Waterjet Ablation Therapy for Endoscopic Resection of prostate tissue trial (WATER) vs WATER II: comparing Aquablation therapy for benign prostatic hyperplasia in 30-80 and 80-150 mL prostates.

OBJECTIVE: To compare the outcomes of Aquablation in 30-80 mL prostates with those in 80-150 mL prostates. Surgical options, especially with short learning curves, are limited when treating large prostates for lower urinary tract symptoms (LUTS) due to benign prostatic hyperplasia (BPH). Aquablation (AquaBeam System, PROCEPT BioRobotics Inc., Redwood City, CA, USA) could solve this issue with global reproducibility, independent of prostate volume. PATIENTS AND METHODS: Waterjet Ablation Therapy for Endoscopic Resection of prostate tissue (WATER [W-I]; NCT02505919) is a prospective, double-blind, multicentre, international clinical trial comparing Aquablation and transurethral resection of the prostate (TURP) for the treatment of LUTS/BPH in prostates between 30 and 80 mL. WATER II (W-II; NCT03123250) is a prospective, multicentre, single-arm international clinical trial of Aquablation in prostates between 80 and 150 mL. We compare baseline parameters and 12-month outcomes in 116 W-I and 101 W-II study patients. Students' t-test or Wilcoxon tests were used for continuous variables and Fisher's test for binary variables. RESULTS: The mean (SD) operative time was 33 (17) and 37 (13) min in W-I and W-II, respectively. Actual treatment time was 4 and 8 min in W-I and W-II, respectively. The mean change in the International Prostate Symptom Score was substantial averaging (at 12 months) 15.1 in W-I and 17.1 in W-II (P = 0.605). By 3 months, Clavien-Dindo grade ≥II events occurred in 19.8% of W-I patients and 34.7% of W-II patients (P = 0.468). CONCLUSION: Aquablation clinically normalises outcomes between patients with 30-80 mL prostates and patients with 80-150 mL prostates treated for LUTS/BPH, with an expected increase in the risk of complications in larger prostates. Long-term outcomes of procedure durability are needed.

Aquablation for benign prostatic hyperplasia in large prostates (80-150 cc): 2-year results.

INTRODUCTION: To report 2-year safety and effectiveness of the Aquablation procedure for the treatment of men with symptomatic benign prostatic hyperplasia (BPH) and large-volume 80-150 cc prostates. MATERIALS AND METHODS: Between September-December 2017, 101 men with moderate-to-severe BPH symptoms and prostate volumes of 80-150 cc underwent an ultrasound-guided robotically executed Aquablation procedure in a prospective multicenter international clinical trial (WATER II). Baseline, procedural and follow up parameters were recorded at baseline and scheduled postoperative visits. Herein we report 2-year safety and efficacy for this cohort. RESULTS: Mean prostate volume was 107 cc (range 80-150 cc). Mean IPSS improved from 23.2 at baseline to 5.8 at 2 years (17-point improvement, p < .0001). Mean IPSS quality of life improved from 4.6 at baseline to 1.1 at 2 years (p < .0001). Maximum urinary flow increased from 8.7 to 18.2 cc/sec. Two subjects underwent a repeat procedure for BPH symptoms over the 2-year follow up period. By 2 years or study exit, all but 2 of 74 subjects stopped taking alpha blockers. Similarly, all but 4 of 32 subjects stopped taking 5α-reductase inhibitors. CONCLUSIONS: Two-year prospective multicenter follow up demonstrated that the Aquablation procedure is safe and effective in the treatment of men with LUTS due to BPH and prostates 80-150 cc with durable treatment efficacy, acceptable safety profile and a low retreatment rate. ClinicalTrials.gov number, NCT03123250.

WATER II (80-150 mL) procedural outcomes.

