Improved stratification of ALS clinical trials using predicted survival

Abstract

IntroductionIn small trials, randomization can fail, leading to differences in patient characteristics across treatment arms, a risk that can be reduced by stratifying using key confounders. In ALS trials, riluzole use (RU) and bulbar onset (BO) have been used for stratification. We hypothesized that randomization could be improved by using a multifactorial prognostic score of predicted survival as a single stratifier.MethodsWe defined a randomization failure as a significant difference between treatment arms on a characteristic. We compared randomization failure rates when stratifying for RU and BO (“traditional stratification”) to failure rates when stratifying for predicted survival using a predictive algorithm. We simulated virtual trials using the PRO‐ACT database without application of a treatment effect to assess balance between cohorts. We performed 100 randomizations using each stratification method – traditional and algorithmic. We applied these stratification schemes to a randomization simulation with a treatment effect using survival as the endpoint and evaluated sample size and power.ResultsStratification by predicted survival met with fewer failures than traditional stratification. Stratifying predicted survival into tertiles performed best. Stratification by predicted survival was validated with an external dataset, the placebo arm from the BENEFIT‐ALS trial. Importantly, we demonstrated a substantial decrease in sample size required to reach statistical power.ConclusionsStratifying randomization based on predicted survival using a machine learning algorithm is more likely to maintain balance between trial arms than traditional stratification methods. The methodology described here can translate to smaller, more efficient clinical trials for numerous neurological diseases.

An Application of Machine Learning in IVF: Comparing the Accuracy of Classification Alogithims for the Prediction of Twins

“…Another type of ensemble learning method uses tree-based classifiers, an approach that has become popular in many analytic projects. Tree-based methods segment the predictor space into a number of simpler regions [6]. Once this division has been accomplished, the prediction is made from the mean or mode of the region where the training observation exists.…”

“…Boosting tends to form smaller trees which may prove helpful for data interpretation. Lastly, a random forest improves accuracy by decorrelating the trees [6].…”

“…When applied to a dataset, trees may be more accurate than linear regression classification methods. As they require neither normalization nor standardization, decision trees are easier to use relative to linear models [6]. The ease of data input offered by decision trees makes them especially advantageous to clinicians.…”

Pilot trial of inosine to elevate urate levels in amyotrophic lateral sclerosis

“…A recent trend in the field of ALS is to explore the utility of prediction algorithms as an adjunct to clinical care and/or clinical trials 21, 50, 51, 52. We evaluated the apparent efficacy of serum urate elevation using a virtual control arm derived using a novel prediction algorithm.…”

“…care and/or clinical trials. 21,[50][51][52] We evaluated the apparent efficacy of serum urate elevation using a virtual control arm derived using a novel prediction algorithm. Observed ALSFRS-R total scores were remarkably close to those predicted for untreated patients based on the baseline characteristics of participants, suggesting no dramatic benefit from serum urate elevation; however, lack of an observable treatment effect of inosine on ALSFRS-R should not be over-interpreted because this trial lacked power to detect all but a very large effect over a short period of observation.…”