Background&Aims: Automated classification of hepatic tumours on liver biopsies remains challenging due to morphological overlaps. We developed deep learning models to discriminate hepatocellular carcinoma (HCC), cholangiocarcinoma (CCA), and metastatic lesions on digitalized whole slide images (WSI).Methods: We retrospectively analysed 210 liver biopsy samples from 204 patients (2020-2025) at the University of Palermo, including HCC (n=84), intrahepatic CCA (n=34), hepatocholangiocarcinoma (n=4), liver metastases (n=65), and non-neoplastic samples (n=23). WSI were digitized in NDPI format. Initially, we applied unsupervised learning: WSI were segmented into 512 × 512px patches, deep features extracted using pre-trained ResNet-50, dimensionality reduction performed with UMAP, and clustering executed with HDBSCAN followed by agglomerative merging. Subsequently, we developed a semi-supervised Mean Teacher model using 30% expert-annotated WSI and 70% unlabeled data with consistency regularization.Results: Unsupervised clustering achieved low purity: Cluster 0 (42% HCC, 11% CCA, 47% negative), Cluster 1 (35% HCC, 16% CCA, 48% negative), demonstrating inadequate class discrimination. The semi-supervised model achieved markedly superior performance: 97.85% training accuracy and 89.24% validation accuracy. Starting from limited annotations (30%), the model effectively generated high-quality pseudo-labels for the unlabeled majority, learning robust features despite class imbalance and successfully discriminating between hepatic tumour types.Conclusions: Semi-supervised learning significantly outperformed unsupervised approaches for hepatic tumour classification. This model shows promise as a first-line screening tool in liver biopsies to optimize diagnostic workflow and guide immunohistochemical panel selection, serving as an assistant for junior pathologists and quality assurance instrument in low-volume centres.

Background & Aims: Biliary abnormalities in autoimmune hepatitis (AIH) and interface hepatitis in primary biliary cholangitis (PBC) occur frequently, and misinterpretation may lead to therapeutic mistakes with a negative impact on patients. This study investigates the use of a deep learning (DL)-based pipeline for the diagnosis of AIH and PBC to aid differential diagnosis. Methods: We conducted a multicenter study across six European referral centers, and built a library of digitized liver biopsy slides dating from 1997 to 2023. A training set of 354 cases (266 AIH and 102 PBC) and an external validation set of 92 cases (62 AIH and 30 PBC) were available for analysis. A novel DL model, the autoimmune liver neural estimator (ALNE), was trained on whole- slide images (WSIs) with H&E staining, without human annotations. The ALNE model was evaluated against clinico-pathological diagnoses and tested for interobserver variability among general pathologists. Results: The ALNE model demonstrated high accuracy in differentiating AIH from PBC, achieving an area under the receiver operating characteristic curve of 0.81 in external validation. Attention heatmaps showed that ALNE tends to focus more on areas with increased inflammation, associating such patterns predominantly with AIH. A multivariate explainable ML model revealed that PBC cases misclassified as AIH more often had ALP values between 1 x upper limit of normal (ULN) and 2 x ULN, coupled with AST values above 1 x ULN. Inconsistency among general pathologists was noticed when evaluating a random sample of the same cases (Fleiss's kappa value 0.09). Conclusions: The ALNE model is the first system generating a quantitative and accurate differential diagnosis between cases with AIH or PBC. (c) 2024 The Author(s). Published by Elsevier B.V. on behalf of European Association for the Study of the Liver (EASL). This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/).

Artificial intelligence applications in liver pathology remain limited, with existing tools either narrowly focused or lacking external validation. This study introduces HOTSPoT, an open-source, validated transformer-based model for automated segmentation of portal tracts in H&E-stained liver biopsy whole slide images. A multi-institutional dataset of 223 cases was used, with annotations by expert hepatopathologists. HOTSPoT achieved high performance with mean Dice scores of 0.92 (train/val) and 0.91 (test), and mean IoUs of 0.86, 0.85, and 0.84, respectively, showing minimal domain shift. Automated portal tract quantification showed strong concordance with manual assessments (κ up to 0.90), and portal area correlated with fibrosis stage (r = 0.87, p < 0.001). The model is available as a TorchScript file with a modified WSInfer library, enabling efficient WSI-level inference and integration with QuPath for advanced pathology analysis.

