Endoscopic detection and diagnosis of gastric cancer using image-enhanced endoscopy: A systematic review and meta-analysis.

Objectives: We aimed to conduct a systematic review and meta-analysis to assess the value of image-enhanced endoscopy including blue laser imaging (BLI), linked color imaging, narrow-band imaging (NBI), and texture and color enhancement imaging to detect and diagnose gastric cancer (GC) compared to that of white-light imaging (WLI). Methods: Studies meeting the inclusion criteria were identified through PubMed, Cochrane Library, and Japan Medical Abstracts Society databases searches. The pooled risk ratio for dichotomous variables was calculated using the random-effects model to assess the GC detection between WLI and image-enhanced endoscopy. A random-effects model was used to calculate the overall diagnostic performance of WLI and magnifying image-enhanced endoscopy for GC. Results: Sixteen studies met the inclusion criteria. The detection rate of GC was significantly improved in linked color imaging compared with that in WLI (risk ratio, 2.20; 95% confidence interval [CI], 1.39-3.25; p < 0.01) with mild heterogeneity. Magnifying endoscopy with NBI (ME-NBI) obtained a pooled sensitivity, specificity, and area under the summary receiver operating curve of 0.84 (95 % CI, 0.80-0.88), 0.96 (95 % CI, 0.94-0.97), and 0.92, respectively. Similarly, ME-BLI showed a pooled sensitivity, specificity, and area under the curve of 0.81 (95 % CI, 0.77-0.85), 0.85 (95 % CI, 0.82-0.88), and 0.95, respectively. The diagnostic efficacy of ME-NBI/BLI for GC was evidently high compared to that of WLI, However, significant heterogeneity among the NBI studies still existed. Conclusions: Our meta-analysis showed a high detection rate for linked color imaging and a high diagnostic performance of ME-NBI/BLI for GC compared to that with WLI.

Effect of texture and color enhancement imaging on the visibility of gastric tumors.

Texture and color enhancement imaging (TXI) may improve the visibility of gastric tumors and allow their early detection. However, few reports have examined the utility of TXI. Between June 2021 and October 2022, 56 gastric tumors in 51 patients undergoing endoscopic submucosal dissection at Fukuchiyama City Hospital were evaluated preoperatively using conventional white light imaging (WLI), narrow-band imaging (NBI), and TXI modes 1 and 2. The color differences of the tumors and surrounding mucosae were evaluated using the CIE 1976 L*a*b color space, Additionally, the visibility scores were scaled. Of the 56 gastric tumors, 45 were early gastric cancers, and 11 were adenomas. Overall, the color difference in TXI mode 1 was considerably higher compared to WLI (16.36 ± 7.05 vs. 10.84 ± 4.05; p < 0.01). Moreover, the color difference in early gastric cancers was considerably higher in TXI mode 1 compared to WLI, whereas no significant difference was found in adenomas. The visibility score in TXI mode 1 was the highest, and it was significantly higher compared to WLI. Regarding adenomas, the visibility score in TXI mode 1 was also significantly higher compared to that in WLI. TXI may provide improved gastric tumor visibility.

Pseudo-nephropathy and hyper-excretion of urinary C-peptide: an overlooked adverse effect of an angiotensin receptor-neprilysin inhibitor (ARNI).

Sacubitril/valsartan, which is a combined angiotensin receptor-neprilysin inhibitor (ARNI), is used for the treatment of chronic heart failure and hypertension. Substrates of neprilysin are numerous, and the systemic effects of an ARNI remain to be determined. Increased urinary C-peptide (UCPR) and urinary albumin (UAlb) excretion has been reported with the use of an ARNI, but the mechanism is still unknown. We report an 84-year-old man with type 2 diabetes and hypertension. His UAlb and UCPR excretion and (to a lesser degree) the estimated glomerular filtration rate were increased after ARNI administration. They returned to basal levels after discontinuing ARNI administration. There was little or no change in glycemic control. Therefore, increased glomerular permeability and filtration could partially explain how neprilysin inhibition led to an elevation in UCPR excretion, in addition to other mechanisms, such as impairment of the renal ability to degrade C-peptide. Physicians must be cautious when interpreting the insulin secretion capability by UCPR and nephropathy by UAlb in ARNI-treated patients with diabetes.

