ChemMORT: an automatic ADMET optimization platform using deep learning and multi-objective particle swarm optimization.

Drug discovery and development constitute a laborious and costly undertaking. The success of a drug hinges not only good efficacy but also acceptable absorption, distribution, metabolism, elimination, and toxicity (ADMET) properties. Overall, up to 50% of drug development failures have been contributed from undesirable ADMET profiles. As a multiple parameter objective, the optimization of the ADMET properties is extremely challenging owing to the vast chemical space and limited human expert knowledge. In this study, a freely available platform called Chemical Molecular Optimization, Representation and Translation (ChemMORT) is developed for the optimization of multiple ADMET endpoints without the loss of potency (https://cadd.nscc-tj.cn/deploy/chemmort/). ChemMORT contains three modules: Simplified Molecular Input Line Entry System (SMILES) Encoder, Descriptor Decoder and Molecular Optimizer. The SMILES Encoder can generate the molecular representation with a 512-dimensional vector, and the Descriptor Decoder is able to translate the above representation to the corresponding molecular structure with high accuracy. Based on reversible molecular representation and particle swarm optimization strategy, the Molecular Optimizer can be used to effectively optimize undesirable ADMET properties without the loss of bioactivity, which essentially accomplishes the design of inverse QSAR. The constrained multi-objective optimization of the poly (ADP-ribose) polymerase-1 inhibitor is provided as the case to explore the utility of ChemMORT.

Structural insights into the N-terminal APHB domain of HrpA: mediating canonical and i-motif recognition.

RNA helicases function as versatile enzymes primarily responsible for remodeling RNA secondary structures and organizing ribonucleoprotein complexes. In our study, we conducted a systematic analysis of the helicase-related activities of Escherichia coli HrpA and presented the structures of both its apo form and its complex bound with both conventional and non-canonical DNAs. Our findings reveal that HrpA exhibits NTP hydrolysis activity and binds to ssDNA and ssRNA in distinct sequence-dependent manners. While the helicase core plays an essential role in unwinding RNA/RNA and RNA/DNA duplexes, the N-terminal extension in HrpA, consisting of three helices referred to as the APHB domain, is crucial for ssDNA binding and RNA/DNA duplex unwinding. Importantly, the APHB domain is implicated in binding to non-canonical DNA structures such as G-quadruplex and i-motif, and this report presents the first solved i-motif-helicase complex. This research not only provides comprehensive insights into the multifaceted roles of HrpA as an RNA helicase but also establishes a foundation for further investigations into the recognition and functional implications of i-motif DNA structures in various biological processes.

Deciphering the stereo-specific catalytic mechanisms of cis-epoxysuccinate hydrolases producing L(+)-tartaric acid.

Microbial epoxide hydrolases, cis-epoxysuccinate hydrolases (CESHs), have been utilized for commercial production of enantiomerically pure L(+)- and D(-)-tartaric acids for decades. However, the stereo-catalytic mechanism of CESH producing L(+)-tartaric acid (CESH[L]) remains unclear. Herein, the crystal structures of two CESH[L]s in ligand-free, product-complexed, and catalytic intermediate forms were determined. These structures revealed the unique specific binding mode for the mirror-symmetric substrate, an active catalytic triad consisting of Asp-His-Glu, and an arginine providing a proton to the oxirane oxygen to facilitate the epoxide ring-opening reaction, which has been pursued for decades. These results provide the structural basis for the rational engineering of these industrial biocatalysts.

Comprehensive assessment of nine target prediction web services: which should we choose for target fishing?