Artificial intelligence (AI) transforms image data analysis across many biomedical fields, such as cell biology, radiology, pathology, cancer biology, and immunology, with object detection, image feature extraction, classification, and segmentation applications. Advancements in deep learning (DL) research have been a critical factor in advancing computer techniques for biomedical image analysis and data mining. A significant improvement in the accuracy of cell detection and segmentation algorithms has been achieved as a result of the emergence of open-source software and innovative deep neural network architectures. Automated cell segmentation now enables the extraction of quantifiable cellular and spatial features from microscope images of cells and tissues, providing critical insights into cellular organization in various diseases. This review aims to examine the latest AI and DL techniques for cell analysis and data mining in microscopy images, aid the biologists who have less background knowledge in AI and machine learning (ML), and incorporate the ML models into microscopy focus images.

Oral and oropharyngeal cancers, caused by persistent human papillomavirus (HPV), have recently increased. Diagnostic methods often fail to assess precancerous lesion risk, delaying oral cancer diagnosis. New molecular biomarkers, particularly DNA methylation, are sought to better stratify patients' risk. The PreCursor-M+ (Fujirebio, Tokyo, Japan), which analyze hypermethylation of the two onco-suppressor FAM19A4 and miR124-2 in cervical samples from high-risk HPV-positive women, was used to assess the methylation level of 111 oral samples distinguished in oral squamous cell carcinomas (OSCC), oral potentially malignant disorders (OPMD) benign lesions (BL), and no lesions (NL). HPV was detected by INNO-LiPA HPV Genotyping Extra II (Fujirebio, Tokyo, Japan). Hypermethylation was correlated with the severity of the diagnosis. A positive result was more common in OSCC (p < 0.0001). HPV positivity correlated with hypermethylation in OSCCs (32.4%, p = 0.0006), although statistical significance was also found in HPV-negatives (p = 0.0007). HPV16-positive OSCC showed higher methylation. Targets' methylation increased from the NL to the BL, OPMD and OSCCs groups. The methylation status of FAM19A4 and miR-124-2 may play an important role in the progression of oral cancer and, consequently, in determining the prognosis of patients with OPMD, for whom hypermethylation would suggest the need for close monitoring. Furthermore, HPV16's association with hypermethylation suggests its involvement in oral carcinogenesis. To confirm these results and gain further insight into HPV's role in methylation impairment, the sample size will be increased.

OBJECTIVE:Hyperferritinaemia is associated with liver fibrosis severity in patients with metabolic dysfunction-associated steatotic liver disease (MASLD), but the longitudinal implications have not been thoroughly investigated. We assessed the role of serum ferritin in predicting long-term outcomes or death. DESIGN:We evaluated the relationship between baseline serum ferritin and longitudinal events in a multicentre cohort of 1342 patients. Four survival models considering ferritin with confounders or non-invasive scoring systems were applied with repeated five-fold cross-validation schema. Prediction performance was evaluated in terms of Harrell's C-index and its improvement by including ferritin as a covariate. RESULTS:Median follow-up time was 96 months. Liver-related events occurred in 7.7%, hepatocellular carcinoma in 1.9%, cardiovascular events in 10.9%, extrahepatic cancers in 8.3% and all-cause mortality in 5.8%. Hyperferritinaemia was associated with a 50% increased risk of liver-related events and 27% of all-cause mortality. A stepwise increase in baseline ferritin thresholds was associated with a statistical increase in C-index, ranging between 0.02 (lasso-penalised Cox regression) and 0.03 (ridge-penalised Cox regression); the risk of developing liver-related events mainly increased from threshold 215.5 µg/L (median HR=1.71 and C-index=0.71) and the risk of overall mortality from threshold 272 µg/L (median HR=1.49 and C-index=0.70). The inclusion of serum ferritin thresholds (215.5 µg/L and 272 µg/L) in predictive models increased the performance of Fibrosis-4 and Non-Alcoholic Fatty Liver Disease Fibrosis Score in the longitudinal risk assessment of liver-related events (C-indices>0.71) and overall mortality (C-indices>0.65). CONCLUSIONS:This study supports the potential use of serum ferritin values for predicting the long-term prognosis of patients with MASLD.

