A proof of concept for microcirculation monitoring using machine learning based hyperspectral imaging in critically ill patients: a monocentric observational study.
Journal
Critical care (London, England)
ISSN: 1466-609X
Titre abrégé: Crit Care
Pays: England
ID NLM: 9801902
Informations de publication
Date de publication:
10 Jul 2024
10 Jul 2024
Historique:
received:
18
04
2024
accepted:
06
07
2024
medline:
11
7
2024
pubmed:
11
7
2024
entrez:
10
7
2024
Statut:
epublish
Résumé
Impaired microcirculation is a cornerstone of sepsis development and leads to reduced tissue oxygenation, influenced by fluid and catecholamine administration during treatment. Hyperspectral imaging (HSI) is a non-invasive bedside technology for visualizing physicochemical tissue characteristics. Machine learning (ML) for skin HSI might offer an automated approach for bedside microcirculation assessment, providing an individualized tissue fingerprint of critically ill patients in intensive care. The study aimed to determine if machine learning could be utilized to automatically identify regions of interest (ROIs) in the hand, thereby distinguishing between healthy individuals and critically ill patients with sepsis using HSI. HSI raw data from 75 critically ill sepsis patients and from 30 healthy controls were recorded using TIVITA® Tissue System and analyzed using an automated ML approach. Additionally, patients were divided into two groups based on their SOFA scores for further subanalysis: less severely ill (SOFA ≤ 5) and severely ill (SOFA > 5). The analysis of the HSI raw data was fully-automated using MediaPipe for ROI detection (palm and fingertips) and feature extraction. HSI Features were statistically analyzed to highlight relevant wavelength combinations using Mann-Whitney-U test and Benjamini, Krieger, and Yekutieli (BKY) correction. In addition, Random Forest models were trained using bootstrapping, and feature importances were determined to gain insights regarding the wavelength importance for a model decision. An automated pipeline for generating ROIs and HSI feature extraction was successfully established. HSI raw data analysis accurately distinguished healthy controls from sepsis patients. Wavelengths at the fingertips differed in the ranges of 575-695 nm and 840-1000 nm. For the palm, significant differences were observed in the range of 925-1000 nm. Feature importance plots indicated relevant information in the same wavelength ranges. Combining palm and fingertip analysis provided the highest reliability, with an AUC of 0.92 to distinguish between sepsis patients and healthy controls. Based on this proof of concept, the integration of automated and standardized ROIs along with automated skin HSI analyzes, was able to differentiate between healthy individuals and patients with sepsis. This approach offers a reliable and objective assessment of skin microcirculation, facilitating the rapid identification of critically ill patients.
Sections du résumé
BACKGROUND
BACKGROUND
Impaired microcirculation is a cornerstone of sepsis development and leads to reduced tissue oxygenation, influenced by fluid and catecholamine administration during treatment. Hyperspectral imaging (HSI) is a non-invasive bedside technology for visualizing physicochemical tissue characteristics. Machine learning (ML) for skin HSI might offer an automated approach for bedside microcirculation assessment, providing an individualized tissue fingerprint of critically ill patients in intensive care. The study aimed to determine if machine learning could be utilized to automatically identify regions of interest (ROIs) in the hand, thereby distinguishing between healthy individuals and critically ill patients with sepsis using HSI.
METHODS
METHODS
HSI raw data from 75 critically ill sepsis patients and from 30 healthy controls were recorded using TIVITA® Tissue System and analyzed using an automated ML approach. Additionally, patients were divided into two groups based on their SOFA scores for further subanalysis: less severely ill (SOFA ≤ 5) and severely ill (SOFA > 5). The analysis of the HSI raw data was fully-automated using MediaPipe for ROI detection (palm and fingertips) and feature extraction. HSI Features were statistically analyzed to highlight relevant wavelength combinations using Mann-Whitney-U test and Benjamini, Krieger, and Yekutieli (BKY) correction. In addition, Random Forest models were trained using bootstrapping, and feature importances were determined to gain insights regarding the wavelength importance for a model decision.
RESULTS
RESULTS
An automated pipeline for generating ROIs and HSI feature extraction was successfully established. HSI raw data analysis accurately distinguished healthy controls from sepsis patients. Wavelengths at the fingertips differed in the ranges of 575-695 nm and 840-1000 nm. For the palm, significant differences were observed in the range of 925-1000 nm. Feature importance plots indicated relevant information in the same wavelength ranges. Combining palm and fingertip analysis provided the highest reliability, with an AUC of 0.92 to distinguish between sepsis patients and healthy controls.