Identification of potential targets for known bioactive compounds and novel synthetic analogs is of considerable significance. In silico target fishing (TF) has become an alternative strategy because of the expensive and laborious wet-lab experiments, explosive growth of bioactivity data and rapid development of high-throughput technologies. However, these TF methods are based on different algorithms, molecular representations and training datasets, which may lead to different results when predicting the same query molecules. This can be confusing for practitioners in practical applications. Therefore, this study systematically evaluated nine popular ligand-based TF methods based on target and ligand-target pair statistical strategies, which will help practitioners make choices among multiple TF methods. The evaluation results showed that SwissTargetPrediction was the best method to produce the most reliable predictions while enriching more targets. High-recall similarity ensemble approach (SEA) was able to find real targets for more compounds compared with other TF methods. Therefore, SwissTargetPrediction and SEA can be considered as primary selection methods in future studies. In addition, the results showed that k = 5 was the optimal number of experimental candidate targets. Finally, a novel ensemble TF method based on consensus voting is proposed to improve the prediction performance. The precision of the ensemble TF method outperforms the individual TF method, indicating that the ensemble TF method can more effectively identify real targets within a given top-k threshold. The results of this study can be used as a reference to guide practitioners in selecting the most effective methods in computational drug discovery.

Graph deep learning enabled spatial domains identification for spatial transcriptomics.

Advancing spatially resolved transcriptomics (ST) technologies help biologists comprehensively understand organ function and tissue microenvironment. Accurate spatial domain identification is the foundation for delineating genome heterogeneity and cellular interaction. Motivated by this perspective, a graph deep learning (GDL) based spatial clustering approach is constructed in this paper. First, the deep graph infomax module embedded with residual gated graph convolutional neural network is leveraged to address the gene expression profiles and spatial positions in ST. Then, the Bayesian Gaussian mixture model is applied to handle the latent embeddings to generate spatial domains. Designed experiments certify that the presented method is superior to other state-of-the-art GDL-enabled techniques on multiple ST datasets. The codes and dataset used in this manuscript are summarized at https://github.com/narutoten520/SCGDL.

DKADE: a novel framework based on deep learning and knowledge graph for identifying adverse drug events and related medications.

Adverse drug events (ADEs) are common in clinical practice and can cause significant harm to patients and increase resource use. Natural language processing (NLP) has been applied to automate ADE detection, but NLP systems become less adaptable when drug entities are missing or multiple medications are specified in clinical narratives. Additionally, no Chinese-language NLP system has been developed for ADE detection due to the complexity of Chinese semantics, despite ˃10 million cases of drug-related adverse events occurring annually in China. To address these challenges, we propose DKADE, a deep learning and knowledge graph-based framework for identifying ADEs. DKADE infers missing drug entities and evaluates their correlations with ADEs by combining medication orders and existing drug knowledge. Moreover, DKADE can automatically screen for new adverse drug reactions. Experimental results show that DKADE achieves an overall F1-score value of 91.13%. Furthermore, the adaptability of DKADE is validated using real-world external clinical data. In summary, DKADE is a powerful tool for studying drug safety and automating adverse event monitoring.

Reducing false positive rate of docking-based virtual screening by active learning.

Machine learning-based scoring functions (MLSFs) have become a very favorable alternative to classical scoring functions because of their potential superior screening performance. However, the information of negative data used to construct MLSFs was rarely reported in the literature, and meanwhile the putative inactive molecules recorded in existing databases usually have obvious bias from active molecules. Here we proposed an easy-to-use method named AMLSF that combines active learning using negative molecular selection strategies with MLSF, which can iteratively improve the quality of inactive sets and thus reduce the false positive rate of virtual screening. We chose energy auxiliary terms learning as the MLSF and validated our method on eight targets in the diverse subset of DUD-E. For each target, we screened the IterBioScreen database by AMLSF and compared the screening results with those of the four control models. The results illustrate that the number of active molecules in the top 1000 molecules identified by AMLSF was significantly higher than those identified by the control models. In addition, the free energy calculation results for the top 10 molecules screened out by the AMLSF, null model and control models based on DUD-E also proved that more active molecules can be identified, and the false positive rate can be reduced by AMLSF.

Comprehensive evaluation of deep and graph learning on drug-drug interactions prediction.

