Early warning and stratification of the elderly cardiopulmonary dysfunction-related diseases: multicentre prospective study protocol
Early warning and stratification of the elderly cardiopulmonary dysfunction-related diseases: multicentre prospective study protocol
In China, there is a lack of standardised clinical imaging databases for multidimensional evaluation of cardiopulmonary diseases. To address this gap, this study protocol launched a project to build a clinical imaging technology integration and a multicentre database for early warning and stratification of cardiopulmonary dysfunction in the elderly.
This study employs a cross-sectional design, enrolling over 6000 elderly participants from five regions across China to evaluate cardiopulmonary function and related diseases. Based on clinical criteria, participants are categorized into three groups: a healthy cardiopulmonary function group, a functional decrease group and an established cardiopulmonary diseases group. All subjects will undergo comprehensive assessments including chest CT scans, echocardiography, and laboratory examinations. Additionally, at least 50 subjects will undergo cardiopulmonary exercise testing (CPET). By leveraging artificial intelligence technology, multimodal data will be integrated to establish reference ranges for cardiopulmonary function in the elderly population, as well as to develop early-warning models and severity grading standard models.
The study has been approved by the local ethics committee of Shanghai Changzheng Hospital (approval number: 2022SL069A). All the participants will sign the informed consent. The results will be disseminated through peer-reviewed publications and conferences.
http://creativecommons.org/licenses/by-nc/4.0/
This is an open access article distributed in accordance with the Creative Commons Attribution Non Commercial (CC BY-NC 4.0) license, which permits others to distribute, remix, adapt, build upon this work non-commercially, and license their derivative works on different terms, provided the original work is properly cited, appropriate credit is given, any changes made indicated, and the use is non-commercial. See: http://creativecommons.org/licenses/by-nc/4.0/.
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.
According to the report of the WHO in 2020, cardiopulmonary dysfunction-related diseases represented by ischaemic heart disease and chronic obstructive pulmonary disease (COPD) rank among the top global causes of morbidity and mortality, making their prevention and control a key priority in chronic disease management worldwide.1 2 As an integrated evaluation method of cardiopulmonary function and metabolism, the cardiopulmonary exercise test (CPET) is the only means to comprehensively examine the overall function of cardiopulmonary metabolism from resting state to exercise state in clinical practice and has the characteristics of non-invasiveness, quantitative evaluation and high sensitivity.3 Its gas exchange parameters hold significant value for the diagnosis and treatment of patients with pulmonary hypertension, heart failure and COPD.4 5 However, the complexity and high cost, as well as the lack of diagnostic criteria for evaluating elderly cardiopulmonary function, limit its widespread application. The assessments of cardiopulmonary diseases in clinical settings mainly focus on subjective evaluations such as heart function grading and dyspnoea scales. Cardiac function assessment is mainly based on the subjective New York classification methods6. Although echocardiography is key for evaluating heart function, it can miss cases with preserved ejection fraction, and 50% of such patients go undiagnosed.7 In clinical practice for COPD, the main reliance is on pulmonary function test (PFT), while the value of imaging and other examinations in disease diagnosis and treatment has not yet received sufficient attention.8 However, PFT cannot accurately assess the location of airflow obstruction and hasn’t been widely used for the screening of COPD. A lot of individuals without significant clinical symptoms have been underdiagnosed.
Current assessments for cardiopulmonary dysfunction-related diseases suffer from insufficient sensitivity and unidimensional evaluation approaches. A comprehensive multidimensional assessment strategy is therefore required to holistically evaluate cardiopulmonary function. Previous studies have validated imaging’s critical role in the early detection of COPD and cardiovascular diseases.9 10 This study aims to establish a multidimensional framework—particularly leveraging CT imaging—for comprehensive diagnosis, prediction and therapeutic guidance of cardiopulmonary diseases in the elderly population. To achieve these objectives, we propose to: (1) develop a multicentre database integrating clinical and imaging data from elderly cohorts; (2) establish age-specific reference ranges for cardiopulmonary evaluation parameters; (3) construct Artifical Intelligence (AI)-driven predictive models for disease progression and severity classification and (4) implement a personalised health management model characterised by comprehensiveness, multidimensionality, standardisation, reproducibility, cost-effectiveness and AI dependency for elderly patients with cardiopulmonary dysfunction.
