Radiomics score predicts acute respiratory distress syndrome based on the initial CT scan after trauma.
Acute respiratory distress syndrome
Multidetector computed tomography
Polytrauma
Radiomics
Thoracic injuries
Journal
European radiology
ISSN: 1432-1084
Titre abrégé: Eur Radiol
Pays: Germany
ID NLM: 9114774
Informations de publication
Date de publication:
Aug 2021
Aug 2021
Historique:
received:
22
05
2020
accepted:
16
12
2020
revised:
02
12
2020
pubmed:
19
3
2021
medline:
14
7
2021
entrez:
18
3
2021
Statut:
ppublish
Résumé
Acute respiratory distress syndrome (ARDS) constitutes a major factor determining the clinical outcome in polytraumatized patients. Early prediction of ARDS is crucial for timely supportive therapy to reduce morbidity and mortality. The objective of this study was to develop and test a machine learning-based method for the early prediction of ARDS derived from the first computed tomography scan of polytraumatized patients after admission to the hospital. One hundred twenty-three patients (86 male and 37 female, age 41.2 ± 16.4) with an injury severity score (ISS) of 16 or higher (31.9 ± 10.9) were prospectively included and received a CT scan within 1 h after the accident. The lungs, including air pockets and pleural effusions, were automatically segmented using a deep learning-based algorithm. Subsequently, we extracted radiomics features from within the lung and trained an ensemble of gradient boosted trees (GBT) to predict future ARDS. Cross-validated ARDS prediction resulted in an area under the curve (AUC) of 0.79 for the radiomics score compared to 0.66 for ISS, and 0.68 for the abbreviated injury score of the thorax (AIS-thorax). Prediction using the radiomics score yielded an f1-score of 0.70 compared to 0.53 for ISS and 0.57 for AIS-thorax. The radiomics score achieved a sensitivity and specificity of 0.80 and 0.76. This study proposes a radiomics-based algorithm for the prediction of ARDS in polytraumatized patients at the time of admission to hospital with an accuracy that competes and surpasses conventional scores despite the heterogeneous, and therefore more realistic, scanning protocols. • Early prediction of acute respiratory distress syndrome in polytraumatized patients is possible, even when using heterogenous data. • Radiomics-based prediction resulted in an area under the curve of 0.79 compared to 0.66 for the injury severity score, and 0.68 for the abbreviated injury score of the thorax. • Highlighting the most relevant lung regions for prediction facilitates the understanding of machine learning-based prediction.
Identifiants
pubmed: 33733689
doi: 10.1007/s00330-020-07635-6
pii: 10.1007/s00330-020-07635-6
pmc: PMC8270830
doi:
Types de publication
Journal Article
Langues
eng
Sous-ensembles de citation
IM
Pagination
5443-5453Commentaires et corrections
Type : ErratumIn
Informations de copyright
© 2021. The Author(s).
Références
Elmali M, Baydin A, Nural MS et al (2007) Lung parenchymal injury and its frequency in blunt thoracic trauma: the diagnostic value of chest radiography and thoracic CT. Diagn Interv Radiol 13:179–182
pubmed: 18092287
Matthay MA, Ware LB, Zimmerman GA (2012) The acute respiratory distress syndrome. J Clin Investig 122:2731–2740
pubmed: 3408735
pmcid: 3408735
Wu J, Sheng L, Ma Y et al (2008) The analysis of risk factors of impacting mortality rate in severe multiple trauma patients with posttraumatic acute respiratory distress syndrome. Am J Emerg Med 26:419–424
pubmed: 18410809
Alisha C, Gajanan G, Jyothi H (2015) Risk factors affecting the prognosis in patients with pulmonary contusion following chest trauma. J Clin Diagn Res 9:OC17–OC19
pubmed: 26435984
pmcid: 4576577
Neher MD, Weckbach S, Flierl MA et al (2011) Molecular mechanisms of inflammation and tissue injury after major trauma-is complement the “bad guy”? J Biomed Sci 18:90
pubmed: 22129197
pmcid: 3247859
Perl M, Lomas-Neira J, Venet F et al (2011) Pathogenesis of indirect (secondary) acute lung injury. Expert Rev Respir Med 5:115–126
pubmed: 21348592
pmcid: 3108849
ARDS Definition Task Force, Ranieri VM, Rubenfeld GD et al (2012) Acute respiratory distress syndrome: the Berlin Definition. JAMA 307:2526–2533