Linked color imaging with artificial intelligence improves the detection of early gastric cancer.

INTRODUCTION: Esophagogastroduodenoscopy (EGD) is the most important tool to detect gastric cancer (GC). In this study, we developed a computer-aided system (CADe) to detect gastric cancer (GC) with white light imaging (WLI) and linked color imaging (LCI) modes and aimed to compare the performance of CADe with that of endoscopists. METHODS: The system was developed based on the deep learning framework from 9021 images in 385 patients between 2017 and 2020. A total of 116 LCI and WLI videos from 110 patients between 2017 and 2023 were used to evaluate per-case sensitivity and per-frame specificity. RESULTS: The per-case sensitivity and per-frame specificity of CADe with a confidence level of 0.5 in detecting GC were 78.6% and 93.4% for WLI and 94.0% and 93.3% for LCI, respectively (P < 0.001). The per-case sensitivities of nonexpert endoscopists for WLI and LCI were 45.8% and 80.4%, whereas those of expert endoscopists were 66.7% and 90.6%, respectively. Regarding detectability between CADe and endoscopists, the per-case sensitivities for WLI and LCI were 78.6% and 94.0% in CADe, respectively, which were significantly higher than those for LCI in experts (90.6%, P = 0.004) and those for WLI and LCI in nonexperts (45.8% and 80.4%, respectively, P < 0.0001); however, no significant difference for WLI was observed between CADe and experts (P = 0.134). CONCLUSIONS: Our CADe system showed a significantly better sensitivity in detecting GC when used in LCI compared with WLI mode. Moreover, the sensitivity of CADe using LCI is significantly higher than those of expert endoscopists using LCI.

Non-polypoid Colorectal Lesions Detection and False Positive Detection by Artificial Intelligence under Blue Laser Imaging and Linked Color Imaging.

Objectives: Artificial intelligence (AI) with white light imaging (WLI) is not enough for detecting non-polypoid colorectal polyps and it still has high false positive rate (FPR). We developed AIs using blue laser imaging (BLI) and linked color imaging (LCI) to detect them with specific learning sets (LS). Methods: The contents of LS were as follows, LS (WLI): 1991 WLI images of lesion of 2-10 mm, LS (IEE): 5920 WLI, BLI, and LCI images of non-polypoid and small lesions of 2-20 mm. LS (IEE) was extracted from videos and included both in-focus and out-of-focus images. We designed three AIs as follows: AI (WLI) finetuned by LS (WLI), AI (IEE) finetuned by LS (WLI)+LS (IEE), and AI (HQ) finetuned by LS (WLI)+LS (IEE) only with images in focus. Polyp detection using a test set of WLI, BLI, and LCI videos of 100 non-polypoid or non-reddish lesions of 2-20 mm and FPR using movies of 15 total colonoscopy were analyzed, compared to 2 experts and 2 trainees. Results: The sensitivity for LCI in AI (IEE) (83%) was compared to that for WLI in AI (IEE) (76%: p=0.02), WLI in AI (WLI) (57%: p<0.01), BLI in AI (IEE) (78%: p=0.14), and LCI in trainees (74%: p<0.01). The sensitivity for LCI in AI (IEE) (83%) was significantly higher than that in AI (HQ) (78%: p<0.01). The FPR for LCI (6.5%) in AI (IEE) were significantly lower than that in AI (HQ) (17.3%: p<0.01). Conclusions: AI finetuned by appropriate LS detected non-reddish and non-polypoid polyps under LCI.

Carfilzomib shows therapeutic potential for reduction of liver fibrosis by targeting hepatic stellate cell activation.