The study includes four major hospitals and an AI company in China, with a focus on the design at a leading hospital. Standardised procedures and a practical questionnaire are created for all hospitals. The same standardised lung function device is used across all participating hospitals, with comprehensive training provided to ensure consistent operation.
A cross-sectional survey of over 6000 elderly people (≥ 60 years) is conducted across five regions — Shanghai, Shenyang, Guangdong, Qingdao and Ningxia — representing four major regions of Northwest, Northeast, East and South China. These regions feature diverse climates, altitudes and ethnic compositions. Based on existing clinical diagnostic criteria like the Global Initiative for Chronic Obstructive Lung Disease (GOLD) and European Society of Cardiology guidelines.11–16 The surveyed population is divided into three cohorts: a healthy cardiopulmonary function cohort, a cardiopulmonary function decreased cohort and an established cardiopulmonary disease cohort. Low-dose chest CT scans, echocardiography, CPET and related laboratory examinations are performed on these populations. Using AI-assisted technology and multidimensional evaluation parameters, a comprehensive reference range for healthy cardiopulmonary function in the elderly will be established based on the baseline healthy cohort. For both the baseline healthy cohort and the cohort with decreased function, a follow-up observation of at least 3 years will be conducted to determine endpoint events, including cardiopulmonary function decline and cardiopulmonary diseases. Cardiopulmonary diseases in this study mainly refer to patients who, after excluding other circulatory and respiratory system diseases according to the International Classification of Diseases 11th Revision, are diagnosed with one or several of the following: COPD, coronary syndromes, pulmonary hypertension and heart failure. Based on these findings, three early-warning models will be constructed, encompassing: (1) progress from a healthy population to cardiopulmonary function decline; (2) progress from a healthy population to cardiopulmonary diseases and (3) progress from cardiopulmonary function decline to cardiopulmonary diseases. Additionally, for the baseline cohort diagnosed with cardiopulmonary diseases, AI will be used to develop a multidimensional severity grading standard. Furthermore, standardised technical operating procedures and corresponding quality control measures will be established. A multicentre structured database is constructed to support diversified big data analysis and AI-driven research on elderly cardiopulmonary dysfunction and related diseases. The flowchart is shown in figure 1.
The recruitment process commenced around January 2023, with various start dates across different centers and hospitals. Local community hospitals are engaged in face-to-face outreach to recruit elderly residents from the community, especially those with a history of smoking, heart disease, and difficulty breathing, who visit the clinics. Screenings have been initiated since August 2023. Follow-up assessments will begin in the third year from the initial baseline visit. At that time, diagnoses of COPD, CVD, and details regarding related treatments for all participants will be systematically gathered through the Hospital Information System of the principal hospital. This method ensures a thorough and organized collection of data, which is essential for gaining insights into the occurrence and management of these health conditions within the community.
Inclusion criteria for participants are: (1) community residents with or without chest symptoms recruited from four hospitals, (2) 60 years or older and (3) no history of lung cancer (self-reported), thoracic surgery and large area pulmonary infection. Inclusion criteria are assessed by a Community Health Service Centre (CHSC) doctor during a face-to-face interview using a questionnaire.
Questionnaire investigation
Before conducting this study, multidisciplinary experts were invited to the assessment of elderly cardiopulmonary function for multiple detailed discussions and designed a questionnaire survey suitable for this study. The main content of the questionnaire survey includes the following: basic information, lifestyle, comprehensive geriatric assessment, cardiopulmonary assessment and each participant signs an informed consent form.
Laboratory examinations
Based on laboratory tests related to elderly cardiopulmonary dysfunction and related diseases, the following relevant indicators will be collected, including complete blood count, lipid profile, electrolyte panel, liver and kidney function tests, blood glucose, erythrocyte sedimentation rate, procalcitonin and interleukin-6.
Chest CT examination and imaging evaluation
Preparation and scanning plan for chest CT examination
Before the examination, verify the examinee’s personal information. Additionally, train the examinee in end-inspiratory and end-expiratory breath-holding to ensure optimal image quality. All metallic objects on the chest surface should be removed, while lead shielding is required for radiation protection of non-irradiated regions. For enhanced examination, follow the guidelines for the use of iodinated contrast agents.17
CT scanning plan: to balance the accuracy of subjective qualitative evaluation and objective quantitative evaluation, and ensure the accuracy of subsequent quantitative CT evaluation, all centres adopt a consistent paired respiratory chest CT scanning.
