Relationship of Bern Score, Spinal Elastance, and Opening Pressure in Patients With Spontaneous Intracranial Hypotension.

BACKGROUND AND OBJECTIVES: Existing tools to diagnose spontaneous intracranial hypotension (SIH), namely spinal opening pressure (OP) and brain MRI, have limited sensitivity. We investigated whether evaluation of brain MRI using the Bern score, combined with calculated craniospinal elastance, would aid in diagnosing SIH and provide insight into its pathophysiology. METHODS: A retrospective chart review was performed of patients who underwent brain MRI and pressure-augmented dynamic CT myelography (dCTM) for suspicion of SIH. Two blinded neuroradiologists assigned Bern scores for each brain MRI. OP and incremental pressure changes after intrathecal saline infusion were recorded to calculate craniospinal elastance. The relationship between Bern score, OP, and elastance and whether a leak was found were analyzed. RESULTS: Seventy-two consecutive dCTMs were performed in 53 patients. Twelve CSF-venous fistulae, 2 ruptured meningeal diverticula, 2 dural defects, and 1 dural bleb were found (17/53, 32%). Among patients with imaging-proven CSF leak/fistula, OP was normal in all but 1 patient and was not significantly different in those with a leak compared with those without (15.1 vs 13.6 cm H2O, p = 0.24, A = 0.40). The average Bern score in individuals with a leak was significantly higher than that in those without (5.35 vs 1.85, p < 0.001, A = 0.85), even when excluding pachymeningeal enhancement from the score (3.77 vs 1.57, p = 0.001, A = 0.78). The average elastance in those with a leak was higher than that in those without, but this difference was not statistically significant (2.05 vs 1.20 mL/cm H2O, p = 0.19, A = 0.40). Increased elastance was significantly associated with an increased Bern score (95% CI -0.55 to 0.12, p < 0.01) and was significantly associated with venous distention, pachymeningeal enhancement, prepontine narrowing, and subdural collections, but not a narrowed mamillopontine or suprasellar distance. DISCUSSION: OP is not an effective predictor for diagnosing CSF leak and if used in isolation would result in misdiagnosis of 94% of patients in our cohort. The Bern score was associated with a higher diagnostic yield of dCTM. Elastance was significantly associated with certain components of the Bern score.

Structured Reporting for Findings of Spontaneous Intracranial Hypotension on Brain MRI.

A voluntary structured reporting template (based on the Bern score) for brain MRI examinations performed for suspected spontaneous intracranial hypotension (SIH) was associated with an increase in reporting of intracranial MRI findings of SIH and a reduction in discordant assessments with respect to a reference reader.

A framework for automated anomaly detection in high frequency water-quality data from in situ sensors.

Monitoring the water quality of rivers is increasingly conducted using automated in situ sensors, enabling timelier identification of unexpected values or trends. However, the data are confounded by anomalies caused by technical issues, for which the volume and velocity of data preclude manual detection. We present a framework for automated anomaly detection in high-frequency water-quality data from in situ sensors, using turbidity, conductivity and river level data collected from rivers flowing into the Great Barrier Reef. After identifying end-user needs and defining anomalies, we ranked anomaly importance and selected suitable detection methods. High priority anomalies included sudden isolated spikes and level shifts, most of which were classified correctly by regression-based methods such as autoregressive integrated moving average models. However, incorporation of multiple water-quality variables as covariates reduced performance due to complex relationships among variables. Classifications of drift and periods of anomalously low or high variability were more often correct when we applied mitigation, which replaces anomalous measurements with forecasts for further forecasting, but this inflated false positive rates. Feature-based methods also performed well on high priority anomalies and were similarly less proficient at detecting lower priority anomalies, resulting in high false negative rates. Unlike regression-based methods, however, all feature-based methods produced low false positive rates and have the benefit of not requiring training or optimization. Rule-based methods successfully detected a subset of lower priority anomalies, specifically impossible values and missing observations. We therefore suggest that a combination of methods will provide optimal performance in terms of correct anomaly detection, whilst minimizing false detection rates. Furthermore, our framework emphasizes the importance of communication between end-users and anomaly detection developers for optimal outcomes with respect to both detection performance and end-user application. To this end, our framework has high transferability to other types of high frequency time-series data and anomaly detection applications.