Artificial intelligence for detection of prostate cancer in biopsies during active surveillance.

OBJECTIVES: To evaluate a cancer detecting artificial intelligence (AI) algorithm on serial biopsies in patients with prostate cancer on active surveillance (AS). PATIENTS AND METHODS: A total of 180 patients in the Prostate Cancer Research International Active Surveillance (PRIAS) cohort were prospectively monitored using pre-defined criteria. Diagnostic and re-biopsy slides from 2011 to 2020 (n = 4744) were scanned and analysed by an in-house AI-based cancer detection algorithm. The algorithm was analysed for sensitivity, specificity, and for accuracy to predict need for active treatment. Prognostic properties of cancer size, prostate-specific antigen (PSA) level and PSA density at diagnosis were evaluated. RESULTS: The sensitivity and specificity of the AI algorithm was 0.96 and 0.73, respectively, for correct detection of cancer areas. Original pathology report diagnosis was used as the reference method. The area of cancer estimated by the pathologists correlated highly with the AI detected cancer size (r = 0.83). By using the AI algorithm, 63% of the slides would not need to be read by a pathologist as they were classed as benign, at the risk of missing 0.55% slides containing cancer. Biopsy cancer content and PSA density at diagnosis were found to be prognostic of whether the patient stayed on AS or was discontinued for active treatment. CONCLUSION: The AI-based biopsy cancer detection algorithm could be used to reduce the pathologists' workload in an AS cohort. The detected cancer amount correlated well with the cancer length measured by the pathologist and the algorithm performed well in finding even small areas of cancer. To our knowledge, this is the first report on an AI-based algorithm in digital pathology used to detect cancer in a cohort of patients on AS.

Outcomes of active surveillance for Japanese patients with prostate cancer (PRIAS-JAPAN).

OBJECTIVE: To report the outcomes of repeat biopsies, metastasis and survival in the Prostate Cancer Research International: Active Surveillance (PRIAS)-JAPAN study, a prospective observational study for Japanese patients, initiated in 2010. PATIENTS AND METHODS: At the beginning, inclusion criteria were initially low-risk patients, prostate-specific antigen (PSA) density (PSAD) <0.2, and ≤2 positive biopsy cores. As from 2014, GS3+4 has also been allowed for patients aged 70 years and over. Since January 2021, the age limit for Gleason score (GS) 3 + 4 cases was removed, and eligibility criteria were expanded to PSA ≤20 ng/mL, PSAD <0.25 nd/mL/cc, unlimited number of positive GS 3 + 3 cores, and positive results for fewer than half of the total number of cores for GS 3 + 4 cases if magnetic resonance imaging fusion biopsy was performed at study enrolment or subsequent follow-up. For patients eligible for active surveillance, PSA tests were performed every 3 months, rectal examination every 6 months, and biopsies at 1, 4, 7 and 10 years, followed by every 5 years thereafter. Patients with confirmed pathological reclassification were recommended for secondary treatments. RESULTS: As of February 2024, 1302 patients were enrolled in AS; 1274 (98%) met the eligibility criteria. The median (interquartile range) age, PSA level, PSAD, and number of positive cores were 69 (64-73) years, 5.3 (4.5-6.6) ng/mL, 0.15 (0.12-0.17) ng/mL, and 1 (1-2), respectively. The clinical stage was T1c in 1089 patients (86%) and T2 in 185 (15%). The rates of acceptance by patients for the first, second, third and fourth re-biopsies were 83%, 64%, 41% and 22%, respectively. The pathological reclassification rates for the first, second, third and fourth re-biopsies were 29%, 30%, 35% and 25%, respectively. The 1-, 5- and 10-year persistence rates were 77%, 45% and 23%, respectively. Six patients developed metastasis, and one patient died from prostate cancer. CONCLUSION: Pathological reclassification was observed in approximately 30% of the patients during biopsy; however, biopsy acceptance rates decreased over time. Although metastasis occurred in six patients, only one death from prostate cancer was recorded.

Digital Image Correlation of Forescatter Detector Images for Simultaneous Strain and Orientation Mapping.