Scans are performed using 16 or more layers of multislice spiral CT and opt for low-dose chest CT to minimise subjects' radiation exposure. Scanning range: from the entrance of the chest to the level of both adrenal glands. Low dose scanning mode: spiral scanning. Tube voltage: for CT scanners equipped with iterative reconstruction technology, scanning parameters of 100–120 kVp and below 30 mAs can be employed, using iterative reconstruction to enhance image quality. For CT systems without iterative reconstruction capability, scanning parameters of 120 kVp and 30–50 mAs should be used. Collimation: ≤1.25 mm. Reconstruction layer thickness: ≤1.25 mm; reconstruction interval: for layer thickness ≤0.625 mm, no interval reconstruction is applied; when the layer thickness is 0.625–1.25 mm, it is recommended to use 80% of the layer thickness for reconstruction interval; scan matrix: 512×512. Reconstruction algorithm or filter function: standard algorithm and lung algorithm reconstruction. Reconstruction mode: iterative reconstruction. Window setting: lung window level −600 to −650 Hounsfield units (HU), window width 1500 HU to 1600 HU; mediastinal window level: 40 to 50 HU, window width: 350 to 380 HU.
Chest CT imaging evaluation
Image analysis is mainly conducted from two aspects: subjective evaluation and objective quantitative evaluation, including emphysema, airway, pulmonary vessels and coronary artery calcification. In this study, the necessity of double reading will be tested in a pilot study involving 500 participants in the group. If consensus double reading shows no benefit in this study either, then one trained radiologist will continue to read the consecutive scans.
Use Picture Archiving and Communication System (PACS) system for image viewing, employing lung windows and combine with multiplanar reformation and minimum intensity projection for image reading. The subjective evaluation of CT images for COPD mainly includes the presence of pulmonary emphysema, types and distribution of emphysema and the morphology of airways and pulmonary vasculature. According to the Fleischner Society guidelines,18 emphysema is classified into three types: central lobular emphysema (CLE), panlobular emphysema and paraseptal emphysema. CLE is further divided into mild CLE, moderate CLE, confluent CLE and progressive-destructive types. Panlobular emphysema is extremely rare in the Chinese population. The severity of emphysema can be assessed using the subjective visual Goddard scoring method19 (table 1). By comparing paired respiratory CT images, the presence and extent of air trapping can be assessed. Since there is a clear correlation between imaging phenotypes and clinical staging of COPD, as well as therapeutic improvement, determining the imaging phenotype is particularly important. The commonly used imaging phenotypes include the following types: Type A, no emphysema or the presence of mild emphysema, regardless of whether there is associated bronchial wall thickening; Type E, the presence of significant emphysema without bronchial wall thickening and Type M, the presence of both significant emphysema and bronchial wall thickening.20 The changes in the pulmonary vasculature are mainly evaluated through the inspiratory CT images by assessing the diameter of the main pulmonary artery and its ratio to the diameter of the ascending aorta at the same level. If the main pulmonary artery diameter is >29 mm and the ratio of the main pulmonary artery to the ascending aorta diameter is >1, it suggests a high likelihood of secondary pulmonary hypertension.
Table 1
Emphysema index score (Goddard score)
Based on the evaluation of low-dose CT scan images, the calcification status of the coronary circulation is categorised into the following levels: (1) no calcification (None): no visible calcification. (2) Mild calcification (Mild): isolated small calcification spots are present in a segment of the coronary artery. (3) Moderate calcification (Moderate): the degree of calcification exceeds the description of mild but is not yet at the level of severe calcification. (4) Severe calcification (Heavy): there is continuous calcification in a segment of the coronary artery21 (table 2). Researchers have found that visual scoring has good consistency among different radiologists.21 Although visual scoring may not be as precise as quantitative methods, it provides a quick and easy assessment tool, especially suitable for initial screening and large-scale screening programmes. The above subjective evaluations are listed in figure 2.
