Better biomarkers of ALS could enable smaller and more targeted clinical trials. Partially to address this aim, the Prize for Life foundation collected de-identified records from ALS sufferers who participated in clinical trials and made them available to researchers in the PRO-ACT database.
PRO-ACT data was used to predict decline in ALS-FRS over the next 9 months with accuracies exceeding chance and clinicians’ predictions and achieving root mean squared errors of ~54% (which is not so good).
Mei-Lyn Ong, Pei Fang Tan and Joanna D. Holbrook  attempted to build models to better predict decline in the ALSFRS-R score and to predict survival. They applied learning algorithms to the demographic, clinical and laboratory parameters in the training set to predict ALSFRS-R decline and the derived fast/slow progression and high/low death risk categories. They say their error in prediction is better than 20%.
What is interesting is that they published their results in 2017 in a comprehensive manner and it is possible to use them to provide information to guide a treatment from a reduced set of common biomarkers.
// Alkaline phosphatase
// Creatine Kinase
// Gamma Glutamyl Transferase
 Ong M-L, Tan PF, Holbrook JD (2017)
Predicting functional decline and survival in
amyotrophic lateral sclerosis. PLoS ONE 12(4):