A framework for longitudinal latent factor modelling of treatment response in clinical trials with applications to Psoriatic Arthritis and Rheumatoid Arthritis.

OBJECTIVE: Clinical trials involve the collection of a wealth of data, comprising multiple diverse measurements performed at baseline and follow-up visits over the course of a trial. The most common primary analysis is restricted to a single, potentially composite endpoint at one time point. While such an analytical focus promotes simple and replicable conclusions, it does not necessarily fully capture the multi-faceted effects of a drug in a complex disease setting. Therefore, to complement existing approaches, we set out here to design a longitudinal multivariate analytical framework that accepts as input an entire clinical trial database, comprising all measurements, patients, and time points across multiple trials. METHODS: Our framework composes probabilistic principal component analysis with a longitudinal linear mixed effects model, thereby enabling clinical interpretation of multivariate results, while handling data missing at random, and incorporating covariates and covariance structure in a computationally efficient and principled way. RESULTS: We illustrate our approach by applying it to four phase III clinical trials of secukinumab in Psoriatic Arthritis (PsA) and Rheumatoid Arthritis (RA). We identify three clinically plausible latent factors that collectively explain 74.5% of empirical variation in the longitudinal patient database. We estimate longitudinal trajectories of these factors, thereby enabling joint characterisation of disease progression and drug effect. We perform benchmarking experiments demonstrating our method's competitive performance at estimating average treatment effects compared to existing statistical and machine learning methods, and showing that our modular approach leads to relatively computationally efficient model fitting. CONCLUSION: Our multivariate longitudinal framework has the potential to illuminate the properties of existing composite endpoint methods, and to enable the development of novel clinical endpoints that provide enhanced and complementary perspectives on treatment response.

VertXNet: an ensemble method for vertebral body segmentation and identification from cervical and lumbar spinal X-rays.

Accurate annotation of vertebral bodies is crucial for automating the analysis of spinal X-ray images. However, manual annotation of these structures is a laborious and costly process due to their complex nature, including small sizes and varying shapes. To address this challenge and expedite the annotation process, we propose an ensemble pipeline called VertXNet. This pipeline currently combines two segmentation mechanisms, semantic segmentation using U-Net, and instance segmentation using Mask R-CNN, to automatically segment and label vertebral bodies in lateral cervical and lumbar spinal X-ray images. VertXNet enhances its effectiveness by adopting a rule-based strategy (termed the ensemble rule) for effectively combining segmentation outcomes from U-Net and Mask R-CNN. It determines vertebral body labels by recognizing specific reference vertebral instances, such as cervical vertebra 2 ('C2') in cervical spine X-rays and sacral vertebra 1 ('S1') in lumbar spine X-rays. Those references are commonly relatively easy to identify at the edge of the spine. To assess the performance of our proposed pipeline, we conducted evaluations on three spinal X-ray datasets, including two in-house datasets and one publicly available dataset. The ground truth annotations were provided by radiologists for comparison. Our experimental results have shown that the proposed pipeline outperformed two state-of-the-art (SOTA) segmentation models on our test dataset with a mean Dice of 0.90, vs. a mean Dice of 0.73 for Mask R-CNN and 0.72 for U-Net. We also demonstrated that VertXNet is a modular pipeline that enables using other SOTA model, like nnU-Net to further improve its performance. Furthermore, to evaluate the generalization ability of VertXNet on spinal X-rays, we directly tested the pre-trained pipeline on two additional datasets. A consistently strong performance was observed, with mean Dice coefficients of 0.89 and 0.88, respectively. In summary, VertXNet demonstrated significantly improved performance in vertebral body segmentation and labeling for spinal X-ray imaging. Its robustness and generalization were presented through the evaluation of both in-house clinical trial data and publicly available datasets.

A deep learning approach to private data sharing of medical images using conditional generative adversarial networks (GANs).

