Journal of Pathology Clinical Research最新文献

Folate receptor alpha (FRα) is a promising therapeutic target due to its high expression in several tumor types and its rare expression in healthy tissue. Recently, the antibody-drug conjugate mirvetuximab soravtansine has been approved for treatment of advanced platinum-resistant high-grade serous carcinoma (HGSC). Immunohistochemical expression of FRα has been extensively studied in HGSC, but most studies conducted before the clinical studies targeting FRα used variable antibodies and scoring criteria, which makes comparison of older literature data with recent studies difficult. Moreover, the data regarding its expression in other types of ovarian and other female genital tract tumors are limited or absent. In our study, we focused on immunohistochemical expression in 923 tubo-ovarian and endometrial tumors (assessed on tissue microarrays), using standardized scoring criteria and the VENTANA FOLR1 CDx assay. The results of our study showed the highest FRα expression in serous carcinomas, specifically HGSC (45% positive cases), followed by low-grade serous carcinoma (25%), endometrial serous carcinoma (11%), and serous borderline tumor (10%). Endometrioid and clear cell ovarian carcinomas showed rare positivity (2% and 1%, respectively). All other tumors examined were negative, including mucinous ovarian tumors, sex cord-stromal tumors, endometrial endometrioid carcinomas, undifferentiated and dedifferentiated carcinomas, and endometrial clear cell carcinomas. In conclusion, these results confirm that FRα expression in HGSC and LGSC reaches similar values compared to published data, and is present in a minority of endometrial serous carcinomas. In other ovarian and endometrial tumors examined, FRα expression is absent or rare.

Myeloid sarcoma (MS) is a tumorous extramedullary proliferation of blast or blast equivalent cells (e.g., promonocytes or promyelocytes). The most frequent cutaneous presentation is often referred to as leukemia cutis (LC). These lesions, especially without the clinical context of a known bone marrow disease, pose a differential diagnostic challenge. In this retrospective multicenter clinico-pathological study on 154 patients with MS or LC, 169 samples were analyzed by morphology, immunohistochemistry, and fluorescent in situ hybridization, and a subset by additional sequencing [TP53]. The majority of cases were lysozyme positive (diffuse in 91% and focal in 5%), 51% showed diffuse and 6% focal expression of CD56, and IRF8 was strongly positive in 31% of the lesions. Lack of myeloperoxidase (MPO), CD117, and CD34 expression was observed in 27%, 39%, and 58%, respectively. PU.1 was positive in almost all instances (95%), but BRAF V600E was consistently negative. CD123 was diffusely (13%) or focally (25%) positive, which, in addition to frequent CD4 (73%) and CD56 expression, pointed to a phenotypic overlap with blastic plasmacytoid dendritic cell neoplasms. Survival analysis revealed that MS occurring at sanctuary sites (CNS, orbit, ovary, and testis) was characterized by excellent survival. Similarly to histiocytoses, there was a prognostic difference between isolated and multisystemic involvement by MS. Patients who underwent allogeneic hematopoietic stem cell transplantation showed significantly improved survival. In conclusion, this multicenter study suggests that most MS are of myelomonocytic/monoblastic origin, a high proportion of them are NPM1 mutated, and may lack expression of MPO and CD34. NPM1 mutation-specific antibodies should be integrated into the diagnostic panels for MS or LC, while IRF8 and PU.1 are not recommended as they cannot distinguish MS from histiocytic neoplasms.

