Updates on grading mesothelioma.

Mesothelioma is a rare disease with an historically poor prognosis. Over the past decade, a grading system has been developed that is a powerful prognostic tool in epithelioid mesothelioma. Grading of epithelioid mesothelioma is now required or strongly recommended by expert consensus, the College of American Pathologists, the World Health Organization, and the International Mesothelioma Interest Group. The original nuclear grading system for epithelioid mesothelioma, developed in the United States, split epithelioid mesotheliomas into three prognostic groups with marked differences in survival. Now, this three-tiered nuclear grading system has been combined with the presence or absence of necrosis to form the currently recommended two-tiered grading system of low- and high-grade epithelioid mesothelioma. This review will focus on the development of this grading system in mesothelioma, the grading system's shortcomings, and the application of the grading system to cytology specimens and other extra-pleural sites. Lastly, this review will briefly discuss alternative grading systems and future considerations.

Long-term Lung Abnormalities Associated with COVID-19 Pneumonia.

In the 3rd year of the SARS-CoV-2 pandemic, much has been learned about the long-term effects of COVID-19 pneumonia on the lungs. Approximately one-third of patients with moderate-to-severe pneumonia, especially those requiring intensive care therapy or mechanical ventilation, have residual abnormalities at chest CT 1 year after presentation. Abnormalities range from parenchymal bands to bronchial dilation to frank fibrosis. Less is known about the long-term pulmonary vascular sequelae, but there appears to be a persistent, increased risk of venothromboembolic events in a small cohort of patients. Finally, the associated histologic abnormalities resulting from SARS-CoV-2 infection are similar to those seen in patients with other causes of acute lung injury.

[68 Ga]Ga-FAPI-46 PET for non-invasive detection of pulmonary fibrosis disease activity.

PURPOSE: The lack of effective molecular biomarkers to monitor idiopathic pulmonary fibrosis (IPF) activity or treatment response remains an unmet clinical need. Herein, we determined the utility of fibroblast activation protein inhibitor for positron emission tomography (FAPI PET) imaging in a mouse model of pulmonary fibrosis. METHODS: Pulmonary fibrosis was induced by intratracheal administration of bleomycin (1 U/kg) while intratracheal saline was administered to control mice. Subgroups from each cohort (n = 3-5) underwent dynamic 1 h PET/CT after intravenously injecting FAPI-46 radiolabeled with gallium-68 ([68 Ga]Ga-FAPI-46) at 7 days and 14 days following disease induction. Animals were sacrificed following imaging for ex vivo gamma counting and histologic correlation. [68 Ga]Ga-FAPI-46 uptake was quantified and reported as percent injected activity per cc (%IA/cc) or percent injected activity (%IA). Lung CT density in Hounsfield units (HU) was also correlated with histologic examinations of lung fibrosis. RESULTS: CT only detected differences in the fibrotic response at 14 days post-bleomycin administration. [68 Ga]Ga-FAPI-46 lung uptake was significantly higher in the bleomycin group than in control subjects at 7 days and 14 days. Significantly (P = 0.0012) increased [68 Ga]Ga-FAPI-46 lung uptake in the bleomycin groups at 14 days (1.01 ± 0.12%IA/cc) vs. 7 days (0.33 ± 0.09%IA/cc) at 60 min post-injection of the tracer was observed. These findings were consistent with an increase in both fibrinogenesis and FAP expression as seen in histology. CONCLUSION: CT was unable to assess disease activity in a murine model of IPF. Conversely, FAPI PET detected both the presence and activity of lung fibrogenesis, making it a promising tool for assessing early disease activity and evaluating the efficacy of therapeutic interventions in lung fibrosis patients.

Heterogeneity in PD-L1 expression in malignant peritoneal mesothelioma with systemic or intraperitoneal chemotherapy.

Programmed death-ligand 1 (PD-L1) expression has been described in patients with malignant peritoneal mesothelioma (MPM), but treatment strategies utilising immune checkpoint inhibition are yet to be defined. Here, we examine levels of PD-L1 expression in MPM patients treated with systemic and/or intraperitoneal chemotherapy using tissue from patient tumour biopsies or resections at multiple time points. We found the mean PD-L1 expression was higher in those with a germline mutation and/or those with a higher somatic mutation burden. Moreover, PD-L1 expression was lower in patients who had received prior chemotherapy as compared to the treatment-naive cohort. Twenty patients who received chemotherapy, either systemic and/or peritoneal, between PD-L1 measurements showed marked heterogeneity. Six (30%) patients demonstrated upregulation of PD-L1, while eight (40%) demonstrated downregulation. Heterogeneity in PD-L1 expression in MPM before and after cytotoxic therapies may present an additional consideration when initiating immune checkpoint inhibition in this rare and challenging disease.

