A public data-based molecular classification of small cell lung cancer by neuroactive signaling networks unveils distinct microenvironment landscapes and immunotherapy-related prognostic biomarkers

Preoperative molecular classification of endometrial cancer: Validation through biopsy and matched hysterectomy specimens.

Primary liver cancer (PLC) exhibits a high incidence and mortality rate. Early diagnosis and effective treatment are crucial for improving patient survival rates. This study aims to identify biomarkers of hepatitis B-related liver cancer and establish a new method for molecular subtype classification based on differential metabolite-related regulatory gene expression profiles. This study collected sterile midstream urine samples from patients with hepatitis B-related liver cancer who had not received standardized systematic antiviral therapy or anticancer therapy, as well as from healthy controls. Potential biomarkers were identified through liquid chromatography-tandem mass spectrometry (LC-MS/MS)-based metabolomics, followed by Kyoto Encyclopedia of Genes and Genomes (KEGG) enrichment analysis performed on the differential metabolites. Gene expression data of 371 hepatocellular carcinoma (HCC) samples in The Cancer Genome Atlas-Liver Hepatocellular Carcinoma (TCGA-LIHC) database were clustered using gene annotations for differential metabolites derived from the Human Metabolome Database (HMDB). The Kaplan-Meier (KM) survival curve was employed to assess the prognosis of different HCC molecular subtypes. Expression differences of subtype-specific genes and their enrichment in Hallmark, KEGG and Gene Ontology (GO) pathways were analyzed. The Tumor Immune Dysfunction and Exclusion (TIDE) scoring tool was used to evaluate the subtypes' response to immunotherapy. Sensitivity to sorafenib was also compared across the different subtypes. A total of 53 differential metabolites were identified (p < 0.01), which were significantly enriched in seven metabolic pathways (p < 0.05). Three potential biomarkers were discovered: Suberic acid, 2'-O-methylcytidine, and 3'-Sialyllactose. Regulatory genes associated with these differential metabolites clustered HCC samples from the TCGA-LIHC database into two molecular subtypes (C1 and C2). KM survival analysis indicated that patients in the C2 subtype exhibited higher overall survival compared to those in C1. Differential genes between the two subtypes were significantly enriched in Hallmark, KEGG and GO pathways. The TIDE scoring tool revealed a higher likelihood of immune escape in C1 subtype patients. Molecular targeted drug analysis suggested that sorafenib may be more effective in patients with the C1 subtype. Suberic acid, 2'-O-methylcytidine, and 3'-Sialyllactose hold promise as metabolic biomarkers for hepatitis B-related liver cancer. Understanding the diversity of the human liver cancer gene expression profile from a metabolomic perspective has potential applications for developing novel clinical treatment strategies.

Epidermal Growth Factor Receptor (EGFR) plays a critical role in the development of several cancers. Thus, modulation/inhibition of EGFR activity is an appealing target of developing novel cancer therapeutics. With the advent of modern machine learning technologies, it is now possible to simulate interactions with high precision between EGFR and small molecules to predict inhibitory/ modulatory activity at an unprecedented scale. In this work, we propose a novel machine-learning method to fast and precise classification of small compounds that are active, intermediate or inactive in inhibiting/modulating EGFR activity. We developed DeepEGFR, a novel multi-class graph neural network (GNN) model, to classify compounds into Active, Inactive, and Intermediate functional categories. DeepEGFR leverages complementary molecular representations, combining SMILES strings and molecular fingerprint matrices (Klekota-Roth and PubChem) to capture both structural and property-based features of compounds. The model constructs an advanced molecular graph representing atom type, formal charge, bond type, and bond order, through nodes and edges. DeepEGFR achieved superior performance compared to baseline machine learning algorithms (e.g., SVM, Random Forest, ANN), with approximately 94% F1-scores across training and test datasets for all activity classes. To ensure interpretability, the top 20 features identified by DeepEGFR were validated against the five key characteristics of FDA-approved EGFR inhibitors (Afatinib, Gefitinib, Osimertinib, Dacomitinib, Erlotinib), confirming the biological relevance of the features. Moreover, DeepEGFR successfully identified 300 underexplored EGFR-targeting compounds, demonstrating its potential to accelerate the discovery of therapeutic agents. These results highlight the effectiveness of graph neural networks in advancing molecular activity classification, setting a potential new benchmark for EGFR inhibitor prediction. These findings demonstrate the DeepEGFR’s ability to highlight the promising EGFR inhibitors, that have received limited prior investigation, thereby supporting its role in facilitating the rational development of targeted therapies for precision oncology.Supplementary InformationThe online version contains supplementary material available at 10.1038/s41598-025-22126-8.