Recent advances and achievements of artificial intelligence (AI) as well as deep and graph learning models have established their usefulness in biomedical applications, especially in drug-drug interactions (DDIs). DDIs refer to a change in the effect of one drug to the presence of another drug in the human body, which plays an essential role in drug discovery and clinical research. DDIs prediction through traditional clinical trials and experiments is an expensive and time-consuming process. To correctly apply the advanced AI and deep learning, the developer and user meet various challenges such as the availability and encoding of data resources, and the design of computational methods. This review summarizes chemical structure based, network based, natural language processing based and hybrid methods, providing an updated and accessible guide to the broad researchers and development community with different domain knowledge. We introduce widely used molecular representation and describe the theoretical frameworks of graph neural network models for representing molecular structures. We present the advantages and disadvantages of deep and graph learning methods by performing comparative experiments. We discuss the potential technical challenges and highlight future directions of deep and graph learning models for accelerating DDIs prediction.

Assembling spatial clustering framework for heterogeneous spatial transcriptomics data with GRAPHDeep.

MOTIVATION: Spatial clustering is essential and challenging for spatial transcriptomics' data analysis to unravel tissue microenvironment and biological function. Graph neural networks are promising to address gene expression profiles and spatial location information in spatial transcriptomics to generate latent representations. However, choosing an appropriate graph deep learning module and graph neural network necessitates further exploration and investigation. RESULTS: In this article, we present GRAPHDeep to assemble a spatial clustering framework for heterogeneous spatial transcriptomics data. Through integrating 2 graph deep learning modules and 20 graph neural networks, the most appropriate combination is decided for each dataset. The constructed spatial clustering method is compared with state-of-the-art algorithms to demonstrate its effectiveness and superiority. The significant new findings include: (i) the number of genes or proteins of spatial omics data is quite crucial in spatial clustering algorithms; (ii) the variational graph autoencoder is more suitable for spatial clustering tasks than deep graph infomax module; (iii) UniMP, SAGE, SuperGAT, GATv2, GCN, and TAG are the recommended graph neural networks for spatial clustering tasks; and (iv) the used graph neural network in the existent spatial clustering frameworks is not the best candidate. This study could be regarded as desirable guidance for choosing an appropriate graph neural network for spatial clustering. AVAILABILITY AND IMPLEMENTATION: The source code of GRAPHDeep is available at https://github.com/narutoten520/GRAPHDeep. The studied spatial omics data are available at https://zenodo.org/record/8141084.

Single Molecular Nanomedicines Based on Macrocyclic Carrier-Drug Conjugates for Concentration-Independent Encapsulation and Precise Activation of Drugs.

Nanomedicines often rely on noncovalent self-assembly and encapsulation for drug loading and delivery. However, challenges such as reproducibility issues due to the multicomponent nature, off-target activation caused by premature drug release, and complex pharmacokinetics arising from assembly dissociation have hindered their clinical translation. In this study, we introduce an innovative design concept termed single molecular nanomedicine (SMNM) based on macrocyclic carrier-drug conjugates. Through the covalent linkage of two chemotherapy drugs to a hypoxia-cleavable macrocyclic carrier, azocalix[4]arene, we obtained two self-included complexes to serve as SMNMs. The intramolecular inclusion feature of the SMNMs has not only demonstrated comprehensive shielding and protection for the drugs but also effectively prevented off-target drug leakage, thereby significantly reducing their side effects and enhancing their antitumor therapeutic efficacy. Additionally, the attributes of being a single component and molecularly dispersed confer advantages such as ease of preparation and good reproducibility for SMNMs, which is desirable for clinical applications.

Efficacy and Safety of Vonoprazan-Amoxicillin Dual Regimen With Varying Dose and Duration for Helicobacter pylori Eradication: A Multicenter, Prospective, Randomized Study.

