Neural network analysis of the volumetric capnogram to detect pulmonary embolism

Manish M. Patel, Daniel B. Rayburn, Jane A. Browning, Jeffrey Kline

Research output: Contribution to journalArticle

12 Citations (Scopus)

Abstract

Background: Pulmonary embolism (PE) produces ventilation/perfusion mismatch that may be manifested in various variables of the volume-based capnogram (VBC). We hypothesized that a neural network (NN) system could detect changes in VBC variables that reflect the presence of a PE. Methods: A commercial VBC system was used to record multiple respiratory variables from consecutive expiratory breaths. Data from 12 subjects (n = 6 PE+ and n = 6 PE-) were used as input to a fully connected back-propagating NN for model development. The derived model was tested in a prospective, observational study at an urban teaching hospital. Volumetric capnograms were then collected on 53 test subjects: 30 subjects with PE confirmed by pulmonary angiography or diagnostic scintillation lung scan; and 23 subjects without PE based on pulmonary angiography. The derived NN model was applied to VBC data from the test population. Results: Seventeen VBC variables were used by the derived NN model to generate a numeric probability of PE. When the derived NN model was applied to VBC data from the 53 test subjects, PE was detected with a sensitivity of 100% (95% CI = 89% to 100%) and a specificity of 48% (95% CI = 27% to 69%). The likelihood ratio positive [LR(+)] for the VBC-NN test was 1.82 and the LR(-) was 0.1. Conclusion: This study demonstrates the feasibility of developing a rapid, noninvasive breath test for diagnosing PE using volumetric capnography and NN analysis.

Original languageEnglish (US)
Pages (from-to)1325-1332
Number of pages8
JournalChest
Volume116
Issue number5
DOIs
StatePublished - 1999
Externally publishedYes

Fingerprint

Pulmonary Embolism
Neural Networks (Computer)
Lung
Angiography
Capnography
Breath Tests
Urban Hospitals
Feasibility Studies
Teaching Hospitals
Observational Studies
Ventilation
Perfusion
Prospective Studies
Population

Keywords

  • Artificial intelligence
  • Capnography
  • Diagnosis
  • Pulmonary embolism
  • Respiratory monitoring
  • Thromboembolism

ASJC Scopus subject areas

  • Pulmonary and Respiratory Medicine

Cite this

Neural network analysis of the volumetric capnogram to detect pulmonary embolism. / Patel, Manish M.; Rayburn, Daniel B.; Browning, Jane A.; Kline, Jeffrey.

In: Chest, Vol. 116, No. 5, 1999, p. 1325-1332.

Research output: Contribution to journalArticle

Patel, Manish M. ; Rayburn, Daniel B. ; Browning, Jane A. ; Kline, Jeffrey. / Neural network analysis of the volumetric capnogram to detect pulmonary embolism. In: Chest. 1999 ; Vol. 116, No. 5. pp. 1325-1332.
@article{0fcabf2f4029437281ac8db26f96c692,
title = "Neural network analysis of the volumetric capnogram to detect pulmonary embolism",
abstract = "Background: Pulmonary embolism (PE) produces ventilation/perfusion mismatch that may be manifested in various variables of the volume-based capnogram (VBC). We hypothesized that a neural network (NN) system could detect changes in VBC variables that reflect the presence of a PE. Methods: A commercial VBC system was used to record multiple respiratory variables from consecutive expiratory breaths. Data from 12 subjects (n = 6 PE+ and n = 6 PE-) were used as input to a fully connected back-propagating NN for model development. The derived model was tested in a prospective, observational study at an urban teaching hospital. Volumetric capnograms were then collected on 53 test subjects: 30 subjects with PE confirmed by pulmonary angiography or diagnostic scintillation lung scan; and 23 subjects without PE based on pulmonary angiography. The derived NN model was applied to VBC data from the test population. Results: Seventeen VBC variables were used by the derived NN model to generate a numeric probability of PE. When the derived NN model was applied to VBC data from the 53 test subjects, PE was detected with a sensitivity of 100{\%} (95{\%} CI = 89{\%} to 100{\%}) and a specificity of 48{\%} (95{\%} CI = 27{\%} to 69{\%}). The likelihood ratio positive [LR(+)] for the VBC-NN test was 1.82 and the LR(-) was 0.1. Conclusion: This study demonstrates the feasibility of developing a rapid, noninvasive breath test for diagnosing PE using volumetric capnography and NN analysis.",
keywords = "Artificial intelligence, Capnography, Diagnosis, Pulmonary embolism, Respiratory monitoring, Thromboembolism",
author = "Patel, {Manish M.} and Rayburn, {Daniel B.} and Browning, {Jane A.} and Jeffrey Kline",
year = "1999",
doi = "10.1378/chest.116.5.1325",
language = "English (US)",
volume = "116",
pages = "1325--1332",
journal = "Chest",
issn = "0012-3692",
publisher = "American College of Chest Physicians",
number = "5",