Hepatocellular carcinoma (HCC) is characterized by distinct molecular and pathological subtypes, each with unique prognostic implications. This review aims to synthesize the imaging features associated with these HCC subtypes and discuss their potential to guide therapeutic decision-making. We searched PubMed and Embase for articles published from September 2004 to December 2024. The search strategy combined terms for imaging modalities ("CT," "MRI"), the primary disease ("hepatocellular carcinoma"), and various molecular and pathological subtypes (e.g., "macrotrabecular-massive," "steatohepatitic," "CK19," and "CTNNB1"). HCC is a biologically heterogeneous malignancy with varied prognosis and sensitivity to treatment. Assessment of its molecular and pathological subtypes relies on invasive histopathological examination, which is subject to sampling errors and often unavailable prior to treatment selection. A growing body of evidence suggests that radiologic features aid in the non-invasive classification of HCC subtypes, thereby informing individualized therapy. Given the substantial overlap between molecular, pathological, and imaging features, this review hypothesize that a comprehensive phenotyping system integrating all these information could significantly enhance personalized prognostication and treatment strategies. Radiologic imaging features not only provide valuable information for identifying molecular and pathological subtypes of HCC but also serve as practical tools to guide individualized therapeutic decision-making. Question Can CT and MRI reliably infer the molecular classification and pathological subtypes that drive prognosis in HCC? Findings Several imaging features have been found to reflect underlying molecular and pathological subtypes, but they do not demonstrate a one-to-one correlation. Clinical relevance An integrated classification system incorporating clinical, imaging, pathological, and molecular data may help mitigate the limitations of histologic and molecular analyses and facilitate individualized prognostication.

Preliminary investigations into the feasibility of Carbamoyl-phosphate synthetase 2, Aspartate transcarbamoylase, and Dihydroorotase (CAD)-targeted therapies have been conducted in a limited range of cancer types in pre-clinical studies. A comprehensive exploration of the diagnostic and prognostic capabilities of CAD, along with an understanding of its underlying biological mechanisms, is needed. A range of bioinformatics tools was employed to produce an extensive pan-cancer analysis of CAD expression. Experimental validation of the role of CAD in enzalutamide resistance in prostate cells was conducted. The molecular classification and drug patents of CAD were reviewed using the Worldwide Espacenet ®. Our study revealed that CAD was upregulated in tumor tissues in most cancer types. The expression of CAD was significantly different in clinical stages, pathological grades, and clinical prognosis. The highest frequency of CAD mutation was shown, but CAD mutations did not affect the clinical outcome of cancer patients. Comprehensive data across different cancer types illustrate the relationship between the expression of CAD and tumor mutation burden (TMB), microsatellite instability (MSI), and homologous recombination deficiency (HRD). Immune infiltration algorithms showed a positive link between CAD level and the prevalence of tumor-associated fibroblasts, MDSC, mast cells, and CD4+T cells. CAD level was positively linked to the immune checkpoint, suggesting a potential synergistic effect between CAD and immunotherapy. The GSEA analysis revealed that CAD expression is significantly associated with angiogenesis and epithelial-mesenchymal transition (EMT) pathways. Finally, we demonstrated that knockdown of CAD inhibits the growth of prostate cancer (PCa) cells and resistance to enzalutamide. This study revealed the diagnostic and prognostic potential of CAD. Notably, CAD exhibits essential functions in PCa cell proliferation and enzalutamide resistance.

