Science & Enterprise subscription

Follow us on Twitter

  • A medical software team designed a mobile app that records and analyzes a person's sounds and sleep positions to de…
    about 5 hours ago
  • New post on Science and Enterprise: Mobile App Screens for Sleep Apnea #Science #Business
    about 5 hours ago
  • Reported in Science & Enterprise on 6 May ... Coronavirus Testing the Cheap, Simple Way
    about 8 hours ago
  • A company discovering therapeutic antibodies identified antibodies considered particularly effective in neutralizin…
    about 23 hours ago
  • New post on Science and Enterprise: Biotech IDs Potent Covid-19 Antibodies #Science #Business
    about 23 hours ago

Please share Science & Enterprise

Challenge Seeks Algorithm to Predict ALS Progression

Illustration of brain (NIDA)

(National Institute of Drug Abuse)

A new challenge competition seeks mathematical tools to predict the progression of amyotrophic lateral sclerosis (ALS) or Lou Gehrig’s disease based on the patient’s current disease status. The competition, handled through the open innovation company InnoCentive, has a prize of $25,000 and a deadline of 15 October 2012.

ALS, which affects 5 out of every 100,000 people worldwide, is a disease of the nerve cells in the brain and spinal cord that control voluntary muscle movement. In ALS, nerve cells waste away or die, and can no longer send messages to muscles, which leads to muscle weakening, twitching, and an inability to move the arms, legs, and body. When the muscles in the chest area stop working, it becomes hard or impossible to breathe on one’s own.

The challenge, known as the DREAM Phil Bowen ALS Prediction Prize4Life, is part of the Dialogue for Reverse Engineering Assessments and Methods  or DREAM project, sponsored by Columbia University Center for Multiscale Analysis Genomic and Cellular Networks, NIH Roadmap Initiative, IBM Computational Biology Center, and New York Academy of Sciences.

ALS patients experience varying degrees of progression of the disease, with some patients suffering complete paralysis in as few as two years, while for others the disease progression is much slower. In the early stages of the disease, it is difficult to determine whether a given patient will experience slow or fast disease progression. The ability to predict disease progression is also critical for those interested in planning ALS clinical trials for potential new treatments.

Information about disease progression is currently not provided to patients because of a lack of specific and reliable predictors. Rough predictions can be made with a scale called the ALS Functional Rating Scale (ALSFRS), but the estimations vary so widely over the short term that ALSFRS is normally not part of the clinical guidance provided to ALS patients.

This challenge aims to forecast disease progression more accurately with an approach that predicts a given patient’s disease status within a year’s time based on three months of data. Disease progression will be calculated as the average change in ALSFRS over a year’s time from enrollment in a clinical trial. At the end of the challenge, the prediction submitted, based on three months of data, will be compared against the actual ALSFRS slope experienced by the patient over a year.

The data available for analysis will include symptom onset date, medical and family history data, demographic data, and visit dates. In addition, the following data collected at multiple times throughout the course of the study will be provided: functional (ALSFRS) measures, body weight and vital signs, and lab data — blood chemistry, hematology, and urinalysis.

Challenge participants will need to propose an algorithm that should can take as input the covariates corresponding to a single patient, and generate an outcome for that patient. To be eligible for the prize, proposed solutions will need to perform better than an established benchmark generated by an off-the-shelf machine learning algorithm.

Participants will need to submit the algorithm in two forms (1) code written in the R statistical language, and (2) a write-up with the documentation to run the algorithm, and an explanation of the methods used to arrive at the submitted predictions.

Challenge participants can validate their model against a subset of patients that are neither part of the training set nor the final test, a step the challenge organizers encourage. InnoCentive will run the submitted R code against the interim validation data set, with the results of the partial validation appearing on a leaderboard along with the relative rankings of participants.

InnoCentive says challenge participants will not be required to transfer intellectual property to receive an award. However, receiving the full award will be contingent upon permission to submit the results and the algorithm to a peer-reviewed scientific journal, for which the participant will be expected to pay the publication costs.

Read more:

*     *     *

Please share Science & Enterprise ...

4 comments to Challenge Seeks Algorithm to Predict ALS Progression