Artificial intelligence-assisted colonoscopy to identify histologic remission and predict the outcomes of patients with ulcerative colitis: A systematic review

This systematic review evaluated the current status of AI-assisted colonoscopy to identify histologic remission and predict the clinical outcomes of patients with ulcerative colitis. The use of artificial intelligence (AI) has increased substantially across several medical fields, including gastrointestinal endoscopy. Evidence suggests that it may be helpful to predict histologic remission and relapse, which would be beneficial because current histological diagnosis is limited by the inconvenience of obtaining biopsies and the high cost and time-intensiveness of pathological diagnosis. MEDLINE and the Cochrane Central Register of Controlled Trials were searched for studies published between January 1, 2000, and October 31, 2023. Nine studies fulfilled the selection criteria and were included; five evaluated the prediction of histologic remission, two assessed the prediction of clinical outcomes, and two evaluated both. Seven were prospective observational or cohort studies, while two were retrospective observational studies. No randomized controlled trials were identified. AI-assisted colonoscopy demonstrated sensitivity between 65 %-98 % and specificity values of 80 %-97 % for identifying histologic remission. Furthermore, it was able to predict future relapse in patients with ulcerative colitis. However, several challenges and barriers still exist to its routine clinical application, which should be overcome before the true potential of AI-assisted colonoscopy can be fully realized.

Effectiveness of transabdominal ultrasonography in predicting clinical relapse of Crohn's disease

BACKGROUND/AIMS

Transabdominal ultrasonography (US) helps evaluate Crohn's disease (CD) activity. We investigated whether the US could predict subsequent adverse outcomes for patients with CD in clinical remission.

METHODS

This single-center retrospective study included patients with CD in clinical remission who underwent US between April 2011 and April 2021, focusing on the predictability of subsequent adverse outcomes within 5 years. We used the US-CD, which was calculated using multiple US findings. Predictive variables were assessed using Cox proportional hazards regression analysis, and the predictive value was evaluated using receiver operating characteristic curves.

RESULTS

Seventy-three patients were included. During a median follow-up of 1,441 days (range, 41-1,825 days), 16.4% (12/73) experienced clinical relapse, 9.6% (7/73) required endoscopic balloon dilation (EBD), 58.9% (43/73) required enhanced treatment, and 20.5% (15/73) underwent surgery. In the multivariate analysis, US-CD was significantly associated with clinical relapse (P= 0.038) and the need for enhanced treatment (P= 0.005). The area under the receiver operating characteristic curve for predicting clinical relapse and the need for EBD was 0.77 and 0.81, respectively, with US-CD (cutoff value = 11), and that for requiring enhanced treatment was 0.74 with US-CD (cutoff value = 6). Patients with US-CD ≥ 11 demonstrated a significantly higher occurrence of clinical relapse (P= 0.001) and EBD (P= 0.002) within 5 years. Patients with US-CD ≥ 6 experienced a significantly higher likelihood of requiring enhanced treatment (P< 0.001) within 5 years.

CONCLUSIONS

High US-CD is associated with subsequent adverse outcomes in patients with CD.

Long-term comparisons of the durability of 6 months versus 12 months antiviral therapy for hepatitis B after chemotherapy cessation

BACKGROUND

Prophylaxis antiviral therapy is recommended for patients with hepatitis B receiving chemotherapy but the ideal treatment duration after chemotherapy cessation needs more evidence for clarification.

AIMS

This study aimed to compare the relapse rate of short finite intervals of 6 months and 12 months of -nucleos(t)ide analogue (NA) therapy in patients stratified by low hepatitis B virus (HBV)-DNA of < 2000 IU/ml or high HBV DNA of ≥ 2000 IU/ml.

METHODS

Patients started tenofovir or entecavir treatment 1 week before chemotherapy and were assigned to different treatment duration groups randomly after stratified by HBV DNA pretreatment: (1) HBV DNA of < 2000 IU/ml at 6-month or 12-month duration; (2)HBV DNA of ≥ 2000 IU/ml at 6-month or 12-month duration. Virological relapse (VR) was defined as HBV DNA of > 2000 IU/ml, and clinical relapse (CR) was defined as HBV DNA of > 2000 IU/ml and alanine aminotransferase of > 80 IU/L during the follow-up period. The primary endpoint was to compare the durability between groups 1 year after antiviral therapy cessation. The secondary endpoint was VR and CR rate at long-term follow-up after antiviral therapy cessation.

