Інформація призначена тільки для фахівців сфери охорони здоров'я, осіб,
які мають вищу або середню спеціальну медичну освіту.

Підтвердіть, що Ви є фахівцем у сфері охорони здоров'я.

Мобильная версия

Регистрация Забыли пароль?

Журнал «Здоровье ребенка» Том 18, №4, 2023

Вернуться к номеру

Чутливість і специфічність методу акустичної діагностики пневмонії пристроєм акустичного моніторингу «Trembita-Corona»

Чутливість і специфічність методу акустичної діагностики пневмонії пристроєм акустичного моніторингу «Trembita-Corona»

Авторы: Yu.V. Marushko, O.V. Khomych
Bogomolets National Medical University, Kyiv, Ukraine

Рубрики: Педиатрия/Неонатология

Разделы: Клинические исследования

Версия для печати


Резюме

Актуальність. Діагностика пневмонії складна та вимагає комплексного обстеження. Перспективним методом діагностики наразі є респіраторна акустика. Мета дослідження: оцінити можливість виявлення пневмонії за середньою потужністю, частотою і амплітудою акустичного сигналу при використанні пристрою акустичного моніторингу «Trembita-Corona». Матеріали та методи. Обстежено 193 дитини віком від 1 місяця до 18 років. Їх було розділено на дві групи: перша — 98 пацієнтів із позалікарняною пневмонією, друга — 95 здорових дітей. Усім дітям було проведено дослідження на пристрої акустичного моніторингу «Trembita-Coronа». Результати. Чутливість методу виявлення пневмонії з використанням пристрою акустичного моніторингу «Trembita-Corona» за потужністю звукового сигналу в 4, 5, 6-й октаві становить від 90,8 до 99 %, а специфічність — від 87,4 до 100 %. Чутливість методу виявлення пневмонії при застосуванні пристрою акустичного моніторингу «Trembita-Corona» за частотою акустичного сигналу в 3-й та 5-й октавах становить понад 60 %, а специфічність — 63,2 % у 3-й октаві та 100 % у 5-й октаві. Чутливість методу виявлення пневмонії з використанням пристрою акустичного моніторингу «Trembita-Corona» за амплітудою акустичного сигналу в 4, 5, 6-й октавах становить 88,9; 99,9 і 86,9 % відповідно, а специфічність — 100 % у 4-й та 5-й октаві і 87,4 % у 6-й октаві. Висновки. Пристрій акустичного моніторингу «Trembita-Corona» демонструє високу специфічність, чутливість та прогностичну цінність позитивного результату та прогностичну цінність негативного результату при заданій поширеності.

Background. Diagnosis of pneumonia is difficult and requires a comprehensive examination. Respiratory acoustics is currently a promising diagnostic method. The purpose was to evaluate the possibility of pneumonia detection by the average signal power, frequency and amplitude of the acoustic signal using the Trembita-Corona acoustic monitoring device. Materials and methods. Onу hundred and ninety-three children aged from 1 month to 18 years were examined. The children were divided into two groups: group I — 98 patients with community-acquired pneumonia, group II — 95 healthy children. All children were examined using the Trembita-Corona acoustic monitoring device. Results. The sensitivity of the method for detecting pneumonia using the acoustic monitoring device Trembita-Corona according to the average signal power in the 4th, 5th and 6th octaves is from 90.8 to 99 %, and the specificity is from 87.4 to 100 %. The sensitivity of the method for detecting pneumonia using the acoustic monitoring device Trembita-Corona according to the frequency of the acoustic signal in the 3rd and 5th octaves is more than 60 %, and the specificity is more than 63 % in the 3rd octave and 100 % in the 5th octave, respectively. The sensitivity of the method for detecting pneumonia using the acoustic monitoring device Trembita-Corona according to the amplitude of the acoustic signal in the 4th, 5th and 6th octaves is 88.9, 99.9 and 86.9 %, respectively, and the specificity is 100 % in the 4th and 5th octaves and 87.4 % in the 6th octave. Conclusions. The acoustic monitoring device Trembita-Corona demonstrates high specificity, sensitivity and positive predictive value and negative predictive value at given prevalence.


