Various morphometric measurements of cancer extent on needle prostatic biopsies: which is predictive of pathologic stage and biochemical recurrence following radical prostatectomy?

Prediction of Adverse Pathology at Radical Prostatectomy in Grade Group 2 and 3 Prostate Biopsies Using Machine Learning

PURPOSE

There is ongoing clinical need to improve estimates of disease outcome in prostate cancer. Machine learning (ML) approaches to pathologic diagnosis and prognosis are a promising and increasingly used strategy. In this study, we use an ML algorithm for prediction of adverse outcomes at radical prostatectomy (RP) using whole-slide images (WSIs) of prostate biopsies with Grade Group (GG) 2 or 3 disease.

METHODS

We performed a retrospective review of prostate biopsies collected at our institution which had corresponding RP, GG 2 or 3 disease one or more cores, and no biopsies with higher than GG 3 disease. A hematoxylin and eosin-stained core needle biopsy from each site with GG 2 or 3 disease was scanned and used as the sole input for the algorithm. The ML pipeline had three phases: image preprocessing, feature extraction, and adverse outcome prediction. First, patches were extracted from each biopsy scan. Subsequently, the pre-trained Visual Geometry Group-16 convolutional neural network was used for feature extraction. A representative feature vector was then used as input to an Extreme Gradient Boosting classifier for predicting the binary adverse outcome. We subsequently assessed patient clinical risk using CAPRA score for comparison with the ML pipeline results.

RESULTS

The data set included 361 WSIs from 107 patients (56 with adverse pathology at RP). The area under the receiver operating characteristic curves for the ML classification were 0.72 (95% CI, 0.62 to 0.81), 0.65 (95% CI, 0.53 to 0.79) and 0.89 (95% CI, 0.79 to 1.00) for the entire cohort, and GG 2 and GG 3 patients, respectively, similar to the performance of the CAPRA clinical risk assessment.

CONCLUSION

We provide evidence for the potential of ML algorithms to use WSIs of needle core prostate biopsies to estimate clinically relevant prostate cancer outcomes.

Prognostic Utility of the Gleason Grading System Revisions and Histopathological Factors Beyond Gleason Grade

Background

The International Society of Urological Pathology (ISUP) revised the Gleason system in 2005 and 2014. The impact of these changes on prostate cancer (PCa) prognostication remains unclear.

Objective

To evaluate if the ISUP 2014 Gleason score (GS) predicts PCa death better than the pre-2005 GS, and if additional histopathological information can further improve PCa death prediction.

Patients and Methods

We conducted a case–control study nested among men in the National Prostate Cancer Register of Sweden diagnosed with non-metastatic PCa 1998–2015. We included 369 men who died from PCa (cases) and 369 men who did not (controls). Two uro-pathologists centrally re-reviewed biopsy ISUP 2014 Gleason grading, poorly formed glands, cribriform pattern, comedonecrosis, perineural invasion, intraductal, ductal and mucinous carcinoma, percentage Gleason 4, inflammation, high-grade prostatic intraepithelial neoplasia (HGPIN) and post-atrophic hyperplasia. Pre-2005 GS was back-transformed using i) information on cribriform pattern and/or poorly formed glands and ii) the diagnostic GS from the registry. Models were developed using Firth logistic regression and compared in terms of discrimination (AUC).

Results

The ISUP 2014 GS (AUC = 0.808) performed better than the pre-2005 GS when back-transformed using only cribriform pattern (AUC = 0.785) or both cribriform and poorly formed glands (AUC = 0.792), but not when back-transformed using only poorly formed glands (AUC = 0.800). Similarly, the ISUP 2014 GS performed better than the diagnostic GS (AUC = 0.808 vs 0.781). Comedonecrosis (AUC = 0.811), HGPIN (AUC = 0.810) and number of cores with ≥50% cancer (AUC = 0.810) predicted PCa death independently of the ISUP 2014 GS.

Conclusion

The Gleason Grading revisions have improved PCa death prediction, likely due to classifying cribriform patterns, rather than poorly formed glands, as Gleason 4. Comedonecrosis, HGPIN and number of cores with ≥50% cancer further improve PCa death discrimination slightly.

