Impact of a real-time automatic quality control system on colorectal polyp and adenoma detection: a prospective randomized controlled study (with videos).

BACKGROUND AND AIMS: Quality control can decrease variations in the performance of colonoscopists and improve the effectiveness of colonoscopy to prevent colorectal cancers. Unfortunately, routine quality control is difficult to carry out because a practical method is lacking. The aim of this study was to develop an automatic quality control system (AQCS) and assess whether it could improve polyp and adenoma detection in clinical practice. METHODS: First, we developed AQCS based on deep convolutional neural network models for timing of the withdrawal phase, supervising withdrawal stability, evaluating bowel preparation, and detecting colorectal polyps. Next, consecutive patients were prospectively randomized to undergo routine colonoscopies with or without the assistance of AQCS. The primary outcome of the study was the adenoma detection rate (ADR) in the AQCS and control groups. RESULTS: A total of 659 patients were enrolled and randomized. A total of 308 and 315 patients were analyzed in the AQCS and control groups, respectively. AQCS significantly increased the ADR (0.289 vs 0.165, P < .001) and the mean number of adenomas per procedure (0.367 vs 0.178, P < .001) compared with the control group. A significant increase was also observed in the polyp detection rate (0.383 vs 0.254, P = .001) and the mean number of polyps detected per procedure (0.575 vs 0.305, P < .001). In addition, the withdrawal time (7.03 minutes vs 5.68 minutes, P < .001) and adequate bowel preparation rate (87.34% vs 80.63%, P = .023) were superior for the AQCS group. CONCLUSIONS: AQCS could effectively improve the performance of colonoscopists during the withdrawal phase and significantly increase polyp and adenoma detection. (Clinical trial registration number: NCT03622281.).

Triglyceride-Glucose Index is an Independent Risk Factor for Hepatocellular Carcinoma Development in Patients with HBV-Related Liver Cirrhosis.

Aim: This study aimed to explore the effects of the triglyceride-glucose (TyG) index on hepatocellular carcinoma (HCC) development in patients with hepatitis B virus (HBV)-related liver cirrhosis (LC). Methods: A total of 242 patients with HBV-related LC were enrolled and followed-up. Logistic regression analysis was performed to investigate risk factors for HCC. Results: The median follow-up time was 37 months (range: 6-123 months). At the end of the follow-up, 11 (11.3%) patients with compensated cirrhosis (CC) and 45 (31.0%) with decompensated cirrhosis (DC) developed HCC. The TyG index was higher in the HCC group than in the non-HCC group (P=0.05). Univariate analysis showed that age (P<0.01), DC (P<0.01), TyG index (P=0.08), albumin (ALB) level (P=0.05), platelet (PLT) count (P<0.01), and HBV DNA positivity (P<0.01) were associated with HCC development. Multivariate analysis revealed that age, DC, TyG index, PLT count, and HBV DNA positivity were independent risk factors for HCC development (P=0.01, 0.01, <0.01, 0.05, and <0.01, respectively). For patients with DC, multivariate logistic regression analysis revealed that age, TyG index, and HBV DNA positivity were independent risk factors for HCC development (all P<0.05). A new model encompassing age, DC, TyG, PLT, and positive HBV DNA had optimal predictive accuracy in patients with DC or CC, with a cutoff value of 0.197. The areas under the receiver operating characteristic curves (AUROCs) of the model for predicting HCC development in patients with LC, DC, and CC were 0.778, 0.721, and 0.783, respectively. Conclusion: TyG index was identified as an independent risk factor for HCC development in patients with LC.

D-dimer is a prognostic marker for 1-year mortality in patients with chronic liver failure and hepatic encephalopathy.

Background and Aim: Hepatic encephalopathy (HE) is a neuropsychiatric complication of liver failure with poor outcomes. The present study aimed to evaluate the predictive values of D-dimer in patients with HE. Materials and Methods: Patients with chronic liver failure (CLF) and HE were enrolled. Univariate and multivariate logistic analysis was performed to investigate the risk factors for 1-year mortality of HE. Results: During the first year after diagnosis, 39.2% (65/166) of the patients died. D-dimer was significantly higher in non-survivors (Z=2.617, p<0.01). Both D-dimer and international normalized ratio (INR) positively correlated with Child-Pugh and MELD scores, and negatively correlated with sodium (all p<0.01). Moreover, there was a negative relationship between D-dimer and HE grades (r=-0.168, p=0.031), while the relationship between INR and HE grades was not significant (r=0.083, p=0.289). Multivariate analysis showed that age (odds ratio (OR):1.035, 95% CI:1.004-1.067, p=0.03), D-dimer (OR=1.138, 95% CI:1.030-1.258, p=0.01), ALT (OR=1.012, 95% CI:1.001-1.022, p=0.03), and sodium (OR=0.920, 95% CI:0.858-0.986, p=0.02) were independent risk factors for 1-year mortality. Then, a new model Model(Age_DD_ALT_Na) incorporating age, D-dimer, ALT, and sodium was developed. AUROC of Model(Age_DD_ALT_Na) was 0.732, which was significantly higher than MELD and Child-Pugh scores (AUROC: 0.602 and 0.599, p=0.013 and 0.022). Conclusion: D-dimer is a prognostic marker for 1-year mortality in patients with CLF and HE.