RESULTS

This study enrolled 61 patients, and 5 patients were lost to follow-up or tumor recurrence. VR and CR rates were 46.4% and 14.3% at 1-year and 55.3% and 16.1%, at long-term follow-up, respectively. VR and CR rates demonstrated no difference between the groups. Pretreatment HBV DNA at ≥ 2000 IU/ml and end-of-treatment hepatitis B surface antigen (HBsAg) at ≥ 500 IU/ml were the predictor of VR (hazard ratio [HR]: 2.98; p = 0.010 and HR: 2.38; p = 0.037).

CONCLUSIONS

Prolongation from 6 months to 12 months of NA consolidation after chemotherapy cessation did not affect the VR or CR of HBV. High pretreatment HBV DNA and end-of-treatment HBsAg levels could predict VR after antiviral therapy cessation for chemotherapy.

HBV relapse rates in patients who discontinue tenofovir disoproxil fumarate with or without switching to tenofovir alafenamide

BACKGROUND/AIMS

The incidence and relapse pattern in patients stopping tenofovir alafenamide (TAF), a prodrug of tenofovir which is more concentrated in hepatocytes, is unknown.

METHODS

HBeAg-negative CHB patients stopping tenofovir disoproxil fumarate (TDF) (off-TDF) or who had switched to TAF more than 3 months before discontinuation (off-TAF) were recruited. The propensity score-matching method (PSM) was used, creating a ratio of 1:3 between the off-TAF versus the off-TDF groups to adjust for associated factors.

RESULTS

After PSM, 180 off-TDF and 60 off-TAF patients were analyzed. The cumulative rates of virological and clinical relapse at 52 weeks were 75.1% and 58.5% respectively in the off-TDF group and 91.1% and 61.6% in the off-TAF group. Patients in the off-TAF group had significantly higher rates of virological relapse than those in the off-TDF group (p = 0.021), but not clinical relapse (p = 0.785). Multivariate cox regression analysis showed that off-TAF group was an independent factor for virological relapse, but not clinical relapse. Severity of clinical relapse and hepatic decompensation rate were comparable between off-TDF and off-TAF groups CONCLUSIONS: The off-TAF group had a higher virological relapse rate than the off-TDF group. The difference in clinical relapse pattern and severity was not clinically important between the two groups.

Predictors of Virological Suppression After Clinical Relapse in Patients Who Discontinued Entecavir or Tenofovir

BACKGROUND

The predictors of persistent virological suppression after clinical relapse remain unclear.

AIMS

To investigate the predictors of retreatment or persistent virological suppression after clinical relapse in chronic hepatitis B (CHB) patients who discontinued entecavir or tenofovir disoproxil fumarate (TDF).

METHODS

A total of 243 hepatitis B e antigen-negative CHB patients without cirrhosis who experienced clinical relapse after entecavir or TDF cessation were enrolled.

RESULTS

Of the 243 CHB patients, 192 received retreatment and 51 did not receive retreatment after clinical relapse. Of the 51 patients without retreatment, 23 achieved persistent virological suppression (persistent HBV DNA < 2000 IU/mL at least 2 years) and 10 experienced hepatitis B surface antigen (HBsAg) loss. The Cox regression analysis showed that short consolidation duration, short duration of the first clinical relapse from the end of treatment (EOT), and high bilirubin and HBV DNA levels at the first clinical relapse were independent predictors of retreatment. Long duration of the first clinical relapse from the EOT and low HBsAg levels at the first clinical relapse were independent factors of patients with persistent virological suppression. The rates of persistent virological suppression at the first clinical relapse among patients with HBsAg < 100 and ≥ 100 IU/mL were 44.4% (12/27) and 5.1% (11/216) (P < 0.001), respectively. Baseline HBsAg levels and no retreatment requirement were independent factors associated with HBsAg loss.

CONCLUSIONS

The HBsAg of 100 IU/mL at the first clinical relapse could predict persistent virological suppression after clinical relapse in patients who discontinued entecavir or TDF therapy.

Combining end-of-treatment HBsAg and baseline hepatitis B core-related antigen reduce HBV relapse rate after tenofovir cessation

BACKGROUND/PURPOSE

The study investigated the role of hepatitis B core-related antigen (HBcrAg) in hepatitis B virus (HBV) relapse after stopping tenofovir disoproxil fumarate (TDF) in HBeAg-negative patients.

METHODS

A total of 185 HBeAg-negative patients without cirrhosis who had stopped TDF treatment for at least 6 months were recruited. All patients fulfilled the stopping criteria proposed by the Asian Pacific Association for the Study of the Liver 2012.