Distinguishing infiltrative basal cell carcinoma (BCC) from poorly differentiated cutaneous squamous cell carcinoma (cSCC) remains a significant histopathological challenge. Automated deep learning approaches hold promise for improving diagnostic reliability, yet robust external validation is essential. In this study, we developed a weakly supervised deep learning model to classify these diagnostically challenging subtypes and evaluated its generalizability across internal and external cohorts, as well as in comparison to a dermatopathology foundation model (HistoGPT). The model employed a multiple-instance learning framework (CLAM) using the histopathology-specific transformer Phikon for feature extraction from whole-slide images. Slide-level ground-truth diagnoses from the collected images (n = 335, University Hospital Erlangen) were derived from routine clinical practice and re-evaluated by two board-certified dermatopathologists. Performance was assessed on an internal test set of 84 whole-slide images (27 cSCC and 57 BCC) and two external datasets: Queensland cohort (n = 10, curated in-distribution cases) and the COBRA cohort (n = 200, broad, partly out-of-distribution cases). Model discrimination was quantified using ROC curves, while accuracy, sensitivity, and specificity were reported alongside 95% Wilson confidence intervals (CIs). On the internal test set, the model achieved perfect classification [area under the receiver operating characteristic (AUC) = 1.0; 100% accuracy, sensitivity, and specificity]. Similarly, strong performance was observed in the Queensland cohort (AUC = 1.0), although limited by sample size. In the more heterogeneous COBRA cohort, discrimination remained high (AUC = 0.923, 95% CI 0.885-0.961), requiring threshold adjustment to correct for marked calibration shift (balanced accuracy 86.5% at Youden's J). Attention heatmaps highlighted histologically meaningful regions. In zero-shot evaluation on the internal test set, HistoGPT achieved an overall accuracy of 77%, with high class-wise sensitivity for BCC (98%, 95% CI 91-100) but markedly reduced sensitivity for cSCC (33%, 95% CI 19-52). Fine-tuning a task-specific classifier on the HistoGPT backbone substantially improved performance, achieving near-perfect discrimination and 98% balanced accuracy. These findings demonstrate that weakly supervised deep learning enables highly accurate classification of diagnostically challenging BCC and cutaneous squamous cell carcinoma subtypes. However, reliable deployment across institutions necessitates careful calibration and domain adaptation, and even powerful foundation models such as HistoGPT benefit from targeted fine-tuning to ensure robust performance in dermatopathology.

This prospective single-center study aimed to investigate whether preserving cell-free ribonucleic acid (cfRNA) in specimens' supernatants (SS) obtained from percutaneous core-needle biopsy (CNB) of non-small cell lung cancer (NSCLC) can improve the performance of genomic testing. Forty-three NSCLC patients who underwent image-guided CNB were enrolled. The SS from each CNB specimen was divided into two parts: one was mixed 1:1 with a cfRNA-protective solution, and the other was left unhandled. For quantitative analysis, 30 patients (regardless of mutation status) were evaluated by comparing cfRNA yield between cfRNA-protected and -unprotected SS. For qualitative analysis, 15 patients with fusion gene alterations were assessed by comparing genotyping results from cfRNA-protected SS to those from paired CNB specimens. Pneumothorax was the most frequent adverse event in CNB procedures, occurring in 20.9% of cases (9/43). No one experienced severe adverse events. In the quantitative analysis, 90.0% (27/30) of cfRNA-protected SS yielded adequate cfRNA, significantly higher than 53.3% (16/30) in cfRNA-unprotected SS (p < 0.01). In the qualitative analysis, despite ineffective cfRNA preservation in three cases, DNA-level mutations were still detected in both the CNB specimens and cfRNA-protected SS. The overall concordance of genotyping results between cfRNA-protected SS and paired CNB specimens was 100%, correctly identifying all ALK fusions, ROS1 rearrangements, and MET-14 skipping alterations. These findings highlight that preserving cfRNA in CNB-SS from NSCLC can improve the performance of genomic testing, particularly for RNA-based assays, without compromising the accuracy of DNA-based assays.