Because hepatic stellate cells (HSCs) play a major role in fibrosis, we focused on HSCs as a potential target for the treatment of liver fibrosis. In this study, we attempted to identify drug candidates to inactivate HSCs and found that several proteasome inhibitors (PIs) reduced HSC viability. Our data showed that a second-generation PI, carfilzomib (CZM), suppressed the expression of fibrotic markers in primary murine HSCs at low concentrations of 5 or 10 nM. Since CZM was not toxic to HSCs up to a concentration of 12.5 nM, we examined its antifibrotic effects further. CZM achieved a clear reduction in liver fibrosis in the carbon tetrachloride (CCl4)-induced mouse model of liver fibrosis without worsening of liver injury. Mechanistically, RNA sequence analysis of primary HSCs revealed that CZM inhibits mitosis in HSCs. In the CCl4-injured liver, amphiregulin, which is known to activate mitogenic signaling pathways and fibrogenic activity and is upregulated in murine and human metabolic dysfunction-associated steatohepatitis (MASH), was downregulated by CZM administration, leading to inhibition of mitosis in HSCs. Thus, CZM and next-generation PIs in development could be potential therapeutic agents for the treatment of liver fibrosis via inactivation of HSCs without liver injury.

[Gastric-type adenocarcinoma consisting of signet-ring cell carcinoma and adenocarcinoma of the fundic gland type:a case report].

After endoscopic treatment for esophageal cancer, a 65-year-old male underwent surveillance esophagogastroduodenoscopy. A 12-mm discolored flat lesion was noted on the greater curvature of the middle gastric body. Magnifying endoscopy with blue laser imaging demonstrated an irregular papillary surface. Biopsy revealed atypical cells with mucus and irregularly distributed nuclei. The lesion was diagnosed as a gastric-type neoplasm with low atypia. Thereafter, endoscopic submucosal dissection (ESD) was conducted and specimen was sent for biopsy. The ESD specimen suggested a signet-ring cell carcinoma with MUC5AC-positive phenotype and adenocarcinoma of the fundic gland type, with MUC6 positivity and pepsinogen I positivity in the shallow and deeper layers, respectively. Moreover, the cervical region of fundic glands demonstrated a transformation zone of the signet-ring cell carcinoma into an adenocarcinoma of the fundic gland type. Herein, we report this rare case along with a literature review.

A Man with Primary Hyperchylomicronemia with Triglyceride Levels Exceeding 11,000 mg/dL was Well Controlled by Pemafibrate Combined with Dietary Therapy: A Case Report.

A 50-year-old man with a triglyceride (TG) level of 11,397 mg/dL was admitted to our hospital. He consumed a high-fat and high-carbohydrate diet as well as more than 100 g of alcohol per day. He had type 2 diabetes and obesity and had previously suffered from severe acute pancreatitis twice. A genetic analysis revealed compound heterozygous mutations in APOA5 (c.56C>G and c.553G>T). In addition to low-fat meals and alcohol cessation, administration of pemafibrate lowered his triglyceride levels to <150 mg/dL.

Urolithin A-mediated augmentation of intestinal barrier function through elevated secretory mucin synthesis.

Maintaining the mucus layer is crucial for the innate immune system. Urolithin A (Uro A) is a gut microbiota-derived metabolite; however, its effect on mucin production as a physical barrier remains unclear. This study aimed to elucidate the protective effects of Uro A on mucin production in the colon. In vivo experiments employing wild-type mice, NF-E2-related factor 2 (Nrf2)-deficient mice, and wild-type mice treated with an aryl hydrocarbon receptor (AhR) antagonist were conducted to investigate the physiological role of Uro A. Additionally, in vitro assays using mucin-producing cells (LS174T) were conducted to assess mucus production following Uro A treatment. We found that Uro A thickened murine colonic mucus via enhanced mucin 2 expression facilitated by Nrf2 and AhR signaling without altering tight junctions. Uro A reduced mucosal permeability in fluorescein isothiocyanate-dextran experiments and alleviated dextran sulfate sodium-induced colitis. Uro A treatment increased short-chain fatty acid-producing bacteria and propionic acid concentration. LS174T cell studies confirmed that Uro A promotes mucus production through the AhR and Nrf2 pathways. In conclusion, the enhanced intestinal mucus secretion induced by Uro A is mediated through the actions of Nrf-2 and AhR, which help maintain intestinal barrier function.

Efficacy of Two Kinds of Scissor-Type Knives for Colorectal Endoscopic Submucosal Dissection: A Retrospective Comparative Study.