BackgroundThe high molecular phenotype heterogeneity of cervical cancer (CC) is the main focus of individualized therapy. Molecular classification may lead to personal treatment and new drug discovery. We summarized the molecular features by establishing a new classification of metabolism-related gene expression profiles.MethodsClinical information and messenger ribonucleic acid (mRNA) expression were downloaded from The Cancer Genome Atlas (TCGA) and Gene Expression Omnibus (GEO). Twenty-two immune cells were detected by CIBERSORT method. K-means clustering algorithm based on 258 metabolism-related genes was used for CC classification. Univariate and multivariate Cox regression analyses were carried out to find out the optimal metabolism-related genes. A predictive model was established to evaluate the overall survival (OS) of patients. Then, a nomogram model was established to predict the OS of patients based on independent prognostic factors.ResultsBased on the expression profiles of 258 survival-related metabolic genes, we identified two metabolism-related subtypes of CC. Cluster_A subtype was characterized with significant glucose metabolism, and had a poor prognosis; and cluster_B subtype exhibited high enrichment of lipid metabolism-related and immune-related signaling pathways. Then, seven metabolism-related genes (CYP4F12, NPL, CH25H, NOS2, SDR16C5, PGK1 and LYZ) were used to establish a metabolism-related risk signature. Patients in high risk groups had a worse prognosis than those in low risk group. Multivariate Cox regression analysis indicted that the metabolism-related risk signature could predict OS as an independent prognostic factor.ConclusionsOur study provides new insight into the metabolic heterogeneity of CC and its relationship with immune landscape. The novel metabolism-related gene signature is an effective potential prognostic signature in the individualized prognosis prediction of CC.

Burkitt lymphoma (BL) remains a prevalent pediatric cancer in sub-Saharan Africa and was the first human cancer identified with a virus when Epstein-Barr virus (EBV) was discovered in a Ugandan BL tumor in 1964. The impact of EBV in BL is highlighted by a new molecular tumor classification of EBV positivity versus negativity which is starting to supersede longstanding epidemiologic classifications. The high incidence of EBV-positive BL in Africa and Papua New Guinea has been linked to Plasmodium falciparum (Pf) malaria coinfections in young children. Epidemiologic studies have yielded insight into early-age EBV infections and have demonstrated direct impacts of Pf malaria infections on EBV reactivation and disruptions in EBV persistence. Moreover, when children residing in malaria holoendemic regions are contending with chronic Pf malaria infections, they undergo immune adaptations to mitigate life-threatening immunopathology. We postulate that this malaria-induced immune conditioning leads to diminished EBV-specific cellular immune surveillance, when combined with higher B cell proliferation, and EBV load creates a permissive environment for BL tumorigenesis.

Advancing Lung Cancer Staging: Integrating IASLC Recommendations and Bioinformatics to Delineate Tumor Origins.

Molecular subtype classification of heterogeneous breast cancer is crucial for personalized therapies yet is limited by the low specificity of conventional single-target diagnosis systems. Herein, we developed a compact and versatile catalytic DNA computing (CDC) circuit as a programmable cancer evaluator for efficient dual-microRNA (miRNA) detection, enabling precise breast cancer subtype identification in clinical samples through a sequentially amplified multiplexed molecular imaging technique. Using an innovative and exquisite probe-concatenating and grafting strategy, the compact CDC system was engineered with minimal strand complexity, incorporating only two tandem-caged probes to form two distinct catalytic hairpin assembly (CHA) circuitry modules: pre-CDC and post-CDC modules. These CHA-based modules were sequentially activated by multiple miRNAs, enabling localized cascade signal amplification for the cancer subtype evaluation. Through systematic experimental validation and complementary theoretical simulations, we elucidated the sequential reaction mechanism and discovered the reaction kinetic confinement of the upstream pre-CDC module on the downstream post-CDC module activation. These findings provided valuable insights into the molecular reaction processes and offered critical guidance for designing efficient CDC probes. With its comprehensive multianalyte recognition and synergistic cascade amplification capabilities, the compact CDC circuit enabled the magnified detection of multiple miRNAs within cancer cells. The CDC platform demonstrated exceptional specificity in identifying clinical cancer tissues, making it a robust cancer cell subtype evaluator for breast cancer. Due to its high accuracy and reliability, this molecular evaluator serves as a promising diagnostic tool with potential applications in clinical diagnosis and disease-related molecular mechanism studies.

