Shafazand and Gould developed a model for predicting mediastinal metastases from a non-small cell lung cancer (NSCLC). This can help to guide further testing. The authors are from the University of Miami, Stanford University and the VA Palo Alto Health Care System.
Patient selection: non-small cell lung cancer (NSCLC)
Outcome: mediastinal metastases
Parameters:
(1) histologic type
(2) mediastinal metastases suspected on chest imaging
(3) location of tumor
(4) symptoms from primary tumor
(5) tumor diameter in centimeters
(6) age in years
Parameter
Finding
beta-coefficient
histologic type
adenocarcinoma or large cell
0.955
squamous
0
mediastinal metastases
no
0
yes
0.876
location
central (no air between the mediastinum and tumor)
0.749
not central
0
symptomatic
no
0
yes
0.485
diameter
<= 3.6 cm
0
> 3.6 cm
0.435
age of the patient
< 65 years
0
>= 65
-0.408
where
• The beta-coefficients were derived as LN(odds ratios) in Table 2.
• The entry for age in Table 2 is < 65 years, so beta-coefficient should be for >= 65 years.
X =
= SUM(points for all 6 parameters) - 1.806
probability of mediastinal metastases =
= 1/(1+EXP((-1)*X))
Performance:
• The area under the ROC curve is 0.70.
To read more or access our algorithms and calculators, please log in or register.
