Clinical Trial Dataset Research Articles

Determining Predictors of Weight Loss in a Behavioral Intervention: A Case Study in the Use of Lasso Regression.

ObjectiveThis study investigates predictors of weight loss among individuals with serious mental illness participating in an 18-month behavioral weight loss intervention, using Lasso regression to select the most powerful predictors.MethodsData were analyzed from the intervention group of the ACHIEVE trial, an 18-month behavioral weight loss intervention in adults with serious mental illness. Lasso regression was employed to identify predictors of at least five-pound weight loss across the intervention time span. Once predictors were identified, classification trees were created to show examples of how to classify participants into having likely outcomes based on characteristics at baseline and during the intervention.ResultsThe analyzed sample contained 137 participants. Seventy-one (51.8%) individuals had a net weight loss of at least five pounds from baseline to 18 months. The Lasso regression selected weight loss from baseline to 6 months as a primary predictor of at least five pound 18-month weight loss, with a standardized coefficient of 0.51 (95% CI: −0.37, 1.40). Three other variables were also selected in the regression but added minimal predictive ability.ConclusionsThe analyses in this paper demonstrate the importance of tracking weight loss incrementally during an intervention as an indicator for overall weight loss, as well as the challenges in predicting long-term weight loss with other variables commonly available in clinical trials. The methods used in this paper also exemplify how to effectively analyze a clinical trial dataset containing many variables and identify factors related to desired outcomes.

Propagating Uncertainty Across Cascaded Medical Imaging Tasks for Improved Deep Learning Inference.

Although deep networks have been shown to perform very well on a variety of medical imaging tasks, inference in the presence of pathology presents several challenges to common models. These challenges impede the integration of deep learning models into real clinical workflows, where the customary process of cascading deterministic outputs from a sequence of image-based inference steps (e.g. registration, segmentation) generally leads to an accumulation of errors that impacts the accuracy of downstream inference tasks. In this paper, we propose that by embedding uncertainty estimates across cascaded inference tasks, performance on the downstream inference tasks should be improved. We demonstrate the effectiveness of the proposed approach in three different clinical contexts: (i) We demonstrate that by propagating T2 weighted lesion segmentation results and their associated uncertainties, subsequent T2 lesion detection performance is improved when evaluated on a proprietary large-scale, multi-site, clinical trial dataset acquired from patients with Multiple Sclerosis. (ii) We show an improvement in brain tumour segmentation performance when the uncertainty map associated with a synthesised missing MR volume is provided as an additional input to a follow-up brain tumour segmentation network, when evaluated on the publicly available BraTS-2018 dataset. (iii) We show that by propagating uncertainties from a voxel-level hippocampus segmentation task, the subsequent regression of the Alzheimer's disease clinical score is improved.

Inference on latent factor models for informative censoring.

This work discusses the problem of informative censoring in survival studies. A joint model for the time to event and the time to censoring is presented. Their hazard functions include a latent factor in order to identify this joint model without sacrificing the flexibility of the parametric specification. Furthermore, a fully Bayesian formulation with a semi-parametric proportional hazard function is provided. Similar latent variable models have been described in literature, but here the emphasis is on the performance of the inferential task of the resulting mixture model with unknown number of components. The posterior distribution of the parameters is estimated using Hamiltonian Monte Carlo methods implemented in Stan. Simulation studies are provided to study its performance and the methodology is implemented for the analysis of the ACTG175 clinical trial dataset yielding a better fit. The results are also compared to the non-informative censoring case to show that ignoring informative censoring may lead to serious biases.

Explainable AI and Interpretable Machine Learning: A Case Study in Perspective

Explainable AI, as the word implies is a type of artificial intelligence which enables the explanation of learning models and focuses on why the system arrived at a particular decision, exploring its logical paradigms, contrary to the inherent black box nature of artificial intelligence. Similarly, machine learning interpretability allows users to comprehend the results of the learning models by providing reasoning for the decisions that it has arrived at. This nature of Explainable AI(XAI) and Interpretable Machine Learning (IML) is particularly helpful in the context of AI applications pertaining to healthcare and medical diagnosis. In this paper, we present a case study wherein we have focused on using the ELI5 XAI toolkit in conjunction with LIME (Local Interpretable Model-agnostic Explanations) and SHAP (SHapley Additive exPlanations) algorithmic frameworks in Python, for determining if a patient is diabetic or not, based on a randomized clinical trial dataset. We also endeavor to point out trends and most vital factors that can help clinicians and researchers in analyzing patient data, in conjunction with machine learning and artificial intelligence outputs. Having explanations for machine learning models allows for higher degree of interpretability and paves the way for accountability and transparency in medical and other fields of data analysis. We explore the aforementioned paradigms in the context of this research paper, paving the way for developing an accountable, transparent and robust data analytics framework using XAI & IML.