INTRODUCTION: Scissor-type knives are spreading as safe devises in endoscopic submucosal dissection (ESD). We evaluated the efficacy of two kinds of scissor-type knives (Clutch Cutter: CC, Fujifilm Co. and SB Knife Jr2: SB, SB-KAWASUMI Laboratories. Inc.) in colorectal ESD. METHODS: This single-center retrospective study analyzed 178 ESD cases treated with CC from January 2020 to August 2021 and 91 cases with SB from September 2021 to December 2023. The two groups were compared through propensity score matching. Therapeutic results, such as ESD procedure time, en bloc resection rate, perioperative bleeding frequency, and complications, were analyzed in each group. Risk factors for long ESD procedure time (≥ 90 min) were also examined. RESULTS: After matching, 87 cases in each group were analyzed. There was no significant difference in the ESD procedure time (min, median [interquartile range]) between the CC and SB groups (54.0 [36.0-72.0] vs. 53.0 [39.0-72.0], p = 0.99). Additionally, there were no differences in the en bloc resection (100% vs. 100%, p = 1.00), perioperative perforation (1.1% vs. 1.1%, p = 1.00), or delayed bleeding (1.1% vs. 0.0%, p = 1.00). There was a significant difference in perioperative bleeding frequency (mean ± standard deviation: 1.8 ± 2.6 vs. 3.0 ± 3.5, p < 0.01). The significant risk factors (odds ratio [95% confidence interval]) for long ESD procedure time in patients treated with CC or SB were antiplatelet (7.51 [1.82-31.00]), large lesion size (1.08 [1.05-1.12]), severe fibrosis (24.30 [7.60-77.90]), and perioperative bleeding frequency (1.34 [1.14-1.56]). CONCLUSIONS: CC and SB in colorectal ESD enabled high en bloc resection and low complication rates. CC showed significantly less perioperative bleeding than SB.

Evolution of Systemic Treatment for Hepatocellular Carcinoma: Changing Treatment Strategies and Concepts.

Systemic therapy for hepatocellular carcinoma (HCC) has undergone substantial advancements. With the advent of atezolizumab plus bevacizumab (ATZ/BEV) combination therapy, followed by durvalumab plus tremelimumab, the era of immunotherapy for HCC has commenced. The emergence of systemic treatment with high response rates has led to improvements in overall survival while enabling conversion to radical surgical resection in some patients with HCC. In patients with intermediate-stage HCC, new treatment strategies combining systemic treatment and transcatheter arterial chemoembolization (TACE) are under development in clinical trials. Moreover, the addition of local therapies, such as TACE, to systemic treatment according to the treatment effect could achieve a certain percentage of complete response. In the IMbrave050 trial, the efficacy of ATZ/BEV combination therapy was validated in patients predicted to have a high risk of recurrence, especially in those who had undergone radical surgery or radiofrequency ablation for HCC. Therefore, systemic treatment for HCC is entering a new phase for all disease stages. The objective of this review is to organize the current position of systemic therapy for each HCC stage and discuss the development of new treatment methods and strategies, with a focus on regimens incorporating immune checkpoint inhibitors, along with future prospects.

Development of high-quality artificial intelligence for computer-aided diagnosis in determining subtypes of colorectal cancer.

