Prognostic impact of cooccurring mutations in FLT3-ITD pediatric acute myeloid leukemia.

ABSTRACT: We sought to define the cooccurring mutational profile of FLT3-ITD-positive (ITDpos) acute myeloid leukemia (AML) in pediatric and young adult patients and to define the prognostic impact of cooperating mutations. We identified 464 patients with FLT3-ITD mutations treated on Children's Oncology Group trials with available sequencing and outcome data. Overall survival, event-free survival (EFS), and relapse risk were determined according to the presence of cooccurring risk stratifying mutations. Among the cohort, 79% of patients had cooccurring alterations across 239 different genes that were altered through mutations or fusions. Evaluation of the prognostic impact of the cooccurring mutations demonstrated that patients with ITDpos AML experienced significantly different outcomes according to the cooccurring mutational profile. Patients with ITDpos AML harboring a cooccurring favorable-risk mutation of NPM1, CEBPA, t(8;21), or inv(16) experienced a 5-year EFS of 64%, which was significantly superior to of 22.2% for patients with ITDpos AML and poor-risk mutations of WT1, UBTF, or NUP98::NSD1 as well to 40.9% for those who lacked either favorable-risk or poor-risk mutation (ITDpos intermediate; P < .001 for both). Multivariable analysis demonstrated that cooccurring mutations had significant prognostic impact, whereas allelic ratio had no impact. Therapy intensification, specifically consolidation transplant in remission, resulted in significant improvements in survival for ITDpos AML. However, patients with ITDpos/NUP98::NSD1 continued to have poor outcomes with intensified therapy, including sorafenib. Cooccurring mutational profile in ITDpos AML has significant prognostic impacts and is critical to determining risk stratification and therapeutic allocation. These clinical trials were registered at www.clinicaltrials.gov as NCT00002798, NCT00070174, NCT00372593, and NCT01371981.

Chromatin Profiling of CBFA2T3-GLIS2 AMLs Identifies Key Transcription Factor Dependencies and BRG1 Inhibition as a Novel Therapeutic Strategy.

Oncogenic fusions involving transcription factors are present in the majority of pediatric leukemias; however, the context-specific mechanisms they employ to drive cancer remain poorly understood. CBFA2T3-GLIS2 (C/G) fusions occur in treatment-refractory acute myeloid leukemias and are restricted to young children. To understand how the C/G fusion drives oncogenesis we applied CUT&RUN chromatin profiling to an umbilical cord blood/endothelial cell (EC) co-culture model of C/G AML that recapitulates the biology of this malignancy. We find C/G fusion binding is mediated by its zinc finger domains. Integration of fusion binding sites in C/G- transduced cells with Polycomb Repressive Complex 2 (PRC2) sites in control cord blood cells identifies MYCN, ZFPM1, ZBTB16 and LMO2 as direct C/G targets. Transcriptomic analysis of a large pediatric AML cohort shows that these genes are upregulated in C/G patient samples. Single cell RNA-sequencing of umbilical cord blood identifies a population of megakaryocyte precursors that already express many of these genes despite lacking the fusion. By integrating CUT&RUN data with CRISPR dependency screens we identify BRG1/SMARCA4 as a vulnerability in C/G AML. BRG1 profiling in C/G patient-derived cell lines shows that the CBFA2T3 locus is a binding site, and treatment with clinically-available BRG1 inhibitors reduces fusion levels and downstream C/G targets including N-MYC, resulting in C/G leukemia cell death and extending survival in a murine xenograft model.

Etiology of oncogenic fusions in 5,190 childhood cancers and its clinical and therapeutic implication.

