Predict Treatment Effects Research Articles

Sex differences in residual somatic symptoms in patients with first-episode depression after acute-phase treatment

BackgroundResidual somatic symptoms (RSS) are common in depressed patients, predicting treatment effectiveness. However, sex differences in RSS have received little systematic study. This study was conducted to compare sex differences of RSS in patients with first-episode depression (FED).MethodsNine hundred eighty-two patients with FED were selected and treated for 8 to 12 weeks. We evaluated the subjects' socio-demographic characteristics and residual depressive symptoms. Using the Patient Health Questionnaire-15 (PHQ-15) scale to assess residual somatic symptoms, the Sheehan Disability Scale (SDS) for the assessment of patients' function, the Quality of Life Enjoyment and Satisfaction Questionnaire-Short Form (Q-LES-Q-SF) for quality of life.ResultsThe incidence of RSS with FED was 46.4%. For patients with residual symptoms, the age and age of onset in females were higher than males, but males had more years of education than females. The degree of "stomach pain" in females was more severe than in males, while "trouble sleeping" in males was more severe than that in females. Multiple regression analysis showed that the total Q-LES-Q-SF score was an independent influencing factor of RSS in both males and females, while the total SDS score only affected female RSS.ConclusionsThe prevalence of RSS in FED after acute-phase treatment is high. The symptom of "stomachache" is more pronounced in females, while "trouble sleeping" is more severe in males. Quality of life plays an essential role in RSS in both genders. Thus, sex needs to be considered when assessing the relationship between RSS and therapeutic effect in depression.

The application of the Limberg flap repair technique in the surgical treatment of pilonidal sinus disease.

Pilonidal sinus disease (PNSD) challenged surgeons for decades. Limberg flap repair (LFR) is a common treatment for PNSD. The purpose of this study was to observe the effect and risk factors of LFR in PNSD. A retrospective study was conducted on the PNSD patients who visited two medical centers and four departments in the People's Liberation Army General Hospital and were taking LFR treatment between 2016 and 2022. The risk factors, the effect of the operation, and complications were observed. The effects of known risk factors on the surgical results were compared. There were 37 PNSD patients: male/female ratio of 35:2, average age: 25.1 ± 7.9 years. Average BMI: 25.2 ± 4.0kg/m2 , average wound healing time: 15.4 ± 3.4 days. 30 patients (81.0%) healed in stage one and 7 (16.3%) had postoperative complications. Only 1 patient (2.7%) had a recurrence while others were healed after dressing-changing. There was no significant difference in age, BMI, preoperative debridement history, preoperative sinus classification, Wound area, Negative pressure drainage tube, prone time (<3d) and treatment effect. Squat defecate and premature defecation were associated with treatment effect, and they were independent predictors of treatment effect in the multivariate analysis. LFR has a stable therapeutic outcome. Compared with other skin flaps, the therapeutic effect of this flap is not significantly different, but the design is simple and is not affected by the known risk factors before operation. However, it is necessary to avoid the influence of two independent risk factors, squatting defecation and premature defecation, on the therapeutic effect.

A Retrospective Study to Evaluate the Role of Dynamic Fracture Mobility in the Conservative Treatment of Osteoporotic Vertebral Compression Fractures

To investigate whether dynamic fracture mobility could affect the outcome of conservative treatment in patients with acute osteoporotic vertebral compression fracture (OVCF). A total of 158 patients who underwent conservative treatment in our hospital for painful OVCFs were included in this study and their data were retrospectively analyzed. According to the degree of pain relief, patients were divided into an excellent efficacy group and a poor efficacy group. Factors that may affect the outcome of conservative treatment were recorded for each patient. Variables with a statistical difference between the 2 groups were entered into multivariate logistic regression analysis to identify the factors influencing the outcome of conservative treatment. Receiver operating characteristic curve analysis was also performed. The result showed that dynamic fracture mobility, overweight, age, and bone mineral density (BMD) (all P < 0.001) were independent factors influencing the outcome of conservative treatment. Receiver operating characteristic curve analysis showed that the cutoff values for age and BMD that predicted treatment effect were 72.5 years and -3.30, respectively. This study confirmed that dynamic fracture mobility could be used as an independent factor predicting the outcome of conservative treatment in patients with acute OVCFs. It was also shown that overweight, age, and BMD were other independent factors influencing the outcome of conservative treatment. A comprehensive evaluation of these related factors can guide the doctor to take appropriate treatment for a unique acute OCVF.

