
Endocervical gastric-type adenocarcinoma (GAS) is one of the most aggressive subtypes of cervical cancer and is frequently underdiagnosed due to morphological ambiguity, leading to delayed diagnosis. Despite the availability of molecular and genomic assays, their high cost, complexity, and limited reproducibility restrict clinical use. This study therefore proposes a highly sensitive artificial intelligence (AI)-assisted diagnostic system for GAS based exclusively on H&E-stained histopathological images. We included 309 slides from 96 GAS cases collected at Peking University Third Hospital from January 2018 to January 2025, representing the largest GAS cohort reported to date for AI research. In addition, we incorporated other morphologically analogous diseases, encompassing a total of 1,320 slides sourced from four categories: normal cervical mucosa (NORM), benign endocervical lesion entities (BELE), HPV-associated adenocarcinoma (HPVA), and endometrioid carcinoma with mucinous differentiation (ECMD). We developed GASPath, based on a novel multiple instance learning framework that efficiently captures fine-grained morphological variations from H&E-stained images. Beyond internal validation, GASPath was evaluated across 12 independent retrospective cohorts and further subjected to large-scale real-world validation on more than 7,000 samples from March 2024 to April 2025. Across three stages, GASPath demonstrated high performance. In internal validation (Stage I), it achieved an accuracy of 0.980 (95% CI 0.977-0.983) and an ROC-AUC of 0.995 (95% CI 0.994-0.997). In external validation (Stage II), the sensitivity reached 0.902 and improved to 0.968 with proposed strategies. For biopsy samples, GASPath achieved an ROC-AUC of 0.990 (95% CI 0.984-0.997). In large-scale real-world deployment (Stage III, n = 7,056), GASPath achieved a balanced accuracy of 0.953, with 100% sensitivity for GAS (45/45 cases correctly identified). The heatmaps highlight morphological features of GAS that are easily underestimated, such as irregular, angulated glands, subtle loss of nuclear polarity, and mild cytologic atypia, which show substantial morphological overlap with other diagnostic categories. GASPath enables high-sensitivity detection of GAS in routine H&E-stained slides, obviating the need for extensive auxiliary testing while preventing underdiagnosis and misdiagnosis. This advancement addresses a critical gap by streamlining diagnostic workflows without compromising accuracy. Its implementation could enable cost-effective, scalable AI-assisted diagnostics, potentially transforming the early detection and management of this aggressive cancer subtype.
Few studies have evaluated the impact of organ-specific tumor microenvironments (TMEs) on clinical outcomes in diffuse large B-cell lymphoma (DLBCL). This study investigated potential molecular markers, the immune cell composition within the TME, and their associations with clinical outcomes in patients with primary intestinal DLBCL (PI-DLBCL). We initially analyzed RNA expression in tumor cells from 19 patients with PI-DLBCL in the experimental cohort. Candidate biomarkers were then assessed in lymphoma tissue from a total of 48 patients in the same cohort, including the 19 with RNA-expression data, and validated in an independent cohort of 29 patients with PI-DLBCL. Associations between these markers and clinicopathological features, event-free survival (EFS), and overall survival (OS) in the two cohorts were analyzed. In the RNA expression panel, CCL5 expression was higher in patients with advanced-stage PI-DLBCL and was associated with inferior EFS and OS. Its principal receptor, CCR5, expressed in approximately one-third of patients, was associated with lower complete response rates to first-line immunochemotherapy: 50% in the experimental cohort and 10% in the validation cohort and correlated with the 7-year worse OS rate in both the experimental (49.8% versus 71.4%, p = 0.082) and validation (0% versus 76.2%, p < 0.001) cohorts. In both the experimental and validation cohorts, CCR5-positive tumors exhibited decreased CD86-positive (M1-like) and increased CD206-positive (M2-like) macrophage infiltration, consistent with an immunosuppressive TME. Multivariate analyses revealed that CCR5 expression was independently associated with worse EFS (p < 0.001) and OS (p = 0.004) in patients with PI-DLBCL (combined experimental and validation cohorts). CCR5 expression indicates a biologically aggressive subtype of PI-DLBCL with poor clinical outcomes and M2 macrophage-dominant immunosuppression.
