Most histological classifications of lung tumors assess information derived
from 2-dimensional surfaces. However, tumors grow in 3 dimensions in a var
iety of architectural patterns that possibly reflect the histological type
or progression status of the tumor. A new mathematically derived variable (
Rratio) is presented, that assesses the overall 3-dimensional shape of a tu
mor and classifies tumors based on their sphericality. Analysis and compari
son of groups of spherical and nonspherical tumors fi-om mouse inbred strai
ns and recombinant congenic strains with a strict set of predetermined hist
ological criteria shows that Rratio is significantly associated with tumor
heterogeneity and that histological differences exist between the spherical
and nonspherical tumors. These results suggest that Rratio may be a valuab
le tool for the characterization of tumors and can be used to explore quali
tative differences in tumors derived from segregating crosses.