Artificial intelligence for diagnosis and grading of prostate cancer in biopsies: a population-based, diagnostic study.

BACKGROUND: An increasing volume of prostate biopsies and a worldwide shortage of urological pathologists puts a strain on pathology departments. Additionally, the high intra-observer and inter-observer variability in grading can result in overtreatment and undertreatment of prostate cancer. To alleviate these problems, we aimed to develop an artificial intelligence (AI) system with clinically acceptable accuracy for prostate cancer detection, localisation, and Gleason grading. METHODS: We digitised 6682 slides from needle core biopsies from 976 randomly selected participants aged 50-69 in the Swedish prospective and population-based STHLM3 diagnostic study done between May 28, 2012, and Dec 30, 2014 (ISRCTN84445406), and another 271 from 93 men from outside the study. The resulting images were used to train deep neural networks for assessment of prostate biopsies. The networks were evaluated by predicting the presence, extent, and Gleason grade of malignant tissue for an independent test dataset comprising 1631 biopsies from 246 men from STHLM3 and an external validation dataset of 330 biopsies from 73 men. We also evaluated grading performance on 87 biopsies individually graded by 23 experienced urological pathologists from the International Society of Urological Pathology. We assessed discriminatory performance by receiver operating characteristics and tumour extent predictions by correlating predicted cancer length against measurements by the reporting pathologist. We quantified the concordance between grades assigned by the AI system and the expert urological pathologists using Cohen's kappa. FINDINGS: The AI achieved an area under the receiver operating characteristics curve of 0·997 (95% CI 0·994-0·999) for distinguishing between benign (n=910) and malignant (n=721) biopsy cores on the independent test dataset and 0·986 (0·972-0·996) on the external validation dataset (benign n=108, malignant n=222). The correlation between cancer length predicted by the AI and assigned by the reporting pathologist was 0·96 (95% CI 0·95-0·97) for the independent test dataset and 0·87 (0·84-0·90) for the external validation dataset. For assigning Gleason grades, the AI achieved a mean pairwise kappa of 0·62, which was within the range of the corresponding values for the expert pathologists (0·60-0·73). INTERPRETATION: An AI system can be trained to detect and grade cancer in prostate needle biopsy samples at a ranking comparable to that of international experts in prostate pathology. Clinical application could reduce pathology workload by reducing the assessment of benign biopsies and by automating the task of measuring cancer length in positive biopsy cores. An AI system with expert-level grading performance might contribute a second opinion, aid in standardising grading, and provide pathology expertise in parts of the world where it does not exist. FUNDING: Swedish Research Council, Swedish Cancer Society, Swedish eScience Research Center, EIT Health.

Utility of Pathology Imagebase for standardisation of prostate cancer grading.

AIMS: Despite efforts to standardise grading of prostate cancer, even among experts there is still a considerable variation in grading practices. In this study we describe the use of Pathology Imagebase, a novel reference image library, for setting an international standard in prostate cancer grading. METHODS AND RESULTS: The International Society of Urological Pathology (ISUP) recently launched a reference image database supervised by experts. A panel of 24 international experts in prostate pathology reviewed independently microphotographs of 90 cases of prostate needle biopsies with cancer. A linear weighted kappa of 0.67 (95% confidence interval = 0.62-0.72) and consensus was reached in 50 cases. The interobserver weighted kappa varied from 0.48 to 0.89. The highest level of agreement was seen for Gleason score (GS) 3 + 3 = 6 (ISUP grade 1), while higher grades and particularly GS 4 + 3 = 7 (ISUP grade 3) showed considerable disagreement. Once a two-thirds majority was reached, images were moved automatically into a public database available for all ISUP members at www.isupweb.org. Non-members are able to access a limited number of cases. CONCLUSIONS: It is anticipated that the database will assist pathologists to calibrate their grading and, hence, decrease interobserver variability. It will also help to identify instances where definitions of grades need to be clarified.

Gleason grade 4 prostate adenocarcinoma patterns: an interobserver agreement study among genitourinary pathologists.

