Automated Hemorrhage Volume Quantification in Aneurysmal Subarachnoid Hemorrhage

Automated Hemorrhage Volume Quantification in Aneurysmal Subarachnoid Hemorrhage | Research Square window.SnipcartSettings = { analytics: { enabled: false } }; (function() { var accessVector = localStorage.getItem('access_vector') || ''; window.dataLayer = window.dataLayer || []; if (accessVector) { window.dataLayer.push({ user: { profile: { profileInfo: { snid: accessVector } } } }); } })(); (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0],j=d.createElement(s),dl=l!='dataLayer'?'&l='+l:'';j.async=true;j.src='https://www.googletagmanager.com/gtm.js?id='+i+dl;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-K279D39R'); Browse Preprints In Review Journals COVID-19 Preprints AJE Video Bytes Research Tools Research Promotion AJE Professional Editing AJE Rubriq About Preprint Platform In Review Editorial Policies Our Team Advisory Board Help Center Sign In Submit a Preprint Cite Share Download PDF Research Article Automated Hemorrhage Volume Quantification in Aneurysmal Subarachnoid Hemorrhage Sebastian Sanchez, Jacob M Miller, Matthew T Jones, Rishi R Patel, and 5 more This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-4308305/v1 This work is licensed under a CC BY 4.0 License Status: Published Journal Publication published 25 Sep, 2024 Read the published version in Neurocritical Care → Version 1 posted 5 You are reading this latest preprint version Abstract Background The volume of hemorrhage is a critical factor in predicting outcomes following aneurysmal subarachnoid hemorrhage (aSAH). Although grading scales such as the Fisher score are extensively used, their subjective nature can lead to inaccuracies in quantifying the total volume of blood. We analyzed a large cohort of patients with aSAH with an automated software for the precise quantification of hemorrhage volume. The primary aim is to identify clear thresholds that correlate with the likelihood of complications post-aSAH, thereby enhancing the predictive accuracy and improving patient management strategies. Methods An automated algorithm was developed to analyze non-contrast computed tomography scans of aSAH patients. The algorithm categorized tissues into blood, gray matter, white matter, and cerebrospinal fluid, isolating the blood for volume quantification. Receiver operating curve analysis was done to establish thresholds for vasospasm, acute hydrocephalus, shunt-dependent hydrocephalus (SDH), and death within 7 days. Additionally, we determined if there is any relationship between the aneurysm size and the amount of hemorrhage. Results A total of 500 aSAH patients and their respective aneurysms were analyzed. Hemorrhage volume was significantly higher in patients with vasospasm (21.7 [10.9, 41.4] vs 10.7 [4.2, 26.9], p < 0.001), acute hydrocephalus (22.7 [9.2, 41.8] vs 5.1 [2.1, 13.5], p < 0.001), SDH (23.8 [11.3, 40.7] vs 11.7 [4.1, 28.2], p < 0.001), and those who died before 7 days (52.8 [34.6, 90.6] mL vs 14.8 [5.0, 32.4] mL, p < 0.001) compared to their counterparts. Notably, specific hemorrhage thresholds were identified for each complication: 15.16 mL for vasospasm (65% sensitivity and 60% specificity), 9.95 mL for acute hydrocephalus (74% sensitivity and 69% specificity), 16.76 mL for SDH (63% sensitivity and 60% specificity), and 33.84 mL for death within 7 days (79% sensitivity and 77% specificity). Conclusion Automated blood volume quantification tools could aid in stratifying complication risk after aSAH. Established thresholds for hemorrhage volume related to complications could be used in clinical practice to aid in management decisions. Figures Figure 1 Figure 2 Figure 3 Introduction Hemorrhage volume plays a crucial role in evaluating the likelihood of complications following aneurysmal subarachnoid hemorrhage (aSAH). 1 The Fisher score 2 and modified Fisher score 3 are radiological grading systems designed to predict the likelihood of vasospasm after aSAH. These scores are based on the thickness and location of hemorrhage and have been validated in stratifying the risk of vasospasm following aSAH. 2 – 4 However, adjudication of the scores is operator dependent and has low interrater reliability. 5 Moreover, they provide scarce information about the risk for acute hydrocephalus, shunt dependent hydrocephalus (SDH) and death in the acute phase. Additionally, clinical grading scales such as the Hunt and Hess (H&H) 6 and the World federation of neurosurgeons (WFNS) 7 have been used to assess surgical and complication risk after aSAH. Nonetheless, the relationship between these scales and the volume of hemorrhage remains to be evaluated. Recently, manual, 8 semiautomated, 9 and automated methods, 10 – 13 that have been used to quantify hemorrhage volume following aSAH. However, there is no consensus on the hemorrhage volume thresholds that can increase the likelihood of complications following aSAH. In this study an automated method was used to quantify hemorrhage volume in a large cohort of patients with aSAH. We hypothesize that an objective and more accurate method to quantify hemorrhage volume could achieve good correlations with complications following aSAH. We specifically aimed to identify thresholds for hemorrhage volume associated with arterial vasospasm, acute hydrocephalus, SDH, and death within 7 days following aSAH. Additionally, we investigated the potential correlation between hemorrhage volume and clinical scales such as WFNS and H&H. Finally, we also studied if there is any potential association between the size of the aneurysm and hemorrhage volume. Methods After institutional review board approval patients diagnosed with an aSAH at our institution from 2009 to 2022 were included in the study. The TriNetX search engine was utilized to identify patients with aSAH. Patients with a Fisher score of 1 where excluded from the study as this score corresponds to no quantifiable blood in the subarachnoid space. 2 Patient demographics, H&H WFNS scores and available imaging were retrieved for analysis. Aneurysmal size was defined as aneurysm maximal diameter, and was determined with the best vascular imaging. 14 Hemorrhage quantification algorithm An automated software was developed to identify and quantify aSAH using MATLAB (MATLAB and Image Processing Toolbox, release 2020b; MathWorks, Inc.). DICOM files containing non-contrast computed tomography (CT) scans were loaded into MATLAB. Hounsfield unit (HU) values below zero were identified as background and air, while HU values above 160 were considered bone, and were excluded. 12 The lower anatomical limit for aSAH detection was the foramen magnum, as established in a previous study by Renan Sales et al. 12 A 3D segmentation of intracranial blood was created by using k-means clustering, and four different tissue groups based on HUs. These areas tentatively corresponded to blood, white matter, gray matter, and cerebrospinal fluid (CSF). The highest HU values were adjudicated as blood and isolated as a blood cluster. MATLAB functions were then used to isolate and exclude individual voxels corresponding to the blood cluster if they were in other tissues as they were likely artifact. Similarly, individual voxels adjudicated as blood were removed if they were adjacent to the skull to avoid potential contamination. The blood segmentation was overlayed on the original CT scan slices for quality control (Fig. 1 ). Finally, the aSAH volume was calculated by the automated algorithm. Detailed technical methods are described in the supplementary materials. Validation of the automated algorithm for aSAH detection Two independent raters manually measured hemorrhage volumes from ten randomly selected non-contrast CT scans using the region of interest (ROI) tool on PACS (Carestream Vue). The areas from the ROIs were summed to quantify the hemorrhage volume in each slice. Then, the hemorrhage volume in all slices were summed and multiplied by the slice thickness (0.5 mm) to obtain the total hemorrhage volume. The automated algorithm was then used to calculate the hemorrhage volume of the same ten CT scans. The Dice Similarity Coefficient was calculated to compare the performance of the automated method with the manual measurements. 15 Dice Coefficients were calculated against each of the raters’ scans, as follows: $$DSC=\frac{2*A\cap B}{A+B}$$ A is the number of voxels in the rater’s manual segmentation, B is the number of voxels in the automatic segmentation, and A ∩ B is the number of voxels the segmentations overlap. Dice scores compared the volumes as well as the spatial distributions of voxels being identified by each method. Clinical outcomes The following clinical outcomes were included in the analysis: symptomatic vasospasm, presence of acute hydrocephalus, SDH and death within seven days. Symptomatic vasospasm was defined as focal neurological impairment that occurred within 3–14 days in the setting of radiographic vasospasm documented on computed tomography angiography or digital subtraction angiography, and not attributed to other causes. 16 Symptomatic vasospasm was also adjudicated if patients required endovascular treatment for vasospasm after aSAH. Acute hydrocephalus was identified as an enlargement of the ventricles due to obstruction of the CSF drainage due to intracranial bleeding. 