NCT04531995

One Million Cancer Treatment Months

Development of an Artificial Intelligence-based Incident Prediction Algorithm to Improve Cancer Patient Care and Patient Safety

• Status: Unknown
• Phase: —
• Study type: Observational
• Enrollment: 166,000 (estimated)
• Sponsor: Cankado GmbH · Industry
• Sex: All
• Age: 18 Years
• Healthy volunteers: Not accepted

Summary

The OMCAT Register aims to provide learning databases in cancer comprising both PRO data using PRO-React and "ground truth" (outcome data verified by the physician during patient examinations). Intelligent learning and knowledge engineering procedures will utilize this PRO data to provide high-quality event prediction algorithms. The ground-truth data enables so-called "supervised learning" techniques of artificial intelligence, because predicted events can be verified with a high level of certainty from ground-truth data.

Detailed description

The next generation of PRO-React by CANKADO is designed to predict impending incident threats at an earlier stage than previously feasible and -- by more timely intervention -- help physicians to eliminate or mitigate the severity of an unfavourable event, reduce the required intensity of countermeasures, or otherwise reduce patient risks. A highly reliable identification of situations classified as "low-risk" by CANKADO could also enable a more focused utilization of resources as well as enhanced patient comfort and decreased stress, e.g., due to less frequent monitoring visits or reduced need for invasive diagnostics. The OMCAT Register aims to provide learning databases in cancer comprising both PRO data using PRO-React and "ground truth" (outcome data verified by the physician during patient examinations). Intelligent learning and knowledge engineering procedures will utilize this PRO data to provide high-quality event prediction algorithms. The ground-truth data enables so-called "supervised learning" techniques of artificial intelligence, because predicted events can be verified with a high level of certainty from ground-truth data. The PRO data of a patient provide what is known in engineering, physics, and statistics as "time series" of observations. The unique feature of PRO time series for applications in cancer is the very high "sampling frequency" (e.g., daily or better) compared to examinations, which generally occur at fixed, and much less frequent intervals. Prediction algorithms based on PRO data would thus be ideally suited to reduce the delay in detecting events, for example, by triggering physician appointments or indicating the need for more intensive medical diagnostics.

Conditions

• Cancer

Interventions

Type	Name	Description
DEVICE	CANKADO PRO-React Onco	CANKADO PRO-React Onco is approved as class I medical device within the European Union (registration number DE /CA59 /371/2020-R/Hd) and is compliant with the FDA classification for Mobile Medical Devices (2015) Appendix B. The purpose of CANKADO PRO-React Onco is to be an automated digital support for patients to help them decide how urgent it is to contact the attending physician based on the symptoms they independently record in the system. It supports patients with cancer under systemic, anti-tumor or anti-hormonal therapy in adjuvant, neoadjuvant, post-neoadjuvant or palliative situations. It is unsuitable for patients undergoing radiotherapy, cell and gene therapy, surgical procedures or alternative healing methods.

Timeline

• Start date: 2022-08-03
• Primary completion: 2025-12-01
• Completion: 2025-12-01
• First posted: 2020-08-31
• Last updated: 2022-09-27

Locations

4 sites across 1 country: Germany