Classification Of Prostate Cancer Research Articles

Analysis of small non-coding RNAs in urinary exosomes to classify prostate cancer into low-grade (GG1) and higher-grade (GG2-5).

277 Background: To develop a new predictive test for prostate cancer, based on the interrogation of small non-coding RNAs (sncRNA) isolated from urinary exosomes. We report the development and performance of the miR Sentinel PCa Test, that distinguishes patients with prostate cancer from those with no evidence of prostate cancer (NEPC) and the miR Sentinel CS Test, that distinguishes low grade from higher grade disease. Methods: Affymetrix miR 4.0 arrays were used to identify informative sncRNAs isolated from urinary exosomes. sncRNA from 233 subjects undergoing a prostate biopsy [89 men with benign biopsies, 88 with grade group 1 (GG1) cancer and 56 patients with grade group 2-5 (GG2-5)] were interrogated on these arrays. A custom OpenArray platform was designed to interrogate the 280 most informative sncRNAs, identified using a data-driven selection algorithm. The platform was designed to categorize patients as either no cancer or cancer using the miR Sentinel PCa Test, and subclassify the patients with cancer into GG1 or GG2-5 cancer using the miR Sentinel CS Test. The performance of the miR Sentinel PCa and CS Tests was validated in an independent cohort. Results: In 233 men, theSentinel PCa Test correctly classified 89/89 subjects with no cancer and 144/144 with cancer. The Sentinel CS Test correctly identified 55/56 patients with GG2-5 and 87/88 patients with GG1. Sensitivity was 98%, Specificity 98%, NPV 98% and PPV 93%. For validation, a prospective observational study of 329 subjects (NEPC = 139; GG1= 88; GG2-5 = 102) with elevated PSA correctly classified 134/139 as no cancer [Sensitivity 98% (195/199); Specificity 96% (134/139), PPV 98% and NPV 97%]. The Sentinel CS Test classified 87/88 as GG1 and 102/102 as GG2-5 [Sensitivity 100% (102/102), Specificity 99% (87/88), PPV 99%, and NPV=100%]. Conclusions: Initial evaluation of the miR Sentinel PCa and CS Tests demonstrated the high precision of these tests to detect prostate cancer and distinguish high grade (GG2-5) disease. Further validation is ongoing.[Table: see text]

Transformation Based Tri-Level Feature Selection Approach Using Wavelets and Swarm Computing for Prostate Cancer Classification

Prostate Cancer is a cancer that occurs in the prostate- a small walnut shaped gland in men. This gland helps in the production of seminal fluid which is used to nourish and transport the sperm. One of the most common types of cancer in men is prostate cancer. A microarray dataset contains the microarray gene expression information. On a genome wide scale, gene expression profiles make it easy to analyze the patterns between genes and cancers, however the analysis of gene expression data is very difficult as it has a high dimensionality and low Signal to Noise Ratio (SNR). In this paper, a transformation-based Tri-level feature selection using wavelets for prostate cancer classification has been proposed. For the input microarray data, initially wavelets are applied and then the essential features are selected. Then the standardized gene selection techniques are implemented such as Relief-F, Fishers Score, Information Gain and SNR for a second level feature selection stage. Finally, before proceeding to classification, a third level feature selection by means of optimization techniques are implemented. The optimization techniques incorporated in this work are Marriage in Honey Bee Optimization Algorithm (MHBOA), Migrating Birds Optimization Algorithm (MBOA), Salp Swarm Optimization Algorithm (SSOA) and Whale Optimization Algorithm (WOA). This kind of an approach is totally new, and the best results show when SNR with WOA is classified with Artificial Neural Network (ANN) giving a classification accuracy of 99.48%. The second highest classification accuracy of 99.22% is obtained when Relief-F test with MBOA is classified with Naive Bayesian Classifier (NBC).

