Álex Ferré1,2,3, María A Poca3,4, María Dolore de la Calzada3, Dulce Moncho1,3, Aintzane Urbizu5, Odile Romero1,2,6, Gabriel Sampol2,6,7, Juan Sahuquillo3,4. 1. Clinical Neurophysiology Department, Vall d'Hebron University Hospital, Universitat Autónoma de Barcelona, Barcelona, Spain. 2. Multidisciplinary Sleep Unit, Vall d'Hebron University Hospital, Universitat Autónoma de Barcelona, Barcelona, Spain. 3. Neurotraumatology and Neurosurgery Research Unit (UNINN), Vall d'Hebron University Hospital, Barcelona, Spain. 4. Neurosurgery Department, Vall d'Hebron University Hospital, Universitat Autónoma de Barcelona, Barcelona, Spain. 5. Conquer Chiari Research Center, Department of Mechanical Engineering, The University of Akron, Ohio. 6. Centro de Investigación Biomédica en Red Enfermedades Respiratorias (CIBERES), Instituto de la Salud Carlos III (ISCIIII), Madrid, Spain. 7. Pneumology Department, Vall d'Hebron Research institute, Vall d'Hebron University Hospital, Universitat Autónoma de Barcelona, Barcelona, Spain.
Abstract
STUDY OBJECTIVES: The aim of this study is to generate and validate supervised machine learning algorithms to detect patients with Chiari malformation (CM) 1 or 1.5 at high risk of the development of sleep-related breathing disorders (SRBD) using clinical and neuroradiological parameters. METHODS: We prospectively included two independent datasets. A training dataset (n = 90) was used to obtain the best model, whereas a second dataset was used to validate it (n = 74). In both cohorts, the same clinical, neuroradiological, and sleep studies were carried out. We used two supervised machine learning approaches, multiple logistic regression (MLR) and the unbiased recursive partitioning technique conditional inference tree (URP-CTREE), to detect patients at high risk of SRBD. We then compared the accuracy, sensitivity, and specificity of the two prediction models. RESULTS: Age (odds ratio [OR] 1.1 95% confidence interval [CI] 1.05-1.17), sex (OR 0.19 95% CI 0.05-0.67), CM type (OR 4.36 95% CI 1.14-18.5), and clivus length (OR 1.14 95% CI 1.01-1.31) were the significant predictor variables for a respiratory disturbance index (RDI) cutoff that was ≥ 10 events/h using MLR. The URP-CTREE model predicted that patients with CM-1 who were age 52 years or older and males with CM-1 who were older than 29 years had a high risk of SRBD. The accuracy of predicting patients with an RDI ≥ 10 events/h was similar in the two cohorts but in the URP-CTREE model, specificity was significantly greater when compared to MLR in both study groups. CONCLUSIONS: Both MLR and URP-CTREE predictive models are useful for the diagnosis of SRBD in patients with CM. However, URP-CTREE is easier to apply and interpret in clinical practice.
STUDY OBJECTIVES: The aim of this study is to generate and validate supervised machine learning algorithms to detect patients with Chiari malformation (CM) 1 or 1.5 at high risk of the development of sleep-related breathing disorders (SRBD) using clinical and neuroradiological parameters. METHODS: We prospectively included two independent datasets. A training dataset (n = 90) was used to obtain the best model, whereas a second dataset was used to validate it (n = 74). In both cohorts, the same clinical, neuroradiological, and sleep studies were carried out. We used two supervised machine learning approaches, multiple logistic regression (MLR) and the unbiased recursive partitioning technique conditional inference tree (URP-CTREE), to detect patients at high risk of SRBD. We then compared the accuracy, sensitivity, and specificity of the two prediction models. RESULTS: Age (odds ratio [OR] 1.1 95% confidence interval [CI] 1.05-1.17), sex (OR 0.19 95% CI 0.05-0.67), CM type (OR 4.36 95% CI 1.14-18.5), and clivus length (OR 1.14 95% CI 1.01-1.31) were the significant predictor variables for a respiratory disturbance index (RDI) cutoff that was ≥ 10 events/h using MLR. The URP-CTREE model predicted that patients with CM-1 who were age 52 years or older and males with CM-1 who were older than 29 years had a high risk of SRBD. The accuracy of predicting patients with an RDI ≥ 10 events/h was similar in the two cohorts but in the URP-CTREE model, specificity was significantly greater when compared to MLR in both study groups. CONCLUSIONS: Both MLR and URP-CTREE predictive models are useful for the diagnosis of SRBD in patients with CM. However, URP-CTREE is easier to apply and interpret in clinical practice.
Authors: Olga Mestres; María A Poca; Elizabeth Solana; Andreea Radoi; Manuel Quintana; Enriqueta Force; Juan Sahuquillo Journal: Rev Neurol Date: 2012-08-01 Impact factor: 0.870
Authors: Chamara V Senaratna; Jennifer L Perret; Caroline J Lodge; Adrian J Lowe; Brittany E Campbell; Melanie C Matheson; Garun S Hamilton; Shyamali C Dharmage Journal: Sleep Med Rev Date: 2016-07-18 Impact factor: 11.609
Authors: Nathaniel S Marshall; Keith K H Wong; Stewart R J Cullen; Matthew W Knuiman; Ronald R Grunstein Journal: J Clin Sleep Med Date: 2014-04-15 Impact factor: 4.062
Authors: Vera Yakovchenko; Edward J Miech; Matthew J Chinman; Maggie Chartier; Rachel Gonzalez; JoAnn E Kirchner; Timothy R Morgan; Angela Park; Byron J Powell; Enola K Proctor; David Ross; Thomas J Waltz; Shari S Rogal Journal: Med Care Date: 2020-05 Impact factor: 2.983
Authors: William H Shuman; Aislyn DiRisio; Alejandro Carrasquilla; Colin D Lamb; Addison Quinones; Aymeric Pionteck; Yang Yang; Mehmet Kurt; Raj K Shrivastava Journal: Neurosurg Rev Date: 2021-07-13 Impact factor: 3.042