A model for lung cancer risk prediction that combines deep learning features from the Sybil model with clinical and epidemiological factors.
First, you need to process a low-dose CT image of the subject to be analyzed using the Sybil model. Then, record the resulting 6-year lung cancer risk prediction value and use it as input in the program below.
To run Sybil-Epi, download the sybil_epi.py file from this repository and run it as indicated below:
python sybil_epi.py --risk_sybil_6_year 0.123949024 --age 63 --bmi 28.88 --copd 0 --education 6 --ethnicity "White" --family_history 0 --personal_history 0 --smoking_duration 40 --smoking_intensity 1.0 --smoking_quit 40 --smoking_status 0
The subject used in the example above presents the following factor values:
| Factor | Value |
|---|---|
| 6-year Risk Sybil | 0.123949024 |
| Age (years) | 63 |
| BMI (kg/m2) | 28.88 |
| COPD (0-yes, 1-no) | 0 |
| Education level | 6 |
| Ethnicity | White |
| Family lung cancer history (0-yes, 1-no) | 0 |
| Personal cancer history (0-yes, 1-no) | 0 |
| Smoking duration (years) | 40 |
| Smoking intensity (cigarrettes per day) | 1.0 |
| Smoking quit time (years) | 40 |
| Smoking status (0-former, 1-current) | 0 |
Further details on how to use sybil_epi.py can be obtained with the command
python sybil_epi.py -h