Machine Learning Could Help Predict Migraineur Response to NSAIDs

Researchers leveraged machine learning to predict which migraineurs may have better responses to nonsteroidal anti-inflammatory drugs (NSAIDs).

Using machine learning (ML) researchers were able to determine disrupted functional connectivity (FC) of the amygdala may predict efficacy of nonsteroidal anti-inflammatory drugs (NSAIDs) among patients with migraine without aura (MwoA). Study findings were published in Frontiers in Molecular Neuroscience.

Although both opioids and sedatives are often used to treat migraine, NSAIDs are considered first-line treatment for the condition. However, due to inter-individual variability, NSAIDS can have unsatisfactory therapeutic outcomes in a considerable percentage of migraineurs, authors explained.

Some patients who overuse NSAIDs can suffer from treatment-resistant headaches and develop cognitive decline or depression.

To better understand interindividual variability affecting responses to NSAIDs and to optimize therapy administration, the researchers decided to use ML algorithms to predict the efficacy of this drug class in patients with MwoA.

Previous research has implicated the amygdala in analgesic response to NSAIDs and noted that antinociceptive tolerance to NSAIDs is associated with disrupted amygdala-related FC patterns, they said.

All 70 patients included in the study had MwoA, had not taken medications for at least 1 month prior to enrollment, and were headache free for at least 3 days prior to and following scanning. Thirty-three age, sex, and education level–matched healthy controls were also enrolled.

Patients were required to keep a headache diary, to undergo cognitive impairment screening, and to complete the Visual Analogue Scale (VAS), Headache Impact Test 6-item scale (HIT-6), and migraine disability assessment (MIDAS). Imaging data were collected via functional MRI scans.

Using these data, researchers developed multivariable logistic regression (MLR) and support vector machine (SVM) models, with 80% of migraineurs assigned to the training group and the remaining 20% to the testing group.

In total, 35 patients with migraine effectively managed with NSAIDs (M-eNSAIDs), 35 patients with MwoA with ineffective response to NSAIDs (M-ieNSAIDs), and 33 healthy participants were assessed.

Analyses revealed:

  • The M-eNSAIDs group exhibited enhanced FC with ipsilateral calcarine sulcus (CAL), superior parietal gyrus, paracentral lobule, and contralateral superior frontal gyrus (SFG) in the left amygdala
  • The M-eNSAIDs group showed decreased FC with ipsilateral caudate nucleus (CAU) compared with the M-ieNSAIDs group
  • The M-eNSAIDs group showed higher FC with left precentral gyrus and postcentral gyrus compared with controls
  • The M-ieNSAIDs group showed lower FC with the left anterior cingulate cortex and right SFG
  • Patients with MwoA showed increased FC with the left middle frontal gyrus in the right amygdala compared with controls
  • The disrupted left amygdala-related FC patterns exhibited significant correlations with migraine characteristics in the M-ieNSAIDs group

“MLR and SVM models discriminated clinical efficacy of NSAIDs with an area under the curve of 0.891 and 0.896, sensitivity of 0.971 and 0.833, and specificity of 0.629 and 0.875, respectively,” the authors wrote.

Data also showed a significant correlation between the FC of the left amygdala-CAL and MIDAS scores, and the FC of the amygdala-CAU and VAS scores in the M-ieNSAIDs group, with no significant correlation found between the FC of the right amygdala and clinical characteristics.

“These results suggest that the neural function of the left amygdala plays an important role in determining the effectiveness of NSAIDs,” while the findings provide insights into migraine’s underlying neurophysiological mechanisms, the researchers said.

Because the MLR model revealed amygdala-visual FCs, shown to positively affect NSAIDs’ efficacy, they hypothesized this finding could indicate amygdala-visual dysfunctional patters potentially affect migraineurs’ response to the drug class.

Due to the study’s cross-sectional nature, causality cannot be proven, while longitudinal investigations are warranted to verify any causal mechanism. These studies should also include patients with different migraine subtypes and a larger sample size.

“Alteration of amygdala-related FC patterns with the higher-level sensory cortex and prefrontal cortex were important central pathological features for distinguishing MwoA patients likely to have effective response to NSAIDs from MwoA patients likely to have ineffective response to NSAIDs,” the authors concluded. “These results may help in further elucidating the functional characteristics of migraine and predicting individualized response to NSAIDs.”

Reference

Wei H, Xu C, Wang J, et al. Disrupted functional connectivity of the amygdala predicts the efficacy of non-steroidal anti-inflammatory drugs in migraineurs without aura. Front Mol Neurosci. Published online February 24, 2022. doi:10.3389/fnmol.2022.819507