Search

Summary: Background & Aims: In recent years, artificial intelligence (AI) models based on chest radiography have gained attention in various fields, including cardiac function, age estimation, and clinical assessment during hospitalization. The Global Leadership Initiative on Malnutrition (GLIM) criteria are widely used in malnutrition diagnosis. The objective of this study is to examine the predictive ability of chest radiographs using deep learning techniques in relation to hematological parameters that are already known to be associated with malnutrition, as well as malnutrition scores, including GLIM criteria. Methods: A total of 3701 older patients (age ≥65 years) were admitted to our hospital from January 2021 to January 2022, and after excluding those with missing data, participants (N=2862) were enrolled. Chest Radiographs, Basic information (height, weight, age, and sex), hematological parameters (albumin, hemoglobin, lymphocyte count, and C-reactive protein), malnutritional assessments: GLIM severity, Geriatric Nutritional Risk Index (GNRI), modified Controlling Nutritional status (mCONUT), and Subjective Global Assessment (SGA), and Nutritional Support Team (NST) intervention were extracted and utilized as training and validation data. We used chest radiographic image's matrix as explanatory variables for numerical (hematological parameters) or categorical (scoring) nutritional data as objective variables. A previously reported deep learning model helped construct a chest radiography-based prediction model from the training data. The predicted data were evaluated by computing the correlation coefficients and area under the curve (AUC). Results: As a numerical variables analysis, albumin and hemoglobin predictions were relatively accurate (R=0.71, 0.74). As a categorical malnutritional prediction, chest radiograph-based AI effectively aided nutritional decisions based on GNRI (AUC: 0.88, 0.83, 0.88, and 0.90), SGA (0.75, 0.71, and 0.88), mCONUT (0.84, 0.88, 0.90, and 0.88), and GLIM severity (0.88, 0.82, and 0.85) class indices. The Class activation maps (CAM) analysis identified variation in the X-ray sites for each malnutritional AI prediction, with some sites in agreement and others in disagreement. Conclusion: Deep learning-based chest radiographic AI has the potential to more accurately reflect malnutrition scoring than hematologic parameters. Furthermore, it can predict outcomes and assess malnutrition, including GLIM criteria. It is anticipated that AI will be integrated into the NST workflow in the future.

Abstract We measured intestinal fatty acid-binding protein (I-FABP) levels, a useful marker of small intestinal mucosal injury, in patients with cirrhosis to determine their relationship with liver function and complications. This cross-sectional study included 71 patients with cirrhosis admitted for treatment of cirrhotic complications or hepatocellular carcinoma (cohort A) and 104 patients with cirrhosis who received direct-acting antiviral therapy for HCV (cohort B). I-FABP levels, measured by ELISA, were evaluated relative to hepatic reserve and compared with non-invasive scoring systems for diagnostic performance in cirrhotic complications. The median I-FABP level in both cohorts were significantly elevated in patients with reduced hepatic reserve (CTP grade A/BC cohort A, 2.33/3.17 ng/mL, p = 0.032; cohort B, 2.46/3.64 ng/mL, p = 0.008) and complications with gastroesophageal varices (GEV; GEV (-)/(+) cohort A, 1.66/3.67 ng/mL, p

Abstract This study aimed to address the diagnostic challenges in distinguishing between alcohol-related liver disease (ALD) and metabolic dysfunction-associated steatotic liver disease (MASLD). We utilized whole-slide imaging technology to conduct a comprehensive digital analysis of liver specimens collected from patients undergoing transplantation. This study included 36 and 17 patients with ALD and MASLD cirrhosis, respectively, who underwent transplantation at our institution. Digital slides were analyzed for fibrosis patterns using FibroNest™. Patient background characteristics were comparable between ALD (n = 36) and MASLD (n = 17) groups, except for sex. The ALD group exhibited thicker collagen per strand, longer and more flexural fibrosis, and a more heterogeneous distribution than the MASLD group. In patients with ALD and concomitant metabolic dysfunction, fiber distribution became relatively uniform, resembling MASLD. Application of the phenotypic fibrosis composite score achieved 100% sensitivity and specificity for ALD/MASLD diagnosis. Digital pathological analysis of the fibrosis patterns showed morphological differences between ALD and MASLD. This approach holds promise for histological differentiation, providing valuable insights beyond the current definitions based solely on alcohol intake. This study emphasizes the potential of digital pathology in refining the diagnostic criteria for hepatic disorders.