17 SDH was adjudicated if the patient required any permanent drainage device such as a ventriculoperitoneal shunt. Finally, death within seven days was adjudicated in the case a patient passed within 7 days of aneurysm rupture. Statistical Analysis IBM SPSS Statistics 27.0 was utilized for statistical analysis. The Shapiro-Wilk test was used to assess distribution among variables. Normally distributed variables are reported as mean and standard deviation (SD), while non-normally distributed variables are presented as median and interquartile range (IQR). A student t-test was used to evaluate differences among normally distributed variables. The Mann-Whitney U test was used to compare non-normally distributed variables. Pearson intraclass correlation coefficients and Spearman’s Rho were used to evaluate correlations in normal and non-normally distributed variables. A Bland-Altman analysis was conducted to compare algorithmic volumes to volumes averaged among raters. Univariate logistic regression analysis was performed to assess the relationship between hemorrhage volume and four separate clinical outcomes: vasospasm, acute hydrocephalus, SDH, and death within 7 days. Receiver operating characteristic (ROC) curves were created to analyze potential thresholds of hemorrhage volume that could discriminate for the presence of the previously described outcomes. These thresholds, along with quantified area under the curve (AUC), were extracted from the ROC curves. The significance level was set at 0.05. Results Validation of the aSAH algorithm The manual raters’ intraclass coefficient was 0.993 (p < 0.001), and the intraclass coefficient between the averaged manual volumes and the automated algorithm was 0.973 (p < 0.001). The average Dice Coefficient between the segmentations was 0.730. The mean difference between manual and automated measurements based on the Bland-Altman analysis was of -3.14 mL, and the mean absolute error was 4.33 mL (Supplementary Fig. 1). Aneurysm Characteristics and hemorrhage volume Five-hundred patients with 500 ruptured aneurysms were included in the analysis. The mean age was 57.9 ± 13.4, 65.8% were female (n = 329), and 88.8% (n = 444) were white (Table 1 ). The median aneurysm size was 5.4 mm [4.0, 7.9], and the median intracranial hemorrhage volume was 18.0 mL [5.7, 37.5] (Supplementary table 1 ). Thirty-one percent (n = 148) of saccular aneurysms had a diameter larger than 7 mm. Aneurysmal size and volume of hemorrhage had a weak linear correlation (Rho = 0.139) (Fig. 2 ). Similarly, aneurysm size did not correlate with the Fisher score (Rho = 0.18), H&H (Rho = 0.186) and WFNS grade (Rho = 0.164). Table 1 Patient demographics Demographics Age (mean ± SD) 57.9 ± 13.4 Female N (%) 329 (65.8) White N (%) 444 (88.8) Fisher grade N (%) Fisher 2 97 (19.4) Fisher 3 121 (24.2) Fisher 4 282 (56.4) Hunt and Hess N (%) Grade 1 86 (17.2) Grade 2 147 (29.4) Grade 3 119 (23.8) Grade 4 64 (12.8) Grade 5 84 (16.8) Word federation of Neurosurgical Societies N (%) Grade 1 200 (40) Grade 2 78 (15.6) Grade 3 14 (2.8) Grade 4 79 (15.8) Grade 5 129 (25.8) Complications N (%) Symptomatic Vasospasm 161 (36.4) Acute hydrocephalus 399 (79.8) Shunt dependent hydrocephalus 127 (28.7) Death 7 mm 148 (30.8) ≤ 7 mm 332 (69.2) Table 2 Hemorrhage volume quantification and aneurysm size measurements Clinical Presentation Volume of Hemorrhage mL Median [IQR] Aneurysm Diameter mm Median [IQR] Fisher Grade Fisher 2 4.5 [1.5, 10.5] 5.0 [4.0, 6.5] Fisher 3 13.4 [5.0, 26.3] 5.0 [3.4, 7.0] Fisher 4 29.5 [13.8, 51.8] 6.0 [4.4, 8.8] Rho 0.514 0.180 Hunt and Hess Grade 1 8.5 [3.1, 15.2] 5.4 [4.3, 7.0] Grade 2 7.0 [3.1, 17.7] 4.9 [3.5, 6.9] Grade 3 20.8 [10.6, 35.5] 5.5 [3.9, 7.5] Grade 4 40.5 [24.8, 64.0] 5.2 [4.0, 8.2] Grade 5 49.8 [32.3, 76.2] 7.4 [5.0, 11.4] Rho 0.604 0.186 World Federation of Neurosurgical Societies Grade 1 7.1 [2.9, 16.7] 5.2 [3.9, 7.0] Grade 2 14.9 [6.7, 30.2] 5.0 [3.5, 7.3] Grade 3 17.7 [10.2, 25.6] 5.1 [3.4, 7.8] Grade 4 25.3 [11.4, 39.2] 5.0 [3.9, 7.3] Grade 5 51.0 [28.6, 73.2] 6.3 [5.0, 10.5] Rho 0.613 0.164 Hemorrhage volume and Clinical Outcomes Hemorrhage volume had a moderate correlation with the Fisher score (Rho = 0.514), and a strong correlation with a higher H&H scale (Rho = 0.604) and higher WFNS grade (Rho = 0.613) (Supplementary Fig. 2). There were significantly higher hemorrhage volumes in patients with vasospasm (21.7 [10.9, 41.4] vs 10.7 [4.2, 26.9], p < 0.001), acute hydrocephalus (22.7 [9.2, 41.8] vs 5.1 [2.1, 13.5]; p < 0.001), SDH (23.8 [11.3, 40.7] vs 11.7 [4.1, 28.2], p < 0.001), and those who died within 7 days (52.8 [34.6, 90.6] mL vs 14.8 [5.0, 32.4] mL, p < 0.001) compared to their counterparts (Supplementary Table 4). For each mL increase in intracranial hemorrhage, there was a 2% increase in vasospasm (OR 1.02, 95% CI: 1.01–1.03; p < 0.001), a 5% increase in hydrocephalus (OR 1.05, 95% CI: 1.03–1.07; p < 0.001), a 2% increase in SDH (OR 1.02, 95% CI: 1.01–1.03; p < 0.001); and 4% higher death within the first week (OR 1.04, 95% CI: 1.03–1.05; p 15.16 mL was associated symptomatic vasospasm (65% sensitivity and 60% specificity, AUC = 0.656), volume > 9.95 mL with acute hydrocephalus (74% sensitivity and 69% specificity, AUC = 0.763), volume > 16.76 mL with SDH (63% sensitivity and 60% specificity, AUC = 0.666), and volume > 33.84 mL with mortality within seven days (79% sensitivity and 77% specificity, AUC = 0.839) (Fig. 3 ). Discussion The use of automated and objective methods for quantifying hemorrhage volumes has the potential for improving the prediction of complications following aSAH compared to subjective grading scales. The Fisher scale only achieved moderate correlation with the objective quantification of hemorrhage, highlighting the poor performance of this subjective scale. Specific volume thresholds are more likely correlated with the occurrence of vasospasm, acute hydrocephalus, SDH, and mortality within seven days. Furthermore, our findings suggest a weak correlation between the size of the aneurysm and the volume of hemorrhage. Hemorrhage volume post-aSAH has traditionally been assessed subjectively, leading to considerable interrater variability. Van der Gagt et al. observed significant variability in the assessment of hemorrhage volume on 159 aSAH CT scans (kappa 0.34–0.66). 18 Similarly, Van Norden et al studied 132 aSAH CT scans and found that the interobserver agreement of the Fisher scale was only mild to moderate (kappa 0.37–0.55). 19 New semiautomated 9 and automated, 10 – 13 methods have been used for hemorrhage volume quantification. A recent meta-analysis conducted by Matsoukas et al. encompassed 40 studies detailing artificial intelligence (AI) algorithms designed for intracranial hemorrhage detection. 20 On average, most AI-driven software tools had a high performance for the diagnosis of intracerebral hemorrhage, however, specific thresholds for aSAH complications prediction were not reported. Quantifying hemorrhage volume in aSAH poses challenges due to blood distribution across subarachnoid spaces, ventricles and parenchyma. Ziljstra et al.’s analysis of 333 patients using an automated detection method found intraventricular hemorrhage to be a predictor of delayed cerebral ischemia (DCI), while intraparenchymal hematoma showed a negative association. 10 In contrast, Platz et al reported that intraparenchymal hemorrhage was associated with DCI. 21 The precise factors leading to DCI remain unclear, but accurately measuring total intracerebral hemorrhage could better predict DCI and complications like symptomatic vasospasm, SDH, and poor functional outcomes following aSAH. 22 Moreover, in the era of AI, defining blood volume thresholds may refine prognostic evaluations and enhance patient outcomes. In AUC analysis, an aSAH volume of 15 mL exhibited moderate accuracy in predicting symptomatic vasospasm (65.2% sensitivity and 60.1% specificity, AUC 0.65). In a smaller cohort of 42 patients, Street et al. reported that a semi-automated method outperform the modified Fisher scale in predicting radiological vasospasm, achieving a higher accuracy (AUC 0.86 vs 0.70). 23 Variability in reported incidences of radiological vasospasm, symptomatic vasospasm, and DCI across studies highlights the lack of standardized definitions and outcomes in this research area. 24 In our analysis, an aSAH volume of 9.95 mL achieved a higher accuracy for the occurrence of hydrocephalus (sensitivity of 74.2% and specificity of 69%). This volume is notably smaller than the volumes associated with vasospasm, and death within 7 days. Only a small volume of hemorrhage may be necessary to obstruct or damage the pathways responsible for CSF circulation and absorption. Hemorrhage volume is also pivotal in assessing the likelihood of SDH. A meta-analysis conducted by Wilson et al. on the risk factors associated with SDH revealed a strong correlation between the presence of intraventricular blood (OR 3.93) and SDH. 25 We report a volume of 16.76 mL for predicting SDH post-aSAH (sensitivity of 64% and specificity of 60%). Previously, Garcia et al. studied a cohort of 168 patients using a manual method for hemorrhage assessment and reported a threshold of 10.3 mL for SDH. This threshold was less sensitive of (50%) but more specific (83%) than the threshold reported in our study. 8 This disparity may be explained by the different clinical criteria in determining the need for a permanent shunt. Mortality in the acute phase of aSAH can be the result of multiple factors, but a larger hemorrhage burden often plays a pivotal role. In our study, a volume of 33 mL of aSAH was very accurate in predicting mortality within 7 days (sensitivity of 79% and specificity of 76.9%). Similarly, Lagares et al. examined 206 patients and found that the proportion of individuals with poor outcomes, defined as Glasgow Outcome Scale score of 3 or lower, increased from 35% when the aSAH total hemorrhage volume was below 20 mL to 86% when it exceeded 20 mL. 22 Aneurysm size and hemorrhage volume It is unclear whether larger aneurysms lead to larger hemorrhages. While it might seem logical to assume that a larger aneurysm could result in more significant bleeding, Wiebers et al. have suggested the possibility of aneurysms diminishing in size upon rupture. 26 Our study found only a weak correlation between aneurysm size and hemorrhage volume (Fig. 2 ). This disparity may stem from the intricate combination of variables at play during aneurysm rupture, including blood pressure, the diameter of the parent artery, and the presence on adjacent structures such as brain parenchyma that would potentially constrain the amount of hemorrhage, among other factors. This study has several limitations. Being a retrospective analysis of a large patient cohort, certain confounding factors that could influence clinical outcomes may not have been precisely accounted for. Consequently, our analysis concentrated on identifying symptomatic vasospasm rather than DCI, since the latter poses more challenges for retrospective determination. In contrast, vasospasm was objectively documented during hospitalization, making it a more reliable focus for our study. Conclusion The relationship between the objective quantification of blood volume and the Fisher scale was moderately correlated at best. Increased hemorrhage volumes, as objectively measured were correlated with an increased risk of complications. Through precise quantification, it was possible to identify specific hemorrhage volume thresholds associated with complications like vasospasm, acute hydrocephalus, SDH, and mortality within seven days. Notably, the size of the aneurysm did not seem to have a significant correlation with the volume of hemorrhage. Declarations Disclosure of Potential Conflict of Interests EAS is consultant for Medtronic, Microvention, Rapid Medical, Cerenovus and iSchemaView. Details page The manuscript complies with all instructions to authors. The authorship requirements have been met and the final manuscript has been approved by all authors. Authors Contribution: Conception and design of the study: EAS and SS. Acquisition and analysis of data: JMM, MTJ, RRP, ES, AG, CD, AV. Manuscript drafting and final approval: all authors. This manuscript has not been published somewhere else and is not under consideration for another journal. We adhered to the ethical guidelines and obtained IRB approval. Any author has conflict of interests to declare. We included the STROBE checklist. We did not use any source of funding for this research. References Harrod CG, Bendok BR, Batjer HH. Prediction of Cerebral Vasospasm in Patients Presenting with Aneurysmal Subarachnoid Hemorrhage: A Review. Neurosurgery . 