RESULTS

The 3-year cumulative incidences of virological relapse, clinical relapse, and hepatitis B surface antigen (HBsAg) loss were 72, 60.1 and 14.5%, respectively. End-of-treatment (EOT) HBsAg level was an independent predictor of virological relapse (hazard ratio (HR): 2.263; 95% confidence interval (CI): 1.779-2.887), clinical relapse (HR 1.773; 95% CI 1.367-2.298), and HBsAg loss (HR 0.179; 95% CI 0.096-0.335). Among patients who had HBsAg < 100 and ≥ 100 IU/mL, the 3-year virological relapse rates were 37.4% and 85.3% (p < 0.001), clinical relapse rates were 30.3 and 71.7% (p < 0.001), and HBsAg loss rates were 40.6 and 2.6% (p < 0.001), respectively. Among the 53 patients with EOT HBsAg level < 100 IU/mL, the 3-year virological relapse rates in patients with baseline HBcrAg levels < 4.7 and ≥ 4.7 log₁₀ U/mL were 20.3 and 60.4% (p = 0.003), and the clinical relapse rates were 10.3 and 59.5% (p < 0.001) respectively. Additionally, the 3-year HBsAg loss rates in patients with baseline HBcrAg ≤ 3 and > 3 log₁₀ U/mL were 42.9 and 7.9% (p < 0.001).

CONCLUSIONS

The combination of EOT HBsAg and baseline HBcrAg levels could further reduce the risk of HBV relapse after stopping TDF therapy in HBeAg-negative patients.

Further research on the clinical relevance of the ulcerative colitis colonoscopic index of severity for predicting 5-year relapse

PURPOSE

The ulcerative colitis colonoscopic index of severity (UCCIS) evaluates the state of the entire colonic mucosa in ulcerative colitis. However, no cut-off values of scores for predicting clinical relapse in patients with ulcerative colitis have been established. This study aimed to determine the cut-off values for predicting clinical relapse in patients with ulcerative colitis.

METHODS

The endoscopic scores (sum of Mayo endoscopic subscores (S-MES) and UCCIS) of 157 patients with ulcerative colitis experiencing clinical remission and their subsequent clinical course were retrospectively reviewed. The optimal cut-off values for predicting relapse and relapse-free rates were analyzed by receiver operating characteristic analysis.

RESULTS

Forty patients with ulcerative colitis experienced relapse within 24 months. The median UCCIS for these patients at the time of study enrollment was significantly higher than that for patients with clinical remission (P < 0.001). The cut-off value of the UCCIS for predicting relapse was 9.8. The relapse-free rate was significantly lower in patients with UCCIS ≥ 9.8 than in those with UCCIS < 9.8 (log-rank test P < 0.001). For patients who experienced relapse within 5 years, the optimal cut-off values for the UCCIS and S-MES were 10.2 and 1, respectively (P = 0.004).

CONCLUSIONS

The data from this study indicate that the USSIC is a more relevant score than the S-MES for predicting the time to relapse in patients with ulcerative colitis in remission.

Genetic variations in the CXCR5 gene decrease the risk of clinical relapse after discontinuation of nucleos(t)ide analogue therapy in patients with chronic hepatitis B

Discontinuation of nucleos(t)ide analogue (NA) therapy in patients with chronic hepatitis B (CHB) remains a global but controversial problem. Clinical outcomes of NA cessation depend on the interplay between viral factors and host immunity. Recent studies have shown that genetic polymorphisms might influence the immune response in chronic HBV infection. A total of 33 single-nucleotide polymorphisms (SNPs) from 16 genes (BCL6, CD40, CD40L, CTLA-4, CXCL13, CXCR5, ICOS, IL-21, HLA-C, NTCP, UBE2L3, STAT4, IFN-λ3, CYP27B1, INST10, and IPS1) were selected and analyzed in 106 CHB patients enrolled in an off-treatment cohort. Significantly unbalanced distributions between patients who experienced clinical relapse and those who did not were found regarding two SNPs, rs676925 in CXCR5 and rs733618 in CTLA-4. Furthermore, the genotype 'GC' of rs676925 were associated with decreased risk of clinical relapse, implicating that rs676925 may serve as a protective factor for HBV control and facilitate a virus-specific immune response. We also compared the expression of CXCR5 in lymphocytes and its ligand CXCL13 in plasma between different genotypes of rs676925. However, no significant differences were observed. In conclusion, this study suggested that the rs676925 'GC' genotype of the CXCR5 gene were associated with decreased risk of clinical relapse after discontinuation of long-term NA therapy in CHB patients.

