medical decision support medical decision support
Sign In

Probability of Cancer in Pulmonary Nodule on First Screening CT

Purpose: To use the full model of McWilliams et al to predict cancer in a nonspiculated mass seen on a first screening CT.

Description

McWilliams et al developed models for predicting cancer in a pulmonary nodule seen on a first screening CT. Two models are for a nonspiculated mass, one parsimonious and one a full model. The authors are from multiple institutions in Canada, the United States and Argentina.


Data modifications:

(1) age: 62 is subtracted from the actual age

(2) nodule size: 4 mm is subtracted and then transformed (below)

(3) number of nodules: 4 is subtracted

 

nodule size transformation =

= (((nodule size after modified) / 10)^(-0.5)) - 1.58113883

 

Parameters for parsimonious model:

(1) sex

(2) nodule size

(3) nodule location

(4) age

(5) family history of lung cancer

(6) emphysema

(7) nodule type

(8) number of nodules

 

Parameter

Findings

Points

sex

male

0

 

female

0.5635

nodule size

 

-5.6693 * (size)

nodule location

upper

0.7116

 

middle or lower lobe

0

age

 

0.0321 * (age)

family history

no

0

 

yes

0.3013

emphysema

no

0

 

yes

0.3462

nodule type

solid

0

 

part solid

0.3395

 

nonsolid

-0.3055

 

wtih ground glass opacity

-0.3055

number nodules

 

-0.0803 * (number)

 

value of X =

= SUM(points for the parameters)-6.8071

 

probability of cancer =

= 1 / (1 + EXP((-1) * X))


To read more or access our algorithms and calculators, please log in or register.

Specialty

Objective

ICD-10