Evaluation of machine learning algorithms for predicting direct-acting antiviral treatment failure among patients with chronic hepatitis C infection.

Despite the availability of efficacious direct-acting antiviral (DAA) therapy, the number of people infected with hepatitis C virus (HCV) continues to rise, and HCV remains a leading cause of liver-related morbidity, liver transplantation, and mortality. We developed and validated machine learning (ML) algorithms to predict DAA treatment failure. Using the HCV-TARGET registry of adults who initiated all-oral DAA treatment, we developed elastic net (EN), random forest (RF), gradient boosting machine (GBM), and feedforward neural network (FNN) ML algorithms. Model performances were compared with multivariable logistic regression (MLR) by assessing C statistics and other prediction evaluation metrics. Among 6525 HCV-infected adults, 308 patients (4.7%) experienced DAA treatment failure. ML models performed similarly in predicting DAA treatment failure (C statistic [95% CI]: EN, 0.74 [0.69-0.79]; RF, 0.74 [0.69-0.80]; GBM, 0.72 [0.67-0.78]; FNN, 0.75 [0.70-0.80]), and all 4 outperformed MLR (C statistic [95% CI]: 0.51 [0.46-0.57]), and EN used the fewest predictors (n = 27). With Youden index, the EN had 58.4% sensitivity and 77.8% specificity, and nine patients were needed to evaluate to identify 1 DAA treatment failure. Over 60% treatment failure were classified in top three risk decile subgroups. EN-identified predictors included male sex, treatment < 8 weeks, treatment discontinuation due to adverse events, albumin level < 3.5 g/dL, total bilirubin level > 1.2 g/dL, advanced liver disease, and use of tobacco, alcohol, or vitamins. Addressing modifiable factors of DAA treatment failure may reduce the burden of retreatment. Machine learning algorithms have the potential to inform public health policies regarding curative treatment of HCV.

Sustainable and equivalent improvements in symptoms and functional well-being following viral cure from ledipasvir/sofosbuvir versus elbasvir/grazoprevir for chronic hepatitis C infection: Findings from the randomized PRIORITIZE trial.

The PRIORITIZE trial (clinicaltrials.gov: NCT02786537) was the first comparative effectiveness study to directly compare ledipasvir/sofosbuvir (LDV/SOF) and elbasvir/grazoprevir (EBR/GZR) for the treatment of chronic hepatitis C virus (HCV). A secondary aim of this study was to compare LDV/SOF and EBR/GZR on sustainable changes in several HCV-associated symptoms and functional well-being in patients who achieved sustained virological response (SVR). PRIORITIZE, a randomized controlled trial conducted between 2016 and 2020, evaluated change in six PROMIS® symptom scores (fatigue, sleep disturbance, cognitive disturbance, nausea, diarrhoea, abdominal pain) and functional well-being using the disease-specific HCV-PRO instrument. Survey assessments were administered at baseline, early post-treatment (median = 6 months) and late post-treatment (median = 21 months). Constrained longitudinal linear mixed-effects models were used to evaluate within-treatment change and between-treatment differences. Data from 793 participants (average 55 years old, 57% male, 44% black, 17% with cirrhosis) were analysed. From baseline to early post-treatment, 5 out of 6 symptoms and functional well-being significantly improved (all p's < .05). In the LDV/SOF arm, mean changes ranged from -3.73 for nausea to -6.41 for fatigue and in the EBR/GZR, mean changes ranged from -2.19 for cognitive impairment to -4.67 for fatigue. Change of >3 points was consider clinically meaningful. Improvements in most symptoms slightly favoured LDV/SOF, although the magnitude of differences between the regimens were small. Both regimens demonstrated significant improvements in symptoms and functional well-being that were sustained during the late post-treatment phase. EBR/GZR and LDV/SOF regimens had clinically equivalent and durable improvements in HCV symptoms and functional well-being up to two years after SVR.

Machine learning algorithms for predicting direct-acting antiviral treatment failure in chronic hepatitis C: An HCV-TARGET analysis.

