Search Results (1,026)

Background: Pancreatic cystic lesions (PCLs), including intraductal papillary mucinous neoplasms (IPMNs) and mucinous cystic neoplasms (MCNs), pose a diagnostic challenge due to their variable malignant potential. Current guidelines, such as Fukuoka and American Gastroenterological Association (AGA), have moderate predictive accuracy and may lead to overtreatment or missed malignancies. Artificial intelligence (AI), incorporating machine learning (ML) and deep learning (DL), offers the potential to improve risk stratification, diagnosis, and management of PCLs by integrating clinical, radiological, and molecular data. This is the first systematic review to evaluate the application, performance, and clinical utility of AI models in the diagnosis, classification, prognosis, and management of pancreatic cysts. Methods: A systematic review was conducted in accordance with PRISMA guidelines and registered on PROSPERO (CRD420251008593). Databases searched included PubMed, EMBASE, Scopus, and Cochrane Library up to March 2025. The inclusion criteria encompassed original studies employing AI, ML, or DL in human subjects with pancreatic cysts, evaluating diagnostic, classification, or prognostic outcomes. Data were extracted on the study design, imaging modality, model type, sample size, performance metrics (accuracy, sensitivity, specificity, and area under the curve (AUC)), and validation methods. Study quality and bias were assessed using the PROBAST and adherence to TRIPOD reporting guidelines. Results: From 847 records, 31 studies met the inclusion criteria. Most were retrospective observational (n = 27, 87%) and focused on preoperative diagnostic applications (n = 30, 97%), with only one addressing prognosis. Imaging modalities included Computed Tomography (CT) (48%), endoscopic ultrasound (EUS) (26%), and Magnetic Resonance Imaging (MRI) (9.7%). Neural networks, particularly convolutional neural networks (CNNs), were the most common AI models (n = 16), followed by logistic regression (n = 4) and support vector machines (n = 3). The median reported AUC across studies was 0.912, with 55% of models achieving AUC ≥ 0.80. The models outperformed clinicians or existing guidelines in 11 studies. IPMN stratification and subtype classification were common focuses, with CNN-based EUS models achieving accuracies of up to 99.6%. Only 10 studies (32%) performed external validation. The risk of bias was high in 93.5% of studies, and TRIPOD adherence averaged 48%. Conclusions: AI demonstrates strong potential in improving the diagnosis and risk stratification of pancreatic cysts, with several models outperforming current clinical guidelines and human readers. However, widespread clinical adoption is hindered by high risk of bias, lack of external validation, and limited interpretability of complex models. Future work should prioritise multicentre prospective studies, standardised model reporting, and development of interpretable, externally validated tools to support clinical integration. Full article

Background: Angiosarcomas (ASs) represent a heterogeneous and highly aggressive subset of tumors that respond poorly to systemic treatments and are associated with short progression-free survival (PFS) and overall survival (OS). The aim of this study was to develop and validate an immune-related prognostic model—termed the AS score—using data from two independent sarcoma cohorts. Methods: A prognostic model was developed using a previously characterized cohort of 25 angiosarcoma samples. Candidate genes were identified via the Maxstat algorithm (Maxstat v0.7-25 for R), combined with log-rank testing. The AS score was then computed by weighing normalized gene expression levels according to Cox regression coefficients. For external validation, transcriptomic data from TCGA Sarcoma cohort (n = 253) were analyzed. The Immunoscore—which reflects the tumor immune microenvironment—was inferred using the ESTIMATE package (v1.0.13) in R. All statistical analyses were performed in RStudio (v 4.0.3). Results: Four genes—IGF1R, MAP2K1, SERPINE1, and TCF12—were ultimately selected to construct the prognostic model. The resulting AS score enabled the classification of angiosarcoma cases into two prognostically distinct groups (p = 0.00012). Cases with high AS score values, which included both cutaneous and non-cutaneous forms, exhibited significantly poorer outcomes, whereas cases with low AS scores were predominantly cutaneous. A significant association was observed between the AS score and the Immunoscore (p = 0.025), with higher Immunoscore values found in high-AS score tumors. Validation using TCGA sarcoma cohort confirmed the prognostic value of both the AS score (p = 0.0066) and the Immunoscore (p = 0.0029), with a strong correlation between their continuous values (p = 2.9 × 10⁻⁸). Further survival analysis, integrating categorized scores into four groups, demonstrated robust prognostic significance (p = 0.00021). Notably, in tumors with a low Immunoscore, AS score stratification was not prognostic. In contrast, among cases with a high Immunoscore, the AS score effectively distinguished outcomes (p < 0.0001), identifying a subgroup with poor prognosis but potential sensitivity to immunotherapy. Conclusions: This combined classification using the AS score and Immunoscore has prognostic relevance in sarcoma, suggesting that angiosarcomas with an immunologically active microenvironment (high Immunoscore) and poor prognosis (high AS score) may be prime candidates for immunotherapy and this approach warrants prospective validation. Full article

