Quantitative analysis of planar technetium-99m-sestamibi myocardial perfusion images using modified background subtraction.

Standard interpolative background subtraction, as used for thallium-201 (201Tl), may create artifacts when applied to planar technetium-99m-Sestamibi (99mTc-Sestamibi) images, apparently because of the oversubtraction of relatively high extra-cardiac activity. A modified background subtraction algorithm was developed and compared to standard background subtraction in 16 patients who had both exercise-delayed 201Tl and exercise-rest 99mTc-Sestamibi imaging. Furthermore, a new normal data base was generated. Normal 99mTc-Sestamibi distribution was slightly different compared to 201Tl. Using standard background subtraction, mean defect reversibility was significantly underestimated by 99mTc-Sestamibi compared to 201Tl (2.8 +/- 4.9 versus -1.8 +/- 8.4, p less than 0.05). Using the modified background subtraction, mean defect reversibility on 201Tl and 99mTc-Sestamibi images was comparable (2.8 +/- 4.9 versus 1.7 +/- 5.2, p = NS). We conclude, that for quantification of 99mTc-Sestamibi images a new normal data base, as well as a modification of the interpolative background subtraction method should be employed to obtain quantitative results comparable to those with 201Tl.