Extent of High-Grade Prostatic Adenocarcinoma in Multiparametric Magnetic Resonance Imaging-Targeted Biopsy Enhances Prediction of Pathologic Stage

CONTEXT.—

Multiparametric magnetic resonance imaging (mpMRI) of prostate with targeted biopsy has enhanced detection of high-grade prostatic adenocarcinoma (HG PCa). However, utility of amount of HG PCa (Gleason pattern 4/5) in mpMRI-targeted biopsies versus standard 12-core biopsies in predicting adverse outcomes on radical prostatectomy (RP) is unknown.

OBJECTIVE.—

To examine the utility of amount of HG PCa in mpMRI-targeted biopsies versus standard 12-core biopsies in predicting adverse RP outcomes.

DESIGN.—

We performed a retrospective review of prostate biopsies, which had corresponding RP, 1 or more mpMRI-targeted biopsy, and grade group 2 disease or higher. For the 169 cases identified, total millimeters of carcinoma and HG PCa, and longest length HG PCa in a single core were recorded for 12-core biopsies and each set of mpMRI-targeted biopsies. For RP specimens, Gleason grade, extraprostatic extension, seminal vesicle involvement, and lymph node metastasis were recorded. The main outcome studied was prostate-confined disease at RP. A logistic regression model was used to test which pre-RP variables related to this outcome.

RESULTS.—

Univariate analysis showed significant associations with adverse RP outcomes in 5 of 8 quantifiable variables; longest millimeter HG PCa in a single 12-core biopsy, highest grade group in any core, and total millimeter HG in mpMRI-targeted biopsies showed no statistical association (P = .54, P = .13, and P = .55, respectively). In multivariate analysis, total millimeter carcinoma in all cores, highest GrGrp in any core, and longest millimeter HG PCa in a single mpMRI-targeted core provided additional predictive value (P < .001, P = .004, and P = .03, respectively).

CONCLUSIONS.—

Quantitation of HG PCa in mpMRI-targeted biopsies provides additional value over 12-core biopsies alone in predicting nonorgan confined prostate cancer at RP. Linear millimeters of HG PCa in mpMRI-targeted biopsies is a significant parameter associated with higher pathologic stage and could be of value in risk models.

The Importance of Tumor Length in Needle Biopsies of the Prostate

OBJECTIVES

To form a composite predictor variable that combines the effects of tumor length, serum prostate-specific antigen (PSA), and International Society of Urologic Pathologists (ISUP) grade on the observation of adverse prostatectomy pathology.

METHODS

Logistic regression analysis was used to demonstrate how tumor length, serum PSA, and ISUP grade related to adverse prostatectomy results and to derive weighting factors for a composite variable, cx.

RESULTS

The composite variable, cx, relates closely to adverse prostatectomy results as well as to observed PSA failure.

CONCLUSIONS

The composite variable cx uses preoperative information that may allow the sorting of patients into low, intermediate, and high risk for adverse outcomes after prostatectomy.

Automated Gleason Scoring and Tumor Quantification in Prostate Core Needle Biopsy Images Using Deep Neural Networks and Its Comparison with Pathologist-Based Assessment

The Gleason grading system, currently the most powerful prognostic predictor of prostate cancer, is based solely on the tumor’s histological architecture and has high inter-observer variability. We propose an automated Gleason scoring system based on deep neural networks for diagnosis of prostate core needle biopsy samples. To verify its efficacy, the system was trained using 1133 cases of prostate core needle biopsy samples and validated on 700 cases. Further, system-based diagnosis results were compared with reference standards derived from three certified pathologists. In addition, the system’s ability to quantify cancer in terms of tumor length was also evaluated via comparison with pathologist-based measurements. The results showed a substantial diagnostic concordance between the system-grade group classification and the reference standard (0.907 quadratic-weighted Cohen’s kappa coefficient). The system tumor length measurements were also notably closer to the reference standard (correlation coefficient, R = 0.97) than the original hospital diagnoses (R = 0.90). We expect this system to assist pathologists to reduce the probability of over- or under-diagnosis by providing pathologist-level second opinions on the Gleason score when diagnosing prostate biopsy, and to support research on prostate cancer treatment and prognosis by providing reproducible diagnosis based on the consistent standards.