Upper Limb Function Recovery by Combined Repetitive Transcranial Magnetic Stimulation and Occupational Therapy in Patients with Chronic Stroke According to Paralysis Severity.

Repetitive transcranial magnetic stimulation (rTMS) with intensive occupational therapy improves upper limb motor paralysis and activities of daily living after stroke; however, the degree of improvement according to paralysis severity remains unverified. Target activities of daily living using upper limb functions can be established by predicting the amount of change after treatment for each paralysis severity level to further aid practice planning. We estimated post-treatment score changes for each severity level of motor paralysis (no, poor, limited, notable, and full), stratified according to Action Research Arm Test (ARAT) scores before combined rTMS and intensive occupational therapy. Motor paralysis severity was the fixed factor for the analysis of covariance; the delta (post-pre) of the scores was the dependent variable. Ordinal logistic regression analysis was used to compare changes in ARAT subscores according to paralysis severity before treatment. We implemented a longitudinal, prospective, interventional, uncontrolled, and multicenter cohort design and analyzed a dataset of 907 patients with stroke hemiplegia. The largest treatment-related changes were observed in the Limited recovery group for upper limb motor paralysis and the Full recovery group for quality-of-life activities using the paralyzed upper limb. These results will help predict treatment effects and determine exercises and goal movements for occupational therapy after rTMS.

Value of growth arrest lines for predicting treatment effect on children with distal tibial epiphysis fractures.

This study aims to explore whether growth arrest lines can predict epiphyseal fracture healing. The data of 234 children with distal tibial epiphysis fractures treated in our hospital from February 2014 to February 2022 were retrospectively analyzed. Imaging data were examined to record epiphyseal grade, fracture type, and the time to appearance of growth arrest lines. Follow-up data were retrieved to record treatment results (i.e., malunion, premature closure, or bone bridge formation). There was a significant difference in the time to appearance of growth arrest lines between patients with epiphyseal grade 0-1 and grade 2-3 (P < 0.05) and between patients with normal healing and patients with a bone bridge (P < 0.05). Among patients with normal healing, there were no significant differences in the time to appearance of growth arrest lines between men and women and between patients with and without surgery (P > 0.05). There was a significant difference in the time to appearance of growth arrest lines between patients with different Salter-Harris fracture types (P < 0.05). For patients with epiphyseal grade 0-1, the time to appearance of growth arrest lines could be useful for predicting the treatment result of a distal tibial epiphyseal fracture.

P783 Symptom Improvement of ulceRative colitis after an Injection dose of Ustekinumab in Japanese clinical practice, measured using patient-reported outcomes (SIRIUS) study, Interim Analysis

Abstract Background Ustekinumab (UST) is treated to moderate to severe adult ulcerative colitis (UC) patients. We conducted the SIRIUS study (Symptom Improvement of ulceRative colitis after an Injection dose of UStekinumab in Japanese clinical practice, measured using patient-reported outcomes (PRO)) to record daily symptoms for 8 weeks after the introduction of intravenous ustekinumab in moderate to severe UC patients, which aimed to reveal early response patterns of symptoms during the induction of remission period with ustekinumab, and to explore symptoms suitable for monitoring and predicting treatment effect. This interim analysis focuses on stool frequency (SF) and rectal bleeding (RB) scores in the PRO-2 during the 8-week induction period. Methods This single-arm prospective observational study conducted in 23 centers in Japan aimed to recruit 140 patients with moderate-to-severe UC between July 2021 and July 2022. RB score, SF score, abdominal pain, nocturnal diarrhea, tenesmus, and patient perception of improvement were recorded daily using a smartphone app. To be eligible for the study, participants must have a documented diagnosis of UC, be ≥16 years of age, have UC (partial Mayo score of 5~9), have had an inadequate response or intolerance to prior UC therapies, and be able to understand and complete PRO instruments using a smartphone. Patients who had prior exposure to UST, history of extensive colectomy, or were hospitalized when UST was started were excluded. Results In this analysis,137 patients were included. Seventy-four (54.4%) patients were male and the median age was 35 years. Seventy-six (55.9%) patients were Bio-naïve. This high proportion of Bio-naïve patients reflects the fact that UST can be used for Bio-naive patients without restrictions in the Japanese health insurance. At weeks 2 and 4, 56 (41.2%) and 57 (41.9%) patients attended outpatient clinics, respectively. The data collection rate of the PROs was 95%. Mean SF and RB scores were demonstrated stratified by prior treatment with Bio (Figure 1&amp;3). Significant decreases in mean SF and RB scores were observed within 1 week in both Bio-naïve and -experienced patients. the time-course change in SF was similar between Bio-naïve and -experienced patients, but the decrease in RB scores was greater in the Bio-naïve patients than the Bio-experienced patients. Conclusion A daily PRO-based analysis was conducted in the Sirius study to monitor symptoms during the first 8 weeks after the introduction of UST in patients with moderate-to-severe UC. This interim analysis revealed that SF and RB scores started to decrease within 1 week, demonstrating a rapid symptomatic response to UST. The analysis of the other PROs and predictors of efficacy will be performed in further studies.

