|
1
|
Zhang H, Zou P, Luo P, Jiang X. Machine Learning for the Early Prediction of Delayed Cerebral Ischemia in Patients With Subarachnoid Hemorrhage: Systematic Review and Meta-Analysis. J Med Internet Res 2025; 27:e54121. [PMID: 39832368 PMCID: PMC11791451 DOI: 10.2196/54121] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [Abstract] [Key Words] [MESH Headings] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Received: 10/31/2023] [Revised: 10/14/2024] [Accepted: 11/26/2024] [Indexed: 01/22/2025] Open
Abstract
BACKGROUND Delayed cerebral ischemia (DCI) is a primary contributor to death after subarachnoid hemorrhage (SAH), with significant incidence. Therefore, early determination of the risk of DCI is an urgent need. Machine learning (ML) has received much attention in clinical practice. Recently, some studies have attempted to apply ML models for early noninvasive prediction of DCI. However, systematic evidence for its predictive accuracy is still lacking. OBJECTIVE The aim of this study was to synthesize the prediction accuracy of ML models for DCI to provide evidence for the development or updating of intelligent detection tools. METHODS PubMed, Cochrane, Embase, and Web of Science databases were systematically searched up to May 18, 2023. The risk of bias in the included studies was assessed using PROBAST (Prediction Model Risk of Bias Assessment Tool). During the analysis, we discussed the performance of different models in the training and validation sets. RESULTS We finally included 48 studies containing 16,294 patients with SAH and 71 ML models with logistic regression as the main model type. In the training set, the pooled concordance index (C index), sensitivity, and specificity of all the models were 0.786 (95% CI 0.737-0.835), 0.77 (95% CI 0.69-0.84), and 0.83 (95% CI 0.75-0.89), respectively, while those of the logistic regression models were 0.770 (95% CI 0.724-0.817), 0.75 (95% CI 0.67-0.82), and 0.71 (95% CI 0.63-0.78), respectively. In the validation set, the pooled C index, sensitivity, and specificity of all the models were 0.767 (95% CI 0.741-0.793), 0.66 (95% CI 0.53-0.77), and 0.78 (95% CI 0.71-0.84), respectively, while those of the logistic regression models were 0.757 (95% CI 0.715-0.800), 0.59 (95% CI 0.57-0.80), and 0.80 (95% CI 0.71-0.87), respectively. CONCLUSIONS ML models appear to have relatively desirable power for early noninvasive prediction of DCI after SAH. However, enhancing the prediction sensitivity of these models is challenging. Therefore, efficient, noninvasive, or minimally invasive low-cost predictors should be further explored in future studies to improve the prediction accuracy of ML models. TRIAL REGISTRATION PROSPERO (CRD42023438399); https://tinyurl.com/yfuuudde.
Collapse
Affiliation(s)
- Haofuzi Zhang
- Department of Neurosurgery, Xijing Hospital, Fourth Military Medical University, Xi'an, China
| | - Peng Zou
- Department of Neurosurgery, Xijing Hospital, Fourth Military Medical University, Xi'an, China
| | - Peng Luo
- Department of Neurosurgery, Xijing Hospital, Fourth Military Medical University, Xi'an, China
| | - Xiaofan Jiang
- Department of Neurosurgery, Xijing Hospital, Fourth Military Medical University, Xi'an, China
| |
Collapse
|
|
2
|
Al-Salihi MM, Gillani SA, Saha R, Abd Elazim A, Al-Jebur MS, Al-Salihi Y, Ayyad A, Nattanmai P, Siddiq F, Gomez CR, Qureshi AI. Clinical Characteristics as Predictors of Early and Delayed Cerebral Infarction in Aneurysmal Subarachnoid Hemorrhage Patients: A Meta-Analysis of 4527 Cases. World Neurosurg 2024; 189:373-380.e3. [PMID: 38906475 DOI: 10.1016/j.wneu.2024.06.060] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [Abstract] [Key Words] [MESH Headings] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Received: 06/08/2024] [Accepted: 06/12/2024] [Indexed: 06/23/2024]
Abstract
BACKGROUND Predictors of delayed cerebral infarction (DCI) and early cerebral infraction (ECI) among aneurysmal subarachnoid hemorrhage (aSAH) patients remain unclear. We aimed to systematically review and synthesize the literature on predictors of ECI and DCI among aSAH patients. METHODS We systematically searched PubMed, EMBASE, Cochrane Library, and Scopus databases comprehensively from inception through January 2024 for observational cohort studies examining predictors of DCI or ECI following aneurysmal SAH. Studies were screened, reviewed, and meta-analyzed, adhering to Preferred Reporting Items for Systematic Reviews and Meta-Analyses and Cochrane guidelines. The data were pooled as Odds ratios (OR) with 95% confidence intervals using Review Manager 5.4 