Christian Rubbert1, Kaustubh R Patil2,3, Kerim Beseoglu4, Christian Mathys5,6, Rebecca May5, Marius G Kaschner5, Benjamin Sigl5, Nikolas A Teichert5, Johannes Boos5, Bernd Turowski5, Julian Caspers5,7. 1. University Düsseldorf, Medical Faculty, Department of Diagnostic and Interventional Radiology, Moorenstr. 5, D-40225, Düsseldorf, Germany. christian.rubbert@med.uni-duesseldorf.de. 2. Institute of Neuroscience and Medicine, Brain & Behaviour (INM-7), Research Centre Jülich, D-52425, Jülich, Germany. 3. Institute of Systems Neuroscience, Medical Faculty, Heinrich Heine University Düsseldorf, D-40225, Düsseldorf, Germany. 4. Department of Neurosurgery, Medical Faculty, Heinrich-Heine-University, D-40225, Düsseldorf, Germany. 5. University Düsseldorf, Medical Faculty, Department of Diagnostic and Interventional Radiology, Moorenstr. 5, D-40225, Düsseldorf, Germany. 6. Institute of Radiology and Neuroradiology, Evangelisches Krankenhaus, University of Oldenburg, D-26122, Oldenburg, Germany. 7. Institute of Neuroscience and Medicine (INM-1), Research Centre Jülich, D-52425, Jülich, Germany.
Abstract
OBJECTIVES: The pathogenesis leading to poor functional outcome after aneurysmal subarachnoid haemorrhage (aSAH) is multifactorial and not fully understood. We evaluated a machine learning approach based on easily determinable clinical and CT perfusion (CTP) features in the course of patient admission to predict the functional outcome 6 months after ictus. METHODS: Out of 630 consecutive subarachnoid haemorrhage patients (2008-2015), 147 (mean age 54.3, 66.7% women) were retrospectively included (Inclusion: aSAH, admission within 24 h of ictus, CTP within 24 h of admission, documented modified Rankin scale (mRS) grades after 6 months. Exclusion: occlusive therapy before first CTP, previous aSAH, CTP not evaluable). A random forests model with conditional inference trees was optimised and trained on sex, age, World Federation of Neurosurgical Societies (WFNS) and modified Fisher grades, aneurysm in anterior vs. posterior circulation, early external ventricular drainage (EVD), as well as MTT and Tmax maximum, mean, standard deviation (SD), range, 75th quartile and interquartile range to predict dichotomised mRS (≤ 2; > 2). Performance was assessed using the balanced accuracy over the training and validation folds using 20 repeats of 10-fold cross-validation. RESULTS: In the final model, using 200 trees and the synthetic minority oversampling technique, median balanced accuracy was 84.4% (SD 0.7) over the training folds and 70.9% (SD 1.2) over the validation folds. The five most important features were the modified Fisher grade, age, MTT range, WFNS and early EVD. CONCLUSIONS: A random forests model trained on easily determinable features in the course of patient admission can predict the functional outcome 6 months after aSAH with considerable accuracy. KEY POINTS: • Features determinable in the course of admission of a patient with aneurysmal subarachnoid haemorrhage (aSAH) can predict the functional outcome 6 months after the occurrence of aSAH. • The top five predictive features were the modified Fisher grade, age, the mean transit time (MTT) range from computed tomography perfusion (CTP), the WFNS grade and the early necessity for an external ventricular drainage (EVD). • The range between the minimum and the maximum MTT may prove to be a valuable biomarker for detrimental functional outcome.