Ключевые слова

діти; позалікарняна пневмонія; діагностика; пристрій акустичного моніторингу «Trembita-Coronа»

children; community-acquired pneumonia; diagnostics; Trembita-Corona acoustic monitoring device

doi: http://dx.doi.org/10.22141/2224-0551.18.4.2023.1597

Introduction

At present, community-acquired pneumonia (CAP) is an urgent problem and the leading cause of morbidity and mortality worldwide [1–4].
Early diagnosis is important for timely diagnosis and appropriate treatment. In recent years, international re–commendations on improving and speeding up the diagnosis of CAP have been updating all over the world and in Ukraine. In Ukraine, the diagnosis of CAP is carried out in accordance with the new order of the Ministry of Health of Ukraine No. 1380 dated August 2, 2022 [5, 6]. Accor–ding to it, the diagnosis of СAP is based on the collection of ana–mnesis, clinical data, application of the results of instrumental and laboratory studies [5, 6]. According to clinical guidelines “Pneumonia in children” (2022), the diagnosis of CAP is made clinically, and not all children need radiological exa–mination [5, 6].
Respiratory acoustics is currently a promising scientific direction all over the world. The main tasks in this field are the development of theories on the generation of sound in the lungs and the creation of acoustic methods that can improve disease diagnosis [7, 8].
Together with leading specialists of the National Aviation University, we developed an experimental model of the Trembita-Corona acoustic monitoring device for diagno–sing acoustic changes in the lungs [9]. The device facilitates the diagnosis of the disease, helps identify areas of lung da–mage and perform acoustic monitoring of the lungs. All these processes can be done in a remote format. One of the main advantages of this method is that it eliminates the subjective assessment of the doctor. This is due to the fact that the method involves an automated system for monitoring and evaluating respiratory sounds, which completely excludes the human factor [10, 11].
In this work, we studied the feasibility of pneumonia detection by the average signal power, frequency and amplitude of the acoustic signal using the Trembita-Corona acoustic monitoring device.
Specifically for the evaluation of pneumonia detection using the Trembita-Corona acoustic monitoring device, we used indicators such as model sensitivity, model specificity, positive predictive value (PPV), and negative predictive va–lue (PPV) [12]. These four indicators provide an assessment of the model’s predictive qualities.
To assess the quality of the logistic regression model, we used the method of construction and analysis of the receiver operating characteristic (ROC) curve. ROC analysis is a method of graphical evaluation of the effectiveness of mo–dels using two indicators — sensitivity and specificity [14].
The purpose of the study was to evaluate the possibi–lity of pneumonia detection by an average signal power, frequency and amplitude of the acoustic signal using the Trembita-Corona acoustic monitoring device.

Materials and methods

One hundred and ninety-three children were examined, including 98 with CAP and 95 healthy children aged from 1 month to 18 years. The children were divided into two groups: group I — 98 patients with community-acquired pneumonia (9.6 ± 0.3 years); group II — 95 healthy children (12.1 ± 0.5 years old).
Inclusion criteria: age from 1 month to 18 years, confirmed CAP, informed consent of the child’s parents or guardians.
Exclusion criteria: congenital pneumonia, endocrine diseases, congenital heart defects, genetic syndromes.
Children from group 1 underwent a comprehensive exa–mination, as prescribed for CAP, which included percussion, chest X-ray, clinical examination (general blood test, biochemical blood test).
Also, all children were examined using the Trembita-Corona acoustic monitoring device for the diagnosis of respiratory sounds and localization of lung damage zones [7–10].
In this work, the method of construction and analysis of the ROC curve was used to assess the quality of the logistic regression model. The quality of the built model was assessed by the area under the ROC curve (AUC). The model will be adequate to the experimental data if the AUC is statistically significantly (p < 0.05) exceeds 0.5. If the AUC does not differ from 0.5, then the model is not adequate to the experimental data [11].
The study was conducted in compliance with the international principles of GCP, GLP for clinical research. The protocol was approved at the meeting of the Commission on Bioethical Expertise at the Bogomolets National Medical University (Protocol No. 138 dated November 10, 2020). Informed consent of parents/guardians was also obtained, which was approved at the same meeting.