Improved plasticity models require simultaneous experimental local strain and microstructural evolution data. Microscopy tools, such as electron backscatter diffraction (EBSD), that can monitor transformation at the relevant length-scale, are often incompatible with digital image correlation (DIC) techniques required to determine local deformation. In this paper, the viability of forescatter detector (FSD) images as the basis for the DIC study is investigated. Standard FSD and an integrated EBSD/FSD approach (Pattern Region of Interest Analysis System: PRIAS™) are analyzed. Simultaneous strain and microstructure maps are obtained for tensile deformation of Q&P 1180 steel up to ~14% strain. Tests on an undeformed sample that is simply shifted indicate a standard deviation of error in strain of around 0.4% without additional complications from a deformed surface. The method resolves strain bands at ~2 µm spacing but does not provide significant sub-grain strain resolution. Similar resolution was obtained for mechanically polished and electropolished samples, despite electropolished surfaces presenting a smoother, simpler topography. While the resolution of the PRIAS approach depends upon the EBSD step size, the 80 nm step size used provides seemingly similar resolution as 8,000× (22.7 nm) FSD images. Surface feature evolution prevents DIC analysis across large strain steps (>6% strain), but restarting DIC, using an FSD reference image from an interim strain step, allows reasonable DIC across the stress­strain curve. Furthermore, the data are obtained easily and provide complementary information for EBSD analysis.

Clinically insignificant prostate cancer suitable for active surveillance according to Prostate Cancer Research International: Active surveillance criteria: Utility of PI-RADS v2.

BACKGROUND: Active surveillance (AS) is an important treatment strategy for prostate cancer (PCa). Prostate Imaging-Reporting and Data System (PI-RADS) v2 has been addressed, but few studies have reported the value of PI-RADS v2 for assessing risk stratification in patients with PCa, especially on selecting potential candidates for AS. PURPOSE: To investigate the utility of PI-RADS v2 and apparent diffusion coefficient (ADC) in evaluating patients with insignificant PCa, who are suitable for AS. STUDY TYPE: Retrospective. SUBJECTS: In all, 238 patients with PCa who met the Prostate Cancer Research International: Active Surveillance criteria underwent radical prostatectomy. FIELD STRENGTH/SEQUENCE: 3.0T, including T2 -weighted, diffusion-weighted, and dynamic contrast-enhanced imaging. ASSESSMENT: Insignificant cancer was defined histopathologically as an organ-confined disease with a tumor volume <0.5 cm3 without Gleason score 4-5. Patients were divided into two groups based on the PI-RADS v2 and tumor ADC: A, PI-RADS score ≤3 and ADC ≥1.095 × 10-3 mm2 /s; and B, PI-RADS score 4-5 or ADC <1.095 × 10-3 mm2 /s. Preoperative clinical and imaging variables were evaluated regarding the associations with insignificant cancer. RESULTS: Of the 238 patients, 101 (42.8%) were diagnosed with insignificant cancer on pathological findings. The number of positive cores, prostate-specific antigen density (PSAD), PI-RADS v2 and tumor ADC were significantly associated with insignificant cancer on univariate analysis (P < 0.05). However, multivariate analysis indicated tumor ADC (odds ratio [OR] = 4.57, P < 0.001) and PI-RADS v2 (OR = 3.60, P < 0.001) were independent predictors of insignificant cancer. Area under the receiver operating characteristics curve (AUC) reached 0.803 when PI-RADS v2 (AUC = 0.747) was combined with tumor ADC (AUC = 0.786). DATA CONCLUSION: The PI-RADS v2 together with tumor ADC may be a useful marker for predicting patients with insignificant PCa when considering AS. LEVEL OF EVIDENCE: 4 Technical Efficacy: Stage 2 J. Magn. Reson. Imaging 2018;47:1072-1079.

Primary Gleason pattern upgrading in contemporary patients with D'Amico low-risk prostate cancer: implications for future biomarkers and imaging modalities.