Military-civilian partner selection for mental health management: An analysis of organisational and technological legitimacy

Traumatic and chronic mental health problems have heightened the socio-economic cost, inducing military-civil collaboration. Since the military-civilian partnership offers conceptual development and technological development, we explored the partner-selection process in the military-civil strategic alliance for therapeutic products. Following the institutional approach, we applied legitimacy to the organisational and technological levels. First, we proposed that the military-civilian collaboration stands for the organisation's legitimacy (experience in the field). Second, we proposed the military-civilian collaboration predicts technological legitimacy (drugs versus psychotherapies). We used a worldwide clinical trial dataset (10,585 projects) on military-civilian cooperation in mental health projects, comprising clinical trials for all types of therapies for mental health management. Especially, PTSD, stress and anxiety were the main concepts in the conditions of the sampled clinical trials since the National Institutes of Health (NIH) in the USA began to maintain the database from 1997 onwards. For the legitimacy of the partner, we find support for the legitimacy of the organisation in the military-civilian partner selection. The military selects partners with an extensive portfolio of completed projects in the field. For the product's legitimacy, contrary to expectations, we find that mental health problems use psychotherapies more than drug therapies. Drug therapy has institutional support and higher legitimacy; psychotherapy is higher in the frequency of usage. The theoretical and practical role of inter-organisational legitimacy and inter-project (inter-technology) legitimacy suggest that the sponsor's prestige is more important than the technological prestige.

Diagnosis of breast cancer using machine learning algorithms based on features selected by Genetic Algorithm: Assessed on five datasets

Breast Cancer is one of the chronic diseases occurred to human beings throughout the world. Early detection of this disease is the most promising way to improve patients’ chances of survival. The strategy employed in this paper is to select the best features from various breast cancer datasets using a genetic algorithm and machine learning algorithm is applied to predict the outcomes. Two machine learning algorithms such as Support Vector Machines and Decision Tree are used along with Genetic Algorithm. The proposed work is experimented on five datasets such as Wisconsin Breast Cancer-Diagnosis Dataset, Wisconsin Breast Cancer-Original Dataset, Wisconsin Breast Cancer-Prognosis Dataset, ISPY1 Clinical trial Dataset, and Breast Cancer Dataset. The results exploit that SVM-GA achieves higher accuracy of 98.16% than DT-GA of 97.44%.

Self-Reported Race as a Social Determinant of Stroke Risk in Observational Versus Clinical Trial Datasets

ObjectivesSelf-reported Black (SRB) Americans are approximately twice as likely to have a stroke as self-reported White (SRW) Americans. While social determinants of health and vascular risk factors account for some of the disparity, half the increased risk remains unexplained and may be related to unmeasured real-world factors of the racialized experience. Materialsand Methods In this cohort study, we compared SRB and SRW participants in the Systolic Blood Pressure Intervention Trial (SPRINT) to the same groups in the observational Atherosclerosis Risk in Communities (ARIC) study to evaluate if clinical trial participation mitigates disparities in stroke risk. We set the ARIC baseline at visit 4 and excluded participants with prior stroke to create an ARIC cohort similar in age to SPRINT participants. The study outcome was incident primary stroke. We report hazard ratios from Cox models and inverse-probability weighted Cox models with propensity score matching on participant age, sex, diabetes, atrial fibrillation, and smoking. ResultsWe included 10,094 patients from ARIC and 8,869 from SPRINT, of which 26.1% were SRB. The risk of stroke between SRW participants in SPRINT versus ARIC was not significantly different (IPW-Weighted HR 0.78 [0.52–1.19]). SRB ARIC participants were twice as likely to have a stroke as SRW ARIC participants (IPW-Weighted HR = 1.96 [1.41–2.71]). However, SRB SPRINT participants did not have higher stroke risk compared to SRW SPRINT or ARIC participants (IPW-Weighted HR 0.99 [0.68–-1.77] and 0.95 [.57–1.59], respectively). SRB SPRINT participants in the intensive BP control group had a lower risk of stroke compared to SRB ARIC participants (IPW-Weighted HR = 0.39 [0.20–0.75]). ConclusionsSRB race, compared to SRW race, is associated with an increase in primary stroke risk in the ARIC study but not in the SPRINT trial. The absence of the racial disparity in stroke incidence in SPRINT indicates that aspects of the disparity are modifiable. Population-based interventions that test this hypothesis deserve further attention.