Arrhythmogenic cardiomyopathies (ACMs) are a phenotypically and etiologically heterogeneous group of myocardial disorders characterized by fibrotic or fibro-fatty replacement of ventricular myocardium, electrical instability, and an elevated risk of sudden cardiac death. Initially identified as a right ventricular disease, ACMs are now recognized to include biventricular and left-dominant forms. Genetic causes account for a substantial proportion of cases and include desmosomal variants, non-desmosomal variants, and familial gene-elusive forms with no identifiable pathogenic mutation. Nongenetic etiologies, including post-inflammatory, autoimmune, and infiltrative mechanisms, may mimic the phenotype. In many patients, the disease remains idiopathic despite comprehensive evaluation. Cardiac magnetic resonance imaging has emerged as a key tool for identifying non-ischemic scar patterns and for distinguishing arrhythmogenic phenotypes from other cardiomyopathies. Emerging classifications propose the unifying concept of scarring cardiomyopathies based on shared structural substrates, although global consensus is evolving. Risk stratification remains challenging, particularly in patients without overt systolic dysfunction or identifiable genetic markers. Advances in tissue phenotyping, multi-omics, and artificial intelligence hold promise for improved prognostic assessment and individualized therapy. Full article

Background/Objectives. This manuscript presents an overview of advances in oncological radiotherapy as an effective treatment method for cancerous tumors, focusing on mechanisms of action within metabolite–antimetabolite systems. The urgency of this topic is underscored by the fact that cancer remains one of the leading causes of death worldwide: as of 2022, approximately 20 million new cases were diagnosed globally, accounting for about 0.25% of the total population. Given prognostic models predicting a steady increase in cancer incidence to 35 million cases by 2050, there is an urgent need for the latest developments in physics, chemistry, molecular biology, pharmacy, and strict adherence to oncological vigilance. The purpose of this work is to demonstrate the relationship between the nature and mechanisms of past diagnostic and therapeutic oncology approaches, their current improvements, and future prospects. Particular emphasis is placed on isotope technologies in the production of therapeutic nuclides, focusing on the mechanisms of formation of simple and complex theranostic compounds and their classification according to target specificity. Methods. The methodology involved searching, selecting, and analyzing information from PubMed, Scopus, and Web of Science databases, as well as from available official online sources over the past 20 years. The search was structured around the structure–mechanism–effect relationship of active pharmaceutical ingredients (APIs). The manuscript, including graphic materials, was prepared using a narrative synthesis method. Results. The results present a sequential analysis of materials related to isotope technology, particularly nucleus stability and instability. An explanation of theranostic principles enabled a detailed description of the action mechanisms of radiopharmaceuticals on various receptors within the metabolite–antimetabolite system using specific drug models. Attention is also given to radioactive nanotheranostics, exemplified by the mechanisms of action of radioactive nanoparticles such as Tc-99m, AuNPs, wwAgNPs, FeNPs, and others. Conclusions. Radiotheranostics, which combines the diagnostic properties of unstable nuclei with therapeutic effects, serves as an effective adjunctive and/or independent method for treating cancer patients. Despite the emergence of resistance to both chemotherapy and radiotherapy, existing nuclide resources provide protection against subsequent tumor metastasis. However, given the unfavorable cancer incidence prognosis over the next 25 years, the development of “preventive” drugs is recommended. Progress in this area will be facilitated by modern medical knowledge and a deeper understanding of ligand–receptor interactions to trigger apoptosis in rapidly proliferating cells. Full article