Pancreatic ductal adenocarcinoma (PDAC) is a lethal disease, with limited therapeutic options, few patients showing targetable molecular changes. New therapeutic strategies are necessary. Antibody-drug conjugates (ADCs) have emerged as alternative therapeutic strategies across various cancer types. Herein, we analyze the expression and spatial heterogeneity (six cores per patients) of three ADC targets (c-MET, NECTIN4, and TROP-2) in a cohort of 62 PDAC patients (1,116 tissue cores) and associate their levels with clinicopathological and genomic parameters, and the expression of immune checkpoints. c-MET exhibited significantly higher expression at the tumor front versus tumor center, along with notable intratumoral heterogeneity. In contrast, NECTIN4 and TROP-2 displayed homogeneous expression patterns, with NECTIN4 being absent in approximately two-thirds of cases, while TROP-2 showed consistently strong positivity across tumor regions (98% 3+). By simulating sampling sufficiency for reliable scoring, we observed that, for c-MET, two tumor samples were sufficient to achieve a maximum score of 1+, while for higher scores (2+ and 3+), four samples were required. For NECTIN4, four samples were necessary to detect scores of 1+ and 2+. For TROP-2, for a 3+ score, just two samples were sufficient to reach the maximum score. c-MET or TROP-2 expression scores were not associated with any clinicopathological parameters. In contrast, NECTIN4 expression showed an association with tumor grade. Correlations with immune checkpoints revealed that high TROP-2 expression was inversely correlated with PD-L1 expression. For all three markers no significant differences in expression were found between SMAD4 wild-type and SMAD4-mutated tumors, nor between TP53 wild-type and TP53-mutated tumors. Furthermore, analysis of lymph node and distant (liver and peritoneal) metastases revealed significantly higher c-MET and NECTIN4 expression in the metastatic setting. In conclusion, TROP-2 is highly expressed in most PDACs, independent of clinicopathological and genomic parameters, and inversely correlating with PD-L1, making TROP-2 an ideal ADC target.

The current melanoma staging system predicts 74% of the variance in survival, with prognostic biomarkers subject to high levels of inter-observer variation. This work assesses whether a previously developed convolutional neural network (CNN) for invasive melanoma segmentation in whole slide images (WSIs) may reveal new insights into melanoma morphology and patient prognosis. This paper uses Cox proportional multivariate regression analyses to evaluate the ability of the CNN outputs to predict patient survival across 745 WSIs from 5 data sources. Five objective histomorphological parameters of tumour size and shape that are independently associated with overall and melanoma-specific survival were created from the CNN: tumour area(log) (HR 1.48 CI 1.30-1.68, p < 0.001), tumour perimeter(log) (HR 1.86 CI 1.48-2.32, p < 0.001), major axis length(log) (HR 1.88 CI 1.42-2.48, p < 0.001), Nodularity Index(log) (HR 1.77 CI 1.28-2.43, p < 0.001) and digital Breslow thickness(log) (HR 2.04, CI 1.63-2.54, p < 0.001). These results indicate that melanoma segmentation of the entire lesion within a WSI may be used to predict patient outcome. Moreover, this technology can be used to make new morphological discoveries to provide information not currently contained within our staging system (e.g. Nodularity Index), as well as provide objectivity and automation of current biomarkers (e.g. digital Breslow thickness). Further work is required to validate this initial discovery and evaluation.

Histopathological analysis is considered the gold standard for the diagnosis and prognostication of cancer. Recent advances in AI, driven by large-scale digitisation and pan-cancer foundation models, are opening new opportunities for clinical integration. However, it remains unclear how robust these foundation models are to real-world sources of variability, particularly in H&E staining and scanners produced by different manufacturers. In this study, we use soft tissue tumours, a rare and morphologically diverse tumour type, as a challenging test case to systematically investigate the colour-related robustness and generalisability of seven AI models. Controlled staining and scanning experiments were utilised to assess model performance across diverse real-world data sources. Foundation models, particularly UNI-v2, Virchow and TITAN, demonstrated encouraging robustness to staining and scanning variation, particularly when a small number of stain-varied slides were included in the training loop, highlighting their potential as adaptable and data-efficient tools for real-world digital pathology workflows.

