We designed software for measuring the volume of cerebral arteriovenous mal
formations from angiography and validated it against prescription volumes i
n radiosurgery. We aimed to create a model for the risk for complications a
s a function of volume, based on established outcome prediction models for
Gamma Knife(R) radiosurgery, but without the need for dose planning. We cre
ated an application for computing the volume of cerebral arteriovenous malf
ormations from the intersection of two X-ray cones in stereotactic space. V
olume measurements were compared with prescription volumes from dose planni
ng, in phantoms and in patients treated with Gamma Knife(R) radiosurgery fo
r cerebral arteriovenous malformations. Previous studies of 1128 treated pa
tients were used to calculate the risk for complication as a function of th
e nidus volume. In 63 patients volumes measured with either method correlat
ed. R-2 = 0.85. Volume as measured with the intersecting cone model (ICM) c
orrelated with predicted Gamma Knife(R) radiosurgery complication rate, R-2
= 0.84. The ICM can thus be used for measurement of AVM volumes less than
10 cm(3) from angiography. Outcome models from Gamma Knife(R) radiosurgery
may be applied, but with reduced exactness. Standardised AVM volume measure
ment is valuable for comparing outcome and for quantification of volume red
uction after therapy, notably embolisation. Thus the optimal management pla
n may be selected in conjunction with diagnostic or therapeutic angiography
.