Longitudinal Image Data for Outcome Modeling.

In oncology, medical imaging is crucial for diagnosis, treatment planning and therapy execution. Treatment responses can be complex and varied and are known to involve factors of treatment, patient characteristics and tumor microenvironment. Longitudinal image analysis is able to track temporal changes, aiding in disease monitoring, treatment evaluation, and outcome prediction. This allows for the enhancement of personalized medicine. However, analyzing longitudinal 2D and 3D images presents unique challenges, including image registration, reliable segmentation, dealing with variable imaging intervals, and sparse data. This review presents an overview of techniques and methodologies in longitudinal image analysis, with a primary focus on outcome modeling in radiation oncology.

ROE (Radiotherapy Outcomes Estimator): An open-source tool for optimizing radiotherapy prescriptions.

BACKGROUND AND OBJECTIVES: Radiotherapy prescriptions currently derive from population-wide guidelines established through large clinical trials. We provide an open-source software tool for patient-specific prescription determination using personalized dose-response curves. METHODS: We developed ROE, a plugin to the Computational Environment for Radiotherapy Research to visualize predicted tumor control and normal tissue complication simultaneously, as a function of prescription dose. ROE can be used natively with MATLAB and is additionally made accessible in GNU Octave and Python, eliminating the need for commercial licenses. It provides a curated library of published and validated predictive models and incorporates clinical restrictions on normal tissue outcomes. ROE additionally provides batch-mode tools to evaluate and select among different fractionation schemes and analyze radiotherapy outcomes across patient cohorts. CONCLUSION: ROE is an open-source, GPL-copyrighted tool for interactive exploration of the dose-response relationship to aid in radiotherapy planning. We demonstrate its potential clinical relevance in (1) improving patient awareness by quantifying the risks and benefits of a given treatment protocol (2) assessing the potential for dose escalation across patient cohorts and (3) estimating accrual rates of new protocols.

Blood- and Imaging-Derived Biomarkers for Oncological Outcome Modelling in Oropharyngeal Cancer: Exploring the Low-Hanging Fruit.

AIMS: To assess whether CT-based radiomics and blood-derived biomarkers could improve the prediction of overall survival (OS) and locoregional progression-free survival (LRPFS) in patients with oropharyngeal cancer (OPC) treated with curative-intent RT. METHODS: Consecutive OPC patients with primary tumors treated between 2005 and 2021 were included. Analyzed clinical variables included gender, age, smoking history, staging, subsite, HPV status, and blood parameters (baseline hemoglobin levels, neutrophils, monocytes, and platelets, and derived measurements). Radiomic features were extracted from the gross tumor volumes (GTVs) of the primary tumor using pyradiomics. Outcomes of interest were LRPFS and OS. Following feature selection, a radiomic score (RS) was calculated for each patient. Significant variables, along with age and gender, were included in multivariable analysis, and models were retained if statistically significant. The models' performance was compared by the C-index. RESULTS: One hundred and five patients, predominately male (71%), were included in the analysis. The median age was 59 (IQR: 52-66) years, and stage IVA was the most represented (70%). HPV status was positive in 63 patients, negative in 7, and missing in 35 patients. The median OS follow-up was 6.3 (IQR: 5.5-7.9) years. A statistically significant association between low Hb levels and poorer LRPFS in the HPV-positive subgroup (p = 0.038) was identified. The calculation of the RS successfully stratified patients according to both OS (log-rank p < 0.0001) and LRPFS (log-rank p = 0.0002). The C-index of the clinical and radiomic model resulted in 0.82 [CI: 0.80-0.84] for OS and 0.77 [CI: 0.75-0.79] for LRPFS. CONCLUSIONS: Our results show that radiomics could provide clinically significant informative content in this scenario. The best performances were obtained by combining clinical and quantitative imaging variables, thus suggesting the potential of integrative modeling for outcome predictions in this setting of patients.

Virtual clinical trial based on outcome modeling with iteratively redistributed extrapolation data.