Objectives: This multicenter study aimed to clarify the recurrence patterns; treatment outcomes; and prognostic factors of thymic carcinoma, a rare cancer. Methods: We analyzed 101 patients with thymic carcinoma who underwent multidisciplinary treatment, including radiotherapy. The median age was 62 years, with 27 patients in stage I–II; 44 in stage III; and 30 in stage IV by the TNM classification. Seventy-two patients underwent surgery with radiotherapy; and 29 patients underwent definitive radiotherapy. Image-guided radiotherapy (IGRT) and elective nodal irradiation (ENI) were used for 35 and 23 patients, respectively. Local recurrence-free survival (LRFS); progression-free survival (PFS); and overall survival (OS) were calculated, and univariate and multivariate analyses were performed. Results: With a median follow-up of 68 months, we observed 17 local recurrences; 27 regional recurrences; and 35 distant metastases. The 5-year LRFS; PFS; and OS were 82%, 41%, and 76%, respectively. Multivariate analysis revealed that stage was the only factor associated with LRFS; PFS; and OS (p = 0.040; p < 0.0001; and p = 0.048, respectively), while treatment modality was associated with only LRFS (p = 0.015). IGRT and ENI were also associated with LRFS (p = 0.002 and 0.013, respectively). PFS and OS of stage IV patients were comparable between the surgery with radiotherapy and definitive radiotherapy groups (p = 0.99 and 0.98, respectively). Conclusions: Our results suggest the importance of stage-specific treatment strategies rather than resectability, especially for stage IV patients. These results should be validated in a prospective study. Our results also suggest that radiotherapy methods influence recurrence Full article

Prognostic markers such as overall survival (OS) and tertiary lymphoid structure (TLS) ratios, alongside diagnostic signatures like primary cancer-type classification, provide critical information for treatment selection, risk stratification, and longitudinal care planning across the oncology continuum. However, extracting these signals solely from sparse, high-dimensional multi-omics data remains a major challenge due to heterogeneity and frequent missingness in patient profiles. To address this challenge, we present SeNMo, a self-normalizing deep neural network trained on five heterogeneous omics layers—gene expression, DNA methylation, miRNA abundance, somatic mutations, and protein expression—along with the clinical variables, that learns a unified representation robust to missing modalities. Trained on more than 10,000 patient profiles across 32 tumor types from The Cancer Genome Atlas (TCGA), SeNMo provides a baseline that can be readily fine-tuned for diverse downstream tasks. On a held-out TCGA test set, the model achieved a concordance index of 0.758 for OS prediction, while external evaluation yielded 0.73 on the CPTAC lung squamous cell carcinoma cohort and 0.66 on an independent 108-patient Moffitt Cancer Center cohort. Furthermore, on Moffitt’s cohort, baseline SeNMo fine-tuned for TLS ratio prediction aligned with expert annotations (p < 0.05) and sharply separated high- versus low-TLS groups, reflecting distinct survival outcomes. Without altering the backbone, a single linear head classified primary cancer type with 99.8% accuracy across the 33 classes. By unifying diagnostic and prognostic predictions in a modality-robust architecture, SeNMo demonstrated strong performance across multiple clinically relevant tasks, including survival estimation, cancer classification, and TLS ratio prediction, highlighting its translational potential for multi-omics oncology applications. Full article