Malignant peritoneal mesothelioma: prognostic significance of clinical and pathologic parameters and validation of a nuclear-grading system in a multi-institutional series of 225 cases.

Malignant peritoneal mesothelioma historically carried a grim prognosis, but outcomes have improved substantially in recent decades. The prognostic significance of clinical, morphologic, and immunophenotypic features remains ill-defined. This multi-institutional cohort comprises 225 malignant peritoneal mesotheliomas, which were assessed for 21 clinical, morphologic, and immunohistochemical parameters. For epithelioid mesotheliomas, combining nuclear pleomorphism and mitotic index yielded a composite nuclear grade, using a previously standardized grading system. Correlation of clinical, morphologic, and immunohistochemical parameters with overall and disease-free survival was examined by univariate and multivariate analyses. On univariate analysis, longer overall survival was significantly associated with diagnosis after 2000 (P = 0.0001), age <60 years (P = 0.0001), ECOG performance status 0 or 1 (P = 0.01), absence of radiographic lymph-node metastasis (P = 0.04), cytoreduction surgery (P < 0.0001), hyperthermic intraperitoneal chemotherapy (P = 0.0001), peritoneal carcinomatosis index <27 (P = 0.01), absence of necrosis (P = 0.007), and epithelioid histotype (P < 0.0001). Among epithelioid malignant mesotheliomas only, longer overall survival was further associated with female sex (P = 0.03), tubulopapillary architecture (P = 0.005), low nuclear pleomorphism (P < 0.0001), low mitotic index (P = 0.0007), and low composite nuclear grade (P < 0.0001). On multivariate analyses, the low composite nuclear grade was independently associated with longer overall and disease-free survival (P < 0.0001). Our data further clarify the interactions of clinical and pathologic features in peritoneal mesothelioma prognosis and validate the prognostic significance of a standardized nuclear-grading system in epithelioid malignant mesothelioma of the peritoneum.

MTAP immunohistochemistry is an accurate and reproducible surrogate for CDKN2A fluorescence in situ hybridization in diagnosis of malignant pleural mesothelioma.

Ancillary studies facilitate accurate diagnosis of morphologically challenging mesothelial proliferations. The current diagnostic algorithm proceeds from BAP1 immunohistochemistry to CDKN2A fluorescence in situ hybridization. While MTAP immunohistochemistry has recently shown promise as a surrogate for CDKN2A fluorescence in situ hybridization, it has been examined in only a few single-institution studies. Furthermore, there are no published reports on interobserver agreement or interlaboratory reproducibility for MTAP immunohistochemistry. We performed MTAP immunohistochemistry on 20 benign mesothelial lesions and 99 malignant mesotheliomas from five mesothelioma centers in four countries, and each MTAP stain was independently interpreted by four pathologists. CDKN2A fluorescence in situ hybridization data were available for a subset of cases, and a subset of cases was subjected in MTAP immunohistochemistry in multiple laboratories to assess interlaboratory reproducibility. Interobserver agreement in MTAP immunostain interpretation was excellent for all mesothelial lesions (kappa: 0.85) and for malignant mesothelioma cases only (kappa: 0.82). Interlaboratory reproducibility was also excellent (kappa values for paired protocols: 0.77-0.89). MTAP loss by immunohistochemistry was 78% sensitive and 96% specific for CDKN2A homozygous deletion. MTAP immunohistochemistry is a reliable surrogate for CDKN2A fluorescence in situ hybridization in diagnosis of malignant mesothelioma. Interobserver agreement is excellent for interpretation of MTAP staining, and protocols performed in different laboratories yield concordant MTAP staining results. Rare cases with immunohistochemical MTAP loss may retain normal CDKN2A copy number, and the MTAP staining results should be correlated with clinicopathologic findings and other ancillary studies.

Paraneoplastic opsoclonus myoclonus syndrome associated with inflammatory myofibroblastic tumor in a pediatric patient.