Cancer core length from targeted biopsy: an index of prostate cancer volume and pathological stage

OBJECTIVE

To study the relationship of maximum cancer core length (MCCL), on targeted biopsy (TB) of magnetic resonance imaging (MRI)-visible index lesions, to volume of that tumour found at radical prostatectomy (RP).

PATIENTS AND METHODS

In all, 205 men undergoing fusion biopsy and RP were divided into two groups: 136 in whom the MCCL came from an index MRI-visible lesion (TB) and 69 in whom MCCL came from a non-targeted lesion (non-targeted biopsy [NTB]). MRI was 3-T multi-parametric and biopsy was via MRI-ultrasonography fusion.

RESULTS

In the TB group, MCCL correlated with volume of clinically significant index tumours (ρ = 0.44-0.60, P < 0.01). The correlation was similar for first and repeat biopsy and for transition and peripheral zone lesions (ρ = 0.42-0.49, P < 0.01). No correlations were found in the NTB group. TB MCCL (6-10 and >10 mm) and MRI lesion diameter (>20 mm) were independently associated with tumour volume. TB MCCLs >10 mm and Gleason scores >7 were each associated with pathological T3 disease (odds ratios 5.73 and 5.04, respectively), but MRI lesion diameter lesion was not.

CONCLUSIONS

MCCL on a TB from an MRI-visible lesion is an independent predictor of both cancer volume and pathological stage. This relationship does not exist for MCCL from a NTB core. Quantifying CCL on MRI-TBs may have a value, not previously described, to risk-stratify patients with prostate cancer before treatment.

The Learning Curve for Magnetic Resonance Imaging/Ultrasound Fusion-guided Prostate Biopsy

BACKGROUND

Magnetic resonance imaging/ultrasound-guided fusion biopsy (FBx) is more accurate at detecting clinically significant prostate cancer than conventional transrectal ultrasound-guided systematic biopsy. However, learning curves for attaining accuracy may limit the generalizability of published outcomes.

OBJECTIVE

To delineate and quantify the learning curve for FBx by assessing the targeted biopsy accuracy and pathological quality of systematic biopsy over time.

DESIGN, SETTING, AND PARTICIPANTS

We carried out a retrospective analysis of 173 consecutive men who underwent Artemis FBx with computer-template systematic sampling between July 2015 and May 2017.

OUTCOME MEASUREMENTS AND STATISTICAL ANALYSIS

The accuracy of targeted biopsy was determined by calculating the distance between planned and actual core trajectories stored on Artemis. Systematic sampling proficiency was assessed via pathological analysis of fibromuscular tissue in all cores and then comparing pathology elements from individual cores from men in the first and last tertiles. Polynomial linear regression models, change-point analysis, and piecewise linear regression were used to quantify the learning curve.

RESULTS AND LIMITATION

A significant improvement in targeted biopsy accuracy occurred up to 98 cases (p<0.01). There was a significant decrease in fibromuscular tissue in the systematic biopsy cores up to 84 cases (p<0.01) and an improvement in pathological quality when comparing systematic cores from the first and third tertiles. Use of a different fusion platform may limit the generalizability of our results.

CONCLUSIONS

There is a significant learning curve for targeted and systemic biopsy using the Artemis platform. Improvements in accuracy of targeted biopsy and better sampling for systematic biopsy can be achieved with greater experience.

PATIENT SUMMARY

We define the learning curve for magnetic resonance imaging/ultrasound-guided fusion biopsy (FBx) using targeted biopsy accuracy and systematic core sampling quality as measures. Our findings underscore the importance of overcoming learning curves inherent to FBx to minimize patient discomfort and biopsy risk and improve the quality of care for accurate risk stratification, active surveillance, and treatment selection.