OBJECTIVE: To retrospectively assess the rate of high-grade primary Gleason upgrading (HGPGU) to primary Gleason pattern 4 or 5 in a contemporary cohort of patients with D'Amico low-risk prostate cancer including those who fulfilled Prostate Cancer Research International Active Surveillance (PRIAS) criteria, and to develop a tool for HGPGU prediction. HGPGU is a contraindication in most active surveillance (AS) and focal therapy protocols. PATIENTS AND METHODS: In all, 10 616 patients with localised prostate cancer were treated at a high-volume European tertiary care centre from 2010 to 2015 with radical prostatectomy. Analyses were restricted to 1 819 patients with D'Amico low-risk prostate cancer (17.1%) with prostate-specific antigen (PSA) levels of <10.0 ng/mL, cT1c-cT2a and Gleason score ≤6, and were repeated within 772 of the men (7.3%) who fulfilled the PRIAS criteria for AS (PSA level of ≤10 ng/mL, T1c-T2, Gleason score ≤6, PSA density (PSAD) of <0.2 ng/mL2 , ≤2 positive cores). Uni- and multivariable logistic regression models were fitted, testing predictors of HGPGU. The final logistic regression model was based on the most informative variables. RESULTS: There was HGPGU in 88 (4.8%) patients with D'Amico low-risk prostate cancer and in 32 (4.1%) of the subgroup who were PRIAS eligible. Multivariable analysis predicting HGPGU for the patients with D'Amico low-risk yielded three independent predictors: age, PSAD, and clinical tumour stage (P = 0.008, P = 0.005 and P = 0.021, respectively). Within the same patients, the model using all vs the most informative variables resulted in area under the curves (AUCs) of 69.2% and 68.3%, respectively. Multivariable analysis of those who were PRIAS eligible, yielded age and number of positive cores as independent predictors of HGPGU (P = 0.002 and P = 0.049, respectively; AUC 64.9%). CONCLUSIONS: The low accuracy (invariably <70%) for HGPGU prediction in both patients with D'Amico low-risk prostate cancer and PRIAS eligibility indicates that these variables have poor predictive ability in contemporary patients. Despite HGPGU being a rare phenomenon, it may have life threatening implications and consequently alternatives such as biomarkers, genetic markers, or imaging modalities at re-biopsy are needed.

A proposal of a new nomogram for predicting upstaging in contemporary D'Amico low-risk prostate cancer patients.

PURPOSE: Unfavorable prostate cancer (PCa) disease at final pathology affects at least 10 % of D'Amico low-risk patients. Thus, conservative therapies including active surveillance may be wrongfully applied. The purposes were to assess the rate of upstaging in a contemporary cohort of D'Amico low-risk PCa patients and to develop and externally validate a nomogram as upstaging prediction tool in two European cohorts. METHODS: Analyses were restricted to 2007 patients who harbored low-risk PCa at ≥10-cores initial biopsy according to D'Amico classification (PSA <10.0 ng/ml, Gleason score <7 and clinical stage ≤T2a). Patients underwent radical prostatectomy at a high-volume center in Hamburg, Germany, from 2010 to 2015. The Hamburg cohort was randomly divided into development (n = 1338) and validation cohorts (n = 669). The development cohort was used to devise a nomogram predicting upstaging, defined as presence of ≥pT3 and/or lymph node invasion. The nomogram was externally validated in two European validation cohorts (Hamburg, n = 669; Milan, n = 465). RESULTS: Upstaging was observed in 187/1338 (14.0 %) of low-risk patients. In multivariable models, four of ten tested variables achieved independent predictor status: age (OR 1.07, 95 % CI 1.04-1.09), PSA (OR 1.21, 95 % CI 1.12-1.31), prostate volume (OR 0.97, 95 % CI 0.96-0.98) and percentage of positive cores (OR 1.02, 95 % CI 1.01-1.03). In external validation, the nomogram demonstrated 70.8 % (Hamburg) and 70.0 % (Milan) accuracy, respectively, with excellent concordance between predicted and observed values. CONCLUSIONS: Our proposed nomogram is capable to accurately identify D'Amico low-risk patients at risk of upstaging, utilizing four routinely available clinical variables, age, PSA, prostate volume and percentage of positive biopsy cores. Unfavorable prostate cancer disease at final pathology affects at least 10 % of D'Amico low-risk patients. Thus, we developed and externally validated a new nomogram based on contemporary low-risk prostate cancer patients to accurately identify D'Amico low-risk patients at risk of upstaging. It utilizes four routine variables, age, PSA, prostate volume and percentage of positive biopsy cores.

Tumor lesion diameter on diffusion weighted magnetic resonance imaging could help predict insignificant prostate cancer in patients eligible for active surveillance: preliminary analysis.