BACKGROUND & AIMS: Previous studies confirm vonoprazan-amoxicillin effectiveness for Helicobacter pylori. This study aims to investigate vonoprazan with varying amoxicillin dose and duration. METHODS: This multicenter, prospective, randomized controlled, noninferiority trial enrolled patients with treatment naive H pylori infection from 5 clinical centers. Eligible participants were randomly assigned to H-VA-10 (vonoprazan 20 mg twice a day (b.i.d.) + amoxicillin 750 mg 4 times a day, 10 days), L-VA-10 (vonoprazan 20 mg b.i.d. + amoxicillin 1000 mg b.i.d, 10 days), and H-VA-14 (vonoprazan 20 mg b.i.d + amoxicillin 750 mg 4 times a day, 14 days) in a 1:1:1 ratio. The eradication rate was assessed using the 13C-urea breath test at least 28 days after treatment. RESULTS: Of the 623 eligible patients, 516 patients were randomized. In both the intention-to-treat and per-protocol analyses, eradication rates were comparable between H-VA-10 and H-VA-14 groups (86.6% vs 89.5% and 90.9% vs 94.5%, P = .021 and .013 for noninferiority, respectively). However, eradication rates were significantly lower in the L-VA-10 group than the H-VA-14 group (79.7% vs 89.5% and 82.0% vs 94.5%, P = .488 and .759, respectively). Rates of study withdrawal, loss to follow-up, and adverse events were similar across study groups. CONCLUSIONS: H-VA-10 and H-VA-14 regimens provide satisfactory efficacy for H pylori infection, and the L-VA-10 regimen was inferior. CLINICALTRIALS: gov number: NCT05719831.

RGA1 regulates grain size, rice quality and seed germination in the small and round grain mutant srg5.

BACKGROUND: Generating elite rice varieties with high yield and superior quality is the main goal of rice breeding programs. Key agronomic traits, including grain size and seed germination characteristics, affect the final yield and quality of rice. The RGA1 gene, which encodes the α-subunit of rice G-protein, plays an important role in regulating rice architecture, seed size and abiotic stress responses. However, whether RGA1 is involved in the regulation of rice quality and seed germination traits is still unclear. RESULTS: In this study, a rice mutant small and round grain 5 (srg5), was identified in an EMS-induced rice mutant library. Systematic analysis of its major agronomic traits revealed that the srg5 mutant exhibited a semi-dwarf plant height with small and round grain and reduced panicle length. Analysis of the physicochemical properties of rice showed that the difference in rice eating and cooking quality (ECQ) between the srg5 mutant and its wild-type control was small, but the appearance quality was significantly improved. Interestingly, a significant suppression of rice seed germination and shoot growth was observed in the srg5 mutant, which was mainly related to the regulation of ABA metabolism. RGA1 was identified as the candidate gene for the srg5 mutant by BSA analysis. A SNP at the splice site of the first intron disrupted the normal splicing of the RGA1 transcript precursor, resulting in a premature stop codon. Additional linkage analysis confirmed that the target gene causing the srg5 mutant phenotype was RGA1. Finally, the introduction of the RGA1 mutant allele into two indica rice varieties also resulted in small and round rice grains with less chalkiness. CONCLUSIONS: These results indicate that RGA1 is not only involved in the control of rice architecture and grain size, but also in the regulation of rice quality and seed germination. This study sheds new light on the biological functions of RGA1, thereby providing valuable information for future systematic analysis of the G-protein pathway and its potential application in rice breeding programs.

Robust and Corrosion-Resistant Overall Water Splitting Electrode Enabled by Additive Manufacturing.

Electrolysis of water has emerged as a prominent area of research in recent years. As a promising catalyst support, copper foam is widely investigated for electrolytic water, yet the insufficient mechanical strength and corrosion resistance render it less suitable for harsh working conditions. To exploit high-performance catalyst supports, various metal supports are comprehensively evaluated, and Ti6 Al4 V (Ti64) support exhibited outstanding compression and corrosion resistance. With this in mind, a 3D porous Ti64 catalyst support is fabricated using the selective laser sintering (SLM) 3D printing technology, and a conductive layer of nickel (Ni) is coated to increase the electrical conductivity and facilitate the deposition of catalysts. Subsequently, Co0.8 Ni0.2 (CO3 )0.5 (OH)·0.11H2 O (CoNiCH) nanoneedles are deposited. The resulting porous Ti64/Ni/CoNiCH electrode displayed an impressive performance in the oxygen evolution reaction (OER) and reached 30 mA cm-2 at an overpotential of only 200 mV. Remarkably, even after being compressed at 15.04 MPa, no obvious structural deformation is observed, and the attenuation of its catalytic efficiency is negligible. Based on the computational analysis, the CoNiCH catalyst demonstrated superior catalytic activity at the Ni site in comparison to the Co site. Furthermore, the electrode reached 30 mA cm-2 at 1.75 V in full water splitting conditions and showed no significant performance degradation even after 60 h of continuous operation. This study presents an innovative approach to robust and corrosion-resistant catalyst design.