Artificial intelligence and radiomics in MRI-based prostate diagnostics

In view of the diagnostic complexity and the large number of examinations, modern radiology is challenged to identify clinically significant prostate cancer (PCa) with high sensitivity and specificity. Meanwhile overdiagnosis and overtreatment of clinically nonsignificant carcinomas need to be avoided. Increasingly, international guidelines recommend multiparametric magnetic resonance imaging (mpMRI) as first-line investigation in patients with suspected PCa. Image interpretation according to the PI-RADS criteria is limited by interobserver variability. Thus, rapid developments in the field of automated image analysis tools, including radiomics and artificial intelligence (AI; machine learning, deep learning), give hope for further improvement in patient care. AI focuses on the automated detection and classification of PCa, but it also attempts to stratify tumor aggressiveness according to the Gleason score. Recent studies present good to very good results in radiomics or AI-supported mpMRI diagnosis. Nevertheless, these systems are not widely used in clinical practice. In order to apply these innovative technologies, agrowing awareness for the need of structured data acquisition, development of robust systems and an increased acceptance of AI as diagnostic support are needed. If AI overcomes these obstacles, it may play akey role in the quantitative and reproducible image-based diagnosis of ever-increasing prostate MRI examination volumes.

Impact of a novel ultrasound microvascular imaging and elastography on prostate cancer classification.

This study was aimed to compare the sensitivity of transrectal ultrasound (TRUS) guided systematic biopsy (TRUS-SB), superb microvascular imaging guided targeted biopsy (SMI-TB), and elastography ultrasound guided targeted biopsy (EU-TB) in detecting prostate cancer (PCa). One hundred and eighty-four patients with suspicious PCa were randomly divided into two groups, SMI-TB group (n=92) and EU-TB group (n=92). All the patients received a 2-core SMI-TB or a 2-core EU-TB before receiving a 12-core TRU-SB depending on the group. The 2-core SMI-TB was conducted at the most increased vascularity areas rated by a radiologist on a semi-quantitative criterion. The 2-core EU-TB was performed at the dark blue areas that identified as suspicious areas. The PCa detection rate was compared among TRUS-SB, SMI-TB, and EU-TB in the total population and in two groups. PCa was detected in 65 out of 184 patients. The SMI-TB group patients received 1,272 cores in total with a positive rate of 23.81% (40/168) for SMI-TB cores and 4.17% (46/1,104) for TRUS-SB cores. The EU-TB group patients received a total of 1,262 cores with a positive rate of 31.01% (49/158) for EU-TB cores and 5.34% (59/1,104) for TRUS-SB cores. Targeted cores (27.30%, 89/326) achieved significantly higher sensitivity in PCa detection than systematic cores (4.76%, 105/2,208) (SMI-TB vs. TRUS-SB: OR 7.188, P<0.001; EU-TB vs. TRUS-SB: OR 7.962, P<0.001). Superb microvascular imaging and elastography ultrasound guided targeted biopsy may result in higher prostate cancer detection rate as adjunct to gray-scale ultrasound guided systematic biopsy. However, targeted biopsy alone is not sufficient to detect PCa, compared with systematic biopsy.

PROSTATE CANCER DETECTION USING HISTOPATHOLOGY IMAGES AND CLASSIFICATION USING IMPROVED RideNN

Medical industry reports prostate cancer as common and high among men and alarms the necessity for detecting prostate cancer for which the required morphology is extracted from the histopathology images. Commonly, the Gleason grading system remains a perfect factor for grading prostate cancer in men, but pathologists suffer from minute inter- and intra-observer variations. Thus, an automatic method for segmenting and classifying prostate cancer is modeled in this paper. The significance of the developed method is that the segmentation and classification are gland-oriented using the Color Space (CS) transformation and Salp Swarm Optimization Algorithm-based Rider Neural Network (SSA-RideNN). The gland region is considered as the morphology for cancer detection from which the maximal significant regions are extracted as features using multiple-kernel scale-invariant feature transform (MK-SIFT). Here, the RideNN classifier is trained optimally using the proposed Salp–Rider Algorithm (SRA), which is the integration of Salp Swarm Optimization Algorithm (SSA) and Rider Optimization Algorithm (ROA). The experimentation is performed using the histopathology images and the analysis based on sensitivity, accuracy, and specificity reveals that the proposed prostate cancer detection method acquired the maximal accuracy, sensitivity, and specificity of 0.8966, 0.8919, and 0.8596, respectively.

