Assessing the Diagnostic Accuracy of Brain MR Spectroscopy for Various Neurologic Pathologies: A Comparative Study with Final Clinical Diagnoses in Patients Referred to Shiraz University Hospitals (2021-2024)

IntroductionMagnetic resonance spectroscopy (MRS) of the brain offers significant potential for monitoring metabolic alterations associated with neurological pathologies. However, to date, its application has predominantly been limited to research settings and a select number of clinical centers. This study aimed to assess the diagnostic accuracy of brain MRS in detecting various neurological diseases. MethodsBrain MRS examinations conducted at public hospitals affiliated with Shiraz University, Iran, from 2021 to 2024 were retrospectively collected. The findings, as reported by attending radiologists, were compared with the final clinical or histopathological diagnoses of the patients. Diagnostic performance metrics, including sensitivity, specificity, positive predictive value, negative predictive value, and overall accuracy, were calculated. ResultsBrain MRS correctly identified the pathology in 61% of cases. The technique demonstrated high specificity, positive predictive value, negative predictive value, and overall accuracy--each exceeding 90%--across a range of brain pathologies. These included both primary and secondary tumoral processes, as well as non-tumoral conditions such as pyogenic brain abscess, toxoplasmosis, demyelinating diseases, and encephalitis. However, sensitivity varied considerably, ranging from 33% to 100%, with the lowest sensitivity observed in cases of pyogenic brain abscesses, brain hemorrhages, and central nervous system (CNS) vasculitis. ConclusionThis study indicates that certain neurological pathologies may exhibit atypical spectroscopic features that are not yet fully characterized. The findings underscore the necessity for further research involving larger patient populations to better delineate the spectroscopic patterns associated with these conditions.