Headache

Purpose To assess the performance of a local open-source large language model (LLM) in various information extraction tasks from real-life emergency brain MRI reports. Materials and Methods All consecutive emergency brain MRI reports written in 2022 from a French quaternary center were retrospectively reviewed. Two radiologists identified MRI scans that were performed in the emergency department for headaches. Four radiologists scored the reports' conclusions as either normal or abnormal. Abnormalities were labeled as either headache-causing or incidental. Vicuna (LMSYS Org), an open-source LLM, performed the same tasks. Vicuna's performance metrics were evaluated using the radiologists' consensus as the reference standard. Results Among the 2398 reports during the study period, radiologists identified 595 that included headaches in the indication (median age of patients, 35 years [IQR, 26-51 years]; 68% [403 of 595] women). A positive finding was reported in 227 of 595 (38%) cases, 136 of which could explain the headache. The LLM had a sensitivity of 98.0% (95% CI: 96.5, 99.0) and specificity of 99.3% (95% CI: 98.8, 99.7) for detecting the presence of headache in the clinical context, a sensitivity of 99.4% (95% CI: 98.3, 99.9) and specificity of 98.6% (95% CI: 92.2, 100.0) for the use of contrast medium injection, a sensitivity of 96.0% (95% CI: 92.5, 98.2) and specificity of 98.9% (95% CI: 97.2, 99.7) for study categorization as either normal or abnormal, and a sensitivity of 88.2% (95% CI: 81.6, 93.1) and specificity of 73% (95% CI: 62, 81) for causal inference between MRI findings and headache. Conclusion An open-source LLM was able to extract information from free-text radiology reports with excellent accuracy without requiring further training. Keywords: Large Language Model (LLM), Generative Pretrained Transformers (GPT), Open Source, Information Extraction, Report, Brain, MRI Supplemental material is available for this article. Published under a CC BY 4.0 license. See also the commentary by Akinci D'Antonoli and Bluethgen in this issue.

Several lines of evidence support a role of the immune system in headache pathogenesis, with particular regard to migraine. Firstly, alterations in cytokine profile and in lymphocyte subsets have been reported in headache patients. Secondly, several genetic and environmental pathogenic factors seem to be frequently shared by headache and immunological/autoimmune diseases. Accordingly, immunological alterations in primary headaches, in particular in migraine, have been suggested to predispose some patients to the development of immunological and autoimmune diseases. On the other hand, pathogenic mechanisms underlying autoimmune disorders, in some cases, seem to favour the onset of headache. Therefore, an association between headache and immunological/autoimmune disorders has been thoroughly investigated in the last years. The knowledge of this possible association may have relevant implications in the clinical practice when deciding diagnostic and therapeutic approaches. The present review summarizes findings to date regarding the plausible relationship between headache and immunological/autoimmune disorders, starting from a description of immunological alteration of primary headaches, and moving onward to the evidence supporting a potential link between headache and each specific autoimmune/immunological disease.

To estimate the frequency of headache in patients with confirmed COVID-19 and characterize the phenotype of headache attributed to COVID-19, comparing patients depending on the need of hospitalization and sex, an observational study was done. We systematically screened all eligible patients from a reference population of 261,431 between March 8 (first case) and April 11, 2020. A physician administered a survey assessing demographic and clinical data and the phenotype of the headache. During the study period, 2194 patients out of the population at risk were diagnosed with COVID-19. Headache was described by 514/2194 patients (23.4%, 95% CI 21.7-25.3%), including 383/1614 (23.7%) outpatients and 131/580 (22.6%) inpatients. The headache phenotype was studied in detail in 458 patients (mean age, 51 years; 72% female; prior history of headache, 49%). Headache was the most frequent first symptom of COVID-19. Median headache onset was within 24 h, median duration was 7 days and persisted after 1 month in 13% of patients. Pain was bilateral (80%), predominantly frontal (71%), with pressing quality (75%), of severe intensity. Systemic symptoms were present in 98% of patients. Headache frequency and phenotype was similar in patients with and without need for hospitalization and when comparing male and female patients, being more intense in females.Trial registration: This study was supported by the Institute of Health Carlos III (ISCIII), code 07.04.467804.74011 and Regional Health Administration, Gerencia Regional de Salud, Castilla y Leon (GRS: 2289/A/2020).

Calcitonin gene-related peptide (CGRP) is a key mediator in migraine pathophysiology. Idiopathic intracranial hypertension (IIH) headache phenotype is now understood to be typically migraine-like, but it is unclear whether CGRP directly provokes IIH headaches or alters intracranial pressure (ICP) dynamics. We conducted a randomized, double-blind, placebo-controlled, two-way crossover trial to address this. Twenty women with IIH and no prior migraine were randomly assigned to receive a 20-min continuous intravenous infusion of CGRP (1.5 μg/min) or placebo (isotonic saline). The primary outcome was the difference in the proportion of participants who developed a provoked headache attack between CGRP and placebo during the 12 h observation after infusion. Secondary outcomes included the area under the curve (AUC) for headache intensity from -10 min to 12 h, the timing and duration of headache features, and baseline-adjusted changes for vital signs, cerebrovascular haemodynamics and ICP. Seventeen participants with mean age (standard deviation) 26.7 (6.4) years completed both visits. Twelve (71%) participants developed a typical IIH headache attack with migraine-like features after CGRP compared with three (18%) after placebo (risk difference 53%; 95% confidence interval, 26-79; P = 0.004). The AUC-10min-12h for headache intensity was higher after CGRP than after placebo (P = 0.016). The mean ICP remained unchanged, whereas ICP amplitude increased significantly after CGRP (P = 0.005). Vital signs and cerebrovascular haemodynamics AUC-10min-90min were significantly altered after CGRP [increased: heart rate (P < 0.001), tissue oxygenation index (P = 0.041), and oxygenated haemoglobin (P < 0.001); decreased: mean arterial pressure (P = 0.010) and middle cerebral artery blood velocity (P = 0.006)]. CGRP reliably provoked typical IIH headache attacks (which have migraine-like features) and increased ICP pulse amplitude, as a measure of intracranial compliance, without altering mean pressure. These findings provide mechanistic support for CGRP involvement in headache attributed to IIH and justify prospective evaluation of CGRP pathway blockade in this population.