Measuring the performance of prediction models to personalize treatment choice.

When data are available from individual patients receiving either a treatment or a control intervention in a randomized trial, various statistical and machine learning methods can be used to develop models for predicting future outcomes under the two conditions, and thus to predict treatment effect at the patient level. These predictions can subsequently guide personalized treatment choices. Although several methods for validating prediction models are available, little attention has been given to measuring the performance of predictions of personalized treatment effect. In this article, we propose a range of measures that can be used to this end. We start by defining two dimensions of model accuracy for treatment effects, for a single outcome: discrimination for benefit and calibration for benefit. We then amalgamate these two dimensions into an additional concept, decision accuracy, which quantifies the model's ability to identify patients for whom the benefit from treatment exceeds a given threshold. Subsequently, we propose a series of performance measures related to these dimensions and discuss estimating procedures, focusing on randomized data. Our methods are applicable for continuous or binary outcomes, for any type of prediction model, as long as it uses baseline covariates to predict outcomes under treatment and control. We illustrate all methods using two simulated datasets and a real dataset from a trial in depression. We implement all methods in the R package predieval. Results suggest that the proposed measures can be useful in evaluating and comparing the performance of competing models in predicting individualized treatment effect.

Haemophagocytic lymphohistiocytosis: case series. Serum ferritin level as an indicator of treatment effectiveness

Introduction: Haemophagocytic lymphohistiocytosis (HLH) is a medical condition associated with over-activation of the immune system. HLH results from the inactivity of NK cells and uncontrolled activity of cytotoxic T lymphocytes and macrophages with the massive cytokine response.[1] Cytohistological examinations might show hemophagocytes in different tissues (e.g., bone marrow). Exceptionally high level of ferritin is often observed in the laboratory findings. Due to aggressive course of the disease mortality rate is extremely high. Material and methods: We presented three patients with acquired form of HLH treated successfully in Heaematooncology Department of Medical University of Lublin from September 2018 to April 2021. Results: In the first patient, HLH developed during pregnancy. Patient 2 was first hospitalized in the Intensive Care Unit (ICU) and 10 therapeutic plasma exchanges (TPE) were carried out. In both patients, stabilization -of ferritin levels and remission of the disease were achieved soon after the application of the treatment according to the HLH-2004 protocol. Case 3 presents a patient in whom HLH was induced by EBV and SARS CoV-2 infections. In the first stage of the treatment, the patient received drugs according to the HLH-2004 protocol, but the serum ferritin did not normalize. The reinfection with SARS CoV-2 virus occurred again in the course of the treatment. Due to the disease resistance and inability to continue the HLH-2004 protocol, it was decided to start ruxolitinib therapy, which resulted in stabilization of the serum ferritin and improvement of the general condition. The patient was qualified for allogeneic bone marrow transplantation. Conclusions: HLH is a difficult and interdisciplinary diagnostic and therapeutic problem. It is essential to popularize knowledge about fast and targeted diagnostics. Among laboratory finding, the ferritin concentration seemed to be particularly helpful as a predictor of treatment effectiveness. Proper diagnosis and treatment implemented as early as possible could save patients’ lives.

Flexible evaluation of surrogacy in platform studies.