Clinical data sharing can facilitate data-driven scientific research, allowing a broader range of questions to be addressed and thereby leading to greater understanding and innovation. However, sharing biomedical data can put sensitive personal information at risk. This is usually addressed by data anonymization, which is a slow and expensive process. An alternative to anonymization is construction of a synthetic dataset that behaves similar to the real clinical data but preserves patient privacy. As part of a collaboration between Novartis and the Oxford Big Data Institute, a synthetic dataset was generated based on images from COSENTYX® (secukinumab) ankylosing spondylitis (AS) clinical studies. An auxiliary classifier Generative Adversarial Network (ac-GAN) was trained to generate synthetic magnetic resonance images (MRIs) of vertebral units (VUs), conditioned on the VU location (cervical, thoracic and lumbar). Here, we present a method for generating a synthetic dataset and conduct an in-depth analysis on its properties along three key metrics: image fidelity, sample diversity and dataset privacy.

Quantification of pre-existing radiographic damage and its relationship with joint activity and long-term clinical outcomes with secukinumab therapy in patients with psoriatic arthritis.

BACKGROUND: Psoriatic arthritis (PsA) patient data from two phase 3 secukinumab trials (FUTURE 1, 5) were analysed to quantify the prevalence and extent of pre-existing radiographic damage (RD) at baseline; investigate the association of RD with swollen/tender joint counts (SJC/TJC) at baseline; and investigate the extent to which RD at baseline correlated with response to secukinumab. METHODS: Pooled data (N = 1554) provided baseline radiographic bone erosion and joint space narrowing (JSN) scores at pre-specified locations per the van der Heijde-modified total Sharp score (vdH-mTSS) for PsA and swollen and tender joint scores in the same joints at multiple visits. Overall patient RD and individual joints RD bone erosion and JSN scores were assessed. The association between joint activity (tenderness, swelling) and vdH-mTSS was assessed at the overall patient-level and individual joint tender, swollen scores (yes/no) and RD joint JSN and bone erosion scores at the individual joint-level. Treatment response was assessed using SJC/TJC at weeks 16 and 52 and the proportion of patients achieving minimal disease activity (MDA) over all assessments within 1 year from FUTURE 5 alone. RESULTS: A substantial prevalence of pre-existing RD with higher prevalence of erosion than JSN was observed (86% and 60% of patients had positive erosion and JSN scores, respectively); higher RD prevalence was associated with longer time since PsA diagnosis. Joint activity was weakly associated with RD at baseline at the patient-level (Pearson's coefficients: range 0.12-0.18), but strongly associated at the individual joint-level, with a higher probability of tender/swollen joints to associate with higher JSN/erosion scores: all 42 analysed joints showed statistical significance at the 0.05 level (unadjusted) for the relationship between joint tenderness (yes/no) and its JSN score, all but one for tenderness and bone erosion scores, and all but 2 for swollen and JSN scores and for swollen and bone erosion score. Secukinumab (150/300 mg), reduced TJC and SJC across all values of baseline erosion and JSN scores at weeks 16 and 52. Patients with higher levels of RD were less likely to achieve zero tender/zero swollen joint status and had lower chance of achieving MDA. CONCLUSIONS: PsA patients showed substantial prevalence of RD at baseline that correlated with time since diagnosis, but patient's individual joint activity was strongly associated with pre-existing RD at those joints. Patients with the highest RD at baseline had a reduced likelihood of achieving zero joint count status.

Clinically relevant patient clusters identified by machine learning from the clinical development programme of secukinumab in psoriatic arthritis.