Ultrasound (US) surveillance is a cornerstone for early diagnosis of HCC, anyway US presentation has undergone significant changes. With the aim of evaluating the effects of US surveillance program in the real-world clinical practice, we wanted to evaluate US presentation of HCCs over the last 30 years and the differences of HCCs presentation according to etiology. 174 patients diagnosed between 1993 and 98 (G1), 96 between 2003 and 08 (G2), 102 between 2013 and 18 (G3), were compared. US patterns were: single, multiple or diffuse nodules. The echo-patterns: iso-, hypo-, hyper-echoic, or mixed. In G1, the HCC diagnosis was mainly histologic; in G2 by EASL 2001 and AASLD 2005, in G3 AASLD 2011, EASL 2012, and AISF 2013 guidelines. HCV was the most frequent etiology, dropping between G1 (81

Background. The prevalence of oral human papillomavirus (HPV) in the healthy population and patients with oral diseases such as oral squamous cell carcinoma (OSCC), oral potentially malignant disorders (OPMDs), and oral benign lesions (BL), is not consistently described in the literature, with scarce and often heterogeneous data. In addition, the efficacy of HPV prophylactic vaccines in preventing HPV-related oral disorders has been scarcely investigated. Methods. The prevalence of HPV and the potential impact of vaccines were analyzed in 1,415 oral rinse specimens, collected over 10 years and grouped into four categories based on histological/clinical diagnosis. Results. HPV prevalence in OSCC, OPMD, and BL patients and in healthy individuals potentially exposed to HPV (HPE) was comparable (12.7 vs. 27.2% vs. 13.5 vs. 9%). Statistical analysis of the vaccine impact involved calculating high and low estimates and showed a significant difference only for the low effect. The nonavalent vaccine had higher low estimates than the bivalent vaccine in OSCC and HPE patients (29.6 vs. 51.9%, p < 0.05; 18.2 vs. 42.4%, p < 0.05), while for OPMD and BL, the frequency of bivalent low estimates was lower than that of quadrivalent and nonavalent (48.6 vs. 68.6%, p < 0.05 and 48.6 vs. 77.1%, p < 0.05; 23.9 vs. 50.7%, p < 0.05, and 23.9 vs. 63.4%, p < 0.05). Conclusions. This study provided new insights into the prevalence of oral HPV and showed that the nonavalent vaccine may provide better protection than the other vaccines in the presence of an OSCC diagnosis. Conversely, the quadrivalent vaccine may be sufficient to prevent OPMD and BL.

In the era of digitalization, the biomedical sector has been affected by the spread of artificial intelligence. In recent years, the possibility of using deep and machine learning methods for clinical diagnostic and therapeutic interventions has been emerging as an essential resource for biomedical imaging. Digital pathology represents innovation in a clinical world that looks for faster and better-performing diagnostic methods, without losing the accuracy of current human-guided analyses. Indeed, artificial intelligence has played a key role in a wide variety of applications that require the analysis of a massive amount of data, including segmentation processes in medical imaging. In this context, artificial intelligence enables the improvement of image segmentation methods, moving towards the development of fully automated systems of analysis able to support pathologists in decision-making procedures. The aim of this review is to aid biologists and clinicians in discovering the most common segmentation open-source tools, including ImageJ (v. 1.54), CellProfiler (v. 4.2.5), Ilastik (v. 1.3.3) and QuPath (v. 0.4.3), along with their customized implementations. Additionally, the tools’ role in the histological imaging field is explored further, suggesting potential application workflows. In conclusion, this review encompasses an examination of the most commonly segmented tissues and their analysis through open-source deep and machine learning tools.