Trial-level surrogates are useful tools for improving the speed and cost effectiveness of trials but surrogates that have not been properly evaluated can cause misleading results. The evaluation procedure is often contextual and depends on the type of trial setting. There have been many proposed methods for trial-level surrogate evaluation, but none, to our knowledge, for the specific setting of platform studies. As platform studies are becoming more popular, methods for surrogate evaluation using them are needed. These studies also offer a rich data resource for surrogate evaluation that would not normally be possible. However, they also offer a set of statistical issues including heterogeneity of the study population, treatments, implementation, and even potentially the quality of the surrogate. We propose the use of a hierarchical Bayesian semiparametric model for the evaluation of potential surrogates using nonparametric priors for the distribution of true effects based on Dirichlet process mixtures. The motivation for this approach is to flexibly model relationships between the treatment effect on the surrogate and the treatment effect on the outcome and also to identify potential clusters with differential surrogate value in a data-driven manner so that treatment effects on the surrogate can be used to reliably predict treatment effects on the clinical outcome. In simulations, we find that our proposed method is superior to a simple, but fairly standard, hierarchical Bayesian method. We demonstrate how our method can be used in a simulated illustrative example (based on the ProBio trial), in which we are able to identify clusters where the surrogate is, and is not useful. We plan to apply our method to the ProBio trial, once it is completed.

Zebrafish patient-derived xenografts accurately and quickly reproduce treatment outcomes in non-small cell lung cancer patients.

Zebrafish patient-derived xenograft (zPDX) models have shown great potential in predicting the short-term treatment response in various types of tumor cases. However, few studies have used zPDX models for drug screening in non-small cell lung cancer (NSCLC). We aimed to compare the treatment responses of patients with NSCLC with those of the corresponding zPDX models. Tumor cells were obtained from pleural fluid or biopsy procedures from patients with NSCLC and injected into the perivitelline space of zebrafish larvae. Then, the same antineoplastic drugs administered to the corresponding patient were tested in the successfully constructed zPDX model, for 3 days. Responses to treatment were compared. A total of 21 patients with advanced NSCLC were enrolled in our study, and 13 corresponding zPDX models were successfully established. Based on the clinical medication of enrolled patients, we provided a corresponding drug treatment to these zebrafish embryos, including epidermal growth factor receptor tyrosine kinase inhibitors (EGFR TKIs), pemetrexed/platinum (AP), or docetaxel/platinum (DP) administration. The chemosensitivity consistency rate between the clinical responses and those obtained from zPDXs was 76.9% (10/13). There was a high correlation between patient responses and the corresponding zPDX drug responses. Thus, zPDX can accurately and quickly reproduce patient responses to treatment with EGFR TKIs, AP, and DP and has a considerable potential to serve as a biological platform for predicting treatment effect on patients with NSCLC.

Why Bayesian analysis of clinical trials may be a good idea– a worked example of the aducanumab studies

AbstractBackgroundThe FDA decision to approve Aducanumab has sparked controversy. Here, we argue that some of these controversies are rooted in the conclusions facilitated by classical frequentist analysis. We suggest that a Bayesian analysis would answer some of the burning questions the aducanumab trials have posed to the field.MethodWe applied Bayesian analysis of model plausibility and effect sizes based on simulated data (the original data are not publicly available) of the two phase 3 trials (trials 301 and 302) of aducanumab in prodromal and mild dementia stages of Alzheimer’s disease.ResultExcluding the rapid progressors from our simulated data did not produce decisive evidence supporting aducanumab’s treatment effect. The Bayesian analysis suggested that the study indicator (study 301 or study 302) was irrelevant as a predictor of treatment effect (Bayes Factor (BF)exc = 42.16). Our simulated data further offered overwhelming evidence against Biogen’s assumption that rapid progressors affected any treatment arm of either study disproportionately (i.e., BFexc = 9910.74). We also found anecdotal evidence against the treatment arms (i.e., BFexc = 2.65) as a predictor of CDR‐SB change (Figure 1). The prior and posterior plots show that if the effect existed, it would be very small, likely ranging from Kendall’s tau 0.007 to 0.052, with a median of 0.029 (Figure 2).ConclusionBayesian analysis can quantify evidence in favor of the presence or absence of an effect, as well as assess the strength of the effect. In the case of our simulated data, Bayesian analysis quantified evidence that studies 301 and 302 do not provide diverging evidence, and that the treatment arm‐specific rapid progressor effect was absent. The Bayesian analysis thus provides more informative and nuanced conclusions than the reported frequentist tests, which leave many questions unanswered.