}

TY - JOUR

T1 - Neural network analysis of the volumetric capnogram to detect pulmonary embolism

AU - Patel, Manish M.

AU - Rayburn, Daniel B.

AU - Browning, Jane A.

AU - Kline, Jeffrey

PY - 1999

Y1 - 1999

N2 - Background: Pulmonary embolism (PE) produces ventilation/perfusion mismatch that may be manifested in various variables of the volume-based capnogram (VBC). We hypothesized that a neural network (NN) system could detect changes in VBC variables that reflect the presence of a PE. Methods: A commercial VBC system was used to record multiple respiratory variables from consecutive expiratory breaths. Data from 12 subjects (n = 6 PE+ and n = 6 PE-) were used as input to a fully connected back-propagating NN for model development. The derived model was tested in a prospective, observational study at an urban teaching hospital. Volumetric capnograms were then collected on 53 test subjects: 30 subjects with PE confirmed by pulmonary angiography or diagnostic scintillation lung scan; and 23 subjects without PE based on pulmonary angiography. The derived NN model was applied to VBC data from the test population. Results: Seventeen VBC variables were used by the derived NN model to generate a numeric probability of PE. When the derived NN model was applied to VBC data from the 53 test subjects, PE was detected with a sensitivity of 100% (95% CI = 89% to 100%) and a specificity of 48% (95% CI = 27% to 69%). The likelihood ratio positive [LR(+)] for the VBC-NN test was 1.82 and the LR(-) was 0.1. Conclusion: This study demonstrates the feasibility of developing a rapid, noninvasive breath test for diagnosing PE using volumetric capnography and NN analysis.

AB - Background: Pulmonary embolism (PE) produces ventilation/perfusion mismatch that may be manifested in various variables of the volume-based capnogram (VBC). We hypothesized that a neural network (NN) system could detect changes in VBC variables that reflect the presence of a PE. Methods: A commercial VBC system was used to record multiple respiratory variables from consecutive expiratory breaths. Data from 12 subjects (n = 6 PE+ and n = 6 PE-) were used as input to a fully connected back-propagating NN for model development. The derived model was tested in a prospective, observational study at an urban teaching hospital. Volumetric capnograms were then collected on 53 test subjects: 30 subjects with PE confirmed by pulmonary angiography or diagnostic scintillation lung scan; and 23 subjects without PE based on pulmonary angiography. The derived NN model was applied to VBC data from the test population. Results: Seventeen VBC variables were used by the derived NN model to generate a numeric probability of PE. When the derived NN model was applied to VBC data from the 53 test subjects, PE was detected with a sensitivity of 100% (95% CI = 89% to 100%) and a specificity of 48% (95% CI = 27% to 69%). The likelihood ratio positive [LR(+)] for the VBC-NN test was 1.82 and the LR(-) was 0.1. Conclusion: This study demonstrates the feasibility of developing a rapid, noninvasive breath test for diagnosing PE using volumetric capnography and NN analysis.

KW - Artificial intelligence

KW - Capnography

KW - Diagnosis

KW - Pulmonary embolism

KW - Respiratory monitoring

KW - Thromboembolism

UR - http://www.scopus.com/inward/record.url?scp=0032696294&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=0032696294&partnerID=8YFLogxK

U2 - 10.1378/chest.116.5.1325

DO - 10.1378/chest.116.5.1325

M3 - Article

C2 - 10559095

AN - SCOPUS:0032696294

VL - 116

SP - 1325

EP - 1332

JO - Chest

JF - Chest

SN - 0012-3692

IS - 5

ER -