Finite nucleos(t)ide analog therapy in HBeAg-negative chronic hepatitis B: an emerging paradigm shift. Hepatol Int. 2019;13:665-673. [PMID: 31559604 DOI: 10.1007/s12072-019-09989-6]

Potent nucleos(t)ide analogs (NUC), such as entecavir and tenofovir disoproxil fumarate, are able to suppress HBV DNA to undetectable level. These agents have no direct action on cccDNA, which is a very stable template for HBV production, hence long-term or even life-long NUC therapy is required in HBeAg-negative patients to maintain HBV suppression and to achieve the ultimate goal of HBsAg loss. However, there are concerns of indefinite or life-long NUC therapy, including drug resistance, financial burden, adherence and willingness for indefinite long-term NUC therapy. Patients lost to follow-up and hence, not monitored may risk severe relapse that may deteriorate to hepatic decompensation or even hepatic failure. This Review integrated the cumulated evidence and assessed the strategy of finite NUC therapy in HBeAg-negative patients which was first tried in early 2000s. Earlier Asian findings that 2-year NUC therapy is feasible and safe have founded APASL stopping rule for patients on NUC therapy over 2-3 years since 2008. Subsequent studies have supported the strategy of finite NUC therapy, which has finally been accepted as an option by American and European liver associations since 2016. More recent studies have further shown greatly increased HBsAg loss rate (up to 5-year 39%) after stopping NUC therapy. The cumulated evidence has shown that the paradigm shift from indefinite long-term therapy to finite NUC therapy in HBeAg-negative patients is emerging. More studies are needed to fine-tuning the strategy including research for the optimal duration of consolidation therapy, timing to stop and to start re-treatment.

Elevated expression of serum soluble ST2 in clinical relapse after stopping long-term Nucleos(t)ide analogue therapy for chronic hepatitis B

Background

The virological or clinical relapse is common in chronic hepatitis B (CHB) patients after stopping long-term nucleos(t)ide analogue (NA) therapy. Soluble growth stimulation expressed gene 2 (sST2), one of the Toll-like/interleukin-1 receptor members, is involved in a variety of inflammatory processes and immune responses. However, the expression and function of serum sST2 in CHB patients after stopping NA treatment remains unknown.

Methods

A total of 91 non-cirrhotic Asian patients with CHB who discontinued NA therapy according to international guidelines were prospectively followed up to 240 weeks. All patients were divided into clinical relapse group and non-clinical relapse (including sustained virological response and only virological relapse) group according HBV DNA and ALT levels. The serum levels of sST2 of all participants were determined by ELISA and compared between each two groups.

Results

Clinical relapse occurred in 26 patients and virological relapse occurred in 57 patients. We found that there was a positive correlation between sST2 expression and HBsAg, ALT, HBV DNA, and anti-HBc levels in CHB patients after discontinuation of NA treatment. Levels of serum sST2 in clinical relapse patients showed a rising trend and most patients showed peak sST2 levels at the point of clinical relapse. Moreover, the sST2 levels of clinical relapse group at week 12, week 24 and week 48 were relatively higher than non-clinical relapse group. However, the level of sST2 at the end of treatment was not an effective biological marker for the early prediction of clinical relapse after discontinuation of long-term NA therapy.

Conclusions

In conclusion, we found that an increase in sST2 in clinical relapse patients might be associated with an inflammation-related immune response after discontinuation of NA treatment.

Trial registration

The trial was retrospectively registered at Chinese Clinical Trial Registry: ChiCTR-OOC-17013970. Registration date: December 15, 2017.

Plasma CXCL13 is a predictive factor for HBsAg loss and clinical relapse after discontinuation of nucleos(t)ide analogue treatment

In this study, we investigated whether plasma cytokine/chemokine levels could predict HBsAg loss or clinical relapse (CR) after stopping nucleos(t)ides analogue (NA) treatment. Theplasma cytokines/chemokines levels were measured at 0, 4, 8, 12, 24 and 48 weeks after NA discontinuation by using the enzyme-linked immunoassay (ELISA) kit. Cox regression analysis revealed that CXCL13 level at the end of treatment (EOT) was an independent predictor for CR (HR 0.26, p < 0.001) and HBsAg loss (HR 3.01, p = 0.008) after treatment cessation. Among the patients with EOT CXCL13 level < 80 pg/ml, the cumulative incidences of CR and HBsAg loss were 65% and 0% at 4 years, respectively. As for the patients with EOT CXCL13 level ≥ 1000 pg/ml, 47.5% cases had HBsAg loss. Our study showed that EOT CXCL13 level was associated with off-treatment response, which may be used to guide cessation of NA treatment in clinical practice.