MSR59 Replication and Extrapolation From RCTs to Observational Data to Predict Treatment Effects: Why, What, How

RCT Duplicate has shown that we can replicate the results of an RCT using observational data. If such a replication is possible, it provides trust in the evidence coming from the used observational data, but it doesn’t leverage the strength of observational data. The replication also provides an estimation of bias introduced by moving from trial to observational data. This trust and the estimate of bias helps to extrapolate the findings beyond the RCT used, building on the richness of observational data.

Individualising intensive systolic blood pressure reduction in hypertension using computational trial phenomaps and machine learning: a post-hoc analysis of randomised clinical trials

The cardiovascular benefits of intensive systolic blood pressure control vary across clinical populations tested in large randomised clinical trials. We aimed to evaluate the application of machine learning to clinical trials of patients without and with type 2 diabetes to define the personalised cardiovascular benefit of intensive control of systolic blood pressure. In SPRINT, a trial of intensive (systolic blood pressure <120 mm Hg) versus standard (systolic blood pressure <140 mm Hg) systolic blood pressure control in patients without type 2 diabetes, we defined a phenotypic representation of the study population using 59 baseline variables. We extracted personalised treatment effect estimates for the primary outcome, time-to-first major adverse cardiovascular event (MACE; cardiovascular death, myocardial infarction or acute coronary syndrome, stroke, and acute decompensated heart failure), through iterative Cox regression analyses providing average hazard ratio (HR) estimates weighted for the phenotypic distance of each participant from the index patient of each iteration. Next, we trained an extreme gradient boosting algorithm (known as XGBoost) to predict the personalised effect of intensive systolic blood pressure control using features most consistently linked to increased personalised benefit, before evaluating its performance in the ACCORD BP trial of patients with type 2 diabetes randomly assigned to receive intensive versus standard systolic blood pressure control. We stratified patients based on their predicted treatment effect, and key demographic groups (age, sex, cardiovascular disease, and smoking). We assessed the presence of heterogeneity with an interaction test, and assessed the performance of the algorithm in a simulation analysis of SPRINT in the presence or absence of an artificially introduced heterogeneous treatment effect. From SPRINT, we included all 9361 study participants (mean age 67·9 years [SD 9·4], 3332 [35·6%] female) who underwent randomisation to either intensive (n=4678) or standard (n=4683) treatment. The median individualised HR for MACE was 0·63 (IQR 0·53-0·78). An eight-feature tool built for this analysis to predict personalised benefit in SPRINT was externally tested in ACCORD BP (4733 participants (mean age 62·7 years [SD 6·7], 2258 [47·7%] female), wherein it successfully identified individuals with differential benefit from intensive versus standard systolic blood pressure control (adjusted HR for MACE of 0·70 [95% CI 0·55-0·90] in individuals with above-median MACE benefit versus 1·05 [95% CI 0·84-1·32] for below-median predicted benefit; pinteraction=0·0184). Subgroup analysis based on age (<65 years: HR 0·89 [95% CI 0·71-1·12]; ≥65 years: 0·85 [0·67-1·09]), sex (male: 0·89 [0·72-1·10]; female: 0·85 [0·65-1·10]), established cardiovascular disease (no: 0·89 [0·70-1·14]; yes: 0·84 [0·67-1·06]), or active smoking (no: 0·85 [0·71-1·02]; yes: 1·01 [0·64-1·60]) did not identify groups with heterogeneity of treatment effect. In a simulation analysis of SPRINT, the proposed algorithm detected groups with heterogeneous treatment effects in the presence, but not absence, of simulated subgroup differences. By use of machine learning to define an individual's personalised benefit through phenotypic representations of clinical trials, we created a practical tool for individualising the selection of intensive versus standard systolic blood pressure control in patients without and with type 2 diabetes. National Heart, Lung, and Blood Institute of the US National Institutes of Health.

One-Year estimated GFR Slope Independently Predicts Clinical Benefit in Immunoglobulin A Nephropathy

Immunoglobulin A nephropathy (IgAN) is a serious, progressive, immune complex–mediated kidney disease. Being a rare disease, a key challenge for the evaluation of potential new treatments for IgAN is that the progression to end-stage kidney disease typically evolves over many years. Therefore, there is interest in defining earlier surrogate endpoints to serve as reliable predictors of treatment effect on long-term IgAN kidney outcomes.