Table 2
Coronary artery calcium
All centres use uniformly configured commercial software (Aview) for the quantitative analysis of CT images. The quantitative analysis of chest CT includes lung volume, emphysema, airways, pulmonary vasculature and coronary artery calcium scores. Initially, the software automatically segments the left and right lungs and five lung lobes; then, physicians review the segmentation results layer by layer. After the review is complete, quantitative evaluation is performed, outputting the average lung density, emphysema volume and proportion and regional differences in lung density for the entire lung, left lung, right lung or individual lobes. Finally, the presence, severity and distribution of emphysema are determined. On inspiratory chest CT, the percentage of lung tissue volume with CT values lower than −950 HU of the entire lung volume is defined as the proportion of low attenuation area (LAA%) or the emphysema index (EI). According to the Fleischner Society guidelines, an EI ≥6% is diagnostic of emphysema.18
The degree of small airway narrowing is often indirectly reflected by the extent of air trapping. End-expiratory CT is the best method for evaluating air trapping in COPD, and most studies quantitatively assess air trapping by evaluating the percentage of LAA% at thresholds of −856 HU or −850 HU (LAAexp-856 or LAAexp-850).
For paired respiratory CT images, after the segmentation of the left and right lungs is completed, quantitative analysis of air trapping can be performed. The relevant parameters include the inspiratory-to-expiratory lung volume ratio, inspiratory-to-expiratory lung attenuation ratio and the relative volume change of voxels with attenuation values between −860 HU and −950 HU at the level of the entire lung, left and right lungs and even the five lung lobes. Alternatively, a parameter response map (PRM) analysis can be conducted to evaluate the condition of functional small airway diseases. The specific parameters include the volumes and percentages of PRMnormal, PRMemphysema and function small airway disease of PRM (PRMfSAD).22 The Coronary Artery Calcification Score (CACS) represents a quantitative evaluation of calcification within the coronary arteries using CT. Previously, it has been verified that Coronary Artery Calcification (CAC) could be assessed with high reliability on chest CT with a high correlation between the CAC scores obtained from chest CT and those obtained from cardiac CT scans.23–27 To assess CACS, several methods are widely recognised, including the Agatston score, volumetric scoring and mass scoring. The Agatston score is calculated by multiplying a calcium density score, derived from the peak HU value within the plaque, by the area of calcification. The HU values are assigned points as follows: 1 point for HU values between 130 and 199, 2 points for 200–299 HU, 3 points for 300–399 HU and 4 points for HU values of 400 and above. This product is then calculated for each calcified area (in mm²), and the scores from all coronary arteries across the CT sections are aggregated to yield the total calcification score.28
CAC scores are typically categorised into various grades that reflect the severity of calcification and correlate with the risk of cardiovascular disease. These grades include no calcification (CAC=0), mild calcification (CAC 1–100), moderate calcification (CAC 101–300), severe calcification (CAC 301–1000), and extreme calcification (CAC>1000)28 (tables 3 and 4).
Table 3
Density factor based on the maximal CT number (HU) of the plaque
Table 4
Coronary artery calcium score (Agatston score)
The calcium volume score, a direct measure, is determined by multiplying the calcified area by the thickness of the CT slice, thus providing an estimate of the total calcified volume. Meanwhile, the calcium density score is calculated by dividing the Agatston score by the area of calcification, which indicates the average severity of the calcification.29 A higher calcium score reflects a more severe degree of coronary atherosclerosis and is associated with an increased risk of adverse vascular events, such as heart attacks and strokes. This study aims to use an improved evaluation method for calcification calculation based on non-ECG-gated chest CT.30 The above quantitative evaluations of the lung and cardiovascular systems are listed in figure 3.
Pulmonary function test
All 18 research centres will use uniformly configured lung function metres to perform PFTs. Doctors who perform lung function tests will receive professional training. Vital capacity, forced vital capacity (FVC), forced expiratory volume in one second (FEV1), the ratio of FEV1 to FVC, ratio of FEV1 to predicted value (FEV1%pred), ratio of residual volume to total lung capacity, peak expiratory flow and maximum ventilatory volume will be assessed. Spirometric parameters are crucial for the diagnosis of COPD, which is defined as a post-bronchodilator FEV1/FVC ratio greater than 0.7. Based on the percentage of FEV1 relative to the predicted value, the GOLD guidelines categorise the severity of COPD into four stages: GOLD Stage 1 (mild, FEV1 ≥80%), GOLD Stage 2 (moderate, 50%≤FEV1 < 80%), GOLD Stage 3 (severe, 30%≤FEV1 < 50%) and GOLD Stage 4 (very severe, FEV1<30%).