In ABO-incompatible (ABOi) kidney transplantation, C4d deposition is associated with accommodation rather than rejection. Isoagglutinins targeting blood group antigens A/B are also classified as donor-specific antibodies (DSA). Therefore, the diagnosis of antibody-mediated rejection (AMR) relies primarily on microvascular inflammation (MVI). We analyzed 66 ABOi and 251 ABO-compatible (ABOc) KTRs concerning anti-HLA DSA development. 46 protocol biopsies from ABOi KTRs were classified according to Banff 2022. In addition, 25 ABOi protocol biopsies were assessed by the Molecular Microscope Diagnostics System (MMDx) and compared to ABOc biopsies: (1) 35 DSA-negative, C4d-negative cases with MVI<2, (2) 16 C4d-positive cases with MVI<2, (3) 35 DSA-positive, C4d-negative cases with MVI=1 (probable AMR), and (4) 87 C4d-negative/positive cases with MVI≥2. ABOi KTRs showed lower rates of de novo anti-HLA DSA (p=0.001) and clinical AMR (p=0.018) than ABOc KTRs. Among 25 ABOi protocol biopsies analyzed with MMDx, 56% met AMR criteria due to anti-A/B DSA: 20% active AMR, 20% probable AMR, 16% chronic AMR. However, molecular AMR was confirmed in only 14% by MMDx (p<0.001). ABOi and DSA-negative, C4d-negative ABOc biopsies with MVI below threshold did not differ in molecular rejection, archetype, and lesion scores (p>0.05) and showed stable graft function. Molecular AMR classifier scores were significantly lower in ABOi and DSA-negative, C4d-negative ABOc cases with MVI below threshold compared to C4d-positive ABOc and ABOc cases with probable AMR (p=0.007). Notably, C4d drives molecular AMR activity in ABOc biopsies already at C4d1 levels by immunofluorescence (p=0.011) and even in the absence of a histological Banff AMR diagnosis (p=0.003). ABOi transplantation reduces the risk of developing de novo anti-HLA DSA. Banff 2022 criteria may over-diagnose AMR. Biopsy-based transcript diagnostics differentiate anti-HLA- and anti-A/B-mediated alloimmune injury from C4d deposition due to accommodation. Interestingly, C4d deposition drives molecular AMR activity in ABOc biopsies.

Abstract Castration-resistant prostate cancer (CRPC) is a considerable clinical challenge, driven by complex molecular mechanisms that enable tumors to evade androgen deprivation therapy. This review explores the molecular mechanisms driving CRPC progression, focusing on androgen receptor (AR) signaling, cancer stem cells (CSCs), and neuroendocrine differentiation (NED). In AR-dependent CRPC, AR signaling remains pivotal in disease progression. Mutations, splice variants, alternative pathways, and transcriptional regulation facilitate sustained AR activation despite androgen deprivation therapy. In addition, CSCs promote tumor recurrence and treatment resistance by maintaining cellular heterogeneity and evading conventional therapies. Furthermore, castration-resistant neuroendocrine prostate cancer, an aggressive subtype of CRPC, is characterized by AR independence and NED, making treatment challenging. These findings underscore the need for therapeutic strategies targeting AR-, CSC-, and NED-specific mechanisms. Crucially, the molecular classification of CRPC into AR-dependent CRPC, stem cell–like CRPC, and castration-resistant neuroendocrine prostate cancer subtypes—based on the interplay between AR signaling, CSCs, and neuroendocrine features—is essential for advancing precision medicine. Tailoring treatments to the molecular subtype and characteristics of each patient offers the potential to substantially improve prognosis and survival in CRPC.

To develop a deep learning model for predicting molecular subgroups of medulloblastoma (MB) using preoperative brain MRI. This study included a cohort of 350 patients with MB for model development. Preoperative multiparametric brain MRIs were acquired, and molecular classification data for tumor samples were analyzed. A dual-task deep learning model, composed of a 3D Swin Transformer backbone and a Transformer-based mask decoder, was developed for the prediction of MB molecular subgroups. The model was jointly optimized with a parallel task of tumor and cerebellum segmentation. Ablation analysis was conducted to verify the effectiveness of the dual-task model design. An independent test cohort of 126 patients with MB was established to validate the predictive performance of the dual-task model. Our dual-task deep learning model demonstrated superior performance for MB molecular subgroup prediction, achieving an AUC of 0.877, accuracy of 88.9%, sensitivity of 71.6%, and specificity of 91.9%. The performance remained robust across both adult and pediatric age populations, with AUCs of 0.915 and 0.871, respectively. Furthermore, our approach exhibited effective generalization to the independent test cohort, yielding an AUC of 0.853, accuracy of 89.7%, sensitivity of 73.5%, and specificity of 92.1%. Ablation analysis demonstrated a significant improvement in AUC of 0.169 (95% CI 0.097-0.244) when using the dual-task model design. In comparison with the radiomics-based model, our deep learning model achieved a higher AUC by 0.156 (95% CI 0.079-0.233). Our proposed dual-task deep learning model enables automated and accurate prediction of MB molecular subgroups.