2005;56. Fisher CM, Kistler JP, Davis JM. Relation of cerebral vasospasm to subarachnoid hemorrhage visualized by computerized tomographic scanning. Neurosurgery . 1980;6:1-9. Claassen J, Bernardini GL, Kreiter K, Bates J, Du YE, Copeland D, Connolly ES, Mayer SA. Effect of cisternal and ventricular blood on risk of delayed cerebral ischemia after subarachnoid hemorrhage: the Fisher scale revisited. Stroke . 2001;32:2012-2020. Kistler JP, Crowell RM, Davis KR, Heros R, Ojemann RG, Zervas T, Fisher CM. The relation of cerebral vasospasm to the extent and location of subarachnoid blood visualized by CT scan: a prospective study. Neurology . 1983;33:424-436. doi: 10.1212/wnl.33.4.424 Melinosky C, Kincaid H, Claassen J, Parikh G, Badjatia N, Morris NA. The Modified Fisher Scale Lacks Interrater Reliability. Neurocrit Care . 2021;35:72-78. doi: 10.1007/s12028-020-01142-8 Hunt WE, Hess RM. Surgical risk as related to time of intervention in the repair of intracranial aneurysms. J Neurosurg . 1968;28:14-20. doi: 10.3171/jns.1968.28.1.0014 Ogungbo B. The World Federation of Neurological Surgeons scale for subarachnoid haemorrhage. Surgical Neurology . 2003;59:236-237. doi: https://doi.org/10.1016/S0090-3019(02)01049-2 García S, Torné R, Hoyos JA, Rodríguez-Hernández A, Amaro S, Llull L, López-Rueda A, Enseñat J. Quantitative versus qualitative blood amount assessment as a predictor for shunt-dependent hydrocephalus following aneurysmal subarachnoid hemorrhage. J Neurosurg . 2018;131:1743-1750. doi: 10.3171/2018.7.Jns18816 Friedman JA, Goerss SJ, Meyer FB, Piepgras DG, Pichelmann MA, McIver JI, Toussaint LG, 3rd, McClelland RL, Nichols DA, Atkinson JL, et al. Volumetric quantification of Fisher Grade 3 aneurysmal subarachnoid hemorrhage: a novel method to predict symptomatic vasospasm on admission computerized tomography scans. J Neurosurg . 2002;97:401-407. doi: 10.3171/jns.2002.97.2.0401 Zijlstra IA, Gathier CS, Boers AM, Marquering HA, Slooter AJ, Velthuis BK, Coert BA, Verbaan D, van den Berg R, Rinkel GJ, et al. Association of Automatically Quantified Total Blood Volume after Aneurysmal Subarachnoid Hemorrhage with Delayed Cerebral Ischemia. AJNR Am J Neuroradiol . 2016;37:1588-1593. doi: 10.3174/ajnr.A4771 Scherer M, Cordes J, Younsi A, Sahin YA, Götz M, Möhlenbruch M, Stock C, Bösel J, Unterberg A, Maier-Hein K, et al. Development and Validation of an Automatic Segmentation Algorithm for Quantification of Intracerebral Hemorrhage. Stroke . 2016;47:2776-2782. doi: 10.1161/strokeaha.116.013779 Renan Sales B, Wessel E, Anna MMB, Zijlstra IJ, Rene, Wassim, Alexandre U, Dagmar V, Peter V, Charles M, et al. Automated segmentation of subarachnoid hemorrhages with convolutional neural networks. Informatics in Medicine Unlocked . 2020;19:100321. doi: https://doi.org/10.1016/j.imu.2020.100321 Xu J, Zhang R, Zhou Z, Wu C, Gong Q, Zhang H, Wu S, Wu G, Deng Y, Xia C, et al. Deep Network for the Automatic Segmentation and Quantification of Intracranial Hemorrhage on CT. Front Neurosci . 2020;14:541817. doi: 10.3389/fnins.2020.541817 Sanchez S, Hickerson M, Patel RR, Ghazaleh D, Tarchand R, Paranjape GS, Pope H, Ortega‐Gutierrez S, Pederson JM, Hasan D, et al. Morphological Characteristics of Ruptured Brain Aneurysms: A Systematic Literature Review and Meta‐Analysis. Stroke: Vascular and Interventional Neurology .0:e000707. doi: doi:10.1161/SVIN.122.000707 Zou KH, Warfield SK, Bharatha A, Tempany CM, Kaus MR, Haker SJ, Wells WM, 3rd, Jolesz FA, Kikinis R. Statistical validation of image segmentation quality based on a spatial overlap index. Acad Radiol . 2004;11:178-189. doi: 10.1016/s1076-6332(03)00671-8 Zanaty M, Allan L, Samaniego EA, Piscopo A, Ryan E, Torner JC, Hasan D. Phase 1/2a Trial of ISPASM. Stroke . 2021;52:3750-3758. doi: doi:10.1161/STROKEAHA.121.034578 Danière F, Gascou G, Menjot de Champfleur N, Machi P, Leboucq N, Riquelme C, Ruiz C, Bonafé A, Costalat V. Complications and follow up of subarachnoid hemorrhages. Diagnostic and Interventional Imaging . 2015;96:677-686. doi: https://doi.org/10.1016/j.diii.2015.05.006 van der Jagt M, Hasan D, Bijvoet HW, Pieterman H, Koudstaal PJ, Avezaat CJ. Interobserver variability of cisternal blood on CT after aneurysmal subarachnoid hemorrhage. Neurology . 2000;54:2156-2158. doi: 10.1212/wnl.54.11.2156 van Norden AG, van Dijk GW, van Huizen MD, Algra A, Rinkel GJ. Interobserver agreement and predictive value for outcome of two rating scales for the amount of extravasated blood after aneurysmal subarachnoid haemorrhage. J Neurol . 2006;253:1217-1220. doi: 10.1007/s00415-006-0205-0 Matsoukas S, Scaggiante J, Schuldt BR, Smith CJ, Chennareddy S, Kalagara R, Majidi S, Bederson JB, Fifi JT, Mocco J, et al. Accuracy of artificial intelligence for the detection of intracranial hemorrhage and chronic cerebral microbleeds: a systematic review and pooled analysis. Radiol Med . 2022;127:1106-1123. doi: 10.1007/s11547-022-01530-4 Platz J, Güresir E, Wagner M, Seifert V, Konczalla J. Increased risk of delayed cerebral ischemia in subarachnoid hemorrhage patients with additional intracerebral hematoma. J Neurosurg . 2017;126:504-510. doi: 10.3171/2015.12.Jns151563 Lagares A, Jiménez-Roldán L, Gomez PA, Munarriz PM, Castaño-León AM, Cepeda S, Alén JF. Prognostic Value of the Amount of Bleeding After Aneurysmal Subarachnoid Hemorrhage: A Quantitative Volumetric Study. Neurosurgery . 2015;77:898-907; discussion 907. doi: 10.1227/neu.0000000000000927 Street JS, Pandit AS, Toma AK. Predicting vasospasm risk using first presentation aneurysmal subarachnoid hemorrhage volume: A semi-automated CT image segmentation analysis using ITK-SNAP. PLoS One . 2023;18:e0286485. doi: 10.1371/journal.pone.0286485 Haedo MG, Grille P, Burghi G, Barbato M. Correlation between tomographic scales and vasospasm and delayed cerebral ischemia in aneurysmal subarachnoid hemorrhage. Crit Care Sci . 2023;35:311-319. doi: 10.5935/2965-2774.20230119-en Wilson CD, Safavi-Abbasi S, Sun H, Kalani MY, Zhao YD, Levitt MR, Hanel RA, Sauvageau E, Mapstone TB, Albuquerque FC, et al. Meta-analysis and systematic review of risk factors for shunt dependency after aneurysmal subarachnoid hemorrhage. J Neurosurg . 2017;126:586-595. doi: 10.3171/2015.11.Jns152094 Wiebers DO, Whisnant JP, Sundt TM, Jr., O'Fallon WM. The significance of unruptured intracranial saccular aneurysms. J Neurosurg . 1987;66:23-29. doi: 10.3171/jns.1987.66.1.0023 Kroon D-J. Region Growing. MATLAB Central File Exchange. https://www.mathworks.com/matlabcentral/fileexchange/19084-region-growing. 2022. Accessed June 14, 2022. Supplementary Files STROBEchecklistv4combined.docx Sup1.jpg Sup2.tif Supplemental.docx Cite Share Download PDF Status: Published Journal Publication published 25 Sep, 2024 Read the published version in Neurocritical Care → Version 1 posted Reviewers agreed at journal 29 Apr, 2024 Reviewers invited by journal 29 Apr, 2024 Editor invited by journal 26 Apr, 2024 Editor assigned by journal 24 Apr, 2024 First submitted to journal 24 Apr, 2024 You are reading this latest preprint version Research Square lets you share your work early, gain feedback from the community, and start making changes to your manuscript prior to peer review in a journal. As a division of Research Square Company, we’re committed to making research communication faster, fairer, and more useful. We do this by developing innovative software and high quality services for the global research community. Our growing team is made up of researchers and industry professionals working together to solve the most critical problems facing scientific publishing. Also discoverable on Platform About Our Team In Review Editorial Policies Advisory Board Help Center Resources Author Services Accessibility API Access RSS feed Manage Cookie Preferences © Research Square 2026 | ISSN 2693-5015 (online) Privacy Policy Terms of Service Do Not Sell My Personal Information {"props":{"pageProps":{"initialData":{"identity":"rs-4308305","acceptedTermsAndConditions":true,"allowDirectSubmit":false,"archivedVersions":[],"articleType":"Research Article","associatedPublications":[],"authors":[{"id":296740814,"identity":"a68df979-3d8f-4ff8-85cd-106eb28d83f8","order_by":0,"name":"Sebastian Sanchez","email":"","orcid":"","institution":"Yale New Haven Hospital","correspondingAuthor":false,"prefix":"","firstName":"Sebastian","middleName":"","lastName":"Sanchez","suffix":""},{"id":296740815,"identity":"f8d6f67b-e091-4bfe-b065-14dabab4e6d1","order_by":1,"name":"Jacob M Miller","email":"","orcid":"","institution":"The University of Iowa","correspondingAuthor":false,"prefix":"","firstName":"Jacob","middleName":"M","lastName":"Miller","suffix":""},{"id":296740816,"identity":"c297bc58-e030-4c37-a8ac-9eb288285db6","order_by":2,"name":"Matthew T Jones","email":"","orcid":"","institution":"The University of Iowa Roy J and Lucille A Carver College of Medicine","correspondingAuthor":false,"prefix":"","firstName":"Matthew","middleName":"T","lastName":"Jones","suffix":""},{"id":296740817,"identity":"868010e4-fd4a-453c-9719-e6a3e745de27","order_by":3,"name":"Rishi R Patel","email":"","orcid":"","institution":"The University of Iowa Roy J and Lucille A Carver College of Medicine","correspondingAuthor":false,"prefix":"","firstName":"Rishi","middleName":"R","lastName":"Patel","suffix":""},{"id":296740818,"identity":"5dd0ef54-f2ae-42be-92f3-99ea9614342f","order_by":4,"name":"Elena Sagues","email":"","orcid":"","institution":"The University of Iowa","correspondingAuthor":false,"prefix":"","firstName":"Elena","middleName":"","lastName":"Sagues","suffix":""},{"id":296740819,"identity":"046c9391-5e08-4901-a59f-f2de79cd6496","order_by":5,"name":"Carlos Dier","email":"","orcid":"","institution":"The University of Iowa","correspondingAuthor":false,"prefix":"","firstName":"Carlos","middleName":"","lastName":"Dier","suffix":""},{"id":296740820,"identity":"a2fd045f-712a-4f3b-aa32-9d9818e112b1","order_by":6,"name":"Andres Gudino","email":"","orcid":"","institution":"The University of Iowa","correspondingAuthor":false,"prefix":"","firstName":"Andres","middleName":"","lastName":"Gudino","suffix":""},{"id":296740821,"identity":"b0904d69-cb3a-4015-b10e-0a24cb522fbc","order_by":7,"name":"Ariel Vargas-Sanchez","email":"","orcid":"","institution":"The University of Iowa","correspondingAuthor":false,"prefix":"","firstName":"Ariel","middleName":"","lastName":"Vargas-Sanchez","suffix":""},{"id":296740822,"identity":"ad9bad9f-9f7a-400f-8a39-ad0fb8be5ebf","order_by":8,"name":"Edgar Andres Samaniego","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAA2ElEQVRIiWNgGAWjYFCCBCjN3gAkDCxI0cJzAKRFghQtEmAGEVr42dMvfq74YxfNP/P51Q0/CiQY+Nu7E/Bqkex5Uyx5ti05d8btnLKbPUCHSZw5uwGvFoMbOQmSjQ3MuQ23c9Ju8AC1GEjk4tdifyMn+WfDn/rc+TfPpN38Q4wWA4n0Y5INbIdzN9xgP3abKFskzrxhs2xsO5678UwO220ZAwkegn7hb09/fLPhT3XuvOPHn91888dGjr+9F78WYBQaoDB4CCgHAfYH6IxRMApGwSgYBagAABc6Tcr3IgiiAAAAAElFTkSuQmCC","orcid":"https://orcid.org/0000-0003-2764-2268","institution":"University of Iowa","correspondingAuthor":true,"prefix":"","firstName":"Edgar","middleName":"Andres","lastName":"Samaniego","suffix":""}],"badges":[],"createdAt":"2024-04-22 22:32:19","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-4308305/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-4308305/v1","draftVersion":[],"editorialEvents":[{"content":"https://doi.org/10.1007/s12028-024-02123-x","type":"published","date":"2024-09-25T15:57:20+00:00"}],"editorialNote":"","failedWorkflow":false,"files":[{"id":56037109,"identity":"7e199861-a774-4f1f-b3e0-67bec756245c","added_by":"auto","created_at":"2024-05-07 18:48:46","extension":"png","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":891803,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eAutomated hemorrhage quantification.