A matching procedure for sequential experiments that iteratively learns which covariates improve power.

We propose a dynamic allocation procedure that increases power and efficiency when measuring an average treatment effect in sequential randomized trials exploiting some subjects' previous assessed responses. Subjects arrive sequentially and are either randomized or paired to a previously randomized subject and administered the alternate treatment. The pairing is made via a dynamic matching criterion that iteratively learns which specific covariates are important to the response. We develop estimators for the average treatment effect as well as an exact test. We illustrate our method's increase in efficiency and power over other allocation procedures in both simulated scenarios and a clinical trial dataset. An R package "SeqExpMatch" for use by practitioners is available on CRAN.

Complete Versus Incomplete Hematologic Recovery after CAR T Cell Therapy: Implications for Relapse Free Survival and Overall Survival in Pediatric and Young Adult Patients with Relapsed/Refractory B-ALL

Background: Anti-CD19 Chimeric Antigen Receptor (CAR) T cell therapy has emerged as a mainstay in the treatment of patients with relapsed/refractory (r/r) B-cell acute lymphoblastic leukemia (ALL). While trials have shown encouraging survival rates, up to 50% of patients receiving CART eventually relapse. Identification of risk factors for subsequent relapse is crucial, as this could allow tailored surveillance and treatment approaches for high-risk patients, potentially leading to improved morbidity and mortality. Complete remission with incomplete hematologic recovery (CRi) has been associated with decreased remission duration and overall survival in acute leukemias. As count suppression following CAR T cell therapy is frequently seen, this study evaluated CRi as a prognostic marker for worse relapse-free (RFS) and overall survival (OS) after CAR T cell therapy in comparison to complete remission with hematologic recovery (CR).Methods: Patients with r/r ALL who achieved a complete morphologic remission after receiving anti-CD19 CAR T cell treatment with the 4-1BB-containing products CTL019 or humanized CART19 in the context of a clinical trial (NCT01626495, NCT02374333, NCT02228096, NCT02435849, NCT02906371) or commercial product (tisagenlecleucel) at Children's Hospital of Philadelphia from 4/2012-4/2019 were identified. Patients who received prior CAR therapy, and those with Trisomy 21 were excluded. Demographic, disease and treatment characteristics, and outcome data were abstracted from the medical record or clinical trial datasets. CR was defined as achieving a morphologic remission with both an absolute neutrophil count (ANC) ≥1,000/µL and a transfusion-independent platelet count ≥100,000/µL at any time between 27-33 days after CAR T cell infusion, whereas those achieving a morphologic remission without complete hematologic recovery were defined as CRi. RFS and OS were described for each cohort. Exposure-outcome association was assessed via the log-rank test and multivariable Cox proportional hazard regression.Results: Of the 206 patients included in the analysis, 104 (51%) achieved CR, 102 (49%) CRi. Forty patients (39%) met criteria for CRi with both ANC <1000/µL and platelets <100,000/µL, 40 (39%) for isolated thrombocytopenia, and 20 (20%) for neutropenia alone. Median age at leukemia diagnosis was 7 years (range 0.2-24), and at CAR T cell infusion was 12 years (range 1-29), with 48% undergoing a prior allogeneic stem cell transplant (HSCT), 32% with high risk cytogenetics, and 10% with prior blinatumomab treatment, none of which differed between the CR and CRi groups (Table 1). The number of relapses prior to CAR T cell referral was different between the groups (p=0.003), with more CR patients having been referred for primary refractory disease (25% vs 9.8%), and more CRi patients for first relapse (34% vs 18%). More CRi patients had >25% bone marrow blasts (43% vs 19%) at infusion, whereas more CR patients were MRD-negative (50% vs 30%) at infusion (p=0.002). CRi patients were also more likely to have Grade 3/4 CRS (38% vs 6.7%). Median length of follow-up for patients with CR was 39 months (range 7-89), which was not statistically significantly different than for those patients with CRi (41 months, 11-98, p=0.875). There was no difference in RFS when stratified by hematologic recovery (Figure 1, p=0.2165), with RFS at 36 months for CR of 57% (47-69) and CRi of 46% (36-59). OS was significantly lower (Figure 2, p=0.0081) for those with CRi, with 36-month OS for CR of 81% (74-89), and for CRi of 63% (54-73). In multivariable analysis adjusting for sex, prior blinatumomab, relapse number, disease burden at infusion, and maximum CRS grade, CR was not associated with either RFS (HR 0.76 [95%CI 0.50-1.17] p=0.2182) or OS (HR 0.74 [95%CI 0.43-1.29] p=0.2908), in comparison to CRi, Table 2.Discussion: Complete remission with incomplete hematologic recovery, manifesting as neutropenia and/or thrombocytopenia, at the first disease assessment following CAR T cell infusion should not be regarded as a harbinger of relapse and demonstrates that patients with CRi have similar probability of durable remission without further therapy. Anticipatory guidance should be provided to patients, and their families, that CRi is more frequently seen in patients who experience high grade CRS, who have high disease burden at infusion, and who are treated in first relapse. [Display omitted] DisclosuresCallahan: Novartis: Speakers Bureau. Rheingold: Optinose: Other: Spouse's current employment; Pfizer: Research Funding. Grupp: Novartis, Roche, GSK, Humanigen, CBMG, Eureka, and Janssen/JnJ: Consultancy; Novartis, Adaptimmune, TCR2, Cellectis, Juno, Vertex, Allogene and Cabaletta: Other: Study steering committees or scientific advisory boards; Novartis, Kite, Vertex, and Servier: Research Funding; Jazz Pharmaceuticals: Consultancy, Other: Steering committee, Research Funding. Maude: Wugen: Consultancy; Novartis Pharmaceuticals Corporation: Consultancy, Research Funding.