Opsoclonus myoclonus syndrome (OMS) is a rare neurological syndrome caused by a paraneoplastic autoimmune process that affects children with neuroblastic tumors. Treatment includes corticosteroids, intravenous gamma globulin (IVIG), rituximab, and other immunosuppressive therapies. Here, we describe a patient diagnosed with OMS associated with a localized inflammatory myofibroblastic tumor. The patient has no evidence of tumor recurrence following surgical resection with 8-month follow-up. The neurologic symptoms resolved with corticosteroids and IVIG. This case demonstrates that in children, neoplasms other than neuroblastoma may be associated with this paraneoplastic syndrome, and highlights the importance of evaluating patients with OMS for underlying malignancies.

An Unusual Lung Mass of Heterotopic Pancreatic Tissue in a Neonate With an Elevated Immunoreactive Trypsinogen on Newborn Screen.

We present a case of a neonate with tracheoesophageal fistula and esophageal atresia along with a suspicious lung mass who had a false-positive newborn screen for cystic fibrosis due to an elevated serum immunoreactive trypsinogen with an additionally elevated serum lipase. The infant's lung mass was found to contain heterotopic pancreatic tissue consisting of acini, ducts, and islet cells, without an associated gastrointestinal duplication cyst. This constellation of congenital abnormalities has not been described in previous literature. Also, this is the first reported case of a neonate with elevated serum pancreatic enzymes in which the underlying etiology was discovered to be heterotopic pancreas.

High-Grade Conventional Osteosarcoma of the Mandible Associated With P53 Germline Mutation.

High-grade conventional osteosarcoma of the mandible is a rare entity that can manifest as a rapidly growing mass. The authors report the case of a young adult patient with right mandible pain and swelling and a history of humeral osteosarcoma treated over 1 decade earlier. Computed tomography and magnetic resonance imaging showed a mass arising from the right mandible that progressed despite induction chemotherapy and the patient underwent surgery. Histopathology revealed the presence of malignant osteoblasts and osteoid, which led to the diagnosis of high-grade conventional osteosarcoma. There were also regional lymph node metastases. Genetic testing revealed a p53 germline mutation. The presence of mandibular osteosarcoma in a young patient should raise the suspicion of an underlying p53 germline mutation. Ultimately, the recurrent tumor regressed with Etoposide and Ifosfamide.

Polyvalvular Dysplasia and Vascular Abnormalities in a Neonate With an FLNA Variant.

There is growing appreciation for inherited structural heart diseases and their genetic causes. One causal gene for congenital cardiac and vascular lesions is FLNA which encodes a critical protein for cytoskeletal and extracellular matrix development. A newborn infant male, with prenatally diagnosed polyvalvular dysfunction, presented with low cardiac output and postnatally detected aortic arch hypoplasia and coarctation. Attempted palliative coarctation intervention resulted in vascular complications that ultimately contributed to his demise. This case report highlights polyvalvular dysplasia, vascular abnormalities, and a likely causal de novo missense variant in the FLNA gene (c.5180 C>T p.P1727L) not previously described.

Microwave dielectric properties of normal, fibroelastotic, and malignant human lung tissue.

Technological development of microwave treatment and detection techniques for lung cancer requires accurate and comprehensive knowledge of the microwave dielectric properties of human lung tissue. We characterize the dielectric properties of room temperature human lung tissue from 0.5 to 10 GHz for three lung tissue groups: normal, fibroelastotic, and malignant. We fit a two-pole Debye model to the measured frequency-dependent complex permittivity and calculate the median Debye parameters for the three groups. We find that malignant lung tissue is approximately 10% higher in relative permittivity and conductivity compared to normal lung tissue; this trend matches previously reported normal versus malignant data for other biological tissues. There is little contrast between benign lung tissue with fibroelastosis and malignant lung tissue. We extrapolate our data from room temperature to 37 °C using a temperature-dependence model for animal lung tissue and use the Maxwell-Garnett dielectric mixing model to predict the dielectric properties of inflation-dynamic human lung tissue; both approximations correspond with previously reported dielectric data of bovine and porcine lung tissue.

Developing a low-cost, open-source, locally manufactured workstation and computational pipeline for automated histopathology evaluation using deep learning.

