학술논문
Clustering identifies endotypes of traumatic brain injury in an intensive care cohort: a CENTER-TBI study
Document Type
Report
Author
Ãkerlund, Cecilia A. I.; Holst, Anders; Stocchetti, Nino; Steyerberg, Ewout W.; Menon, David K.; Ercole, Ari; Nelson, David W.; Ãkerlund, Cecilia; Amrein, Krisztina; Andelic, Nada; Andreassen, Lasse; Anke, Audny; Antoni, Anna; Audibert, Gérard; Azouvi, Philippe; Azzolini, Maria Luisa; Bartels, Ronald; Barzó, Pál; Beauvais, Romuald; Beer, Ronny; Bellander, Bo-Michael; Belli, Antonio; Benali, Habib; Berardino, Maurizio; Beretta, Luigi; Blaabjerg, Morten; Bragge, Peter; Brazinova, Alexandra; Brinck, Vibeke; Brooker, Joanne; Brorsson, Camilla; Buki, Andras; Bullinger, Monika; Cabeleira, Manuel; Caccioppola, Alessio; Calappi, Emiliana; Calvi, Maria Rosa; Cameron, Peter; Lozano, Guillermo Carbayo; Carbonara, Marco; Cavallo, Simona; Chevallard, Giorgio; Chieregato, Arturo; Citerio, Giuseppe; Clusmann, Hans; Coburn, Mark; Coles, Jonathan; Cooper, Jamie D.; Correia, Marta; ÄoviÄ, Amra; Curry, Nicola; Czeiter, Endre; Czosnyka, Marek; DahyotFizelier, Claire; Dark, Paul; Dawes, Helen; De Keyser, Véronique; Degos, Vincent; Corte, Francesco Della; Boogert, Hugo den; Depreitere, Bart; Äilvesi, Äula; Dixit, Abhishek; Donoghue, Emma; Dreier, Jens; Dulière, GuyLoup; Esser, Patrick; Ezer, Erzsébet; Fabricius, Martin; Feigin, Valery L.; Foks, Kelly; Frisvold, Shirin; Furmanov, Alex; Gagliardo, Pablo; Galanaud, Damien; Gantner, Dashiell; Gao, Guoyi; George, Pradeep; Ghuysen, Alexandre; Giga, Lelde; Glocker, Ben; Golubovic, JagoÅ¡; Gomez, Pedro A.; Gratz, Johannes; Gravesteijn, Benjamin; Grossi, Francesca; Gruen, Russell L.; Gupta, Deepak; Haagsma, Juanita A.; Haitsma, Iain; Helbok, Raimund; Helseth, Eirik; Horton, Lindsay; Huijben, Jilske; Hutchinson, Peter J.; Jacobs, Bram; Jankowski, Stefan; Jarrett, Mike; Jiang, Jiyao; Johnson, Faye; Jones, Kelly; Karan, Mladen; Kolias, Angelos G.; Kompanje, Erwin; Kondziella, Daniel; Kornaropoulos, Evgenios; Koskinen, LarsOwe; Kovács, Noémi; Kowark, Ana; Lagares, Alfonso; Lanyon, Linda; Laureys, Steven; Lecky, Fiona; Ledoux, Didier; Lefering, Rolf; Legrand, Valerie; Lejeune, Aurelie; Levi, Leon; Lightfoot, Roger; Lingsma, Hester; Maas, Andrew I. R.; CastaéoLeón, Ana M.; Maegele, Marc; Majdan, Marek; Manara, Alex; Manley, Geoffrey; Martino, Costanza; Maréchal, Hugues; Mattern, Julia; McMahon, Catherine; Melegh, Béla; Menon, David; Menovsky, Tomas; Mikolic, Ana; Misset, Benoit; Muraleedharan, Visakh; Murray, Lynnette; Negru, Ancuta; Nelson, David; Newcombe, Virginia; Nieboer, Daan; Nyirádi, József; Olubukola, Otesile; Oresic, Matej; Ortolano, Fabrizio; Palotie, Aarno; Parizel, Paul