P084 - Application of Raman spectroscopy coupled with chemometrics to the early diagnostics and classification of prostate cancer

P084 - Application of Raman spectroscopy coupled with chemometrics to the early diagnostics and classification of prostate cancer

Random forest-based modelling to detect biomarkers for prostate cancer progression

BackgroundThe clinical course of prostate cancer (PCa) is highly variable, demanding an individualized approach to therapy. Overtreatment of indolent PCa cases, which likely do not progress to aggressive stages, may be associated with severe side effects and considerable costs. These could be avoided by utilizing robust prognostic markers to guide treatment decisions.ResultsWe present a random forest-based classification model to predict aggressive behaviour of prostate cancer. DNA methylation changes between PCa cases with good or poor prognosis (discovery cohort with n = 70) were used as input. DNA was extracted from formalin-fixed tumour tissue, and genome-wide DNA methylation differences between both groups were assessed using Illumina HumanMethylation450 arrays. For the random forest-based modelling, the discovery cohort was randomly split into a training (80%) and a test set (20%). Our methylation-based classifier demonstrated excellent performance in discriminating prognosis subgroups in the test set (Kaplan-Meier survival analyses with log-rank p value < 0.0001). The area under the receiver operating characteristic curve (AUC) for the sensitivity analysis was 95%. Using the ICGC cohort of early- and late-onset prostate cancer (n = 222) and the TCGA PRAD cohort (n = 477) for external validation, AUCs for sensitivity analyses were 77.1% and 68.7%, respectively. Cancer progression-related DNA hypomethylation was frequently located in ‘partially methylated domains’ (PMDs)—large-scale genomic areas with progressive loss of DNA methylation linked to mitotic cell division. We selected several candidate genes with differential methylation in gene promoter regions for additional validation at the protein expression level by immunohistochemistry in > 12,000 tissue micro-arrayed PCa cases. Loss of ZIC2 protein expression was associated with poor prognosis and correlated with significantly shorter time to biochemical recurrence. The prognostic value of ZIC2 proved to be independent from established clinicopathological variables including Gleason grade, tumour stage, nodal stage and prostate-specific-antigen.ConclusionsOur results highlight the prognostic relevance of methylation loss in PMD regions, as well as of several candidate genes not previously associated with PCa progression. Our robust and externally validated PCa classification model either directly or via protein expression analyses of the identified top-ranked candidate genes will support the clinical management of prostate cancer.

Precision Medicine in Prostate Cancer

Abstract Prostate cancer is generally known as cancer with a good prognosis, and even in metastatic cases, the 10-year survival rate in Japan is around 38%. Since gene mutations constantly change over a relatively long course of treatment, the clinical utility of Liquid Biopsy, which can evaluate mutations from cancer cells in the blood, is attracting clinical attention. As a result of analysis using a next-generation sequencer, the most frequent mutation is androgen receptor amplification (AR Amp), followed by P53 mutation and PTEN mutation. Based on the above background, the prognostic significance of AR Amp was studied in 102 samples (38 cases) of Cell-free DNA (cfDNA) derived from prostate cancer patients treated at our institution. As a result of performing multivariate analysis including PSA, a clinical marker of prostate cancer, AR Amp predicts independently over PSA in progression-free survival (PFS) and overall survival (OS). In prostate cancer, in addition to AR, DNA repair gene mutations such as BRCA and ATM have been found around 10%, and in cases with BRCA deletion, PARP inhibitors have been reported to prolong the prognosis by six months. In cases of prostate cancer with a high degree of microbial instability(MSI-H), although as low as around 2% of the total prostate cancer patients, the immune checkpoint inhibitor showed a decent response. Currently, the risk classification of prostate cancer is determined based on the pathological grade (Gleason Score), PSA, size of the metastatic lesion (High Volume). However, the time will come to choose the treatment policy based on new risk classification includes the genome information.