Abstract Although the use of immune checkpoint inhibitors (ICIs)-targeted agents for unresectable hepatocellular carcinoma (HCC) is promising, individual response variability exists. Therefore, we developed an artificial intelligence (AI)-based model to predict treatment efficacy using pre-ICIs contrast-enhanced computed tomography (CT) imaging characteristics. We evaluated the efficacy of atezolizumab and bevacizumab in 43 patients at the Nagasaki University Hospital from 2020 to 2022 using the modified Response Evaluation Criteria in Solid Tumors. A total of 197 Progressive Disease (PD), 271 Partial Response (PR), and 342 Stable Disease (SD) contrast CT images of HCC were used for training. We used ResNet-18 as the Convolutional Neural Network (CNN) model and YOLOv5, YOLOv7, YOLOv8 as the You Only Look Once (YOLO) model with precision-recall curves and class activation maps (CAMs) for diagnostic performance evaluation and model interpretation, respectively. The 3D t-distributed Stochastic Neighbor Embedding was used for image feature analysis. The YOLOv7 model demonstrated Precision 53.7%, Recall 100%, F1 score 69.8%, mAP@0.5 99.5% for PD, providing accurate and clinically versatile predictions by identifying decisive points. The ResNet-18 model had Precision 100% and Recall 100% for PD. However, the CAMs sites did not align with the tumors, suggesting the CNN model is not predicting that a given CT slice is PD, PR, or SD, but that it accurately predicts Individual Patient's CT slices. Preparing substantial training data for tumor drug effect prediction models is challenging compared to general tumor diagnosis models; hence, large-scale validation using an efficient YOLO model is warranted.

Abstract Aim After pancreaticoduodenectomy, 20–40% of patients develop steatotic liver disease (SLD), and steatohepatitis can be a problem. Although patatin‐like phospholipase domain‐containing 3 protein (PNPLA3) and transmembrane 6 superfamily member 2 (TM6SF2) polymorphisms are involved in SLD and steatohepatitis development, whether this is the case after pancreaticoduodenectomy is unclear. Methods and Results Forty‐three patients with pancreatic cancer who underwent pancreaticoduodenectomy at our hospital between April 1, 2018, and March 31, 2021, were included. We extracted DNA from noncancerous areas of residual specimens after pancreaticoduodenectomy and determined PNPLA3 and TM6SF2 gene polymorphisms using real‐time polymerase chain reaction. SLD was defined as a liver with an attenuation value of ≤40 HU or a liver‐to‐spleen ratio of ≤0.9 on computed tomography. We defined high hepatic fibrosis indexes (HFI) instead of steatohepatitis as a Fibrosis‐4 index of ≥2.67 or nonalcoholic fatty liver disease fibrosis score of ≥0.675 in patients with SLD. The cumulative incidence of SLD (P = 0.299) and high HFI (P = 0.987) after pancreaticoduodenectomy were not significantly different between the PNPLA3 homozygous and minor allele groups. The incidences of high HFI at 1 year after pancreaticoduodenectomy were 16.8% and 27.0% in the TM6SF2 major homozygous and minor allele groups, respectively, with a significant difference in the cumulative incidence (P = 0.046). Conclusion The TM6SF2 minor allele may contribute to steatohepatitis development after pancreaticoduodenectomy.

Summary: Background & Aims: Malnutrition in the elderly, frequently and significantly affects both physical functioning and cognition, as well as incurs direct and indirect costs to society. Guidelines recommend rapid nutritional intervention and initiation of nutritional therapy within 24–48 hours of admission. Height and weight information is essential for proper nutritional assessment; however, it is difficult to obtain individual the height and weight of bedridden elderly patients directly. This study aimed to illustrate the potential of a convolutional neural network model to assess the height and weight based on chest radiographs. Methods: We retrospectively evaluated radiographs obtained over 15 years of follow-up. Overall, 6,453 radiographs from male patients, and 7,879 from female patients were included in the analysis. A convolutional neural network was used to predict the height and weight of the patients (Juzen NST). A ResNet152 classifier was trained using Fastai (V1.0) running on PyTorch to predict the height and weight. Training was performed for four epochs using validation without augmentation. Results: The correlation coefficients between the predicted and measured values using the height prediction model for males and females were R=0.855 and R=0.81, respectively. The correlation coefficients between the values predicted by the weight prediction model and measured values were R=0.793 and R=0.86, respectively. Conclusion: Our chest radiographic prediction model has a high correlation with actual height and weight and can be combined with information regarding clinical nutrition factors for rapid assessment of risk for malnutrition. By training the prediction model using chest radiographs from each hospital, it can be optimized for the most common ethnic groups in the area. Importantly, our model is an illustration of the potential of automated imaging artificial intelligence for proper nutrition prediction models in older adults.

