Colorectal cancer histopathological grading relies on the accurate segmentation of glandular structures. Current deep learning–based methods depend heavily on large-scale pixel-level annotations that are labor-intensive and not amenable to clinical practice. Weakly supervised semantic segmentation offers a promising alternative; yet, existing class activation map–based weakly supervised semantic segmentation approaches often produce incomplete, low-quality pseudo-masks that overemphasize discriminative regions and fail to provide reliable supervision for unannotated glandular structures, limiting their suitability for dense histopathology segmentation under sparse supervision. We propose a novel weakly supervised teacher–student framework that leverages sparse pathologists’ annotations and an Exponential Moving Average–stabilized teacher network to generate refined pseudo-masks.

Our framework integrates confidence-based filtering, adaptive fusion of teacher predictions with limited ground truth, and curriculum-guided refinement, enabling the student network to progressively delineate and accurately segment unannotated glandular regions. We validated our framework on an institutional colorectal cancer cohort from The Ohio State University Wexner Medical Center, consisting of 60 hematoxylin and eosin-stained whole-slide images from independent patients with varying degrees of gland differentiation, as well as on public benchmarks including the Gland Segmentation dataset (derived from stage T3–T4 colorectal adenocarcinomas), TCGA-COAD, TCGA-READ, and SPIDER.

The proposed framework achieved strong performance on the institutional dataset despite limited annotations. On the Gland Segmentation dataset, it demonstrated competitive performance compared to both weakly and fully supervised approaches, achieving a mean Intersection over Union of 80.10% ± 1.52 and a mean Dice coefficient of 89.10% ± 2.10. Moreover, cross-cohort evaluations showed robust generalization on TCGA-COAD and TCGA-READ without requiring additional annotations, while reduced performance on SPIDER reflected pronounced domain shift.

Our framework provides an annotation-efficient and generalizable paradigm for accurate gland segmentation in colorectal histopathology, offering a practical pathway toward significantly reducing annotation burdens while preserving high segmentation fidelity.

Metaplastic breast carcinoma, a rare entity (<1% of breast neoplasms), lacks comprehensive spectroscopic characterization. This study aimed to address this gap by providing a qualitative and quantitative spectroscopic profile of metaplastic carcinoma in comparison to ductal carcinoma in situ (DCIS) and invasive ductal carcinoma (IDC).

A retrospective analysis was conducted on archival tissue blocks of metaplastic carcinoma (n = 10), DCIS (n = 12), and IDC (n = 31). Sections were stained with hematoxylin and eosin for histological confirmation. Attenuated total reflectance Fourier-transform infrared spectroscopy was performed on adjacent unstained sections, with normal breast tissue (n = 10) serving as the control. Spectral data were analyzed using t-tests to identify significant differences in peak intensities and ratios. Hierarchical clustering analysis and receiver operating characteristic curves were generated to assess the diagnostic potential of selected spectral features.

Spectral analysis revealed that mean peak intensities were generally lower in all carcinoma subtypes compared to normal breast tissue. Specific ratios, including A1237/A1080 (phosphate; p < 0.01), A1043/1543 (glycogen; p < 0.01), and A1080/A1632 (nucleocytoplasmic index; p < 0.03), were significantly elevated in carcinomatous tissues. Receiver operating characteristic analysis identified peak 3,280 (area under the curve (AUC) = 0.93–0.96) as highly effective in differentiating normal from carcinomatous tissues. Peak 2,922 showed specificity for distinguishing normal tissue from IDC (AUC ≈ 0.7). Peak 1,744 effectively discriminated between DCIS and metaplastic carcinoma (AUC = 0.7). The ratio 1,080/1,632 (nucleocytoplasmic ratio) demonstrated exceptional diagnostic accuracy, distinguishing normal from carcinomatous tissues (AUC ≈ 1.0), DCIS from IDC (AUC ≈ 0.86), and DCIS from metaplastic carcinoma (AUC ≈ 0.8).

