A pathologist-AI collaboration framework for enhancing diagnostic accuracies and efficiencies.

In pathology, the deployment of artificial intelligence (AI) in clinical settings is constrained by limitations in data collection and in model transparency and interpretability. Here we describe a digital pathology framework, nuclei.io, that incorporates active learning and human-in-the-loop real-time feedback for the rapid creation of diverse datasets and models. We validate the effectiveness of the framework via two crossover user studies that leveraged collaboration between the AI and the pathologist, including the identification of plasma cells in endometrial biopsies and the detection of colorectal cancer metastasis in lymph nodes. In both studies, nuclei.io yielded considerable diagnostic performance improvements. Collaboration between clinicians and AI will aid digital pathology by enhancing accuracies and efficiencies.

Detection of FOXL2 C134W Mutation Status by a Novel BaseScope In Situ Hybridization Assay is Highly Sensitive and Specific for Adult Granulosa Cell Tumors.

Adult granulosa cell tumors (AGCTs) are a molecularly distinct group of malignant ovarian sex cord-stromal tumors (SCSTs) characterized by a nearly ubiquitous c.402C>G/p.C134W mutation in FOXL2 (hereafter referred to as "C134W"). In some cases, AGCT exhibits marked morphologic overlap with other SCSTs and has an identical immunophenotype, and molecular testing may be necessary to help confirm the diagnosis. However, molecular testing is time consuming, relatively expensive, and unavailable in many pathology laboratories. We describe the development and validation of an in situ hybridization (ISH) custom BaseScope assay for the detection of the FOXL2 C134W mutation. We evaluated 106 ovarian SCSTs, including 78 AGCTs, 9 juvenile granulosa cell tumors, 18 fibromas (cellular and conventional), and 1 SCST, not otherwise specified, as well as 53 epithelial ovarian tumors (42 endometrioid carcinomas and 11 carcinosarcomas) and 1 STK11 adnexal tumor for the presence or absence of FOXL2 wild-type and FOXL2 C134W RNA expression via BaseScope-ISH. Fifty-one tumors had previously undergone DNA sequencing of the FOXL2 gene. Across the entire cohort, the FOXL2 C134W probe staining was positive in 77 of 78 (98.7%) AGCTs. Two of 81 (2.5%) non-AGCTs also showed positive staining, both of which were epithelial ovarian tumors. The assay worked in tissue from blocks >20 years old. There was 100% concordance between the FOXL2 sequencing and BaseScope-ISH results. Overall, assessment of FOXL2 mutation status by custom BaseScope-ISH demonstrated 98.7% sensitivity and 97.5% specificity for the diagnosis of AGCT. BaseScope-ISH for FOXL2 C134W represents a reasonable alternative to sequencing, is quicker and less expensive, and is more easily incorporated than molecular testing into many pathology laboratories. It also has the advantage of requiring less tissue, and the neoplastic cells can be directly visualized on stained sections.

Does Milan affect management? A retrospective analysis of resection rate and time to surgery among Milan categories.

INTRODUCTION: The Milan System for Reporting Salivary Gland Cytopathology (MSRSGC) is an established system with reproducible risk of malignancies (ROM) for salivary gland fine needle aspiration (SGFNA). No studies have reviewed the relationship between Milan categories and the resection rate (RR) and time to resection (TTR). METHODS: We searched our database (January 1, 2011 to January 4, 2021) for non-lymphoma SGFNAs and assigned appropriate MSRSGC categories. RR and TTR were calculated and compared for each category. A literature search was performed; RRs and TTRs were compared. RESULTS: Seven hundred and eighty SGFNAs were identified, 333 with follow-up. RR was highest in suspicious for malignancy (SUS, V; 70.6%, n = 12/17), followed by the salivary gland neoplasm of uncertain malignant potential (SUMP, IVb; 69.6%, n = 80/115) and malignant (M, VI; 55.6%, n = 75/135). Among M, primary tumors had a higher RR (65.1%, n = 41/63) than metastases (47.2%, n = 34/72, p = .36). In literature review, SUS had the highest RR (69.3%, n = 233/336) followed by M (61.6%, n = 821/1332) and SUMP (60.2%, n = 632/1050). TTR was shorter in SUS (mean = 32.3 days, median = 25 days). Within the benign neoplasms (BN, IVa), Pleomorphic adenomas (PAs) had a higher RR than Warthin tumors (WTs) (66.3% vs. 37.2%, p < .00001), and a shorter TTR (median = 63 days vs. 90 days). CONCLUSIONS: Tumors classified as SUS had higher RR and at shorter intervals than those classified as SUMP. PAs have higher RRs and more expedient surgery than WTs. Cases classified as M are less likely to undergo follow-up than SUS, perhaps due to a lower RR for metastases.

Nondiagnostic salivary gland FNAs are associated with decreased risk of malignancy compared with "all-comer" patients: Analysis of the Milan System for Reporting Salivary Gland Cytopathology with a focus on Milan I: Nondiagnostic.

BACKGROUND: The Milan System for Reporting Salivary Gland Cytopathology (MSRSGC) reports a 25% rate of malignancy (ROM) for the Milan I: Nondiagnostic (ND) category. We clarify the ROM of ND salivary gland fine-needle aspirations (SGFNAs) based on our institutional experience and review of the literature. METHODS: Overall risk of malignancy (OROM) and that for those with surgical/flow cytometric follow-up (FROM) for each category and "all-comers" were calculated for Emory SGFNAs from January 2010 through March 2021. From a literature review of 50 articles using MSRSGC, distribution of diagnoses, rates of follow-up, FROM, and OROM by category were calculated. FROMs and OROMs between ND FNAs and all-comers were compared. Milan I rate was compared with the ratio of Milan I OROM to all-comer OROM. RESULTS: Of 819 SGFNAs at Emory, 12.8% (n = 105/819) were ND. Thirty-two had known follow-up, with 12 (37.5%) being malignant. Nonmucinous cyst contents accounted for 26.7% of ND SGFNAs (n = 28/105); all 7 with surgical follow-up were benign. Of 50 MSRSGC studies, 18.2% (n = 2384/13,129) of SGFNAs were classified as ND, 26.6% (n = 635/2384) with known follow-up. Total FROM and OROM for ND FNAs (15.7% and 4.1%, respectively) were significantly lower than those for all-comers (24.9% and 11.4%, respectively) (p < .001). There was no relationship between rate of ND SGFNA and ND ROM. CONCLUSIONS: The ND category is associated with a lower ROM than that of all-comer SGFNA patients. The "true" ROM for ND SGFNAs is likely best estimated by the 4.1% OROM. SGFNAs showing nonmucinous cyst contents have a particularly low ROM. Rate of ND SGFNAs does not influence ND ROM.