OBJECTIVES: To present early safety and feasibility data from a multicentre prospective study (WATER II) of aquablation in the treatment of symptomatic men with large-volume benign prostatic hyperplasia (BPH). METHODS: Between September and December 2017, 101 men with moderate-to-severe BPH symptoms and prostate volume of 80-150 mL underwent aquablation in a prospective multicentre international clinical trial. Baseline demographics and standardized postoperative management variables were carefully recorded in a central independently monitored database. Surgeons answered analogue scale questionnaires on intra-operative technical factors and postoperative management. Adverse events up to 1 month were adjudicated by an independent clinical events committee. RESULTS: The mean (range) prostate volume was 107 (80-150) mL. The mean (range) operating time was 37 (15-97) min and aquablation resection time was 8 (3-15) min. Adequate adenoma resection was achieved with a single pass in 34 patients and with additional passes in 67 patients (mean 1.8 treatment passes), all in a single operating session. Haemostasis was achieved using either a Foley balloon catheter placed in the bladder under traction (n = 98, mean duration 18 h) or direct tamponade using a balloon inflated in the prostate fossa (n = 3, mean duration 15 h). No patient required electrocautery for haemostasis at the time of the primary procedure. The mean length of stay after the procedure was 1.6 days (range same day to 6 days). The Clavien-Dindo grade ≥2 event rate observed at 1 month was 29.7%. Bleeding complications were recorded in 10 patients (9.9%) during the index procedure hospitalization prior to discharge, and included six (5.9%) peri-operative transfusions. CONCLUSIONS: Aquablation is feasible and safe in treating men with men with large prostates (80-150 mL). The 6-month efficacy data are being accrued and will be presented in future publications (ClinicalTrials.gov number, NCT03123250).

Aquablation for benign prostatic hyperplasia in large prostates (80-150 mL): 6-month results from the WATER II trial.

OBJECTIVE: To present 6-month safety and effectiveness data from a multicentre prospective study of aquablation in men with lower urinary tract symptoms (LUTS) attributable to benign prostatic hyperplasia (BPH) with prostate volumes between 80 and 150 mL. METHODS: Between September and December 2017, 101 men with LUTSattributable to BPHwere prospectively enrolled at 16 centres in Canada and the USA. RESULTS: The mean prostate volume was 107 mL. The mean length of hospital stay after the aquablation procedure was 1.6 days (range: same day to 6 days). The primary safety endpoint (Clavien-Dindo grade 2 or higher or any grade 1 event resulting in persistent disability) at 3 months occurred in 45.5% of men, which met the study design goal of < 65% (P < 0.001). At 6 months, 22% of the patients had experienced a Clavien-Dindo grade 2, 14% a grade 3 and 5% a grade 4 adverse event. Bleeding complications requiring intervention and/or transfusion were recorded in eight patients prior to discharge and in six patients after discharge. The mean International Prostate Symptom Score improved from 23.2 ± 6.3 at baseline to 6.7 ± 5.1 at 3 months, meeting the study's primary efficacy endpoint goal (P < 0.001). The maximum urinary flow rate increased from 8.7 to 18.8 mL/s (P < 0.001) and post-void residual urine volume decreased from 131 at baseline to 47 at 6 months (P < 0.0001). At 6 months, prostate-specific antigen concentration reduced from 7.1 ± 5.9 ng/mL at baseline to 4.0 ± 3.9 ng/mL, a 44% reduction. CONCLUSIONS: Aquablation is safe and effective in treating men with larger prostates (80-150 mL), without significant increase in procedure or resection time.

Correction to: Patterns and predictors of registration and participation at a supportive care program for prostate cancer survivors.

The name of Stirling Bryan was incorrectly captured in the original manuscript.

Patterns and predictors of registration and participation at a supportive care program for prostate cancer survivors.

