Laboratory Investigation最新文献

{"title":"Differential Expression of P-Element–Induced Wimpy Testis (PIWI)–Interacting RNAs in Glioblastoma Stem Cells Affects Their Biological Features: Implications for Tumor Progression and Patient Survival","authors":"Frantisek Siegl ,&nbsp;Elena Garcia-Borja ,&nbsp;Rosana Mateu ,&nbsp;Karolina Trachtova ,&nbsp;Matej Jasik ,&nbsp;Martin Kopecky ,&nbsp;Marek Vecera ,&nbsp;Lenka Radova ,&nbsp;Michal Hendrych ,&nbsp;Tomas Kazda ,&nbsp;Pavel Fadrus ,&nbsp;Aleksi Sedo ,&nbsp;Ondrej Slaby ,&nbsp;Petr Busek ,&nbsp;Jiri Sana","doi":"10.1016/j.labinv.2026.106072","DOIUrl":"10.1016/j.labinv.2026.106072","url":null,"abstract":"<div><div>Glioblastoma isocitrate dehydrogenase–wild-type (GBM) is the most lethal and, concurrently, the most common malignant primary brain tumor. Its aggressive behavior is associated with the presence of a subpopulation of glioblastoma stem cells (GSCs) that drive resistance to therapy and early relapse. P-element–induced wimpy testis–interacting RNAs (piRNAs), which are characteristically expressed in undifferentiated cells, are thought to play an important role in GSC biology. To identify piRNAs associated with GSCs, we selected 24 paired GSC and non-GSC cultures from patients with GBM based on the expression of stem cell markers CD133 and Sox2, ability of GSCs to form spheres in culture, and to differentiate and replicate tumors in immunodeficient mice. Global piRNA profiling using next-generation sequencing revealed 98 significantly differentially expressed piRNAs in GSCs compared with non-GSCs, including piR-9491, whose dysregulation in GBM has been previously described. Downregulation of piR-9491 in GSCs led to decreased viability, growth, invasion, and increased apoptosis. Significance of piRNA pathway for GBM pathology was further highlighted by identification of 34 piRNAs connected to patients' survival. These molecules were further used for the establishment of a piRNA signature predicting survival of patients with GBM. Our results not only confirm the important role of piR-9491 in GSCs but also indicate a potential significant role of piRNAs in the biology of GBM.</div></div>","PeriodicalId":17930,"journal":{"name":"Laboratory Investigation","volume":"106 3","pages":"Article 106072"},"PeriodicalIF":4.2,"publicationDate":"2026-01-13","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145989821","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Erratum to “Establishing and characterizing the molecular profiles, cellular features, and clinical utility of a patient-derived xenograft model using benign prostatic tissues” (Lab Invest 2024 Oct;104(10):102129)","authors":"Alexandra Lapat Polasko ,&nbsp;Dalin Zhang ,&nbsp;Avanti Ramraj ,&nbsp;Chun-Lung Chiu ,&nbsp;Fernando J. Garcia-Marques ,&nbsp;Abel Bermudez ,&nbsp;Kathryn Kapp ,&nbsp;Eric Peterson ,&nbsp;Zhengyuan Qiu ,&nbsp;Anna S. Pollack ,&nbsp;Hongjuan Zhao ,&nbsp;Jonathan R. Pollack ,&nbsp;Sharon J. Pitteri ,&nbsp;James D. Brooks","doi":"10.1016/j.labinv.2025.104275","DOIUrl":"10.1016/j.labinv.2025.104275","url":null,"abstract":"","PeriodicalId":17930,"journal":{"name":"Laboratory Investigation","volume":"106 2","pages":"Article 104275"},"PeriodicalIF":4.2,"publicationDate":"2026-01-09","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145929259","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Diagnostic Potential of Tn-MUC1 in Breast Cancer: A Novel Immunohistochemical Marker Reflecting Tumor Progression","authors":"Ai Shimizu ,&nbsp;Kanako C. Hatanaka ,&nbsp;Ayae Nange ,&nbsp;Asami Okumura ,&nbsp;Masanori Takehashi ,&nbsp;Yoshiki Shinomiya ,&nbsp;Kentaro Naruchi ,&nbsp;Masaharu Sato ,&nbsp;Hiroshi Kase ,&nbsp;Tomoko Mitsuhashi ,&nbsp;Hiroko Yamashita ,&nbsp;Yutaka Hatanaka ,&nbsp;Yoshihiro Matsuno","doi":"10.1016/j.labinv.2025.106070","DOIUrl":"10.1016/j.labinv.2025.106070","url":null,"abstract":"<div><div>Mucin 1 (MUC1) is a highly O-glycosylated transmembrane glycoprotein. Tumor-associated MUC1, characterized by aberrant O-linked glycans, is overexpressed in cancer cells; however, conventional MUC1 antibodies show limited specificity for tumor-associated glycan structures. Recently, a novel epitope-defined antibody (MUC1-Tn antigen epitope-defined antibody [MUC1-Tn ED Ab]) that specifically recognizes the Tn-MUC1 antigen was developed. In this study, we evaluated the potential of MUC1-Tn ED Abs as diagnostic markers of breast