PURPOSE: We analyzed the pathological outcomes of candidates for active surveillance according to tumor lesion diameter on diffusion weighted magnetic resonance imaging. MATERIALS AND METHODS: We retrospectively analyzed 188 candidates for active surveillance who had undergone diffusion weighted magnetic resonance imaging before radical prostatectomy between 2006 and 2012. We measured the diameter of the suspicious tumor lesion on diffusion weighted magnetic resonance imaging and stratified the cohort into 2 groups. Group 1 included patients with normal magnetic resonance imaging or a suspicious tumor lesion smaller than 1 cm and group 2 included patients with a suspicious tumor lesion larger than 1 cm. We compared pathological outcomes including insignificant prostate cancer in each group and analyzed whether different tumor diameters resulted in a change in insignificant prostate cancer rates. RESULTS: Group 1 consisted of 115 (61.2%) patients and group 2 included 73 (38.8%) patients. In group 1 magnetic resonance imaging was normal in 72 patients. Mean ± SD diameter of suspicious tumor lesions was 12.0 ± 5.58 mm. Tumor volume was significantly different between the groups (0.73 ± 0.86 vs 1.09 ± 1.07 cm(3), p = 0.018), as was the rate of insignificant prostate cancer (48.7% vs 24.7%, p = 0.001). The rate of insignificant prostate cancer decreased as tumor diameter increased over 1 cm. On multivariate logistic regression analysis the diameter of suspicious tumor lesions was an important predictor of insignificant prostate cancer (OR 0.319, p = 0.014). CONCLUSIONS: Our analysis demonstrates that the simple measurement of the diameter of suspicious tumor lesions on diffusion weighted magnetic resonance imaging could improve the prediction of insignificant prostate cancer in candidates for active surveillance.

Electron imaging with an EBSD detector.

Electron Backscatter Diffraction (EBSD) has proven to be a useful tool for characterizing the crystallographic orientation aspects of microstructures at length scales ranging from tens of nanometers to millimeters in the scanning electron microscope (SEM). With the advent of high-speed digital cameras for EBSD use, it has become practical to use the EBSD detector as an imaging device similar to a backscatter (or forward-scatter) detector. Using the EBSD detector in this manner enables images exhibiting topographic, atomic density and orientation contrast to be obtained at rates similar to slow scanning in the conventional SEM manner. The high-speed acquisition is achieved through extreme binning of the camera-enough to result in a 5 × 5 pixel pattern. At such high binning, the captured patterns are not suitable for indexing. However, no indexing is required for using the detector as an imaging device. Rather, a 5 × 5 array of images is formed by essentially using each pixel in the 5 × 5 pixel pattern as an individual scattered electron detector. The images can also be formed at traditional EBSD scanning rates by recording the image data during a scan or can also be formed through post-processing of patterns recorded at each point in the scan. Such images lend themselves to correlative analysis of image data with the usual orientation data provided by and with chemical data obtained simultaneously via X-Ray Energy Dispersive Spectroscopy (XEDS).

Percentage of cancer involvement in positive cores can predict unfavorable disease in men with low-risk prostate cancer but eligible for the prostate cancer international: active surveillance criteria.

OBJECTIVES: To identify predictive factors of unfavorable disease and of biochemical failure in patients treated with radical prostatectomy but eligible for active surveillance (AS) according to Prostate Cancer Research International: Active Surveillance (PRIAS) criteria. We aimed to introduce and validate the percentage of cancer involvement in positive cores (CIPC) as potential worse predictive factor. METHODS: From January 2002 to December 2007, 750 consecutive subjects underwent radical prostatectomy at a single institution. We identified 147 (19.05%) patients who were eligible for AS based on PRIAS criteria: clinical stage T1c or T2 disease, prostate-specific antigen level of ≤ 10 ng/ml, Gleason score ≤ 6, prostate-specific antigen-D of<0.2 ng/ml(2), and fewer than 3 positive biopsy cores. CIPC was included in the analysis. RESULTS: Of the 147 patients, 95 (66.43%) patients had favorable disease, whereas 48 (33.57%) had unfavorable disease. In multivariate logistic regression, maximum cancer length (odds ratio 12.52, P<0.01) and CIPC (odds ratio 1.70, P<0.01) represented independent predictors of unfavorable prostate cancer. The area under the receiver operating characteristics curve analysis revealed significantly higher performance after including CIPC to the PRIAS criteria (0.61 vs. 0.94, P<0.01). A cutoff of 0.4mm of CIPC was set to predict unfavorable disease with 93% specificity, 76% sensibility, and 87% accuracy based on the receiver operating characteristics curve analysis. Finally, the 3- and 5-years biochemical recurrence (BCR)-free survival were significantly lower in subjects with CIPC ≥ 0.4mm, 88.4 % and 81.0% vs. 97.8% and 95.7%, respectively (P< 0.01). CONCLUSIONS: Our findings suggest that the inclusion of CIPC to the prostate biopsy features could be helpful to avoid misclassification in patients eligible for AS according to the PRIAS criteria.