AIMS: To assess the interobserver reproducibility of individual Gleason grade 4 growth patterns. METHODS AND RESULTS: Twenty-three genitourinary pathologists participated in the evaluation of 60 selected high-magnification photographs. The selection included 10 cases of Gleason grade 3, 40 of Gleason grade 4 (10 per growth pattern), and 10 of Gleason grade 5. Participants were asked to select a single predominant Gleason grade per case (3, 4, or 5), and to indicate the predominant Gleason grade 4 growth pattern, if present. 'Consensus' was defined as at least 80% agreement, and 'favoured' as 60-80% agreement. Consensus on Gleason grading was reached in 47 of 60 (78%) cases, 35 of which were assigned to grade 4. In the 13 non-consensus cases, ill-formed (6/13, 46%) and fused (7/13, 54%) patterns were involved in the disagreement. Among the 20 cases where at least one pathologist assigned the ill-formed growth pattern, none (0%, 0/20) reached consensus. Consensus for fused, cribriform and glomeruloid glands was reached in 2%, 23% and 38% of cases, respectively. In nine of 35 (26%) consensus Gleason grade 4 cases, participants disagreed on the growth pattern. Six of these were characterized by large epithelial proliferations with delicate intervening fibrovascular cores, which were alternatively given the designation fused or cribriform growth pattern ('complex fused'). CONCLUSIONS: Consensus on Gleason grade 4 growth pattern was predominantly reached on cribriform and glomeruloid patterns, but rarely on ill-formed and fused glands. The complex fused glands seem to constitute a borderline pattern of unknown prognostic significance on which a consensus could not be reached.

Baseline Prostate Atrophy is Associated with Reduced Risk of Prostate Cancer in Men Undergoing Repeat Prostate Biopsy.

PURPOSE: We evaluated whether the presence and severity of baseline prostate atrophy in men with initial biopsy negative for prostate cancer was associated the risk of subsequent prostate cancer detection in a clinical trial with scheduled study mandated biopsies. MATERIALS AND METHODS: We retrospectively analyzed the records of 3,084 men 50 to 75 years old with prostate specific antigen between 2.5 and 10 ng/ml, and a prior negative biopsy in the placebo arm of the REDUCE (Reduction by Dutasteride of Prostate Cancer Events) study who completed at least 1 per-protocol biopsy. Prostate cancer (defined as present or absent) and prostate atrophy (graded as absent, mild or moderate/marked) was assessed by central pathology review. The association of baseline atrophy with positive 2 and 4-year repeat biopsies was evaluated with logistic regression, controlling for baseline covariates. RESULTS: Baseline prostate atrophy was detected in 2,143 men (70%) and graded as mild and moderate/marked in 1,843 (60%) and 300 (10%) baseline biopsies, respectively. Patients with atrophy were older and had a larger prostate, and more acute and chronic prostate inflammation. At 2-year biopsy the prostate cancer incidence was 17% (508 cases). Baseline atrophy was significantly associated with lower prostate cancer risk (univariable and multivariable OR 0.60, 95% CI 0.50-0.74 and OR 0.67, 95% CI 0.54-0.83, p <0.001, respectively) at the 2-year biopsy. These results were similar at the 4-year biopsy (univariable and multivariable OR 0.70, 95% CI 0.53-0.93 and OR 0.72, 95% CI 0.53-0.97, p = 0.03, respectively). Relative to no atrophy the prostate cancer risk at the 2-year repeat biopsy was lower for mild atrophy (OR 0.69, 95% CI 0.55-0.86) and moderate/marked atrophy (OR 0.51, 95% CI 0.34-0.76, each p <0.001). CONCLUSIONS: Baseline prostate atrophy in men with a prostate biopsy negative for prostate cancer was independently associated with subsequent lower prostate cancer detection.

Correlation of HOXD3 promoter hypermethylation with clinical and pathologic features in screening prostate biopsies.