ABC-Net: a divide-and-conquer based deep learning architecture for SMILES recognition from molecular images.

Structural information for chemical compounds is often described by pictorial images in most scientific documents, which cannot be easily understood and manipulated by computers. This dilemma makes optical chemical structure recognition (OCSR) an essential tool for automatically mining knowledge from an enormous amount of literature. However, existing OCSR methods fall far short of our expectations for realistic requirements due to their poor recovery accuracy. In this paper, we developed a deep neural network model named ABC-Net (Atom and Bond Center Network) to predict graph structures directly. Based on the divide-and-conquer principle, we propose to model an atom or a bond as a single point in the center. In this way, we can leverage a fully convolutional neural network (CNN) to generate a series of heat-maps to identify these points and predict relevant properties, such as atom types, atom charges, bond types and other properties. Thus, the molecular structure can be recovered by assembling the detected atoms and bonds. Our approach integrates all the detection and property prediction tasks into a single fully CNN, which is scalable and capable of processing molecular images quite efficiently. Experimental results demonstrate that our method could achieve a significant improvement in recognition performance compared with publicly available tools. The proposed method could be considered as a promising solution to OCSR problems and a starting point for the acquisition of molecular information in the literature.

BioNet: a large-scale and heterogeneous biological network model for interaction prediction with graph convolution.

MOTIVATION: Understanding chemical-gene interactions (CGIs) is crucial for screening drugs. Wet experiments are usually costly and laborious, which limits relevant studies to a small scale. On the contrary, computational studies enable efficient in-silico exploration. For the CGI prediction problem, a common method is to perform systematic analyses on a heterogeneous network involving various biomedical entities. Recently, graph neural networks become popular in the field of relation prediction. However, the inherent heterogeneous complexity of biological interaction networks and the massive amount of data pose enormous challenges. This paper aims to develop a data-driven model that is capable of learning latent information from the interaction network and making correct predictions. RESULTS: We developed BioNet, a deep biological networkmodel with a graph encoder-decoder architecture. The graph encoder utilizes graph convolution to learn latent information embedded in complex interactions among chemicals, genes, diseases and biological pathways. The learning process is featured by two consecutive steps. Then, embedded information learnt by the encoder is then employed to make multi-type interaction predictions between chemicals and genes with a tensor decomposition decoder based on the RESCAL algorithm. BioNet includes 79 325 entities as nodes, and 34 005 501 relations as edges. To train such a massive deep graph model, BioNet introduces a parallel training algorithm utilizing multiple Graphics Processing Unit (GPUs). The evaluation experiments indicated that BioNet exhibits outstanding prediction performance with a best area under Receiver Operating Characteristic (ROC) curve of 0.952, which significantly surpasses state-of-theart methods. For further validation, top predicted CGIs of cancer and COVID-19 by BioNet were verified by external curated data and published literature.

Molecular Recognition with Macrocyclic Receptors for Application in Precision Medicine.