Prostate cancer pathology: What has changed in the last 5 years.

Prostate cancer is the most frequent non-cutaneous malignancy in men in the United States. In the last few years, many recommendations have been made available from the 2014 International Society of Urologic Pathology consensus conference, 2016 World Health Organization blue book and 2018 8th edition of American Joint Committee on Cancer Staging System. Here, we focus on four topics which are considered relevant on the basis of their common appearance in routine practice, clinical importance and 'need to improve communication between pathology reports and clinicians': prostate cancer classification, prostate cancer grading, prostate cancer staging, and current definition of clinically significant prostate cancer. Tissue biomarkers that can predict significant disease and/or upgrading and tissue-based genomics for the purpose of diagnosis and prognosis are mentioned briefly.

Interfocal heterogeneity challenges the clinical usefulness of molecular classification of primary prostate cancer

Prostate cancer is a highly heterogeneous disease and typically multiple distinct cancer foci are present at primary diagnosis. Molecular classification of prostate cancer can potentially aid the precision of diagnosis and treatment. A promising genomic classifier was published by The Cancer Genome Atlas (TCGA), successfully classifying 74% of primary prostate cancers into seven groups based on one cancer sample per patient. Here, we explore the clinical usefulness of this classification by testing the classifier’s performance in a multifocal context. We analyzed 106 cancer samples from 85 distinct cancer foci within 39 patients. By somatic mutation data from whole-exome sequencing and targeted qualitative and quantitative gene expression assays, 31% of the patients were uniquely classified into one of the seven TCGA classes. Further, different samples from the same focus had conflicting classification in 12% of the foci. In conclusion, the level of both intra- and interfocal heterogeneity is extensive and must be taken into consideration in the development of clinically useful molecular classification of primary prostate cancer.

Semi-automatic classification of prostate cancer on multi-parametric MR imaging using a multi-channel 3D convolutional neural network.

To present a deep learning-based approach for semi-automatic prostate cancer classification based on multi-parametric magnetic resonance (MR) imaging using a 3D convolutional neural network (CNN). Two hundred patients with a total of 318 lesions for which histological correlation was available were analyzed. A novel CNN was designed, trained, and validated using different combinations of distinct MRI sequences as input (e.g., T2-weighted, apparent diffusion coefficient (ADC), diffusion-weighted images, and K-trans) and the effect of different sequences on the network's performance was tested and discussed. The particular choice of modeling approach was justified by testing all relevant data combinations. The model was trained and validated using eightfold cross-validation. In terms of detection of significant prostate cancer defined by biopsy results as the reference standard, the 3D CNN achieved an area under the curve (AUC) of the receiver operating characteristics ranging from 0.89 (88.6% and 90.0% for sensitivity and specificity respectively) to 0.91 (81.2% and 90.5% for sensitivity and specificity respectively) with an average AUC of 0.897 for the ADC, DWI, and K-trans input combination. The other combinations scored less in terms of overall performance and average AUC, where the difference in performance was significant with a p value of 0.02 when using T2w and K-trans; and 0.00025 when using T2w, ADC, and DWI. Prostate cancer classification performance is thus comparable to that reported for experienced radiologists using the prostate imaging reporting and data system (PI-RADS). Lesion size and largest diameter had no effect on the network's performance. The diagnostic performance of the 3D CNN in detecting clinically significant prostate cancer is characterized by a good AUC and sensitivity and high specificity. • Prostate cancer classification using a deep learning model is feasible and it allows direct processing of MR sequences without prior lesion segmentation. • Prostate cancer classification performance as measured by AUC is comparable to that of an experienced radiologist. • Perfusion MR images (K-trans), followed by DWI and ADC, have the highest effect on the overall performance; whereas T2w images show hardly any improvement.