PURPOSE: To examine registration rates, and the timing/intensity of follow-up with a prostate cancer supportive care (PCSC) program, and to explore clinical and sociodemographic factors associated with participation and non-participation. METHODS: We used electronic medical records for a cohort of men diagnosed with prostate cancer (PC) who attended a PC-related appointment at the Vancouver Prostate Centre, January 2013-December 2016. We used multivariate logistic regression to quantify the effect of diagnostic treatment and sociodemographic characteristic PCSC program registration and subsequent attendance. We produced Kaplan-Meier estimators to assess the probability of program attendance over the disease trajectory for those who registered. RESULTS: Almost 15% of the men who registered for the program did not end up using any services. An additional 23% attended only one session/clinical appointment. Surgical and radiation treatments increased the odds and intensity of participation. Long travel distance decreased the odds of registering and participating. Low income decreased the odds of registration but not subsequent participation. CONCLUSIONS: While the use of supportive care services can help address the detrimental effects of prostate cancer diagnosis and treatment, one in six men who register for supportive care do not end up using any. Offering these services at no cost and alongside treatment appears to be insufficient to ensuring access for all patients. Additional research is needed to understand barriers and facilitators of accessing supportive care in this population.

Evaluating an educational intervention to alleviate distress amongst men with newly diagnosed prostate cancer and their partners.

OBJECTIVE: To determine whether an education session alleviates distress for both patients with prostate cancer and their partners; and whether their partner's attendance at the session; and disease, treatment, and sociodemographic characteristics affect changes in distress levels. PATIENTS, SUBJECTS AND METHODS: We identified men with untreated prostate cancer at the Vancouver Prostate Centre between February 2015 and March 2016 who agreed to attend our education session. The session consisted of a didactic presentation covering the biology of prostate cancer, treatment options, and side-effects, followed by a private joint session with a urologist and radiation oncologist. We assessed distress using the Distress Thermometer (DT) and compared pre- and post-session distress, and change in distress between patients and partners using matched and unmatched t-tests, respectively. We also assessed pre-session anxiety using the seven-item Generalised Anxiety Disorder measure, and decisional certainty using the Decisional Conflict Scale. RESULTS: In all, 71 patients and 48 partners participated in the study. Attending the session led to a significant reduction in the median DT score for patients (4.0-3.0, P < 0.01) and partners (5.0-4.0, P = 0.02). Partners reported higher distress both before and after the session (4.9 vs 3.8, P = 0.03 pre-session and 4.2 vs 3.2, P = 0.03 post-session). The presence of a partner at the session did not affect patients' pre- or post-session distress or the success of the session at alleviating distress. Sociodemographic and clinical characteristics had little effect on distress levels. CONCLUSIONS: An interdisciplinary education session is equally effective at alleviating distress for both patients with prostate cancer and their female partners.

Intraoperative registered transrectal ultrasound guidance for robot-assisted laparoscopic radical prostatectomy.

PURPOSE: To provide unencumbered real-time ultrasound image guidance during robot-assisted laparoscopic radical prostatectomy, we developed a robotic transrectal ultrasound system that tracks the da Vinci® Surgical System instruments. We describe our initial clinical experience with this system. MATERIALS AND METHODS: After an evaluation in a canine model, 20 patients were enrolled in the study. During each procedure the transrectal ultrasound transducer was manually positioned using a brachytherapy stabilizer to provide good imaging of the prostate. Then the transrectal ultrasound was registered to the da Vinci robot by a previously validated procedure. Finally, automatic rotation of the transrectal ultrasound was enabled such that the transrectal ultrasound imaging plane safely tracked the tip of the da Vinci instrument controlled by the surgeon, while real-time transrectal ultrasound images were relayed to the surgeon at the da Vinci console. Tracking was activated during all critical stages of the surgery. RESULTS: The transrectal ultrasound robot was easy to set up and use, adding 7 minutes (range 5 to 14) to the procedure. It did not require an assistant or additional control devices. Qualitative feedback was acquired from the surgeons, who found transrectal ultrasound useful in identifying the urethra while passing the dorsal venous complex suture, defining the prostate-bladder interface during bladder neck dissection, identifying the seminal vesicles and their location with respect to the rectal wall, and identifying the distal prostate boundary at the apex. CONCLUSIONS: Real-time, registered robotic transrectal ultrasound guidance with automatic instrument tracking during robot-assisted laparoscopic radical prostatectomy is feasible and potentially useful. The results justify further studies to establish whether the approach can improve procedure outcomes.

Heterosexual couples and prostate cancer support groups: a gender relations analysis.

