The impact of ROI definition on estimator performance on FBP reconstructedGa-67SPECT images

Quantitative accuracy of reconstructed images depends not only on the recon struction strategy, but also on the definition of the estimator. We investi gate estimators based on: 1) a dilation of the true ROI, 2) thresholding th e reconstructed image, 3) edge detection on the reconstructed image, and 4) a quasi-Gauss-Markov (QGM) estimator which partially compensates for the i mage acquisition and reconstruction process. The Gauss-Markov estimator is also briefly compared to these first four for a simpler imaging problem. Th e task is to find the activity in small (1 cm diameter) lesions in simulate d images of thoracic gallium SPECT scans. We consider 35 lesion locations. Reconstruction is done with Chang-corrected FBP. We consider both the best accuracy achievable with each estimator and how sensitive this accuracy is to errors in the specification of the lesion's size and position. A signal- to-noise ratio (SNR) which combines both bias and variance is used as the b asis for comparison. The results show that while the estimators based on a dilation of the true ROI or the thresholded image can produce very high SNR , they are both very sensitive to errors in lesion size and position. The e dge-detection estimator and the QGM estimator both have lower optimal SNR, but are less sensitive to some lesion-specification