OBJECTIVES: Identify distinct clusters of psoriatic arthritis (PsA) patients based on their baseline articular, entheseal and cutaneous disease manifestations and explore their clinical and therapeutic value. METHODS: Pooled baseline data in PsA patients (n=1894) treated with secukinumab across four phase 3 studies (FUTURE 2-5) were analysed to determine phenotypes based on clusters of clinical indicators. Finite mixture models methodology was applied to generate clinical clusters and mean longitudinal responses were compared between secukinumab doses (300 vs 150 mg) across identified clusters and clinical indicators through week 52 using machine learning (ML) techniques. RESULTS: Seven distinct patient clusters were identified. Cluster 1 (very-high (VH) - SWO/TEN (swollen/tender); n=187) was characterised by VH polyarticular burden for both tenderness and swelling of joints, while cluster 2 (H (high) - TEN; n=251) was marked by high polyarticular burden in tender joints and cluster 3 (H - Feet - Dactylitis; n=175) by high burden in joints of feet and dactylitis. For cluster 4 (L (Low) - Nails - Skin; n=209), cluster 5 (L - skin; n=283), cluster 6 (L - Nails; n=294) and cluster 7 (L; n=495) articular burden was low but nail and skin involvement was variable, with cluster 7 marked by mild disease activity across all domains. Greater improvements in the longitudinal responses for enthesitis in cluster 2, enthesitis and Psoriasis Area and Severity Index (PASI) in cluster 4 and PASI in cluster 6 were shown for secukinumab 300 mg compared with 150 mg. CONCLUSIONS: PsA clusters identified by ML follow variable response trajectories indicating their potential to predict precise impact on patients' outcomes. TRIAL REGISTRATION NUMBERS: NCT01752634, NCT01989468, NCT02294227, NCT02404350.

Secukinumab Efficacy in Psoriatic Arthritis: Machine Learning and Meta-analysis of Four Phase 3 Trials.

BACKGROUND: Using a machine learning approach, the study investigated if specific baseline characteristics could predict which psoriatic arthritis (PsA) patients may gain additional benefit from a starting dose of secukinumab 300 mg over 150 mg. We also report results from individual patient efficacy meta-analysis (IPEM) in 2049 PsA patients from the FUTURE 2 to 5 studies to evaluate the efficacy of secukinumab 300 mg, 150 mg with and without loading regimen versus placebo at week 16 on achievement of several clinically relevant difficult-to-achieve (higher hurdle) endpoints. METHODS: Machine learning employed Bayesian elastic net to analyze baseline data of 2148 PsA patients investigating 275 predictors. For IPEM, results were presented as difference in response rates versus placebo at week 16. RESULTS: Machine learning showed secukinumab 300 mg has additional benefits in patients who are anti-tumor necrosis factor-naive, treated with 1 prior anti-tumor necrosis factor agent, not receiving methotrexate, with enthesitis at baseline, and with shorter PsA disease duration. For IPEM, at week 16, all secukinumab doses had greater treatment effect (%) versus placebo for higher hurdle endpoints in the overall population and in all subgroups; 300-mg dose had greater treatment effect than 150 mg for all endpoints in overall population and most subgroups. CONCLUSIONS: Machine learning identified predictors for additional benefit of secukinumab 300 mg compared with 150 mg dose. Individual patient efficacy meta-analysis showed that secukinumab 300 mg provided greater improvements compared with 150 mg in higher hurdle efficacy endpoints in patients with active PsA in the overall population and most subgroups with various levels of baseline disease activity and psoriasis.

Secukinumab provides sustained low rates of radiographic progression in psoriatic arthritis: 52-week results from a phase 3 study, FUTURE 5.