Prostate cancer volume estimations based on transrectal ultrasonography-guided biopsy in order to predict clinically significant prostate cancer

INTRODUCTION

Tumor diameter is a reliable parameter to estimate tumor volume in solid organ cancers; its use in prostate cancer is controversial since it exhibits a more irregular pattern of growth. This study aimed to examine the association between the tumor volume estimations based on transrectal ultrasound (TRUS) guided biopsy results and the tumor volume measured on the pathological specimen.

MATERIALS AND METHODS

A total of 237 patients who underwent radical retropubic prostatectomy (RRP) were included in this retrospective study. The differences and correlations between cancer volume estimations based on TRUS guided biopsy findings and cancer volume estimations based on post-prostatectomy pathology specimens were examined. In addition, diagnostic value of TRUS guided biopsy-based volume estimations in order to predict clinically significant cancer (>0.5 cc) were calculated.

RESULTS

The mean cancer volume estimated using TRUS biopsy results was lower (5.5±6.5 cc) than the mean cancer volume calculated using prostatectomy specimens (6.4±7.6 cc) (p<0.041).TRUS guided biopsy examination resulted in 5 false positive and 15 false negative cases. There was a significant but weak correlation between the two parameters (r=0.62, p<0.001). The sensitivity and specificity of TRUS guided biopsy in predicting the presence of clinically significant cancer was 93.4% (95% CI, 89.1-96.1) and 50.0% (95% CI, 20.1-79.9), respectively.

CONCLUSIONS

TRUS guided biopsy-derived estimations seem to have a limited value to predict pathologically established tumor volume. Further studies are warranted to identify additional methods that may more accurately predict actual pathological characteristics and prognosis of prostate cancer.

Predictive criteria of insignificant prostate cancer: what is the correspondence of linear extent to percentage of cancer in a single core?

Objective

The aim of active surveillance of early prostate cancer is to individualize therapy by selecting for curative treatment only patients with significant cancer. Epstein’s criteria for prediction of clinically insignificant cancer in surgical specimens are widely used. Epstein’s criterion “no single core with >50% cancer” has no correspondence in linear extent. The aim of this study is to find a possible correspondence.

Materials and Methods

From a total of 401 consecutive patients submitted to radical prostatectomy, 17 (4.2%) met criteria for insignificant cancer in the surgical specimen. The clinicopathologic findings in the correspondent biopsies were compared with Epstein’s criteria for insignificant cancer. Cancer in a single core was evaluated in percentage as well as linear extent in mm.

Results

Comparing the clinicopathologic findings with Epstein’s criteria predictive of insignificant cancer, there was 100% concordance for clinical stage T1c, no Gleason pattern 4 or 5, ≤2 cores with cancer, and no single core with >50% cancer. However, only 25% had density ≤0.15. The mean, median and range of the maximum length of cancer in a single core in mm were 1.19, 1, and 0.5-2.5, respectively. Additionally, the mean, median, and range of length of cancer in all cores in mm were 1.47, 1.5, and 0.5-3, respectively.

Conclusion

To pathologists that use Epstein’s criteria predictive of insignificant cancer and measure linear extent in mm, our study favors that “no single core with >50% cancer” may correspond to >2.5 mm in linear extent.

The critical role of the pathologist in determining eligibility for active surveillance as a management option in patients with prostate cancer: consensus statement with recommendations supported by the College of American Pathologists, International Society of Urological Pathology, Association of Directors of Anatomic and Surgical Pathology, the New Zealand Society of Pathologists, and the Prostate Cancer Foundation

CONTEXT

Prostate cancer remains a significant public health problem. Recent publications of randomized trials and the US Preventive Services Task Force recommendations have drawn attention to overtreatment of localized, low-risk prostate cancer. Active surveillance, in which patients undergo regular visits with serum prostate-specific antigen tests and repeat prostate biopsies, rather than aggressive treatment with curative intent, may address overtreatment of low-risk prostate cancer. It is apparent that a greater awareness of the critical role of pathologists in determining eligibility for active surveillance is needed.

OBJECTIVES

To review the state of current knowledge about the role of active surveillance in the management of prostate cancer and to provide a multidisciplinary report focusing on pathologic parameters important to the successful identification of patients likely to succeed with active surveillance, to determine the role of molecular tests in increasing the safety of active surveillance, and to provide future directions.