\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e\u003cem\u003eCT= Computed tomography\u003c/em\u003e\u003c/p\u003e\n\u003cp\u003eRaw non-contrast CT images with intraventricular blood (A) and parenchymal blood (C). Automatic identification of hemorrhage voxels overlayed onto the CT images (B, D). The automated software is able to detect and quantify both intraparenchymal and intraventricular hemorrhages (B). The detected hemorrhage area was 1460.5 mm\u003csup\u003e2\u003c/sup\u003e and 1481.4 mm\u003csup\u003e2\u003c/sup\u003e in panels B and D, respectively.\u003c/p\u003e","description":"","filename":"Figure1.png","url":"https://assets-eu.researchsquare.com/files/rs-4308305/v1/15787f9044dfc8a7684ff51c.png"},{"id":56037116,"identity":"ce1b2f0f-1eaa-439f-8e49-dd6eb81f5488","added_by":"auto","created_at":"2024-05-07 18:48:47","extension":"png","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":2339020,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eCorrelation between aneurysmal size and hemorrhage volume.\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e\u003cem\u003eCT= Computed tomography\u003c/em\u003e\u003c/p\u003e\n\u003cp\u003eSize had a weak correlation with hemorrhage volume. Hemorrhage volume can be lower in big aneurysms (A, B) compared to smaller aneurysms (C, D). Moreover, aneurysms that are similar in size (F, H) can have very different hemorrhage volumes (E, G).\u003c/p\u003e","description":"","filename":"Figure2.png","url":"https://assets-eu.researchsquare.com/files/rs-4308305/v1/abc0c6555d11a45e679ecdf0.png"},{"id":56037114,"identity":"5a38dd19-9a08-4970-aae9-e9a22210942d","added_by":"auto","created_at":"2024-05-07 18:48:47","extension":"png","order_by":3,"title":"Figure 3","display":"","copyAsset":false,"role":"figure","size":1071666,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eROC analysis of aneurysmal subarachnoid hemorrhage volume as a predictor of outcomes.\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e\u003cem\u003eAUC= Area under the curve, ROC=Receiver Operating Characteristics, SDH= Shunt dependent hydrocephalus\u003c/em\u003e\u003c/p\u003e\n\u003cp\u003e(A) Specific thresholds for vasospasm (15 mL), acute hydrocephalus (10 mL), shunt-dependent hydrocephalus (SDH) (17 mL), and death within 7 days (34 mL) are presented. The automated algorithm facilitates accurate quantification of hemorrhage in patients with varying Fisher scores. (B) A patient with a Fisher score of II and a hemorrhage volume of 7 mL (below all established thresholds) did not experience any of the studied complications. (C) A patient with a Fisher score of III and a hemorrhage volume of 15.36 mL developed vasospasm (threshold for vasospasm = 15 mL). Notably, complications appear to vary more based on intracranial hemorrhage volume than Fisher score. (D) A patient with a Fisher score of IV had 10.79 mL of hemorrhage and experienced complications related to hydrocephalus (threshold for hydrocephalus = 10 mL). (E) Another patient with a Fisher score of IV and a hemorrhage volume of 34.14 mL, died within 7 days (threshold for death = 34 mL).\u003c/p\u003e","description":"","filename":"Figure3.png","url":"https://assets-eu.researchsquare.com/files/rs-4308305/v1/46829d987e31724152f3015e.png"},{"id":65627162,"identity":"2cd33718-bf12-4570-a38c-c18724c1b93f","added_by":"auto","created_at":"2024-09-30 16:12:41","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":6119049,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-4308305/v1/eb55f6c0-37e7-4802-9719-fc237ec82600.pdf"},{"id":56037110,"identity":"fb50e7b3-da5e-4c52-8b16-bad01bc6de51","added_by":"auto","created_at":"2024-05-07 18:48:46","extension":"docx","order_by":1,"title":"","display":"","copyAsset":false,"role":"supplement","size":34069,"visible":true,"origin":"","legend":"","description":"","filename":"STROBEchecklistv4combined.docx","url":"https://assets-eu.researchsquare.com/files/rs-4308305/v1/6c913715df8da558f81ef403.docx"},{"id":56037111,"identity":"804e4c44-1ade-4938-b147-7dff343243b5","added_by":"auto","created_at":"2024-05-07 18:48:47","extension":"jpg","order_by":2,"title":"","display":"","copyAsset":false,"role":"supplement","size":14733,"visible":true,"origin":"","legend":"","description":"","filename":"Sup1.jpg","url":"https://assets-eu.researchsquare.com/files/rs-4308305/v1/125364817ed59359841295be.jpg"},{"id":56037117,"identity":"d556a31d-abf0-469a-9a36-06c37391e583","added_by":"auto","created_at":"2024-05-07 18:48:47","extension":"tif","order_by":3,"title":"","display":"","copyAsset":false,"role":"supplement","size":590240,"visible":true,"origin":"","legend":"","description":"","filename":"Sup2.tif","url":"https://assets-eu.researchsquare.com/files/rs-4308305/v1/16483fb6ab43a0d1329f91a2.tif"},{"id":56038027,"identity":"f9c8b3b1-1d63-4771-88a0-18d917233ed5","added_by":"auto","created_at":"2024-05-07 18:56:47","extension":"docx","order_by":4,"title":"","display":"","copyAsset":false,"role":"supplement","size":21981,"visible":true,"origin":"","legend":"","description":"","filename":"Supplemental.docx","url":"https://assets-eu.researchsquare.com/files/rs-4308305/v1/158d15dd2aed7fc19b23da64.docx"}],"financialInterests":"","formattedTitle":"Automated Hemorrhage Volume Quantification in Aneurysmal Subarachnoid Hemorrhage","fulltext":[{"header":"Introduction","content":"\u003cp\u003eHemorrhage volume plays a crucial role in evaluating the likelihood of complications following aneurysmal subarachnoid hemorrhage (aSAH).\u003csup\u003e\u003cspan citationid=\"CR1\" class=\"CitationRef\"\u003e1\u003c/span\u003e\u003c/sup\u003e The Fisher score\u003csup\u003e\u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e2\u003c/span\u003e\u003c/sup\u003e and modified Fisher score\u003csup\u003e\u003cspan citationid=\"CR3\" class=\"CitationRef\"\u003e3\u003c/span\u003e\u003c/sup\u003e are radiological grading systems designed to predict the likelihood of vasospasm after aSAH. These scores are based on the thickness and location of hemorrhage and have been validated in stratifying the risk of vasospasm following aSAH.\u003csup\u003e\u003cspan additionalcitationids=\"CR3\" citationid=\"CR2\" class=\"CitationRef\"\u003e2\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR4\" class=\"CitationRef\"\u003e4\u003c/span\u003e\u003c/sup\u003e However, adjudication of the scores is operator dependent and has low interrater reliability.\u003csup\u003e\u003cspan citationid=\"CR5\" class=\"CitationRef\"\u003e5\u003c/span\u003e\u003c/sup\u003e Moreover, they provide scarce information about the risk for acute hydrocephalus, shunt dependent hydrocephalus (SDH) and death in the acute phase. Additionally, clinical grading scales such as the Hunt and Hess (H\u0026amp;H)\u003csup\u003e\u003cspan citationid=\"CR6\" class=\"CitationRef\"\u003e6\u003c/span\u003e\u003c/sup\u003e and the World federation of neurosurgeons (WFNS)\u003csup\u003e\u003cspan citationid=\"CR7\" class=\"CitationRef\"\u003e7\u003c/span\u003e\u003c/sup\u003e have been used to assess surgical and complication risk after aSAH. Nonetheless, the relationship between these scales and the volume of hemorrhage remains to be evaluated. Recently, manual,\u003csup\u003e\u003cspan citationid=\"CR8\" class=\"CitationRef\"\u003e8\u003c/span\u003e\u003c/sup\u003e semiautomated,\u003csup\u003e\u003cspan citationid=\"CR9\" class=\"CitationRef\"\u003e9\u003c/span\u003e\u003c/sup\u003e and automated methods,\u003csup\u003e\u003cspan additionalcitationids=\"CR11 CR12\" citationid=\"CR10\" class=\"CitationRef\"\u003e10\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR13\" class=\"CitationRef\"\u003e13\u003c/span\u003e\u003c/sup\u003e that have been used to quantify hemorrhage volume following aSAH. However, there is no consensus on the hemorrhage volume thresholds that can increase the likelihood of complications following aSAH.\u003c/p\u003e \u003cp\u003eIn this study an automated method was used to quantify hemorrhage volume in a large cohort of patients with aSAH. We hypothesize that an objective and more accurate method to quantify hemorrhage volume could achieve good correlations with complications following aSAH. We specifically aimed to identify thresholds for hemorrhage volume associated with arterial vasospasm, acute hydrocephalus, SDH, and death within 7 days following aSAH. Additionally, we investigated the potential correlation between hemorrhage volume and clinical scales such as WFNS and H\u0026amp;H. Finally, we also studied if there is any potential association between the size of the aneurysm and hemorrhage volume.\u003c/p\u003e"},{"header":"Methods","content":"\u003cp\u003e After institutional review board approval patients diagnosed with an aSAH at our institution from 2009 to 2022 were included in the study. The TriNetX search engine was utilized to identify patients with aSAH. Patients with a Fisher score of 1 where excluded from the study as this score corresponds to no quantifiable blood in the subarachnoid space.\u003csup\u003e\u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e2\u003c/span\u003e\u003c/sup\u003e Patient demographics, H\u0026amp;H WFNS scores and available imaging were retrieved for analysis. Aneurysmal size was defined as aneurysm maximal diameter, and was determined with the best vascular imaging.