OBJECTIVE: To evaluate the effect of secukinumab on radiographic progression through 52 weeks in patients with PsA from the FUTURE 5 study. METHODS: Patients with active PsA, stratified by prior anti-TNF use (naïve or inadequate response), were randomized to s.c. secukinumab 300 mg load (300 mg), 150 mg load (150 mg), 150 mg no load regimens or placebo at baseline, at weeks 1, 2 and 3 and every 4 weeks starting at week 4. Radiographic progression was assessed by change in van der Heijde-modified total Sharp score (vdH-mTSS; mean of two readers). Statistical analysis used a linear mixed-effects model (random slope) at weeks 24 and 52, and observed data at week 52. Assessments at week 52 included additional efficacy endpoints (non-responders imputation and mixed-effects models for repeated measures) and safety. RESULTS: The majority (86.6%) of patients completed 52 weeks of treatment. The proportion of patients with no radiographic progression (change from baseline in vdH-mTSS ⩽0.5) was 91.8, 85.2 and 87.2% in 300, 150 and 150 mg no load groups, respectively, at week 52. The change in vdH-mTSS from baseline to week 52 using random slope [mean change (s.e.)] was -0.18 (0.17), 0.11 (0.18) and -0.20 (0.18) in 300, 150 and 150 mg no load groups, respectively; the corresponding observed data [mean change (s.d.)] was -0.09 (1.02), 0.13 (1.39) and 0.21 (1.15). Clinical efficacy endpoints were sustained, and no new or unexpected safety signals were reported through 52 weeks. CONCLUSION: Secukinumab 300 and 150 mg with or without s.c. loading regimen provided sustained low rates of radiographic progression through 52 weeks of treatment. TRIAL REGISTRATION: ClinicalTrials.gov, http://clinicaltrials.gov, NCT02404350.

Secukinumab, a human anti-interleukin-17A monoclonal antibody, in patients with psoriatic arthritis (FUTURE 2): a randomised, double-blind, placebo-controlled, phase 3 trial.

BACKGROUND: Interleukin 17A is a proinflammatory cytokine that is implicated in the pathogenesis of psoriatic arthritis. We assessed the efficacy and safety of subcutaneous secukinumab, a human anti-interleukin-17A monoclonal antibody, in patients with psoriatic arthritis. METHODS: In this phase 3, double-blind, placebo-controlled study undertaken at 76 centres in Asia, Australia, Canada, Europe, and the USA, adults (aged ≥18 years old) with active psoriatic arthritis were randomly allocated in a 1:1:1:1 ratio with computer-generated blocks to receive subcutaneous placebo or secukinumab 300 mg, 150 mg, or 75 mg once a week from baseline and then every 4 weeks from week 4. Patients and investigators were masked to treatment assignment. The primary endpoint was the proportion of patients achieving at least 20% improvement in the American College of Rheumatology response criteria (ACR20) at week 24. Analysis was by intention to treat. This study is registered with ClinicalTrials.gov, number NCT01752634. FINDINGS: Between April 14, and Nov 25, 2013, 397 patients were randomly assigned to receive secukinumab 300 mg (n=100), 150 mg (n=100), 75 mg (n=99), or placebo (n=98). A significantly higher proportion of patients achieved an ACR20 at week 24 with secukinumab 300 mg (54 [54%] patients; odds ratio versus placebo 6·81, 95% CI 3·42-13·56; p<0·0001), 150 mg (51 [51%] patients; 6·52, 3·25-13·08; p<0·0001), and 75 mg (29 [29%] patients; 2·32, 1·14-4·73; p=0·0399) versus placebo (15 [15%] patients). Up to week 16, the most common adverse events were upper respiratory tract infections (four [4%], eight [8%], ten [10%], and seven [7%] with secukinumab 300 mg, 150 mg, 75 mg, and placebo, respectively) and nasopharyngitis (six [6%], four [4%], six [6%], and eight [8%], respectively). Serious adverse events were reported by five (5%), one (1%), and four (4%) patients in the secukinumab 300 mg, 150 mg, and 75 mg groups, respectively, compared with two (2%) in the placebo group. No deaths were reported. INTERPRETATION: Subcutaneous secukinumab 300 mg and 150 mg improved the signs and symptoms of psoriatic arthritis, suggesting that secukinumab is a potential future treatment option for patients with this disorder. FUNDING: Novartis.

One-year efficacy and safety results of secukinumab in patients with rheumatoid arthritis: phase II, dose-finding, double-blind, randomized, placebo-controlled study.