Body Composition and Risk of All-Cause Mortality in Men Treated With Radiation Therapy for Prostate Cancer: A Pooled Analysis of NRG/RTOG 9406 and NRG/RTOG 0126

To validate the association between body composition and all-cause mortality in men treated with radiation therapy for localized prostate cancer (PCa). Secondarily, to integrate body composition as a factor to help classify patients by risk of all-cause mortality.All participants of NRG/RTOG 9406 and NRG/RTOG 0126 with archived computed tomography that extended cranially to include the L4-L5 interface were included. Muscle mass and muscle density were estimated by measuring the cross-sectional area and average attenuation of the paired psoas muscles on a single slice at the level of L4-L5. Bone density was estimated by measuring the average attenuation of the vertebral body cancellous bone on a single slice at mid-L5. Adipose tissue density was estimated by measuring the average attenuation of the subcutaneous adipose tissue on a single slice at L4-L5. Survival analyses, including Cox proportional hazards models, were performed to assess the relationship between body composition variables and all-cause mortality. Recursive partitioning (RPA) was utilized to create a classification tree to classify NRG/RTOG 0126 participants by risk of death, and the discriminant ability of the classification model was validated using the NRG/RTOG 9406 data set.Data from 2,066 men was included in this study (864 from NRG/RTOG 9406 and 1,202 from NRG/RTOG 0126). A total of 648 men died in the follow-up period and 51 (7.9%) were due to PCa. Psoas area, psoas density, and vertebral body density were individually associated with overall survival. In the final multivariable model, psoas area, comorbidity score, and age were associated with overall survival (Table). The RPA yielded a classification tree with 4 prognostic groups determined by age, comorbidity, and psoas cross-sectional area. When the RPA classification was applied to the NRG/RTOG 9406 validation set the discriminant ability was preserved (P < 0.001 groupwise log-rank).The results of this study strongly support that body composition is related to all-cause mortality in men with localized PCa, with most deaths due to causes other than PCa. The inclusion of psoas cross-sectional area in the RPA classification tree suggests that body composition provides additive information to age and comorbidity status for mortality prediction. This study also confirms the feasibility of performing body composition analysis using archived CT scans using NRG Oncology clinical trial data sets. These methods can be applied to other NRG Oncology data sets to further explore how body composition is related to patient outcomes.