BACKGROUND AND AIMS: We aimed to develop and validate machine learning algorithms to predict direct-acting antiviral (DAA) treatment failure among patients with HCV infection. APPROACH AND RESULTS: We used HCV-TARGET registry data to identify HCV-infected adults receiving all-oral DAA treatment and having virologic outcome. Potential pretreatment predictors (n = 179) included sociodemographic, clinical characteristics, and virologic data. We applied multivariable logistic regression as well as elastic net, random forest, gradient boosting machine (GBM), and feedforward neural network machine learning algorithms to predict DAA treatment failure. Training (n = 4894) and validation (n = 1631) patient samples had similar sociodemographic and clinical characteristics (mean age, 57 years; 60% male; 66% White; 36% with cirrhosis). Of 6525 HCV-infected adults, 95.3% achieved sustained virologic response, whereas 4.7% experienced DAA treatment failure. In the validation sample, machine learning approaches performed similarly in predicting DAA treatment failure (C statistic [95% CI]: GBM, 0.69 [0.64-0.74]; random forest, 0.68 [0.63-0.73]; feedforward neural network, 0.66 [0.60-0.71]; elastic net, 0.64 [0.59-0.70]), and all four outperformed multivariable logistic regression (0.51 [0.46-0.57]). Using the Youden index to identify the balanced risk score threshold, GBM had 66.2% sensitivity and 65.1% specificity, and 12 individuals were needed to evaluate to identify 1 DAA treatment failure. Over 55% of patients with treatment failure were classified by the GBM in the top three risk decile subgroups (positive predictive value: 6%-14%). The top 10 GBM-identified predictors included albumin, liver enzymes (aspartate aminotransferase, alkaline phosphatase), total bilirubin levels, sex, HCV viral loads, sodium level, HCC, platelet levels, and tobacco use. CONCLUSIONS: Machine learning algorithms performed effectively for risk prediction and stratification of DAA treatment failure.

A Pragmatic, Randomized Controlled Trial of Oral Antivirals for the Treatment of Chronic Hepatitis C: The PRIORITIZE Study.

BACKGROUND AND AIMS: Multiple direct-acting antiviral (DAA) regimens are available to treat HCV genotype 1 infection. However, comparative effectiveness from randomized controlled trials of DAA regimens is unavailable. APPROACH AND RESULTS: We conducted a pragmatic randomized controlled trial (NCT02786537) to compare the effectiveness of DAAs for HCV genotype 1a or 1b on viral response, safety, tolerability, and medication nonadherence. Adults with compensated liver disease, HCV genotype 1, not pregnant or breastfeeding, and with health insurance likely to cover ledipasvir/sofosbuvir (LDV/SOF) were recruited from 34 US viral hepatitis clinics. Participants were randomized (± ribavirin) to LDV/SOF, elbasvir/grazoprevir (EBR/GZR), and paritaprevir/ritonavir/ombitasvir+dasabuvir (PrOD; treatment arm stopped early). Primary outcomes included sustained viral response at 12 weeks (SVR12), clinician-recorded adverse events, patient-reported symptoms, and medication nonadherence. Between June 2016 and March 2018, 1,609 participants were randomized. Among 1,128 participants who received ≥1 dose of EBR/GZR or LDV/SOF (± ribavirin), SVR12 was 95.2% (95% CI, 92.8%-97.6%) and 97.4% (95% CI, 95.5%-99.2%), respectively, with a difference estimate of 2.2% (-0.5% to 4.7%), falling within the "equivalence" interval (-5% to 5%). While most (56%) participants experienced adverse events, few were serious (4.2%) or severe (1.8%). In the absence of ribavirin, discontinuations due to adverse events were rare. Patient-reported symptoms and medication nonadherence were similar. Study limitations were dropout due to insurance denial and loss to follow-up after treatment, limiting the ability to measure SVR12. CONCLUSIONS: This pragmatic trial demonstrated high SVR12 for participants treated with EBR/GZR and LDV/SOF with few adverse effects. Overall, the two regimens were equivalent in effectiveness. The results support current HCV guidelines that do not distinguish between ribavirin-free EBR/GZR and LDV/SOF.

Linkage of resistance-associated substitutions in GT1 sofosbuvir + NS5A inhibitor failures treated with glecaprevir/pibrentasvir.