Prostate imaging reporting and data system correlation with Gleason score: Pathological aspects of magnetic resonance imaging findings.

Prostate cancer has a high prevalence and mortality, being the most diagnosed urologic cancer. Prostatic magnetic resonance imaging showed high sensitivity in the detection of clinically significant neoplasia and agreement with the Gleason score. Therefore, we attempted to evaluate the diagnostic accuracy of the prostate imaging reporting and data system, using biopsy and prostatectomy as the reference standard. The secondary goal of correlating prostatic magnetic resonance imaging findings and anatomopathological samples is obtained. We retrospectively analyzed seventy-nine 1.5 Tesla prostatic magnetic resonance imaging scans in patients aged 31 to 86 years, performed at the Clinical Hospital of the Federal University of Paraná between January 2015 and February 2018. Considering all 79 patients, prostatic magnetic resonance imaging was able to diagnose tumor in 47 patients (59.4%). Considering the peripheral zone, the prostatic magnetic resonance imaging had a sensitivity of 75.0% (95% confidence interval: 52.1%-98.0%), specificity of 89.5% (95% confidence interval: 66.0%-100%), 94.4% positive predictive value (95% confidence interval: 71.0%-100%), 66.7% negative predictive value (95% confidence interval: 43.0%-69.0%), 83.8% Positive Likelihood Ratio (PVR) (95% confidence interval: 60.0%-100%), 27.9% Negative Likelihood Ratio (RVN) (95% confidence interval: 5.0%-50.0 %), and accuracy of 86.3% (95% confidence interval: 63.0%-100%). The receiver operating characteristic curve obtained demonstrated the sensitivity variation according to the prostate imaging reporting and data system score of the patients, obtaining an area under the curve of 84.8 for a prostate imaging reporting and data system cutoff of 3. The use of the prostate imaging reporting and data system score is useful for the screening and classification of prostate cancer, due to its easy reproducibility, even in a population with an unknown prostate cancer prevalence, which can be easily correlated with biopsy studies and/or radical prostatectomy.

Abstract 453: A circulating tumor cell assay for dynamic assessment of drug sensitivity in metastatic castration-resistant prostate cancer

Abstract Background: Tissue-based gene signatures can predict clinical behavior in prostate cancer (PC). Our objective was to extend their application to circulating tumor cells (CTCs) and to show that changes in the signature were associated with changes in clinical behavior. Methods: Our approach combined the Thermoresponsive(TR)-NanoVelcro CTC purification system with the Nanostring nCounter system for cellular purification and transcriptomic analysis. The Prostate Cancer Classification System (PCS) panel was modified for use in CTCs. We selected 31 blood samples from 23 PC patients receiving androgen receptor signaling inhibitors (ARSI) and measured the PCS1 Z score (probability). These findings were compared with clinical outcome data (responsiveness/resistance). Results: A modified, 16-gene PCS1 signature was established and validated through a rigorous bioinformatics process. We performed analytical validation of our combined CTC-RNA system to ensure reproducibility and specificity. In patient bloods, ARSI-resistant samples (ARSI-R, n=14) had significantly higher PCS1 Z scores as compared with ARSI-sensitive samples (ARSI-S, n=17) (Rank-sum test, P=0.003). The analyzed bloods contained samples from 8 patients who developed resistance to an ARSI allowing for dynamic measurement of gene expression. Our analysis found that the PCS1 Z score increased at the time that ARSI-resistance emerged (Pairwise T-test, P=0.016). Conclusions:Using this new methodology, contemporary, clinically-relevant gene signatures such as PCS could be measured non-invasively in CTCs. These findings can be used to relate gene expression to clinical drug response. This approach also allowed for measurement of dynamic variations of gene expression in individual patients over time that correlated to ARSI sensitivity. Citation Format: Pai-Chi Teng, Yu Jen Jan, Junhee Yoon, Jie-Fu Chen, Pin-Jung Chen, Nu Yao, Shirley Cheng, Amber Lozano, Michael R. Freeman, Sungyong You, Hsian-Rong Tseng, Edwin M. Posadas. A circulating tumor cell assay for dynamic assessment of drug sensitivity in metastatic castration-resistant prostate cancer [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2019; 2019 Mar 29-Apr 3; Atlanta, GA. Philadelphia (PA): AACR; Cancer Res 2019;79(13 Suppl):Abstract nr 453.