\u003csup\u003e\u003cspan citationid=\"CR14\" class=\"CitationRef\"\u003e14\u003c/span\u003e\u003c/sup\u003e\u003c/p\u003e \u003cdiv id=\"Sec3\" class=\"Section2\"\u003e \u003ch2\u003eHemorrhage quantification algorithm\u003c/h2\u003e \u003cp\u003eAn automated software was developed to identify and quantify aSAH using MATLAB (MATLAB and Image Processing Toolbox, release 2020b; MathWorks, Inc.). DICOM files containing non-contrast computed tomography (CT) scans were loaded into MATLAB. Hounsfield unit (HU) values below zero were identified as background and air, while HU values above 160 were considered bone, and were excluded.\u003csup\u003e\u003cspan citationid=\"CR12\" class=\"CitationRef\"\u003e12\u003c/span\u003e\u003c/sup\u003e The lower anatomical limit for aSAH detection was the foramen magnum, as established in a previous study by Renan Sales et al.\u003csup\u003e\u003cspan citationid=\"CR12\" class=\"CitationRef\"\u003e12\u003c/span\u003e\u003c/sup\u003e\u003c/p\u003e \u003cp\u003eA 3D segmentation of intracranial blood was created by using k-means clustering, and four different tissue groups based on HUs. These areas tentatively corresponded to blood, white matter, gray matter, and cerebrospinal fluid (CSF). The highest HU values were adjudicated as blood and isolated as a blood cluster. MATLAB functions were then used to isolate and exclude individual voxels corresponding to the blood cluster if they were in other tissues as they were likely artifact. Similarly, individual voxels adjudicated as blood were removed if they were adjacent to the skull to avoid potential contamination. The blood segmentation was overlayed on the original CT scan slices for quality control (Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003e). Finally, the aSAH volume was calculated by the automated algorithm. Detailed technical methods are described in the supplementary materials.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec4\" class=\"Section2\"\u003e \u003ch2\u003eValidation of the automated algorithm for aSAH detection\u003c/h2\u003e \u003cp\u003eTwo independent raters manually measured hemorrhage volumes from ten randomly selected non-contrast CT scans using the region of interest (ROI) tool on PACS (Carestream Vue). The areas from the ROIs were summed to quantify the hemorrhage volume in each slice. Then, the hemorrhage volume in all slices were summed and multiplied by the slice thickness (0.5 mm) to obtain the total hemorrhage volume. The automated algorithm was then used to calculate the hemorrhage volume of the same ten CT scans. The Dice Similarity Coefficient was calculated to compare the performance of the automated method with the manual measurements.\u003csup\u003e\u003cspan citationid=\"CR15\" class=\"CitationRef\"\u003e15\u003c/span\u003e\u003c/sup\u003e Dice Coefficients were calculated against each of the raters\u0026rsquo; scans, as follows:\u003cdiv id=\"Equa\" class=\"Equation\"\u003e\u003cdiv format=\"TEX\" class=\"mathdisplay\" id=\"FileID_Equa\" name=\"EquationSource\"\u003e\n$$DSC=\\frac{2*A\\cap B}{A+B}$$\u003c/div\u003e\u003c/div\u003e\u003c/p\u003e \u003cp\u003eA is the number of voxels in the rater\u0026rsquo;s manual segmentation, B is the number of voxels in the automatic segmentation, and A \u0026cap; B is the number of voxels the segmentations overlap. Dice scores compared the volumes as well as the spatial distributions of voxels being identified by each method.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec5\" class=\"Section2\"\u003e \u003ch2\u003eClinical outcomes\u003c/h2\u003e \u003cp\u003eThe following clinical outcomes were included in the analysis: symptomatic vasospasm, presence of acute hydrocephalus, SDH and death within seven days. Symptomatic vasospasm was defined as focal neurological impairment that occurred within 3\u0026ndash;14 days in the setting of radiographic vasospasm documented on computed tomography angiography or digital subtraction angiography, and not attributed to other causes.\u003csup\u003e\u003cspan citationid=\"CR16\" class=\"CitationRef\"\u003e16\u003c/span\u003e\u003c/sup\u003e Symptomatic vasospasm was also adjudicated if patients required endovascular treatment for vasospasm after aSAH. Acute hydrocephalus was identified as an enlargement of the ventricles due to obstruction of the CSF drainage due to intracranial bleeding.\u003csup\u003e\u003cspan citationid=\"CR17\" class=\"CitationRef\"\u003e17\u003c/span\u003e\u003c/sup\u003e SDH was adjudicated if the patient required any permanent drainage device such as a ventriculoperitoneal shunt. Finally, death within seven days was adjudicated in the case a patient passed within 7 days of aneurysm rupture.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec6\" class=\"Section2\"\u003e \u003ch2\u003eStatistical Analysis\u003c/h2\u003e \u003cp\u003eIBM SPSS Statistics 27.0 was utilized for statistical analysis. The Shapiro-Wilk test was used to assess distribution among variables. Normally distributed variables are reported as mean and standard deviation (SD), while non-normally distributed variables are presented as median and interquartile range (IQR). A student t-test was used to evaluate differences among normally distributed variables. The Mann-Whitney U test was used to compare non-normally distributed variables. Pearson intraclass correlation coefficients and Spearman\u0026rsquo;s Rho were used to evaluate correlations in normal and non-normally distributed variables. A Bland-Altman analysis was conducted to compare algorithmic volumes to volumes averaged among raters. Univariate logistic regression analysis was performed to assess the relationship between hemorrhage volume and four separate clinical outcomes: vasospasm, acute hydrocephalus, SDH, and death within 7 days. Receiver operating characteristic (ROC) curves were created to analyze potential thresholds of hemorrhage volume that could discriminate for the presence of the previously described outcomes. These thresholds, along with quantified area under the curve (AUC), were extracted from the ROC curves. The significance level was set at 0.05.\u003c/p\u003e \u003c/div\u003e"},{"header":"Results","content":"\u003cdiv id=\"Sec8\" class=\"Section2\"\u003e \u003ch2\u003eValidation of the aSAH algorithm\u003c/h2\u003e \u003cp\u003eThe manual raters\u0026rsquo; intraclass coefficient was 0.993 (p\u0026thinsp;\u0026lt;\u0026thinsp;0.001), and the intraclass coefficient between the averaged manual volumes and the automated algorithm was 0.973 (p\u0026thinsp;\u0026lt;\u0026thinsp;0.001). The average Dice Coefficient between the segmentations was 0.730. The mean difference between manual and automated measurements based on the Bland-Altman analysis was of -3.14 mL, and the mean absolute error was 4.33 mL (Supplementary Fig.\u0026nbsp;1).\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec9\" class=\"Section2\"\u003e \u003ch2\u003eAneurysm Characteristics and hemorrhage volume\u003c/h2\u003e \u003cp\u003eFive-hundred patients with 500 ruptured aneurysms were included in the analysis. The mean age was 57.9\u0026thinsp;\u0026plusmn;\u0026thinsp;13.4, 65.8% were female (n\u0026thinsp;=\u0026thinsp;329), and 88.8% (n\u0026thinsp;=\u0026thinsp;444) were white (Table\u0026nbsp;\u003cspan refid=\"Tab1\" class=\"InternalRef\"\u003e1\u003c/span\u003e). The median aneurysm size was 5.4 mm [4.0, 7.9], and the median intracranial hemorrhage volume was 18.0 mL [5.7, 37.5] (Supplementary table \u003cspan refid=\"MOESM1\" class=\"InternalRef\"\u003e1\u003c/span\u003e). Thirty-one percent (n\u0026thinsp;=\u0026thinsp;148) of saccular aneurysms had a diameter larger than 7 mm. Aneurysmal size and volume of hemorrhage had a weak linear correlation (Rho\u0026thinsp;=\u0026thinsp;0.139) (Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003e). Similarly, aneurysm size did not correlate with the Fisher score (Rho\u0026thinsp;=\u0026thinsp;0.18), H\u0026amp;H (Rho\u0026thinsp;=\u0026thinsp;0.186) and WFNS grade (Rho\u0026thinsp;=\u0026thinsp;0.164).\u003c/p\u003e \u003cp\u003e \u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab1\" border=\"1\"\u003e \u003ccaption language=\"En\"\u003e \u003cdiv class=\"CaptionNumber\"\u003eTable 1\u003c/div\u003e \u003cdiv class=\"CaptionContent\"\u003e \u003cp\u003ePatient demographics\u003c/p\u003e \u003c/div\u003e \u003c/caption\u003e \u003ccolgroup cols=\"2\"\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e \u003cthead\u003e \u003ctr\u003e \u003cth align=\"left\" colspan=\"2\" nameend=\"c2\" namest=\"c1\"\u003e \u003cp\u003eDemographics\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003c/thead\u003e \u003ctbody\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eAge (mean\u0026thinsp;\u0026plusmn;\u0026thinsp;SD)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e57.9\u0026thinsp;\u0026plusmn;\u0026thinsp;13.4\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eFemale N (%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e329 (65.8)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eWhite N (%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e444 (88.8)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c2\" namest=\"c1\"\u003e \u003cp\u003e\u003cb\u003eFisher grade N (%)\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eFisher 2\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e97 (19.4)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eFisher 3\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e121 (24.2)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eFisher 4\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e282 (56.4)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c2\" namest=\"c1\"\u003e \u003cp\u003e\u003cb\u003eHunt and Hess N (%)\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eGrade 1\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e86 (17.2)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eGrade 2\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e147 (29.4)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eGrade 3\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e119 (23.8)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eGrade 4\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e64 (12.8)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eGrade 5\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e84 (16.8)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c2\" namest=\"c1\"\u003e \u003cp\u003e\u003cb\u003eWord federation of Neurosurgical Societies N (%)\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eGrade 1\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e200 (40)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eGrade 2\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e78 (15.6)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eGrade 3\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e14 (2.8)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eGrade 4\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e79 (15.8)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eGrade 