Lung Cancer Screening With Low-dose CT Scan in Women : Implementation Study (CASCADE)

The purpose of this study is to determine whether reading of low-dose thoracic CT scans can be done by a single general radiologist who has been trained to lung cancer screening, and will evaluate the performance in comparison with double reading by experts. The study will enroll women between 50 and 74 years old, at risk for lung cancer due to their smoking history.

Lung cancer is the leading cause of cancer death, worldwide. Several randomized studies have demonstrated that annual or biennial low-dose CT screening reduces lung cancer mortality. However, these studies involved expert chest radiologists, with double reading being performed in most studies. Furthermore, none of the published studies have evaluated the role of artificial intelligence to serve as second of concurrent reader. Women with at least 25 pack-year smoking history who quitted smoking less than 15 years ago will be enrolled to have baseline, 1-year and 2-year low-dose CT of the chest. The CT scans will be read on site by a general radiologist trained to lung cancer screening according to the European lung cancer screening certification program, first without then with the aid of an artificial algorithm trained to lung nodule detection (Veye Chest, Aidence). All CT scans will also be read by 2 chest experts, who will resolve their disagreement by a consensus reading if necessary. Patient management will rely on the double reading by expert. The criteria for positive screen result are as follows: solid nodule > 500 mm3 (10 mm) or growing (30% volume increase), part-solid nodule with > 8 mm solid component or new or growing, pure ground glass nodule developing a solid portion.

lung neoplasms, multidetector computed tomography, early detection of cancer, artificial intelligence, lung cancer screening

Proportion of participants with a positive screening test and proportion of confirmed diagnosis of cancer

Analysis of the diagnostic performance of the reading without detection software, in order to assess its incremental value

Sensitivity, specificity, predictive values and likelihood ratios of the initial reading compared to the expert readings, using histological diagnosis as gold standard for positive screens and stability at 2 years for negative screens

Number of participants related to the number of eligible women, having completed all the required scans (3 to 6), speed of inclusion in the study

HADS (Hospital Anxiety and Depression Scale) questionnaire at each scan, Cancer worry scale, Satisfaction with Decision scale at inclusion and end of study, all translated into French

Number of participants related to the number of women included in the study for whom treatment is initiated (bronchodilators/ statins or revascularization/ osteoporosis treatment)

Presence of at least one thoracic vertebral fracture and measurement of trabecular attenuation of the T8 vertebral body

Inclusion Criteria: Women aged from 50 to 74 years who had smoked at least 20 pack years and quit less than 15 years ago Had given their consent and accepted the need for a 2-year follow-up Affiliated to the social security system Exclusion Criteria Presence of clinical symptoms suggesting malignancy (weight loss, hemoptysis) or ongoing infection (cough with fever) Evolving cancer History of lung cancer A 2-year follow-up not possible Chest CT scan performed within 2 years prior to inclusion