Clinical Validation of Deep Learning Algorithms for Lung Cancer Radiotherapy Targeting

<h3>Purpose/Objective(s)</h3> Automated target segmentation for non-small cell lung cancer (NSCLC) patients has the potential to support radiation treatment planning. Artificial intelligence (AI) has demonstrated great promise in medical image segmentation tasks. However, most studies have been confined to <i>in silico</i> validation in small internal cohorts, lacking data on real-world clinical utility. In this study, we developed primary tumor and involved lymph node segmentation algorithms in computed tomography (CT) images. Validation is performed in multiple large multi-institutional cohorts to assess model generalizability. <h3>Materials/Methods</h3> Simulation CTs and ground truth annotations were collected from multiple public and private sources (total n = 2584). We employed the following benchmarks: Inter-observer (6 radiation oncologists, n = 20, median volumetric dice 0.83, 95% CI 0.82-0.84) and intra-observer (1 radiation oncologist, 3 reads, n = 21, median volumetric dice 0.88, 95% CI 0.84-0.9). We developed two segmentation algorithms: seed-point assisted and fully automated. Model training data (n = 787) comprised NSCLC-Radiomics (stages I-IIIB, n = 422) and LungRT-1 (stages IA-IV, n = 365). Validation was first performed in an internal dataset annotated by a single thoracic radiation oncologist (LungRT-1, n = 136). Additional validation included: (1) an internal dataset annotated by other radiation oncologists, including generalists, in our center (LungRT-2, n = 1075), (2) an external clinical trial dataset from 185 different institutions (RTOG-0617, n = 403), and (3) a dataset of early-stage surgical patients annotated for diagnostic purposes by radiologists (NSCLC-Radiogenomics, n = 142). Volumetric dice, using expert manual segmentations as ground truth, was used as an evaluation metric. <h3>Results</h3> The model performance is comparable to the benchmarks when validated on internal data, with degrading performance in cohorts annotated by other radiation oncologists. <h3>Conclusion</h3> The results highlight the importance of assessing segmentation style among annotators and understanding model generalizability in external cohorts, all while cautioning against degrading performance in increasingly external data. Differences between radiologists and radiation oncologists performing the same segmentation task underscore the importance of clinical context in AI model deployment. Further validation includes studying the dosimetric impact of AI-generated segmentations, and conducting human subject experiments to assess AI output acceptance and time savings.

Differentiating severe and non-severe lower respiratory tract illness in patients hospitalized with influenza: Development of the Influenza Disease Evaluation and Assessment of Severity (IDEAS) scale.

BackgroundExperimental studies have shown that vaccination can reduce viral replication to attenuate progression of influenza-associated lower respiratory tract illness (LRTI). However, clinical studies are conflicting, possibly due to use of non-specific outcomes reflecting a mix of large and small airway LRTI lacking specificity for acute lung or organ injury.MethodsWe developed a global ordinal scale to differentiate large and small airway LRTI in hospitalized adults with influenza using physiologic features and interventions (PFIs): vital signs, laboratory and radiographic findings, and clinical interventions. We reviewed the literature to identify common PFIs across 9 existing scales of pneumonia and sepsis severity. To characterize patients using this scale, we applied the scale to an antiviral clinical trial dataset where these PFIs were measured through routine clinical care in adults hospitalized with influenza-associated LRTI during the 2010–2013 seasons.ResultsWe evaluated 12 clinical parameters among 1020 adults; 210 (21%) had laboratory-confirmed influenza, with a median severity score of 4.5 (interquartile range, 2–8). Among influenza cases, median age was 63 years, 20% were hospitalized in the prior 90 days, 50% had chronic obstructive pulmonary disease, and 22% had congestive heart failure. Primary influencers of higher score included pulmonary infiltrates on imaging (48.1%), heart rate ≥110 beats/minute (41.4%), oxygen saturation <93% (47.6%) and respiratory rate >24 breaths/minute (21.0%). Key PFIs distinguishing patients with severity < or ≥8 (upper quartile) included infiltrates (27.1% vs 90.0%), temperature ≥ 39.1°C or <36.0°C (7.1% vs 27.1%), respiratory rate >24 breaths/minute (7.9% vs 47.1%), heart rate ≥110 beats/minute (29.3% vs 65.7%), oxygen saturation <90% (14.3% vs 31.4%), white blood cell count >15,000 (5.0% vs 27.2%), and need for invasive or non-invasive mechanical ventilation (2.1% vs 15.7%).ConclusionWe developed a scale in adults hospitalized with influenza-associated LRTI demonstrating a broad distribution of physiologic severity which may be useful for future studies evaluating the disease attenuating effects of influenza vaccination or other therapeutics.