M.; Payen, JeanFrançois; Perera, Natascha; Perlbarg, Vincent; Persona, Paolo; Peul, Wilco; Piippo-Karjalainen, Anna; Pirinen, Matti; Pisica, Dana; Ples, Horia; Polinder, Suzanne; Pomposo, Inigo; Posti, Jussi P.; Puybasset, Louis; Radoi, Andreea; Ragauskas, Arminas; Raj, Rahul; Rambadagalla, Malinka; Helmrich, Isabel Retel; Rhodes, Jonathan; Richardson, Sylvia; Richter, Sophie; Ripatti, Samuli; Rocka, Saulius; Roe, Cecilie; Roise, Olav; Rosand, Jonathan; Rosenfeld, Jeffrey V.; Rosenlund, Christina; Rosenthal, Guy; Rossaint, Rolf; Rossi, Sandra; Rueckert, Daniel; Rusnák, Martin; Sahuquillo, Juan; Sakowitz, Oliver; SanchezPorras, Renan; Sandor, Janos; Schäfer, Nadine; Schmidt, Silke; Schoechl, Herbert; Schoonman, Guus; Schou, Rico Frederik; Schwendenwein, Elisabeth; Sewalt, Charlie; Singh, Ranjit D.; Skandsen, Toril; Smielewski, Peter; Sorinola, Abayomi; Stamatakis, Emmanuel; Stanworth, Simon; Stevens, Robert; Stewart, William; Sundström, Nina; Takala, Riikka; Tamás, Viktória; Tamosuitis, Tomas; Taylor, Mark Steven; Ao, Braden Te; Tenovuo, Olli; Theadom, Alice; Thomas, Matt; Tibboel, Dick; Timmers, Marjolein; Tolias, Christos; Trapani, Tony; Tudora, Cristina Maria; Unterberg, Andreas; Vajkoczy, Peter; Vallance, Shirley; Valeinis, Egils; Vámos, Zoltán; van der Jagt, Mathieu; Van der Steen, Gregory; van der Naalt, Joukje; van Dijck, Jeroen T. J. M.; van Erp, Inge A.; van Essen, Thomas A.; Van Hecke, Wim; van Heugten, Caroline; Van Praag, Dominique; van Veen, Ernest; Vyvere, Thijs Vande; van Wijk, Roel P. J.; Vargiolu, Alessia; Vega, Emmanuel; Velt, Kimberley; Verheyden, Jan; Vespa, Paul M.; Vik, Anne; Vilcinis, Rimantas; Volovici, Victor; von Steinbüchel, Nicole; Voormolen, Daphne; Vulekovic, Petar; Wang, Kevin K. W.; Whitehouse, Daniel; Wiegers, Eveline; Williams, Guy; Wilson, Lindsay; Winzeck, Stefan; Wolf, Stefan; Yang, Zhihui; Ylén, Peter; Younsi, Alexander; Zeiler, Frederick A.; Zelinkova, Veronika; Ziverte, Agate; Zoerle, Tommaso
Source
Critical Care. July 27, 2022, Vol. 26 Issue 1
Subject
Language
English
ISSN
1364-8535
Abstract
Author(s): Cecilia A. I. Ãkerlund[sup.1,2] , Anders Holst[sup.2] , Nino Stocchetti[sup.3] , Ewout W. Steyerberg[sup.4] , David K. Menon[sup.5] , Ari Ercole[sup.5,6] , David W. Nelson[sup.1] , Cecilia Ãkerlund[sup.] , [...]