software. Methodologic quality was assessed with the Newcastle-Ottawa Scale. RESULTS Our meta-analysis included 12 moderate to high-quality cohort studies comprising 4527 patients. Regarding DCI predictors, Higher severity scores (OR = 1.49, 95% confidence interval [1.12, 1.97], P = 0.005) and high Fisher scores (OR = 2.23, 95% confidence interval [1.28, 3.89], P = 0.005) on presentation were significantly associated with an increased risk of DCI. Also, the female sex and the presence of vasospasm were significantly associated with an increased risk of DCI (OR = 3.04, 95% confidence interval [1.35, 6.88], P = 0.007). In contrast, preexisting hypertension (P = 0.94), aneurysm treatment (P = 0.14), and location (P = 0.16) did not reliably predict DCI risk. Regarding ECI, the pooled analysis demonstrated no significant associations between sex (P = 0.51), pre-existing hypertension (P = 0.63), severity (P = 0.51), or anterior aneurysm location versus posterior (P = 0.86) and the occurrence of ECI. CONCLUSION Female sex, admission disease severity, presence of vasospasm and Fisher grading can predict DCI risk post-aSAH. Significant knowledge gaps exist for ECI predictors. Further large standardized cohorts are warranted to guide prognosis and interventions.
Collapse
Affiliation(s)
- Mohammed Maan Al-Salihi
- Department of Neurological Surgery, School of Medicine and Public Health, University of Wisconsin, Madison, Wisconsin, USA.
| | - Syed A Gillani
- Zeenat Qureshi Stroke Institute, University of Missouri, Columbia, Missouri, USA; Department of Neurology, University of Missouri, Columbia, Missouri, USA
| | - Ram Saha
- Department of Neurology, School of Medicine, Virginia Commonwealth University, Richmond, Virginia, USA
| | - Ahmed Abd Elazim
- Department of Neurology, University of South Dakota, Sioux Falls, South Dakota, USA
| | | | | | - Ali Ayyad
- Department of Neurosurgery, Hamad General Hospital, Doha, Qatar
| | | | - Farhan Siddiq
- Department of Neurosurgery, University of Missouri, Columbia, Missouri, USA
| | - Camilo R Gomez
- Department of Neurology, University of Missouri, Columbia, Missouri, USA
| | - Adnan I Qureshi
- Zeenat Qureshi Stroke Institute, University of Missouri, Columbia, Missouri, USA; Department of Neurology, University of Missouri, Columbia, Missouri, USA
| |
Collapse
|
|
3
|
Deininger MM, Weiss M, Wied S, Schlycht A, Haehn N, Marx G, Hoellig A, Schubert GA, Breuer T. Value of Glycemic Indices for Delayed Cerebral Ischemia after Aneurysmal Subarachnoid Hemorrhage: A Retrospective Single-Center Study. Brain Sci 2024; 14:849. [PMID: 39335345 PMCID: PMC11430037 DOI: 10.3390/brainsci14090849] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [Abstract] [Key Words] [Grants] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Received: 07/29/2024] [Revised: 08/19/2024] [Accepted: 08/21/2024] [Indexed: 09/30/2024] Open
Abstract
Delayed cerebral ischemia (DCI) is a severe complication following aneurysmal subarachnoid hemorrhage (aSAH), linked to poor functional outcomes and prolonged intensive care unit (ICU) stays. Timely DCI diagnosis is crucial but remains challenging. Dysregulated blood glucose, commonly observed after aSAH, may impair the constant glucose supply that is vital for brain function, potentially contributing to DCI. This study aimed to assess whether glucose indices could help identify at-risk patients and improve DCI detection. This retrospective, single-center observational study examined 151 aSAH patients between 2016 and 2019. Additionally, 70 of these (46.4%) developed DCI and 81 did not (no-DCI). To determine the value of glycemic indices for DCI, they were analyzed separately in patients in the period before (pre-DCI) and after DCI (post-DCI). The time-weighted average glucose (TWAG, p = 0.024), mean blood glucose (p = 0.033), and novel time-unified dysglycemic rate (TUDR140, calculated as the ratio of dysglycemic to total periods within a glucose target range of 70-140 mg/dL, p = 0.042), showed significantly higher values in the pre-DCI period of the DCI group than in the no-DCI group. In the time-series analysis, significant increases in TWAG and TUDR140 were observed at the DCI onset. In conclusion, DCI patients showed elevated blood glucose levels before and a further increase at the DCI onset. Prospective studies are needed to confirm these findings, as this retrospective, single-center study cannot completely exclude confounders and limitations. In the future blood glucose indices might become valuable parameters in multiparametric models to identify patients at risk and detect DCI onset earlier.