Can machine learning augment clinician adjudication of events in cardiovascular trials? A case study of major adverse cardiovascular events (MACE) across CVRM trials

Abstract Background and introduction Accurate identification of clinical outcome events is critical to obtaining reliable results in cardiovascular outcomes trials (CVOTs). Current processes for event adjudication are expensive and hampered by delays. As part of a larger project to more reliably identify outcomes, we evaluated the use of machine learning to automate event adjudication using data from the SOCRATES trial (NCT01994720), a large randomized trial comparing ticagrelor and aspirin in reducing risk of major cardiovascular events after acute ischemic stroke or transient ischemic attack (TIA). Purpose We studied whether machine learning algorithms could replicate the outcome of the expert adjudication process for clinical events of ischemic stroke and TIA. Could classification models be trained on historical CVOT data and demonstrate performance comparable to human adjudicators? Methods Using data from the SOCRATES trial, multiple machine learning algorithms were tested using grid search and cross validation. Models tested included Support Vector Machines, Random Forest and XGBoost. Performance was assessed on a validation subset of the adjudication data not used for training or testing in model development. Metrics used to evaluate model performance were Receiver Operating Characteristic (ROC), Matthews Correlation Coefficient, Precision and Recall. The contribution of features, attributes of data used by the algorithm as it is trained to classify an event, that contributed to a classification were examined using both Mutual Information and Recursive Feature Elimination. Results Classification models were trained on historical CVOT data using adjudicator consensus decision as the ground truth. Best performance was observed on models trained to classify ischemic stroke (ROC 0.95) and TIA (ROC 0.97). Top ranked features that contributed to classification of Ischemic Stroke or TIA corresponded to site investigator decision or variables used to define the event in the trial charter, such as duration of symptoms. Model performance was comparable across the different machine learning algorithms tested with XGBoost demonstrating the best ROC on the validation set for correctly classifying both stroke and TIA. Conclusions Our results indicate that machine learning may augment or even replace clinician adjudication in clinical trials, with potential to gain efficiencies, speed up clinical development, and retain reliability. Our current models demonstrate good performance at binary classification of ischemic stroke and TIA within a single CVOT with high consistency and accuracy between automated and clinician adjudication. Further work will focus on harmonizing features between multiple historical clinical trials and training models to classify several different endpoint events across trials. Our aim is to utilize these clinical trial datasets to optimize the delivery of CVOTs in further cardiovascular drug development. Funding Acknowledgement Type of funding sources: Private company. Main funding source(s): AstraZenca Plc

Automatic radiotherapy delineation quality assurance on prostate MRI with deep learning in a multicentre clinical trial

Volume delineation quality assurance (QA) is particularly important in clinical trial settings where consistent protocol implementation is required, as outcomes will affect future as well current patients. Currently, where feasible, this is conducted manually, which is time consuming and resource intensive. Although previous studies mostly focused on automating delineation QA on CT, magnetic resonance imaging (MRI) is being increasingly used in radiotherapy treatment. In this work, we propose to perform automatic delineation QA on prostate MRI for both the clinical target volume (CTV) and organs-at-risk (OARs) by using delineations generated by 3D Unet variants as benchmarks for QA. These networks were trained on a small gold standard atlas set and applied on a multicentre radiotherapy clinical trial dataset to generate benchmark delineations. Then, a QA stage was designed to recommend ‘pass’, ‘minor correction’ and ‘major correction’ for each manual delineation in the trial set by thresholding its Dice similarity coefficient to the network generated delineation. Among all 3D Unet variants explored, the Unet with anatomical gates in an AtlasNet architecture performed the best in delineation QA, achieving an area under the receiver operating characteristics curve of 0.97, 0.92, 0.89 and 0.97 for identifying unacceptable (major correction) delineations with a sensitivity of 0.93, 0.73, 0.74 and 0.90 at a specificity of 0.93, 0.86, 0.86 and 0.95 for bladder, prostate CTV, rectum and gel spacer respectively. To the best of our knowledge, this is the first study to propose automated delineation QA for a multicentre radiotherapy clinical trial with treatment planning MRI. The methods proposed in this work can potentially improve the accuracy and consistency of CTV and OAR delineation in radiotherapy treatment planning.