Hepatocellular carcinoma (HCC) is the most common variant of primary liver cancer, characterized by high mortality and unfavorable prognosis. Global incidence and mortality from HCC continue to rise despite progress in the treatment of viral hepatitis due to the increasing prevalence of metabolic dysfunction-associa­ted steatotic liver disease and obesity. Based on the analysis of lite­rature sources from the Pubmed, MedLine, The Cochrane Library, Embase databases, the review summarizes current data on the epidemiology, key risk factors, pathogenesis and molecular classification of HCC. Special attention is paid to modern approaches to diagnosis, in particular imaging methods, the Liver Imaging Reporting and Data System, the role of non-invasive biomarkers and morphological verification. The importance of screening programs in high-risk groups and timely interdisciplinary interaction is emphasized, which allows optimizing the strategy of patient management in accordance with international clinical guidelines.

BackgroundEarly identification of molecular subtypes and WHO grades in adult-type diffuse gliomas (ADGs) provides critical evidence for prognostic evaluation and personalized therapeutic decision-making. This study aims to develop and validate radiopathomics models for the prediction of molecular subtypes and WHO grades in ADGs, addressing the limitations of unimodal approaches.MethodsIn this retrospective multicenter study, 499 consecutive ADG patients from three centers (training set: n = 306, testing set: n = 132, external validation set: n = 61) were included. Radiomics features were extracted from preoperative MRI sequences (T2-FLAIR and CE-T1WI), while pathomics features were derived from whole-slide images (WSIs). Feature selection methods and Multilayer Perceptron (MLP) classifier were performed to construct radiomics, pathomics, and radiopathomics models for molecular subtype classification and ADG grading. The performance of the model was evaluated using receiver operating characteristic (ROC) curves, area under the curve (AUC), accuracy, sensitivity, specificity, and F1 score. Decision curve analysis (DCA) was performed to assess clinical efficacy. The Shapley Additive Explanation (SHAP) analysis was employed to explore the interpretability of models.ResultsFor discriminating molecular subtypes, the radiopathomics model demonstrated superior performance compared to standalone radiomics or pathomics models, achieving AUCs (macro/micro) of 0.847/0.864 in the testing set, and AUCs (macro/micro) of 0.858/0.867 in the external validation set. For differentiating WHO grades, the radiopathomics model achieved superior performance compared to models based solely on radiomics or pathomics features. The AUCs for the radiopathomics model were 0.849 (95% CI 0.775–0.915) in the testing set and 0.855 (95% CI 0.748–0.945) in the external validation set. DCA confirmed superior net clinical benefit across wider risk thresholds compared to unimodal alternatives. SHAP analysis provided interpretable insights into the predictive significance and contributions of individual features.ConclusionThe proposed radiopathomics models demonstrate robust diagnostic performance by synergizing cross-scale features, offering a clinically actionable tool for ADG stratification.Supplementary InformationThe online version contains supplementary material available at 10.1186/s12967-025-07073-2.

Precision oncology in gastric cancer: Shaping the future of personalized treatment.

Molecular sequence analysis is crucial for understanding several biological processes, including protein-protein interactions, functional annotation, and disease classification. The large number of sequences and the inherently complicated nature of protein structures make it challenging to analyze such data. Finding patterns and enhancing subsequent research requires the use of dimensionality reduction and feature selection approaches. Recently, a method called Correlated Clustering and Projection (CCP) has been proposed as an effective method for biological sequencing data. The CCP technique remains computationally expensive, despite its effectiveness for sequence visualization. Furthermore, its utility for classifying molecular sequences is still uncertain. To solve these two problems, we present a Nearest-Neighbor Correlated Clustering and Projection (CCP-NN)-based technique for efficiently preprocessing molecular sequence data. To group related molecular sequences and produce representative supersequences, CCP makes use of sequence-to-sequence correlations. As opposed to conventional methods, CCP does not rely on matrix diagonalization, therefore, it can be applied to a range of machine-learning problems. We estimate the density map and compute the correlation using a nearest-neighbor search technique. We perform a molecular sequence classification using CCP and CCP-NN representations to assess the efficacy of our proposed approach. Our findings show that CCP-NN considerably improves the accuracy of the classification task and significantly outperforms CCP in terms of computational runtime.