Oncogenic fusions formed through chromosomal rearrangements are hallmarks of childhood cancer that define cancer subtype, predict outcome, persist through treatment, and can be ideal therapeutic targets. However, mechanistic understanding of the etiology of oncogenic fusions remains elusive. Here we report a comprehensive detection of 272 oncogenic fusion gene pairs by using tumor transcriptome sequencing data from 5190 childhood cancer patients. We identify diverse factors, including translation frame, protein domain, splicing, and gene length, that shape the formation of oncogenic fusions. Our mathematical modeling reveals a strong link between differential selection pressure and clinical outcome in CBFB-MYH11. We discover 4 oncogenic fusions, including RUNX1-RUNX1T1, TCF3-PBX1, CBFA2T3-GLIS2, and KMT2A-AFDN, with promoter-hijacking-like features that may offer alternative strategies for therapeutic targeting. We uncover extensive alternative splicing in oncogenic fusions including KMT2A-MLLT3, KMT2A-MLLT10, C11orf95-RELA, NUP98-NSD1, KMT2A-AFDN and ETV6-RUNX1. We discover neo splice sites in 18 oncogenic fusion gene pairs and demonstrate that such splice sites confer therapeutic vulnerability for etiology-based genome editing. Our study reveals general principles on the etiology of oncogenic fusions in childhood cancer and suggests profound clinical implications including etiology-based risk stratification and genome-editing-based therapeutics.

Verification of prognostic expression biomarkers is improved by examining enriched leukemic blasts rather than mononuclear cells from acute myeloid leukemia patients.

BACKGROUND: Studies have not systematically compared the ability to verify performance of prognostic transcripts in paired bulk mononuclear cells versus viable CD34-expressing leukemic blasts from patients with acute myeloid leukemia. We hypothesized that examining the homogenous leukemic blasts will yield different biological information and may improve prognostic performance of expression biomarkers. METHODS: To assess the impact of cellular heterogeneity on expression biomarkers in acute myeloid leukemia, we systematically examined paired mononuclear cells and viable CD34-expressing leukemic blasts from SWOG diagnostic specimens. After enrichment, patients were assigned into discovery and validation cohorts based on availability of extracted RNA. Analyses of RNA sequencing data examined how enrichment impacted differentially expressed genes associated with pre-analytic variables, patient characteristics, and clinical outcomes. RESULTS: Blast enrichment yielded significantly different expression profiles and biological pathways associated with clinical characteristics (e.g., cytogenetics). Although numerous differentially expressed genes were associated with clinical outcomes, most lost their prognostic significance in the mononuclear cells and blasts after adjusting for age and ELN risk, with only 11 genes remaining significant for overall survival in both cell populations (CEP70, COMMD7, DNMT3B, ECE1, LNX2, NEGR1, PIK3C2B, SEMA4D, SMAD2, TAF8, ZNF444). To examine the impact of enrichment on biomarker verification, these 11 candidate biomarkers were examined by quantitative RT/PCR in the validation cohort. After adjusting for ELN risk and age, expression of 4 genes (CEP70, DNMT3B, ECE1, and PIK3CB) remained significantly associated with overall survival in the blasts, while none met statistical significance in mononuclear cells. CONCLUSIONS: This study provides insights into biological information gained/lost by examining viable CD34-expressing leukemic blasts versus mononuclear cells from the same patient and shows an improved verification rate for expression biomarkers in blasts.

Long Noncoding RNA Expression Independently Predicts Outcome in Pediatric Acute Myeloid Leukemia.

PURPOSE: Optimized strategies for risk classification are essential to tailor therapy for patients with biologically distinctive disease. Risk classification in pediatric acute myeloid leukemia (pAML) relies on detection of translocations and gene mutations. Long noncoding RNA (lncRNA) transcripts have been shown to associate with and mediate malignant phenotypes in acute myeloid leukemia (AML) but have not been comprehensively evaluated in pAML. METHODS: To identify lncRNA transcripts associated with outcomes, we evaluated the annotated lncRNA landscape by transcript sequencing of 1,298 pediatric and 96 adult AML specimens. Upregulated lncRNAs identified in the pAML training set were used to establish a regularized Cox regression model of event-free survival (EFS), yielding a 37 lncRNA signature (lncScore). Discretized lncScores were correlated with initial and postinduction treatment outcomes using Cox proportional hazards models in validation sets. Predictive model performance was compared with standard stratification methods by concordance analysis. RESULTS: Training set cases with positive lncScores had 5-year EFS and overall survival rates of 26.7% and 42.7%, respectively, compared with 56.9% and 76.3% with negative lncScores (hazard ratio, 2.48 and 3.16; P < .001). Pediatric validation cohorts and an adult AML group yielded comparable results in magnitude and significance. lncScore remained independently prognostic in multivariable models, including key factors used in preinduction and postinduction risk stratification. Subgroup analysis suggested that lncScores provide additional outcome information in heterogeneous subgroups currently classified as indeterminate risk. Concordance analysis showed that lncScore adds to overall classification accuracy with at least comparable predictive performance to current stratification methods that rely on multiple assays. CONCLUSION: Inclusion of the lncScore enhances predictive power of traditional cytogenetic and mutation-defined stratification in pAML with potential, as a single assay, to replace these complex stratification schemes with comparable predictive accuracy.