Collapse
Affiliation(s)
- Matthias Manfred Deininger
- Department of Intensive and Intermediate Care, Medical Faculty, RWTH Aachen University, 52074 Aachen, Germany
| | - Miriam Weiss
- Department of Neurosurgery, Medical Faculty, RWTH Aachen University, 52074 Aachen, Germany
- Department of Neurosurgery, Cantonal Hospital Aarau, 5001 Aarau, Switzerland
| | - Stephanie Wied
- Institute of Medical Statistics, RWTH Aachen University, 52074 Aachen, Germany
| | - Alexandra Schlycht
- Department of Intensive and Intermediate Care, Medical Faculty, RWTH Aachen University, 52074 Aachen, Germany
| | - Nico Haehn
- Department of Intensive and Intermediate Care, Medical Faculty, RWTH Aachen University, 52074 Aachen, Germany
| | - Gernot Marx
- Department of Intensive and Intermediate Care, Medical Faculty, RWTH Aachen University, 52074 Aachen, Germany
| | - Anke Hoellig
- Department of Neurosurgery, Medical Faculty, RWTH Aachen University, 52074 Aachen, Germany
| | - Gerrit Alexander Schubert
- Department of Neurosurgery, Medical Faculty, RWTH Aachen University, 52074 Aachen, Germany
- Department of Neurosurgery, Cantonal Hospital Aarau, 5001 Aarau, Switzerland
| | - Thomas Breuer
- Department of Intensive and Intermediate Care, Medical Faculty, RWTH Aachen University, 52074 Aachen, Germany
| |
Collapse
|
|
4
|
Galijasevic M, Steiger R, Treichl SA, Ho WM, Mangesius S, Ladenhauf V, Deeg J, Gruber L, Ouaret M, Regodic M, Lenhart L, Pfausler B, Grams AE, Petr O, Thomé C, Gizewski ER. Could Phosphorous MR Spectroscopy Help Predict the Severity of Vasospasm? A Pilot Study. Diagnostics (Basel) 2024; 14:841. [PMID: 38667486 PMCID: PMC11049300 DOI: 10.3390/diagnostics14080841] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [Abstract] [Key Words] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Received: 02/22/2024] [Revised: 04/02/2024] [Accepted: 04/07/2024] [Indexed: 04/28/2024] Open
Abstract
One of the main causes of the dismal prognosis in patients who survive the initial bleeding after aneurysmal subarachnoidal hemorrhage is the delayed cerebral ischaemia caused by vasospasm. Studies suggest that cerebral magnesium and pH may potentially play a role in the pathophysiology of this adverse event. Using phosphorous magnetic resonance spectrocopy (31P-MRS), we calculated the cerebral magnesium (Mg) and pH levels in 13 patients who suffered from aSAH. The values between the group that developed clinically significant vasospasm (n = 7) and the group that did not (n = 6) were compared. The results of this study show significantly lower cerebral Mg levels (p = 0.019) and higher pH levels (p < 0.001) in the cumulative group (all brain voxels together) in patients who developed clinically significant vasospasm. Further clinical studies on a larger group of carefully selected patients are needed in order to predict clinically significant vasospasm.
Collapse
Affiliation(s)
- Malik Galijasevic
- Department of Radiology, Medical University of Innsbruck, 6020 Innsbruck, Austria; (M.G.); (S.M.); (V.L.); (J.D.); (L.G.); (M.O.); (M.R.); (L.L.); (A.E.G.); (E.R.G.)
- Neuroimaging Research Core Facility, Medical University of Innsbruck, 6020 Innsbruck, Austria
| | - Ruth Steiger
- Department of Radiology, Medical University of Innsbruck, 6020 Innsbruck, Austria; (M.G.); (S.M.); (V.L.); (J.D.); (L.G.); (M.O.); (M.R.); (L.L.); (A.E.G.); (E.R.G.)
