Machine Learning-Driven Metabolomic Evaluation of Cerebrospinal Fluid: Insights Into Poor Outcomes After Aneurysmal Subarachnoid Hemorrhage

Artificial Intelligence and Machine Learning Applications in Critically Ill Brain Injured Patients

“…156 Furthermore, ML has also been instrumental in discovering novel biomarkers that could enhance DCI and clinical outcome models and provide insights into the molecular and physiological changes in this complex condition. 157,158 Moving beyond static inputs, studies have explored the predictive value of continuous physiological data in forecasting DCI up to 12 hours before onset with high accuracy (83%) and PPV (2.7 true events for every false alert). 159,160 A growing number of studies continue to emphasize the potential role of ML in SAH DCI prediction, with a recent metaanalysis involving 1,828 patients showing that advanced ML models using standard clinical and imaging data surpass traditional logistic regression in DCI prediction.…”

Commentary: Machine Learning-Driven Metabolomic Evaluation of Cerebrospinal Fluid: Insights Into Poor Outcomes After Aneurysmal Subarachnoid Hemorrhage

“…To date, most metabolomics studies, including the present one in aSAH patients, primarily focus on identifying markers associated with the occurrence of aSAH or aSAH outcome. 9,4,23,24 Of the vast number of biomarkers associated with aSAH, only a few actually have any clinical implication. 25 It should be considered that correlating markers with outcome does not impact patient care if those markers do not point toward a therapeutic target.…”

“…8 The authors of the current study used cerebrospinal fluid (CSF) samples from 81 patients with aSAH and an external ventricular drain to measure and quantify 138 metabolites. 9 As a control group, CSF samples from 16 patients were used who had undergone elective clipping of unruptured intracranial aneurysms. The association of the metabolomics with functional neurological outcomes was determined using a machine learning algorithm.…”

Commentary: Serum Levels of Myo-inositol Predicts Clinical Outcome 1 Year After Aneurysmal Subarachnoid Hemorrhage

“…9 Metabolomics has been used in to predict aSAH outcomes in other papers with widely varying results and each with distinct significant markers leading to difficulty in applying the findings. 10-14 Of note, none of the previous papers mentioned MI as a metabolic predictor of aSAH outcomes. The results of this study must be interpreted with caution as a mechanism for the association between MI and vasospasm or DCI has not been established.…”

“…Despite these flaws, this paper contributes to our understanding of aSAH and provides a novel way to prognosticate the long-term outcomes following aSAH. Other papers using metabolomics found that outcomes are correlated with various biomarkers such as taurine and other amino acids; 12,14 symmetric dimethylarginine, diabetes mellitus cardiovascular marker, and ornithine; 10 and chemokines. 13 Given that these papers also measured these levels using venous blood draws, further work could examine trends in aggregate of all these markers.…”