OR: 0.025; p = 0.03 and OR: 0.001; p = 0.004, respectively). At the ROC analysis, the AUC of mean ADC and ADC ratio to predict the presence of PCP in patients with PZ suspicious lesion at the mpMRI were 0.69 (95% CI 0.56-0.81P, p = 0.003) and 0.72 (95% CI 0.62-0.82P, p = 0.001), respectively.

CONCLUSIONS

The mpMRI may correctly identify PCP tumors of the PZ and the mean ADC value and ADC ratio can predict the presence of the cribriform pattern in the PCa.

Clinicopathological Analysis of the ISUP Grade Group And Other Parameters in Prostate Cancer: Elucidation of Mutual Impact of the Various Parameters

Background

Prostate cancer has become increasingly common worldwide. Although Grade group (GG) is widely accepted as an indicator of prostate cancer grade, there are malignancies that cannot be defined by GG alone. Moreover, the relationship between GG and other parameters remains unclear. Herein, we aimed to explore the biological characteristics of prostate cancer.

Methods

This study included 299 radical prostatectomy cases. The Chi-square test and analysis of variance were used to analyze the association of GG with binary and continuous variables. We then conducted morphological analyses. Multivariate analyses were performed to extract the data on risk factors for biochemical recurrence (BCR) and lymph node metastasis.

Results

The lymphatic, venous, perineural, and seminal vesicle invasion rates were 37/299 (12.4%), 25/299 (8.4%), 280/299 (93.6%), and 23/299 (7.7%), respectively. The extraprostatic extension (EPE), positive surgical margin, tertiary Gleason pattern 5, intraductal carcinoma of the prostate gland, and lymph node metastasis rates were 89/299 (29.8%), 106/299 (35.5%), 33/260 (12.7%), 56/299 (18.7%), and 23/299 (7.7%), respectively. As GG increased, various parameters became easier to visualize; however, there were differences between the parameters. Postoperative BCR was observed in 31/242 (12.8%) cases without preoperative hormone therapy; GG2, GG3, GG4, and GG5 accounted for 4, 7, 7, and 13 cases, respectively. Multivariate analyses revealed that GG and tumor diameter were significant risk factors for early BCR, whereas lymphatic invasion, EPE, and seminal vesicle invasion were significant risk factors for lymph node metastasis. For BCR, the odds ratios (ORs) for GG and tumor diameter were 2.253 (95% confidence interval (CI]): 1.297–3.912; P=0.004) and 1.074 (95% CI: 1.011–1.142; P=0.022), respectively. For lymph node metastasis, ORs for the presence of lymphatic invasion, EPE, and seminal vesicle invasion were 7.425 (95% CI: 1.688–22.583; P=0.004), 4.391 (95% CI: 1.037–18.589; P=0.044), and 5.755 (95% CI: 1.308–25.316; P=0.021), respectively.

Conclusions

We summarized various parameters correlating with each GG. Through multivariate analyses, we established the independent risk factors for early BCR and lymph node metastasis. In addition to GG, other important indices of malignancy were determined and weighted to provide a basis for future investigations.

Single-cell Spatial Proteomic Revelations on the Multiparametric MRI Heterogeneity of Clinically Significant Prostate Cancer

PURPOSE

Multiparametric MRI (mpMRI) has become an indispensable radiographic tool in diagnosing prostate cancer. However, mpMRI fails to visualize approximately 15% of clinically significant prostate cancer (csPCa). The molecular, cellular, and spatial underpinnings of such radiographic heterogeneity in csPCa are unclear.

EXPERIMENTAL DESIGN

We examined tumor tissues from clinically matched patients with mpMRI-invisible and mpMRI-visible csPCa who underwent radical prostatectomy. Multiplex immunofluorescence single-cell spatial imaging and gene expression profiling were performed. Artificial intelligence-based analytic algorithms were developed to examine the tumor ecosystem and integrate with corresponding transcriptomics.

RESULTS

More complex and compact epithelial tumor architectures were found in mpMRI-visible than in mpMRI-invisible prostate cancer tumors. In contrast, similar stromal patterns were detected between mpMRI-invisible prostate cancer and normal prostate tissues. Furthermore, quantification of immune cell composition and tumor-immune interactions demonstrated a lack of immune cell infiltration in the malignant but not in the adjacent nonmalignant tissue compartments, irrespective of mpMRI visibility. No significant difference in immune profiles was detected between mpMRI-visible and mpMRI-invisible prostate cancer within our patient cohort, whereas expression profiling identified a 24-gene stromal signature enriched in mpMRI-invisible prostate cancer. Prostate cancer with strong stromal signature exhibited a favorable survival outcome within The Cancer Genome Atlas prostate cancer cohort. Notably, five recurrences in the 8 mpMRI-visible patients with csPCa and no recurrence in the 8 clinically matched patients with mpMRI-invisible csPCa occurred during the 5-year follow-up post-prostatectomy.