BACKGROUND: Molecular markers that can discriminate indolent cancers from aggressive ones may improve the management of prostate cancer and minimize unnecessary treatment.Aberrant DNA methylation is a common epigenetic event in cancers and HOXD3 promoter hypermethylation (H3PH) has been found in prostate cancer. Our objective was to evaluate the relationship between H3PH and clinicopathologic features in screening prostate biopsies. METHODS: Ninety-two patients who underwent a prostate biopsy at our institution between October 2011 and May 2012 were included in this study. The core with the greatest percentage of the highest grade disease was analyzed for H3PH by methylation-specific PCR. Correlational analysis was used to analyze the relationship between H3PH and various clinical parameters. Chi-square analysis was used to compare H3PH status between benign and malignant disease. RESULTS: Of the 80 biopsies with HOXD3 methylation status assessable, 66 sets were confirmed to have cancer. In the 14 biopsies with benign disease there was minimal H3PH with the mean percentage of methylation reference (PMR) of 0.7%. In contrast, the HOXD3 promoter was hypermethylated in 16.7% of all cancers and in 50% of high risk tumors with an average PMR of 4.3% (P=0.008). H3PH was significantly correlated with age (P=0.013), Gleason score (P=0.031) and the maximum involvement of the biopsy core (P=0.035). CONCLUSIONS: H3PH is associated with clinicopathologic features. The data indicate that H3PH is more common in older higher risk patients. More research is needed to determine the role of this marker in optimizing management strategies in men with newly diagnosed prostate cancer.

Intraductal carcinoma of the prostate: interobserver reproducibility survey of 39 urologic pathologists.

The diagnosis of intraductal carcinoma (IDC) of the prostate remains subjective because 3 sets of diagnostic criteria are in use. An internet survey was compiled from 38 photomicrographs showing duct proliferations: 14 signed out as high-grade prostatic intraepithelial neoplasia (HGPIN), 17 IDC, and 7 invasive cribriform/ductal carcinoma. Each image was assessed for the presence of 9 histologic criteria ascribed to IDC. Thirty-nine respondents were asked to rate images as (1) benign/reactive, (2) HGPIN, (3) borderline between HGPIN and IDC, (4) IDC, or (5) invasive cribriform/ductal carcinoma. Intraclass correlation coefficient was 0.68. There was 70% overall agreement with HGPIN, 43% with IDC, and 73% with invasive carcinoma (P < .001, χ(2)). Respondents considered 19 (50%) of 38 cases as IDC candidates, of which 5 (26%) had a two-thirds consensus for IDC; two-thirds consensus for either borderline or IDC was reached in 9 (47%). Two-thirds consensus other than IDC was reached in the remaining 19 of 38 cases, with 15 supporting HGPIN and 4 supporting invasive carcinoma. Findings that differed across diagnostic categories were lumen-spanning neoplastic cells (P < .001), 2× benign duct diameters (P < .001), duct space contours (round, irregular, and branched) (P < .001), papillary growth (P = .048), dense cribriform or solid growth (both P = .023), and comedonecrosis (P = .015). When the 19 of 38 images that attained consensus for HGPIN or invasive carcinoma were removed from consideration, lack of IDC consensus was most often attributable to only loose cribriform growth (5/19), central nuclear maturation (5/19), or comedonecrosis (3/19). Of the 9 histologic criteria, only 1 retained significant correlation with a consensus diagnosis of IDC: the presence of solid areas (P = .038). One case that attained IDC consensus had less than 2× duct enlargement yet still had severe nuclear atypia and nucleomegaly. Six fold nuclear enlargement was not significant (P = .083), although no image had both 6× nuclei and papillary or loose cribriform growth: a combination postulated as sufficient criteria for IDC. Finally, 20.5% of respondents agreed that an isolated diagnosis of IDC on needle biopsy warrants definitive therapy, 20.5% disagreed, and 59.0% considered the decision to depend upon clinicopathologic variables. Although IDC diagnosis remains challenging, we propose these criteria: a lumen-spanning proliferation of neoplastic cells in preexisting ducts with a dense cribriform or partial solid growth pattern. Solid growth, in any part of the duct space, emerges as the most reproducible finding to rule in a diagnosis of IDC. Comedonecrosis is a rarer finding, but in most cases, it should rule in IDC. Duct space enlargement to greater than 2× the diameter of the largest, adjacent benign spaces is usually present in IDC, although there may be rare exceptions.

Effect of dutasteride on the risk of prostate cancer.