BACKGROUND: Deployment and access to state-of-the-art precision medicine technologies remains a fundamental challenge in providing equitable global cancer care in low-resource settings. The expansion of digital pathology in recent years and its potential interface with diagnostic artificial intelligence algorithms provides an opportunity to democratize access to personalized medicine. Current digital pathology workstations, however, cost thousands to hundreds of thousands of dollars. As cancer incidence rises in many low- and middle-income countries, the validation and implementation of low-cost automated diagnostic tools will be crucial to helping healthcare providers manage the growing burden of cancer. METHODS: Here we describe a low-cost ($230) workstation for digital slide capture and computational analysis composed of open-source components. We analyze the predictive performance of deep learning models when they are used to evaluate pathology images captured using this open-source workstation versus images captured using common, significantly more expensive hardware. Validation studies assessed model performance on three distinct datasets and predictive models: head and neck squamous cell carcinoma (HPV positive versus HPV negative), lung cancer (adenocarcinoma versus squamous cell carcinoma), and breast cancer (invasive ductal carcinoma versus invasive lobular carcinoma). FINDINGS: When compared to traditional pathology image capture methods, low-cost digital slide capture and analysis with the open-source workstation, including the low-cost microscope device, was associated with model performance of comparable accuracy for breast, lung, and HNSCC classification. At the patient level of analysis, AUROC was 0.84 for HNSCC HPV status prediction, 1.0 for lung cancer subtype prediction, and 0.80 for breast cancer classification. INTERPRETATION: Our ability to maintain model performance despite decreased image quality and low-power computational hardware demonstrates that it is feasible to massively reduce costs associated with deploying deep learning models for digital pathology applications. Improving access to cutting-edge diagnostic tools may provide an avenue for reducing disparities in cancer care between high- and low-income regions. FUNDING: Funding for this project including personnel support was provided via grants from NIH/NCIR25-CA240134, NIH/NCIU01-CA243075, NIH/NIDCRR56-DE030958, NIH/NCIR01-CA276652, NIH/NCIK08-CA283261, NIH/NCI-SOAR25CA240134, SU2C (Stand Up to Cancer) Fanconi Anemia Research Fund - Farrah Fawcett Foundation Head and Neck Cancer Research Team Grant, and the European UnionHorizon Program (I3LUNG).

Guidelines for Pathologic Diagnosis of Mesothelioma.

CONTEXT.­: Mesothelioma is an uncommon tumor that can be difficult to diagnose. OBJECTIVE.­: To provide updated, practical guidelines for the pathologic diagnosis of mesothelioma. DATA SOURCES.­: Pathologists involved in the International Mesothelioma Interest Group and others with expertise in mesothelioma contributed to this update. Reference material includes peer-reviewed publications and textbooks. CONCLUSIONS.­: There was consensus opinion regarding guidelines for (1) histomorphologic diagnosis of mesothelial tumors, including distinction of epithelioid, biphasic, and sarcomatoid mesothelioma; recognition of morphologic variants and patterns; and recognition of common morphologic pitfalls; (2) molecular pathogenesis of mesothelioma; (3) application of immunohistochemical markers to establish mesothelial lineage and distinguish mesothelioma from common morphologic differentials; (4) application of ancillary studies to distinguish benign from malignant mesothelial proliferations, including BAP1 and MTAP immunostains; novel immunomarkers such as Merlin and p53; fluorescence in situ hybridization (FISH) for homozygous deletion of CDKN2A; and novel molecular assays; (5) practical recommendations for routine reporting of mesothelioma, including grading epithelioid mesothelioma and other prognostic parameters; (6) diagnosis of mesothelioma in situ; (7) cytologic diagnosis of mesothelioma, including use of immunostains and molecular assays; and (8) features of nonmalignant peritoneal mesothelial lesions.

The 1000 Mitoses Project: A Consensus-Based International Collaborative Study on Mitotic Figures Classification.

Introduction. The identification of mitotic figures is essential for the diagnosis, grading, and classification of various different tumors. Despite its importance, there is a paucity of literature reporting the consistency in interpreting mitotic figures among pathologists. This study leverages publicly accessible datasets and social media to recruit an international group of pathologists to score an image database of more than 1000 mitotic figures collectively. Materials and Methods. Pathologists were instructed to randomly select a digital slide from The Cancer Genome Atlas (TCGA) datasets and annotate 10-20 mitotic figures within a 2 mm2 area. The first 1010 submitted mitotic figures were used to create an image dataset, with each figure transformed into an individual tile at 40x magnification. The dataset was redistributed to all pathologists to review and determine whether each tile constituted a mitotic figure. Results. Overall pathologists had a median agreement rate of 80.2% (range 42.0%-95.7%). Individual mitotic figure tiles had a median agreement rate of 87.1% and a fair inter-rater agreement across all tiles (kappa = 0.284). Mitotic figures in prometaphase had lower percentage agreement rates compared to other phases of mitosis. Conclusion. This dataset stands as the largest international consensus study for mitotic figures to date and can be utilized as a training set for future studies. The agreement range reflects a spectrum of criteria that pathologists use to decide what constitutes a mitotic figure, which may have potential implications in tumor diagnostics and clinical management.

