We employed factor analysis to quantify the degree of histologic heter
ogeneity of childhood infratentorial neuroglial tumors. Our data were
26 reliably ascertained histologic features in 1068 children in the Ch
ildhood Brain Tumor Consortium database. The factor analysis identifie
d five uncorrelated quantitative ''factors,'' each derived from a diff
erent linear combination of the 26 histologic features, that accounted
for much of the histologic variation. Histologic features differed in
their importance in each factor. The most important features in each
factor were used for naming using simple histologic, familiar descript
ive terms: Spongy, Proliferative, Ring Fibrillary and Nuclear. Each tu
mor has a score on each factor. Two-thirds of tumors had high scores f
or at least two factors, indicating frequent histologic heterogeneity
among these tumors. Ninety-five percent of tumors were allocated to 1
of 11 nonoverlapping histologically homogeneous groups. The Jive quant
itative factors complement standard qualitative taxonomies by making e
xplicit the histologic heterogeneity or homogeneity of individual tumo
rs and provide the pathologist with a method that takes advantage of m
ore of the histology of each tumor than conventional nomenclatures. Hi
stologically homogeneous groups of tumors are likely to be of value in
clinical trials and biologic research. Prognostic models based on the
se factors have been published.