Combination possibility and deep learning model as clinical decision-aided approach for prostate cancer.

This study aims to introduce as proof of concept a combination model for classification of prostate cancer using deep learning approaches. We utilized patients with prostate cancer who underwent surgical treatment representing the various conditions of disease progression. All possible combinations of significant variables from logistic regression and correlation analyses were determined from study data sets. The combination possibility and deep learning model was developed to predict these combinations that represented clinically meaningful patient's subgroups. The observed relative frequencies of different tumor stages and Gleason score Gls changes from biopsy to prostatectomy were available for each group. Deep learning models and seven machine learning approaches were compared for the classification performance of Gleason score changes and pT2 stage. Deep models achieved the highest F1 scores by pT2 tumors (0.849) and Gls change (0.574). Combination possibility and deep learning model is a useful decision-aided tool for prostate cancer and to group patients with prostate cancer into clinically meaningful groups.

Classification of prostate cancer based on clinical and omics data using neural networks techniques to improve prognostic power.

e16569 Background: In 2017, prostate cancer (PCa) was the second most common cancer in men after lung cancer. While there are different courses of action to treat the disease, its mortality in Peru is higher than 50%. Conventionally, PCa is diagnosed by evaluating tissue biopsies, and classified according to the Gleason grading system. Novel molecular classifications of PCa have been proposed for diagnostic and prognostic purposes. The main goal of this work is to implement a tool predicting the disease free time of patient according to the genomic expression and highlight the genes playing an influential role on the prediction. Methods: Modern techniques to classify data keep getting broader and more accurate, in particular with the introduction of Neural Networks(NN). We implement an Artificial Neural Network automatic genomic classification strategy based on a Local Interpretable Model-Agnostic Explanations (LIME) algorithm because it allows the network to choose the features of major discriminative significance. As a proof-of-concept, we selected a sub-set of 3530 genes related to recurrence from 499 PCa genomes to build the neural networks. Results: The resulting neural network, trained and tested on cancer cell 2010 database and validate on the MSKCC data the can predict the time of recurrence within a range of three months based on the genomic expression with an accuracy of 96,9% and a loss of less than 9%. Using the implemented LIME algorithm, our results indicate that this subset of genes is informative of recurrence and plays a substantial role in the prediction. Conclusions: Instead of using a classic fully connected layer, we implemented different types of Deep Learning networks where the final network provides the predicted survival rate or time to recurrence. This information will allow the doctors to propose the best course of treatment. Our method is able to generate an augmented score, enabling a more accurate evaluation of risk and personalized treatment strategy

A circulating tumor cell specific RNA assay for assessment of androgen receptor signaling inhibitor sensitivity in metastatic castration-resistant prostate cancer.