OBJECTIVE: To evaluate the longer-term safety and efficacy of secukinumab, a fully human monoclonal antiinterleukin-17A antibody, in patients with rheumatoid arthritis. METHODS: In this 52-week, double-blind, placebo-controlled (up to Week 20) study (NCT00928512), patients responding inadequately to disease-modifying antirheumatic drugs (DMARD) or biologics were randomized to receive monthly subcutaneous injections of secukinumab (25, 75, 150, or 300 mg), or placebo. The efficacy and safety results up to Week 20 have been reported previously. Here, efficacy results from Week 20 to 52 and safety results from Week 20 to 60 are presented. RESULTS: Of 237 patients randomized, 174 (73.4%) completed the study. Patients with improved American College of Rheumatology (ACR) and 28-joint Disease Activity Score (DAS28) C-reactive protein (CRP) responses at Week 16 sustained their responses through Week 52. In patients taking 150 mg of secukinumab, responses were improved through Week 52 (ACR50: Week 16 = 45%, Week 52 = 55%; DAS28-CRP ≤ 2.6: Week 16 = 25%, Week 52 = 40%). The rate of adverse events (AE) from weeks 20 to 60 was 64.8%, with most AE being mild to moderate in severity. The overall rate of infections was 31.9%, most being mild. The most predominant infection was nasopharyngitis, and was not associated with dose or concurrent neutropenia. Serious AE were reported in 21 patients (8.9%). There were 3 reports of malignancies (ovarian, lung, basal cell), and no deaths between weeks 20 and 60. CONCLUSION: Patients with active RA who failed to respond to DMARD and other biologics showed an improvement after longterm treatment with 150 mg of secukinumab. The frequency of AE remained stable over time and secukinumab had a consistent safety profile over 60 weeks.

Long-term tegaserod treatment for dysmotility-like functional dyspepsia: results of two identical 1-year cohort studies.

BACKGROUND: Functional dyspepsia (FD) is a chronic disorder that adversely affects health-related quality of life (HRQoL). Published information on its long-term management is minimal and treatment options are limited. AIM: The aim of this study was to evaluate safety, efficacy and HRQoL with tegaserod 6 mg twice daily over 1 year in women with FD who completed one of two 6-week, randomized, placebo-controlled, double-blind studies. METHODS: About 780 patients received tegaserod 6 mg twice daily in two identical 1-year extension studies. Scheduled assessments included adverse events, the Short-Form Nepean Dyspepsia Index (SF-NDI), Work Productivity and Activity Impairment-Dyspepsia (WPAI-Dyspepsia) questionnaire, and patient perceptions of treatment efficacy. RESULTS: Mean tegaserod treatment duration in the two studies was 236 and 222 days. Most adverse events occurred in the first 6 months, were similar to previous reports (commonly diarrhea), and were transient and well tolerated. SF-NDI, WPAI-Dyspepsia scores and perceived symptom relief improved from baseline over the 1-year evaluation. CONCLUSIONS: The long-term safety profile of tegaserod in women with FD was consistent with that of short-term treatment and accompanied by improvements in HRQoL, work productivity and symptom relief. These long-term results add to the clinical experience with FD and support the potential value of a 5-HT(4) agonist in the management of FD.

Tegaserod treatment for dysmotility-like functional dyspepsia: results of two randomized, controlled trials.

OBJECTIVES: Therapies for dysmotility-like functional dyspepsia (FD) are limited. We studied tegaserod, a selective serotonin type 4 receptor agonist, in patients with FD. METHODS: Two identical multicenter, double-blind, randomized, placebo-controlled trials enrolled women >/=18 yr with recurring mid-upper abdominal discomfort characterized by postprandial fullness, early satiety, and/or bloating. Patients were randomized to tegaserod 6 mg b.i.d. or placebo. Two patient-reported primary variables were assessed: percentage of days with satisfactory symptom relief, and symptom severity using the composite average daily severity score (CADSS). RESULTS: In total, 2,667 women were randomized with no differences between trials in terms of recruitment method, Helicobacter pylori status, heartburn, or medication use. Mean percentage of days with satisfactory symptom relief for tegaserod versus placebo in Trial 1: 32.2%versus 26.6% (95% CI of treatment difference 2.82, 9.27; P < 0.01), Trial 2: 31.9%versus 29.4% (95% CI of treatment difference -0.21, 6.53; P= 0.066). Mean CADSS in Trial 1: 3.14 versus 3.35 (95% CI of treatment difference -0.29, -0.10; P < 0.0001), Trial 2: 3.15 versus 3.23 (95% CI of treatment difference -0.18, 0.01; P= 0.094). Meta-analysis showed significant benefit for both end points: increase in days with satisfactory relief 4.6% (95% CI 2.29, 6.96); decrease in CADSS 0.14 (95% CI 0.21, 0.07). Treatment effect was greater in patients with severe baseline symptoms. Diarrhea requiring study discontinuation was more common with tegaserod than placebo (4.1%vs 0.3%). CONCLUSIONS: Some improvement in dysmotility-like FD was observed with tegaserod treatment. The clinical implication of this improvement is uncertain.