Characterizing the distribution of alterations in mesothelioma and their correlation to morphology.

OBJECTIVES: Mesothelioma is a lethal disease that arises from the serosal lining of organ cavities. Several recurrent alterations have been observed in pleural and peritoneal -mesotheliomas, including in BAP1, NF2, and CDKN2A. Although specific histopathologic parameters have been correlated with prognosis, it is not as well known whether genetic alterations correlate with histologic findings. METHODS: We reviewed 131 mesotheliomas that had undergone next-generation sequencing (NGS) at our institutions after pathologic diagnosis. There were 109 epithelioid mesotheliomas, 18 biphasic mesotheliomas, and 4 sarcomatoid mesotheliomas. All our biphasic and sarcomatoid cases arose in the pleura. Of the epithelioid mesotheliomas, 73 were from the pleura and 36 were from the peritoneum. On average, patients were 66 years of age (range, 26-90 years) and predominantly male (92 men, 39 women). RESULTS: The most common alterations identified were in BAP1, CDKN2A, NF2, and TP53. Twelve mesotheliomas did not show a pathogenic alteration on NGS. For epithelioid mesotheliomas in the pleura, the presence of an alteration in BAP1 correlated with low nuclear grade (P = .04), but no correlation was found in the peritoneum (P = .62). Similarly, there was no correlation between the amount of solid architecture in epithelioid mesotheliomas and any alterations in the pleura (P = .55) or peritoneum (P = .13). For biphasic mesotheliomas, cases with either no alteration detected or with an alteration in BAP1 were more likely to be epithelioid predominant (>50% of the tumor, P = .0001), and biphasic mesotheliomas with other alterations detected and no alteration in BAP1 were more likely to be sarcomatoid predominant (>50% of the tumor, P = .0001). CONCLUSIONS: This study demonstrates a significant association between morphologic features associated with a better prognosis and an alteration in BAP1.

Central lung gene expression associates with myofibroblast features in idiopathic pulmonary fibrosis.

RATIONALE: Contribution of central lung tissues to pathogenesis of idiopathic pulmonary fibrosis (IPF) remains unknown. OBJECTIVE: To ascertain the relationship between cell types of IPF-central and IPF-peripheral lung explants using RNA sequencing (RNA-seq) transcriptome. METHODS: Biopsies of paired IPF-central and IPF-peripheral along with non-IPF lungs were selected by reviewing H&E data. Criteria for differentially expressed genes (DEG) were set at false discovery rate <5% and fold change >2. Computational cell composition deconvolution was performed. Signature scores were computed for each cell type. FINDINGS: Comparison of central IPF versus non-IPF identified 1723 DEG (1522 upregulated and 201 downregulated). Sixty-two per cent (938/1522) of the mutually upregulated genes in central IPF genes were also upregulated in peripheral IPF versus non-IPF. Moreover, 85 IPF central-associated genes (CAG) were upregulated in central IPF versus both peripheral IPF and central non-IPF. IPF single-cell RNA-seq analysis revealed the highest CAG signature score in myofibroblasts and significantly correlated with a previously published activated fibroblasts signature (r=0.88, p=1.6×10-4). CAG signature scores were significantly higher in IPF than in non-IPF myofibroblasts (p=0.013). Network analysis of central-IPF genes identified a module significantly correlated with the deconvoluted proportion of myofibroblasts in central IPF and anti-correlated with inflammation foci trait in peripheral IPF. The module genes were over-represented in idiopathic pulmonary fibrosis signalling pathways. INTERPRETATION: Gene expression in central IPF lung regions demonstrates active myofibroblast features that contributes to disease progression. Further elucidation of pathological transcriptomic state of cells in the central regions of the IPF lung that are relatively spared from morphological rearrangements may provide insights into molecular changes in the IPF progression.

Endometrial Carcinomas With Subclonal Loss of Mismatch Repair Proteins: A Clinicopathologic and Genomic Study.