The aim of the present study consists of the evaluation of the biodistribution of a novel 68Ga-labeled radiopharmaceutical, [68Ga]Ga-NODAGA-Z360, injected into Balb/c nude mice through histopathological analysis on bioptic samples and radiomics analysis of positron emission tomography/computed tomography (PET/CT) images. The 68Ga-labeled radiopharmaceutical was designed to specifically bind to the cholecystokinin receptor (CCK2R). This receptor, naturally present in healthy tissues such as the stomach, is a biomarker for numerous tumors when overexpressed. In this experiment, Balb/c nude mice were xenografted with a human epidermoid carcinoma A431 cell line (A431 WT) and overexpressing CCK2R (A431 CCK2R+), while controls received a wild-type cell line. PET images were processed, segmented after atlas-based co-registration and, consequently, 112 radiomics features were extracted for each investigated organ / tissue. To confirm the histopathology at the tissue level and correlate it with the degree of PET uptake, the studies were supported by digital pathology. As a result of the analyses, the differences in radiomics features in different body districts confirmed the correct targeting of the radiopharmaceutical. In preclinical imaging, the methodology confirms the importance of a decision-support system based on artificial intelligence algorithms for the assessment of radiopharmaceutical biodistribution.

In the submitted and accepted version of this paper, five author names were omitted in error. These authors were: Ville Männistö, Jussi Pihlajamäki, Sami Qadri, Hannele Yki-Järvinen, Stefano Romeo. The correct author and affilitation list is provided in this corrigendum. We apologise for any inconvenience caused.

In cervical biopsies, for diagnosis of Human Papilloma Virus (HPV) related conditions, the immunohistochemical staining for p16 has a diagnostic value only if diffusely and strongly positive, pattern named “block-like”. “Weak and/or focal (w/f) p16 expression” is commonly considered nonspecific. In our previous study, we demonstrated the presence of high-risk HPV (hrHPV) DNA by LiPa method in biopsies showing w/f p16 positivity. The aim of the present study was to investigate the presence of hrHPV-DNA by CISH in the areas showing w/f p16 expression. We assessed the presence of hrHPV16, 18, 31, 33, 51 by CISH in a group of 20 cervical biopsies showing w/f p16 expression, some with increased Ki67, and in 10 cases of block-like expression, employed as control. The immunohistochemical p16 expression was also assessed by digital pathology. hrHPV-CISH nuclear positivity was encountered in 12/20 cases of w/f p16 expression (60%). Different patterns of nuclear positivity were identified, classified as punctate, diffuse and mixed, with different epithelial distributions. Our results, albeit in a limited casuistry, show the presence of HPV in an integrated status highlighted by CISH in w/f p16 positive cases. This could suggest the necessity of a careful follow-up of the patients with “weak” and/or “focal” immunohistochemical patterns of p16, mainly in cases of increased Ki67 cell proliferation index, supplemented with molecular biology examinations.

To assess the performance and the reproducibility of ultrasound-guided attenuation parameter (UGAP) and two-dimensional shear wave elastography (2D-SWE) in patients with biopsy-proven metabolic dysfunction-associated steatotic liver disease (MASLD). This study included consecutive adult patients with MASLD who underwent ultrasound with UGAP, 2D-SWE and percutaneous liver biopsy. The median values of 12 consecutive UGAP measurements were acquired by two independent radiologists (R1 and R2). Hepatic steatosis was graded by liver biopsy as: (0) < 5

(1) Background: Nonalcoholic Steatohepatitis/Nonalcoholic Fatty Liver Disease (NASH/NAFLD) is the most recurrent chronic liver disease. NASH could present with a cholestatic (C) or hepatic (H) pattern of damage. Recently, we observed that increased Epithelial Cell Adhesion Molecule (EpCAM) expression was the main immunohistochemical feature to distinguish C from H pattern in NASH. (2) Methods: In the present study, we used digital pathology to compare the quantitative results of digital image analysis by QuPath software (Q-results), with the semi-quantitative results of observer assessment (S-results) for cytokeratin 7 and 19, (CK7, CK19) as well as EpCAM expression. Patients were classified into H or C group on the basis of the ratio between alanine transaminase (ALT) and alkaline phosphatase (ALP) values, using the “R-ratio formula”. (3) Results: Q- and S-results showed a significant correlation for all markers (p < 0.05). Q-EpCAM expression was significantly higher in the C group than in the H group (p < 0.05). Importantly ALP, an indicator of hepatobiliary disorder, was the only biochemical parameter significantly correlated with Q-EpCAM. Instead, Q-CK7, but not Q-CK19, correlated only with γGlutamyl-Transferase (γGT). Of note, Stage 4 fibrosis correlated with Q-EpCAM, Q-CK19, and ALP but not with γGT or ALT. Conclusions: Image analysis confirms the relation between cholestatic-like pattern, associated with a worse prognosis, with increased ALP values, EpCAM positive biliary metaplasia, and advanced fibrosis. These preliminary data could be useful for the implementation of AI algorithms for the assessment of cholestatic NASH.

