AI brain scan can predict whether antidepressant drug sertraline will work

Maarten G. Poirot, Mississippi

Credit: Research Gate

Thanks to artificial intelligence (AI), patients no longer have to wait weeks to see if antidepressants improve their symptoms. depressive symptoms– Brain scans programmed with an AI algorithm can predict whether the antidepressant drug sertraline will work within a week, according to a new study.¹

“This is important news for patients,” researcher Liesbeth Lenemann, professor of neuroradiology at UMC Amsterdam, said in a press release.² “It usually takes six to eight weeks to find out if antidepressants are working.”

Sertraline is one of the most frequently prescribed drugs in the United States and Europe and can cause unpleasant side effects such as dry mouth, nausea, headache, insomnia, and diarrhea.³ For some patients, sertraline may not improve symptoms of depression. Currently, antidepressant prescriptions are test-and-trial, with patients prescribed sertraline and tested six to eight weeks later, or up to several months later, to see if the drug is working.¹

If your symptoms do not go away, you will be prescribed another antidepressant. Trying a new antidepressant can take several months and can prevent an individual from fully participating in life.

Researchers led by Maarten G. Poirot, MS, at the University of Amsterdam, are conducting a second round of the Establishing Moderators and Biosignatures of Antidepressant Response in Clinical Treatment (EMBARC) study, a multicenter, double-blind, placebo-controlled, randomized trial. The analysis was carried out. . The primary study evaluated MRI scans and clinical data in 229 patients with major depressive disorder in the United States before and after one week of treatment with sertraline or a placebo. The mean age of the sample was her 38 years and 66% were female.

Researchers collected data on how well brain scans predicted response and remission after taking sertraline for eight weeks, using balanced accuracy and area under the receiver operating characteristic curve scores. did.

“The algorithm shows that the blood flow in the anterior cingulate cortex, an area of ​​the anterior cingulate cortex, [the] “The brain, which is involved in emotional regulation, could predict the effectiveness of drugs,” Radoboudumuk psychiatrist Eric Rouhe said in a press release.² “And a second measurement, one week after the start, showed that the severity of symptoms was even more predictable.”

Ultimately, analysis revealed that 1/3 of patients responded to sertraline and 2/3 did not respond.

“This approach has already prevented two-thirds of 'erroneous' sertraline prescriptions, allowing us to provide higher quality care to our patients,” said Leneman.

Therefore, instead of wasting time and money taking sertraline, patients can try newer antidepressants that may be more effective in terms of treatment response and fewer adverse events.

When predicting sertraline response, the accuracy of the brain scan algorithm was balanced at 68%, with an area under the receiver operating characteristic curve of 0.73. This suggests that the algorithm predicted sertraline response and remission significantly better than chance.¹ Balanced accuracy for patients receiving placebo was 62% and area under the receiver operating characteristic curve was 0.66. Furthermore, the balanced accuracy for patients switching from placebo to sertraline was 65% and the area under the receiver operating characteristic curve was 0.63, demonstrating a difference between sertraline and placebo treatment.

After follow-up, researchers found that one in three people with depression did not see their symptoms improve despite trying many antidepressants. Therefore, the research team emphasized the importance of predicting the effects of antidepressants as early as one week. Investigators plan to enhance the algorithm with additional information soon.

“Study results confirm that early sertraline treatment response can be predicted. The model is sertraline specific compared to placebo. Integration of multimodal MRI data and clinical data enables prediction. . and perfusion images contribute most to these predictions,” the researchers wrote. “Using this approach, we have the potential to personalize sertraline treatment plans and improve psychiatric care with lean and effective protocols.”

^References

1. ^{Poirot MG, Rouhe HG, Muzarz HMM, et al. Predicting treatment response in major depressive disorder using multimodal MRI and clinical data: a secondary analysis of a randomized clinical trial. J Psychiatry. Published online on February 7, 2024. doi:10.1176/appi.ajp.20230206}
2. ^{Artificial intelligence can help predict whether antidepressants will work for patients. Yurek Alert! February 7, 2024. https://www.eurekalert.org/news-releases/1033410. Accessed February 14, 2024.}
3. ^{Sertraline side effects. National Health Service. https://www.nhs.uk/. Accessed February 14, 2024.}