CONCLUSION
CONCLUSIONS
Based on this proof of concept, the integration of automated and standardized ROIs along with automated skin HSI analyzes, was able to differentiate between healthy individuals and patients with sepsis. This approach offers a reliable and objective assessment of skin microcirculation, facilitating the rapid identification of critically ill patients.
Identifiants
pubmed: 38987802
doi: 10.1186/s13054-024-05023-w
pii: 10.1186/s13054-024-05023-w
doi:
Types de publication
Journal Article
Observational Study
Langues
eng
Sous-ensembles de citation
IM
Pagination
230Subventions
Organisme : Stiftung Universitätsmedizin Essen
ID : 20204699120
Informations de copyright
© 2024. The Author(s).
Références
De Backer D, Aissaoui N, Cecconi M, Chew MS, Denault A, Hajjar L, Hernandez G, Messina A, Myatra SN, Ostermann M, et al. How can assessing hemodynamics help to assess volume status? Intensive Care Med. 2022;48(10):1482–94.
doi: 10.1007/s00134-022-06808-9
pubmed: 35945344
pmcid: 9363272
Evans L, Rhodes A, Alhazzani W, Antonelli M, Coopersmith CM, French C, Machado FR, McIntyre L, Ostermann M, Prescott HC, et al. Surviving sepsis campaign: international guidelines for management of sepsis and septic shock 2021. Intensive Care Med. 2021;47(11):1181–247.
doi: 10.1007/s00134-021-06506-y
pubmed: 34599691
pmcid: 8486643
Duranteau J, De Backer D, Donadello K, Shapiro NI, Hutchings SD, Rovas A, Legrand M, Harrois A, Ince C. The future of intensive care: the study of the microcirculation will help to guide our therapies. Crit Care. 2023;27(1):190.
doi: 10.1186/s13054-023-04474-x
pubmed: 37193993
pmcid: 10186296
Ince C. Hemodynamic coherence and the rationale for monitoring the microcirculation. Crit Care. 2015;19(Suppl 3):S8.
doi: 10.1186/cc14726
pubmed: 26729241
pmcid: 4699073
Tafner P, Chen FK, Rabello RF, Correa TD, Chaves RCF, Serpa AN. Recent advances in bedside microcirculation assessment in critically ill patients. Rev Bras Ter Intensiva. 2017;29(2):238–47.
doi: 10.5935/0103-507X.20170033
pubmed: 28977264
pmcid: 5496759
Hilty MP, Favaron E, Wendel Garcia PD, Ahiska Y, Uz Z, Akin S, Flick M, Arbous S, Hofmaenner DA, Saugel B, et al. Microcirculatory alterations in critically ill COVID-19 patients analyzed using artificial intelligence. Crit Care. 2022;26(1):311.
doi: 10.1186/s13054-022-04190-y
pubmed: 36242010
pmcid: 9568900
Cui R, Yu H, Xu T, Xing X, Cao X, Yan K, Chen J. Deep learning in medical hyperspectral images: a review. Sensors (Basel). 2022;22(24):9790.
doi: 10.3390/s22249790
pubmed: 36560157
Lu G, Fei B. Medical hyperspectral imaging: a review. J Biomed Opt. 2014;19(1):10901.
doi: 10.1117/1.JBO.19.1.010901
pubmed: 24441941
Dietrich M, Marx S, von der Forst M, Bruckner T, Schmitt FCF, Fiedler MO, Nickel F, Studier-Fischer A, Muller-Stich BP, Hackert T, et al. Bedside hyperspectral imaging indicates a microcirculatory sepsis pattern—an observational study. Microvasc Res. 2021;136:104164.
doi: 10.1016/j.mvr.2021.104164
pubmed: 33831406
Dietrich M, Marx S, von der Forst M, Bruckner T, Schmitt FCF, Fiedler MO, Nickel F, Studier-Fischer A, Muller-Stich BP, Hackert T, et al. Hyperspectral imaging for perioperative monitoring of microcirculatory tissue oxygenation and tissue water content in pancreatic surgery—an observational clinical pilot study. Perioper Med (Lond). 2021;10(1):42.
doi: 10.1186/s13741-021-00211-6
pubmed: 34847953
Kazune S, Caica A, Luksevics E, Volceka K, Grabovskis A. Impact of increased mean arterial pressure on skin microcirculatory oxygenation in vasopressor-requiring septic patients: an interventional study. Ann Intensive Care. 2019;9(1):97.
doi: 10.1186/s13613-019-0572-1
pubmed: 31468202
pmcid: 6715757
Kuhlmann H, Garczarek L, Kunne D, Pattberg K, Skarabis A, Frank M, Schmidt B, Arends S, Herbstreit F, Brenner T, et al. Bedside hyperspectral imaging and organ dysfunction severity in critically ill COVID-19 patients-a prospective, monocentric observational study. Bioengineering (Basel). 2023;10(10):1167.