INTRODUCTION: Men diagnosed with prostate cancer (PCa) can receive supportive care from an array of sources including female partners and prostate cancer support groups (PCSGs). However, little is known about how heterosexual gender relations and supportive care play out among couples who attend PCSGs. Distilling such gender relation patterns is a key to understanding and advancing supportive care for men who experience PCa and their families. PURPOSE: This study describes connections between heterosexual gender relations and PCa supportive care among couples who attend PCSGs. METHOD: In-depth, individual interviews with 30 participants (15 men treated for PCa and their female partners) were analyzed using interpretive descriptive methods. Couples were asked about their relationships, supportive care needs, and attendance at PCSGs. A heterosexual gender relations framework was used to theorize the findings. RESULTS: Findings showed that traditional heterosexual gender relations guided most couples' PCa-related support both in and out of PCSGs. Three themes were inductively derived: "Not pushing too hard"-balancing women's support with men's autonomy, "Confreres"-men supporting men at PCSGs, and "Women are better at reassuring"-support from and for women. CONCLUSIONS: Couples both aligned to and resisted traditional gender roles to accommodate, explain, and rationalize how, as a couple, they approached PCa supportive care needs.

Multi-scale relational graph convolutional network for multiple instance learning in histopathology images.

Graph convolutional neural networks have shown significant potential in natural and histopathology images. However, their use has only been studied in a single magnification or multi-magnification with either homogeneous graphs or only different node types. In order to leverage the multi-magnification information and improve message passing with graph convolutional networks, we handle different embedding spaces at each magnification by introducing the Multi-Scale Relational Graph Convolutional Network (MS-RGCN) as a multiple instance learning method. We model histopathology image patches and their relation with neighboring patches and patches at other scales (i.e., magnifications) as a graph. We define separate message-passing neural networks based on node and edge types to pass the information between different magnification embedding spaces. We experiment on prostate cancer histopathology images to predict the grade groups based on the extracted features from patches. We also compare our MS-RGCN with multiple state-of-the-art methods with evaluations on several source and held-out datasets. Our method outperforms the state-of-the-art on all of the datasets and image types consisting of tissue microarrays, whole-mount slide regions, and whole-slide images. Through an ablation study, we test and show the value of the pertinent design features of the MS-RGCN.

A multi-institutional evaluation of active surveillance for low risk prostate cancer.

PURPOSE: For select men with low risk prostate cancer active surveillance is more often being considered a management strategy. In a multicenter retrospective study we evaluated the actuarial rates and predictors of remaining on active surveillance, the incidence of cancer progression and the pathological findings of delayed radical prostatectomy. MATERIALS AND METHODS: A cohort of 262 men from 4 institutions met the inclusion criteria of age 75 years or younger, prostate specific antigen 10 ng/ml or less, clinical stage T1-T2a, biopsy Gleason sum 6 or less, 3 or less positive cores at diagnostic biopsy, repeat biopsy before active surveillance and no treatment for 6 months following the repeat biopsy. Active surveillance started on the date of the second biopsy. Actuarial rates of remaining on active surveillance were calculated and univariate Cox regression was used to assess predictors of discontinuing active surveillance. RESULTS: With a median followup of 29 months 43 patients ultimately received active treatment. The 2 and 5-year probabilities of remaining on active surveillance were 91% and 75%, respectively. Patients with cancer on the second biopsy (HR 2.23, 95% CI 1.23-4.06, p = 0.007) and a higher number of cancerous cores from the 2 biopsies combined (p = 0.002) were more likely to undergo treatment. Age, prostate specific antigen, clinical stage, prostate volume and number of total biopsy cores sampled were not predictive of outcome. Skeletal metastases developed in 1 patient 38 months after starting active surveillance. Of the 43 patients undergoing delayed treatment 41 (95%) are without disease progression at a median of 23 months following treatment. CONCLUSIONS: With a median followup of 29 months active surveillance for select patients appears to be safe and associated with a low risk of systemic progression. Cancer at restaging biopsy and a higher total number of cancerous cores are associated with a lower likelihood of remaining on active surveillance. A restaging biopsy should be strongly considered to finalize eligibility for active surveillance.