cancer. Tissue microarray sections from 124 patients with invasive breast carcinoma (IBC) and 26 whole tissue sections, including multiple neoplastic lesions—flat epithelial atypia (n = 24), ductal carcinoma in situ (n = 26), and IBC (n = 16)—were analyzed. Immunohistochemical distributions of Tn-MUC1 and MUC1 were assessed using the MUC1-Tn ED Ab (clone SN102) and a conventional antibody (clone Ma552), respectively. In tissue microarray analysis, Tn-MUC1 exhibited minimal immunoreactivity in nonneoplastic areas and high specificity for IBC. In IBC tissues, immunoreactivity with Tn-MUC1 was predominantly cytoplasmic, unlike conventional MUC1 staining observed in both the cytoplasm and membrane. In multilesion analysis, cytoplasmic Tn-MUC1 expression was rarely detected in nonneoplastic areas but progressively increased across flat epithelial atypia, ductal carcinoma in situ, and IBC. Knockdown assays in breast cancer cell lines demonstrated that core 1 β1,3-galactosyltransferase 1 (C1GALT1), an enzyme involved in galactosylation of the Tn antigen, significantly influenced the cellular localization of Tn-MUC1. This study demonstrates that Tn-MUC1, as detected by the MUC1-Tn ED Ab, has high specificity for breast cancer and may act as a novel immunohistochemical marker reflecting tumor progression.</div></div>","PeriodicalId":17930,"journal":{"name":"Laboratory Investigation","volume":"106 3","pages":"Article 106070"},"PeriodicalIF":4.2,"publicationDate":"2026-01-06","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145934348","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"The Value of Internal Positive Controls and Synthetic Calibrators for Performance Assessment of Breast Carcinoma Estrogen Receptor Immunohistochemistry Assay","authors":"Emina E. Torlakovic ,&nbsp;Seshi R. Sompuram ,&nbsp;Søren Nielsen ,&nbsp;Steven A. Bogen","doi":"10.1016/j.labinv.2025.106065","DOIUrl":"10.1016/j.labinv.2025.106065","url":null,"abstract":"<div><div>In 2010, the American Society of Clinical Oncology and the College of American Pathologists jointly issued guidelines for estrogen receptor (ER) and progesterone receptor testing and an update in 2020. Both guidelines recommended that ER and progesterone receptor assays be interpreted in the context of internal and external controls. Despite the considerable importance of internal controls for ER testing by immunohistochemistry (IHC), there are no studies that assess their performance and effectiveness. The Canadian Immunohistochemistry Quality Control performed an educational proficiency testing run for 70 laboratories using a tissue microarray with 80 samples of breast carcinoma enriched for tumors weakly positive for ER. The tissue microarray contained several internal positive controls and was run together with ER calibrators from Boston Cell Standards. The internal positive controls were scored as negative in 0/70 (0%), weak in 12/70 (17%), moderate in 29/70 (41%), and strong in 29/70 (41%) laboratories secondary to differences in analytical sensitivity of IHC protocols. For calibrators, only SP1 clone results (52 participants) are reported because of the small number of laboratories using other primary antibody clones. Thirty-nine laboratories (75%) had IHC protocols with (limit of detection) LOD &lt; 25,000. Both internal positive controls and Boston Cell Standard ER calibrators showed excellent correlation with overall H-score mean, percent 3+ positive cells, and percent negative cells (<em>P</em> &lt; .0001). The difference in diagnostic sensitivity was significant between weak and moderate and between weak and strong internal positive controls (<em>t</em> test, <em>P</em> = .015 and <em>P</em> = .009, respectively) and between IHC protocols with LOD &lt; 25,000 and all others (<em>t</em> test, <em>P</em> = .001). The measured LOD had a significant association with diagnostic accuracy at 1% and 10% cutoffs, whereas the internal positive control score was significantly associated only with diagnostic accuracy at the 10% cutoff (1-way analysis of variance, <em>P</em> = .009). Negative results of the ER IHC assay should not be reported when internal positive controls are either negative or weak. Further studies with more advanced calibrators are needed in order to establish the recommended LOD range for optimal ER IHC assay performance with SP1 and other anti-ER primary antibody clones.