An AI model can help predict adverse events from new drug combinations

Credit: Unsplash / CC0 Public Domain

Preliminary data from the human intelligence model may potentially predict the consequences of new collaborative approaches, according to the results presented at the AACR 2022 Annual Meeting, held April 8-13.

“Physicians have challenged the real problem that new therapies can lead to unpredictable results,” said Bart Westerman, Ph.D., lead author of the study and a professor at the Cancer Center Amsterdam. “Our methods can help us understand the relationship between the effects of different drugs depending on the nature of the disease.”

Many types of cancer are increasingly being treated with a combination of therapies, through which doctors are trying to improve the quality and reduce the likelihood of treatment. However, such collaborative approaches can add multiple medications simultaneously to the complex list of medications available to my patients. Clinical trials that test new drugs or interactions are difficult and the list of other medications that a patient may take outside of the experimental system is tested.

“My patients seeking treatment use four to six pills a day, which makes it difficult to decide if it is a new combination. far Westerman said, “It is difficult to determine whether the effective effect of combination therapy will prove its adverse effect on a patient.”

Westerman and colleagues – including graduate student Aslı Küçükosmanoğlu, who presented the study – sought to use a learning device to better predict the adverse events leading to new drug therapies. They collect data from the US Food and Drug Administration’s Adverse Event Reporting System (FAERS), a database containing more than 15 million incidents. Using a method called growth retardation, they collected recurring events to facilitate research and strengthen associations between the drug and its description of its effects.

The researchers then fed the data to a mobile network system, a type of machine learning that simulated the process. human brain make movements between data. Unsatisfactory results have been used for individual therapies to train the algorithm, which detects common signs between drugs and their side effects. The detected elements are grouped into the so-called “hidden space” which simplifies the calculation by representing each negative profile as a line of 225 numbers between 0 and 1, which can be converted to the original profile.

To test their model, the researchers provided a data set of the invisible elements of the therapeutic agents for their system, called “non-atlas-like atlas”, to see if can identify these new profiles and cut them well using space data. This suggests that the model can recognize these new patterns, suggesting that the joint measurement scale can be reversed to that of any treatment in the joint.

This, Westerman said, shows that there are serious consequences hade far can be easily predicted. “We were able to assess the total effect of human medicine by calculating the simple algebra of space designers,” he said. “Since this method reduces the amount of noise in the data because the algorithm is trained to detect global patterns, it can take the standard elements of a combination of therapies.”

Westerman and colleagues further validated their model by comparing the prognosis of clinical events to those seen at the clinic. Used data from FAERS and US clinical trial dataThe researchers suggested that the model could adapt standardized data for events for other commonly used collaborative methods.

Another complication of the combination therapy is the new, potentially unexpected side effects that can develop when medications are combined. Using additional markers as the model found, the researchers were able to differentiate the underlying effects from the combined effects of the drug combination. This, Westerman said, could help them understand what might happen if complex data sets of events are intervened.

Researchers are developing a computational model to quantify the accuracy of their design. “After that my landscape remedy The interaction is very complex and consists of many molecular, macromolecular, cellular, and organ systems, and our system is unlikely to lead to black and white decisions, “Westerman said. The finding is that we have been able to obtain images of drug interactions, illnesses, and human body as stated by millions of patients. “

Limitations of this study include the possibility of difficulties in comparing these data with the minimum data, as well as the application specifications. role model you medicine until further confirmation.

The study was funded by the University of Amsterdam, and Health ~ Holland. Westerman received public and private funding from Health ~ Holland for a review of a design project where Medstone is an independent party.

Combination therapies may improve outcomes due to independent, rather than collaborative or additional, drug action

Learn more:

hintThe AI ​​model can help predict the adverse events from the new pharmacokinetics (2022, April 8) retrieved 8 April 2022 from events-drug-combinations.html

This document is copyrighted. Apart from any genuine transaction for the purpose of personal analysis or investigation, no part may be reproduced without our written permission. Content is provided for informational purposes only.

An AI model can help predict adverse events from new drug combinations Source link An AI model can help predict adverse events from new drug combinations

Related Articles

Back to top button