This research is being done to determine if an image-based deep learning model (Sybil) can accurately predict the likelihood of future lung cancer based on chest computed tomography (CT) imaging from individuals.
This non-therapeutic study will enroll individuals who have family history of lung cancer. Participants will undergo a low-dose non-contrast computed tomography of the chest (LDCT) and may also send images from any chest CT scan(s) obtained as part of routine clinical care, outside of the study. The images and data collected will be analyzed by an image-based deep learning model (Sybil). Sybil is a type of artificial intelligence model that has been shown to accurately predict individuals' future risk of lung cancer based solely on images from a CT Chest scan, but it remains unclear whether Sybil works well in people with a family history of lung cancer. The goals of this study are: 1) to obtain CT Chest images from individuals with a family history of lung cancer in order to test whether Sybil continues to work well, and 2) offer free screening CT scans to qualifying individuals. It is expected that 250 people will take part in this research study.
Study Type
INTERVENTIONAL
Allocation
NA
Purpose
SCREENING
Masking
NONE
Enrollment
250
Massachusetts General Hospital
Boston, Massachusetts, United States
Sybil's performance in predicting future lung cancer diagnoses
All subjects will be followed for lung cancer diagnosis scan for up to 5 years following the baseline scan. Sybil's performance in predicting future lung cancer diagnoses across the study population will be calculated using the area under the receiver operating curve (AUROC), which is a measure of a risk prediction model's ability to discriminate between cases and controls. Sybil's output corresponds to the cumulative annual risk of lung cancer for up to 6 years following a given scan.
Time frame: Annually, from time of initial CT scan to up to 5 years after the scan.
Compare the distribution of Sybil lung cancer risk scores in this trial to the distribution of Sybil risk scores from the NLST clinical trial
Investigators will compare the distribution of Sybil scores (ranging from 0-1) from participants in this study with the distribution of Sybil scores from historical data from participants in the National Lung Screening Trial.
Time frame: Initial provided CT scan will represent time 0. Additional provided CT scans will vary between individuals and will be measured in years relative to time 0 (e.g., time -3.5 years, time +2 years, etc). Sybil risk scores will be calculated for each scan.
Incidence and prevalence of lung cancer in the study population
Investigators will estimate the incidence and prevalence of lung cancer in the LEGACY population. Incidence will be reported per person per year. Prevalence will be reported separately as a measure over the 5-year study follow up period.
Time frame: Annually, from time of initial CT scan to up to 5 years after the scan.
Incidence of lung nodules in this population
Investigators will estimate the incidence of lung nodules in the LEGACY population. Incidence will be measured per person per year.
Time frame: Annually, from time of initial CT scan to up to 5 years after the scan.
This platform is for informational purposes only and does not constitute medical advice. Always consult a qualified healthcare professional.
Prevalence of lung nodules in this population
Investigators will estimate the prevalence of lung nodules in the LEGACY population. This will be measured over the 5-year study follow up period.
Time frame: Annually, from time of initial CT scan to up to 5 years after the scan.
Describe the characteristics of lung nodules in this population
Investigators will describe the characteristics of lung nodules in the study population, including but not limited to size, location, and attenuation.
Time frame: At time of each provided CT scan to up to 5 years after the scan.