ConspectusMacrocyclic receptors can serve as alternatives to natural recognition systems as recognition tools. They provide effectively preorganized cavities to encapsulate guests via host-guest interactions, thereby affecting the physiochemical properties of the guests. Macrocyclic receptors exhibit chemical and thermal stabilities higher than those of natural receptors and thus are expected to resist degradation inside the body. This reduces the risk of harmful degradation byproducts and ensures optimal levels of effectiveness. Macrocyclic receptors have precise molecular weights and well-defined structures; this ensures their batch-to-batch reproducibility, which is critical for ensuring quality and effectiveness levels. Moreover, macrocyclic receptors exhibit broad modification tunabilities, rendering them adaptable to various guests. Molecular recognition is the basis of numerous biological processes. Macrocyclic receptors may display considerable potential for application in diagnosing and treating diseases, depending on the host-guest recognition of bioactive molecules. However, the binding affinities and selectivities of macrocyclic receptors toward bioactive molecules are generally insufficient, which may lead to problems such as low diagnosis accuracies, off-target leaking, and interference with normal functions. Therefore, addressing the challenge of the strong and specific complexation of bioactive molecules and macrocyclic receptors is imperative.To overcome this challenge, we proposed the innovative strategies of longitudinal cavity extension and coassembled heteromultivalent recognition for application in the recognition of small molecules and biomacromolecules, respectively. The deepened cavity provides a stronger hydrophobic effect and a larger interaction area while maintaining the framework rigidity. By coassembling two macrocyclic amphiphiles into one ensemble, we achieved the desired heteromultivalent recognition. This strategy affords the necessary binding properties while preventing the requirement of tedious steps and site mismatch in covalent synthesis. Using these two strategies, we achieved specific and strong binding of macrocyclic receptors to various bioactive molecules including biomarkers, drugs, and disease-related peptides/proteins. We then applied these macrocyclic receptor-based recognition systems in biosensing and bioimaging, drug delivery, and therapeutics.In this Account, we summarize the strategies we used in the recognition of small molecules and biomacromolecules. Thereafter, we discuss their applications in precision medicine, involving the (1) sensing of biomarkers and imaging of lesion sites, which are critical in the early screening of diseases and accurate diagnoses; (2) precise loading and targeted delivery of drugs, which are crucial in improving their therapeutic efficacies and reducing their side effects; and (3) capture and removal of disease-related biomacromolecules, which are significant for precise intervention in life processes. Finally, we propose recommendations for the further development of macrocyclic receptor-based recognition systems in biomedicine. Macrocyclic receptors exhibit considerable potential for research, and continued investigation may not only expand the applications of supramolecular chemistry but also open novel avenues for the development of precision medicine.

Heterozygous Seryl-tRNA Synthetase 1 Variants Cause Charcot-Marie-Tooth Disease.

OBJECTIVE: Despite the increasing number of genes associated with Charcot-Marie-Tooth (CMT) disease, many patients currently still lack appropriate genetic diagnosis for this disease. Autosomal dominant mutations in aminoacyl-tRNA synthetases (ARSs) have been implicated in CMT. Here, we describe causal missense mutations in the gene encoding seryl-tRNA synthetase 1 (SerRS) for 3 families affected with CMT. METHODS: Whole-exome sequencing was performed in 16 patients and 14 unaffected members of 3 unrelated families. The functional impact of the genetic variants identified was investigated using bioinformatic prediction tools and confirmed using cellular and biochemical assays. RESULTS: Combined linkage analysis for the 3 families revealed significant linkage (Zmax LOD = 6.9) between the genomic co-ordinates on chromosome 1: 108681600-110300504. Within the linkage region, heterozygous SerRS missense variants segregated with the clinical phenotype in the 3 families. The mutant SerRS proteins exhibited reduced aminoacylation activity and abnormal SerRS dimerization, which suggests the impairment of total protein synthesis and induction of eIF2α phosphorylation. INTERPRETATION: Our findings suggest the heterozygous SerRS variants identified represent a novel cause for autosomal dominant CMT. Mutant SerRS proteins are known to impact various molecular and cellular functions. Our findings provide significant advances on the current understanding of the molecular mechanisms associated with ARS-related CMT. ANN NEUROL 2023;93:244-256.

Cellular Uptake of Cell-Penetrating Peptides Activated by Amphiphilic p-Sulfonatocalix[4]arenes.