Attenuated total reflectance Fourier-transform infrared spectroscopy, particularly using peak 3,280 (Amide A) and the 1,080/1,632 ratio (nucleocytoplasmic index), offers a promising approach for discriminating between normal breast tissue and carcinoma, as well as differentiating pre-IDC from metaplastic carcinoma. These spectral markers demonstrate both statistical significance and diagnostic potential.

Primary hepatocellular carcinoma (HCC) remains one of the leading causes of cancer-related mortality worldwide, with curative options still limited for patients with advanced disease. As an emerging modality of cancer immunotherapy, tumor vaccines represent a promising approach that activates the host immune system to recognize and eliminate malignant cells. Multiple vaccine platforms, including peptide vaccines, dendritic-cell vaccines, nucleic-acid vaccines, and viral-vector vaccines, have been explored for HCC. Among these, peptide- and dendritic-cell-based vaccines are supported by the most extensive clinical data, demonstrating favorable safety and immunogenicity profiles. The advent of personalized therapeutic cancer vaccines based on tumor-specific antigens has further refined the precision of vaccine design. Nevertheless, several major challenges persist, including immune suppression within the tumor immune microenvironment, marked tumor heterogeneity, immune-escape mechanisms, and limited vaccine immunogenicity, all of which hinder clinical efficacy. In addition, issues related to standardization, large-scale production, and regulatory oversight remain unresolved. Recent advances in sequencing technology, nanotechnology, and artificial intelligence have opened new avenues for optimizing vaccine platforms and delivery strategies. Combination therapies that integrate cancer vaccines with immune checkpoint inhibitors, chemotherapy, or locoregional treatments are also being actively investigated to improve patient outcomes. In summary, although vaccine-based immunotherapy for HCC is still at an early stage, its integration with personalized medicine and multimodal therapeutic strategies holds great potential for improving the long-term prognosis of patients with HCC. Therefore, this review aims to systematically summarize current advances in tumor vaccine–based immunotherapy for hepatocellular carcinoma, with a particular focus on vaccine platforms, target antigens, clinical trial outcomes, and future challenges for clinical translation.

Breast cancer (BCA) is one of the most common cancers worldwide, with a high rate of incidence and mortality. This review provides global information on BCA therapy using curcumin. Chemotherapy, as an effective treatment for different stages of BCA, and curcumin, generally regarded as safe compound and an alternative to synthetic drugs, have been described for the treatment of BCA. A few parameters, including nano-curcumin versus bulk curcumin and its encapsulated form versus its corresponding free form, have been discussed. Curcumin, a safe and edible compound with antitumor properties, is a promising medicinal compound for the treatment of BCA. Encapsulation of curcumin enhances its stability and anticancer efficiency. Nano-curcumin exhibits superior properties when compared to its bulk counterparts, leading to notable interactions and effects.

Acromegaly requires multimodal management. While surgery is first-line, many patients have persistent/recurrent disease. Gamma knife radiosurgery (GKRS) offers precise radiation, but data on its use as initial therapy remain limited. This study aimed to review the outcomes and report on our experience in treating patients with acromegaly using initial GKRS.

We retrospectively identified 33 patients with acromegaly who underwent GKRS from 1993 until 2016 at the Department of Radiotherapy, the Second Affiliated Hospital of Guangzhou Medical University. These patients had complete endocrine, radiological, and imaging data before and after GKRS. Furthermore, univariate and multivariate analysis was utilized to analyze the potential prognostic factors of endocrine remission and new-onset hypopituitarism.

Thirty-three patients were enrolled in the study. Fifteen patients (45.5%) were males and 18 (54.5%) were females. The median age was 44.0 years (range, 24.9–66.2 years). During a median follow-up of 65.6 months (range, 12.9–297.6), the median margin dose for GKRS was 15.0 Gy (range, 10.8–20.3 Gy). Endocrine remission was achieved in nine of the 33 patients (27.3%) over a mean follow-up of 85.1 months (range, 12.9–161.3). No prognostic factors demonstrated a significant association with endocrine remission. New-onset hypopituitarism occurred in eight patients (24.2%) after GKRS. The tumor control rate was 100%. Only one patient developed worsening visual dysfunction. No new cranial neuropathy was noted.