Results and discussion

All children were examined using the Trembita-Corona acoustic monitoring device for the diagnosis of respiratory sounds.
We determined criteria such as average signal power, frequency and amplitude of acoustic signals in 12 octaves for all patients using the Trembita-Corona acoustic monitoring device.
Table 1 shows the results of the evaluation of the coefficients of the model for detecting pneumonia from the ave–rage signal power using the Trembita-Corona acoustic monitoring device.
As can be seen from Table 1, the AUC in the 4th, 5th and 6th octaves is more than 0.9, which proves that the qua–lity of the model can conditionally be rated as excellent. In the 3rd octave, AUC is 0.71 (95% CI 0.636–0.784), which indicates the good quality of the model. The quality of the model is satis–factory in the 0th, 1st, 7th, 8th, 9th octaves (0.6 ≤ AUC ≤ 0.7).
For each octave, an analysis of the ROC curve of the final one-factor model for detecting pneumonia in children was performed depending on the average signal power. Sensitivity, specificity, positive predictive value, negative predictive value were determined.
It was found that at the selected decision threshold (Pborder, which is individual for each octave), the sensitivity of the method for detecting pneumonia based on the ave–rage signal power in the 4th, 5th and 6th octaves is more than 90 %, and the specificity is above 92 %. Thus, the diagnosis of pneumonia can best be made by the 4th, 5th and 6th octaves, respectively, which is presented in Table 2.
Taking into account the prevalence of the disease in 3.5 %, the PPV in the 5th octave according to the average signal power is 100 %, and the NPV is also 100 %.
Table 3 shows the results of the evaluation of the coefficients of the model for detecting pneumonia from the frequency of the acoustic signal using the Trembita-Corona acoustic monitoring device.
It was found that at the Pborder, the sensitivity of the me–thod for detecting pneumonia based on the frequency of the acoustic signal in the 3rd and 5th octaves is more than 60 %, and the specificity is more than 63 % in the 3rd octave and 100 % in the 5th octave. Thus, the diagnosis of pneumonia based on the frequency of the acoustic signal can best be made in the 3rd and 5th octaves, respectively, which is presented in Table 4.
Taking into account the prevalence of the disease in 3.5 %, the PPV in the 5th octave according to the frequency of the acoustic signal is 100 %, and the NPV is 98.6 %.
Table 5 shows the results of the evaluation of the coefficients of the model for detecting pneumonia from the amplitude of the acoustic signal using the Trembita-Corona acoustic monitoring device.
It was found that at the Pborder, the sensitivity of the me–thod for detecting pneumonia based on the amplitude of the acoustic signal in the 4th, 5th and 6th octaves is 88.9, 99.9 and 86.9 %, respectively, and the specificity is 100 % in the 4th and 5th octave and 87.4 % in the 6th octave. Thus, the diagnosis of pneumonia based on the amplitude of the acoustic signal can best be made in the 4th, 5th and 6th octaves, respectively, which is presented in Table 6.
Taking into account the prevalence of the disease in 3.5 %, the PPV in the 4th octave according to the amplitude of the acoustic signal is 100 %, and the NPV is 99.6 %; the PPV in the 5th octave according to the amplitude of the acoustic signal is 100 %, and the NPV is also 100 %.
Therefore, the acoustic monitoring device Trembita-Corona demonstrates high specificity, sensitivity, PPV and NPV, especially in the 4th, 5th and 6th octaves by the average signal power, in the 3rd and 5th octaves by the frequency of the acoustic signal and in the 4th, 5th and 6th octaves by the amplitude of the acoustic signal.

Conclusions

1. The sensitivity of the method for detecting pneumonia using the acoustic monitoring device Trembita-Corona according to the average signal power in the 4th, 5th and 6th octaves is from 90.8 to 99 %, and the specificity is from 87.4 to 100 %. The sensitivity of the method for detecting pneumonia using the acoustic monitoring device Trembita-Corona according to the frequency of the acoustic signal in the 3rd and 5th octaves is from 63.2 to 63.2 %, and the specificity is 63.2 % in the 3rd octave and 100 % in the 5th octave. Thus, the diagnosis of pneumonia based on the frequency of the acoustic signal can best be made in the 3rd and 5th octaves. The sensiti–vity of the method for detecting pneumonia using the acoustic mo–nitoring device Trembita-Corona according to the amplitude of the acoustic signal in the 4th, 5th and 6th octaves is 88.9, 99.9 and 86.9 %, respectively, and the specificity is 100 % in the 4th and 5th octave and 87.4 % in the 6th octave. Thus, the diagnosis of pneumonia based on the amplitude of the acoustic signal can best be made in the 4th, 5th and 6th octaves.
2. The acoustic monitoring device Trembita-Corona demonstrates high specificity, sensitivity, positive predictive value and negative predictive value at given prevalence. We can trace this particularly well by the average signal power in the 4th, 5th and 6th octaves, by the frequency of the acoustic signal in the 3rd, 5th octaves, by the amplitude of the acoustic signal in the 4th, 5th and 6th octaves.
Prospectives for further research. 1. The Trembita-Corona acoustic monitoring device is a new and promising acoustic method for diagnosing pneumonia. 2. There is a need to use the Trembita-Corona acoustic monitoring device to investigate the sensitivity and specificity of the me–thod by the average signal power, the frequency and amplitude of the acoustic signal in third octaves.
 