Cardiac ultrasonography and evaluation
Echocardiography, through the measurement of cardiac chamber size and volume, can assess whether the heart chambers are dilated or constricted. By measuring the valve flow velocity and blood flow direction, it is possible to determine whether there is valve stenosis or regurgitation. The left ventricular ejection fraction (LVEF) is a key indicator for assessing the heart’s pumping function and is crucial for diagnosing and monitoring diseases such as heart failure. An LVEF of less than 52% for males and less than 53% for females suggests abnormal left ventricular systolic function.31 LVEF ranges from 40% to 52% indicating a mild reduction, 30% to 40% indicates a moderate reduction and less than 30% indicates a severe reduction. Stroke volume (SV) and cardiac index reflect the heart’s systolic function and overall circulatory status. LVEF is the standard measure for assessing the systolic function of the left ventricle, representing the volume of blood ejected by the left ventricle with each heartbeat. The normal values for SV are 33–78 mL for males and 29–63 mL for females.31 Evaluation of left ventricular diastolic function includes Doppler echocardiographic indices such as E/e' and E/A.32 A tricuspid valve E/A ratio of less than 0.8 indicates impaired right ventricular relaxation; an E/A ratio between 0.8 and 2.1 with an E/e' greater than 6 or prominent diastolic flow in the hepatic veins, suggests moderate impairment of right ventricular diastolic function (pseudonormalisation). An E/A ratio greater than 2.1 with a deceleration time of less than 120 ms indicates a restrictive filling pattern of the right ventricle.
Cardiopulmonary exercise testing
The specific procedure includes initially ruling out contraindications for the test and obtaining detailed medical history and informed consent from the subject. Then, in a suitable testing environment, basic static pulmonary function measurements are conducted on the subject. Afterwards, an appropriate exercise device (treadmill or bicycle) is selected based on the subject’s condition, and the subject performs gradually increasing levels of exercise under supervision, while key data such as heart rate (HR), blood pressure, gas exchange metrics and ECG are recorded. During the test, the subject’s symptoms are closely monitored, and immediate action is taken in the event of adverse reactions. After exercise completion, subjects are interviewed about their termination rationale, with subsequent report generation derived from analytical data.33 Exercise tolerance and cardiovascular function can be reflected by parameters such as peak oxygen uptake (peak VO2), maximum oxygen pulse (peak VO2/HR), as well as ECG, heart rate and blood pressure changes during the test period. The ventilatory function is indicated by breathing reserve (BR), the dead space volume/tidal volume ratio (VD/VT) and breathing frequency (Bf). Lastly, gas exchange is reflected by the oxygen ventilation equivalent (VE/VO2), carbon dioxide ventilation equivalent (VE/VCO2), VD/VT, alveolar-arterial oxygen pressure difference (P(A-a)O2) and peripheral blood oxygen saturation (SaO2). In data analysis, four main indicators—peak VO2, anaerobic threshold (AT), BR and the slope of VE/VCO2—can help differentiate between pulmonary vascular and cardiovascular diseases. (1) If peak VO2 is normal and the patient subjectively feels exercise limitation: if VD/VT, P(A-a)O2 and P(a-ET)CO2 are abnormal, it indicates pulmonary/pulmonary vascular disease; otherwise, it suggests cardiovascular disease. (2) If peak VO2 and AT are reduced with normal or increased BR, an abnormal ECG indicates myocardial ischaemia. However, if the ECG is normal, the reduction may be due to insufficient effort or muscle disease. On the other hand, if BR is decreased, it indicates lung disease. CPET can assess the prognosis and diagnosis of heart failure patients, those diagnosed or suspected with hypertrophic cardiomyopathy, confirmed pulmonary arterial hypertension/secondary pulmonary hypertension, COPD or interstitial lung disease, categorising them into low risk, medium risk, medium-high risk and high risk.34
The detailed contents of the questionnaire and other examination items are already listed in table 5. All participants are required to undergo the aforementioned examinations and complete the questionnaire survey, provided that their conditions permit. Regarding CPET, due to equipment constraints and limited operator expertise, not all participants will undergo the test, but no fewer than 50 cases will be included.