</div></div>","PeriodicalId":17930,"journal":{"name":"Laboratory Investigation","volume":"106 3","pages":"Article 106065"},"PeriodicalIF":4.2,"publicationDate":"2026-01-06","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145934308","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"","authors":"","doi":"","DOIUrl":"","url":null,"abstract":"","PeriodicalId":17930,"journal":{"name":"Laboratory Investigation","volume":"106 2","pages":"Article 104270"},"PeriodicalIF":4.2,"publicationDate":"2026-01-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146024080","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Comprehensive Spatial Methods (CSM): a Toolbox for Spatially Analyzing Tissues in Histopathology","authors":"Alvaro Lopez-Janeiro ,&nbsp;Eduardo Miraval-Wong ,&nbsp;Paulo Perez-Dominguez ,&nbsp;Raluca Alexandru ,&nbsp;Maria Guadalupe García Vazquez ,&nbsp;David Hardisson ,&nbsp;Alberto Peláez-Garcia ,&nbsp;David Ruiz-Guillamon ,&nbsp;Ignacio Melero ,&nbsp;Carlos E. de Andrea","doi":"10.1016/j.labinv.2025.104276","DOIUrl":"10.1016/j.labinv.2025.104276","url":null,"abstract":"<div><div>The image-based determination of proteins with spatial context has revolutionized our understanding of biology in different fields, including developmental biology, immunology, and oncology. The popularization of multiplex and high-plex tissue imaging methods has allowed researchers to simultaneously interrogate multiple tissue proteins with high spatial resolution in a single tissue section. Although these technologies offer many opportunities, analytical challenges have also emerged. Currently, no single analytical pipeline covers the entire spectrum of analyses required to harness the potential of these spatial platforms. Here, we present Comprehensive Spatial Methods (CSM), an R-based analysis toolbox designed to analyze multiplex and high-plex omics with high spatial resolution. CSM covers all the analytical steps, from cell and tissue segmentation and protein expression normalization to cell phenotyping, spatial heterogeneity analysis, cell-to-cell spatial interaction determination, and cellular neighborhood analysis. CSM relies on top-performing R libraries to deliver a user-friendly experience. We test the performance of CSM on a set of multiplex and high-plex images of endometrial, breast, and colorectal carcinomas and nontumoral lymph node, skin, and lung tissue. We demonstrate that the ability of CSM to phenotype and quantify cells is better than that of other state-of-the-art resources. In addition, we show that the different approaches implemented in CSM for assessing cell phenotypes, spatial heterogeneity, cell-to-cell interactions, and tissue neighborhoods cover a broad range of analytical scenarios. The freely available CSM toolbox covers many of the analytical needs of researchers working with spatially resolved histopathology data.</div></div>","PeriodicalId":17930,"journal":{"name":"Laboratory Investigation","volume":"106 2","pages":"Article 104276"},"PeriodicalIF":4.2,"publicationDate":"2025-12-26","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145849464","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Roles for the Nerve Growth Factor-Farnesoid X Receptor-Retinol Dehydrogenase 16 Axis in Hepatocellular Carcinoma Prognosis and Chemosensitivity","authors":"Po-Han Chen ,&nbsp;Yuan-Chieh Yang ,&nbsp;Huoy-Rou Chang ,&nbsp;Po-Huang Lee ,&nbsp;Yu-Chun Lin ,&nbsp;Ming-Huei Chou ,&nbsp;Der-An Tsao ,&nbsp;Cheuk-Kwan Sun ,&nbsp;Ming-Hong Tai ,&nbsp;Ying-Hsien Kao","doi":"10.1016/j.labinv.2025.104273","DOIUrl":"10.1016/j.labinv.2025.104273","url":null,"abstract":"<div><div>Nerve growth factor (NGF) has been shown to upregulate expression of retinol dehydrogenase 16 (RDH16) and farnesoid X receptor (FXR), thereby conferring hepatoprotection against cholestatic and oxidative injury. This study investigated the expression levels of NGF, RDH16, and FXR in human hepatocellular carcinoma (HCC) tissues and explored their relationships, as well as their roles in tumorigenesis and resistance to antitumor chemotherapy. Transcriptomic data set and Kaplan-Meier survival analyses revealed that HCC patients with higher <em>NGF</em> and <em>RDH16</em> expression exhibited better survival rates. Clinically, NGF protein expression was positively correlated with RDH16 levels in human HCC tissues, whereas RDH16 showed a negative correlation with proliferating cell nuclear antigen levels. In