BACKGROUND & AIMS: Retreatment with glecaprevir/pibrentasvir (G/P) resulted in a rate of sustained virologic response 12 weeks after treatment completion (SVR12) of >90% in HCV genotype 1 (GT1) patients who previously failed a regimen of sofosbuvir plus an NS5A inhibitor (NS5Ai). This study investigated the prevalence and impact of baseline NS3 and NS5A resistance-associated substitutions (RASs) on the efficacy of G/P in prior GT1 sofosbuvir+NS5Ai failures and the persistence of treatment-emergent RASs. METHODS: Longitudinal samples from 177 patients enrolled in a phase IIIb, randomized pragmatic clinical trial were analyzed. Patients without cirrhosis were randomized to 12 or 16 weeks of G/P, and patients with compensated cirrhosis were randomized to G/P and ribavirin for 12 weeks or G/P for 16 weeks. Linkage of RAS was identified using Primer-ID next-generation sequencing at a 15% cut-off. RESULTS: Of 177 patients, 169 (95.5%) were PI-naïve. All 33 GT1b-infected patients achieved SVR12. In GT1a-infected patients, baseline NS5A RASs were prevalent (74.5%, 105/141) but NS3 RASs were uncommon. Baseline NS3 RASs had no impact on G/P efficacy and patients with baseline NS5A RASs showed a numerically but not statistically significantly lower SVR12 rate compared to those without NS5A RASs (89% vs. 97%). SVR12 was achieved in 34 of 35 (97%) patients without NS5A baseline substitution, and 53 of 57 (93%), 35 of 40 (88%), 5 of 8 (63%) with single, double-linked, and triple-linked NS5A substitutions, respectively. Among 13 patients with virologic failure, 4 acquired treatment-emergent NS3 RASs and 10 acquired NS5A RASs. CONCLUSION: Baseline NS5A RASs were highly prevalent. The presence of an increasing number of linked NS5A RASs in GT1a showed a trend in decreasing SVR12 rates, although no specific NS5A RASs or their linkage pattern were associated with lower SVR12 rates. LAY SUMMARY: Direct-acting antivirals have revolutionized the treatment of chronic hepatitis C infection, but treatment failure occurs in some patients. Retreatment of patients who previously failed a regimen consisting of sofosbuvir and an NS5A inhibitor with a regimen of glecaprevir and pibrentasvir (G/P) is >90% effective. Herein, we analyzed samples from these patients and showed that retreatment efficacy with G/P is lower in patients with double- or triple-linked NS5A resistance mutations than in patients with single or no NS5A resistance mutations. CLINICAL TRIAL NUMBER: NCT03092375.

Efficacy of Glecaprevir and Pibrentasvir in Patients With Genotype 1 Hepatitis C Virus Infection With Treatment Failure After NS5A Inhibitor Plus Sofosbuvir Therapy.

BACKGROUND & AIMS: Treatment options are limited for patients with hepatitis C (HCV) infection with treatment failure after sofosbuvir plus an NS5A inhibitor. There are some data for the efficacy of glecaprevir/pibrentasvir (G/P) in these patients. We performed a randomized trial of the safety and efficacy of 12 and 16 weeks of G/P, with or without ribavirin, in patients with HCV genotype 1 infection with treatment failure after sofosbuvir and an NS5A inhibitor. METHODS: We performed a phase 3b, open-label study of patients with chronic HCV genotype 1 infection who received previous treatment with sofosbuvir plus an NS5A inhibitor. Patients without cirrhosis were randomly assigned to groups that received G/P for 12 weeks (n = 78, group A) or 16 weeks (n = 49, group B). Patients with compensated cirrhosis were randomly assigned to groups that received G/P and ribavirin for 12 weeks (n = 21, group C) or G/P for 16 weeks (n = 29, group D). The primary end point was a sustained virologic response 12 weeks after treatment. Samples collected at baseline and at time of treatment failure were sequenced for resistance-associated substitutions in NS3 and NS5A. RESULTS: Of the 177 patients in the 4 groups, 81% were men, 79% had HCV genotype 1a infection, and 44% were black. Proportions of patients with sustained virologic response 12 weeks after treatment in groups A, B, C, and D were 90%, 94%, 86%, and 97%, respectively. The treatment failed in 13 (7.3%) patients with HCV genotype 1a infection, 6 (7.9%) in group A, 3 (6.1%) in group B, 3 (6.1%) in group C (6.1%), and 1 (3.4%) in group D. Most patients had baseline resistance-associated substitutions in NS5A. Treatment-emergent resistance-associated substitutions in NS3 and NS5A were observed in 9 and 10 patients with treatment failure, respectively. G/P was well tolerated. Ribavirin increased adverse events but did not increase efficacy. CONCLUSIONS: In a randomized study of patients with chronic HCV genotype 1 infection who received previous treatment with sofosbuvir plus an NS5A inhibitor, 16 weeks treatment with G/P produced sustained virologic response 12 weeks after treatment in >90% of patients, including those with compensated cirrhosis. ClinicalTrials.gov, Number: NCT03092375.