Brenner Tumor of the Ovary: A 10-Year Single Institution Experience and Comprehensive Review of the Literature.

Brenner tumors (BTs) are surface-epithelial stromal cell tumors that are categorized by the World Health Organization as benign, borderline, and malignant. Due to the rarity of BTs, the published literature on these tumors is comprised primarily of case reports and small retrospective studies. We performed a pathology database review spanning the last ten years at our institution revealing nine reported benign BTs. We collected the clinical and pathological data of patients associated with those BTs, describing the clinical presentation and imaging results, and assessing the possible risk factors associated with them. The average age at diagnosis was 58 years. BTs were discovered incidentally in 7/9 cases. The tumor was multifocal and bilateral in 1/9 cases and ranged in size from 0.2 cm to 7.5 cm. Associated Walthard rests were found in 6/9 cases and transitional metaplasia of surface ovarian and/or tubal epithelium was found in 4/9 cases. One patient had an associated mucinous cystadenoma in the ipsilateral ovary. Another patient had an associated mucinous cystadenoma in the contralateral ovary. In conclusion, we found that Walthard rests and transitional metaplasia are common findings in association with BTs. Additionally, pathologists and surgeons need to be aware of the association between mucinous cystadenomas and BTs.

Outcome of patients with biochemical recurrence of prostate cancer after PSMA PET/CT-directed radiotherapy or surgery without systemic therapy.

BACKGROUND: Radiotherapy (RT) and surgery are potential treatment options in patients with biochemical recurrence (BCR) following primary prostate cancer treatment. This study examines the value of prostate-specific membrane antigen (PSMA) positron emission tomography/computed tomography (PET/CT)-informed surgery and RT in patients with BCR treated without systemic therapy. METHODS: This is a post-hoc subgroup analysis of a prospective clinical trial. Inclusion criteria were: histologically proven prostate cancer at initial curative-intent treatment, BCR after primary treatment with curative intent, having five or fewer lesions identified on [18F]DCFPyL PET/CT, and treatment with either PET/CT-directed RT or surgery without systemic therapy. The biochemical progression-free survival after PSMA ligand PET/CT-directed RT and surgery was determined. Uni- and multivariate Cox regression analyses were performed for the association of patients' characteristics, tumor-specific variables, and PSMA PET/CT imaging results with biochemical progression at the last follow-up. RESULTS: Fifty-eight patients (30 in surgery and 28 in radiotherapy groups) met the inclusion criteria. A total of 87 PSMA-positive lesions were detected: 16 local recurrences (18.4%), 54 regional lymph nodes (62.1%), 6 distant lymph nodes (6,8%), and 11 osseous lesions (12.7%). A total of 85.7% (24 of 28) and 70.0% (21 of 30) of patients showed a ≥ 50% decrease in prostate-specific antigen (PSA) levels after RT and surgery, respectively. At a median follow-up time of 21 months (range, 6-32 months), the median biochemical progression-free survival was 19 months (range, 4 to 23 months) in the radiotherapy group, as compared with 16.5 months (range, 4 to 28 months) in the surgery group. On multivariate Cox regression analysis, the number of PSMA positive lesions (2-5 lesions compared to one lesion), and the anatomic location of the detected lesions (distant metastasis vs. local relapse and pelvic nodal relapse) significantly correlated with biochemical progression at the last follow-up, whereas other clinical, tumor-specific, and imaging parameters did not. CONCLUSIONS: This study suggests that RT or surgery based on [18F]DCFPyL PET/CT are associated with high PSA response rates. The number and site of lesions detected on the PSMA PET/CT were predictive of biochemical progression on follow-up. Further studies are needed to assess the impact of targeting these sites on patient relevant outcomes. TRIAL REGISTRATION: Registered September 14, 2016; NCT02899312; https://clinicaltrials.gov/ct2/show/NCT02899312.

Transatlantic Consensus Group on active surveillance and focal therapy for prostate cancer.