Hurthle cell (HC), anaplastic (AC), and medullary (MC) carcinomas are low frequency thyroid tumors that pose several challenges for physicians and pathologists due to the scarcity of cases, information, and histopathological images, especially in the many areas around the world in which sophisticated molecular and genetic diagnostic facilities are unavailable. It is, therefore, cogent to provide tools for microscopists to achieve accurate diagnosis, such as histopathological images with reliable biomarkers, which can help them to reach a differential diagnosis. We are investigating whether components of the chaperone system (CS), such as the molecular chaperones, can be considered dependable biomarkers, whose levels and distribution inside and outside cells in the tumor tissue could present a distinctive histopathological pattern for each tumor type. Here, we report data on the chaperones Hsp27, Hsp60, and Hsp90. They presented quantitative levels and distribution patterns that were different for each tumor and differed from those of a benign thyroid pathology, goiter (BG). Therefore, the reported methodology can be beneficial when the microscopist must differentiate between HC, AC, MC, and BG.

Pembrolizumab has received approval as a first-line treatment for unresectable/metastatic triple-negative breast cancer (mTNBC) with a PD-L1 combined positive score (CPS) of ≥ 10. However, assessing CPS in mTNBC poses challenges. Firstly, it represents a novel analysis for breast pathologists. Secondly, the heterogeneity of PD-L1 expression in mTNBC further complicates the assessment. Lastly, the lack of standardized assays and staining platforms adds to the complexity. In KEYNOTE trials, PD-L1 expression was evaluated using the IHC 22C3 pharmDx kit as a companion diagnostic test. However, both the 22C3 pharmDx and VENTANA PD-L1 (SP263) assays are validated for CPS assessment. Consequently, assay-platform choice, staining conditions, and scoring methods can significantly impact the testing outcomes. This consensus paper aims to discuss the intricacies of PD-L1 CPS testing in mTNBC and provide practical recommendations for pathologists. Additionally, we present findings from a nationwide Italian survey elucidating the state-of-the-art in PD-L1 CPS testing in mTNBC.

(1) Background: Previous studies have reported a correlation between serum anti-Thyroglobulin-antibodies (TgAb) and papillary thyroid carcinoma. The aim of our study was to evaluate whether serum TgAb and anti-thyroid-peroxidase antibody (TPO) positivity was also related to pre-neoplastic histological changes such as papillary-like nuclear features (PLNF) and with the presence of lymphocytic infiltrate (LI) in thyroid surgical specimens. (2) Methods: The study was retrospectively carried out on 70 consecutively recruited patients who underwent thyroidectomy for benign process and whose TgAb and TPOAb values were retrieved from clinical records. Histological sections of thyroid surgical samples were revised, looking for PLNF and lymphocytic infiltrate. HBME1 expression was assessed by immunohistochemistry. (3) Results: Our results showed a significant association between TgAb, PLNF, and lymphocytic infiltrate. The presence of TgAb was highly specific, but less sensitive, in predicting the presence of PLNF (sensitivity = 0.6, specificity = 0.9; positive predictive value (PPV) = 0.88; negative predictive value (NPV) = 0.63). TgAb positivity showed a good association with the presence of lymphocytic infiltrate (sensitivity = 0.62, specificity = 0.9; PPV = 0.88 and NPV = 0.68). HBME1 immunoreactivity was observed in the colloid of follicles showing PLNF and/or closely associated with LI. (4) Conclusions: The presence of PLNF and LI is associated with serum TgAb positivity. The presence of TgAb and of LI could be triggered by an altered thyroglobulin contained in the HBME1-positive colloid, and could be a first defense mechanism against PLNF that probably represent early dysplastic changes in thyrocytes.