Background: Renal trauma accounts for approximately 3–5% of all trauma cases, predominantly affecting young males. The most common etiology is blunt trauma, particularly due to road traffic accidents, and it frequently occurs as part of polytrauma involving multiple organ systems. Management strategies are primarily dictated by hemodynamic stability, overall clinical condition, comorbidities, and injury severity graded according to the AAST classification. This study aimed to evaluate the effectiveness of non-operative management (NOM) in high-grade renal trauma (AAST grades III–V), beyond its established role in low-grade injuries (grades I–II). Secondary endpoints included the identification of independent prognostic factors for NOM failure and in-hospital mortality. Methods: We conducted a retrospective observational study including patients diagnosed with blunt renal trauma who presented to the Emergency Department of Policlinico Umberto I in Rome between 1 January 2013 and 30 April 2024. Collected data comprised demographics, trauma mechanism, vital signs, hemodynamic status (shock index), laboratory tests, blood gas analysis, hematuria, number of transfused RBC units in the first 24 h, AAST renal injury grade, ISS, associated injuries, treatment approach, hospital length of stay, and mortality. Statistical analyses, including multivariable logistic regression, were performed using SPSS v28.0. Results: A total of 244 patients were included. Low-grade injuries (AAST I–II) accounted for 43% (n = 105), while high-grade injuries (AAST III–V) represented 57% (n = 139). All patients with low-grade injuries were managed non-operatively. Among high-grade injuries, 124 patients (89%) were treated with NOM, including observation, angiography ± angioembolization, stenting, or nephrostomy. Only 15 patients (11%) required nephrectomy, primarily due to persistent hemodynamic instability. The overall mortality rate was 13.5% (33 patients) and was more closely associated with the overall injury burden than with renal injury severity. Multivariable analysis identified shock index and active bleeding on CT as independent predictors of NOM failure, whereas ISS and age were significant predictors of in-hospital mortality. Notably, AAST grade did not independently predict either outcome. Conclusions: In line with the current international literature, our study confirms that NOM is the treatment of choice not only for low-grade renal injuries but also for carefully selected hemodynamically stable patients with high-grade trauma. Our findings highlight the critical role of physiological parameters and overall ISS in guiding management decisions and underscore the need for individualized assessment to minimize unnecessary nephrectomies and optimize patient outcomes. Full article

Gastroenteropancreatic neuroendocrine neoplasms (GEP-NENs) are rare tumors with different clinical and biological characteristics. Ki-67 staining and mitotic counts are the most commonly used prognostic markers, but these methods are time-consuming and lack reproducibility, highlighting the need for innovative approaches that improve histological evaluation and prognosis. In our previous study, we observed that the microRNA (miRNA) expression profile of GEP-NENs correlates with the three grades of GEP-NENs. This study aimed to characterize a group of miRNAs that discriminate well-differentiated GEP-NENs grading 1 (G1) and grading (G2). Fifty formalin-fixed and paraffin-embedded tissue specimens from well-differentiated GEP-NENs G1 and G2 tissues were used for this study. The expression levels of 21 miRNAs were examined using qRT-PCR, while FGFR2 and FGF1 protein expression were evaluated through immunohistochemistry (IHC). We identified four miRNAs (hsa-miR-133, hsa-miR-150-5p, hsa-miR-143-3p and hsa-miR-378a-3p) that are downregulated in G2 GEP-NENs compared to G1. Bioinformatic analysis revealed that these miRNAs play a key role in modulating the FGF/FGFR signaling pathway. Consistent with this observation, we found that fibroblast growth factor receptor 2 (FGFR2) expression is markedly higher in G2 NENs patients, whereas its expression remains low in G1 NENs. Our findings highlight the potential use of miRNAs to confirm the histological evaluation of GEP-NENs by employing them as biomarkers for improving histological evaluation and tumor classification. Full article

Myelodysplastic syndromes/neoplasms (MDS) represent a biologically and clinically diverse group of myeloid malignancies marked by cytopenias, morphological dysplasia, and an inherent risk of progression to acute myeloid leukemia. Over the past two decades, the field has made significant advances in characterizing the molecular landscape of MDS, leading to refined classification systems to reflect the underlying genetic and biological diversity. In 2025, the treatment of MDS is increasingly individualized, guided by integrated clinical, cytogenetic, and molecular risk stratification tools. For lower-risk MDS, the treatment paradigm has evolved beyond erythropoiesis-stimulating agents (ESAs) with the introduction of novel effective agents such as luspatercept and imetelstat, as well as shortened schedules of hypomethylating agents (HMAs). For higher-risk disease, monotherapy with HMAs continue to be the standard of care as combination therapies of HMAs with novel agents have, to date, failed to redefine treatment paradigms. The recognition of precursor states like clonal hematopoiesis of indeterminate potential (CHIP) and the increasing use of molecular monitoring will hopefully enable earlier intervention/prevention strategies. This review provides a comprehensive overview of the current treatment approach for MDS, highlighting new classifications, prognostic tools, evolving therapeutic options, and ongoing challenges. We discuss evidence-based recommendations, treatment sequencing, and emerging clinical trials, with a focus on translating biological insights into improved outcomes for patients with MDS. Full article