Tegaserod for female patients suffering from IBS with mixed bowel habits or constipation: a randomized controlled trial.

OBJECTIVES: Though the greatest proportion of irritable bowel syndrome (IBS) patients report a mixed bowel pattern (IBS-Mixed), no available therapies have been rigorously evaluated in this subgroup. This study aimed to evaluate the efficacy and safety of the 5-HT(4) agonist tegaserod in women with IBS-Mixed and IBS with constipation (IBS-C). METHODS: This prospective, double-blind, randomized, placebo-controlled, multicenter study was conducted in 100 centers in North America, South America, and Europe. Women with IBS-Mixed or IBS-C received tegaserod 6 mg or placebo twice daily. The primary efficacy variable was the patient's assessment of satisfactory relief over the 4-wk treatment period. The proportion of patients reporting satisfactory relief for >/=3 of 4 treatment weeks (75% rule) and individual IBS symptoms were assessed. RESULTS: In total, 661 women were randomized (IBS-Mixed 324, IBS-C 337). Baseline symptom assessments identified clear differences between the two cohorts. Tegaserod provided significant improvement in satisfactory relief of IBS symptoms over 4 wk (OR 1.75, 95% CI 1.35-2.25, P < 0.001) in both IBS-Mixed and IBS-C patients. Using the 75% rule, 52.3% of tegaserod-receiving IBS-M patients and 43.3% of IBS-C patients were responders (vs 36.3, OR 1.88, 95% CI 1.16-3.04, P < 0.010; and 28.9, OR 1.90, 95% CI 1.19-3.05, P < 0.008 for placebo, respectively). The most frequent adverse events leading to study discontinuation in tegaserod-treated patients were diarrhea (1.5%) and abdominal pain (0.9%). Overall 7% of IBS-C patients reported diarrhea compared to 12% of IBS-Mixed (placebo 2.4%, 1.8%, respectively). CONCLUSIONS: Tegaserod is effective in treating overall IBS symptoms in patients with IBS-Mixed and IBS-C.

Omeprazole-induced slowing of gastrointestinal transit in mice can be countered with tegaserod.

Omeprazole, besides suppressing gastric acid, causes delayed gastric emptying, which may be associated with aggravated dyspeptic symptoms. Effects of omeprazole on small intestinal transit are unknown. In this study, we evaluated in mice if (a) omeprazole affects transit of a meal through the stomach and small intestine and (b) co-treatment with the promotility agent, tegaserod, can prevent the slowing effect of omeprazole. Omeprazole (40-150 mg/kg, i.p. once daily for 5 days) delayed gastric emptying of the meal in a dose-related manner. Small intestinal transit was then evaluated at the lowest dose of omeprazole (40 mg/kg) that did not retard gastric emptying. Such transit was significantly delayed after this dose of omeprazole compared with vehicle-treated controls. When tegaserod (0.10 mg/kg) was administered concomitantly with the omeprazole, small intestinal transit of the meal was not slowed and was not different from controls. These results show that omeprazole reduces aboral transit of luminal contents through the stomach and small bowel of mice and that this delay is reversed by tegaserod.