Virtual clinical trials (VCTs) can potentially simulate clinical trials on a computer, but their application with a limited number of past clinical cases is challenging due to the biased estimation of the statistical population. In this study, we developed ExMixup, a novel training technique based on machine learning, using iteratively redistributed extrapolated data. Information obtained from 100 patients with prostate cancer and 385 patients with oropharyngeal cancer was used to predict the recurrence after radiotherapy. Model performance was evaluated by developing outcome prediction models based on three types of training methods: training with original data (baseline), interpolation data (Mixup), and interpolation + extrapolation data (ExMixup). Two types of VCTs were conducted to predict the treatment response of patients with distinct characteristics compared to the training data obtained from patient cohorts categorized under risk classification or cancer stage. The prediction models developed with ExMixup yielded concordance indices (95% confidence intervals) of 0.751 (0.719-0.818) and 0.752 (0.734-0.785) for VCTs on the prostate and oropharyngeal cancer datasets, respectively, which significantly outperformed the baseline and Mixup models (P < 0.01). The proposed approach could enhance the ability of VCTs to predict treatment results in patients excluded from past clinical trials.

Improved Survival Prediction by Combining Radiological Imaging and S-100B Levels Into a Multivariate Model in Metastatic Melanoma Patients Treated With Immune Checkpoint Inhibition.

Purpose: We explored imaging and blood bio-markers for survival prediction in a cohort of patients with metastatic melanoma treated with immune checkpoint inhibition. Materials and Methods: 94 consecutive metastatic melanoma patients treated with immune checkpoint inhibition were included into this study. PET/CT imaging was available at baseline (Tp0), 3 months (Tp1) and 6 months (Tp2) after start of immunotherapy. Radiological response at Tp2 was evaluated using iRECIST. Total tumor burden (TB) at each time-point was measured and relative change of TB compared to baseline was calculated. LDH, CRP and S-100B were also analyzed. Cox proportional hazards model and logistic regression were used for survival analysis. Results: iRECIST at Tp2 was significantly associated with overall survival (OS) with C-index=0.68. TB at baseline was not associated with OS, whereas TB at Tp1 and Tp2 provided similar predictive power with C-index of 0.67 and 0.71, respectively. Appearance of new metastatic lesions during follow-up was an independent prognostic factor (C-index=0.73). Elevated LDH and S-100B ratios at Tp2 were significantly associated with worse OS: C-index=0.73 for LDH and 0.73 for S-100B. Correlation of LDH with TB was weak (r=0.34). A multivariate model including TB change, S-100B, and appearance of new lesions showed the best predictive performance with C-index=0.83. Conclusion: Our analysis shows only a weak correlation between LDH and TB. Additionally, baseline TB was not a prognostic factor in our cohort. A multivariate model combining early blood and imaging biomarkers achieved the best predictive power with regard to survival, outperforming iRECIST.

Alternative analytic and matching approaches for the prevalent new-user design: A simulation study.

PURPOSE: To describe the creation of prevalent new user (PNU) cohorts and compare the relative bias and computational efficiency of several alternative analytic and matching approaches in PNU studies. METHODS: In a simulated cohort, we estimated the effect of a treatment of interest vs a comparator among those who switched to the treatment of interest using the originally proposed time-conditional propensity score (TCPS) matching, standardized morbidity ratio weighting (SMRW), disease risk scores (DRS), and several alternative propensity score matching approaches. For each analytic method, we compared the average RR (across 2000 replicates) to the known risk ratio (RR) of 1.00. RESULTS: SMRW and DRS yielded unbiased results (RR = 0.998 and 0.997, respectively). TCPS matching with replacement was also unbiased (RR = 0.999). TCPS matching without replacement was unbiased when matches were identified starting with patients with the shortest treatment history as initially proposed (RR = 0.999), but it resulted in very slight bias (RR = 0.983) when starting with patients with the longest treatment history. Similarly, creating a match pool without replacement starting with patients with the shortest treatment history yielded an unbiased estimate (RR = 0.997), but matching with the longest treatment history first resulted in substantial bias (RR = 0.903). The most biased strategy was matching after selecting one random comparator observation per individual that continued on the comparator (RR = 0.802). CONCLUSIONS: Multiple analytic methods can estimate treatment effects without bias in a PNU cohort. Still, researchers should be wary of introducing bias when selecting controls for complex matching strategies beyond the initially proposed TCPS.

Predictive Outcome Modeling of Preoperative Clinical Symptoms and Electrodiagnostic Data in Tarsal Tunnel Surgery.

