The following is a summary of “Exploring potential neuroimaging biomarkers for the response to non-steroidal anti-inflammatory drugs in episodic migraine,” published in the June 2024 issue of Pain by Le Wei, et al.
Non-steroidal anti-inflammatory drugs (NSAIDs) are a common first-line treatment for migraines, but their effectiveness varies greatly from person to person.
Researchers concluded a retrospective study to explore if a machine learning model based on brain scans (percentage of amplitude oscillations and gray matter volume) could predict how well patients responded to NSAIDs for migraine treatment.
They employed propensity score matching to pair patients with migraine-responding and not-responding to NSAIDs, ensuring clinical and migraine-related feature consistency. Utilizing multimodal MRI, PerAF and GMV were extracted, employing least absolute shrinkage and selection operator (LASSO) regression and recursive feature elimination for feature selection. Constructing multiple predictive models, they selected the one with the smallest predictive residuals. Model performance metrics included the receiver operating characteristic (ROCAUC) curve, area under the precision-recall curve (PRAUC), balance accuracy (BACC), sensitivity, F1 score, positive predictive value (PPV), and negative predictive value (NPV). External validation utilized a publicly available database. Subsequently, a correlation analysis between neuroimaging predictors and migraine clinical features was conducted.
The results showed 118 patients with migraine (59 responders and 59 non-responders). Six features (PerAF of left insula and left transverse temporal gyrus and GMV of right superior frontal gyrus, left postcentral gyrus, right postcentral gyrus, and left precuneus) were observed. The random forest model with the lowest predictive residuals was chosen, yielding training and testing metrics (ROCAUC, PRAUC, BACC, sensitivity, F1 score, PPV, and NPV) of 0.982, 0.983, 0.927, 0.976, 0.930, 0.889, and 0.973; and 0.711, 0.648, 0.639, 0.667, 0.649, 0.632, and 0.647, respectively. External validation metrics were 0.631, 0.651, 0.611, 0.808, 0.656, 0.553, and 0.706. Moreover, a significant positive correlation was noted between GMV of the left precuneus and attack time in non-responders.
Investigators found that brain imaging techniques (combining different measures) could potentially predict how well NSAIDs worked for migraines, offering new clues about migraine causes and better treatment approaches.
Source: thejournalofheadacheandpain.biomedcentral.com/articles/10.1186/s10194-024-01812-4