5\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e129 (25.8)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c2\" namest=\"c1\"\u003e \u003cp\u003e\u003cb\u003eComplications N (%)\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eSymptomatic Vasospasm\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e161 (36.4)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eAcute hydrocephalus\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e399 (79.8)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eShunt dependent hydrocephalus\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e127 (28.7)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eDeath\u0026thinsp;\u0026lt;\u0026thinsp;7 days\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e58 (11.6)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c2\" namest=\"c1\"\u003e \u003cp\u003e\u003cb\u003eLocation N (%)\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eAnterior\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e336 (67.2)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003ePosterior\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e164 (32.8)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c2\" namest=\"c1\"\u003e \u003cp\u003e\u003cb\u003eType N (%)\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eSaccular\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e484 (96.8)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c2\" namest=\"c1\"\u003e \u003cp\u003e\u003cb\u003eSaccular Aneurysm Size N (%)\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u0026gt;\u0026thinsp;7 mm\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e148 (30.8)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u0026le;\u0026thinsp;7 mm\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e332 (69.2)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003c/tbody\u003e \u003c/colgroup\u003e \u003c/table\u003e\u003c/div\u003e \u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003e \u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab2\" border=\"1\"\u003e \u003ccaption language=\"En\"\u003e \u003cdiv class=\"CaptionNumber\"\u003eTable 2\u003c/div\u003e \u003cdiv class=\"CaptionContent\"\u003e \u003cp\u003eHemorrhage volume quantification and aneurysm size measurements\u003c/p\u003e \u003c/div\u003e \u003c/caption\u003e \u003ccolgroup cols=\"3\"\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e \u003cthead\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c1\"\u003e \u003cp\u003eClinical Presentation\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c2\"\u003e \u003cp\u003eVolume of Hemorrhage mL\u003c/p\u003e \u003cp\u003eMedian [IQR]\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c3\"\u003e \u003cp\u003eAneurysm Diameter mm\u003c/p\u003e \u003cp\u003eMedian [IQR]\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003ctr\u003e \u003cth align=\"left\" colspan=\"3\" nameend=\"c3\" namest=\"c1\"\u003e \u003cp\u003eFisher Grade\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003c/thead\u003e \u003ctbody\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eFisher 2\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e4.5 [1.5, 10.5]\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e5.0 [4.0, 6.5]\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eFisher 3\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e13.4 [5.0, 26.3]\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e5.0 [3.4, 7.0]\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eFisher 4\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e29.5 [13.8, 51.8]\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e6.0 [4.4, 8.8]\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eRho\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e\u003cb\u003e0.514\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e\u003cb\u003e0.180\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colspan=\"3\" nameend=\"c3\" namest=\"c1\"\u003e \u003cp\u003e\u003cb\u003eHunt and Hess\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eGrade 1\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e8.5 [3.1, 15.2]\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e5.4 [4.3, 7.0]\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eGrade 2\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e7.0 [3.1, 17.7]\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e4.9 [3.5, 6.9]\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eGrade 3\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e20.8 [10.6, 35.5]\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e5.5 [3.9, 7.5]\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eGrade 4\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e40.5 [24.8, 64.0]\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e5.2 [4.0, 8.2]\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eGrade 5\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e49.8 [32.3, 76.2]\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e7.4 [5.0, 11.4]\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eRho\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e\u003cb\u003e0.604\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e\u003cb\u003e0.186\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colspan=\"3\" nameend=\"c3\" namest=\"c1\"\u003e \u003cp\u003e\u003cb\u003eWorld Federation of Neurosurgical Societies\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eGrade 1\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e7.1 [2.9, 16.7]\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e5.2 [3.9, 7.0]\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eGrade 2\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e14.9 [6.7, 30.2]\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e5.0 [3.5, 7.3]\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eGrade 3\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e17.7 [10.2, 25.6]\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e5.1 [3.4, 7.8]\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eGrade 4\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e25.3 [11.4, 39.2]\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e5.0 [3.9, 7.3]\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eGrade 5\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e51.0 [28.6, 73.2]\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e6.3 [5.0, 10.5]\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eRho\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e\u003cb\u003e0.613\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e\u003cb\u003e0.164\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003c/tbody\u003e \u003c/colgroup\u003e \u003c/table\u003e\u003c/div\u003e \u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec10\" class=\"Section2\"\u003e \u003ch2\u003eHemorrhage volume and Clinical Outcomes\u003c/h2\u003e \u003cp\u003eHemorrhage volume had a moderate correlation with the Fisher score (Rho\u0026thinsp;=\u0026thinsp;0.514), and a strong correlation with a higher H\u0026amp;H scale (Rho\u0026thinsp;=\u0026thinsp;0.604) and higher WFNS grade (Rho\u0026thinsp;=\u0026thinsp;0.613) (Supplementary Fig.\u0026nbsp;2). There were significantly higher hemorrhage volumes in patients with vasospasm (21.7 [10.9, 41.4] vs 10.7 [4.2, 26.9], p\u0026thinsp;\u0026lt;\u0026thinsp;0.001), acute hydrocephalus (22.7 [9.2, 41.8] vs 5.1 [2.1, 13.5]; p\u0026thinsp;\u0026lt;\u0026thinsp;0.001), SDH (23.8 [11.3, 40.7] vs 11.7 [4.1, 28.2], p\u0026thinsp;\u0026lt;\u0026thinsp;0.001), and those who died within 7 days (52.8 [34.6, 90.6] mL vs 14.8 [5.0, 32.4] mL, p\u0026thinsp;\u0026lt;\u0026thinsp;0.001) compared to their counterparts (Supplementary Table\u0026nbsp;4). For each mL increase in intracranial hemorrhage, there was a 2% increase in vasospasm (OR 1.02, 95% CI: 1.01\u0026ndash;1.03; p\u0026thinsp;\u0026lt;\u0026thinsp;0.001), a 5% increase in hydrocephalus (OR 1.05, 95% CI: 1.03\u0026ndash;1.07; p\u0026thinsp;\u0026lt;\u0026thinsp;0.001), a 2% increase in SDH (OR 1.02, 95% CI: 1.01\u0026ndash;1.03; p\u0026thinsp;\u0026lt;\u0026thinsp;0.001); and 4% higher death within the first week (OR 1.04, 95% CI: 1.03\u0026ndash;1.05; p\u0026thinsp;\u0026lt;\u0026thinsp;0.001).\u003c/p\u003e \u003cp\u003eROC analysis showed that a hemorrhage volume \u0026gt; 15.16 mL was associated symptomatic vasospasm (65% sensitivity and 60% specificity, AUC\u0026thinsp;=\u0026thinsp;0.656), volume \u0026gt; 9.95 mL with acute hydrocephalus (74% sensitivity and 69% specificity, AUC\u0026thinsp;=\u0026thinsp;0.763), volume \u0026gt; 16.76 mL with SDH (63% sensitivity and 60% specificity, AUC\u0026thinsp;=\u0026thinsp;0.666), and volume \u0026gt; 33.84 mL with mortality within seven days (79% sensitivity and 77% specificity, AUC\u0026thinsp;=\u0026thinsp;0.839) (Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003e).\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003c/div\u003e"},{"header":"Discussion","content":"\u003cp\u003eThe use of automated and objective methods for quantifying hemorrhage volumes has the potential for improving the prediction of complications following aSAH compared to subjective grading scales. The Fisher scale only achieved moderate correlation with the objective quantification of hemorrhage, highlighting the poor performance of this subjective scale. Specific volume thresholds are more likely correlated with the occurrence of vasospasm, acute hydrocephalus, SDH, and mortality within seven days. Furthermore, our findings suggest a weak correlation between the size of the aneurysm and the volume of hemorrhage.\u003c/p\u003e \u003cp\u003eHemorrhage volume post-aSAH has traditionally been assessed subjectively, leading to considerable interrater variability. Van der Gagt et al. observed significant variability in the assessment of hemorrhage volume on 159 aSAH CT scans (kappa 0.34\u0026ndash;0.66).\u003csup\u003e\u003cspan citationid=\"CR18\" class=\"CitationRef\"\u003e18\u003c/span\u003e\u003c/sup\u003e Similarly, Van Norden et al studied 132 aSAH CT scans and found that the interobserver agreement of the Fisher scale was only mild to moderate (kappa 0.37\u0026ndash;0.55).\u003csup\u003e\u003cspan citationid=\"CR19\" class=\"CitationRef\"\u003e19\u003c/span\u003e\u003c/sup\u003e New semiautomated\u003csup\u003e\u003cspan citationid=\"CR9\" class=\"CitationRef\"\u003e9\u003c/span\u003e\u003c/sup\u003e and automated,\u003csup\u003e\u003cspan additionalcitationids=\"CR11 CR12\" citationid=\"CR10\" class=\"CitationRef\"\u003e10\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR13\" class=\"CitationRef\"\u003e13\u003c/span\u003e\u003c/sup\u003e methods have been used for hemorrhage volume quantification. A recent meta-analysis conducted by Matsoukas et al. encompassed 40 studies detailing artificial intelligence (AI) algorithms designed for intracranial hemorrhage detection.