Background The relationship between tarsal tunnel syndrome (TTS), electrodiagnostic (Edx) findings, and surgical outcome is unknown. Analysis of TTS surgical release outcome patient satisfaction and comparison to Edx nerve conduction studies (NCSs) is important to improve outcome prediction when deciding who would benefit from TTS release. Methods Retrospective study of 90 patients over 7 years that had tarsal tunnel (TT) release surgery with outcome rating and preoperative tibial NCS. Overall, 64 patients met study inclusion criteria with enough NCS data to be classified into one of the following three groups: (1) probable TTS, (2) peripheral polyneuropathy, or (3) normal. Most patients had preoperative clinical provocative testing including diagnostic tibial nerve injection, tibial Phalen's sign, and/or Tinel's sign and complaints of plantar tibial neuropathic symptoms. Outcome measure was percentage of patient improvement report at surgical follow-up visit. Results Patient-reported improvement was 92% in the probable TTS group ( n = 41) and 77% of the non-TTS group ( n = 23). Multivariate modeling revealed that three out of eight variables predicted improvement from surgical release, NCS consistent with TTS ( p = 0.04), neuropathic symptoms ( p = 0.045), and absent Phalen's test ( p = 0.001). The R 2 was 0.21 which is a robust result for this outcome measurement process. Conclusion The best predictors of improvement in patients with TTS release were found in patients that had preoperative Edx evidence of tibial neuropathy in the TT and tibial nerve plantar symptoms. Determining what factors predict surgical outcome will require prospective evaluation and evaluation of patients with other nonsurgical modalities.

Novel methodology to assess the effect of contouring variation on treatment outcome.

PURPOSE: Contouring variation is one of the largest systematic uncertainties in radiotherapy, yet its effect on clinical outcome has never been analyzed quantitatively. We propose a novel, robust methodology to locally quantify target contour variation in a large patient cohort and find where this variation correlates with treatment outcome. We demonstrate its use on biochemical recurrence for prostate cancer patients. METHOD: We propose to compare each patient's target contours to a consistent and unbiased reference. This reference was created by auto-contouring each patient's target using an externally trained deep learning algorithm. Local contour deviation measured from the reference to the manual contour was projected to a common frame of reference, creating contour deviation maps for each patient. By stacking the contour deviation maps, time to event was modeled pixel-wise using a multivariate Cox proportional hazards model (CPHM). Hazard ratio (HR) maps for each covariate were created, and regions of significance found using cluster-based permutation testing on the z-statistics. This methodology was applied to clinical target volume (CTV) contours, containing only the prostate gland, from 232 intermediate- and high-risk prostate cancer patients. The reference contours were created using ADMIRE® v3.4 (Elekta AB, Sweden). Local contour deviations were computed in a spherical coordinate frame, where differences between reference and clinical contours were projected in a 2D map corresponding to sampling across the coronal and transverse angles every 3°. Time to biochemical recurrence was modeled using the pixel-wise CPHM analysis accounting for contour deviation, patient age, Gleason score, and treated CTV volume. RESULTS: We successfully applied the proposed methodology to a large patient cohort containing data from 232 patients. In this patient cohort, our analysis highlighted regions where the contour variation was related to biochemical recurrence, producing expected and unexpected results: (a) the interface between prostate-bladder and prostate-seminal vesicle interfaces where increase in the manual contour relative to the reference was related to a reduction of risk of biochemical recurrence by 4-8% per mm and (b) the prostate's right, anterior and posterior regions where an increase in the manual contour relative to the reference contours was related to an increase in risk of biochemical recurrence by 8-24% per mm. CONCLUSION: We proposed and successfully applied a novel methodology to explore the correlation between contour variation and treatment outcome. We analyzed the effect of contour deviation of the prostate CTV on biochemical recurrence for a cohort of more than 200 prostate cancer patients while taking basic clinical variables into account. Applying this methodology to a larger dataset including additional clinically important covariates and externally validating it can more robustly identify regions where contour variation directly relates to treatment outcome. For example, in the prostate case we use to demonstrate our novel methodology, external validation will help confirm or reject the counter-intuitive results (larger contours resulting in higher risk). Ultimately, the results of this methodology could inform contouring protocols based on actual patient outcomes.

A unified multi-activation (UMA) model of cell survival curves over the entire dose range for calculating equivalent doses in stereotactic body radiation therapy (SBRT), high dose rate brachytherapy (HDRB), and stereotactic radiosurgery (SRS).