Estimated glomerular filtration rate partially recovered in patients without Hypertriglyceridemia during Tenofovir disoproxil fumarate treatment.

BACKGROUND: Decrease in estimated glomerular filtration rate (eGFR) during Tenofovir disoproxil fumarate (TDF) treatment remains a concern, and few patients experience partial recovery of eGFR. This study aimed to investigate the risk factors for eGFR recovery in patients with and without hypertriglyceridemia. METHODS: A total of 203 patients with chronic HBV infection were prospectively recruited and followed up for three years. Data were collected at baseline, first, second, and third years during TDF treatment. RESULTS: Most patients achieved normal ALT (80.0% vs. 82.5%) and undetectable HBV DNA (95.0% vs. 95.6%) in both groups (p > 0.05). For patients with hypertriglyceridemia, eGFR and cholesterol did not change significantly during the 3-year follow-up, while triglyceride (TG) decreased significantly in the first year and persisted at a lower level in the subsequent two years. For patients without hypertriglyceridemia, eGFR declined significantly in the first year of treatment, then gradually recovered during the subsequent two years, and eGFR was negatively correlated with TG at the four time points. Fifteen (15/183, 8.2%) patients without hypertriglyceridemia experienced eGFR partial recovery in the third year. Univariate and multivariate analyses showed that baseline eGFR <90 mL/(min·1.73 m2) (p < 0.01; 95% CI: 0.019-0.284) and age (p < 0.01; 95% CI: 0.817-0.960) were independent risk factors for eGFR recovery. CONCLUSION: eGFR partially recovered in patients without hypertriglyceridemia during TDF treatment, and TG regulation might be a useful strategy to hinder renal function decline, although larger, confirmatory studies are necessary to validate our findings.Key messagesFor patients with normal triglyceride, eGFR declined significantly at the first year of TDF treatment, then gradually recovered during the subsequent two years, and eGFR was negatively correlated with TG. Baseline eGFR <90 mL/(min·1.73 m2) and age were independent risk factors for eGFR recovery.

Dynamic changes of estimated glomerular filtration rate are conversely related to triglyceride in non-overweight patients.

BACKGROUND: Correlation between Triglyceride (TG) and estimated glomerular filtration rate (eGFR) remains largely unknown in overweight and non-overweight patients. AIM: To investigated the dynamic changes of eGFR and lipid profiles during 3-year tenofovir disoproxil fumarate (TDF) treatment in patients with chronic hepatitis B (CHB) and overweight. METHODS: A total of 202 CHB patients who received TDF treatment at the Third People's Hospital of Changzhou (Changzhou, China) and Nanjing Drum Tower Hospital (Nanjing, China) between January 2016 and May 2018 were retrospectively enrolled. According to the body mass index (BMI) at the initiation of TDF treatment, CHB patients were divided into overweight (BMI ≥ 25 kg/m2) and non-overweight (BMI < 25 kg/m2) groups. Logistic regression was applied for the analysis of risk factors for eGFR < 90 mL/(min·1.73 m2). RESULTS: There is no significant difference in hepatitis B virus DNA (HBV DNA) negativity and hepatitis Be antigen (HBeAg) loss between patients with overweight and non-overweight (both P > 0.05). More patients in non-overweight group achieved alanine aminotransferase normalization compared with those in overweight group (χ 2 = 11.036, P < 0.01). In non-overweight patients, the eGFR significantly declined in the 1st year (P < 0.01), then remained at a relatively lower level. TG significantly declined in the 2nd year (P = 0.02) and increased in the 3rd year. Moreover, TG was negatively correlated with GFR at the four-time points (P = 0.002, 0.030, 0.007, 0.008, respectively). In overweight patients, eGFR and TG remained relatively stable during the 3-year treatment, and eGFR showed no significant relationship with TG. Moreover, multivariate analysis showed that age [P < 0.01, 95%CI (0.97-1.005)] and baseline eGFR [P < 0.01, 95%CI (5.056-33.668)] were independent risk factors for eGFR < 90 mL/(min·1.73 m2) at the 3rd year. CONCLUSION: Dynamic changes in renal function were conversely related to TG during TDF treatment in patients with CHB and normal BMI, but not with overweight.

Application of artificial intelligence in gastrointestinal endoscopy.

With recent significant improvements in artificial intelligence (AI), especially in the field of deep learning, an increasing number of studies have evaluated the use of AI in endoscopy to detect and diagnose gastrointestinal (GI) lesions. The present review summarizes current publications on the use of AI in GI endoscopy and focuses on the challenges and future of AI-aided systems. We expect AI to provide an effective and practical method for endoscopists in lesion detection and characterization as well as in quality control in endoscopy. However, so far, most studies have remained at the preclinical stage. More attention should be paid in the future to the use of AI in real-life clinical applications.