Diversity and Inclusion in Drug Development: Rethinking Intrinsic and Extrinsic Factors with Patient Centricity.

Diversity and Inclusion in Drug Development: Rethinking Intrinsic and Extrinsic Factors with Patient Centricity.

CauchyCP: A powerful test under non-proportional hazards using Cauchy combination of change-point Cox regressions.

Non-proportional hazards data are routinely encountered in randomized clinical trials. In such cases, classic Cox proportional hazards model can suffer from severe power loss, with difficulty in interpretation of the estimated hazard ratio since the treatment effect varies over time. We propose CauchyCP, an omnibus test of change-point Cox regression models, to overcome both challenges while detecting signals of non-proportional hazards patterns. Extensive simulation studies demonstrate that, compared to existing treatment comparison tests under non-proportional hazards, the proposed CauchyCP test (a) controls the type I error better at small levels (); (b) increases the power of detecting time-varying effects; and (c) is more computationally efficient than popular methods like MaxCombo for large-scale data analysis. The superior performance of CauchyCP is further illustrated using retrospective analyses of two randomized clinical trial datasets and a pharmacogenetic biomarker study dataset. The R package CauchyCP is publicly available on CRAN.

Emergency medical services employing intra-arrest transport less frequently for out-of-hospital cardiac arrest have higher survival and favorable neurological outcomes

BackgroundThere is substantial regional variation in out-of-hospital cardiac arrest (OHCA) survival. We investigated whether regional emergency medical services (EMS) intra-arrest transport (IAT) practices are associated with patient outcomes. MethodsWe performed a secondary analysis of a multi-center North American clinical trial dataset, which enrolled EMS-treated adult OHCA cases from 49 regional population-based clusters. The exposure of interest was regional-level intra-arrest transport (IAT), calculated as the proportion of cases in each cluster transported to hospital prior to return of spontaneous circulation, examined as quartiles and as a continuous variable. Multilevel mixed-effects logistic regression modeling estimated the association between regional IAT with survival to hospital discharge and favorable neurologic status (modified Rankin Scale ≤ 3) at hospital discharge. ResultsOf 26,148 subjects (median age 68 years; 36% female; 23% shockable initial rhythm) 2424 (9.3%), survived to hospital discharge and 1993 (7.6%) had favourable neurological outcomes. Across regional clusters, IAT ranged from 0.84% to 75% (quartiles <6.2%, 6.2–19.6%, 19.6–30.4%, and ≥30.4%). For each quartile, 13.3%, 7.9%, 7.4%, and 4.8% survived, and 10.4%, 7.8%, 7.4%, and 4.8% had favourable neurological status. Regional IAT (per 10% change) was associated with decreased probability of survival (AOR 0.86, 95% CI 0.82–0.91) and favorable neurological outcome (AOR 0.80, 95% CI 0.76–0.85). ConclusionTreatment within a region that utilizes IAT less frequently was associated with improved clinical outcomes at hospital discharge. These findings may account for some of the known regional variation in OHCA outcomes.

Associations Between Movement Impairments and Function, Treatment Recommendations, and Treatment Plans for People With Femoroacetabular Impingement Syndrome.