PURPOSE: Application of linear-quadratic (LQ) model to large fractional dose treatments is inconsistent with observed cell survival curves having a straight portion at high doses. We have proposed a unified multi-activation (UMA) model to fit cell survival curves over the entire dose range that allows us to calculate EQD2 for hypofractionated SBRT, SRT, SRS, and HDRB. METHODS: A unified formula of cell survival S = n / e D D o + n - 1 using only the extrapolation number of n and the dose slope of Do was derived. Coefficient of determination, R2 , relative residuals, r, and relative experimental errors, e, normalized to survival fraction at each dose point, were calculated to quantify the goodness in modeling of a survival curve. Analytical solutions for α and ß, the coefficients respectively describe the linear and quadratic parts of the survival curve, as well as the α/ß ratio for the LQ model and EQD2 at any fractional doses were derived for tumor cells undertaking any fractionated radiation therapy. RESULTS: Our proposed model fits survival curves of in-vivo and in-vitro tumor cells with R2  > 0.97 and r < e. The predicted α, ß, and α/ß ratio are significantly different from their values in the LQ model. Average EQD2 of 20-Gy SRS of glioblastomas and melanomas metastatic to the brain, 10-Gy × 5 SBRT of the lung cancer, and 7-Gy × 5 HDRB of endometrial and cervical carcinomas are 36.7 (24.3-48.5), 114.1 (86.6-173.1),, and 45.5 (35-52.6) Gy, different from the LQ model estimates of 50.0, 90.0, and 49.6 Gy, respectively. CONCLUSION: Our UMA model validated through many tumor cell lines can fit cell survival curves over the entire dose range within their experimental errors. The unified formula theoretically indicates a common mechanism of cell inactivation and can estimate EQD2 at all dose levels.

Injury severity analysis of pedestrian and bicyclist trespassing crashes at non-crossings: A hybrid predictive text analytics and heterogeneity-based statistical modeling approach.

Non-motorists involved in rail-trespassing crashes are usually more vulnerable to receiving major or fatal injuries. Previous research has used traditional quantitative crash data for understanding factors contributing to injury outcomes of non-motorists in train involved collisions. However, usually overlooked crash narratives can provide useful and unique contextual crash-specific information regarding factors associated with injury outcomes. The main objective of this study is to harness the rapid advancements in more sophisticated qualitative analysis procedures for identifying thematic concepts in unstructured crash narrative data. A two-staged hybrid approach is proposed where text mining is applied first to extract valuable information from crash narratives followed by inclusion of the new variables derived from text mining in formulation of advanced statistical models for injury outcomes. By using ten-year (2006-2015) non-motorist non-crossing trespassing injury data obtained from the Federal Railroad Administration, statistical procedures and advanced machine learning text analytics are applied to extract unique information on contributory factors of trespassers' injury outcomes. The key concepts are systematically categorized into trespasser, injury, train, medical, and location related factors. A total of 13 unique variables are extracted from the thematic concepts that are not present in traditional tabular crash data. The analysis reveals a positive statistically significant association between presence of crash narrative and trespasser's injury outcome (coded as minor, major, and fatal injury). Compared to crashes with minor injuries, crashes involving major and fatal injuries are more likely to be reported with crash narratives. A crosstabulation of new variables derived from text mining with injury outcomes revealed that trespassers with confirmed suicide attempts, trespassers wearing headphones, or talking on cell phones are more likely to receive fatal injuries. Among other factors identified, trespassers under alcohol influence, trespasser hit by commuter train, and advance warnings by engineer are associated with more severe (major and fatal) trespasser injury outcomes. Accounting for unobserved heterogeneity and controlling for other factors, fixed and random parameter discrete outcome models are developed to understand the heterogeneous correlations between trespasser injury outcomes and the new crash specific explanatory variables derived from text mining - providing deeper insights. Practical implications and future research directions are discussed.

Registering Study Analysis Plans (SAPs) Before Dissecting Your Data-Updating and Standardizing Outcome Modeling.

