- Email: [email protected]
ACCURACY OF INTERPRETATION OF COMPUTERIZED MULTICHANNEL LUNG SOUND ANALYSES
October 2008, Vol 134, No. 4_MeetingAbstracts Abstract: Poster Presentations | October 2008
ACCURACY OF INTERPRETATION OF COMPUTERIZED MULTICHANNEL LUNG SOUND ANALYSES Raymond L. Murphy, MD*; Andrey Vyshedskiy, PhD; Anna Wong-Tse, RN Brigham and Women's/Faulkner Hospitals, Boston, MA Chest Chest. 2008;134(4_MeetingAbstracts):p32002. doi:10.1378/chest.134.4_MeetingAbstracts.p32002
Abstract PURPOSE: The purpose of this investigation was to determine the diagnostic accuracy of the interpretation of waveforms and computerized analyses obtained by a multichannel lung sound analyzer. METHODS: 54 patients seen at a community teaching hospital were included in this study. Their clinical diagnoses included pneumonia, obstructive lung disease, pneumothorax, vocal cord dysfunction, congestive heart failure and patients with no recognized cardiopulmonary disorder. The patients underwent examination with a multichannel lung sound analyzer (Stethographics STG16). Two observers, blinded as to the diagnoses made on these patients, reviewed the waveforms and computerized analyses obtained. The computerized analyses included quantification and localization of sound amplitude and adventitious sounds such as wheezes, crackles and rhonchi. An example of the information presented to the observers is presented in Fig. 1. Figure 1A: waveforms of sounds of a single breath displayed in relation to the anatomic location where they were detected. 1B: Crackle counts made by the computer; circle size is proportional to crackle counts. 1C: Crackle timing in inspiration (grey) and expiration (black), bar size is proportional to the crackle count. A consensus view was reached and compared to the clinical diagnosis. RESULTS: The interpretation by the observers agreed with the clinical diagnosis in 84% of the cases. The most common error was in the misclassification of pneumonia and congestive heart failure. CONCLUSION: Data collected with a multichannel lung sound analyzer agreed surprisingly well with the clinical diagnoses. CLINICAL IMPLICATIONS: Computerized multichannel lung sound analysis offers the promise of providing clinically useful diagnostic information at the bedside. DISCLOSURE: Raymond Murphy, Grant monies (from sources other than industry) Supported in part by grant from NIH SBIR (1R43HL70480–01); Grant monies (from industry related sources) Supported in part by grant from Stethographics, Inc.; Shareholder Dr. Murphy and Dr.
Vyshedskiy have financial interests in Stethographics, Inc.; Employee Dr. Murphy and Dr. Vyshedskiy have financial interests in Stethographics, Inc.; No Product/Research Disclosure Information Tuesday, October 28, 2008 1:00 PM - 2:15 PM
ACCURACY OF INTERPRETATION OF COMPUTERIZED MULTICHANNEL LUNG SOUND ANALYSES Raymond L. Murphy, MD*; Andrey Vyshedskiy, PhD; Anna Wong-Tse, RN Brigham and Women's/Faulkner Hospitals, Boston, MA Chest Chest. 2008;134(4_MeetingAbstracts):p32002. doi:10.1378/chest.134.4_MeetingAbstracts.p32002
Abstract PURPOSE: The purpose of this investigation was to determine the diagnostic accuracy of the interpretation of waveforms and computerized analyses obtained by a multichannel lung sound analyzer. METHODS: 54 patients seen at a community teaching hospital were included in this study. Their clinical diagnoses included pneumonia, obstructive lung disease, pneumothorax, vocal cord dysfunction, congestive heart failure and patients with no recognized cardiopulmonary disorder. The patients underwent examination with a multichannel lung sound analyzer (Stethographics STG16). Two observers, blinded as to the diagnoses made on these patients, reviewed the waveforms and computerized analyses obtained. The computerized analyses included quantification and localization of sound amplitude and adventitious sounds such as wheezes, crackles and rhonchi. An example of the information presented to the observers is presented in Fig. 1. Figure 1A: waveforms of sounds of a single breath displayed in relation to the anatomic location where they were detected. 1B: Crackle counts made by the computer; circle size is proportional to crackle counts. 1C: Crackle timing in inspiration (grey) and expiration (black), bar size is proportional to the crackle count. A consensus view was reached and compared to the clinical diagnosis. RESULTS: The interpretation by the observers agreed with the clinical diagnosis in 84% of the cases. The most common error was in the misclassification of pneumonia and congestive heart failure. CONCLUSION: Data collected with a multichannel lung sound analyzer agreed surprisingly well with the clinical diagnoses. CLINICAL IMPLICATIONS: Computerized multichannel lung sound analysis offers the promise of providing clinically useful diagnostic information at the bedside. DISCLOSURE: Raymond Murphy, Grant monies (from sources other than industry) Supported in part by grant from NIH SBIR (1R43HL70480–01); Grant monies (from industry related sources) Supported in part by grant from Stethographics, Inc.; Shareholder Dr. Murphy and Dr.
Vyshedskiy have financial interests in Stethographics, Inc.; Employee Dr. Murphy and Dr. Vyshedskiy have financial interests in Stethographics, Inc.; No Product/Research Disclosure Information Tuesday, October 28, 2008 1:00 PM - 2:15 PM