Calfee CS, Janz DR, Bernard GR et al (2015) Distinct molecular phenotypes of direct vs indirect ARDS in single-center and multicenter studies. Chest 147:1539–1548
pubmed: 26033126
Garcia CSNB, Pelosi P, Rocco PRM (2008) Pulmonary and extrapulmonary acute respiratory distress syndrome: are they different? Rev Bras Ter Intensiva 20:178–183
pubmed: 25307007
Baker SP, O’Neill B, Haddon W Jr, Long WB (1974) The injury severity score: a method for describing patients with multiple injuries and evaluating emergency care. J Trauma 14:187–196
pubmed: 4814394
Miller PR, Croce MA, Kilgo PD et al (2002) Acute respiratory distress syndrome in blunt trauma: identification of independent risk factors. Am Surg 68:845–850 discussion 850–1
pubmed: 12412708
Afshar M, Smith GS, Cooper RS et al (2016) Trauma indices for prediction of acute respiratory distress syndrome. J Surg Res 201:394–401
pubmed: 27020824
Fahr M, Jones G, O’Neal H et al (2017) Acute respiratory distress syndrome incidence, but not mortality, has decreased nationwide: a National Trauma Data Bank Study. Am Surg 83:323–331
pubmed: 28424124
O’Leary MP, Keeley JA, Yule A et al (2016) Clinical predictors of early acute respiratory distress syndrome in trauma patients. Am J Surg 212:1096–1100
pubmed: 27780558
Daher P, Teixeira PG, Coopwood TB et al (2018) Mild to moderate to severe: what drives the severity of ARDS in trauma patients? Am Surg 84:808–812
pubmed: 29981606
Rubenfeld GD, Caldwell E, Peabody E et al (2005) Incidence and outcomes of acute lung injury. N Engl J Med 353:1685–1693
pubmed: 16236739
van Wessem KJP, Leenen LPH (2018) Incidence of acute respiratory distress syndrome and associated mortality in a polytrauma population. Trauma Surg Acute Care Open 3:e000232
pubmed: 30623025
pmcid: 6307585
Herridge MS, Tansey CM, Matté A et al (2011) Functional disability 5 years after acute respiratory distress syndrome. N Engl J Med 364:1293–1304
pubmed: 21470008
Daurat A, Millet I, Roustan J-P et al (2016) Thoracic Trauma Severity score on admission allows to determine the risk of delayed ARDS in trauma patients with pulmonary contusion. Injury 47:147–153
pubmed: 26358517
Wang S, Ruan Z, Zhang J, Jin W (2011) The value of pulmonary contusion volume measurement with three-dimensional computed tomography in predicting acute respiratory distress syndrome development. Ann Thorac Surg 92:1977–1983
pubmed: 21967818
Strumwasser A, Chu E, Yeung L et al (2011) A novel CT volume index score correlates with outcomes in polytrauma patients with pulmonary contusion. J Surg Res 170:280–285
pubmed: 21601877
Beitler JR, Schoenfeld DA, Taylor Thompson B (2014) Preventing ARDS. Chest 146:1102–1113
pubmed: 25288000
pmcid: 4188145
Yilmaz M, Keegan MT, Iscimen R et al (2007) Toward the prevention of acute lung injury: protocol-guided limitation of large tidal volume ventilation and inappropriate transfusion. Crit Care Med 35:1660–1666
pubmed: 17507824
Johnson SB (2008) Tracheobronchial injury. Semin Thorac Cardiovasc Surg 20:52–57
pubmed: 18420127
Raghavendran K, Davidson BA, Woytash JA et al (2005) The evolution of isolated bilateral lung contusion from blunt chest trauma in rats: cellular and cytokine responses. Shock 24:132–138
pubmed: 16044083
pmcid: 3034634
Sangster GP, González-Beicos A, Carbo AI et al (2007) Blunt traumatic injuries of the lung parenchyma, pleura, thoracic wall, and intrathoracic airways: multidetector computer tomography imaging findings. Emerg Radiol 14:297–310
pubmed: 17623115
Cohn SM (1997) Pulmonary contusion: review of the clinical entity. J Trauma 42:973–979
pubmed: 9191684
deSouza NM, Achten E, Alberich-Bayarri A et al (2019) Validated imaging biomarkers as decision-making tools in clinical trials and routine practice: current status and recommendations from the EIBALL* subcommittee of the European Society of Radiology (ESR). Insights Imaging 10:87
pubmed: 31468205
pmcid: 6715762
Lambin P, Rios-Velazquez E, Leijenaar R et al (2012) Radiomics: extracting more information from medical images using advanced feature analysis. Eur J Cancer 48:441–446
pubmed: 22257792
pmcid: 4533986