doi: 10.3390/bioengineering10101167
pubmed: 37892897
Holmer A, Marotz J, Wahl P, Dau M, Kammerer PW. Hyperspectral imaging in perfusion and wound diagnostics—methods and algorithms for the determination of tissue parameters. Biomed Tech (Berl). 2018;63(5):547–56.
doi: 10.1515/bmt-2017-0155
pubmed: 30028724
Holmer A, Tetschke F, Marotz J, Malberg H, Markgraf W, Thiele C, Kulcke A. Oxygenation and perfusion monitoring with a hyperspectral camera system for chemical based tissue analysis of skin and organs. Physiol Meas. 2016;37(11):2064–78.
doi: 10.1088/0967-3334/37/11/2064
pubmed: 27786164
Kulcke A, Holmer A, Wahl P, Siemers F, Wild T, Daeschlein G. A compact hyperspectral camera for measurement of perfusion parameters in medicine. Biomed Tech (Berl). 2018;63(5):519–27.
doi: 10.1515/bmt-2017-0145
pubmed: 29522415
Kazune S, Caica A, Volceka K, Suba O, Rubins U, Grabovskis A. Relationship of mottling score, skin microcirculatory perfusion indices and biomarkers of endothelial dysfunction in patients with septic shock: an observational study. Crit Care. 2019;23(1):311.
doi: 10.1186/s13054-019-2589-0
pubmed: 31511042
pmcid: 6739999
Dietrich M, Özdemir B, Gruneberg D, Petersen C, Studier-Fischer A, von der Forst M, Schmitt FCF, Fiedler MO, Nickel F, Müller-Stich BP, et al. Hyperspectral imaging for the evaluation of microcirculatory tissue oxygenation and perfusion quality in haemorrhagic shock: a porcine study. Biomedicines. 2021;9(12):1829.
doi: 10.3390/biomedicines9121829
pubmed: 34944645
pmcid: 8698916
Studier-Fischer A, Seidlitz S, Sellner J, Ozdemir B, Wiesenfarth M, Ayala L, Odenthal J, Knodler S, Kowalewski KF, Haney CM, et al. Spectral organ fingerprints for machine learning-based intraoperative tissue classification with hyperspectral imaging in a porcine model. Sci Rep. 2022;12(1):11028.
doi: 10.1038/s41598-022-15040-w
pubmed: 35773276
pmcid: 9247052
Vasilevskis EE, Pandharipande PP, Graves AJ, Shintani A, Tsuruta R, Ely EW, Girard TD. Validity of a modified sequential organ failure assessment score using the richmond agitation-sedation scale. Crit Care Med. 2016;44(1):138–46.
doi: 10.1097/CCM.0000000000001375
pubmed: 26457749
pmcid: 4748963
Lugaresi C, Tang J, Nash H, McClanahan C, Uboweja E, Hays M, Zhang F, Chang C-L, Yong M, Lee J. Mediapipe: a framework for perceiving and processing reality. In: Third workshop on computer vision for AR/VR at IEEE computer vision and pattern recognition (CVPR). 2019.
Pölsterl S. Scikit-survival: a library for time-to-event analysis built on top of scikit-learn. J Mach Learn Res. 2020;21(212):1–6.
Virtanen P, Gommers R, Oliphant TE, Haberland M, Reddy T, Cournapeau D, Burovski E, Peterson P, Weckesser W, Bright J, et al. SciPy 1.0: fundamental algorithms for scientific computing in Python. Nat Methods. 2020;17(3):261–72.
doi: 10.1038/s41592-019-0686-2
pubmed: 32015543
pmcid: 7056644
Seabold S, Perktold J: Statsmodels: Econometric and statistical modeling with python. In: SciPy (2010).