A Pilot randomized trial to examine effects of a hybrid closed-loop insulin delivery system on neurodevelopmental and cognitive outcomes in adolescents with type 1 diabetes

Type 1 diabetes (T1D) is associated with lower scores on tests of cognitive and neuropsychological function and alterations in brain structure and function in children. This proof-of-concept pilot study (ClinicalTrials.gov Identifier NCT03428932) examined whether MRI-derived indices of brain development and function and standardized IQ scores in adolescents with T1D could be improved with better diabetes control using a hybrid closed-loop insulin delivery system. Eligibility criteria for participation in the study included age between 14 and 17 years and a diagnosis of T1D before 8 years of age. Randomization to either a hybrid closed-loop or standard diabetes care group was performed after pre-qualification, consent, enrollment, and collection of medical background information. Of 46 participants assessed for eligibility, 44 met criteria and were randomized. Two randomized participants failed to complete baseline assessments and were excluded from final analyses. Participant data were collected across five academic medical centers in the United States. Research staff scoring the cognitive assessments as well as those processing imaging data were blinded to group status though participants and their families were not. Forty-two adolescents, 21 per group, underwent cognitive assessment and multi-modal brain imaging before and after the six month study duration. HbA1c and sensor glucose downloads were obtained quarterly. Primary outcomes included metrics of gray matter (total and regional volumes, cortical surface area and thickness), white matter volume, and fractional anisotropy. Estimated power to detect the predicted treatment effect was 0.83 with two-tailed, α = 0.05. Adolescents in the hybrid closed-loop group showed significantly greater improvement in several primary outcomes indicative of neurotypical development during adolescence compared to the standard care group including cortical surface area, regional gray volumes, and fractional anisotropy. The two groups were not significantly different on total gray and white matter volumes or cortical thickness. The hybrid closed loop group also showed higher Perceptual Reasoning Index IQ scores and functional brain activity more indicative of neurotypical development relative to the standard care group (both secondary outcomes). No adverse effects associated with study participation were observed. These results suggest that alterations to the developing brain in T1D might be preventable or reversible with rigorous glucose control. Long term research in this area is needed.

Dominance by non-native grasses suppresses long-term shifts in plant species composition and productivity in response to global change.

Climate warming and increased atmospheric nitrogen deposition are both predicted to alter the primary productivity of grass-dominated systems in the coming decades. In field experiments, while both factors can have substantial effects on productivity in the initial years, further changes can be delayed by lags in plant species composition responses. However, the effects of experiment age can be confounded by annual variability in environmental conditions, and a replicate experiment established at a later time is therefore needed to separate these effects. We added new warming and nitrogen plots to an ongoing (14year-old) field experiment in a grass-dominated temperate old field to compare the short-term vs. long-term treatment effects on plant productivity and species composition, while controlling for interannual environmental variability. We predicted treatment effects on relative species abundances would be most pronounced in the old plots. Although treatment responses of productivity (specifically to N addition) were highest in the old plots in the first year, by the second year there were no interactions between treatment and plot age. Moreover, there were no plot age effects on plant species composition, which was associated with continued dominance of non-native grasses. Our results therefore suggest that despite initial increases in productivity in response to global change, dominance by non-native grasses can suppress the emergence of further long-term treatment effects on productivity by inhibiting the responses of other species.

GP.2 Deep learning prediction of response to disease modifying therapy in primary progressive multiple sclerosis

Background: Only one disease modifying therapy (DMT), ocrelizumab, was found to slow disability progression in primary progressive multiple sclerosis (PPMS). Modeling the conditional average treatment effect (CATE) using deep learning could identify individuals more responsive to DMTs, allowing for predictive enrichment to increase the power of future clinical trials. Methods: Baseline clinical and MRI data were acquired as part of three placebo-controlled randomized clinical trials: ORATORIO (ocrelizumab), OLYMPUS (rituximab) and ARPEGGIO (laquinimod). Data from ORATORIO and OLYMPUS was separated into a training (70%) and testing (30%) set, while ARPEGGIO served as additional validation. An ensemble of multitask multilayer perceptrons was trained to predict the rate of disability progression on both treatment and placebo to estimate CATE. Results: The model could separate individuals based on their predicted treatment effect. The top 25% of individuals predicted to respond most have a larger effect size (HR 0.442, p=0.0497) than the entire group (HR 0.787, p=0.292). The model could also identify responders to laquinimod. A simulated study where the 50% most responsive individuals are randomized would require 6-times less participants to detect a significant effect. Conclusions: Individuals with PPMS who respond favourably to DMTs can be identified using deep learning based on their baseline clinical and imaging characteristics.