BACKGROUND: We conducted a study to determine whether dutasteride reduces the risk of incident prostate cancer, as detected on biopsy, among men who are at increased risk for the disease. METHODS: In this 4-year, multicenter, randomized, double-blind, placebo-controlled, parallel-group study, we compared dutasteride, at a dose of 0.5 mg daily, with placebo. Men were eligible for inclusion in the study if they were 50 to 75 years of age, had a prostate-specific antigen (PSA) level of 2.5 to 10.0 ng per milliliter, and had had one negative prostate biopsy (6 to 12 cores) within 6 months before enrollment. Subjects underwent a 10-core transrectal ultrasound-guided biopsy at 2 and 4 years. RESULTS: Among 6729 men who underwent a biopsy or prostate surgery, cancer was detected in 659 of the 3305 men in the dutasteride group, as compared with 858 of the 3424 men in the placebo group, representing a relative risk reduction with dutasteride of 22.8% (95% confidence interval, 15.2 to 29.8) over the 4-year study period (P<0.001). Overall, in years 1 through 4, among the 6706 men who underwent a needle biopsy, there were 220 tumors with a Gleason score of 7 to 10 among 3299 men in the dutasteride group and 233 among 3407 men in the placebo group (P=0.81). During years 3 and 4, there were 12 tumors with a Gleason score of 8 to 10 in the dutasteride group, as compared with only 1 in the placebo group (P=0.003). Dutasteride therapy, as compared with placebo, resulted in a reduction in the rate of acute urinary retention (1.6% vs. 6.7%, a 77.3% relative reduction). The incidence of adverse events was similar to that in studies of dutasteride therapy for benign prostatic hyperplasia, except that in our study, as compared with previous studies, the relative incidence of the composite category of cardiac failure was higher in the dutasteride group than in the placebo group (0.7% [30 men] vs. 0.4% [16 men], P=0.03). CONCLUSIONS: Over the course of the 4-year study period, dutasteride reduced the risk of incident prostate cancer detected on biopsy and improved the outcomes related to benign prostatic hyperplasia. (ClinicalTrials.gov number, NCT00056407.)

Radiofrequency identification specimen tracking in anatomical pathology: pilot study of 1067 consecutive prostate biopsies.

Improved methods such as radiofrequency identification (RFID) are needed to optimize specimen tracking in anatomical pathology. We undertook a study of RFID in an effort to optimize specimen tracking and patient identification, including the following: (1) creation of workflow process maps, (2) evaluation of existing RFID hardware technologies, (3) creation of Web-based software to support the RFID-enabled workflow, and (4) assessment of the impact with a series of prostate biopsies. We identified multiple steps in the workflow process in which RFID enhanced specimen tracking. Multiple product choices were found that could withstand the harsh heat and chemical environments encountered in pathology processing, and software that was compatible with our laboratory information system was designed in-house. A total of 1067 prostate biopsies were received, and 78.3% were successfully processed with the RFID system. Radiofrequency identification allowed dynamic specimen tracking throughout the workflow process in anatomical pathology.

Adequate histologic sectioning of prostate needle biopsies.

No standard method exists for sampling prostate needle biopsies, although most reports claim to embed 3 cores per block and obtain 3 slices from each block. This study was undertaken to determine the extent of histologic sectioning necessary for optimal examination of prostate biopsies. We prospectively compared the impact on cancer yield of submitting 1 biopsy core per cassette (biopsies from January 2010) with 3 cores per cassette (biopsies from August 2010) from a large national reference laboratory. Between 6 and 12 slices were obtained with the former 1-core method, resulting in 3 to 6 slices being placed on each of 2 slides; for the latter 3-core method, a limit of 6 slices was obtained, resulting in 3 slices being place on each of 2 slides. A total of 6708 sets of 12 to 18 core biopsies were studied, including 3509 biopsy sets from the 1-biopsy-core-per-cassette group (January 2010) and 3199 biopsy sets from the 3-biopsy-cores-percassette group (August 2010). The yield of diagnoses was classified as benign, atypical small acinar proliferation, high-grade prostatic intraepithelial neoplasia, and cancer and was similar with the 2 methods: 46.2%, 8.2%, 4.5%, and 41.1% and 46.7%, 6.3%, 4.4%, and 42.6%, respectively (P = .02). Submission of 1 core or 3 cores per cassette had no effect on the yield of atypical small acinar proliferation, prostatic intraepithelial neoplasia, or cancer in prostate needle biopsies. Consequently, we recommend submission of 3 cores per cassette to minimize labor and cost of processing.