Abstract Background and Aim Reports have indicated that a surface area of 4 mm2 or more of collected tissue sections could provide the recommended total DNA for the OncoGuide NCC Oncopanel system, which is a cancer gene panel test developed in Japan. We wished to compare the percentage of tissue sections collected by endoscopic ultrasound‐assisted tissue acquisition (EUS‐TA) with surface areas of ≥4 mm2 between a conventional needle, namely the EZ Shot 3 Plus (Olympus Medical Japan, Tokyo, Japan) (EZ3), and the recent SonoTip TopGain (MediGlobe, Rohrdorf, Germany) (TopGain). Method From April 2010 to December 2021, among 693 EUS‐TA cases, EZ3 was used in 390 cases and TopGain in 45. The EZ3 and TopGain groups were matched in a 1:1 ratio with a tolerance of 0.2, with 35 patients each matched using propensity score analysis. Results The TopGain group had a significantly higher percentage of cases with a tissue area of ≥4 mm2 than the EZ3 group (42.9% vs 68.6%, P = 0.030). Multivariate analysis revealed an association between TopGain and tissue areas of ≥4 mm2 (odds ratio 2.996, 95% confidence interval 1.068–8.403, P = 0.037). Conclusions EUS‐TA using TopGain significantly collected more ≥4 mm2 tissue area compared with EZ3, suggesting its usefulness for cancer gene panel testing.

Abstract Impaired glucose tolerance, glucose fluctuations, and hypoglycemia have been observed in patients with chronic liver disease (CLD). The flash glucose monitoring (FGM) system, which recognises continuous and dynamic glucose changes in real time, is used in daily clinical practice. This study aimed to examine the association between glucose fluctuations and hypoglycemia, as measured by the FGM system, and liver-related events. Seventy-two patients with CLD and type 2 DM who had their blood glucose measured using Freestyle Libre Pro between April 2017 and July 2018 at our institution were enrolled in this retrospective study. We assessed the results of the FGM system measurements and liver-related events, as defined by gastrointestinal bleeding, infection, ascites, encephalopathy, and liver-related death. The standard deviation (SD) of mean glucose as measured by the FGM system was 41.55 mg/dl, and hypoglycemia was observed in 48.6% (35/72) of the patients. Liver-related event-free survival was not significant when stratified based on SD; however, the event-free survival was significantly lower when stratified by hypoglycemia (p = 0.007). In a multivariate analysis using the Cox proportional hazards model, Child–Pugh class B [Hazards ratio (HR) 2.347 (95% confidence interval (CI): 1.042–5.283), p = 0.039] and hypoglycemia [HR 2.279 (95% CI: 1.064–4.881), p = 0.034] were identified as factors contributing to event-free survival. Hypoglycemia, as determined by the FGM system, was identified as a significant factor that was closely associated with liver-related events. In addition to measuring glucose levels, the FGM system is useful in predicting the occurrence of liver-related events.

Abstract Aim Immune checkpoint inhibitors (ICIs) have proven to be effective treatments for various cancers, but can also elicit immune‐related adverse events (irAEs). Given that severe irAEs can be life‐threatening, biomarkers that can predict the occurrence of irAEs are of paramount importance. ICIs affect the dynamics of lymphocytes, and alterations in these dynamics may play a role in the development and severity of irAEs. The aim of this study was to investigate the correlation between irAEs and changes in lymphocyte counts. Methods Information on irAEs was collected from 226 ICI cases from 2014 to 2020. We compared lymphocyte counts before treatment and at the onset of irAE and investigated the association between lymphocyte count fluctuations and the presence and severity of irAE, the course after steroid treatment, and overall survival. Results Of the 226 cases, 27 patients developed grade 3 or higher irAE. Compared to the other groups, the lymphocyte count in this group was significantly decreased at the time of irAE (p