vitro experiments demonstrated that recombinant NGF treatment modestly increased proliferation of SK-Hep1 HCC cells. However, NGF pretreatment significantly enhanced cisplatin-induced and doxorubicin-induced cytotoxicity, accompanied by preserved constitutive FXR expression under chemotherapeutic stress, suggesting an association between NGF-regulated FXR expression and enhanced chemosensitivity in HCC cells. Chromatin immunoprecipitation-qPCR analysis further confirmed that FXR directly binds to the <em>RDH16</em> promoter, providing mechanistic evidence of transcriptional regulation. The correlation between higher <em>FXR</em> expression and improved HCC patient survival, along with the in vitro enhancement of chemosensitivity by FXR agonist GW4064, supported the tumor-suppressive role of FXR in HCC. Conversely, small interfering RNA–mediated FXR gene silencing significantly induced drug resistance and completely abolished the NGF-enhanced chemosensitivity. In conclusion, hepatic NGF expression may maintain higher FXR and RDH16 levels, thereby modulating drug metabolism machinery and survival signaling in HCC cells.</div></div>","PeriodicalId":17930,"journal":{"name":"Laboratory Investigation","volume":"106 2","pages":"Article 104273"},"PeriodicalIF":4.2,"publicationDate":"2025-12-19","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145804857","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Comprehensive Transcriptome and miRNome Profiling in Metachronous Colorectal Liver Metastasis: Insight Into the Prognostic and Molecular Subtypes","authors":"Bhavana Hemantha Rao ,&nbsp;Veronika Boušková ,&nbsp;Lucie Heczko ,&nbsp;Petr Holý ,&nbsp;Karolína Šeborová ,&nbsp;Václav Liška ,&nbsp;Ondřej Vyčítal ,&nbsp;Ondřej Fiala ,&nbsp;Pavel Souček ,&nbsp;Viktor Hlaváč","doi":"10.1016/j.labinv.2025.104274","DOIUrl":"10.1016/j.labinv.2025.104274","url":null,"abstract":"<div><div>In colorectal cancer, approximately 50% to 70% of metastases go into the liver; however, their molecular signature remains unknown. We aimed to investigate the transcriptome and miRNome profiles of metachronous colorectal liver metastasis (mCLM) within the hepatic microenvironment and to identify key deregulated genes, microRNAs (miRNAs), pathways, and their clinical relevance. We performed differential expression analysis on 36 mCLM and adjacent nonmalignant liver tissue pairs using RNA sequencing. Gene set enrichment analysis and consensus molecular subtype (CMS) classification helped to explore pathways. Tumor samples were stratified based on their <em>KRAS</em> mutation status. miRNA-mRNA interactions were investigated through coexpression and correlation analysis, with prognostic relevance assessed using survival analysis. Validation of key interactions was accomplished using multiMiR. We identified 1809 upregulated and 1639 downregulated genes and 108 upregulated and 92 downregulated miRNAs in mCLM compared with the adjacent nonmalignant liver. Upregulated genes were associated with epithelial-to-mesenchymal transition, G2M checkpoints, and E2F targets. About 47% of samples belonged to CMS2 and 22% to mesenchymal CMS4, with distinct mutational patterns. mRNA coexpression identified 4 clusters (associated with metabolism, cell cycle, DNA metabolism, and oncogenic signaling pathways), and miRNA coexpression identified 6 clusters. The hub miRNAs hsa-let-7c, hsa-miR-21-5p, hsa-miR-106a-5p, hsa-miR-139-5p, hsa-miR-101-3p, and hsa-miR-20b-5p were among the inversely correlated miRNA-mRNA clusters. An integrative analysis highlighted <em>PEA15</em> interaction with hsa-miR-320b/c, <em>TEX2/CTSO</em> with hsa-miR-103a-3p, and <em>PHLDA3</em> with hsa-miR-1304, and prognostic relevance for <em>ZNF441, CTSO, TEX2, EID1, CMC1,</em> hsa-miR-4634, hsa-miR-3184-5p, has-miR-320b, hsa-miR-1304-3p, hsa-miR-7-1-3p, hsa-miR-144-3p, hsa-miR-1303, and hsa-miR-660-3p. The miRNA-mRNA interactions were validated using real-time PCR in independent patient cohorts. This study revealed a complex molecular landscape of mCLM within the hepatic microenvironment and novel miRNA-mRNA interactions with potential prognostic and therapeutic implications.