Received 03.04.2023
Revised 26.04.2023
Accepted 01.05.2023

Список литературы

  1. Le Roux D.M., Zar H.J. Community-acquired pneumonia in children — a changing spectrum of disease. Pediatr. Radiol. 2017. 47(11). 1392-1398. doi: 10.1007/s00247-017-3827-8.
  2. Yun K.W., Wallihan R., Juergensen A., Mejias A., Ramilo O. Community-Acquired Pneumonia in Children. Myths and Facts. Am. J. Perinatol. 2019. 36(2). 54-S57. doi: 10.1055/s-0039-1691801.
  3. De Benedictis F.M., Kerem E., Chang A.B., Colin A.A., Zar H.J., Bush A. Complicated pneumonia in children. Lancet. 2020. 396(10253). 786-798. doi: 10.1016/S0140-6736(20)31550-6.
  4. Nascimento-Carvalho C.M. Community-acquired pneumonia among children: the latest evidence for an updated management. J. Pediatr. 2020. 96(1). 29-38. doi: 10.1016/j.jped.2019.08.003.
  5. Order of the Ministry of Health of Ukraine No. 1380 dated 02.08.2022. Standards of medical care “Community-acquired pneumonia in children” (in Ukrainian).
  6. Clinical guideline based on evidence “Pneumonia in children” dated 02.08.2022 (in Ukrainian).
  7. Isaienko V., Kharchenko V., Astanin V., Shchegel G., Olefir V. et al. System for acoustic diagnostics and symptomatic assistance to –COVID-19 patients for use in extremal conditions “TREMBITA-CORONА NAU”. Advances in Aerospace Technology. 2020. 1(82). 58-63. doi: 10.18372/2306-1472.82.14612.
  8. Marushko Y.V., Khomych O.V., Hyshchak T.V., Tarynska O.L., Shchegel G.O. State of the art usage of radiological and acoustic methods for the diagnosis of pneumonia, including those caused by SARS-COV-2 virus, in children. Review. Medical Science of Ukraine. 2021. 17(2). 114-124. doi: 10.32345/2664-4738.2.2021.16 (in Ukrainian).
  9. Pavlenko P.M., Marushko Yu.V., Olefir O.I., Khomych O.V., Shchehel H.O. et al. Patent of Ukraine for utility model No. 148836. Biul. No. 38. Date of publication 09/22/2021. Acoustic monitoring device with axial directional diagram (in Ukrainian).
  10. Marushko Y.V., Khomych O.V. Frequency characteristics of acoustic features of sound signals in the lungs of children with pneumonia using a new acoustic diagnostic device “Trembita-Corona”. Neonatology, Surgery, Perinatal Medicine. 2022. 4(46). 59-66. doi: 10.24061/2413-4260.XIІ.4.46.2022.10.
  11. Marushko Yu.V., Khomych O.V. Characterization of the ave–rage power, frequency and amplitude of acoustic signal peaks over the lungs in children with community-acquired pneumonia using the new device “TREMBITA-CORONA”. Medical Science of Ukraine. 2023. 19(1). 53-69. doi: 10.32345/2664-4738.1.2023.08.
  12. Hurianov V.H., Liakh Yu.Ie., Parii V.D. Handbook of biostatistics. Analysis of medical research results in EZR packages (R-statistics). Study guide. 2018. 145-175 (in Ukrainian).

Вернуться к номеру