CONCLUSIONS

Our study identified distinct molecular, cellular, and structural characteristics associated with mpMRI-visible csPCa, whereas mpMRI-invisible tumors were similar to normal prostate tissue, likely contributing to mpMRI invisibility.

The Impact of Histopathological Features of Prostate Cancerous Lesions on Multiparametric Magnetic Resonance Imaging Findings using PI-RADS Version 2

OBJECTIVES

To determine the square measure threshold of prostate cancer lesions in pathological specimens showing PI-RADS categories 3 to 5, and to identify the pathological characteristics of cancerous lesions over the threshold.

METHODS

Cancer foci detected in horizontal sections of specimens were defined as pathological cancerous lesions, in which square measure, lesion location (peripheral or transition zone), Gleason pattern (GP), GP4-5 component percentages, and GP 4 subtypes were assessed. A receiver operating characteristic curve was used to determine the threshold of the square measure of pathological specimens that distinguishes between lesions of PI-RADS categories 1 and 2 and those of 3 to 5. Univariable and multivariable analyses were performed to determine the histopathological features associated with PI-RADS categories 3 to 5.

RESULTS

A total of 100 consecutive patients underwent multiparametric magnetic resonance imaging before robotic-assisted laparoscopic prostatectomy. A total of 1366 pathological cancerous lesions were detected, 217 of which were classified as PI-RADS categories 3 to 5. A square measure of 40 mm² on pathological specimens was the threshold for PI-RADS categories 3 to 5. Of the 415 lesions that were over 40 mm², 211 lesions exhibited PI-RADS categories 1, 2 and 204 lesions exhibited PI-RADS categories 3 to 5. Multiple logistic regression analysis showed that square measure, fused glands, and cribriform glands were independently associated with PI-RADS categories 3 to 5.

CONCLUSION

Cancerous lesions over 40 mm² showing PI-RADS categories 3 to 5 are associated with square measure, fused glands, and cribriform glands.

Multi-resolution open-top light-sheet microscopy to enable efficient 3D pathology workflows

Open-top light-sheet (OTLS) microscopes have been developed for user-friendly and versatile high-throughput 3D microscopy of thick specimens. As with all imaging modalities, spatial resolution trades off with imaging and analysis times. A hierarchical multi-scale imaging workflow would therefore be of value for many volumetric microscopy applications. We describe a compact multi-resolution OTLS microscope, enabled by a novel solid immersion meniscus lens (SIMlens), which allows users to rapidly transition between air-based objectives for low- and high-resolution 3D imaging. We demonstrate the utility of this system by showcasing an efficient 3D analysis workflow for a diagnostic pathology application.

Going deeper through the Gleason scoring scale: An automatic end-to-end system for histology prostate grading and cribriform pattern detection

BACKGROUND AND OBJECTIVE

Prostate cancer is one of the most common diseases affecting men worldwide. The Gleason scoring system is the primary diagnostic and prognostic tool for prostate cancer. Furthermore, recent reports indicate that the presence of patterns of the Gleason scale such as the cribriform pattern may also correlate with a worse prognosis compared to other patterns belonging to the Gleason grade 4. Current clinical guidelines have indicated the convenience of highlight its presence during the analysis of biopsies. All these requirements suppose a great workload for the pathologist during the analysis of each sample, which is based on the pathologist's visual analysis of the morphology and organisation of the glands in the tissue, a time-consuming and subjective task. In recent years, with the development of digitisation devices, the use of computer vision techniques for the analysis of biopsies has increased. However, to the best of the authors' knowledge, the development of algorithms to automatically detect individual cribriform patterns belonging to Gleason grade 4 has not yet been studied in the literature. The objective of the work presented in this paper is to develop a deep-learning-based system able to support pathologists in the daily analysis of prostate biopsies. This analysis must include the Gleason grading of local structures, the detection of cribriform patterns, and the Gleason scoring of the whole biopsy.