DESIGN

Systematic review of literature on active surveillance for low-risk prostate cancer, pathologic parameters important for appropriate stratification, and issues regarding interobserver reproducibility. Expert panels were created to delineate the fundamental questions confronting the clinical and pathologic aspects of management of men on active surveillance.

RESULTS

Expert panelists identified pathologic parameters important for management and the related diagnostic and reporting issues. Consensus recommendations were generated where appropriate.

CONCLUSIONS

Active surveillance is an important management option for men with low-risk prostate cancer. Vital to this process is the critical role pathologic parameters have in identifying appropriate candidates for active surveillance. These findings need to be reproducible and consistently reported by surgical pathologists with accurate pathology reporting.

Greatest percentage of involved core length and the risk of death from prostate cancer in men with highest Gleason score ≥ 7

INTRODUCTION/BACKGROUND

Men with highest GS ≥ 7 and a differing, lower GS core (ComboGS) have decreased PC-specific mortality (PCSM) risk after RT or RT and androgen deprivation therapy (ADT). Whether the greatest percentage of involved core length (GPC) modulates this risk is unknown.

PATIENTS AND METHODS

Men with GS ≥ 7 PC (n = 333) consecutively treated between December 1989 and July 2000 using RT (n = 268; 80%) or RT and 6 months of ADT (n = 65; 20%) comprised the study cohort. The GPC was calculated using biopsy core and tumor lengths. We used competing risks regression to assess whether increasing GPC was associated with increased PCSM risk in men with or without ComboGS adjusting for risk group, age, and treatment.

RESULTS

After a median follow-up of 5.36 years (interquartile range, 3.22-7.61 years), 92 (28%) men died, 28 (30%) of PC. Increasing GPC was significantly associated with increased risk of PCSM (adjusted hazard ratio, 1.02; 95% confidence interval, 1.01-1.03; P = .005). Men with GPC ≥ 50% versus < 50% had significantly greater PCSM estimates when ComboGS was present (P < .001) versus absent (P = .55). Of the 127 men with ComboGS and GPC < 50%, 83% were treated with RT alone and 2 PC deaths were observed; neither in men with GS 7 and favorable intermediate-risk PC.

CONCLUSION

Men treated with RT for ComboGS, GPC < 50%, GS 7, and favorable intermediate-risk PC have a very low risk of early PCSM. The RTOG 0815 trial will establish whether ADT is necessary to optimize curability in these men.

Urinary engrailed-2 (EN2) levels predict tumour volume in men undergoing radical prostatectomy for prostate cancer

What's known on the subject? and What does the study add? There are a lot of potential prostate cancer biomarkers being evaluated. All aim to improve on the sensitivity and specificity of PSA. EN2 was recently shown by our group to have better sensitivity and specificity than PSA. EN2 is a simple ELISA test and is not dependent on other parameters, even PSA, unlike all the other current biomarkers under evaluation. To date, no marker correlates with the amount of cancer present - the present study shows this positive correlation with EN2 in men undergoing prostatectomy. The potential utility of this work is that by knowing that the level of EN2 corresponds to the amount of cancer present, irrelevant of tumour grade and number of cancer foci, we can define an EN2 level corresponding to small cancers, which can then undergo surveillance. We are conducting a further study that is aimed at determining whether the levels of EN2 in urine can indicate 'significant' vs 'non-significant cancer' using the threshold of 0.5 mL cancer (after Epstein's work).

OBJECTIVES

To evaluate the relationship between levels of a recently described prostate cancer biomarker engrailed-2 (EN2) in urine and cancer volume in men who had undergone radical prostatectomy (RP) for prostate cancer. To date, prostate-specific antigen (PSA) levels have not reliably predicted prostate cancer volume. Reliable volume indicator biomarker(s) may aid management decisions, e.g. active treatment vs active surveillance.

PATIENTS AND METHODS

Archived patient samples from the Aarhus Prostate Cancer Project, Denmark, were assessed. Pre-treatment mid-stream urines, without preceding prostatic massage, were collected and stored at -80 °C. Urinary EN2 levels were measured by a recently published enzyme-linked immunosorbent assay.