UNLABELLED: What's known on the subject? and What does the study add? Active surveillance for prostate cancer is gaining increasing acceptance for low risk prostate cancer. Focal therapy is an emerging tissue preservation strategy that aims for treat only areas of cancer. Early phase trials have shown that side-effects can be significantly reduced using focal therapy. There is significant uncertainty in both active surveillance and focal therapy. This consensus group paper provides a road-map for clinical practice and research for both tissue-preserving strategies in the areas of patient population, tools for risk stratification and cancer localisation, treatment interventions as well as comparators and outcome measures in future comparative trials. OBJECTIVE: To reach consensus on key issues for clinical practice and future research in active surveillance and focal therapy in managing localized prostate cancer. PATIENTS AND METHODS: A group of expert urologists, oncologists, radiologists, pathologists and computer scientists from North America and Europe met to discuss issues in patient population, interventions, comparators and outcome measures to use in both tissue-preserving strategies of active surveillance and focal therapy. Break-out sessions were formed to provide agreement or highlight areas of disagreement on individual topics which were then collated by a writing group into statements that formed the basis of this report and agreed upon by the whole Transatlantic Consensus Group. RESULTS: The Transatlantic group propose that emerging diagnostic tools such as precision imaging and transperineal prostate mapping biopsy can improve prostate cancer care. These tools should be integrated into prostate cancer management and research so that better risk stratification and more effective treatment allocation can be applied. The group envisaged a process of care in which active surveillance, focal therapy, and radical treatments lie on a continuum of complementary therapies for men with a range of disease grades and burdens, rather than being applied in the mutually exclusive and competitive way they are now. CONCLUSION: The changing landscape of prostate cancer epidemiology requires the medical community to re-evaluate the entire prostate cancer diagnostic and treatment pathway in order to minimize harms resulting from over-diagnosis and over-treatment. Precise risk stratification at every point in this pathway is required alongside paradigm shifts in our thinking about what constitutes cancer in the prostate.

Standardized Reporting of Machine Learning Applications in Urology: The STREAM-URO Framework.

The Standardized Reporting of Machine Learning Applications in Urology (STREAM-URO) framework was developed to provide a set of recommendations to help standardize how machine learning studies in urology are reported. This framework serves three purposes: (1) to promote high-quality studies and streamline the peer review process; (2) to enhance reproducibility, comparability, and interpretability of results; and (3) to improve engagement and literacy of machine learning within the urological community.

WATER versus WATER II 2-Year Update: Comparing Aquablation Therapy for Benign Prostatic Hyperplasia in 30-80-cm3 and 80-150-cm3 Prostates.

BACKGROUND: Surgical options are limited when treating large (>80 cm3) prostates for lower urinary tract symptoms (LUTS) due to benign prostatic hyperplasia (BPH). Open simple prostatectomy remains the most common procedure performed for large prostates. There is a need for novel surgical approaches with shorter learning curves and effective treatment. Aquablation could be this novel tool. OBJECTIVE: To compare the outcome of Aquablation for 30-80-cm3 prostates with the outcome for 80-150-cm3 prostates at 2-yr follow-up. DESIGN SETTING AND PARTICIPANTS: We used data from two trials. WATER is a prospective, double-blind, multicenter, international clinical trial comparing the safety and efficacy of Aquablation and transurethral resection of the prostate in the treatment of LUTS/BPH in men aged 45-80 yr with a prostate of 30-80 cm3. WATER II is a prospective, multicenter, single-arm international clinical trial of Aquablation in men with a prostate of 80-150 cm3. INTERVENTION: Aquablation, an ultrasound-guided, robotically executed waterjet ablative procedure. OUTCOME MEASUREMENTS AND STATISTICAL ANALYSIS: We compared 24-mo outcomes between 116 WATER and 101 WATER II study subjects. Student's t test or a Wilcoxon test was used to compare continuous variables and Fisher's test for categorical variables. RESULTS AND LIMITATIONS: The International Prostate Symptom Score (IPSS) reductions at 24 mo was 14.5 points for WATER and 17.4 points for WATER II (p = 0.31). At baseline, the maximum urinary flow rate (Qmax) was 9.4 and 8.7 cm3/s in WATER and WATER II, improving to 20.5 and 18.2 cm3/s, respectively (p = 0.60) at 24 mo. Improvements in both IPSS and Qmax were immediate and sustained throughout follow-up. At 2 yr, the surgical retreatment rate was 4% in WATER and 2% in WATER II. CONCLUSIONS: Aquablation is effective in patients with a prostate of 30-80 cm3 and patients with a prostate of 80-150 cm3 treated for LUTS/BPH, with comparable outcomes in both groups. It has low complication and retreatment rates at 2 yr of follow-up, with durable improvements in functional outcome. PATIENT SUMMARY: Outcomes of Aquablation for both small-to-moderately-sized and large prostates are similar and sustainable at 2 yr of follow-up.

