Evaluation of Lung Nodule and Lung Cancer Detection With Artificial Intelligence Assisted Computed Tomography Among People Living in North China: a Prospective Single-arm Multicentre Study of Screening
试验速览
- 阶段
- 不适用
- 入组人数
- 5,000
- 主要终点
- Detection rate of lung nodule
研究概览
简要总结
Lung cancer is one of the leading cause of cancer related death in China. Lung cancer screening with low-dose computed tomography was considered as a better approach than radiography. However, the role of Lung cancer screening with Low-dose CT (LDCT) among Chinese people remains unclear. With rapid development of artificial intelligence (AI),the application of AI in detection and diagnosis of diseases has become research focus. Moreover, patients' psychological status also plays an important role in diagnosis and treatment.
This study focuses on detection and natural history management of lung nodule and lung cancer with AI assisted chest CT among people living in North China, and aims to investigate epidemiological results, patients' medical records and social psychological status.
详细描述
Lu'an Municipal Hospital and North China Petroleum Bureau General Hospital initialed the lung cancer screening by LDCT a few years ago. People living in North China who are administrated by these hospitals routinely took a chest CT every year. This study is to the best of our knowledge the first one designed to combine lung nodule and lung cancer screening with the application of artificial intelligence in China.
Methods: Firstly, the study acquires epidemiological, medical information and psychological status of people recruited, and investigates the data acquired from past several years of CT scans using AI to develop a model for lung nodule detection. Secondly, evaluating the performance of models and apply it to analyse the CT scans from the North China population recruited. Thirdly, improving the model and adding function for lung nodule prediction of natural history and probability of malignancy.
Aims: To depict the epidemiological results about the incidence of lung nodules and lung cancer in North China population; To evaluate association between people 's epidemiological, medical and psychological profiles and incidence, diagnosis and treatment of lung nodule; To develop an artificial intelligence assisted lung nodule diagnosis and management software to assist strategies of CT screening.
研究设计
- 研究类型
- Observational
- 观察模型
- Cohort
- 时间视角
- Prospective
入排标准
- 年龄范围
- 40 Years 至 —（Adult, Older Adult）
- 性别
- All
- 接受健康志愿者
- 否
入选标准
- •Aged 40 years or older
- •Routinely conducting chest CT scan at a low-dose setting (120kVp, 40-80mA, slice thickness of 1.25 mm or less) yearly in Lu'an Municipal Hospital and North China Petroleum Bureau General Hospital in at least the past 4 years up to December 2017, willing to continue routine yearly LDCT scan.
- •Chest CT data are available for DICOM format.
- •Signed Informed Consent Form.
排除标准
- •Pregnant woman and the disabled
- •Past thoracic surgery history, except for diagnostic thoracoscopy
- •Poor physical status without sufficient respiratory reserve to undergo lobectomy if necessary
- •Shortened life expectancy less than 10 years
- •Malignant tumor history within the past 5 years, except for the following conditions: cured skin basal cell carcinoma, superficial bladder carcinoma. and uterine cervix cancer in situ.
- •Past history of interstitial lung disease, pulmonary bulla and lung tuberculosis.
- •Other circumstances which is deemed inappropriate for enrollment by the researchers.
结局指标
主要结局
Detection rate of lung nodule
时间窗: 3 months
Study participants undergo baseline LDCT. Images are reviewed via AI software independently to identify lung nodules with diameters greater than 4mm. The software is developed by our computer technology collaborator. A radiologist then reviews the images, reports lung nodules with diameters greater than 4mm and any other abnormalities. The radiologist's findings will be conveyed to the study participants or their primary care physicians within 3 weeks. The process was conducted via double-blind method and detection rates of AI and radiologist will be recorded respectively. Unit of measurement: Percentage (number of participants with detected lung nodules over the total number of participants).
Profile of detected lung nodule
时间窗: 3 months
All lung nodules detected will be classified as 4 classes by the density and composition of nodule: 1. pure ground-glass nodule (pGGN); 2. part-solid nodule; 3. solid nodule; 4. uncertain nodule. The number and proportion of each class and the diameter and location of each nodule will be recorded. Unit of measurement: Percentage (number of nodules in each class over the total number of nodules); Numerical value (average value±standard deviation of nodules in each class); Percentage (number of nodules in each lobe over the total number of nodules).
次要结局
- Life quality and health status(3 months)
- Sensitivity in the detection of clinically actionable lung nodules(3 months)
- Growth of lung nodule(3 years)
- Lung cancer detection rate(3 months)
- Anxiety and depression level(3 months)
研究者
Jun Wang
Principal Investigator, Clinical Professor
Peking University People's Hospital