Schizophrenia and Alarmins.

Schizophrenia, consisting of a group of severe psychiatric disorders with a complex etiology, is a leading cause of disability globally. Due to the lack of objective indicators, accurate diagnosis and selection of effective treatments for schizophrenia remain challenging. The association between schizophrenia and alarmins levels has been proposed for many years, but without solid evidence. Alarmins are prestored molecules that do not require processing and can be released upon cell death or damage, making them an ideal candidate for an early initiator of inflammation. Immunological biomarkers seem to be related to disease progression and treatment effectiveness. Several studies suggest strong associations among the high-mobility group box 1 protein (HMGB1), interleukin-1α, interleukin-33, S100B, heat-shock proteins, and uric acid with schizophrenic disorders. The purpose of this review is to discuss the evidence of central and peripheral immune findings in schizophrenia, their potential causes, and the effects of immunomodulatory therapies on symptoms and outline potential applications of these markers in managing the illness. Although there are currently no effective markers for diagnosing or predicting treatment effects in patients with schizophrenia, we believe that screening immune-inflammatory biomarkers that are closely related to the pathological mechanism of schizophrenia can be used for early clinical identification, diagnosis, and treatment of schizophrenia, which may lead to more effective treatment options for people with schizophrenia.

Higher Resting Cardiovagal Activity Predicts Larger Decrease of Depressive Symptoms in Inpatients Treated for Stress-Related Depression

Abstract. Depression is one of the most prevalent mental disorders, with treatment outcomes generally being unsatisfactory. The identification of outcome predictors could contribute to improving diagnosis, treatment, and outcome. Heart rate variability (HRV), an index of cardiovagal activity, has been proposed as a potential correlate of depression as well as a predictor of treatment effectiveness. The aim of the present study was to examine if HRV at baseline could predict the outcome of inpatient treatment for stress-related depressive disorder (SRDD). Depressive symptoms of n = 57 inpatients with an SRDD, who were treated in a specialized burnout ward, were assessed using the Beck Depression Inventory (BDI) at the beginning, the end of treatment, and at 3-month follow-up. HRV (i.e., RMSSD, the root mean square of successive RR interval differences) was determined from a five-minute measurement in the supine position. RMSSD was not significantly associated with the BDI score at the beginning, end, and follow-up. Higher RMSSD was revealed to be a significant predictor of a stronger decrease in depressive severity from the beginning to the end of the treatment. Thereby, the regression model explained 7.6% of the total variance in the BDI decrease. The results revealed initial HRV to predict a larger decrease in depressive severity. Therefore, resting HRV represents a physiological resource and index of successful neurovisceral interaction, which supports inpatients in benefitting from specialized treatment.

Automated interpretable discovery of heterogeneous treatment effectiveness: A COVID-19 case study

Testing multiple treatments for heterogeneous (varying) effectiveness with respect to many underlying risk factors requires many pairwise tests; we would like to instead automatically discover and visualize patient archetypes and predictors of treatment effectiveness using multitask machine learning. In this paper, we present a method to estimate these heterogeneous treatment effects with an interpretable hierarchical framework that uses additive models to visualize expected treatment benefits as a function of patient factors (identifying personalized treatment benefits) and concurrent treatments (identifying combinatorial treatment benefits). This method achieves state-of-the-art predictive power for COVID-19 in-hospital mortality and interpretable identification of heterogeneous treatment benefits. We first validate this method on the large public MIMIC-IV dataset of ICU patients to test recovery of heterogeneous treatment effects. Next we apply this method to a proprietary dataset of over 3000 patients hospitalized for COVID-19, and find evidence of heterogeneous treatment effectiveness predicted largely by indicators of inflammation and thrombosis risk: patients with few indicators of thrombosis risk benefit most from treatments against inflammation, while patients with few indicators of inflammation risk benefit most from treatments against thrombosis. This approach provides an automated methodology to discover heterogeneous and individualized effectiveness of treatments.