Optimized 18F-FDG PET-CT Method to Improve Accuracy of Diagnosis of Metastatic Cancer.

The diagnosis of cancer by FDG PET-CT is often inaccurate owing to subjectivity of interpretation. We compared the accuracy of a novel normalized (standardized) method of interpretation with conventional non-normalized SUV. Patients (n = 393) with various malignancies were studied with FDG PET/CT to determine the presence or absence of cancer. Target lesions were assessed by two methods: (1) conventional SUVmax (conSUVmax) and (2) a novel method that combined multiple factors to optimize SUV (optSUVmax), including the patient's normal liver SUVmax, a liver constant (k) derived from a review of the literature, and use of site-specific thresholds for malignancy. The two methods were compared to pathology findings in 154 patients being evaluated for mediastinal and/or hilar lymph node (MHLNs) metastases, 143 evaluated for extra-thoracic lymph node (ETLNs) metastases, and 96 evaluated for liver metastases. OptSUVmax was superior to conSUVmax for all patient groups. For MHLNs, sensitivity was 83.8% vs. 80.7% and specificity 88.7% vs. 9.6%, respectively; for ETLNs, sensitivity was 92.1% vs. 77.8% and specificity 80.1% vs. 27.6%, respectively; and for lesions in the liver parenchyma, sensitivity was 96.1% vs. 82.3% and specificity 88.8% vs. 23.0%, respectively. Optimized SUVmax increased diagnostic accuracy of FDG PET-CT for cancer when compared with conventional SUVmax interpretation.

Diagnostic performance of PCA3 to detect prostate cancer in men with increased prostate specific antigen: a prospective study of 1,962 cases.

PURPOSE: The detection of prostate cancer relies primarily on abnormal digital rectal examination or increased serum prostate specific antigen concentration. However, low positive predictive values result in many men with increased prostate specific antigen and/or suspicious digital rectal examination having a negative biopsy. We investigated the value of the PCA3 (prostate cancer gene 3) urine test in predicting the likelihood of diagnosis of cancer before biopsy. MATERIALS AND METHODS: We performed a prospective, community based clinical trial to evaluate PCA3 score before any biopsy. This trial was conducted at 50 urology practices in the United States. Samples were obtained from 1,962 men with increased serum prostate specific antigen (greater than 2.5 ng/ml) and/or abnormal digital rectal examination before transrectal prostate needle biopsy. Study samples (urinary PCA3 and biopsies) were processed and analyzed by a central laboratory. Sensitivity-specificity analyses were conducted. RESULTS: A total of 1,913 urine samples (97.5%) were adequate for PCA3 testing. Of 802 cases diagnosed with prostate cancer 222 had high grade prostatic intraepithelial neoplasia or atypical small acinar proliferation and were suspicious for cancer, whereas 889 cases were benign. The traditional PCA3 cutoff of 35 reduced the number of false-positives from 1,089 to 249, a 77.1% reduction. However, false-negatives (missed cancers) increased significantly from 17 to 413, an increase of more than 2,300%. Lowering the PCA3 cutoff to 10 reduced the number of false-positives 35.4% and false-negatives only increased 5.6%. CONCLUSIONS: Urinary PCA3 testing in conjunction with prostate specific antigen has the potential to significantly decrease the number of unnecessary prostate biopsies.

Precursors of prostate cancer.