</div></div>","PeriodicalId":17930,"journal":{"name":"Laboratory Investigation","volume":"106 2","pages":"Article 104274"},"PeriodicalIF":4.2,"publicationDate":"2025-12-19","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145804897","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Letter to the Editor (Correspondence): Re: Prognostic Value of PARP1 and PARP2 Copy Number Alterations in Prostate Cancer","authors":"Jiang Shi,&nbsp;Lin Jiang,&nbsp;ShiWang Yuan,&nbsp;Peng Chen,&nbsp;Tao Li","doi":"10.1016/j.labinv.2025.104265","DOIUrl":"10.1016/j.labinv.2025.104265","url":null,"abstract":"","PeriodicalId":17930,"journal":{"name":"Laboratory Investigation","volume":"106 1","pages":"Article 104265"},"PeriodicalIF":4.2,"publicationDate":"2025-12-18","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145786480","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Using Consensus-Based Reasoning and Large Language Models to Extract Structured Data From Surgical Pathology Reports","authors":"Aakash Tripathi ,&nbsp;Asim Waqas ,&nbsp;Kavya Venkatesan ,&nbsp;Ehsan Ullah ,&nbsp;Asma Khan ,&nbsp;Farah Khalil ,&nbsp;Wei-Shen Chen ,&nbsp;Zarifa Gahramanli Ozturk ,&nbsp;Daryoush Saeed-Vafa ,&nbsp;Marilyn M. Bui ,&nbsp;Matthew B. Schabath ,&nbsp;Ghulam Rasool","doi":"10.1016/j.labinv.2025.104272","DOIUrl":"10.1016/j.labinv.2025.104272","url":null,"abstract":"&lt;div&gt;&lt;div&gt;Surgical pathology reports provide essential diagnostic information critical for cancer staging, treatment planning, and cancer registry documentation. However, their writing styles and formats vary widely, reflecting each pathologist’s stylistic choices, institutional norms, and inherited practices from residency training. When performing large-scale data analysis, this unstructured nature and variability across tumor types and institutions pose significant hurdles for automated data extraction. To overcome these challenges, we present a consensus-driven, reasoning-based framework that adapts multiple locally deployed large language models (LLMs) to extract both standard diagnostic variables (such as site, laterality, histology, stage, grade, and behavior) and organ-specific biomarkers. Each LLM generates structured outputs, accompanied by justifications, which are subsequently evaluated for accuracy and coherence by 3 separate reasoning models (DeepSeek-R1-large, Qwen3-32B, and QWQ-32B). Final consensus values are determined through aggregation. Board-certified pathologists conducted expert validation. This framework was applied to &gt;6100 pathology reports from The Cancer Genome Atlas (TCGA), spanning 10 organ systems, and 510 reports from Moffitt Cancer Center. For TCGA data set, automated evaluation demonstrated a mean accuracy of 84.9% ± 7.3%, with histology (88%), site (87%), stage (84%), and behavior (84%) showing the highest extraction accuracy averaged across all models. Expert review of randomly selected 138 reports confirmed high agreement for behavior (100.0%), histology (99%), grade (96%), and site (95%) in TCGA data set, with slightly lower performance for stage (89%) and laterality (88%). In Moffitt Cancer Center reports (brain, breast, and lung), accuracy remained high (88.2% ± 7.2%), with behavior (99%), histology (97%), laterality (96%), grade (94%), and site (93%) achieving strong agreement. Biomarker extraction achieved 70.6% ± 7.9% overall accuracy, with TP53 (84%) on brain tumor, Ki-67 (68%) on breast cancer, and ROS1 (82%) on lung cancer showing the highest accuracy. Interevaluator agreement analysis revealed high concordance (correlation values ≥0.93) across the 3 evaluation models. Statistical analyses revealed significant main effects of model type (F = 1716.82, &lt;em&gt;P&lt;/em&gt; &lt; .001), variable (F = 3236.68, &lt;em&gt;P&lt;/em&gt; &lt; .001), and organ system (F = 1946.43, &lt;em&gt;P&lt;/em&gt; &lt; .001), as well as model × variable × organ interactions (F = 24.74, &lt;em&gt;P&lt;/em&gt; &lt; .001), emphasizing the role of clinical context in model performance. These results highlight the potential of stratified, multiorgan evaluation frameworks with multievaluator consensus in LLM benchmarking for clinical applications. Overall, this consensus-based approach demonstrated that locally deployed LLMs can provide a transparent, accurate, and auditable solution for integration into real-world pathology workflows, such as synoptic reporting and can","PeriodicalId":17930,"journal":{"name":"Laboratory Investigation","volume":"106 2","pages":"Article 104272"},"PeriodicalIF":4.2,"publicationDate":"2025-12-16","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145781600","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}