Hand DJ, Christen P, Kirielle N. F*: an interpretable transformation of the F-measure. Mach Learn. 2021;110(3):451–6.
doi: 10.1007/s10994-021-05964-1
pubmed: 33746357
pmcid: 7958589
Kattan E, Hernández G, Ospina-Tascón G, Valenzuela ED, Bakker J, Castro R. The A-SSI, the Latin America Intensive Care N: a lactate-targeted resuscitation strategy may be associated with higher mortality in patients with septic shock and normal capillary refill time: a post hoc analysis of the ANDROMEDA-SHOCK study. Ann Intensive Care. 2020;10(1):114.
doi: 10.1186/s13613-020-00732-1
pubmed: 32845407
pmcid: 7450018
Leone M, Einav S, Antonucci E, Depret F, Lakbar I, Martin-Loeches I, Wieruszewski PM, Myatra SN, Khanna AK. Multimodal strategy to counteract vasodilation in septic shock. Anaesth Crit Care Pain Med. 2023;42(3):101193.
doi: 10.1016/j.accpm.2023.101193
pubmed: 36621622
Hariri G, Joffre J, Leblanc G, Bonsey M, Lavillegrand JR, Urbina T, Guidet B, Maury E, Bakker J, Ait-Oufella H. Narrative review: clinical assessment of peripheral tissue perfusion in septic shock. Ann Intensive Care. 2019;9(1):37.
doi: 10.1186/s13613-019-0511-1
pubmed: 30868286
pmcid: 6419794
Kattan E, Ospina-Tascón GA, Teboul J-L, Castro R, Cecconi M, Ferri G, Bakker J, Hernández G, Hernandez G, Ospina-Tascón G, et al. Systematic assessment of fluid responsiveness during early septic shock resuscitation: secondary analysis of the ANDROMEDA-SHOCK trial. Crit Care. 2020;24(1):23.
doi: 10.1186/s13054-020-2732-y
pubmed: 31973735
pmcid: 6979284
de Keijzer IN, Massari D, Sahinovic M, Flick M, Vos JJ, Scheeren TWL. What is new in microcirculation and tissue oxygenation monitoring? J Clin Monit Comput. 2022;36(2):291–9.
doi: 10.1007/s10877-022-00837-x
pubmed: 35275312
pmcid: 9122846
Guerraty M, Bhargava A, Senarathna J, Mendelson AA, Pathak AP. Advances in translational imaging of the microcirculation. Microcirculation. 2021;28(3):e12683.
doi: 10.1111/micc.12683
pubmed: 33524206
pmcid: 8647298
Zhang Y, Wu X, He L, Meng C, Du S, Bao J, Zheng Y. Applications of hyperspectral imaging in the detection and diagnosis of solid tumors. Transl Cancer Res. 2020;9(2):1265–77.
doi: 10.21037/tcr.2019.12.53
pubmed: 35117471
pmcid: 8798535
Barberio M, Benedicenti S, Pizzicannella M, Felli E, Collins T, Jansen-Winkeln B, Marescaux J, Viola MG, Diana M. Intraoperative guidance using hyperspectral imaging: a review for surgeons. Diagnostics. 2021;11(11):2066.
doi: 10.3390/diagnostics11112066
pubmed: 34829413
pmcid: 8624094
Huang H-Y, Hsiao Y-P, Karmakar R, Mukundan A, Chaudhary P, Hsieh S-C, Wang H-C. A review of recent advances in computer-aided detection methods using hyperspectral imaging engineering to detect skin cancer. Cancers. 2023;15(23):5634.
doi: 10.3390/cancers15235634
pubmed: 38067338
pmcid: 10705122
Macdonald SP, Brown SG. Near-infrared spectroscopy in the assessment of suspected sepsis in the emergency department. Emerg Med J. 2015;32(5):404–8.
doi: 10.1136/emermed-2013-202956
pubmed: 24154943
Seidlitz S, Sellner J, Odenthal J, Ozdemir B, Studier-Fischer A, Knodler S, Ayala L, Adler TJ, Kenngott HG, Tizabi M, et al. Robust deep learning-based semantic organ segmentation in hyperspectral images. Med Image Anal. 2022;80:102488.
doi: 10.1016/j.media.2022.102488
pubmed: 35667327
Maktabi M, Kohler H, Ivanova M, Jansen-Winkeln B, Takoh J, Niebisch S, Rabe SM, Neumuth T, Gockel I, Chalopin C. Tissue classification of oncologic esophageal resectates based on hyperspectral data. Int J Comput Assist Radiol Surg. 2019;14(10):1651–61.
doi: 10.1007/s11548-019-02016-x
pubmed: 31222672
He Q, Wang R. Hyperspectral imaging enabled by an unmodified smartphone for analyzing skin morphological features and monitoring hemodynamics. Biomed Opt Express. 2020;11(2):895–910.
doi: 10.1364/BOE.378470
pubmed: 32133229
pmcid: 7041456
Stuart MB, McGonigle AJS, Davies M, Hobbs MJ, Boone NA, Stanger LR, Zhu C, Pering TD, Willmott JR. Low-cost hyperspectral imaging with a smartphone. J Imaging. 2021;7(8):136.
doi: 10.3390/jimaging7080136
pubmed: 34460772
pmcid: 8404918