BACKGROUND AND AIM: There are no previous studies in which computer-aided diagnosis (CAD) diagnosed colorectal cancer (CRC) subtypes correctly. In this study, we developed an original CAD for the diagnosis of CRC subtypes. METHODS: Pretraining for the CAD based on ResNet was performed using ImageNet and five open histopathological pretraining image datasets (HiPreD) containing 3 million images. In addition, sparse attention was introduced to improve the CAD compared to other attention networks. One thousand and seventy-two histopathological images from 29 early CRC cases at Kyoto Prefectural University of Medicine from 2019 to 2022 were collected (857 images for training and validation, 215 images for test). All images were annotated by a qualified histopathologist for segmentation of normal mucosa, adenoma, pure well-differentiated adenocarcinoma (PWDA), and moderately/poorly differentiated adenocarcinoma (MPDA). Diagnostic ability including dice sufficient coefficient (DSC) and diagnostic accuracy were evaluated. RESULTS: Our original CAD, named Colon-seg, with the pretraining of both HiPreD and ImageNET showed a better DSC (88.4%) compared to CAD without both pretraining (76.8%). Regarding the attentional mechanism, Colon-seg with sparse attention showed a better DSC (88.4%) compared to other attentional mechanisms (dual: 79.7%, ECA: 80.7%, shuffle: 84.7%, SK: 86.9%). In addition, the DSC of Colon-seg (88.4%) was better than other types of CADs (TransUNet: 84.7%, MultiResUnet: 86.1%, Unet++: 86.7%). The diagnostic accuracy of Colon-seg for each histopathological type was 94.3% for adenoma, 91.8% for PWDA, and 92.8% for MPDA. CONCLUSION: A deep learning-based CAD for CRC subtype differentiation was developed with pretraining and fine-tuning of abundant histopathological images.

Ramucirumab for advanced hepatocellular carcinoma in the current real world: a Japanese single-arm study post-REACH-2 (The R-evolution study).

This study aimed to complement the results of the REACH-2 study by prospectively evaluating the safety and efficacy of ramucirumab in advanced hepatocellular carcinoma (HCC) in a real-world setting. This was an open-label, nonrandomized, multicenter, prospective study conducted at 13 institutions in Japan (jRCTs031190236). The study included Child-Pugh Class A patients with advanced HCC who had received pretreatment with atezolizumab plus bevacizumab (Atez/Bev) or lenvatinib. Ramucirumab was introduced as a second-line treatment after Atez/Bev or lenvatinib and as a third-line treatment after Atez/Bev and lenvatinib. Between May 2020 and July 2022, we enrolled 19 patients, including 17 who received ramucirumab. Additionally, seven patients received lenvatinib, another seven patients received Atez/Bev, and three patients received Atez/Bev followed by lenvatinib as prior treatment. The primary endpoint was a 6-month progression-free survival (PFS) rate, which was 14.3%. The median PFS and overall survival were 3.7 and 12.0 months, respectively. The most common grade ≥ 3 adverse events (AEs) were hypertension (23.5%), proteinuria (17.6%), and neutropenia (11.8%). The discontinuation rate due to AEs was 29.4%. Six patients progressed from Child-Pugh A to B after treatment with ramucirumab. Thirteen patients were eligible for post-ramucirumab treatment, including systemic therapy. Despite the limited number of patients, the efficacy of ramucirumab was comparable to that observed in the REACH-2 study when used after lenvatinib and Atez/Bev. However, the incidence of AEs was higher than that in the REACH-2 study.

The Comparison of Diagnostic Ability between Blue Laser/Light Imaging and Narrowband Imaging for Sessile Serrated Lesions with or without Dysplasia.

Objectives: Diagnostic ability of sessile serrated lesions (SSL) and SSL with dysplasia (SSLD) using blue laser/light imaging (BLI) has not been well examined. We analyzed the diagnostic accuracy of BLI for SSL and SSLD using several endoscopic findings compared to those of narrow band imaging (NBI). Materials and Methods: This was a subgroup analysis of prospective studies. 476 suspiciously serrated lesions of ≥2 mm on the proximal colon showing serrated change with magnified NBI or BLI in our institution between 2014 and 2021 were examined histopathologically. After propensity score matching, we evaluated the diagnostic ability of SSL and SSLD of the NBI and BLI groups regarding various endoscopic findings. For WLI findings, granule, depression, and reddish were examined for diagnosing SSLD. For NBI/BLI findings, expanded crypt opening (ECO) or thick and branched vessels (TBV) were examined for diagnosing SSL. Network vessels (NV) and white dendritic change (WDC) defined originally were examined for diagnosing SSLD. Results: Among matched 176 lesions, the sensitivity of lesions with either ECO or TBV for SSL in the NBI/BLI group was 97.5%/98.5% (p = 0.668). Those with either WDC or NV for diagnosing SSLD in the groups were 81.0%/88.9% (p = 0.667). Regarding the rates of endoscopic findings among 30 SSLD and 290 SSL, there were significant differences in WDC (66.4% vs. 8.6%, p < 0.001), NV (55.3% vs. 1.4%, p < 0.001), and either WDC or NV (86.8% vs. 9.0%, p < 0.001). Conclusions: The diagnostic ability of BLI for SSL and SSLD was not different from NBI. NV and WDC were useful for diagnosing SSLD.

