Purpose: Breast cancers are three dimensional solids but very few are
spherical. We hypothesized that calculations based on the greatest dia
meter would not accurately reflect tumor volume and that three dimensi
onal measurements would affect tumor staging. Materials and methods: 1
65 invasive carcinomas measuring 2.5 cm or less and having three measu
red diameters (a greater than or equal to b greater than or equal to c
) noted were evaluated. Tumor volume was calculated using four geometr
ic models: the spherical 4/3 pi (a/2)(3), prolate spheroid 4/3 pi (a/2
) (c/2)(2), oblate spheroid 4/3 pi (a/2)(2) (b/2), and ellipsoid 4/3 p
i (a/2 x b/2 x c/2. The ellipsoid correctly determined the volume for
any tumor shape. All cases were stratified according to the TNM stagin
g system. Differences in mean volume calculated as a sphere and ellips
oid for each tumor subclass were analyzed using Student's T test. The
reclassification of tumors by the ellipsoid formula was determined. Re
sults: Seventy-six (46.1%) had tumors with three different diameters w
hile only six (3.6%) were true spheres having three identical diameter
s, Mean tumor volume analysis of T-1a, T-1b, T-1c, and T-2 tumors demo
nstrated a statistically significant overestimation of volume when uti
lizing the sphere formula instead of the ellipsoid formula (p < 0.05).
The differences in volume were more dramatic as the diameters increas
ed. A total of 41 tumors were moved into smaller T subclasses includin
g 10 node positive patients. Conclusions: Tumor volume analysis demons
trates that use of only the greatest diameter poorly reflects the true
volume of a lesion and consistently overestimates volume. The ellipso
id formula accurately calculates volume for these three dimensional tu
mors and when utilized has significant relevance to staging small inva
sive breast cancers.