Background: Pathological complete response (pCR) serves as a prognostic surrogate endpoint for long-term clinical outcomes in breast cancer patients receiving neoadjuvant systemic therapy (NAST). This study aims to develop and evaluate machine learning-based biomarkers for predicting pCR and recurrence-free survival (RFS). Methods: This retrospective monocentric study included 235 women (mean age 46 ± 11 years) with non-metastatic breast cancer treated with NAST. We developed various machine learning models using clinical features (age, genetic mutations, TNM stage, hormonal receptor expression, HER2 status, and histological grade), along with morphological features (size, T2 signal, and surrounding edema) and radiomics data extracted from pre-treatment MRI. Patients were divided into training and test groups with different MRI models. A customized machine learning pipeline was implemented to handle these diverse data types, consisting of feature selection and classification components. Results: The models demonstrated superior prediction ability using radiomics features, with the best model achieving an AUC of 0.72. Subgroup analysis revealed optimal performance in triple-negative breast cancer (AUC of 0.80) and HER2-positive subgroups (AUC of 0.65). Conclusion: Machine learning models incorporating clinical, qualitative, and radiomics data from pre-treatment MRI can effectively predict pCR in breast cancer patients receiving NAST, particularly among triple-negative and HER2-positive breast cancer subgroups. Full article

Endometrial carcinoma (EC) is the most common gynaecological malignancy in developed nations, exhibiting significant molecular heterogeneity that impacts prognosis and treatment response, particularly in advanced or recurrent settings. Traditional classification is increasingly supplemented by molecular subtyping (POLE-ultramutated, MSI-high/dMMR, NSMP, p53-mutated/CNH), which provides crucial prognostic information and predicts benefit from immunotherapy. This review summarizes the landscape of predictive biomarkers for immune checkpoint inhibitor (ICI) therapy in EC, emphasizing a new therapeutic scenario for advanced and recurrent EC. Mismatch repair deficiency (dMMR) or high microsatellite instability (MSI-H), leading to high tumor mutational burden (TMB) and increased neoantigen production, is the most established predictor, resulting in FDA approvals for pembrolizumab and dostarlimab in this subgroup. POLE mutations also confer hypermutation and high immunogenicity, predicting a favorable ICI response. Other biomarkers, including PD-L1 expression and TMB, show variable correlation with response and require further standardization. The tumor immune microenvironment, including tumor-infiltrating lymphocytes (TILs), also influences treatment outcomes. Clinical trials have demonstrated significant survival benefits for ICIs combined with chemotherapy (e.g., dostarlimab/pembrolizumab + carboplatin/paclitaxel) in first-line settings, especially for dMMR/MSI-H EC, and for ICI combinations with targeted agents (e.g., lenvatinib + pembrolizumab) in previously treated patients. Integrating molecular classification and validated biomarkers is essential for optimizing patient selection and developing personalized immunotherapy strategies for EC. Full article

Pulmonary hypertension (PH) encompasses a group of conditions characterized by elevated pulmonary arterial pressure, with pulmonary arterial hypertension (PAH) representing a distinct and severe subset. This review provides a comprehensive overview of the current classification system, highlighting the five clinical groups of PH and the specific hemodynamic criteria defining PAH. We discuss the complex pathophysiological mechanisms underlying PAH, including vascular remodeling, endothelial dysfunction, and genetic predisposition. Advances in diagnostic approaches are explored. Current treatment strategies targeting key molecular pathways such as endothelin, nitric oxide, and prostacyclin are reviewed alongside novel and investigational therapies. Prognostic indicators and risk stratification tools are evaluated to guide clinical management. Finally, we underscore the critical role of expert centers in accurate diagnosis, multidisciplinary care, and enrollment in clinical trials, which collectively improve patient outcomes in this challenging disease spectrum. Full article