Prevalence and impact of baseline resistance-associated substitutions on the efficacy of ledipasvir/sofosbuvir or simeprevir/sofosbuvir against GT1 HCV infection.

Baseline resistance-associated substitutions (RASs) have variable impacts in clinical trials but their prevalence and impact in real-world patients remains unclear. We performed baseline resistance testing using a commercial assay (10% cutoff) for 486 patients treated with LDV/SOF or SMV/SOF, with or without ribavirin, in the multi-center, observational HCV-TARGET cohort. Linkage of RASs was evaluated in selected samples using a novel quantitative single variant sequencing assay. Our results showed that the prevalence of NS3, NS5A, NS5B RASs was 45%, 13%, and 8%, respectively, and 10% of patients harbored RASs in 2 or more drug classes. Baseline LDV RASs in GT1a, TE, and cirrhosis LDV/SOF subgroup was associated with 2-4% lower SVR12 rates. SMV RASs was associated with lower SVR12 rates in GT1a, treatment-experienced, cirrhotics SMV/SOF subgroup. Pooled analysis of all patients with baseline RASs revealed that SVR12 was 100% (19/19) in patients treated for longer than 98 days but was 87% (81/93) in patients treated for shorter than 98 days. These results demonstrate that RASs prevalence and their impact in real world practice are in general agreement with registration trials, and suggest that longer treatment duration may overcome the negative impact of baseline RASs on SVR12 rates in clinical practice.

Hepatitis C virus: how to provide the best treatment with what I have.

The therapeutic landscape for the treatment of chronic hepatitis C virus infection has been rapidly evolving, and by 2016 there will be six approved, all-oral regimens for use in patients in the USA and most of Western Europe. However, as many as patient populations will have limited access to new direct acting antiviral regimens, patients and physicians are often faced with the challenge of selecting the best regimen available, as opposed to the optimal treatment. In this paper, the challenges and opportunities in developing a high cure regimen for different patient populations will be discussed and highlighted through case-based scenarios.

Optimal interferon-free therapy in treatment-experienced chronic hepatitis C patients.

Over the past year, interferon (IFN) free dosing regimens have become available to treat chronic hepatitis C. Offering high rates of sustained virological response (SVR), short treatment and improved tolerability, IFN-free treatment now represents the paradigm for both treatment-naïve and -experienced patients. Patients with prior treatment failure, in particular those with cirrhosis, still represent some of the most difficult to treat, but the availability of multiple agents that can interrupt several steps of the HCV lifecycle affords providers and patients with options that can be combined and individually tailored to each patient's unique needs to obtain high rates of SVR.

Safety profile of boceprevir and telaprevir in chronic hepatitis C: real world experience from HCV-TARGET.

BACKGROUND & AIMS: The safety profiles of boceprevir and telaprevir in the treatment of chronic hepatitis C, administered in academic and community centres across the United States, were evaluated. METHODS: In 90 medical centres, patients with chronic HCV received pegylated interferon, ribavirin, and either telaprevir or boceprevir per local standard of care. Demographic, adverse event, clinical, and virological data were collected during treatment and follow-up. RESULTS: A total of 2084 patients (97% HCV genotype 1) received at least one dose of a protease inhibitor. At baseline, 38% of patients had cirrhosis, and 57% had received at least one prior treatment for hepatitis C. Serious adverse events occurred in 12% of patients receiving protease inhibitor therapy. Overall, 66% of patients experienced anaemia, leading to frequent ribavirin dose reductions (42%) and erythropoietin use (37%); 11% received blood transfusion. More than 90% of patients had adverse events that led to a prescription, treatment, or dosage change, and 39% of patients discontinued treatment early, most commonly because of adverse events (18%) or lack of efficacy (16%). Hepatic decompensation events occurred in 3% of all patients. Age, female gender, cirrhosis, HCV genotype 1 subtype, creatinine clearance, platelet levels, albumin levels and haemoglobin levels were independent predictors of anaemia. Five deaths occurred. Overall, 52% of all patients achieved a sustained virologic response. CONCLUSIONS: In academic and community centres, where chronic hepatitis C patients commonly have advanced liver disease, triple therapy was associated with a high rate of adverse events and involved frequent treatment modifications and adverse event management.