Initial GKRS for acromegaly provided effective tumor control and partial endocrine remission with a favorable safety profile, notably a low rate of new-onset hypopituitarism, representing a viable treatment option.

Hepatic metastasis (HM) and lymph node metastasis in pancreatic ductal adenocarcinoma (PDAC) are associated with worse overall survival, largely due to the immunosuppressive microenvironment. However, the key immunosuppressive cells within this microenvironment remain inadequately defined. This study aimed to identify the cells contributing to HM and lymph node metastasis in PDAC and to investigate their regulatory mechanisms.

Single-cell RNA sequencing was used to profile the tumor microenvironment in HM, lymph node-negative, and lymph node-positive (LNP) PDAC tissues. Bioinformatic analyses revealed subtypes of immunosuppressive myeloid-derived suppressor cells (MDSCs). Immunofluorescence and flow cytometry were performed to detect the distribution and proportion of interleukin-1 receptor antagonist (IL1RA+) MDSCs. The immunosuppressive and pro-tumorigenic functions of IL1RA+ MDSCs were analyzed using enzyme-linked immunosorbent assay, quantitative reverse transcription polymerase chain reaction, Western blotting, and Transwell assays. Patient-derived xenograft mouse models were employed to validate the role of IL1RA+ MDSCs in vivo.

Polymorphonuclear-MDSCs were found to be recruited to metastatic PDAC tissues. Among these, IL1RA+ MDSCs were enriched in HM/LNP tissues and correlated with poorer prognosis. IL1RA+ MDSCs promoted M2 macrophage polarization and suppressed the activity of natural killer cells and cytotoxic T cells. Furthermore, IL1RA+ MDSCs accelerated PDAC migration and progression by upregulating epithelial–mesenchymal transition-related proteins in both in vitro and in vivo models.

IL1RA+ MDSCs represent a key immunosuppressive and pro-tumorigenic subtype in HM/LNP PDAC, providing a solid theoretical basis for prognostic prediction and the development of immunotherapeutic strategies targeting these cells in HM/LNP PDAC.

Chronic hepatitis B (CHB) poses a major global health burden, with China particularly affected. Effective antiviral therapy is crucial to prevent disease progression, but responses may vary by Hepatitis B virus (HBV) genotype. This prospective study aimed to compare genotype-specific responses to 144-week entecavir (ETV) therapy in HBeAg-positive CHB patients, with particular emphasis on histological improvement assessed through paired liver biopsies.

We enrolled 49 treatment-naïve CHB patients (HBV DNA ≥ 20,000 IU/mL, alanine transaminase (ALT) > 2× ULN, and Scheuer system G ≥ 2) who received ETV 0.5 mg/day. HBV genotyping was performed using Polymerase Chain Reaction and fragment length analysis. The primary endpoint was histological improvement (i.e., ≥ 2-grade reduction in necroinflammatory activity without fibrosis progression), evaluated via paired biopsies (baseline and week 144) by blinded pathologists. Secondary endpoints included virological response (i.e., serum HBV DNA < 100 IU/mL), HBeAg seroconversion, and ALT normalization.

The cohort included 24 genotype B and 24 genotype C patients (one genotype A patient was excluded from genotype-specific analyses). Genotype B showed significantly higher histological improvement rates (91.3% vs. 63.2%, P = 0.027) and greater inflammation resolution (0 ≤ G < 1: 56.5% vs. 26.3%, P = 0.048). Virological suppression was excellent in both groups (100% vs. 100%). HBeAg seroconversion trended higher in genotype C (29.2% vs. 50.0%, P = 0.140). All patients achieved ALT normalization by week 48, with no safety concerns.

HBV genotype B demonstrates superior histological responses to ETV therapy compared with genotype C, supporting the clinical value of HBV genotyping for personalized CHB management. These findings highlight the importance of considering viral genotype when evaluating treatment outcomes.