Compartment models are the basis for most physiologically based quantificat
ion of nuclear medicine data. Although some software packages are available
for this purpose, many are expensive, run on relatively few types of compu
ters or are of limited capability, and cannot be extended because of the un
availability of source code. Consequently, institutions with modeling exper
tise often develop software for themselves, which has the disadvantages of
lack of standardization and possible replication of effort. Therefore, gene
ral-purpose compartment modeling software distributed with source code woul
d be a welcome resource for the nuclear medicine community. Methods: We for
mulated a mathematic framework within which compartment models containing u
nimolecular and bimolecular (receptor saturation) kinetics can be described
. We implemented this framework within MATLAB and call the resultant softwa
re COMKAT (Compartment Model Kinetic Analysis Tool). Results: COMKAT simpli
fies the process of defining and solving standard: blood flow,F-18-FDG, and
receptor models as well as models of a user's own design. In particular, C
OMKAT automatically defines and implements state, analytic sensitivity, and
Jacobian equations. Given these, COMKAT can perform simulations in which m
odel outputs are solved for specified parameter values, thereby allowing th
e user to predict how sensitive data are to these parameters. In addition,
COMKAT can be used to estimate values for the parameters by fitting model o
utput to experimental data. COMKAT is equipped with command-line and graphi
c user interfaces from which the user can access these features. Examples o
f these applications are presented along with validation and performance su
mmaries. Conclusion: COMKAT is-a useful software tool and is available with
out cost to researchers, at www.nuclear.uhrad.com/comkat.