RESULTS

In all, 88 of the whole cohort of 125 men (70%) were positive for EN2 in their urine (>42.5 µg/L); 38/58 (65%) men where cancer volume data was available. There was no statistical relationship between urinary EN2 levels and serum PSA levels. PSA levels did not correlate with tumour stage, combined Gleason grade, total prostatic weight or cancer volume. There was a strong statistical relationship between urinary EN2 and prostate cancer volume by linear regression (P = 0.006). Higher EN2 levels correlated with tumour stage T1 vs T2 (P = 0.027).

CONCLUSIONS

Pre-surgical urinary EN2 levels were associated with increasing tumour stage and closely reflected the volume of cancer in RP specimens. Given the ease of collection (no prostatic massage required) and the simplicity, low cost and robustness of the assay, EN2 may become a useful biomarker in not only identifying which patients have prostate cancer but may also facilitate risk stratification by indicating the burden of tumour volume.

Prognostic significance of tumor volume in radical prostatectomy and needle biopsy specimens

PURPOSE

This review addresses the controversies that persist relating to the prognosis and reporting of tumor volume in adenocarcinoma of the prostate.

MATERIALS AND METHODS

A search was performed using the MEDLINE database and referenced lists of relevant studies to obtain articles addressing the quantification of cancer on radical prostatectomy and needle biopsy.

RESULTS

In the 2010 TNM classification system T2 tumor at radical prostatectomy is subdivided into pT2a (unilateral tumor occupying less than ½ a lobe), pT2b (unilateral tumor greater than ½ a lobe) and pT2c (bilateral tumor). This pathological substaging of T2 disease fails on several accounts. In most studies pT2b disease almost does not exist. By the time a tumor is so large that it microscopically occupies more than ½ a lobe, in the majority of cases there is bilateral (pT2c) tumor. An even greater flaw of the substaging system for stage pT2 disease is the lack of prognostic significance. In reporting pathologically organ confined cancer, it should be merely noted as pT2 without further subclassification. The data are conflicting as to the independent prognostic significance of objective measurements of tumor volume in radical prostatectomy specimens. The most likely explanation for the discordant results lies in the strong correlation of tumor volume with other prognostic markers such as extraprostatic extension and positive margins. In studies where it is statistically significant on multivariate analysis, it is unlikely that knowing tumor volume improves prediction of prognosis beyond routinely reported parameters to the degree that it would be clinically useful for an individual patient. An alternative is to record tumor volume as minimal, moderate or extensive, which gives some indication to the urologist as to the extent of disease. Not only does providing an objective measurement not add useful prognostic information beyond what is otherwise routinely reported by the pathologist, but many objective measurements done in routine practice will likely not be an accurate indicator of the true tumor volume. There is also a lack of consensus regarding the best method of measuring tumor length when there are multiple foci in a single core separated by benign intervening prostatic stroma. Some pathologists, this author included, consider discontinuous foci of cancer as if it was 1 uninterrupted focus, the rationale being that these discontinuous foci are undoubtedly the same cancer going in and out of the plane of section. Measuring the cancer from where it starts to where it ends on the core gives the minimal length of cancer in the prostate. Others measure each focus individually, and the sum of these measurements is considered the cancer length on the core. Quantifying cancer with an ocular micrometer to record the total length or percent length of cancer is time-consuming, and the data are conflicting whether this is superior to other, simpler methods and whether any potential differences in predictive accuracy would translate into changes in clinical management. It is recommended that at a minimum the number of positive cores be recorded, unless fragmented involved cores preclude evaluation, along with at least 1 other more detailed measurement such as the percent of core involvement or length of cancer.

CONCLUSIONS

Consensus has been reached on some of the issues relating to quantifying tumor volume in prostate cancer, such as the lack of utility of substaging pT2 disease. Other questions such as whether to include or subtract intervening benign prostate tissue on prostate needle cores will require additional studies. Finally, matters such as the need to quantify cancer at radical prostatectomy or which method of quantifying cancer on needle biopsy is superior will likely remain contentious due to the close interrelationship and redundancy of prognostic variables.