University of Twente Student Theses
Ultrasonographic optic nerve sheath diameter measurements in comatose patients after cardiac arrest additional predictive value and semi-automatic measurements
Visser, I.M. (2023) Ultrasonographic optic nerve sheath diameter measurements in comatose patients after cardiac arrest additional predictive value and semi-automatic measurements.
|
PDF
2MB |
| Abstract: | Background: Neurological outcome prediction after cardiac arrest is possible in only 28-47% of patients. Ultrasonographic measurements of the optic nerve sheath diameter (ONSD) may contribute to neurological outcome prediction. Our first objective was to evaluate the predictive value of ultrasonographic ONSD measurements, in addition to continuous electroencephalography (EEG) measurements, somatosensory evoked potentials (SSEP), and pupillary light reflexes (PLR) for neurological outcome in comatose patients after cardiac arrest. Our second objective was to develop and test a method to semi-automatically measure the ONSD from transorbital sonograms of comatose patients after cardiac arrest. Methods: We performed a prospective observational cohort study in adult comatose patients after cardiac arrest. ONSD was measured on days 1 to 3 using ultrasound. Continuous EEG, SSEP, and PLR were collected as part of standard care. Neurological outcome was classified using the Cerebral Performance Category (CPC) at 6 months (CPC 1-2 = good, CPC 3-5 = poor). For estimation of the additional predictive value of ONSD measurements, logistic regression models predicting neurological outcome were created based on EEG and SSEP, with and without ONSD. The additional predictive value of ONSD measurements was assessed by the increase in sensitivities for a poor (at 100% specificity) and good (at 90% specificity) neurological outcome. Semi-automatic ONSD measurements were performed using an active contour model. Agreement between manual and automatic ONSD measurements was assessed by visual inspection of segmentations, correlation, and Bland-Altman plots. Manual measurements were considered the gold standard. Results: We included 95 patients, of whom 41 (43.2%) died due to post-anoxic encephalopathy. ONSD measured on day 1 was larger in patients with a poor neurological outcome (6.40 [6.15 – 6.88] mm) than in those with a good neurological outcome (6.25 [5.68 – 6.63] mm) (p = 0.023). When adding ONSD measurements to predictions, sensitivity for a poor neurological outcome increased from 25% (95% confidence interval (CI): 0% – 50%) to 45% (95% CI: 25% – 65%) at 100% specificity. Sensitivity for a good neurological outcome raised from 8% (95% CI: 0 % – 23%) to 18% (95% CI: 5% – 36%) at 100% specificity. The PLR was not included as a predictor because of the low incidence of an absent PLR after 72 hours. The active contour model had a mean square error of 4.15 mm2 and feasibility of 86.1%. A good estimation of the ONSD with an absolute error of ≤ 0.5 mm between manual and automatic ONSD measurements was obtained in 16.7% of sonograms. Bland-Altman plots showed a bias of -0.267 mm and 95% limits agreements between -4.230 and 3.696 mm. Manual and automatic ONSD measurements did not correlate. Conclusion: Ultrasonographic ONSD measurements on days 1 to 3 after cardiac arrest hold the potential to add predictive value in neurological outcome prediction in addition to EEG and SSEP recordings. Our semi-automatic method based on active contours is not suited for automatic ONSD measurements. |
| Item Type: | Essay (Master) |
| Clients: | Rijnstate Hospital, Arnhem, Netherlands |
| Faculty: | TNW: Science and Technology |
| Subject: | 44 medicine |
| Programme: | Technical Medicine MSc (60033) |
| Link to this item: | https://purl.utwente.nl/essays/94193 |
| Export this item as: | BibTeX EndNote HTML Citation Reference Manager |
Show download statistics for this publication
Show download statistics for this publicationRepository Staff Only: item control page