Subclonal loss of mismatch repair (MMR) proteins has been described in a small subset of endometrial carcinomas (ECs), but the genomic basis for this phenomenon has received limited attention. Herein, we retrospectively evaluated all ECs with MMR immunohistochemistry (n=285) for subclonal loss, and in those (n=6), performed a detailed clinicopathologic and genomic comparison of the MMR-deficient and MMR-proficient components. Three tumors were FIGO stage IA, and one each stage IB, II, and IIIC2. Patterns of subclonal loss were as follows: (1) 3 FIGO grade 1 endometrioid carcinomas with subclonal MLH1/PMS2, MLH1 promoter hypermethylation, and no MMR gene mutations; (2) POLE -mutated FIGO grade 3 endometrioid carcinoma with subclonal PMS2, and PMS2 and MSH6 mutations limited to the MMR-deficient component; (3) dedifferentiated carcinoma with subclonal MSH2/MSH6, as well as complete loss of MLH1/PMS2, MLH1 promoter hypermethylation, and PMS2 and MSH6 mutations in both components; (4) dedifferentiated carcinoma with subclonal MSH6, and somatic and germline MSH6 mutations in both components, but with a higher allele frequency in MMR-deficient foci. Recurrences occurred in 2 patients, one consisted of the MMR-proficient component from a FIGO 1 endometrioid carcinoma, while the other was from the MSH6 -mutated dedifferentiated endometrioid carcinoma. At the last follow-up (median: 44 mo), 4 patients were alive and disease-free and 2 were alive with disease. In summary, subclonal MMR loss reflects subclonal and often complex genomic and epigenetic alterations, which may have therapeutic implications and therefore must be reported when present. In addition, subclonal loss can occur in both POLE -mutated and Lynch syndrome-associated ECs.

Deep learning generates synthetic cancer histology for explainability and education.

Artificial intelligence methods including deep neural networks (DNN) can provide rapid molecular classification of tumors from routine histology with accuracy that matches or exceeds human pathologists. Discerning how neural networks make their predictions remains a significant challenge, but explainability tools help provide insights into what models have learned when corresponding histologic features are poorly defined. Here, we present a method for improving explainability of DNN models using synthetic histology generated by a conditional generative adversarial network (cGAN). We show that cGANs generate high-quality synthetic histology images that can be leveraged for explaining DNN models trained to classify molecularly-subtyped tumors, exposing histologic features associated with molecular state. Fine-tuning synthetic histology through class and layer blending illustrates nuanced morphologic differences between tumor subtypes. Finally, we demonstrate the use of synthetic histology for augmenting pathologist-in-training education, showing that these intuitive visualizations can reinforce and improve understanding of histologic manifestations of tumor biology.

Successful Lung Transplantation for Severe Post-COVID-19 Pulmonary Fibrosis.

Lung transplantation has been well described for patients with coronavirus disease 2019 (COVID-19) in the acute setting, but less so for the resulting pulmonary sequelae. This report describes a case of lung transplantation for post-COVID-19 pulmonary fibrosis. A 52-year-old woman contracted COVID-19 in July 2020 and mounted a partial recovery, but she went on to have declining function over the ensuing 3 months, with development of fibrocystic lung changes. She underwent bilateral lung transplantation and recovered rapidly, was discharged home on postoperative day 14, and has done well in follow-up. This case report demonstrates that lung transplantation is an acceptable therapy for post-COVID-19 pulmonary fibrosis.

Uncertainty-informed deep learning models enable high-confidence predictions for digital histopathology.

A model's ability to express its own predictive uncertainty is an essential attribute for maintaining clinical user confidence as computational biomarkers are deployed into real-world medical settings. In the domain of cancer digital histopathology, we describe a clinically-oriented approach to uncertainty quantification for whole-slide images, estimating uncertainty using dropout and calculating thresholds on training data to establish cutoffs for low- and high-confidence predictions. We train models to identify lung adenocarcinoma vs. squamous cell carcinoma and show that high-confidence predictions outperform predictions without uncertainty, in both cross-validation and testing on two large external datasets spanning multiple institutions. Our testing strategy closely approximates real-world application, with predictions generated on unsupervised, unannotated slides using predetermined thresholds. Furthermore, we show that uncertainty thresholding remains reliable in the setting of domain shift, with accurate high-confidence predictions of adenocarcinoma vs. squamous cell carcinoma for out-of-distribution, non-lung cancer cohorts.