B-cell precursor acute lymphoblastic leukemia (preB-ALL) is characterized by pathogenic variants currently used in precision oncology. However, the mutational landscape of Mexican children with preB-ALL has not yet been thoroughly explored and defined in terms of the clinical significance. We used a custom-designed next-generation sequencing exome panel, along with high-resolution chromosome microarrays and gene fusion-targeted assays, to characterize the mutational landscape of 73 Mexican children with preB-ALL, exploring the profile of clinically significant variants. We classified the variants following the AMP-ASCO-CAP 2017 and ACMG-CG 2019 guidelines recommendations. The mutational landscape includes a broad molecular spectrum of variants affecting cell cycle regulation, B-cell development, kinase signaling, and epigenetic regulation genes. Tier 1 diagnostic variants allowed the identification of pre-B ALL genetic subtypes, diminishing the preB-ALL NOS group from 63% to 37%. Furthermore, 37% of cases presented Tier 1 variants conferring intermediate to adverse prognoses (primarily involving CRLF2 gene fusions, as well as PAX5, IKZF1, and TP53 inactivating mutations). Notably, these patients exhibited high-risk clinical features and a lower event free survival rate than patients without these variants (60% versus 41.2%; p = 0.046; 95% CI). Between 19% and 42% of patients had Tier 2 variants targetable with JAK-STAT or RAS-MAPK signaling inhibitors. These patients showed a lower overall survival rate than patients without these variants (64% versus 90%; p = 0.048; 95% CI). Tier 1 potential germline variants in cancer predisposition genes (mainly BRCA1/2 and CHEK2) were observed in 10% of patients, some of whom had a family history of cancer. This highlights the importance of genetic counseling for patients and their families. Finally, 33% of patients had Tier 3 variants that were predicted to be deleterious and potentially upgraded to pathogenic with plausible clinical relevance. In conclusion, the mutational landscape analysis revealed variants useful for oncologic management and genetic counseling of Mexican children with preB-ALL.
Thyroid transcription factor-1 (TTF-1) is a widely used immunohistochemical marker for tumors of thyroid and pulmonary origin; however, its expression profile and diagnostic relevance in primary central nervous system (CNS) tumors remain incompletely characterized, particularly following the revisions in the 2021 WHO Classification of CNS Tumors. We systematically evaluated TTF-1 expression in a cohort of 214 primary CNS tumors diagnosed between 2021 and 2024 at The First Affiliated Hospital of China Medical University and classified according to current WHO criteria. Immunohistochemistry for TTF-1 was performed, and expression patterns were correlated with histological tumor type, molecular classification, and WHO grade. Statistical analyses were conducted to compare expression frequencies across tumor subtypes. TTF-1 showed tumor-specific expression patterns. It was highly sensitive in central neurocytomas (89.47%) and fully specific for subependymal giant cell astrocytoma (SEGA) among circumscribed gliomas. Focal TTF-1 positivity was sporadically detected in high-grade diffuse gliomas, without significant links to molecular alterations, anatomical sites, or other common immunohistochemical indicators. TTF-1 was consistently negative in multiple newly classified and rare CNS tumors, whereas partial intracranial mesenchymal tumors with FET-CREB fusions presented TTF-1 positivity. In conclusion, TTF-1 serves as a reliable sensitive marker for central neurocytomas and a highly specific marker for SEGA. Given its occasional positivity in high-grade gliomas and certain intracranial mesenchymal tumors, pathologists should interpret TTF-1 results prudently to avoid misdiagnosing primary CNS lesions as metastatic tumors. TTF-1 is therefore a valuable auxiliary biomarker for difficult differential diagnosis of CNS tumors.
The B7-H family of immune checkpoint proteins plays a central role in tumor immune regulation, yet their prognostic and predictive significance in metastatic clear cell renal cell carcinoma (ccRCC) remains incompletely defined. We analyzed a retrospective cohort of primary tumor specimens from 145 patients with synchronous or metachronous metastatic ccRCC to evaluate the expression of B7-H3, B7-H4, B7-H5, and B7-H7 and their association with tumor aggressiveness, treatment response, and long-term survival. Immunohistochemistry revealed marked intratumoral heterogeneity and distinct compartmentalized expression patterns, with B7-H3 detected in both tumor cells and stromal compartments, B7-H4 exclusively expressed in tumor cells, B7-H5 localized to tumor cells and tumor-infiltrating lymphocytes (TILs), and B7-H7 detected in both tumor cells and TILs. We did not detect any overlap in expression between B7-H proteins and PD-L1 positivity in TILs. High B7-H3 expression in tumor and stroma was significantly associated with higher histological grade, increased local invasion, lymph node involvement, and advanced stage. B7-H5 expression in TILs correlated with impaired Eastern Cooperative Oncology Group performance status. Elevated B7-H3 predicted unfavorable response to tyrosine kinase inhibitors, mechanistic target of rapamycin inhibitors, and immune checkpoint inhibitors, whereas B7-H5 negativity in tumor cells associated with more favorable treatment outcomes. B7-H4 and B7-H7 expression in tumor cells were associated with features of tumor aggressiveness and immune suppression, further highlighting the potential role of multiple B7-H family checkpoints in metastatic ccRCC. In multivariable Cox regression models, B7-H3 expression in tumor and stroma independently predicted shorter disease-free and overall survival. B7-H5 expression in TILs independently predicted reduced overall survival. These findings identify B7-H3 as a robust biomarker of tumor aggressiveness, therapeutic resistance, and adverse prognosis in metastatic ccRCC, while B7-H5 in TILs provides complementary prognostic information. Integrated profiling of B7-H immune checkpoints in primary tumors may refine clinical risk stratification and support the development of biomarker-guided immunotherapeutic strategies.