Safety/efficacy of atezolizumab + bevacizumab during anti-platelet/anticoagulation therapy in unresectable hepatocellular carcinoma.

BACKGROUND AND AIMS: This study aimed to determine the safety and efficacy of atezolizumab + bevacizumab therapy in hepatocellular carcinoma patients receiving anti-platelet agents or anticoagulants. METHODS: Patients were divided into those using (IM out) and those not using (IM in) anti-platelet agents or anticoagulants, who violated the exclusion criteria of the IMbrave150 trial, and were retrospectively examined. RESULTS: The study included 185 patients (IM in: 157; IM out: 28). For first-line treatment, progression-free survival was 184 days for IM in and 266 days for IM out (p = .136). Overall survival was 603 days for IM in and not reached for IM out (p = .265), with no significant between-group difference. Similarly, there were no significant between-group differences in progression-free survival or overall survival for later-line treatment. Haemorrhagic adverse events of ≥grade 3 were observed in 11 IM in patients and 3 IM out patients. No significant factors associated with haemorrhagic adverse events of ≥grade 3 were identified in the multivariate analysis including IM out classification, whose p value was .547. Regarding thrombotic/embolic adverse events in the IM out group, one case of exacerbation of portal vein thrombosis was observed. No deaths were directly attributable to bleeding events or exacerbations of thrombosis. CONCLUSION: Atezolizumab + bevacizumab therapy shows similar safety and efficacy in patients receiving and those not receiving anti-platelet agents or anticoagulants; therefore, it can be considered for patients with hepatocellular carcinoma receiving anti-platelet agents or anticoagulants.

Pathophysiological Implications of Protein Lactylation in Pancreatic Epithelial Tumors.

Protein lactylation is a post-translational modification associated with glycolysis. Although recent evidence indicates that protein lactylation is involved in epigenetic gene regulation, its pathophysiological significance remains unclear, particularly in neoplasms. Herein, we investigated the potential involvement of protein lactylation in the molecular mechanisms underlying benign and malignant pancreatic epithelial tumors, as well as its role in the response of pancreatic cancer (PC) cells to gemcitabine. Increased lactylation was observed in the nuclei of intraductal papillary mucinous adenoma, non-invasive intraductal papillary mucinous carcinoma, and invasive carcinoma, in parallel to the upregulation of hypoxia-inducible factor-1α. This observation indicated that a hypoxia-associated increase in nuclear protein lactylation could be a biochemical hallmark in pancreatic epithelial tumors. The standard PC chemotherapy drug gemcitabine suppressed histone lactylation in vitro, suggesting that histone lactylation might be relevant to its mechanism of action. Taken together, our findings suggest that protein lactylation may be involved in the development of pancreatic epithelial tumors and could represent a potential therapeutic target for PC.

Detailed Superiority of the CAD EYE Artificial Intelligence System over Endoscopists for Lesion Detection and Characterization Using Unique Movie Sets.

Objectives: Detailed superiority of CAD EYE (Fujifilm, Tokyo, Japan), an artificial intelligence for polyp detection/diagnosis, compared to endoscopists is not well examined. We examined endoscopist's ability using movie sets of colorectal lesions which were detected and diagnosed by CAD EYE accurately. Methods: Consecutive lesions of ≤10 mm were examined live by CAD EYE from March-June 2022 in our institution. Short unique movie sets of each lesion with and without CAD EYE were recorded simultaneously using two recorders for detection under white light imaging (WLI) and linked color imaging (LCI) and diagnosis under blue laser/light imaging (BLI). Excluding inappropriate movies, 100 lesions detected and diagnosed with CAD EYE accurately were evaluated. Movies without CAD EYE were evaluated first by three trainees and three experts. Subsequently, movies with CAD EYE were examined. The rates of accurate detection and diagnosis were evaluated for both movie sets. Results: Among 100 lesions (mean size: 4.7±2.6 mm; 67 neoplastic/33 hyperplastic), mean accurate detection rates of movies without or with CAD EYE were 78.7%/96.7% under WLI (p<0.01) and 91.3%/97.3% under LCI (p<0.01) for trainees and 85.3%/99.0% under WLI (p<0.01) and 92.6%/99.3% under LCI (p<0.01) for experts. Mean accurate diagnosis rates of movies without or with CAD EYE for BLI were 85.3%/100% for trainees (p<0.01) and 92.3%/100% for experts (p<0.01), respectively. The significant risk factors of not-detected lesions for trainees were right-sided, hyperplastic, not-reddish, in the corner, halation, and inadequate bowel preparation. Conclusions: Unique movie sets with and without CAD EYE could suggest it's efficacy for lesion detection/diagnosis.