Villitis of unknown etiology (VUE) is a chronic placental inflammatory lesion characterized by lymphohistiocytic infiltration and destruction of villous architecture in the absence of infection. Although VUE is well recognized for its association with fetal growth restriction and adverse pregnancy outcomes, its clinicopathologic correlates and molecular basis remain poorly understood. VUE cases were identified from the pathology database of Jinan Maternal and Child Health Hospital (2020-2023) and classified as high-grade or low-grade according to the Amsterdam criteria. Clinical data, maternal complications, and neonatal outcomes were collected from electronic medical records. Multivariable logistic regression was used to determine independent risk factors. RNA sequencing was performed on 28 placental samples (24 VUE and 4 controls) to identify differentially expressed genes and pathways. A total of 970 cases (381 high-grade and 589 low-grade) and 980 controls were included. VUE prevalence was 5.8%. Hypertensive disorders of pregnancy (HDP) were independently associated with VUE occurrence (odds ratio 1.67, 95% CI 1.28-2.19, p < 0.001). VUE placentas frequently exhibited chronic intervillositis, chronic deciduitis, and avascular villi, whereas maternal vascular malperfusion showed no significant difference from controls. High-grade VUE was significantly associated with low-birth-weight, small-for-gestational-age infants, and increased neonatal intensive care unit admissions, indicating a severity-dependent impact on neonatal outcomes. Transcriptomic profiling revealed robust upregulation of interferon-inducible, cytotoxic, and chemokine genes - most notably GBP5, CXCL9, and CXCL10 - with enrichment of interferon-γ, IL-6/JAK/STAT3, TNF-α/NF-κB, and antigen presentation pathways. VUE, particularly its high-grade form, is a significant placental lesion associated with HDP, adverse neonatal outcomes, and avascular villi. Its distinct interferon-rich molecular profile, consistent with a maternal anti-fetal T-cell-mediated process, underscores its clinical and biological importance. GBP5 emerges as a potential biomarker of interferon-driven inflammation, providing new mechanistic insight and diagnostic relevance for placental immune injury.

Fibroblastic proliferation in various tumor microenvironments influences cancer survival through complex interactions with diverse immune responses. This study investigated the impact of histologically unique activated cancer-associated fibroblasts (aCAFs) on survival outcomes and immune responses and examined their association with various pathophysiological mechanisms. We analyzed a total of 1,024 colorectal adenocarcinoma patients from two cohorts. aCAFs were evaluated based on hematoxylin and eosin-stained whole-slide images, and their associations with clinicopathological features, immune cell infiltration, and survival were assessed. We developed a machine learning-based survival prediction model incorporating aCAFs and clinicopathologic parameters. Additionally, we performed differential gene expression analysis, functional enrichment analyses, and in vitro drug screening of aCAF-related genes. aCAFs were associated with advanced T stage, lymphovascular invasion, perineural invasion, and decreased CD8+ and CD4+ T cell infiltration. aCAFs were also associated with worse overall and disease-free survival in both univariate and multivariate analyses. Functional enrichment analysis revealed that aCAF-related genes were implicated in immunosuppressive signaling, oxidative stress regulation, and tumor progression pathways. Survival prediction models based on machine learning and incorporating aCAFs demonstrated superior prognostic accuracy for overall survival and disease-free survival compared to models excluding aCAFs. Our analysis of aCAFs' association with immune responses through bioinformatics-based genomic analysis and machine learning provides a foundation for future research in CRC patients.

Recent advancements in antibody-drug conjugates, including FDA-approved therapies targeting Nectin-4 and Trop-2, have transformed the cancer treatment landscape, including upper tract urothelial carcinoma (UTUC). However, varied treatment effects and drug-associated adverse effects raise the question of whether patients should be selected based on a biomarker study to achieve optimal outcomes. A better understanding of the patterns and clinicopathological significance of Nectin-4 and Trop-2 expression in UTUC remains to be achieved. We generated tissue microarrays (TMAs) with 120 UTUC specimens from patients who underwent nephroureterectomy at our institution and evaluated the expression of Nectin-4 and Trop-2 in tumor and non-tumor tissue. Nectin-4 expression was significantly higher in both invasive and noninvasive high-grade UTUC compared to noninvasive low-grade tumors. In contrast, Trop-2 expression did not vary significantly between noninvasive low-grade and high-grade tumors. When analyzed by stage, Nectin-4 expression was significantly elevated in tumors of higher stages than in early-stage tumors, similar to Trop-2 expression. Although both Nectin-4 and Trop-2 were broadly expressed in tumor and adjacent non-tumor urothelium, a subset of patients demonstrated low expression in non-tumor tissue but high expression in tumor tissue. Nectin-4 expression, but not Trop-2, was significantly correlated with the Ki-67 index, indicating that they may have different roles in tumor proliferation. The differential expression of Nectin-4 and Trop-2 by tumor grade and stage highlights their potential relevance in guiding targeted therapy for UTUC. Notably, a subset of patients exhibits high expression in tumor tissue, accompanied by low expression in adjacent non-tumor urothelium, suggesting a favorable therapeutic index for antibody-drug conjugate therapy. These findings support the need for further biomarker-driven studies to optimize patient selection and treatment outcomes.