BackgroundHepatocellular carcinoma (HCC) faces challenges in early diagnosis, prognosis, and treatment stratification due to molecular heterogeneity. This study aimed to establish a plasma exosomal long non-coding RNA (lncRNA)-based framework for molecular classification, prognostication, and therapeutic guidance in HCC.MethodsThe transcriptomic data from 230 plasma exosomes and 831 HCC tissues were integrated. A competitive endogenous RNA (ceRNA) network was constructed via the miRcode, miRTarBase, TargetScan, and miRDB databases to define exosome-related genes (ERGs). Unsupervised consensus clustering was used to stratify HCC patients on the basis of ERG profiles. Prognostic models were developed and optimized via 10 machine learning algorithms with 10-fold cross-validation. Treatment responses were predicted via the SubMap, TIDE, and oncoPredict algorithms. RT-qPCR experiments were conducted to validate the expression of model genes.ResultsWe identified 22 dysregulated plasma exosomal lncRNAs in HCC. The upregulated lncRNAs formed a ceRNA network regulating 61 ERGs and were significantly enriched in cell cycle regulation, TGF-β signaling, the p53 pathway, and ferroptosis. ERG expression stratified HCC into three subtypes (C1–C3). The C3 subtype exhibited the poorest overall survival, advanced grade and stage, an immunosuppressive microenvironment (increased Treg infiltration, elevated PD-L1/CTLA4 expression, highest TIDE score), and hyperactivation of proliferation (MYC, E2F targets) and metabolic pathways (glycolysis, mTORC1). A random survival forest-derived 6-gene risk score (G6PD, KIF20A, NDRG1, ADH1C, RECQL4, MCM4) demonstrated high prognostic accuracy. High-risk patients presented increased TP53/TTN mutations and increased tumor mutational burdens. Risk model analysis predicted differential treatment responses: low-risk patients exhibited superior anti-PD-1 immunotherapy responses, whereas high-risk patients showed increased sensitivity to DNA-damaging agents (e.g., the Wee1 inhibitor MK-1775) and sorafenib. Experimental validation confirmed consistent dysregulation of the six-gene signature (G6PD, KIF20A, NDRG1, ADH1C, RECQL4, MCM4) in HCC cell lines, reinforcing the model’s biological relevance.ConclusionPlasma exosomal lncRNAs enable robust molecular subtyping, accurate prognostic stratification, and treatment response prediction in HCC. The ERG-centric classification system and validated 6-gene risk model provide clinically actionable tools for precision oncology.

Homologous recombination deficiency (HRD) has emerged as a functional biomarker reflecting genome-wide DNA repair defects and genomic instability. While the Cancer Genome Atlas (TCGA) molecular classification provides valuable prognostic guidance in endometrial cancer (EC), it lacks resolution for DNA repair competency and therapeutic responsiveness. This study aimed to investigate whether HRD subtyping could complement TCGA classification for improved prognostic stratification and therapeutic decision-making. A total of 142 EC patients were analysed using a next-generation sequencing panel and genomic scar-based HRD scoring (loss of heterozygosity, telomeric allelic imbalance, large-scale state transitions). Unsupervised clustering stratified patients into HRD-High, -Middle, and -Low groups. Maximally selected rank statistics were used to identify prognostic thresholds for HRD scores; the tumour-immune microenvironment was characterised by RNA-based immune gene expression profiling and multiplex immunohistochemistry. A support vector machine (SVM) model was developed for recurrence prediction. HRD subtyping identified distinct genomic, pathological, and immunological features. HRD-High tumours were associated with advanced FIGO stages, TP53 mutations, higher chromosomal instability, and elevated CD8⁺PD-1⁺ T-cell infiltration. HRD subtyping independently predicted disease-free survival and showed superior prognostic accuracy (C-index = 0.857) compared to TCGA subtyping (C-index = 0.751). Integrating HRD and TCGA classifiers further improved predictive performance (C-index = 0.903). An SVM model incorporating HRD score and immune features achieved an AUC of 0.733 for recurrence prediction. HRD subtyping refines risk stratification beyond traditional TCGA classification and identifies patients potentially responsive to immune checkpoint or DNA damage-targeted therapies. Integrating HRD-based genomic instability metrics with molecular and immune profiling supports precision oncology in endometrial cancer.

Amyloidosis is a heterogeneous group of rare diseases characterized by the deposition of misfolded and insoluble proteins in tissues. Lung involvement, airways or parenchyma, is relatively common, but usually mild. Some patients may develop chronic progressive forms or experience life-threatening complications. In this review, we summarize the most recent advances in the comprehension of molecular mechanisms and classification of amyloidosis. We then illustrate the different forms of lung amyloidosis and the current treatment options. Increased awareness in the medical community and the creation of referral centers' networks are of paramount importance to ensure adequate management and access to treatment.