Public preregistration of study analysis plans (SAPs) is widely recognized for clinical trials, but adopted to a much lesser extent in observational studies. Registration of SAPs prior to analysis is encouraged to not only increase transparency and exactness but also to avoid positive finding bias and better standardize outcome modeling. Efforts to generally standardize outcome modeling, which can be based on clinical trial and/or observational data, have recently spurred. We suggest a three-step SAP concept in which investigators are encouraged to (1) Design the SAP and circulate it among the co-investigators, (2) Log the SAP with a public repository, which recognizes the SAP with a digital object identifier (DOI), and (3) Cite (using the DOI), briefly summarize and motivate any deviations from the SAP in the associated manuscript. More specifically, the SAP should include the scope (brief data and study description, co-investigators, hypotheses, primary outcome measure, study title), in addition to step-by-step details of the analysis (handling of missing data, resampling, defined significance level, statistical function, validation, and variables and parameterization).

Machine learning for radiation outcome modeling and prediction.

AIMS: This review paper intends to summarize the application of machine learning to radiotherapy outcome modeling based on structured and un-structured radiation oncology datasets. MATERIALS AND METHODS: The most appropriate machine learning approaches for structured datasets in terms of accuracy and interpretability are identified. For un-structured datasets, deep learning algorithms are explored and a critical view of the use of these approaches in radiation oncology is also provided. CONCLUSIONS: We discuss the challenges in radiotherapy outcome prediction, and suggest to improve radiation outcome modeling by developing appropriate machine learning approaches where both accuracy and interpretability are taken into account.

SpineCloud: image analytics for predictive modeling of spine surgery outcomes.

Purpose: Data-intensive modeling could provide insight on the broad variability in outcomes in spine surgery. Previous studies were limited to analysis of demographic and clinical characteristics. We report an analytic framework called "SpineCloud" that incorporates quantitative features extracted from perioperative images to predict spine surgery outcome. Approach: A retrospective study was conducted in which patient demographics, imaging, and outcome data were collected. Image features were automatically computed from perioperative CT. Postoperative 3- and 12-month functional and pain outcomes were analyzed in terms of improvement relative to the preoperative state. A boosted decision tree classifier was trained to predict outcome using demographic and image features as predictor variables. Predictions were computed based on SpineCloud and conventional demographic models, and features associated with poor outcome were identified from weighting terms evident in the boosted tree. Results: Neither approach was predictive of 3- or 12-month outcomes based on preoperative data alone in the current, preliminary study. However, SpineCloud predictions incorporating image features obtained during and immediately following surgery (i.e., intraoperative and immediate postoperative images) exhibited significant improvement in area under the receiver operating characteristic (AUC): AUC = 0.72 ( CI 95 = 0.59 to 0.83) at 3 months and AUC = 0.69 ( CI 95 = 0.55 to 0.82) at 12 months. Conclusions: Predictive modeling of lumbar spine surgery outcomes was improved by incorporation of image-based features compared to analysis based on conventional demographic data. The SpineCloud framework could improve understanding of factors underlying outcome variability and warrants further investigation and validation in a larger patient cohort.

Combining handcrafted features with latent variables in machine learning for prediction of radiation-induced lung damage.