\u003csup\u003e\u003cspan citationid=\"CR20\" class=\"CitationRef\"\u003e20\u003c/span\u003e\u003c/sup\u003e On average, most AI-driven software tools had a high performance for the diagnosis of intracerebral hemorrhage, however, specific thresholds for aSAH complications prediction were not reported.\u003c/p\u003e \u003cp\u003eQuantifying hemorrhage volume in aSAH poses challenges due to blood distribution across subarachnoid spaces, ventricles and parenchyma. Ziljstra et al.\u0026rsquo;s analysis of 333 patients using an automated detection method found intraventricular hemorrhage to be a predictor of delayed cerebral ischemia (DCI), while intraparenchymal hematoma showed a negative association.\u003csup\u003e\u003cspan citationid=\"CR10\" class=\"CitationRef\"\u003e10\u003c/span\u003e\u003c/sup\u003e In contrast, Platz et al reported that intraparenchymal hemorrhage was associated with DCI.\u003csup\u003e\u003cspan citationid=\"CR21\" class=\"CitationRef\"\u003e21\u003c/span\u003e\u003c/sup\u003e The precise factors leading to DCI remain unclear, but accurately measuring total intracerebral hemorrhage could better predict DCI and complications like symptomatic vasospasm, SDH, and poor functional outcomes following aSAH.\u003csup\u003e\u003cspan citationid=\"CR22\" class=\"CitationRef\"\u003e22\u003c/span\u003e\u003c/sup\u003e Moreover, in the era of AI, defining blood volume thresholds may refine prognostic evaluations and enhance patient outcomes. In AUC analysis, an aSAH volume of 15 mL exhibited moderate accuracy in predicting symptomatic vasospasm (65.2% sensitivity and 60.1% specificity, AUC 0.65). In a smaller cohort of 42 patients, Street et al. reported that a semi-automated method outperform the modified Fisher scale in predicting radiological vasospasm, achieving a higher accuracy (AUC 0.86 vs 0.70).\u003csup\u003e\u003cspan citationid=\"CR23\" class=\"CitationRef\"\u003e23\u003c/span\u003e\u003c/sup\u003e Variability in reported incidences of radiological vasospasm, symptomatic vasospasm, and DCI across studies highlights the lack of standardized definitions and outcomes in this research area.\u003csup\u003e\u003cspan citationid=\"CR24\" class=\"CitationRef\"\u003e24\u003c/span\u003e\u003c/sup\u003e\u003c/p\u003e \u003cp\u003eIn our analysis, an aSAH volume of 9.95 mL achieved a higher accuracy for the occurrence of hydrocephalus (sensitivity of 74.2% and specificity of 69%). This volume is notably smaller than the volumes associated with vasospasm, and death within 7 days. Only a small volume of hemorrhage may be necessary to obstruct or damage the pathways responsible for CSF circulation and absorption. Hemorrhage volume is also pivotal in assessing the likelihood of SDH. A meta-analysis conducted by Wilson et al. on the risk factors associated with SDH revealed a strong correlation between the presence of intraventricular blood (OR 3.93) and SDH.\u003csup\u003e\u003cspan citationid=\"CR25\" class=\"CitationRef\"\u003e25\u003c/span\u003e\u003c/sup\u003e We report a volume of 16.76 mL for predicting SDH post-aSAH (sensitivity of 64% and specificity of 60%). Previously, Garcia et al. studied a cohort of 168 patients using a manual method for hemorrhage assessment and reported a threshold of 10.3 mL for SDH. This threshold was less sensitive of (50%) but more specific (83%) than the threshold reported in our study.\u003csup\u003e\u003cspan citationid=\"CR8\" class=\"CitationRef\"\u003e8\u003c/span\u003e\u003c/sup\u003e This disparity may be explained by the different clinical criteria in determining the need for a permanent shunt.\u003c/p\u003e \u003cp\u003eMortality in the acute phase of aSAH can be the result of multiple factors, but a larger hemorrhage burden often plays a pivotal role. In our study, a volume of 33 mL of aSAH was very accurate in predicting mortality within 7 days (sensitivity of 79% and specificity of 76.9%). Similarly, Lagares et al. examined 206 patients and found that the proportion of individuals with poor outcomes, defined as Glasgow Outcome Scale score of 3 or lower, increased from 35% when the aSAH total hemorrhage volume was below 20 mL to 86% when it exceeded 20 mL.\u003csup\u003e\u003cspan citationid=\"CR22\" class=\"CitationRef\"\u003e22\u003c/span\u003e\u003c/sup\u003e\u003c/p\u003e \u003cdiv id=\"Sec12\" class=\"Section2\"\u003e \u003ch2\u003eAneurysm size and hemorrhage volume\u003c/h2\u003e \u003cp\u003eIt is unclear whether larger aneurysms lead to larger hemorrhages. While it might seem logical to assume that a larger aneurysm could result in more significant bleeding, Wiebers et al. have suggested the possibility of aneurysms diminishing in size upon rupture.\u003csup\u003e\u003cspan citationid=\"CR26\" class=\"CitationRef\"\u003e26\u003c/span\u003e\u003c/sup\u003e Our study found only a weak correlation between aneurysm size and hemorrhage volume (Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003e). This disparity may stem from the intricate combination of variables at play during aneurysm rupture, including blood pressure, the diameter of the parent artery, and the presence on adjacent structures such as brain parenchyma that would potentially constrain the amount of hemorrhage, among other factors.\u003c/p\u003e \u003cp\u003eThis study has several limitations. Being a retrospective analysis of a large patient cohort, certain confounding factors that could influence clinical outcomes may not have been precisely accounted for. Consequently, our analysis concentrated on identifying symptomatic vasospasm rather than DCI, since the latter poses more challenges for retrospective determination. In contrast, vasospasm was objectively documented during hospitalization, making it a more reliable focus for our study.\u003c/p\u003e \u003c/div\u003e"},{"header":"Conclusion","content":"\u003cp\u003eThe relationship between the objective quantification of blood volume and the Fisher scale was moderately correlated at best. Increased hemorrhage volumes, as objectively measured were correlated with an increased risk of complications. Through precise quantification, it was possible to identify specific hemorrhage volume thresholds associated with complications like vasospasm, acute hydrocephalus, SDH, and mortality within seven days. Notably, the size of the aneurysm did not seem to have a significant correlation with the volume of hemorrhage.\u003c/p\u003e "},{"header":"Declarations","content":"\u003cp\u003e\u003cstrong\u003eDisclosure of Potential Conflict of Interests\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eEAS is consultant for Medtronic, Microvention, Rapid Medical, Cerenovus and iSchemaView.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eDetails page\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe manuscript complies with all instructions to authors. The authorship requirements have been met and the final manuscript has been approved by all authors. Authors Contribution: Conception and design of the study: EAS and SS. Acquisition and analysis of data: JMM, MTJ, RRP, ES, AG, CD, AV. Manuscript drafting and final approval: all authors. This manuscript has not been published somewhere else and is not under consideration for another journal. We adhered to the ethical guidelines and obtained IRB approval. Any author has conflict of interests to declare. We included the STROBE checklist. We did not use any source of funding for this research.\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\n\u003cli\u003eHarrod CG, Bendok BR, Batjer HH. Prediction of Cerebral Vasospasm in Patients Presenting with Aneurysmal Subarachnoid Hemorrhage: A Review. \u003cem\u003eNeurosurgery\u003c/em\u003e. 2005;56. \u003c/li\u003e\n\u003cli\u003eFisher CM, Kistler JP, Davis JM. Relation of cerebral vasospasm to subarachnoid hemorrhage visualized by computerized tomographic scanning. \u003cem\u003eNeurosurgery\u003c/em\u003e. 1980;6:1-9. \u003c/li\u003e\n\u003cli\u003eClaassen J, Bernardini GL, Kreiter K, Bates J, Du YE, Copeland D, Connolly ES, Mayer SA. Effect of cisternal and ventricular blood on risk of delayed cerebral ischemia after subarachnoid hemorrhage: the Fisher scale revisited. \u003cem\u003eStroke\u003c/em\u003e. 2001;32:2012-2020. \u003c/li\u003e\n\u003cli\u003eKistler JP, Crowell RM, Davis KR, Heros R, Ojemann RG, Zervas T, Fisher CM. The relation of cerebral vasospasm to the extent and location of subarachnoid blood visualized by CT scan: a prospective study. \u003cem\u003eNeurology\u003c/em\u003e. 1983;33:424-436. doi: 10.1212/wnl.33.4.424\u003c/li\u003e\n\u003cli\u003eMelinosky C, Kincaid H, Claassen J, Parikh G, Badjatia N, Morris NA. The Modified Fisher Scale Lacks Interrater Reliability. \u003cem\u003eNeurocrit Care\u003c/em\u003e. 2021;35:72-78. doi: 10.1007/s12028-020-01142-8\u003c/li\u003e\n\u003cli\u003eHunt WE, Hess RM. Surgical risk as related to time of intervention in the repair of intracranial aneurysms. \u003cem\u003eJ Neurosurg\u003c/em\u003e. 1968;28:14-20. doi: 10.3171/jns.1968.28.1.0014\u003c/li\u003e\n\u003cli\u003eOgungbo B. The World Federation of Neurological Surgeons scale for subarachnoid haemorrhage. \u003cem\u003eSurgical Neurology\u003c/em\u003e. 2003;59:236-237. doi: https://doi.org/10.1016/S0090-3019(02)01049-2\u003c/li\u003e\n\u003cli\u003eGarc\u0026iacute;a S, Torn\u0026eacute; R, Hoyos JA, Rodr\u0026iacute;guez-Hern\u0026aacute;ndez A, Amaro S, Llull L, L\u0026oacute;pez-Rueda A, Ense\u0026ntilde;at J. Quantitative versus qualitative blood amount assessment as a predictor for shunt-dependent hydrocephalus following aneurysmal subarachnoid hemorrhage. \u003cem\u003eJ Neurosurg\u003c/em\u003e. 2018;131:1743-1750. doi: 10.3171/2018.7.Jns18816\u003c/li\u003e\n\u003cli\u003eFriedman JA, Goerss SJ, Meyer FB, Piepgras DG, Pichelmann MA, McIver JI, Toussaint LG, 3rd, McClelland RL, Nichols DA, Atkinson JL, et al. Volumetric quantification of Fisher Grade 3 aneurysmal subarachnoid hemorrhage: a novel method to predict symptomatic vasospasm on admission computerized tomography scans. \u003cem\u003eJ Neurosurg\u003c/em\u003e. 2002;97:401-407. doi: 10.3171/jns.2002.97.2.0401\u003c/li\u003e\n\u003cli\u003eZijlstra IA, Gathier CS, Boers AM, Marquering HA, Slooter AJ, Velthuis BK, Coert BA, Verbaan D, van den Berg R, Rinkel GJ, et al. Association of Automatically Quantified Total Blood Volume after Aneurysmal Subarachnoid Hemorrhage with Delayed Cerebral Ischemia. \u003cem\u003eAJNR Am J Neuroradiol\u003c/em\u003e. 2016;37:1588-1593. doi: 10.3174/ajnr.A4771\u003c/li\u003e\n\u003cli\u003eScherer M, Cordes J, Younsi A, Sahin YA, G\u0026ouml;tz M, M\u0026ouml;hlenbruch M, Stock C, B\u0026ouml;sel J, Unterberg A, Maier-Hein K, et al. Development and Validation of an Automatic Segmentation Algorithm for Quantification of Intracerebral Hemorrhage. \u003cem\u003eStroke\u003c/em\u003e. 