Sphingosine 1-phosphate (S1P) is a bioactive sphingolipid associated with nonalcoholic steatohepatitis (NASH). Immune cell-driven inflammation is a key determinant of NASH progression. Macrophages, monocytes, NK cells, T cells, NKT cells, and B cells variably express S1P receptors from a repertoire of 5 receptors termed S1P1 – S1P5. We have previously demonstrated that non-specific S1P receptor antagonism ameliorates NASH and attenuates hepatic macrophage accumulation. However, the effect of S1P receptor antagonism on additional immune cell populations in NASH remains unknown. We hypothesized that S1P receptor specific modulation may ameliorate NASH by altering leukocyte recruitment. A murine NASH model was established by dietary feeding of C57BL/6 male mice with a diet high in fructose, saturated fat, and cholesterol (FFC) for 24 weeks. In the last 4 weeks of dietary feeding, the mice received the S1P1,4,5 modulator Etrasimod or the S1P1 modulator Amiselimod, daily by oral gavage. Liver injury and inflammation were determined by histological and gene expression analyses. Intrahepatic leukocyte populations were analyzed by flow cytometry, immunohistochemistry, and mRNA expression. Alanine aminotransferase, a sensitive circulating marker for liver injury, was reduced in response to Etrasimod and Amiselimod treatment. Liver histology showed a reduction in inflammatory foci in Etrasimod-treated mice. Etrasimod treatment substantially altered the intrahepatic leukocyte populations through a reduction in the frequency of T cells, B cells, and NKT cells and a proportional increase in CD11b+ myeloid cells, polymorphonuclear cells, and double negative T cells in FFC-fed and control standard chow diet (CD)-fed mice. In contrast, FFC-fed Amiselimod-treated mice showed no changes in the frequencies of intrahepatic leukocytes. Consistent with the improvement in liver injury and inflammation, hepatic macrophage accumulation and the gene expression of proinflammatory markers such as Lgals3 and Mcp-1 were decreased in Etrasimod-treated FFC-fed mice. Etrasimod treated mouse livers demonstrated an increase in non-inflammatory (Marco) and lipid associated (Trem2) macrophage markers. Thus, S1P1,4,5 modulation by Etrasimod is more effective than S1P1 antagonism by Amiselimod, at the dose tested, in ameliorating NASH, likely due to the alteration of leukocyte trafficking and recruitment. Etrasimod treatment results in a substantial attenuation of liver injury and inflammation in murine NASH.

Extracellular vesicles (EVs) are emerging mediators of intercellular communication in nonalcoholic steatohepatitis (NASH). Palmitate, a lipotoxic saturated fatty acid, activates hepatocellular endoplasmic reticulum stress, which has been demonstrated to be important in NASH pathogenesis, including in the release of EVs. We have previously demonstrated that the release of palmitate-stimulated EVs is dependent on the de novo synthesis of ceramide, which is trafficked by the ceramide transport protein, STARD11. The trafficking of ceramide is a critical step in the release of lipotoxic EVs, as cells deficient in STARD11 do not release palmitate-stimulated EVs. Here, we examined the hypothesis that protein cargoes are trafficked to lipotoxic EVs in a ceramide-dependent manner. We performed quantitative proteomic analysis of palmitate-stimulated EVs in control and STARD11 knockout hepatocyte cell lines. Proteomics was performed on EVs isolated by size exclusion chromatography, ultracentrifugation, and density gradient separation, and EV proteins were measured by mass spectrometry. We also performed human EV proteomics from a control and a NASH plasma sample, for comparative analyses with hepatocyte-derived lipotoxic EVs. Size exclusion chromatography yielded most unique EV proteins. Ceramide-dependent lipotoxic EVs contain damage-associated molecular patterns and adhesion molecules. Haptoglobin, vascular non-inflammatory molecule-1, and insulin-like growth factor-binding protein complex acid labile subunit were commonly detected in NASH and hepatocyte-derived ceramide-dependent EVs. Lipotoxic EV proteomics provides novel candidate proteins to investigate in NASH pathogenesis and as diagnostic biomarkers for hepatocyte-derived EVs in NASH patients.

Monocyte homing to the liver and adhesion to the liver sinusoidal endothelial cells (LSECs) are key elements in nonalcoholic steatohepatitis (NASH) pathogenesis. We reported previously that VCAM-1 mediates monocyte adhesion to LSECs. However, the pathogenic role of VCAM-1 in NASH is unclear. Herein, we report that VCAM-1 was a top upregulated adhesion molecule in the NASH mouse liver transcriptome. Open chromatin landscape profiling combined with genome-wide transcriptome analysis showed robust transcriptional upregulation of LSEC VCAM-1 in murine NASH. Moreover, LSEC VCAM-1 expression was significantly increased in human NASH. LSEC VCAM-1 expression was upregulated by palmitate treatment in vitro and reduced with inhibition of the mitogen-activated protein 3 kinase (MAP3K) mixed lineage kinase 3 (MLK3). Likewise, LSEC VCAM-1 expression was reduced in the Mlk3-/- mice with diet-induced NASH. Furthermore, VCAM-1 neutralizing Ab or pharmacological inhibition attenuated diet-induced NASH in mice, mainly via reducing the proinflammatory monocyte hepatic population as examined by mass cytometry by time of flight (CyTOF). Moreover, endothelium-specific Vcam1 knockout mice were also protected against NASH. In summary, lipotoxic stress enhances the expression of LSEC VCAM-1, in part, through MLK3 signaling. Inhibition of VCAM-1 was salutary in murine NASH and might serve as a potential therapeutic strategy for human NASH. [ABSTRACT FROM AUTHOR]