PURPOSE: There has been burgeoning interest in applying machine learning methods for predicting radiotherapy outcomes. However, the imbalanced ratio of a large number of variables to a limited sample size in radiation oncology constitutes a major challenge. Therefore, dimensionality reduction methods can be a key to success. The study investigates and contrasts the application of traditional machine learning methods and deep learning approaches for outcome modeling in radiotherapy. In particular, new joint architectures based on variational autoencoder (VAE) for dimensionality reduction are presented and their application is demonstrated for the prediction of lung radiation pneumonitis (RP) from a large-scale heterogeneous dataset. METHODS: A large-scale heterogeneous dataset containing a pool of 230 variables including clinical factors (e.g., dose, KPS, stage) and biomarkers (e.g., single nucleotide polymorphisms (SNPs), cytokines, and micro-RNAs) in a population of 106 nonsmall cell lung cancer (NSCLC) patients who received radiotherapy was used for modeling RP. Twenty-two patients had grade 2 or higher RP. Four methods were investigated, including feature selection (case A) and feature extraction (case B) with traditional machine learning methods, a VAE-MLP joint architecture (case C) with deep learning and lastly, the combination of feature selection and joint architecture (case D). For feature selection, Random forest (RF), Support Vector Machine (SVM), and multilayer perceptron (MLP) were implemented to select relevant features. Specifically, each method was run for multiple times to rank features within several cross-validated (CV) resampled sets. A collection of ranking lists were then aggregated by top 5% and Kemeny graph methods to identify the final ranking for prediction. A synthetic minority oversampling technique was applied to correct for class imbalance during this process. For deep learning, a VAE-MLP joint architecture where a VAE aimed for dimensionality reduction and an MLP aimed for classification was developed. In this architecture, reconstruction loss and prediction loss were combined into a single loss function to realize simultaneous training and weights were assigned to different classes to mitigate class imbalance. To evaluate the prediction performance and conduct comparisons, the area under receiver operating characteristic curves (AUCs) were performed for nested CVs for both handcrafted feature selections and the deep learning approach. The significance of differences in AUCs was assessed using the DeLong test of U-statistics. RESULTS: An MLP-based method using weight pruning (WP) feature selection yielded the best performance among the different hand-crafted feature selection methods (case A), reaching an AUC of 0.804 (95% CI: 0.761-0.823) with 29 top features. A VAE-MLP joint architecture (case C) achieved a comparable but slightly lower AUC of 0.781 (95% CI: 0.737-0.808) with the size of latent dimension being 2. The combination of handcrafted features (case A) and latent representation (case D) achieved a significant AUC improvement of 0.831 (95% CI: 0.805-0.863) with 22 features (P-value = 0.000642 compared with handcrafted features only (Case A) and P-value = 0.000453 compared to VAE alone (Case C)) with an MLP classifier. CONCLUSION: The potential for combination of traditional machine learning methods and deep learning VAE techniques has been demonstrated for dealing with limited datasets in modeling radiotherapy toxicities. Specifically, latent variables from a VAE-MLP joint architecture are able to complement handcrafted features for the prediction of RP and improve prediction over either method alone.

Artificial Neural Network with Composite Architectures for Prediction of Local Control in Radiotherapy.

In this study, we investigated the application of artificial neural networks (ANNs) with composite architectures into the prediction of local control (LC) of lung cancer patients after radiotherapy. The motivation of this study was to take advantage of the temporal associations among longitudinal (sequential) data to improve the predictive performance of outcome models under the circumstance of limited sample sizes. Two composite architectures: (1) a one dimension (1D) convolutional + fully connected and (2) a locally-connected+ fully connected architectures were implemented for this purpose. Compared with the fully-connected architecture (multi-layer perceptron [MLP]), our composite architectures yielded better predictive performance of LC in lung cancer patients who received radiotherapy. Specifically, in a cohort of 98 patients (29 patients failed locally), the composite architecture of 1D convolutional layers and fully-connected layers achieved an AUC (area under receiver operating characteristic curve) of 0.83 (95% confidence interval (CI): 0.807~0.841) with 18 features (14 features are longitudinal data). Whereas, the composite architecture of locally- connected layers and fully-connected layers achieved an AUC of 0.80 (95%CI: 0.775~0.811). Both outperformed an MLP in the prediction performance with the same set of features, which achieved an AUC of 0.78 (95%CI: 0.751~0.790); (P-values for differences in AUC using the DeLong tests were 1.609 × 10-14and 1.407 × 10-4, respectively).

Embracing Biological and Methodological Variance in a New Approach to Pre-Clinical Stroke Testing.

High-profile failures in stroke clinical trials have discouraged clinical translation of neuroprotectants. While there are several plausible explanations for these failures, we believe that the fundamental problem is the way clinical and pre-clinical studies are designed and analyzed for heterogeneous disorders such as stroke due to innate biological and methodological variability that current methods cannot capture. Recent efforts to address pre-clinical rigor and design, while important, are unable to account for variability present even in genetically homogenous rodents. Indeed, efforts to minimize variability may lessen the clinical relevance of pre-clinical models. We propose a new approach that recognizes the important role of baseline stroke severity and other factors in influencing outcome. Analogous to clinical trials, we propose reporting baseline factors that influence outcome and then adapting for the pre-clinical setting a method developed for clinical trial analysis where the influence of baseline factors is mathematically modeled and the variance quantified. A new therapy's effectiveness is then evaluated relative to the pooled outcome variance at its own baseline conditions. In this way, an objective threshold for robustness can be established that must be overcome to suggest its effectiveness when expanded to broader populations outside of the controlled environment of the PI's laboratory. The method is model neutral and subsumes sources of variance as reflected in baseline factors such as initial stroke severity. We propose that this new approach deserves consideration for providing an objective method to select agents worthy of the commitment of time and resources in translation to clinical trials.