Deep Learning-Based Gleason Grading of Prostate Cancer From Histopathology Images-Role of Multiscale Decision Aggregation and Data Augmentation.

Visual inspection of histopathology images of stained biopsy tissue by expert pathologists is the standard method for grading of prostate cancer (PCa). However, this process is time-consuming and subject to high inter-observer variability. Machine learning-based methods have the potential to improve efficient throughput of large volumes of slides while decreasing variability, but they are not easy to develop because they require substantial amounts of labeled training data. In this paper, we propose a deep learning-based classification technique and data augmentation methods for accurate grading of PCa in histopathology images in the presence of limited data. Our method combines the predictions of three separate convolutional neural networks (CNNs) that work with different patch sizes. This enables our method to take advantage of the greater amount of contextual information in larger patches as well as greater quantity of smaller patches in the labeled training data. The predictions produced by the three CNNs are combined using a logistic regression model, which is trained separately after the CNN training. To effectively train our models, we propose new data augmentation methods and empirically study their effects on the classification accuracy. The proposed method achieves an accuracy of [Formula: see text] in classifying cancerous patches versus benign patches and an accuracy of [Formula: see text] in classifying low-grade (i.e., Gleason grade 3) from high-grade (i.e., Gleason grades 4 and 5) patches. The agreement level of our automatic grading method with expert pathologists is within the range of agreement between pathologists. Our experiments indicate that data augmentation is necessary for achieving expert-level performance with deep learning-based methods. A combination of image-space augmentation and feature-space augmentation leads to the best results. Our study shows that well-designed and properly trained deep learning models can achieve PCa Gleason grading accuracy that is comparable to an expert pathologist.

Improving prostate cancer classification in H&E tissue micro arrays using Ki67 and P63 histopathology.

Histopathology of Hematoxylin and Eosin (H&E)-stained tissue obtained from biopsy is commonly used in prostate cancer (PCa) diagnosis. Automatic PCa classification of digitized H&E slides has been developed before, but no attempts have been made to classify PCa using additional tissue stains registered to H&E. In this paper, we demonstrate that using H&E, Ki67 and p63-stained (3-stain) tissue improves PCa classification relative to H&E alone. We also show that we can infer PCa-relevant Ki67 and p63 information from the H&E slides alone, and use it to achieve H&E-based PCa classification that is comparable to the 3-stain classification. Reported improvements apply to classifying benign vs. malignant tissue, and low grade (Gleason group 2) vs. high grade (Gleason groups 3,4,5) cancer. Specifically, we conducted four classification tasks using 333 tissue samples extracted from 231 radical prostatectomy patients: regression tree-based classification using either (i) 3-stain features, with a benign vs malignant area under the curve (AUC = 92.9%), or (ii) real H&E features and H&E features learned from Ki67 and p63 stains (AUC = 92.4%), as well as deep learning classification using either (iii) real 3-stain tissue patches (AUC = 94.3%) and (iv) real H&E patches and generated Ki67 and p63 patches (AUC = 93.0%) using a deep convolutional generative adversarial network. Classification performance was assessed with Monte Carlo cross validation and quantified in terms of the Area Under the Curve, Brier score, sensitivity, and specificity. Our results are interpretable and indicate that the standard H&E classification could be improved by mimicking other stain types.