5059 Background: Our objective is to develop a circulating tumor cell (CTC)-RNA assay for characterizing clinically relevant RNA signatures for the assessment of androgen receptor signaling inhibitors (ARSIs) sensitivity in metastatic castration-resistant prostate cancer (mCRPC) patients. Methods: We developed NanoVelcro CTC-RNA Assay by combining Thermoresponsive(TR)-NanoVelcro CTC purification system with NanoString nCounter platform for CTC purification and RNA analysis. Based on the well-validated, tissue-based Prostate Cancer Classification System (PCS), we selected the most aggressive and ARSI-resistant subtype- the PCS1, for CTC analysis. We applied a rigorous bioinformatic process to develop a CTC-PCS1 panel that is specific to PC CTCs. We validated NanoVelcro CTC-RNA Assay and CTC-PCS1 panel with PC cell lines to demonstrate sensitivity and specificity of the PCS1 Z score (the likelihood estimate of the PCS1 subtype) for identifying PCS1 subtype and ARSI resistance. We then selected 31 blood samples from 23 PC patients receiving ARSIs to test in our assay. The PCS1 Z score of each sample was computed and compared with ARSI treatment sensitivity. Results: We established a 16-gene CTC-PCS1 panel that consists of CTC-specific RNA signatures. The validation studies using PC cell lines showed that the assay can detect the RNA transcripts with high sensitivity and scalability in the range of 1-100 cells. We also showed that the genes in CTC-PCS1 panel is highly expressed in PC cells. We further demonstrated that the CTC-PCS1 panel is highly specific in identifying PCS1-like samples, and the high PCS1 Z score is associated with ARSI resistance. In patient bloods, ARSI-resistant samples (ARSI-R, n=14) had significantly higher PCS1 Z scores as compared with ARSI-sensitive samples (ARSI-S, n=17) (Rank-sum test, P=0.003). In 8 patients who were initially sensitive to ARSI (ARSI-S) and later developed resistance (ARSI-R), we found that the PCS1 Z score increased from the time of ARSI-S to the time of ARSI-R (Pairwise T-test, P=0.016). Conclusions: Using our new methodology, we developed a first-in-class CTC-RNA assay and demonstrated the feasibility of transforming clinically-relevant tissue-based RNA profiling into CTC tests. This approach allows for detecting RNA expression relevant to clinical drug resistance in a non-invasive fashion, which can facilitate patient-specific treatment selection and early detection of drug resistance- a goal in precision oncology.

Re: Classification of Prostate Cancer Using a Protease Activity Nanosensor Library.

Re: Classification of Prostate Cancer Using a Protease Activity Nanosensor Library.

Prostate pathological evaluation after radical prostatectomy

The article presents information about the prostate's zonal structure, prostate preparation after radical prostatectomy, cancer incidence in different prostate areas. This provides an overview of the WHO classification (2016) of prostate tumors, the changes are noted, approaches to the diagnosis of cancer variants, Gleason score assessment, gradation determination, clinical prognostic risk groups for prostate cancer, primary tumor assessment according to the international classification of prostate cancer according to the TNM system (AjCC 8th ed.), which was supplemented and revised in 2017. Prognostic factors for acinar carcinoma and the role of immunohistochemical diagnosis in the morphological material were determined. In conclusion, the main significant parameters are indicated, which should be reflected in the Protocol of in vivo pathoanatomical study of the prostate samples after radical prostatectomy.

Deep neural maps for unsupervised visualization of high-grade cancer in prostate biopsies.

Prostate cancer (PCa) is the most frequent noncutaneous cancer in men. Early detection of PCa is essential for clinical decision making, and reducing metastasis and mortality rates. The current approach for PCa diagnosis is histopathologic analysis of core biopsies taken under transrectal ultrasound guidance (TRUS-guided). Both TRUS-guided systematic biopsy and MR-TRUS-guided fusion biopsy have limitations in accurately identifying PCa, intraoperatively. There is a need to augment this process by visualizing highly probable areas of PCa. Temporal enhanced ultrasound (TeUS) has emerged as a promising modality for PCa detection. Prior work focused on supervised classification of PCa verified by gold standard pathology. Pathology labels are noisy, and data from an entire core have a single label even when significantly heterogeneous. Additionally, supervised methods are limited by data from cores with known pathology, and a significant portion of prostate data is discarded without being used. We provide an end-to-end unsupervised solution to map PCa distribution from TeUS data using an innovative representation learning method, deep neural maps. TeUS data are transformed to a topologically arranged hyper-lattice, where similar samples are closer together in the lattice. Therefore, similar regions of malignant and benign tissue in the prostate are clustered together. Our proposed method increases the number of training samples by several orders of magnitude. Data from biopsy cores with known labels are used to associate the clusters with PCa. Cancer probability maps generated using the unsupervised clustering of TeUS data help intuitively visualize the distribution of abnormal tissue for augmenting TRUS-guided biopsies.