Comprehensive molecular and clinical characterization of NUP98 fusions in pediatric acute myeloid leukemia.

NUP98 fusions comprise a family of rare recurrent alterations in AML, associated with adverse outcomes. In order to define the underlying biology and clinical implications of this family of fusions, we performed comprehensive transcriptome, epigenome, and immunophenotypic profiling of 2,235 children and young adults with AML and identified 160 NUP98 rearrangements (7.2%), including 108 NUP98-NSD1 (4.8%), 32 NUP98-KDM5A (1.4%) and 20 NUP98-X cases (0.9%) with 13 different fusion partners. Fusion partners defined disease characteristics and biology; patients with NUP98-NSD1 or NUP98-KDM5A had distinct immunophenotypic, transcriptomic, and epigenomic profiles. Unlike the two most prevalent NUP98 fusions, NUP98-X variants are typically not cryptic. Furthermore, NUP98-X cases are associated with WT1 mutations, and have epigenomic profiles that resemble either NUP98-NSD1 or NUP98-KDM5A. Cooperating FLT3-ITD and WT1 mutations define NUP98-NSD1, and chromosome 13 aberrations are highly enriched in NUP98-KDM5A. Importantly, we demonstrate that NUP98 fusions portend dismal overall survival, with the noteworthy exception of patients bearing abnormal chromosome 13 (clinicaltrials gov. Identifiers: NCT00002798, NCT00070174, NCT00372593, NCT01371981).

Global gene expression of human malaria parasite liver stages throughout intrahepatocytic development.

Plasmodium falciparum (Pf) is causing the greatest malaria burden, yet the liver stages (LS) of this most important parasite species have remained poorly studied. Here, we used a human liver-chimeric mouse model in combination with a novel fluorescent PfNF54 parasite line (PfNF54cspGFP) to isolate PfLS-infected hepatocytes and generate transcriptomes that cover the major LS developmental phases in human hepatocytes. RNA-seq analysis of early Pf LS trophozoites two days after infection, revealed a central role of translational regulation in the transformation of the extracellular invasive sporozoite into intracellular LS. The developmental time course gene expression analysis indicated that fatty acid biosynthesis, isoprenoid biosynthesis and iron metabolism are sustaining LS development along with amino acid metabolism and biosynthesis. Countering oxidative stress appears to play an important role during intrahepatic LS development. Furthermore, we observed expression of the variant PfEMP1 antigen-encoding var genes, and we confirmed expression of PfEMP1 protein during LS development. Transcriptome comparison of the late Pf liver stage schizonts with P. vivax (Pv) late liver stages revealed highly conserved gene expression profiles among orthologous genes. A notable difference however was the expression of genes regulating sexual stage commitment. While Pv schizonts expressed markers of sexual commitment, the Pf LS parasites were not sexually committed and showed expression of gametocytogenesis repression factors. Our results provide the first comprehensive gene expression profile of the human malaria parasite Pf LS isolated during in vivo intrahepatocytic development. This data will inform biological studies and the search for effective intervention strategies that can prevent infection.

Inflammatory bone marrow signaling in pediatric acute myeloid leukemia distinguishes patients with poor outcomes.