2016;47:2776-2782. doi: 10.1161/strokeaha.116.013779\u003c/li\u003e\n\u003cli\u003eRenan Sales B, Wessel E, Anna MMB, Zijlstra IJ, Rene, Wassim, Alexandre U, Dagmar V, Peter V, Charles M, et al. Automated segmentation of subarachnoid hemorrhages with convolutional neural networks. \u003cem\u003eInformatics in Medicine Unlocked\u003c/em\u003e. 2020;19:100321. doi: https://doi.org/10.1016/j.imu.2020.100321\u003c/li\u003e\n\u003cli\u003eXu J, Zhang R, Zhou Z, Wu C, Gong Q, Zhang H, Wu S, Wu G, Deng Y, Xia C, et al. Deep Network for the Automatic Segmentation and Quantification of Intracranial Hemorrhage on CT. \u003cem\u003eFront Neurosci\u003c/em\u003e. 2020;14:541817. doi: 10.3389/fnins.2020.541817\u003c/li\u003e\n\u003cli\u003eSanchez S, Hickerson M, Patel RR, Ghazaleh D, Tarchand R, Paranjape GS, Pope H, Ortega‐Gutierrez S, Pederson JM, Hasan D, et al. Morphological Characteristics of Ruptured Brain Aneurysms: A Systematic Literature Review and Meta‐Analysis. \u003cem\u003eStroke: Vascular and Interventional Neurology\u003c/em\u003e.0:e000707. doi: doi:10.1161/SVIN.122.000707\u003c/li\u003e\n\u003cli\u003eZou KH, Warfield SK, Bharatha A, Tempany CM, Kaus MR, Haker SJ, Wells WM, 3rd, Jolesz FA, Kikinis R. Statistical validation of image segmentation quality based on a spatial overlap index. \u003cem\u003eAcad Radiol\u003c/em\u003e. 2004;11:178-189. doi: 10.1016/s1076-6332(03)00671-8\u003c/li\u003e\n\u003cli\u003eZanaty M, Allan L, Samaniego EA, Piscopo A, Ryan E, Torner JC, Hasan D. Phase 1/2a Trial of ISPASM. \u003cem\u003eStroke\u003c/em\u003e. 2021;52:3750-3758. doi: doi:10.1161/STROKEAHA.121.034578\u003c/li\u003e\n\u003cli\u003eDani\u0026egrave;re F, Gascou G, Menjot de Champfleur N, Machi P, Leboucq N, Riquelme C, Ruiz C, Bonaf\u0026eacute; A, Costalat V. Complications and follow up of subarachnoid hemorrhages. \u003cem\u003eDiagnostic and Interventional Imaging\u003c/em\u003e. 2015;96:677-686. doi: https://doi.org/10.1016/j.diii.2015.05.006\u003c/li\u003e\n\u003cli\u003evan der Jagt M, Hasan D, Bijvoet HW, Pieterman H, Koudstaal PJ, Avezaat CJ. Interobserver variability of cisternal blood on CT after aneurysmal subarachnoid hemorrhage. \u003cem\u003eNeurology\u003c/em\u003e. 2000;54:2156-2158. doi: 10.1212/wnl.54.11.2156\u003c/li\u003e\n\u003cli\u003evan Norden AG, van Dijk GW, van Huizen MD, Algra A, Rinkel GJ. Interobserver agreement and predictive value for outcome of two rating scales for the amount of extravasated blood after aneurysmal subarachnoid haemorrhage. \u003cem\u003eJ Neurol\u003c/em\u003e. 2006;253:1217-1220. doi: 10.1007/s00415-006-0205-0\u003c/li\u003e\n\u003cli\u003eMatsoukas S, Scaggiante J, Schuldt BR, Smith CJ, Chennareddy S, Kalagara R, Majidi S, Bederson JB, Fifi JT, Mocco J, et al. Accuracy of artificial intelligence for the detection of intracranial hemorrhage and chronic cerebral microbleeds: a systematic review and pooled analysis. \u003cem\u003eRadiol Med\u003c/em\u003e. 2022;127:1106-1123. doi: 10.1007/s11547-022-01530-4\u003c/li\u003e\n\u003cli\u003ePlatz J, G\u0026uuml;resir E, Wagner M, Seifert V, Konczalla J. Increased risk of delayed cerebral ischemia in subarachnoid hemorrhage patients with additional intracerebral hematoma. \u003cem\u003eJ Neurosurg\u003c/em\u003e. 2017;126:504-510. doi: 10.3171/2015.12.Jns151563\u003c/li\u003e\n\u003cli\u003eLagares A, Jim\u0026eacute;nez-Rold\u0026aacute;n L, Gomez PA, Munarriz PM, Casta\u0026ntilde;o-Le\u0026oacute;n AM, Cepeda S, Al\u0026eacute;n JF. Prognostic Value of the Amount of Bleeding After Aneurysmal Subarachnoid Hemorrhage: A Quantitative Volumetric Study. \u003cem\u003eNeurosurgery\u003c/em\u003e. 2015;77:898-907; discussion 907. doi: 10.1227/neu.0000000000000927\u003c/li\u003e\n\u003cli\u003eStreet JS, Pandit AS, Toma AK. Predicting vasospasm risk using first presentation aneurysmal subarachnoid hemorrhage volume: A semi-automated CT image segmentation analysis using ITK-SNAP. \u003cem\u003ePLoS One\u003c/em\u003e. 2023;18:e0286485. doi: 10.1371/journal.pone.0286485\u003c/li\u003e\n\u003cli\u003eHaedo MG, Grille P, Burghi G, Barbato M. Correlation between tomographic scales and vasospasm and delayed cerebral ischemia in aneurysmal subarachnoid hemorrhage. \u003cem\u003eCrit Care Sci\u003c/em\u003e. 2023;35:311-319. doi: 10.5935/2965-2774.20230119-en\u003c/li\u003e\n\u003cli\u003eWilson CD, Safavi-Abbasi S, Sun H, Kalani MY, Zhao YD, Levitt MR, Hanel RA, Sauvageau E, Mapstone TB, Albuquerque FC, et al. Meta-analysis and systematic review of risk factors for shunt dependency after aneurysmal subarachnoid hemorrhage. \u003cem\u003eJ Neurosurg\u003c/em\u003e. 2017;126:586-595. doi: 10.3171/2015.11.Jns152094\u003c/li\u003e\n\u003cli\u003eWiebers DO, Whisnant JP, Sundt TM, Jr., O\u0026apos;Fallon WM. The significance of unruptured intracranial saccular aneurysms. \u003cem\u003eJ Neurosurg\u003c/em\u003e. 1987;66:23-29. doi: 10.3171/jns.1987.66.1.0023\u003c/li\u003e\n\u003cli\u003eKroon D-J. Region Growing. MATLAB Central File Exchange. https://www.mathworks.com/matlabcentral/fileexchange/19084-region-growing. 2022. Accessed June 14, 2022.\u003c/li\u003e\n\u003c/ol\u003e"}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":false,"hideJournal":false,"highlight":"","institution":"","isAcceptedByJournal":true,"isAuthorSuppliedPdf":false,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":true,"isPdf":false,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"neurocritical-care","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"neca","sideBox":"Learn more about [Neurocritical Care](http://link.springer.com/journal/12028)","snPcode":"12028","submissionUrl":"https://www.editorialmanager.com/neca/default2.aspx","title":"Neurocritical Care","twitterHandle":"","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"em","reportingPortfolio":"Springer Hybrid","inReviewEnabled":true,"inReviewRevisionsEnabled":false},"keywords":"","lastPublishedDoi":"10.21203/rs.3.rs-4308305/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-4308305/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003ch2\u003eBackground\u003c/h2\u003e \u003cp\u003eThe volume of hemorrhage is a critical factor in predicting outcomes following aneurysmal subarachnoid hemorrhage (aSAH). Although grading scales such as the Fisher score are extensively used, their subjective nature can lead to inaccuracies in quantifying the total volume of blood. We analyzed a large cohort of patients with aSAH with an automated software for the precise quantification of hemorrhage volume. The primary aim is to identify clear thresholds that correlate with the likelihood of complications post-aSAH, thereby enhancing the predictive accuracy and improving patient management strategies.\u003c/p\u003e\u003ch2\u003eMethods\u003c/h2\u003e \u003cp\u003eAn automated algorithm was developed to analyze non-contrast computed tomography scans of aSAH patients. The algorithm categorized tissues into blood, gray matter, white matter, and cerebrospinal fluid, isolating the blood for volume quantification. Receiver operating curve analysis was done to establish thresholds for vasospasm, acute hydrocephalus, shunt-dependent hydrocephalus (SDH), and death within 7 days. Additionally, we determined if there is any relationship between the aneurysm size and the amount of hemorrhage.\u003c/p\u003e\u003ch2\u003eResults\u003c/h2\u003e \u003cp\u003eA total of 500 aSAH patients and their respective aneurysms were analyzed. Hemorrhage volume was significantly higher in patients with vasospasm (21.7 [10.9, 41.4] vs 10.7 [4.2, 26.9], p\u0026thinsp;\u0026lt;\u0026thinsp;0.001), acute hydrocephalus (22.7 [9.2, 41.8] vs 5.1 [2.1, 13.5], p\u0026thinsp;\u0026lt;\u0026thinsp;0.001), SDH (23.8 [11.3, 40.7] vs 11.7 [4.1, 28.2], p\u0026thinsp;\u0026lt;\u0026thinsp;0.001), and those who died before 7 days (52.8 [34.6, 90.6] mL vs 14.8 [5.0, 32.4] mL, p\u0026thinsp;\u0026lt;\u0026thinsp;0.001) compared to their counterparts. Notably, specific hemorrhage thresholds were identified for each complication: 15.16 mL for vasospasm (65% sensitivity and 60% specificity), 9.95 mL for acute hydrocephalus (74% sensitivity and 69% specificity), 16.76 mL for SDH (63% sensitivity and 60% specificity), and 33.84 mL for death within 7 days (79% sensitivity and 77% specificity).\u003c/p\u003e\u003ch2\u003eConclusion\u003c/h2\u003e \u003cp\u003eAutomated blood volume quantification tools could aid in stratifying complication risk after aSAH. Established thresholds for hemorrhage volume related to complications could be used in clinical practice to aid in management decisions.\u003c/p\u003e","manuscriptTitle":"Automated Hemorrhage Volume Quantification in Aneurysmal Subarachnoid Hemorrhage","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2024-05-07 18:48:42","doi":"10.21203/rs.3.rs-4308305/v1","editorialEvents":[{"type":"communityComments","content":0},{"type":"reviewerAgreed","content":"","date":"2024-04-29T12:38:28+00:00","index":0,"fulltext":""},{"type":"reviewersInvited","content":"","date":"2024-04-29T12:33:58+00:00","index":"","fulltext":""},{"type":"editorInvited","content":"Neurocritical Care","date":"2024-04-26T15:39:41+00:00","index":"","fulltext":""},{"type":"editorAssigned","content":"","date":"2024-04-24T17:46:03+00:00","index":"","fulltext":""},{"type":"submitted","content":"Neurocritical Care","date":"2024-04-24T07:08:07+00:00","index":"","fulltext":""}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"neurocritical-care","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"neca","sideBox":"Learn more about [Neurocritical Care](http://link.springer.com/journal/12028)","snPcode":"12028","submissionUrl":"https://www.editorialmanager.com/neca/default2.aspx","title":"Neurocritical Care","twitterHandle":"","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"em","reportingPortfolio":"Springer Hybrid","inReviewEnabled":true,"inReviewRevisionsEnabled":false}}],"origin":"","ownerIdentity":"70c845e5-c462-402c-9939-b2b22ef35268","owner":[],"postedDate":"May 7th, 2024","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"published-in-journal","subjectAreas":[],"tags":[],"updatedAt":"2024-09-30T16:01:45+00:00","versionOfRecord":{"articleIdentity":"rs-4308305","link":"https://doi.org/10.1007/s12028-024-02123-x","journal":{"identity":"neurocritical-care","isVorOnly":false,"title":"Neurocritical Care"},"publishedOn":"2024-09-25 15:57:20","publishedOnDateReadable":"September 25th, 2024"},"versionCreatedAt":"2024-05-07 18:48:42","video":"","vorDoi":"10.1007/s12028-024-02123-x","vorDoiUrl":"https://doi.org/10.1007/s12028-024-02123-x","workflowStages":[]},"version":"v1","identity":"rs-4308305","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-4308305","identity":"rs-4308305","version":["v1"]},"buildId":"qtupq5eGEP_6zYnWcrvyt","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}