Background/Objectives: Acute myeloid leukemia (AML) with FLT3 internal tandem duplication (FLT3-ITD) mutations carries a poor prognosis. While FLT3 inhibitors like midostaurin show benefits in combination with chemotherapy, the role of allelic ratio (AR), NPM1 mutation status, and hematopoietic stem cell transplantation (HSCT) remains uncertain. Real-world data can help refine prognostic classification and treatment strategies. Methods: We retrospectively analyzed 37 fit patients with FLT3-ITD AML treated with standard “7+3” chemotherapy, with and without midostaurin, between 2013 and 2022. Patients were stratified by FLT3-ITD AR, NPM1 status, and treatment approach. Outcomes assessed included complete remission (CR), disease-free survival (DFS), and overall survival (OS). Results: Overall, 67.6% achieved CR/CRi. Response rates did not differ significantly by AR (low vs. high: 66.7% vs. 69.2%) or midostaurin use (72.6% vs. 60%; p = 0.49). NPM1 mutations were associated with improved DFS (10.3 vs. 3 months, p = 0.036) but not OS. HSCT, performed in 54.1% of patients, mainly in first remission (CR1), significantly prolonged DFS (not reached vs. 5.3 months, p = 0.005) and remained an independent predictor in multivariate analysis (HR: 0.160, p = 0.039). OS (median 15.1 months) did not vary significantly across subgroups. Among patients achieving CR1, OS was significantly longer in those who underwent HSCT after midostaurin-based induction compared to those not transplanted (median OS not reached vs. 12.8 months; 95% CI, 6.9–18.7; p = 0.045), whereas no significant benefit was observed after standard induction. In a landmark analysis restricted to patients transplanted in CR1, those who had received midostaurin-based induction showed a trend toward improved OS compared to those treated with standard induction (median OS not reached vs. 11.5 months; 95% CI, 0.5–25.0; p = 0.086). Conclusions: This real-life study supports the importance of NPM1 mutations and HSCT in CR1, especially in the midostaurin era, for improving DFS in FLT3-ITD AML. These findings support updated guidelines for reducing the prognostic weight of AR and highlight the need for improved post-remission strategies in this setting. Full article

Background: Cognitive heterogeneity in Parkinson’s disease (PD) remains a diagnostic and prognostic challenge, particularly in early stages. In this cross-sectional study, we aimed to identify clinically relevant cognitive subtypes in early PD by integrating neuropsychological profiles with regional brain atrophy assessed via visual MRI scales. Methods: Eighty-one de novo PD patients (≤36 months from diagnosis) and twenty healthy controls underwent 3T MRI with visual atrophy ratings and completed an extensive neuropsychological battery. Results: Using a mixed a priori–a posteriori approach, we defined eight anatomocognitive subtypes reflecting distinct patterns of regional vulnerability: frontosubcortical, posterior cortical, left/right hippocampal, global, and preserved cognition. Specific MRI markers correlated with cognitive deficits in executive, visuospatial, memory, and language domains. Cluster analyses supported subtype validity (AUC range: 0.68–0.95). Conclusions: These results support a practical classification model linking cognitive performance to brain structural changes in early PD. This scalable approach may improve early patient stratification and guide personalized management strategies. Longitudinal studies are needed to assess progression patterns and therapeutic implications. Full article

HIV-associated lymphoma (HAL) is a heterogeneous and highly aggressive group of malignancies. Although antiretroviral therapy (ART) has significantly prolonged the survival of people living with HIV (PLWH), the risk of malignancy secondary to HIV infection remains higher than in HIV-negative individuals, with HAL being among the most frequent. The pathogenesis of HAL is complex, involving multifactorial interactions. In current clinical practice, HAL faces a double challenge: the lack of effective biological risk warning systems and the lack of precise prognostic stratification tools. In recent years, the construction of multidimensional biomarker systems has shown critical value in the comprehensive management of HAL. This review aims to systematically summarize recent advances in circulating biomarkers for HAL, focusing on the potential applications of immune environment indicators, such as inflammatory cytokine profiles and microbial translocation markers, as well as serum protein profiles, lymphocyte subsets, extracellular vesicles (EVs), circulating microRNAs (miRNAs), and viral biomarkers. These biomarkers offer promising avenues for early risk prediction, therapeutic monitoring, and prognostic evaluation. Developing an assessment system based on multidimensional biomarkers will optimize early risk stratification, enable precise prognostic classification, and support personalized therapeutic strategies, thereby providing a novel theoretical basis and practical direction for the clinical management of HAL. Full article