High-grade prostatic intraepithelial neoplasia (PIN) is the only accepted precursor of prostatic adenocarcinoma, according to numerous studies of animal models and man; other proposed precursors include atrophy and malignancy-associated changes (with no morphologic changes). PIN is characterized by progressive abnormalities of phenotype and genotype that are intermediate between benign prostatic epithelium and cancer, indicating impairment of cell differentiation and regulatory control with advancing stages of prostatic carcinogenesis. The only method of detection of PIN is biopsy because it does not significantly elevate serum prostate-specific antigen concentration and cannot be detected by ultrasonography. The mean incidence of PIN in biopsies is 9% (range, 4%-16%), representing about 115,000 new cases of isolated PIN diagnosed each year in the United States. The clinical importance of PIN is its high predictive value as a marker for adenocarcinoma, and its identification warrants repeat biopsy for concurrent or subsequent carcinoma, especially when multifocal or observed in association with atypical small acinar proliferation (ASAP). Carcinoma develops in most patients with PIN within 10 years. Androgen deprivation therapy and radiation therapy decrease the prevalence and extent of PIN, suggesting that these forms of treatment may play a role in prevention of subsequent cancer. Multiple clinical trials to date of men with PIN have had modest success in delaying or preventing subsequent cancer.

Staging of prostate cancer.

Prostatic carcinoma (PCa) is a significant cause of cancer morbidity and mortality worldwide. Accurate staging is critical for prognosis assessment and treatment planning for PCa. Despite the large volume of clinical activity and research, the challenge to define the most appropriate and clinically relevant staging system remains. The pathologically complex and uncertain clinical course of prostate cancer further complicates the design of staging classification and a substaging system suitable for individualized care. This review will focus on recent progress and controversial issues related to prostate cancer staging. The 2010 revision of the American Joint Committee on Cancer/Union Internationale Contre le Cancer (AJCC/UICC) tumour, node and metastasis (TNM) system is the most widely used staging system at this time. Despite general acceptance of the system as a whole, there is controversy and uncertainty about its application, particularly for T2 subclassification. The three-tiered T2 classification system for organ-confined prostate cancer is superfluous, considering the biology and anatomy of PCa. A tumour size-based substaging system may be considered in the future TNM subclassification of pT2 cancer. Lymph node status is one of the most important prognostic factors for prostate cancer. Nevertheless, clinical outcomes in patients with positive lymph nodes are variable. Identification of patients at the greatest risk of systemic progression helps in the selection of appropriate therapy. The data suggest that the inherent aggressiveness of metastatic prostate cancer is closely linked to the tumour volume of lymph node metastasis. We recommend that a future TNM staging system should consider subclassification of node-positive cancer on the basis of nodal cancer volume, using the diameter of the largest nodal metastasis and/or the number of positive nodes.

Understanding the pathological features of focality, grade and tumour volume of early-stage prostate cancer as a foundation for parenchyma-sparing prostate cancer therapies: active surveillance and focal targeted therapy.

OBJECTIVE: • To better understand the biology and incidence of the cancer foci within the prostate through a comprehensive literature review and a review of our own data, to establish the current level of knowledge regarding the pathological foundation for active surveillance (AS) or focal therapy (FT). PATIENTS AND METHODS: • A systematic review of the literature was performed, searching PubMed® from January 1994 to July 2009. • Electronic searches were limited to the English language using the keywords 'prostate cancer', 'histopathology', 'radical prostatectomy', 'pathological stage' and 'focal therapy'. • The authors' own data were also analysed and are presented. RESULTS: • Recent data have shown a significant pathological stage migration towards earlier disease comprising unilateral pT2a/b prostate cancer (PCa). • The cancer volume of the clinically significant tumour (index lesion) has been proposed as a driving force of PCa progression and therefore should be identified and treated at an early stage. • In general, most satellite lesions do not appear to be life-threatening. • Other pathological features, such as Gleason score, extraprostatic extension and the spatial distribution of PCa within the prostate, remain important selective criteria for AS or FT. CONCLUSION: • The present study reviews the current knowledge of cancer focality, aggression and tumour volume. Further research is needed to better understand the biologic behaviour of each of the tumour foci within a cancerous prostate, and to employ this information to selected patients for no therapy (AS), parenchyma-preserving approaches (FT) or whole gland radical therapy.

Prostatic stromal proliferations: a review.