Comparison of Artificial Intelligence Techniques to Evaluate Performance of a Classifier for Automatic Grading of Prostate Cancer From Digitized Histopathologic Images

ImportanceProper evaluation of the performance of artificial intelligence techniques in the analysis of digitized medical images is paramount for the adoption of such techniques by the medical community and regulatory agencies.ObjectivesTo compare several cross-validation (CV) approaches to evaluate the performance of a classifier for automatic grading of prostate cancer in digitized histopathologic images and compare the performance of the classifier when trained using data from 1 expert and multiple experts.Design, Setting, and ParticipantsThis quality improvement study used tissue microarray data (333 cores) from 231 patients who underwent radical prostatectomy at the Vancouver General Hospital between June 27, 1997, and June 7, 2011. Digitized images of tissue cores were annotated by 6 pathologists for 4 classes (benign and Gleason grades 3, 4, and 5) between December 12, 2016, and October 5, 2017. Patches of 192 µm2 were extracted from these images. There was no overlap between patches. A deep learning classifier based on convolutional neural networks was trained to predict a class label from among the 4 classes (benign and Gleason grades 3, 4, and 5) for each image patch. The classification performance was evaluated in leave-patches-out CV, leave-cores-out CV, and leave-patients-out 20-fold CV. The analysis was performed between November 15, 2018, and January 1, 2019.Main Outcomes and MeasuresThe classifier performance was evaluated by its accuracy, sensitivity, and specificity in detection of cancer (benign vs cancer) and in low-grade vs high-grade differentiation (Gleason grade 3 vs grades 4-5). The statistical significance analysis was performed using the McNemar test. The agreement level between pathologists and the classifier was quantified using a quadratic-weighted κ statistic.ResultsOn 333 tissue microarray cores from 231 participants with prostate cancer (mean [SD] age, 63.2 [6.3] years), 20-fold leave-patches-out CV resulted in mean (SD) accuracy of 97.8% (1.2%), sensitivity of 98.5% (1.0%), and specificity of 97.5% (1.2%) for classifying benign patches vs cancerous patches. By contrast, 20-fold leave-patients-out CV resulted in mean (SD) accuracy of 85.8% (4.3%), sensitivity of 86.3% (4.1%), and specificity of 85.5% (7.2%). Similarly, 20-fold leave-cores-out CV resulted in mean (SD) accuracy of 86.7% (3.7%), sensitivity of 87.2% (4.0%), and specificity of 87.7% (5.5%). Results of McNemar tests showed that the leave-patches-out CV accuracy, sensitivity, and specificity were significantly higher than those for both leave-patients-out CV and leave-cores-out CV. Similar results were observed for classifying low-grade cancer vs high-grade cancer. When trained on a single expert, the overall agreement in grading between pathologists and the classifier ranged from 0.38 to 0.58; when trained using the majority vote among all experts, it was 0.60.Conclusions and RelevanceResults of this study suggest that in prostate cancer classification from histopathologic images, patch-wise CV and single-expert training and evaluation may lead to a biased estimation of classifier’s performance. To allow reproducibility and facilitate comparison between automatic classification methods, studies in the field should evaluate their performance using patient-based CV and multiexpert data. Some of these conclusions may be generalizable to other histopathologic applications and to other applications of machine learning in medicine.