The morphological interpretation of histologic sections forms the basis of diagnosis and prognostication for cancer. In the diagnosis of carcinomas, pathologists perform a semiquantitative analysis of a small set of morphological features to determine the cancer's histologic grade. Physicians use histologic grade to inform their assessment of a carcinoma's aggressiveness and a patient's prognosis. Nevertheless, the determination of grade in breast cancer examines only a small set of morphological features of breast cancer epithelial cells, which has been largely unchanged since the 1920s. A comprehensive analysis of automatically quantitated morphological features could identify characteristics of prognostic relevance and provide an accurate and reproducible means for assessing prognosis from microscopic image data. We developed the C-Path (Computational Pathologist) system to measure a rich quantitative feature set from the breast cancer epithelium and stroma (6642 features), including both standard morphometric descriptors of image objects and higher-level contextual, relational, and global image features. These measurements were used to construct a prognostic model. We applied the C-Path system to microscopic images from two independent cohorts of breast cancer patients [from the Netherlands Cancer Institute (NKI) cohort, n = 248, and the Vancouver General Hospital (VGH) cohort, n = 328]. The prognostic model score generated by our system was strongly associated with overall survival in both the NKI and the VGH cohorts (both log-rank P ≤ 0.001). This association was independent of clinical, pathological, and molecular factors. Three stromal features were significantly associated with survival, and this association was stronger than the association of survival with epithelial characteristics in the model. These findings implicate stromal morphologic structure as a previously unrecognized prognostic determinant for breast cancer.

For prostate cancer patients, the Gleason score is one of the most important prognostic factors, potentially determining treatment independent of the stage. However, Gleason scoring is based on subjective microscopic examination of tumor morphology and suffers from poor reproducibility. Here we present a deep learning system (DLS) for Gleason scoring whole-slide images of prostatectomies. Our system was developed using 112 million pathologist-annotated image patches from 1226 slides, and evaluated on an independent validation dataset of 331 slides. Compared to a reference standard provided by genitourinary pathology experts, the mean accuracy among 29 general pathologists was 0.61 on the validation set. The DLS achieved a significantly higher diagnostic accuracy of 0.70 ( p = 0.002) and trended towards better patient risk stratification in correlations to clinical follow-up data. Our approach could improve the accuracy of Gleason scoring and subsequent therapy decisions, particularly where specialist expertise is unavailable. The DLS also goes beyond the current Gleason system to more finely characterize and quantitate tumor morphology, providing opportunities for refinement of the Gleason system itself.

Despite the advantages of providing an early presumptive diagnosis, fungal classification by histopathology can be difficult and may lead to diagnostic error. To assess the accuracy of histologic diagnosis of fungal infections vs culture ("gold standard"), we performed a 10-year retrospective review at our institution. Of the 47 of 338 positive mold and yeast cultures with concurrent surgical pathology evaluation without known history of a fungal infection, 37 (79%) were correctly identified based on morphologic features in histologic and/or cytologic specimens. The 10 discrepant diagnoses (21%) included misidentification of septate and nonseptate hyphal organisms and yeast forms. Errors resulted from morphologic mimics, use of inappropriate terminology, and incomplete knowledge in mycology. The accuracy did not correlate with preceding antifungal therapy (P = .14) or use of special stains (P = .34) and was not operator-dependent. Among 8 discrepancies with clinical follow-up available, 2 potential adverse clinical consequences resulted. While histopathologic identification of fungi in tissue sections and cytologic preparations is prone to error, implementation of a standardized reporting format should improve diagnostic accuracy and prevent adverse outcomes.

Aim To describe a group of distinct low‐grade oncocytic renal tumours that demonstrate CD117 negative/cytokeratin (CK) 7‐positive immunoprofile. Methods and results We identified 28 such tumours from four large renal tumour archives. We performed immunohistochemistry for: CK7, CD117, PAX8, CD10, AMACR, e‐cadherin, CK20, CA9, AE1/AE3, vimentin, BerEP4, MOC31, CK5/6, p63, HMB45, melan A, CD15 and FH. In 14 cases we performed array CGH, with a successful result in nine cases. Median patient age was 66 years (range 49–78 years) with a male‐to‐female ratio of 1:1.8. Median tumour size was 3 cm (range 1.1–13.5 cm). All were single tumours, solid and tan‐brown, without a syndromic association. On microscopy, all cases showed solid and compact nested growth. There were frequent areas of oedematous stroma with loosely arranged cells. The tumour cells had oncocytic cytoplasm with uniformly round to oval nuclei, but without significant irregularities, and showed only focal perinuclear halos. Negative CD117 and positive CK7 reactivity were present in all cases (in two cases there was focal and very weak CD117 reactivity). Uniform reactivity was found for PAX8, AE1/AE3, e‐cadherin, BerEP4 and MOC31. Negative stains included CA9, CK20, vimentin, CK5/6, p63, HMB45, Melan A and CD15. CD10 and AMACR were either negative or focally positive; FH was retained. On array CGH, there were frequent deletions at 19p13.3 (seven of nine), 1p36.33 (five of nine) and 19q13.11 (four of nine); disomic status was found in two of nine cases. On follow‐up (mean 31.8 months, range 1–118), all patients were alive with no disease progression. Conclusion Low‐grade oncocytic tumours that are CD117‐negative/CK7‐positive demonstrate consistent and readily recognisable morphology, immunoprofile and indolent behaviour.