High levels of the inflammatory cytokine IL-6 in the bone marrow are associated with poor outcomes in pediatric acute myeloid leukemia (pAML), but its etiology remains unknown. Using RNA-seq data from pre-treatment bone marrows of 1489 children with pAML, we show that > 20% of patients have concurrent IL-6, IL-1, IFNα/ß, and TNFα signaling activity and poorer outcomes. Targeted sequencing of pre-treatment bone marrow samples from affected patients (n = 181) revealed 5 highly recurrent patterns of somatic mutation. Using differential expression analyses of the most common genomic subtypes (~60% of total), we identify high expression of multiple potential drivers of inflammation-related treatment resistance. Regardless of genomic subtype, we show that JAK1/2 inhibition reduces receptor-mediated inflammatory signaling by leukemic cells in-vitro. The large number of high-risk pAML genomic subtypes presents an obstacle to the development of mutation-specific therapies. Our findings suggest that therapies targeting inflammatory signaling may be effective across multiple genomic subtypes of pAML.

CBFA2T3-GLIS2 model of pediatric acute megakaryoblastic leukemia identifies FOLR1 as a CAR T cell target.

The CBFA2T3-GLIS2 (C/G) fusion is a product of a cryptic translocation primarily seen in infants and early childhood and is associated with dismal outcome. Here, we demonstrate that the expression of the C/G oncogenic fusion protein promotes the transformation of human cord blood hematopoietic stem and progenitor cells (CB HSPCs) in an endothelial cell coculture system that recapitulates the transcriptome, morphology, and immunophenotype of C/G acute myeloid leukemia (AML) and induces highly aggressive leukemia in xenograft models. Interrogating the transcriptome of C/G-CB cells and primary C/G AML identified a library of C/G-fusion-specific genes that are potential targets for therapy. We developed chimeric antigen receptor (CAR) T cells directed against one of the targets, folate receptor α (FOLR1), and demonstrated their preclinical efficacy against C/G AML using in vitro and xenograft models. FOLR1 is also expressed in renal and pulmonary epithelium, raising concerns for toxicity that must be addressed for the clinical application of this therapy. Our findings underscore the role of the endothelial niche in promoting leukemic transformation of C/G-transduced CB HSPCs. Furthermore, this work has broad implications for studies of leukemogenesis applicable to a variety of oncogenic fusion-driven pediatric leukemias, providing a robust and tractable model system to characterize the molecular mechanisms of leukemogenesis and identify biomarkers for disease diagnosis and targets for therapy.

Integrated stem cell signature and cytomolecular risk determination in pediatric acute myeloid leukemia.

Relapsed or refractory pediatric acute myeloid leukemia (AML) is associated with poor outcomes and relapse risk prediction approaches have not changed significantly in decades. To build a robust transcriptional risk prediction model for pediatric AML, we perform RNA-sequencing on 1503 primary diagnostic samples. While a 17 gene leukemia stem cell signature (LSC17) is predictive in our aggregated pediatric study population, LSC17 is no longer predictive within established cytogenetic and molecular (cytomolecular) risk groups. Therefore, we identify distinct LSC signatures on the basis of AML cytomolecular subtypes (LSC47) that were more predictive than LSC17. Based on these findings, we build a robust relapse prediction model within a training cohort and then validate it within independent cohorts. Here, we show that LSC47 increases the predictive power of conventional risk stratification and that applying biomarkers in a manner that is informed by cytomolecular profiling outperforms a uniform biomarker approach.

Integrated Genomic Analysis Identifies UBTF Tandem Duplications as a Recurrent Lesion in Pediatric Acute Myeloid Leukemia.

The genetics of relapsed pediatric acute myeloid leukemia (AML) has yet to be comprehensively defined. Here, we present the spectrum of genomic alterations in 136 relapsed pediatric AMLs. We identified recurrent exon 13 tandem duplications (TD) in upstream binding transcription factor (UBTF) in 9% of relapsed AML cases. UBTF-TD AMLs commonly have normal karyotype or trisomy 8 with cooccurring WT1 mutations or FLT3-ITD but not other known oncogenic fusions. These UBTF-TD events are stable during disease progression and are present in the founding clone. In addition, we observed that UBTF-TD AMLs account for approximately 4% of all de novo pediatric AMLs, are less common in adults, and are associated with poor outcomes and MRD positivity. Expression of UBTF-TD in primary hematopoietic cells is sufficient to enhance serial clonogenic activity and to drive a similar transcriptional program to UBTF-TD AMLs. Collectively, these clinical, genomic, and functional data establish UBTF-TD as a new recurrent mutation in AML. SIGNIFICANCE: We defined the spectrum of mutations in relapsed pediatric AML and identified UBTF-TDs as a new recurrent genetic alteration. These duplications are more common in children and define a group of AMLs with intermediate-risk cytogenetic abnormalities, FLT3-ITD and WT1 alterations, and are associated with poor outcomes. See related commentary by Hasserjian and Nardi, p. 173. This article is highlighted in the In This Issue feature, p. 171.