Deep learning and digital pathology powers prediction of HCC development in steatotic liver disease.

BACKGROUND AND AIMS: Identifying patients with steatotic liver disease who are at a high risk of developing HCC remains challenging. We present a deep learning (DL) model to predict HCC development using hematoxylin and eosin-stained whole-slide images of biopsy-proven steatotic liver disease. APPROACH AND RESULTS: We included 639 patients who did not develop HCC for ≥7 years after biopsy (non-HCC class) and 46 patients who developed HCC <7 years after biopsy (HCC class). Paired cases of the HCC and non-HCC classes matched by biopsy date and institution were used for training, and the remaining nonpaired cases were used for validation. The DL model was trained using deep convolutional neural networks with 28,000 image tiles cropped from whole-slide images of the paired cases, with an accuracy of 81.0% and an AUC of 0.80 for predicting HCC development. Validation using the nonpaired cases also demonstrated a good accuracy of 82.3% and an AUC of 0.84. These results were comparable to the predictive ability of logistic regression model using fibrosis stage. Notably, the DL model also detected the cases of HCC development in patients with mild fibrosis. The saliency maps generated by the DL model highlighted various pathological features associated with HCC development, including nuclear atypia, hepatocytes with a high nuclear-cytoplasmic ratio, immune cell infiltration, fibrosis, and a lack of large fat droplets. CONCLUSIONS: The ability of the DL model to capture subtle pathological features beyond fibrosis suggests its potential for identifying early signs of hepatocarcinogenesis in patients with steatotic liver disease.

Differential Effects of Genetic Polymorphism on Comorbid Disease in Metabolic Dysfunction-Associated Steatotic Liver Disease.

BACKGROUND & AIMS: PNPLA3 rs738409, TM6SF2 rs58542926, and HSD17B13 rs72613567 have been associated with an increased risk of liver-related events (LREs) in patients with metabolic dysfunction-associated steatotic liver disease (MASLD). In this study, we investigated the combined effects of these variants on LREs. METHODS: The longitudinal multicenter cohort study enrolled 1178 patients with biopsy-proven MASLD. We calculated the genetic risk of hepatic fibrosis and LRE according to the impact of these variants. RESULTS: Patients with genetic fibrosis scores of 2, 3, and 4 or 5 were at greater risk than patients with scores of 0 or 1, with odds ratios of 2.45 (95% CI, 1.27-4.74), 2.14 (95% CI, 1.17-3.94), and 2.54 (95% CI, 1.35-4.77), respectively. Multivariate analysis revealed that PNPLA3 and TM6SF2, but not HSD17B13, were associated significantly with LRE development. The hazard ratio of the genetic high-risk group for LRE was 1.91 (95% CI, 1.20-3.04). The higher risk of LRE development in the genetic high-risk group also was seen in patients with F ≥ 3 or Fibrosis-4 index > 2.67. The hazard ratios of the genetic high-risk group for LRE were greater in patients without obesity, without diabetes, and of younger age compared with patients with obesity, with diabetes, or of older age, respectively. CONCLUSIONS: This combination of MASLD-related genetic variants is useful for predicting LREs in Japanese patients with MASLD. The genetic risk according to these variants is useful for LRE risk assessment, especially in patients without metabolic risk factors or in younger patients in Japan.