- Neuroimaging Research Core Facility, Medical University of Innsbruck, 6020 Innsbruck, Austria
| | - Stephanie Alice Treichl
- Department of Neurosurgery, Medical University of Innsbruck, 6020 Innsbruck, Austria; (S.A.T.); (W.M.H.); (O.P.); (C.T.)
| | - Wing Man Ho
- Department of Neurosurgery, Medical University of Innsbruck, 6020 Innsbruck, Austria; (S.A.T.); (W.M.H.); (O.P.); (C.T.)
| | - Stephanie Mangesius
- Department of Radiology, Medical University of Innsbruck, 6020 Innsbruck, Austria; (M.G.); (S.M.); (V.L.); (J.D.); (L.G.); (M.O.); (M.R.); (L.L.); (A.E.G.); (E.R.G.)
- Neuroimaging Research Core Facility, Medical University of Innsbruck, 6020 Innsbruck, Austria
| | - Valentin Ladenhauf
- Department of Radiology, Medical University of Innsbruck, 6020 Innsbruck, Austria; (M.G.); (S.M.); (V.L.); (J.D.); (L.G.); (M.O.); (M.R.); (L.L.); (A.E.G.); (E.R.G.)
- Neuroimaging Research Core Facility, Medical University of Innsbruck, 6020 Innsbruck, Austria
| | - Johannes Deeg
- Department of Radiology, Medical University of Innsbruck, 6020 Innsbruck, Austria; (M.G.); (S.M.); (V.L.); (J.D.); (L.G.); (M.O.); (M.R.); (L.L.); (A.E.G.); (E.R.G.)
- Neuroimaging Research Core Facility, Medical University of Innsbruck, 6020 Innsbruck, Austria
| | - Leonhard Gruber
- Department of Radiology, Medical University of Innsbruck, 6020 Innsbruck, Austria; (M.G.); (S.M.); (V.L.); (J.D.); (L.G.); (M.O.); (M.R.); (L.L.); (A.E.G.); (E.R.G.)
- Neuroimaging Research Core Facility, Medical University of Innsbruck, 6020 Innsbruck, Austria
| | - Miar Ouaret
- Department of Radiology, Medical University of Innsbruck, 6020 Innsbruck, Austria; (M.G.); (S.M.); (V.L.); (J.D.); (L.G.); (M.O.); (M.R.); (L.L.); (A.E.G.); (E.R.G.)
- Neuroimaging Research Core Facility, Medical University of Innsbruck, 6020 Innsbruck, Austria
| | - Milovan Regodic
- Department of Radiology, Medical University of Innsbruck, 6020 Innsbruck, Austria; (M.G.); (S.M.); (V.L.); (J.D.); (L.G.); (M.O.); (M.R.); (L.L.); (A.E.G.); (E.R.G.)
- Neuroimaging Research Core Facility, Medical University of Innsbruck, 6020 Innsbruck, Austria
| | - Lukas Lenhart
- Department of Radiology, Medical University of Innsbruck, 6020 Innsbruck, Austria; (M.G.); (S.M.); (V.L.); (J.D.); (L.G.); (M.O.); (M.R.); (L.L.); (A.E.G.); (E.R.G.)
- Neuroimaging Research Core Facility, Medical University of Innsbruck, 6020 Innsbruck, Austria
| | - Bettina Pfausler
- Department of Neurology, Medical University of Innsbruck, 6020 Innsbruck, Austria;
| | - Astrid Ellen Grams
- Department of Radiology, Medical University of Innsbruck, 6020 Innsbruck, Austria; (M.G.); (S.M.); (V.L.); (J.D.); (L.G.); (M.O.); (M.R.); (L.L.); (A.E.G.); (E.R.G.)
- Neuroimaging Research Core Facility, Medical University of Innsbruck, 6020 Innsbruck, Austria
| | - Ondra Petr
- Department of Neurosurgery, Medical University of Innsbruck, 6020 Innsbruck, Austria; (S.A.T.); (W.M.H.); (O.P.); (C.T.)
| | - Claudius Thomé
- Department of Neurosurgery, Medical University of Innsbruck, 6020 Innsbruck, Austria; (S.A.T.); (W.M.H.); (O.P.); (C.T.)
| | - Elke Ruth Gizewski
- Department of Radiology, Medical University of Innsbruck, 6020 Innsbruck, Austria; (M.G.); (S.M.); (V.L.); (J.D.); (L.G.); (M.O.); (M.R.); (L.L.); (A.E.G.); (E.R.G.)
- Neuroimaging Research Core Facility, Medical University of Innsbruck, 6020 Innsbruck, Austria
| |
Collapse
|