Röhrich S, Hofmanninger J, Prayer F et al (2020) Prospects and challenges of radiomics by using nononcologic routine chest CT. Radiology: Cardiothoracic Imaging 2:e190190
pmcid: 7978018
Negrin LL, Prosch H, Kettner S et al (2017) The clinical benefit of a follow-up thoracic computed tomography scan regarding parenchymal lung injury and acute respiratory distress syndrome in polytraumatized patients. J Crit Care 37:211–218
pubmed: 27969573
Hofmanninger J, Prayer F, Pan J, Röhrich S, Prosch H, Langs G (2020) Automatic lung segmentation in routine imaging is primarily a data diversity problem, nota methodology problem. Eur Radiol Exp 4(1):50. https://doi.org/10.1186/s41747-020-00173-2
Shafiq-Ul-Hassan M, Zhang GG, Latifi K et al (2017) Intrinsic dependencies of CT radiomic features on voxel size and number of gray levels. Med Phys 44:1050–1062
pubmed: 28112418
Friedman JH (2001) Greedy function approximation: a gradient boosting machine. Ann Stat 29:1189–1232
Pedregosa F, Varoquaux G, Gramfort A et al (2011) Scikit-learn: machine learning in Python. J Mach Learn Res 12:2825–2830. https://scikit-learn.org/stable/index.html . Accessed 31 March 2020
Virtanen P, Gommers R, Oliphant TE et al (2020) SciPy 1.0: fundamental algorithms for scientific computing in Python. Nat Methods 17:261–272
pubmed: 32015543
pmcid: 7056644
Gennarelli TA, Wodzin E (2006) AIS 2005: a contemporary injury scale. Injury 37:1083–1091
pubmed: 17092503
Bayer J, Lefering R, Reinhardt S et al (2017) Thoracic trauma severity contributes to differences in intensive care therapy and mortality of severely injuredpatients: analysis based on the TraumaRegister DGU®. World J Emerg Surg 12:43. https://doi.org/10.1186/s13017-017-0154-1
Ciesla DJ, Moore EE, Johnson JL et al (2005) The role of the lung in postinjury multiple organ failure. Surgery 138:749–758
pubmed: 16269305
Raghavendran K, Notter RH, Davidson BA et al (2009) LUNG CONTUSION. Shock 32:122–130
pubmed: 19174738
pmcid: 2711988
Becher RD, Colonna AL, Enniss TM et al (2012) An innovative approach to predict the development of adult respiratory distress syndrome in patients with blunt trauma. J Trauma Acute Care Surg 73:1229–1235
pubmed: 22914080
Miller PR, Croce MA, Bee TK et al (2001) ARDS after pulmonary contusion: accurate measurement of contusion volume identifies high-risk patients. J Trauma 51:223–230
pubmed: 11493778
Galan G, Peñalver JC, París F et al (1992) Blunt chest injuries in 1696 patients. Eur J Cardiothorac Surg 6:284–287
pubmed: 1616723
Mowery NT, Gunter OL, Collier BR et al (2011) Practice management guidelines for management of hemothorax and occult pneumothorax. J Trauma 70:510–518
pubmed: 21307755
Daly M, Miller PR, Jeffrey Carr J et al (2008) Traumatic pulmonary pathology measured with computed tomography and a semiautomated analytic method. Clin Imaging 32:346–354
pubmed: 18760721
Desai SR, Wells AU, Suntharalingam G et al (2001) Acute respiratory distress syndrome caused by pulmonary and extrapulmonary injury: a comparative CT study. Radiology 218:689–693
pubmed: 11230641
Tignanelli CJ, Hemmila MR, Rogers MAM, Raghavendran K (2019) Nationwide cohort study of independent risk factors for acute respiratory distress syndrome after trauma. Trauma Surg Acute Care Open 4:e000249
pubmed: 30899792
pmcid: 6407565
Weikert T, Noordtzij LA, Bremerich J et al (2020) Assessment of a deep learning algorithm for the detection of rib fractures on whole-body trauma computed tomography. Korean J Radiol 21:891–899
pubmed: 32524789
Zeiberg D, Prahlad T, Nallamothu BK et al (2019) Machine learning for patient risk stratification for acute respiratory distress syndrome. PLoS One 14:e0214465
pubmed: 30921400
pmcid: 6438573
Meyer A, Zverinski D, Pfahringer B et al (2018) Machine learning for real-time prediction of complications in critical care: a retrospective study. Lancet Respir Med 6:905–914
pubmed: 30274956
Kim H, Park CM, Park SJ et al (2016) Temporal changes of texture features extracted from pulmonary nodules on dynamic contrast-enhanced chest computed tomography. Invest Radiol 51:569–574
pubmed: 27504795
Mackin D, Fave X, Zhang L et al (2015) Measuring computed tomography scanner variability of radiomics features. Invest Radiol 50:757–765
pubmed: 26115366
pmcid: 4598251