New therapeutic strategies in HCV: second-generation protease inhibitors.

Telaprevir and boceprevir are the first direct-acting antiviral agents approved for use in HCV treatment and represent a significant advance in HCV therapy. However, these first-generation drugs also have significant limitations related to thrice-daily dosing, clinically challenging side-effect profiles, low barriers to resistance and a lack of pan-genotype activity. A second wave of protease inhibitors are in phase II and III trials and promise to provide a drug regimen with a better dosing schedule and improved tolerance. These second-wave protease inhibitors will probably be approved in combination with PEG-IFN and Ribavirin (RBV), as well as future all-oral regimens. The true second-generation protease inhibitors are in earlier stages of development and efficacy data are anxiously awaited as they may provide pan-genotypic antiviral activity and a high genetic barrier to resistance.

Betaine for nonalcoholic fatty liver disease: results of a randomized placebo-controlled trial.

UNLABELLED: Based on animal studies and pilot studies in humans, betaine, a methyl donor for the remethylation of homocysteine, may be a therapeutic agent for nonalcoholic steatohepatitis (NASH). We evaluated the safety and efficacy of betaine for patients with NASH and whether betaine positively modified factors postulated to be "second hits" and underlying mechanisms of NASH. We conducted a randomized placebo-control study of 55 patients with biopsy-proven NASH who received either oral betaine (20 g daily) or placebo for 12 months. Pre- and posttreatment variables were analyzed using the paired t test or Wilcoxon rank test. Treatment groups were comparable at baseline. Of the 35 patients (17 betaine, 18 placebo) who completed the study, 34 patients (16 betaine, 18 placebo) underwent posttreatment liver biopsy. Patients randomized to betaine had a decrease in steatosis grade. No intra- or intergroup differences or changes in nonalcoholic fatty liver disease activity score or fibrosis stage were noted. Elevations of insulin, glucose, and proinflammatory cytokines and the reduced antioxidant status noted in NASH patients did not improve with betaine therapy. The antiinflammatory agent adiponectin was significantly reduced in both groups and did not change with therapy. Lastly, S-adenosylhomocysteine was approximately twice normal and was not reduced by betaine therapy. CONCLUSION: Compared to placebo, betaine did not improve hepatic steatosis but may protect against worseningsteatosis [corrected]. High-dose betaine supplementation failed to reduce S-adenosylhomocysteine and did not positively affect any of the second hit mechanisms postulated to contribute to NASH that we studied. Although betaine has been proven effective in treating hepatic steatosis in several animal models, translating novel therapeutic options noted in animal studies to humans with NASH will prove challenging.

Hepatitis C in adults and adolescents with hemophilia: a randomized, controlled trial of interferon alfa-2b and ribavirin.

Adolescents and adults with inherited disorders of coagulation have one of the highest prevalence rates of hepatitis C among known risk groups. Few data are available on the use of combination therapy with interferon and ribavirin in this population. Patients 13 years of age and older who were positive for hepatitis C virus (HCV) RNA by polymerase chain reaction and negative for human immunodeficiency virus were randomized to receive interferon alfa-2b (3 million units 3 times a week) plus ribavirin (1,000 mg/day) or interferon alfa-2b alone for 48 weeks with 24 weeks of posttreatment follow-up. Patients started on interferon alone who remained positive for HCV RNA at week 12 crossed over to treatment with interferon plus ribavirin. A total of 113 patients were treated. Thirty-seven patients were younger than 18 years. At the end of treatment, 18 of 56 (32%) treated with interferon plus ribavirin and 6 of 57 (11%) treated with interferon alone were negative for HCV RNA (P =.005). Sustained virologic response in the combination arm was 29% (16 of 56) compared with 7% (4 of 57) for those started on interferon alone (P =.027). Among adolescents younger than 18 years who were treated with combination therapy, 10 of 17 (59%) had sustained response compared with 6 of 39 (15%) of adult patients on the same regimen (P =.001). In conclusion, in this U.S. multicenter, randomized trial of therapy for HCV in patients with inherited bleeding disorders, sustained virologic response rate was significantly improved for patients treated with interferon and ribavirin compared with those started on interferon alone. Adolescents treated with combination therapy had a significantly higher sustained response than adults did on the same regimen.