The purpose of this study was to describe movement impairments for persons with femoroacetabular impingement syndrome and their association with function, treatment recommendations, and treatment plans. This report is a secondary, observational analysis of a clinical trial dataset in which participants received an interdisciplinary evaluation from a surgeon and physical therapist. The therapist documented frontal and sagittal plane movement impairments across 6 functional tasks. Associations between number of impairments in each plane and function (33-item International Hip Outcome Tool [iHOT33]) were evaluated using Pearson or Spearman correlations. Joint provider recommendations (physical therapist and surgeon) and participant-reported treatment plans were dichotomized based on the inclusion of physical therapy or not. Logistic regressions were used to examine the effects of (1) iHOT33, total movement impairments, and previous physical therapist treatment on joint provider recommendation and (2) these same variables along with joint provider recommendation on participant treatment plan; prevalence ratios and 95% CIs were reported for significant contributors. Thirty-nine participants demonstrated an average iHOT33 of 35.0 (SD=19.5) and presented with a median 5 frontal and 3 sagittal plane impairments. More frontal plane impairments were associated with worse iHOT33 scores. Twenty-seven participants received a joint provider recommendation that included physical therapy; no significant contributors to these recommendations were identified. Twenty-four of the 27 participants with a physical therapist recommendation included physical therapy in their treatment plan. Two additional participants did not receive a physical therapist recommendation but included physical therapy in their plan. Joint provider recommendation was the only significant contributor to the participant-reported plan (prevalence ratio=7.06; 95% CI = 3.25-7.97). Persons with femoroacetabular impingement syndrome displayed clinically observable movement impairments that were associated with worse function. Joint provider recommendations strongly influenced participants' treatment plans to pursue physical therapy. Physical therapists contribute new information to surgical examinations regarding movement. Joint recommendations from the physical therapist and the surgeon can influence patients' decisions to pursue physical therapy.

Collagen fiber orientation disorder from H&E images is prognostic for early stage breast cancer: clinical trial validation

Collagen fiber organization has been found to be implicated in breast cancer prognosis. In this study, we evaluated whether computerized features of Collagen Fiber Orientation Disorder in Tumor-associated Stroma (CFOD-TS) on Hematoxylin & Eosin (H&E) slide images were prognostic of Disease Free Survival (DFS) in early stage Estrogen Receptor Positive (ER+) Invasive Breast Cancers (IBC). A Cox regression model named MCFOD-TS, was constructed using cohort St (N = 78) to predict DFS based on CFOD-TS features. The prognostic performance of MCFOD-TS was validated on cohort Sv (N = 219), a prospective clinical trial dataset (ECOG 2197). MCFOD-TS was prognostic of DFS in both St and Sv, independent of clinicopathological variables. Additionally, the molecular pathways regarding cell cycle regulation were identified as being significantly associated with MCFOD-TS derived risk scores. Our results also found that collagen fiber organization was more ordered in patients with short DFS. Our study provided a H&E image-based pipeline to derive a potential prognostic biomarker for early stage ER+ IBC without the need of special collagen staining or advanced microscopy techniques.

Distinct patterns of disease activity over time in patients with active SLE revealed using latent class trajectory models

BackgroundSystemic lupus erythematosus (SLE) is a heterogeneous systemic autoimmune condition for which there are limited licensed therapies. Clinical trial design is challenging in SLE due at least in part to imperfect outcome measures. Improved understanding of how disease activity changes over time could inform future trial design. The aim of this study was to determine whether distinct trajectories of disease activity over time occur in patients with active SLE within a clinical trial setting and to identify factors associated with these trajectories.MethodsLatent class trajectory models were fitted to a clinical trial dataset of a monoclonal antibody targeting CD22 (Epratuzumab) in patients with active SLE using the numerical BILAG-2004 score (nBILAG). The baseline characteristics of patients in each class and changes in prednisolone over time were identified. Exploratory PK-PD modelling was used to examine cumulative drug exposure in relation to latent class membership.ResultsFive trajectories of disease activity were identified, with 3 principal classes: non-responders (NR), slow responders (SR) and rapid-responders (RR). In both the SR and RR groups, significant changes in disease activity were evident within the first 90 days of the trial. The SR and RR patients had significantly higher baseline disease activity, exposure to epratuzumab and activity in specific BILAG domains, whilst NR had lower steroid use at baseline and less change in steroid dose early in the trial.ConclusionsLongitudinal nBILAG scores reveal different trajectories of disease activity and may offer advantages over fixed endpoints. Corticosteroid use however remains an important confounder in lupus trials and can influence early response. Changes in disease activity and steroid dose early in the trial were associated with the overall disease activity trajectory, supporting the feasibility of performing adaptive trial designs in SLE.