Prostatic stromal proliferations account for the majority of benign tumour-like lesions in the prostate. The most common is nodular hyperplasia, seen in a majority of elderly men. Diagnostic difficulty is encountered with some variants, including stromal hyperplasia with atypia, characterised by degenerative changes of myofibroblasts. In contrast with benign stromal tumours, malignant stromal tumours of the prostate are rare, accounting for less than 0.1% of all prostatic malignancies. The most common are rhabdomyosarcoma (paediatric) and leiomyosarcoma (adults); others include phyllodes tumour and stromal sarcoma. Some authors lump malignant tumours with poor outcome (e.g., phyllodes tumour and stromal sarcoma) with benign stromal tumours (e.g., stromal hyperplasia with atypia, leiomyoma), considering them collectively to be of uncertain malignant potential, but this approach is discouraged. This review presents a contemporary approach to classification and diagnosis of prostatic stromal tumours.

Atypical sclerosing adenosis of the prostate: a rare mimic of adenocarcinoma.

AIMS: Sclerosing adenosis of the prostate is a benign, small, acinar proliferation in dense spindle cell stroma, with a distinct immunohistochemical profiles. It is incidentally found in about 2% of transurethral resection specimens. The aim was to describe cases with significant cytological atypia mimicking cancer, which have not been previously reported. METHODS AND RESULTS: We describe five cases of sclerosing adenosis with significant cytological atypia, referred to as atypical sclerosing adenosis (ASA), which were initially considered suspicious or diagnostic of adenocarcinoma. Seven other cases of typical sclerosing adenosis were used as controls. All cases of typical and atypical sclerosing adenosis displayed an intact basal cell layer, which was immunoreactive for high-molecular-weight keratin, S100 protein, smooth muscle actin, and prostate-specific antigen, with no differences between ASA and the control group. Alpha-methylacyl-coenzyme A racemase was negative. Three of four cases of ASA had aneuploid DNA content by digital image analysis. All cases of typical sclerosing adenosis were diploid. During a mean follow-up of 33 months (range 5-73 months), none developed recurrence or prostatic cancer. CONCLUSIONS: ASA is an unusual small, acinar proliferation of the prostate that may be mistaken for adenocarcinoma, and should be distinguished from other mimics, including atypical adenomatous hyperplasia, mesonephric remnant hyperplasia, and post-atrophic hyperplasia. ASA is a benign lesion and aggressive treatment is unwarranted.

Identification of areas of grading difficulties in prostate cancer and comparison with artificial intelligence assisted grading.

The International Society of Urological Pathology (ISUP) hosts a reference image database supervised by experts with the purpose of establishing an international standard in prostate cancer grading. Here, we aimed to identify areas of grading difficulties and compare the results with those obtained from an artificial intelligence system trained in grading. In a series of 87 needle biopsies of cancers selected to include problematic cases, experts failed to reach a 2/3 consensus in 41.4% (36/87). Among consensus and non-consensus cases, the weighted kappa was 0.77 (range 0.68-0.84) and 0.50 (range 0.40-0.57), respectively. Among the non-consensus cases, four main causes of disagreement were identified: the distinction between Gleason score 3 + 3 with tangential cutting artifacts vs. Gleason score 3 + 4 with poorly formed or fused glands (13 cases), Gleason score 3 + 4 vs. 4 + 3 (7 cases), Gleason score 4 + 3 vs. 4 + 4 (8 cases) and the identification of a small component of Gleason pattern 5 (6 cases). The AI system obtained a weighted kappa value of 0.53 among the non-consensus cases, placing it as the observer with the sixth best reproducibility out of a total of 24. AI may serve as a decision support and decrease inter-observer variability by its ability to make consistent decisions. The grading of these cancer patterns that best predicts outcome and guides treatment warrants further clinical and genetic studies. Results of such investigations should be used to improve calibration of AI systems.

Intraductal carcinoma of the prostate is an aggressive form of invasive carcinoma and should be graded.

Infiltration of the prostatic ducts by prostatic adenocarcinoma occurs relatively frequently, being most commonly associated with high grade disease. It is now recognised that intraductal carcinoma of the prostate (IDCP) has an associated poor prognosis and this is reflected in its histological, molecular and immunohistochemical features. The current recommendation of the World Health Organization is that IDCP not be taken into consideration when grading prostate adenocarcinoma. It is apparent that Gleason did not differentiate between IDCP and stromal invasive carcinoma when developing and validating his grading system, and recent studies suggest that the incorporation of IDCP grading into the overall grading of the specimen provides additional prognostic information.