METHODS

The methodological core of this work is a patch-wise predictive model based on convolutional neural networks able to determine the presence of cancerous patterns based on the Gleason grading system. In particular, we train from scratch a simple self-design architecture with three filters and a top model with global-max pooling. The cribriform pattern is detected by retraining the set of filters of the last convolutional layer in the network. Subsequently, a biopsy-level prediction map is reconstructed by bi-linear interpolation of the patch-level prediction of the Gleason grades. In addition, from the reconstructed prediction map, we compute the percentage of each Gleason grade in the tissue to feed a multi-layer perceptron which provides a biopsy-level score.

RESULTS

In our SICAPv2 database, composed of 182 annotated whole slide images, we obtained a Cohen's quadratic kappa of 0.77 in the test set for the patch-level Gleason grading with the proposed architecture trained from scratch. Our results outperform previous ones reported in the literature. Furthermore, this model reaches the level of fine-tuned state-of-the-art architectures in a patient-based four groups cross validation. In the cribriform pattern detection task, we obtained an area under ROC curve of 0.82. Regarding the biopsy Gleason scoring, we achieved a quadratic Cohen's Kappa of 0.81 in the test subset. Shallow CNN architectures trained from scratch outperform current state-of-the-art methods for Gleason grades classification. Our proposed model is capable of characterising the different Gleason grades in prostate tissue by extracting low-level features through three basic blocks (i.e. convolutional layer + max pooling). The use of global-max pooling to reduce each activation map has shown to be a key factor for reducing complexity in the model and avoiding overfitting. Regarding the Gleason scoring of biopsies, a multi-layer perceptron has shown to better model the decision-making of pathologists than previous simpler models used in the literature.

Recutting Blocks of Prostate Needle Biopsies: How Much Diagnostic Yield Is Gained?

Objectives. The criteria for "active surveillance" depend in part on quantification of tumor extent and grade on prostate biopsies. It is known that false-negative biopsies may occur from incomplete sectioning of cores within the paraffin blocks. Methods. We retrospectively analyzed a prostate biopsy series, which were subjected to a second round of sections, in order to determine the rate of missed cancers. Results. Of 1324 sets of prostate biopsies, 4.5% (60) showed additional involved cores or higher grade tumor on recut sections. In 27 patients (2.0%), the changed diagnosis resulted in a potential mild increase in National Comprehensive Cancer Network (NCCN) risk, from negative to very low (12), very low to low (12), and low to favorable intermediate (3). In 3 patients (0.2%), the changed diagnosis resulted in a significant increase in NCCN risk. Comparison of the initial sets of slides to the recuts demonstrated areas of absent tissue in many of the cases in which tumor segments were missed. In 2/3 cases with the significant grade increase, gaps were present in one that should have alerted the pathologist to incomplete sections, and the tumor was fragmented at the edge of the core appearing incompletely sampled. Conclusions. A significant increase in risk was seen in this study in 0.2% of patients when blocks were recut for further sampling, with minor increases in 2%. While embedding issues only rarely resulted in clinically significant sampling error, the 3 significantly underdiagnosed cases underscore the need for pathologists to be alert to incomplete sections of prostate cores.

Defining clinically significant prostate cancer on the basis of pathological findings

The definition of clinically significant prostate cancer is a dynamic process that was initiated many decades ago, when there was already evidence that a great proportion of patients with prostate cancer diagnosed at autopsy never had any clinical symptoms. Autopsy studies led to examinations of radical prostatectomy (RP) specimens and the establishment of the definition of significant cancer at RP: tumour volume of 0.5 cm³ , Gleason grade 6 [Grade Group (GrG) 1], and organ-confined disease. RP studies were then used to develop prediction models for significant cancer by the use of needle biopsies. The first such model was used to delineate the first active surveillance (AS) criteria, known as the 'Epstein' criteria, in which patients with a cancer Gleason score of 3 + 3 = 6 (GrG1) involving fewer than two cores, and <50% of any given core, and a prostate-specific antigen density of <0.15 ng/ml per cm³ had a minimal risk of significant cancer at RP. These were adopted as components of the 'very-low-risk category' of the National Comprehensive Cancer Network guidelines, in which AS is supported as a management option. With the increase in the popularity of AS, much research has been carried out to better define significant/insignificant cancer, in order to be able to safely offer AS to a larger proportion of patients without the risk of undertreatment. Research has focused on allowing higher volume tumours, focal extraprostatic extension, and a limited amount of Gleason pattern 4, and the significance of different morphological patterns of Gleason 4. Other areas of research that will probably impact on the field but that are not covered in this review include the molecular classification of tumours and imaging techniques.