Table 5
Assessment in baseline and follow-up
The study will establish a standardised disease-specific database integrating clinical, imaging and CPET data, ensuring data quality through open Electronic Health Record (EHR) dual-layer modelling. In the cross-sectional study, multivariate statistical analysis will be performed on quantitative parameters derived from chest CT and CPET results. Generalised Additive Models for Location, Scale and Shape will be employed to construct age-specific multidimensional reference intervals for cardiopulmonary function in the elderly population. Using the Aview software derived from deep learning architectures for quantitative computed tomography (QCT) parameter extraction, and applying machine learning (ML) to over 6000 samples, an early warning model will be developed and trained using a 7:2:1 split: 70% of samples as the training set (approximately 3500 cases), 20% as the internal validation set (approximately 1000 cases for hyperparameter tuning) and 10% as the independent test set (approximately 500 cases for assessing generalisation capability). Additionally, approximately 1000 cases from one centre are reserved for external validation to evaluate generalisation performance. Furthermore, if conditions permit, multicentre validation will be used during implementation. Decision curve analysis will be incorporated to assess clinical utility. Among confirmed patients, the same methodology will be applied to develop a ML-based severity stratification model for classifying disease progression. Ultimately, through the integration of the database platform, Aview software and ML models, automated feature extraction, quantitative stratification and predictive risk assessment will be achieved, constructing a comprehensive clinical decision support system for screening and long-term management. All statistical analyses will be conducted using Python, R and SPSS.
Ethical approval for this study was obtained from the Research Ethics Committee of Shanghai Changzheng Hospital (Approval No. 2022SL069A). Prior to any evaluations, participants will receive comprehensive information regarding the study objectives and research procedures. Written informed consent will be secured from all participants, authorising their participation and data usage. On study completion, each participant will receive structured diagnostic reports including chest CT scans, echocardiograms, PFT and laboratory results, along with comprehensive cardiopulmonary assessment reports complemented by personalised health management consultations and evidence-based intervention plans. Furthermore, consistent with the study’s primary objectives, peer-reviewed scientific evidence will be established regarding reference ranges for cardiopulmonary function in elderly populations, early-warning indicators of functional decline and an AI-based integrated diagnostic and stratification pathway for cardiopulmonary diseases. These findings will enable public policymakers and healthcare organisations to develop novel clinical protocols, update professional guidelines and ultimately enhance the quality of life for elderly individuals with cardiopulmonary conditions.
All data are obtained through various examinations and surveys and is imported into the database after collection, hence the database is updated in real-time. Researchers with different levels of authorisation have varying access rights to the database, with those holding higher permissions able to access all content within the database.
To ensure the quality of the collected data, all hospitals will take multiple quality control measures: (1) professional technical personnel from the hospital and project management team will be responsible for monitoring the quality of CT image acquisition technology; (2) researchers will receive formal training to ensure the quality of image analysis and (3) all four research centres use a standardised data management software (Specialised Disease Library Research Platform System Software), and the input data will be repeatedly checked by senior researchers to ensure quality.
All personal identification information of participants (including name, participant number and date of birth) is removed from the dataset. Data for the study is recorded by the radiologist who is on duty for data entry. In accordance with the Cybersecurity Law in China, the source data have to be stored in the leading hospital, and the CT images for the research are separately stored in the Specialised Disease Library research platform system software in the hospital.
Given the technical nature of the research procedures, such as CT scans, echocardiograms and CPET, the extent of direct patient and public involvement in the study was limited. However, the study is conducted with respect for the dignity and autonomy of all participants. Physicians from the CHSC, who are in regular contact with the elderly community, play a crucial role in the initial assessment and informed consent process. They help to explain the study in a manner that was accessible and comprehensible to the participants. The questionnaire, being relatively detailed and complex, requires the patience and cooperation of the patients. Their responses provide valuable insights that have been instrumental to our research.
This study will construct a large database of elderly people’s heart and lung function, determine the normal range of cardiopulmonary function in different age groups, warning standards and severity grading standards for cardiopulmonary diseases and form an integrated and popularised evaluation system for cardiopulmonary function.