We report the synthesis of a series of amphiphilic p-sulfonatocalix[4]arenes with varying alkyl chain lengths (CX4-Cn) and their application as efficient counterion activators for membrane transport of cell-penetrating peptides (CPPs). The enhanced membrane activity is confirmed with the carboxyfluorescein (CF) assay in vesicles and by the direct cytosolic delivery of CPPs into CHO-K1, HCT 116, and KTC-1 cells enabling excellent cellular uptake of the CPPs into two cancer cell lines. Intracellular delivery was confirmed by fluorescence microscopy after CPP entry into live cells mediated by CX4-Cn, which was also quantified after cell lysis by fluorescence spectroscopy. The results present the first systematic exploration of structure-activity relationships for calixarene-based counterion activators and show that CX4-Cn are exceptionally effective in cellular delivery of CPPs. The dodecyl derivative, CX4-C12, serves as best activator. A first mechanistic insight is provided by efficient CPP uptake at 4 °C and in the presence of the endocytosis inhibitor dynasore, which indicates a direct translocation of the CPP-counterion complexes into the cytosol and highlights the potential benefits of CX4-Cn for efficient and direct translocation of CPPs and CPP-conjugated cargo molecules into the cytosol of live cells.

Characteristics of patients with recurrent retinoblastoma: a survival analysis.

BACKGROUND: Management guidelines and corresponding survival data for patients with recurrent retinoblastoma (RB) are lacking. This study aimed to summarize the clinical characteristics of patients with recurrent RB and analyze their survival outcomes. METHODS: We retrospectively analyzed 68 patients with recurrent RB who underwent treatment in our institution from January 2016 to December 2020. Patients were grouped according to location of recurrence: intraocular, orbital, and distant metastasis. RESULTS: The male:female ratio was 1.3:1 and the median age at recurrence was 37.5 months (range, 30.3-62.8). The number of patients in the intraocular recurrence, orbital recurrence, and metastasis groups was 13 (19.1%), 23 (33.8%), and 32 (47.1%), respectively. Thirty patients died, 36 were alive at last follow-up, and two were lost to follow-up. Eye enucleation was performed in 94.1% of patients. Five-year overall survival in patients with intraocular recurrence, orbital recurrence, and metastasis was 84.6%, 69.6%, and 31.3%, respectively (P = 0.001). Most deaths occurred within 2 years of recurrence. Presence of high-risk pathological factors, central nervous system invasion, and absence of combination therapy were independent predictors of worse 5-year overall survival. CONCLUSION: The rate of eye preservation in survivors of recurrent RB was very low. Although 5-year overall survival in patients who underwent treatment for intraocular and orbital recurrence was high, it was low in those with metastasis. RB patients may need lifelong follow-up for recurrence and secondary malignancy.

Prognostic value of the Naples prognostic score in patients with intrahepatic cholangiocarcinoma after hepatectomy.

BACKGROUND: The Naples Prognostic Score (NPS), integrating inflammatory and nutritional biomarkers, has been reported to be associated with the prognosis of various malignancies, but there is no report on intrahepatic cholangiocarcinoma (ICC). This study aimed to explore the prognostic value of NPS in patients with ICC. METHODS: Patients with ICC after hepatectomy were collected, and divided into three groups. The prognosis factors were determined by Cox regression analysis. Predictive efficacy was evaluated by the time-dependent receiver operating characteristic (ROC) curves. RESULTS: A total of 174 patients were included (Group 1: 33 (19.0%) patients; Group 2: 83 (47.7%) patients; and Group 3: 58 (33.3%) patients). The baseline characteristics showed the higher the NPS, the higher the proportion of patients with cirrhosis and Child-Pugh B, and more advanced tumors. The Kaplan-Meier curves reflect higher NPS were associated with poor survival. Multivariable analysis showed NPS was an independent risk factor of overall survival (NPS group 2 vs. 1: HR = 1.671, 95% CI: 1.022-3.027, p = 0.009; NPS group 3 vs. 1: HR = 2.208, 95% CI: 1.259-4.780, p = 0.007) and recurrence-free survival (NPS group 2 vs. 1: HR = 1.506, 95% CI: 1.184-3.498, p = 0.010; NPS group 3 vs. 1: HR = 2.141, 95% CI: 2.519-4.087, P = 0.001). The time ROC indicated NPS was superior to other models in predicting prognosis. CONCLUSIONS: NPS is a simple and effective tool for predicting the long-term survival of patients with ICC after hepatectomy. Patients with high NPS require close follow-up, and improving NPS may prolong the survival time.