A unique renal neoplasm characterized by eosinophilic cytoplasm and solid and cystic growth was recently reported in patients with tuberous sclerosis complex (TSC). We searched multiple institutional archives and consult files in an attempt to identify a sporadic counterpart. We identified 16 morphologically identical cases, all in women, without clinical features of TSC. The median age was 57 years (range, 31 to 75 y). Macroscopically, tumors were tan and had a solid and macrocystic (12) or only solid appearance (4). Average tumor size was 50 mm (median, 38.5 mm; range, 15 to 135 mm). Microscopically, the tumors showed solid areas admixed with variably sized macrocysts and microcysts that were lined by cells with a pronounced hobnail arrangement. The cells had voluminous eosinophilic cytoplasm with prominent granular cytoplasmic stippling and round to oval nuclei with prominent nucleoli. Scattered histiocytes and lymphocytes were invariably present. Thirteen of 16 patients were stage pT1; 2 were pT2, and 1 was pT3a. The cells demonstrated a distinct immunoprofile: nuclear PAX8 expression, predominant CK20-positive/CK7-negative phenotype, patchy AMACR staining, but no CD117 reactivity. Thirteen of 14 patients with follow-up were alive and without disease progression after 2 to 138 months (mean: 53 mo; median: 37.5 mo); 1 patient died of other causes. Although similar to a subset of renal cell carcinomas (RCCs) seen in TSC, we propose that sporadic "eosinophilic, solid, and cystic RCC," which occurs predominantly in female individuals and is characterized by distinct morphologic features, predominant CK20-positive/CK7-negative immunophenotype, and indolent behavior, represents a novel subtype of RCC.

IMPORTANCE For prostate cancer, Gleason grading of the biopsy specimen plays a pivotal role in determining case management. However, Gleason grading is associated with substantial interobserver variability, resulting in a need for decision support tools to improve the reproducibility of Gleason grading in routine clinical practice. OBJECTIVE To evaluate the ability of a deep learning system (DLS) to grade diagnostic prostate biopsy specimens. DESIGN, SETTING, AND PARTICIPANTS The DLS was evaluated using 752 deidentified digitized images of formalin-fixed paraffin-embedded prostate needle core biopsy specimens obtained from 3 institutions in the United States, including 1 institution not used for DLS development. To obtain the Gleason grade group (GG), each specimen was first reviewed by 2 expert urologic subspecialists from a multi-institutional panel of 6 individuals (years of experience: mean, 25 years; range, 18-34 years). A third subspecialist reviewed discordant cases to arrive at a majority opinion. To reduce diagnostic uncertainty, all subspecialists had access to an immunohistochemical-stained section and 3 histologic sections for every biopsied specimen. Their review was conducted from December 2018 to June 2019. MAIN OUTCOMES AND MEASURES The frequency of the exact agreement of the DLS with the majority opinion of the subspecialists in categorizing each tumor-containing specimen as 1 of 5 categories: nontumor, GG1, GG2, GG3, or GG4-5. For comparison, the rate of agreement of 19 general pathologists' opinions with the subspecialists' majority opinions was also evaluated. RESULTS For grading tumor-containing biopsy specimens in the validation set (n = 498), the rate of agreement with subspecialists was significantly higher for the DLS (71.7%; 95% CI, 67.9%-75.3%) than for general pathologists (58.0%; 95% CI, 54.5%-61.4%) (P < .001). In subanalyses of biopsy specimens from an external validation set (n = 322), the Gleason grading performance of the DLS remained similar. For distinguishing nontumor from tumor-containing biopsy specimens (n = 752), the rate of agreement with subspecialists was 94.3% (95% CI, 92.4%-95.9%) for the DLS and similar at 94.7% (95% CI, 92.8%-96.3%) for general pathologists (P = .58). CONCLUSIONS AND RELEVANCE In this study, the DLS showed higher proficiency than general pathologists at Gleason grading prostate needle core biopsy specimens and generalized to an independent institution. Future research is necessary to evaluate the potential utility of using the DLS as a decision support tool in clinical workflows and to improve the quality of prostate cancer grading for therapy decisions.

