Months of trial and error and unwanted side effects in finding the right antidepressant medication may soon be a thing of the past for people with severe depression, thanks to a new artificial intelligence model that can determine how well a drug works for an individual within a week. "This is important news for patients. Normally, it takes six to eight weeks to know whether antidepressants are effective," said Liesbeth Reneman, professor of neuroradiology at the University Medical Center Amsterdam (UMC).

In the latest clinical application of artificial intelligence, researchers at Amsterdam University Medical Center (UMC) and Radboud University Medical Center (RadboudUMC) have developed an algorithm that can determine whether a certain antidepressant drug is effective in the long term based on patients' MRI scans and other data. An estimated 11% of people in the United States have a prescription to manage depression, but about 60% cannot find the right medication on their first try. This is enough to put many people away from trying it a second or third time due to the time invested and side effects experienced.

The researchers set out to investigate whether this artificial intelligence model could first be used to predict the efficacy of the selective serotonin reuptake inhibitor (SSRI) sertraline (also known as Zoloft).

For the study, the team used results from a previous US study of 229 patients with depression, using MRI brain scans and clinical data recordings before patients were given sertraline or a placebo. This data is then fed into artificial intelligence, and the algorithm focuses specifically on the anterior cingulate cortex and the severity of symptoms.

"The algorithm showed that those with greater blood flow to the anterior cingulate cortex, a region of the brain involved in emotion regulation, would be helped by the drug," said Eric Ruch, a psychiatrist at Radboud University Medical Center. "The severity of their symptoms was confirmed in the second measurement, a week after starting treatment."

Using this approach, the AI ​​determined that sertraline was only effective in one-third of the participants, sparing two-thirds of the participants a two-month wait and see period. While antidepressants can take up to six months to be most effective, serious side effects can last for a long time and have a huge impact on daily life, often as much as the depression itself.

"With this approach, we have been able to avoid two-thirds of the 'erroneous' prescriptions of sertraline, thereby providing patients with higher quality treatment, because this drug also has side effects," Reineman said.

Treatment of major depressive disorder is extremely challenging due to its complexity. Treatment is more difficult due to the wide variety of available medical interventions, including SSRIs, serotonin and norepinephrine reuptake inhibitors (SNRIs), atypical antidepressants, tricyclic antidepressants, and monoamine oxidase inhibitors (MAOIs).

While the algorithm is specific to sertraline, the researchers hope to not only tweak it to be more personalized but also apply the same model to a variety of drugs used to treat depression.

The study was published in the American Journal of Psychiatry.