A B-cell developmental gene regulatory network is activated in infant AML.

Infant Acute Myeloid Leukemia (AML) is a poorly-addressed, heterogeneous malignancy distinguished by surprisingly few mutations per patient but accompanied by myriad age-specific translocations. These characteristics make treatment of infant AML challenging. While infant AML is a relatively rare disease, it has enormous impact on families, and in terms of life-years-lost and life limiting morbidities. To better understand the mechanisms that drive infant AML, we performed integrative analyses of genome-wide mRNA, miRNA, and DNA-methylation data in diagnosis-stage patient samples. Here, we report the activation of an onco-fetal B-cell developmental gene regulatory network in infant AML. AML in infants is genomically distinct from AML in older children/adults in that it has more structural genomic aberrations and fewer mutations. Differential expression analysis of ~1500 pediatric AML samples revealed a large number of infant-specific genes, many of which are associated with B cell development and function. 18 of these genes form a well-studied B-cell gene regulatory network that includes the epigenetic regulators BRD4 and POU2AF1, and their onco-fetal targets LIN28B and IGF2BP3. All four genes are hypo-methylated in infant AML. Moreover, micro-RNA Let7a-2 is expressed in a mutually exclusive manner with its target and regulator LIN28B. These findings suggest infant AML may respond to bromodomain inhibitors and immune therapies targeting CD19, CD20, CD22, and CD79A.

Development of siRNA-Loaded Lipid Nanoparticles Targeting Long Non-Coding RNA LINC01257 as a Novel and Safe Therapeutic Approach for t(8;21) Pediatric Acute Myeloid Leukemia.

Standard of care therapies for children with acute myeloid leukemia (AML) cause potent off-target toxicity to healthy cells, highlighting the need to develop new therapeutic approaches that are safe and specific for leukemia cells. Long non-coding RNAs (lncRNAs) are an emerging and highly attractive therapeutic target in the treatment of cancer due to their oncogenic functions and selective expression in cancer cells. However, lncRNAs have historically been considered 'undruggable' targets because they do not encode for a protein product. Here, we describe the development of a new siRNA-loaded lipid nanoparticle for the therapeutic silencing of the novel oncogenic lncRNA LINC01257. Transcriptomic analysis of children with AML identified LINC01257 as specifically expressed in t(8;21) AML and absent in healthy patients. Using NxGen microfluidic technology, we efficiently and reproducibly packaged anti-LINC01257 siRNA (LNP-si-LINC01257) into lipid nanoparticles based on the FDA-approved Patisiran (Onpattro®) formulation. LNP-si-LINC01257 size and ζ-potential were determined by dynamic light scattering using a Malvern Zetasizer Ultra. LNP-si-LINC01257 internalization and siRNA delivery were verified by fluorescence microscopy and flow cytometry analysis. lncRNA knockdown was determined by RT-qPCR and cell viability was characterized by flow cytometry-based apoptosis assay. LNP-siRNA production yielded a mean LNP size of ~65 nm with PDI ≤ 0.22 along with a >85% siRNA encapsulation rate. LNP-siRNAs were efficiently taken up by Kasumi-1 cells (>95% of cells) and LNP-si-LINC01257 treatment was able to successfully ablate LINC01257 expression which was accompanied by a significant 55% reduction in total cell count following 48 h of treatment. In contrast, healthy peripheral blood mononuclear cells (PBMCs), which do not express LINC01257, were unaffected by LNP-si-LINC01257 treatment despite comparable levels of LNP-siRNA uptake. This is the first report demonstrating the use of LNP-assisted RNA interference modalities for the silencing of cancer-driving lncRNAs as a therapeutically viable and non-toxic approach in the management of AML.