Background While the Glasgow coma scale (GCS) is one of the strongest outcome predictors, the current classification of traumatic brain injury (TBI) as 'mild', 'moderate' or 'severe' based on this fails to capture enormous heterogeneity in pathophysiology and treatment response. We hypothesized that data-driven characterization of TBI could identify distinct endotypes and give mechanistic insights. Methods We developed an unsupervised statistical clustering model based on a mixture of probabilistic graphs for presentation (< 24 h) demographic, clinical, physiological, laboratory and imaging data to identify subgroups of TBI patients admitted to the intensive care unit in the CENTER-TBI dataset (N = 1,728). A cluster similarity index was used for robust determination of optimal cluster number. Mutual information was used to quantify feature importance and for cluster interpretation. Results Six stable endotypes were identified with distinct GCS and composite systemic metabolic stress profiles, distinguished by GCS, blood lactate, oxygen saturation, serum creatinine, glucose, base excess, pH, arterial partial pressure of carbon dioxide, and body temperature. Notably, a cluster with 'moderate' TBI (by traditional classification) and deranged metabolic profile, had a worse outcome than a cluster with 'severe' GCS and a normal metabolic profile. Addition of cluster labels significantly improved the prognostic precision of the IMPACT (International Mission for Prognosis and Analysis of Clinical trials in TBI) extended model, for prediction of both unfavourable outcome and mortality (both p < 0.001). Conclusions Six stable and clinically distinct TBI endotypes were identified by probabilistic unsupervised clustering. In addition to presenting neurology, a profile of biochemical derangement was found to be an important distinguishing feature that was both biologically plausible and associated with outcome. Our work motivates refining current TBI classifications with factors describing metabolic stress. Such data-driven clusters suggest TBI endotypes that merit investigation to identify bespoke treatment strategies to improve care. Trial registration The core study was registered with ClinicalTrials.gov, number NCT02210221, registered on August 06, 2014, with Resource Identification Portal (RRID: SCR_015582). Keywords: Traumatic brain injury, Endotypes, Intensive care unit, Critical care, Unsupervised clustering, Machine learning
Background While the Glasgow coma scale (GCS) is one of the strongest outcome predictors, the current classification of traumatic brain injury (TBI) as 'mild', 'moderate' or 'severe' based on this fails to capture enormous heterogeneity in pathophysiology and treatment response. We hypothesized that data-driven characterization of TBI could identify distinct endotypes and give mechanistic insights. Methods We developed an unsupervised statistical clustering model based on a mixture of probabilistic graphs for presentation (< 24 h) demographic, clinical, physiological, laboratory and imaging data to identify subgroups of TBI patients admitted to the intensive care unit in the CENTER-TBI dataset (N = 1,728). A cluster similarity index was used for robust determination of optimal cluster number. Mutual information was used to quantify feature importance and for cluster interpretation. Results Six stable endotypes were identified with distinct GCS and composite systemic metabolic stress profiles, distinguished by GCS, blood lactate, oxygen saturation, serum creatinine, glucose, base excess, pH, arterial partial pressure of carbon dioxide, and body temperature. Notably, a cluster with 'moderate' TBI (by traditional classification) and deranged metabolic profile, had a worse outcome than a cluster with 'severe' GCS and a normal metabolic profile. Addition of cluster labels significantly improved the prognostic precision of the IMPACT (International Mission for Prognosis and Analysis of Clinical trials in TBI) extended model, for prediction of both unfavourable outcome and mortality (both p < 0.001). Conclusions Six stable and clinically distinct TBI endotypes were identified by probabilistic unsupervised clustering. In addition to presenting neurology, a profile of biochemical derangement was found to be an important distinguishing feature that was both biologically plausible and associated with outcome. Our work motivates refining current TBI classifications with factors describing metabolic stress. Such data-driven clusters suggest TBI endotypes that merit investigation to identify bespoke treatment strategies to improve care. Trial registration The core study was registered with ClinicalTrials.gov, number NCT02210221, registered on August 06, 2014, with Resource Identification Portal (RRID: SCR_015582). Keywords: Traumatic brain injury, Endotypes, Intensive care unit, Critical care, Unsupervised clustering, Machine learning