OBJECTIVES: The pathogenesis leading to poor functional outcome after aneurysmal subarachnoid haemorrhage (aSAH) is multifactorial and not fully understood. We evaluated a machine learning approach based on easily determinable clinical and CT perfusion (CTP) features in the course of patient admission to predict the functional outcome 6 months after ictus. METHODS: Out of 630 consecutive subarachnoid haemorrhagepatients (2008-2015), 147 (mean age 54.3, 66.7% women) were retrospectively included (Inclusion: aSAH, admission within 24 h of ictus, CTP within 24 h of admission, documented modified Rankin scale (mRS) grades after 6 months. Exclusion: occlusive therapy before first CTP, previous aSAH, CTP not evaluable). A random forests model with conditional inference trees was optimised and trained on sex, age, World Federation of Neurosurgical Societies (WFNS) and modified Fisher grades, aneurysm in anterior vs. posterior circulation, early external ventricular drainage (EVD), as well as MTT and Tmax maximum, mean, standard deviation (SD), range, 75th quartile and interquartile range to predict dichotomised mRS (≤ 2; > 2). Performance was assessed using the balanced accuracy over the training and validation folds using 20 repeats of 10-fold cross-validation. RESULTS: In the final model, using 200 trees and the synthetic minority oversampling technique, median balanced accuracy was 84.4% (SD 0.7) over the training folds and 70.9% (SD 1.2) over the validation folds. The five most important features were the modified Fisher grade, age, MTT range, WFNS and early EVD. CONCLUSIONS: A random forests model trained on easily determinable features in the course of patient admission can predict the functional outcome 6 months after aSAH with considerable accuracy. KEY POINTS: • Features determinable in the course of admission of a patient with aneurysmal subarachnoid haemorrhage (aSAH) can predict the functional outcome 6 months after the occurrence of aSAH. • The top five predictive features were the modified Fisher grade, age, the mean transit time (MTT) range from computed tomography perfusion (CTP), the WFNS grade and the early necessity for an external ventricular drainage (EVD). • The range between the minimum and the maximum MTT may prove to be a valuable biomarker for detrimental functional outcome.
Authors: Ahmed E Othman; Saif Afat; Marc A Brockmann; Omid Nikoubashman; Carolin Brockmann; Konstantin Nikolaou; Martin Wiesmann Journal: J Neuroradiol Date: 2015-12-10 Impact factor: 3.447
Authors: Digna R Velez; Bill C White; Alison A Motsinger; William S Bush; Marylyn D Ritchie; Scott M Williams; Jason H Moore Journal: Genet Epidemiol Date: 2007-05 Impact factor: 2.135
Authors: H J Wittsack; A M Wohlschläger; E K Ritzl; R Kleiser; M Cohnen; R J Seitz; U Mödder Journal: Comput Med Imaging Graph Date: 2008-01 Impact factor: 4.790
Authors: K Takagi; A Tamura; T Nakagomi; H Nakayama; O Gotoh; K Kawai; M Taneda; N Yasui; H Hadeishi; K Sano Journal: J Neurosurg Date: 1999-04 Impact factor: 5.115
Authors: Charlotte H P Cremers; Pieter C Vos; Irene C van der Schaaf; Birgitta K Velthuis; Mervyn D I Vergouwen; Gabriel J E Rinkel; Jan Willem Dankbaar Journal: Neuroradiology Date: 2015-06-02 Impact factor: 2.804
Authors: Z Shi; B Hu; U J Schoepf; R H Savage; D M Dargis; C W Pan; X L Li; Q Q Ni; G M Lu; L J Zhang Journal: AJNR Am J Neuroradiol Date: 2020-03-12 Impact factor: 3.825
Authors: Axel Meineke; Christian Rubbert; Lino M Sawicki; Christoph Thomas; Yan Klosterkemper; Elisabeth Appel; Julian Caspers; Oliver T Bethge; Patric Kröpil; Gerald Antoch; Johannes Boos Journal: Eur Radiol Date: 2019-02-19 Impact factor: 5.315
Authors: Vittorio Stumpo; Victor E Staartjes; Giuseppe Esposito; Carlo Serra; Luca Regli; Alessandro Olivi; Carmelo Lucio Sturiale Journal: Acta Neurochir Suppl Date: 2022