Tuberous sclerosis complex (TSC) is an autosomal dominant disorder with characteristic tumors involving multiple organ systems. Whereas renal angiomyolipoma (AML) is common in TSC, renal cell carcinoma (RCC) is rarely reported. Fifty-seven RCCs from 13 female and 5 male TSC patients were reviewed. Age at surgery ranged from 7 to 65 years (mean: 42 y). Nine patients (50%) had multiple synchronous and/or metachronous RCCs (range of 2 to 20 RCCs) and 5 had bilateral RCCs (28%). Seventeen patients (94%) had histologically confirmed concurrent renal AMLs, including 15 with multiple AMLs (88%) and 9 (50%) with AMLs with epithelial cysts. None of the 15 patients with available clinical follow-up information had evidence of distant metastatic disease from 6 to 198 months after their initial surgery (mean: 52 mo). The 57 RCCs exhibited 3 major distinct morphologies: (1) 17 RCCs (30%) had features similar to tumors previously described as "renal angiomyoadenomatous tumor" or "RCC with smooth muscle stroma"; (2) 34 RCCs (59%) showed features similar to chromophobe RCC; and (3) 6 RCCs (11%) showed a granular eosinophilic-macrocystic morphology. Distinct histologic changes were also commonly present in the background kidney parenchyma and included cysts or renal tubules lined by epithelial cells with prominent eosinophilic cytoplasm, nucleomegaly, and nucleoli. Immunohistochemically, all RCCs tested showed strong nuclear reactivity for PAX8 and HMB45 negativity. Compared with sporadic RCCs, TSC-associated RCCs have unique clinicopathologic features including female predominance, younger age at diagnosis, multiplicity, association with AMLs, 3 recurring histologic patterns, and an indolent clinical course. Awareness of the morphologic and clinicopathologic spectrum of RCC in this setting will allow surgical pathologists to better recognize clinically unsuspected TSC patients.

The diagnosis of metastatic clear cell renal cell carcinoma (CC-RCC) can be difficult because of its morphologic heterogeneity and the increasing use of small image-guided biopsies that yield scant diagnostic material. This is further complicated by the degree of morphologic and immunophenotypic overlap with non-renal neoplasms and tissues, such as adrenal cortex. In this study, a detailed immunoprofile of 63 adrenal cortical lesions, which included 54 cortical neoplasms, was compared with 185 metastatic CC-RCC using traditional [anti-calretinin, CD10, anti-chromogranin, anti-EMA, anti-inhibin, anti-melanA, anti-cytokeratins (AE1/AE3 and AE1/CAM5.2), anti-renal cell carcinoma marker (RCCma), and anti-synaptophysin)] and novel [anti-carbonic anhydrase-IX (CAIX), anti-hepatocyte nuclear factor-1b (HNF-1b), anti-human kidney injury molecule-1 (hKIM-1), anti-PAX-2, anti-PAX-8, anti-steroidogenic factor-1 (SF-1), and anti-T cell immunoglobulin mucin-1 (TIM-1)] antibodies. Tissue microarray methodology was used to simulate small image-guided biopsies. Staining extent and intensity were scored semiquantitatively for each antibody. In comparing different intensity thresholds required for a ‘‘positive’’ result, ≥2+ was identified as optimal for diagnostic sensitivity/specificity. For the distinction of adrenal cortical lesions from metastatic CC-RCC, immunoreactivity for the adrenal cortical antigens SF-1 (86% adrenal; 0% CC-RCC), calretinin (89% adrenal; 10% CC-RCC), inhibin (86% adrenal; 9% CC-RCC), and melanA (86% adrenal; 10% CC-RCC) and the renal epithelial antigens hKIM-1 (0% adrenal; 83% CC-RCC), PAX-8 (0% adrenal; 83% CC-RCC), HNF-1b (0% adrenal; 76% CC-RCC), EMA (0% adrenal; 78% CC-RCC), and CAIX (3% adrenal; 87% CC-RCC) had the most potential utility. The use of the novel renal epithelial markers hKIM-1 (clone AKG7) and/or PAX-8, and the adrenocortical marker SF-1 in an immunohistochemical panel for distinguishing adrenal cortical lesions from metastatic CC-RCC offers improved diagnostic sensitivity and specificity.