Skip to content

An improved model for lung cancer risk prediction that combines deep learning features from the Sybil model with clinical and epidemiological factors

License

Notifications You must be signed in to change notification settings

hung-lab/Sybil-Epi

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

53 Commits
 
 
 
 
 
 
 
 

Repository files navigation

Sybil-Epi

A model for lung cancer risk prediction that combines deep learning features from the Sybil model with clinical and epidemiological factors.

How to use it

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 --age 66.08055556 --bmi 29.64582054 --copd 0 --education 6 --ethnicity White --family_history 0 --personal_history 1 --smoking_duration 43 --smoking_intensity 0.8 --smoking_quit 0 --smoking_status 0 --risk_sybil_6_year 0.034103291

The subject used in the example above presents the following factor values1:

Factor Value
Age (years) 66.08055556
BMI (kg/m2) 29.64582054
COPD (0-no, 1-yes) 0
Education level2 6
Ethnicity White
Family lung cancer history (0-no, 1-yes) 0
Personal cancer history (0-no, 1-yes) 1
Smoking duration (years) 43
Smoking intensity (cigarrettes per day) 0.8
Smoking quit time (years) 0
Smoking status (0-former, 1-current) 0
6-year Risk Sybil3 0.034103291

Further details on how to use sybil_epi.py can be obtained with the command python sybil_epi.py -h

1All factors were measured using the units indicated in the PLCOm2012 model.

2Education was measured in six ordinal levels: less than high-school graduate (level 1), high-school graduate (level 2), some training after high school (level 3), some college (level 4), college graduate (level 5), and postgraduate or professional degree (level 6).

3The 6-year Risk Sybil value can be calculated from a single low-dose CT image, analyzed using the Sybil model.

About

An improved model for lung cancer risk prediction that combines deep learning features from the Sybil model with clinical and epidemiological factors

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Languages