CEBPA-bZip mutations are associated with favorable prognosis in de novo AML: a report from the Children's Oncology Group.

Biallelic CEBPA mutations are associated with favorable outcomes in acute myeloid leukemia (AML). We evaluated the clinical and biologic implications of CEBPA-basic leucine zipper (CEBPA-bZip) mutations in children and young adults with newly diagnosed AML. CEBPA-bZip mutation status was determined in 2958 patients with AML enrolled on Children's Oncology Group trials (NCT00003790, NCT0007174, NCT00372593, NCT01379181). Next-generation sequencing (NGS) was performed in 1863 patients (107 with CEBPA mutations) to characterize the co-occurring mutations. CEBPA mutational status was correlated with disease characteristics and clinical outcomes. CEBPA-bZip mutations were identified in 160 (5.4%) of 2958 patients, with 132 (82.5%) harboring a second CEBPA mutation (CEBPA-double-mutated [CEBPA-dm]) and 28 (17.5%) had a single CEBPA-bZip only mutation. The clinical and laboratory features of the 2 CEBPA cohorts were very similar. Patients with CEBPA-dm and CEBPA-bZip experienced identical event-free survival (EFS) of 64% and similar overall survival (OS) of 81% and 89%, respectively (P = .259); this compared favorably to EFS of 46% and OS of 61% in patients with CEBPA-wild-type (CEBPA-WT) (both P < .001). Transcriptome analysis demonstrated similar expression profiles for patients with CEBPA-bZip and CEBPA-dm. Comprehensive NGS of patients with CEBPA mutations identified co-occurring CSF3R mutations in 13.1% of patients and GATA2 mutations in 21.5% of patients. Patients with dual CEBPA and CSF3R mutations had an EFS of 17% vs 63% for patients with CEBPA-mutant or CSF3R-WT (P < .001) with a corresponding relapse rate (RR) of 83% vs 22%, respectively (P < .001); GATA2 co-occurrence did not have an impact on outcome. CEBPA-bZip domain mutations are associated with favorable clinical outcomes, regardless of monoallelic or biallelic status. Co-occurring CSF3R and CEBPA mutations are associated with a high RR that nullifies the favorable prognostic impact of CEBPA mutations.

Mesothelin is a novel cell surface disease marker and potential therapeutic target in acute myeloid leukemia.

In an effort to identify acute myeloid leukemia (AML)-restricted targets for therapeutic development in AML, we analyzed the transcriptomes of 2051 children and young adults with AML and compared the expression profile with normal marrow specimens. This analysis identified a large cohort of AML-restricted genes with high expression in AML, but low to no expression in normal hematopoiesis. Mesothelin (MSLN), a known therapeutic target in solid tumors, was shown to be highly overexpressed in 36% of the AML cohort (range, 5-1077.6 transcripts per million [TPM]) and virtually absent in normal marrow (range, 0.1-10.7 TPM). We verified MSLN transcript expression by quantitative reverse transcription polymerase chain reaction, confirmed cell surface protein expression on leukemic blasts by multidimensional flow cytometry, and demonstrated that MSLN expression was associated with promoter hypomethylation. MSLN was highly expressed in patients with KMT2A rearrangements (P < .001), core-binding factor fusions [inv(16)/t(16;16), P < .001; t(8;21), P < .001], and extramedullary disease (P = .001). We also demonstrated the presence of soluble MSLN in diagnostic serum specimens using an MSLN-directed enzyme-linked immunosorbent assay. In vitro and in vivo preclinical efficacy of the MSLN-directed antibody-drug conjugates (ADCs) anetumab ravtansine and anti-MSLN-DGN462 were evaluated in MSLN+ leukemia cell lines in vitro and in vivo, as well as in patient-derived xenografts. Treatment with ADCs resulted in potent target-dependent cytotoxicity in MSLN+ AML. In this study, we demonstrate that MSLN is expressed in a significant proportion of patients with AML and holds significant promise as a diagnostic and therapeutic target in AML, and that MSLN-directed therapeutic strategies, including ADCs, warrant further clinical investigation.

Development and Validation of a Deep Learning Algorithm for Gleason Grading of Prostate Cancer From Biopsy Specimens.

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.