An Evaluation of Lag Schedules of Reinforcement During Functional Communication Training: Effects on Varied Mand Responding and Challenging Behavior

We evaluated the effects of lag schedules of reinforcement during functional communication training (FCT) on the varied use of mands and challenging behavior by two individuals diagnosed with autism. Specifically, we compared the effects of Lag 0 and Lag 1 schedules of reinforcement during FCT. The results showed that each participant exhibited increases in varied mand responding during FCT with the Lag 1 schedule of reinforcement relative to Lag 0; challenging behavior remained low during both FCT lag conditions relative to baseline. Results are discussed in terms of treatment implications relating to FCT and the potential prevention and/or mitigation of clinical relapse during challenges to treatment.

High-mobility group box 1 in multiple sclerosis

This study is one in series determining the potential of RAGE axis (receptor for advanced glycation end products, isoforms, ligands) as a biomarker in multiple sclerosis (MS). We evaluated serum levels of RAGE ligand, the high-mobility group box (HMGB)1 in MS patients, and assessed the correlation between HMGB1 serum levels and the use of disease-modifying drugs (DMDs), and between HMGB1 serum levels and indicators of MS disease severity. HMGB1 serum levels were compared between 96 (23 males) MS patients and 34 age- and gender-matched healthy controls (HCs) using enzyme-linked immunosorbent assays. DMD-naïve MS patients had significantly higher HMGB1 serum levels compared with DMD-treated (P = 0.04) and compared with HCs (P = 0.01). HMGB1 serum levels were not significantly different between total MS patients (DMD-naïve plus DMD-treated) and HCs (P = 0.09). DMD-naïve MS patients in clinical relapse tended to have lower HMGB1 serum levels than clinically stable RRMS patients (P = 0.07). HMGB1 serum levels showed 0.65 area under the curve (95 % CI 0.55-0.95) sensitivity/specificity for MS clinical relapse. The role of HMGB1 in MS disease pathology and DMD modulation of this protein warrant further investigations.

Effect of mucosal healing (Mayo 0) on clinical relapse in patients with ulcerative colitis in clinical remission

OBJECTIVE

The aim of this study was to identify the effect of mucosal healing (MH) on clinical relapse in patients with ulcerative colitis (UC) who are in clinical remission, with special reference to Mayo endoscopic subscore 0.

METHODS

Between November 2005 and December 2013, medical records from a total of 215 patients with UC who underwent colonoscopic examination at the time of clinical remission were retrospectively reviewed. Endoscopic MH was defined as a '0 point' of Mayo endoscopic subscore (Mayo 0). Patients were categorized into two groups according to Mayo endoscopic subscore and then analyzed.

RESULTS

The baseline characteristics of both groups (MH vs. no-MH), including age at diagnosis, gender, and initial clinical and colonoscopic findings, were not significantly different. The median follow-up duration was 80 (12-118) months. Factors predictive of longer clinical remission duration were age ≥30 years at diagnosis (≥30 years vs. <30 years; hazard ratio [HR] 3.16, 95% CI 1.88-5.30, p < 0.001), shorter interval between diagnosis and clinical remission (<15 months vs. ≥15 months; HR 1.93, 95% CI 1.13-3.28, p = 0.015), and presence of MH at clinical remission (HR 1.95, 95% CI 1.15-3.32, p = 0.014). With a Cox regression model, patients with MH at clinical remission were more likely to have longer duration of clinical remission than patients without MH.

CONCLUSION

The achievement of MH, Mayo 0 in particular, in patients with UC who are in clinical remission is important in predicting a favorable disease course prognosis.

How to predict clinical relapse in inflammatory bowel disease patients

Inflammatory bowel diseases have a natural course characterized by alternating periods of remission and relapse. Disease flares occur in a random way and are currently unpredictable for the most part. Predictors of benign or unfavourable clinical course are required to facilitate treatment decisions and to avoid overtreatment. The present article provides a literature review of the current evidence on the main clinical, genetic, endoscopic, histologic, serologic and fecal markers to predict aggressiveness of inflammatory bowel disease and discuss their prognostic role, both in Crohn’s disease and ulcerative colitis. No single marker seems to be reliable alone as a flare predictor, even in light of promising evidence regarding the role of fecal markers, in particular fecal calprotectin, which has reported good results recently. In order to improve our daily clinical practice, validated prognostic scores should be elaborated, integrating clinical and biological markers of prognosis. Finally, we propose an algorithm considering clinical history and biological markers to intercept patients with high risk of clinical relapse.