Current cardiopulmonary function screening in the elderly primarily includes medical history collection, static cardiopulmonary function assessments, general examinations and psychological evaluations.21 35 While ECGs cannot directly identify organic cardiovascular lesions, echocardiographic assessment of cardiac function depends on endocardial clarity. Although CPET provides comprehensive functional evaluation, its utility in critical risk management remains limited.36 Additionally, although PFT is regarded as the gold standard for pulmonary function evaluation, it is unable to differentiate between the pathological types of airflow obstruction, emphysema and small airway lesions.37 Visual CT (conventional CT) has been widely used in clinical practice due to its comprehensive anatomical information and early detection of abnormal lesions, but it has limitations including a lack of further functional information and the potential misunderstanding of some CT features that may reflect normal changes. In the elderly population, some CT features like subpleural reticular patterns, bronchial dilation and bronchial wall thickness are likely to be the normal spectrum of lung ageing.38 39 Therefore, it is necessary to integrate valuable examinations to establish comprehensive evaluation standards, while also mining more information like quantitative evaluation.
The normal range of visual and quantitative CT parameters for emphysema and air trapping has been determined in a cohort of young male subjects40. Furthermore, there was a study about normal regional CT lung density in children under 5 years old using ventilation CT techniques.41 The reference range for quantitative CT values in healthy elderly individuals is still not established, yet CT has the potential to define it. In our study, elderly participants can even have their heart and lung functions assessed in a single chest CT scan. Cardiac function is evaluated through measurements of large vessels and coronary artery calcium scores, while lung function is assessed by looking at emphysema, airways, vessels and PRM. Using cardiopulmonary CT for integrated assessment can boost efficiency and cut healthcare costs.
AI is increasingly used to transform the diagnosis, treatment, risk prediction, clinical care and prognosis of cardiopulmonary decline and related diseases in the elderly. Previous studies have explored radiomic and quantitative features from CT, X-ray and MRI using ML to evaluate and predict cardiopulmonary-related diseases such as heart failure, coronary heart disease, acute respiratory failure, asthma and COPD.42–47 Hamatani et al.48 used ML to predict incident heart failure in patients with atrial fibrillation. Recently, deep learning techniques have been applied to medical images for segmentation, lesion detection and disease classification.44 49–51 In this study, ML is used to construct an intelligent diagnostic model for elderly patients with cardiopulmonary dysfunction and related diseases.
Due to insufficient understanding of chronic diseases such as COPD, many people rarely undergo PFT before clinical symptoms appear.52 Additionally, not all primary healthcare facilities are equipped with spirometry devices, and the utilisation rate of these devices in many community hospitals is insufficient.53 Hence, in most health economies, similar screenings for COPD are not widely implemented.54 Fortunately, with increased awareness of pulmonary nodules, the usage of screening chest CT is expanding in the general population.55 The study attempts to conduct a comprehensive assessment of cardiopulmonary status in the elderly and detect early signs of cardiopulmonary dysfunction. By establishing a standardised assessment system to ensure consistency and comparability of assessment results across different regions. Finally, by using AI to improve the accuracy and efficiency of assessments.42 45 Intelligent diagnostic systems can quickly identify abnormal indicators, support clinical decision-making and potentially uncover new biomarkers and disease patterns.49 51 The design of this technological system has taken into account simplicity and cost-effectiveness, making it suitable for widespread use in primary healthcare facilities. In addition, through regular assessments and long-term follow-up, changes in cardiopulmonary function can be monitored promptly, offering personalised health management and intervention plans.
There are several limitations to this study: (1) the complexity of CPET equipment and procedures limits its use in large-scale studies, resulting in a small sample size of only 500 participants; (2) during the follow-up process, there may be lost follow-up, so relevant measures are developed to reduce the lost follow-up rate and (3) multicentre research may not be completely consistent in terms of machine models, but scanning plans for different models have been developed to avoid a decrease in image quality due to this.
In summary, the study aims to establish a standardised AI-based database by conducting cardiopulmonary function assessments on elderly populations across multiple regions of China, from epidemiological surveys to cohort studies—which will support China’s continuous healthcare management and proactive health services.
Not applicable.
