Using the Hotelling observer on multislice and multiview simulated SPECT myocardial images

A two-layer model which includes a channelized Hotelling observer (CHO) followed by a Hotelling observer (HO) is proposed to operate on three-dimensional images containing multiple slices and multiple views similar to clinical displays of SPECT myocardial perfusion images. The observer was applied to simulated data of an idealized heart model. The task is the detection of cold defects in the myocardial wall. A CHO was computed for each slice of the multislice image and an HO was then applied on the decision variable arrays to obtain a multislice CHO-HO lesion detectability index. Results are compared to those of a single-slice CHO applied to the slice through the center of the lesion and also to human-observer results. The result suggests the ability of the multislice method to separate better these images into their correct classes as compared to the single-slice method, by combining information from multiple slices. The multislice method provides a single detectability scalar for studies in which the lesion could be present in multiple slices. It may be that a multislice CHO-HO model could more closely predict the performance of radiologists in tasks where the radiologist evaluates multislice and multiview images.

Duke Scholars

Author Ronald J. Jaszczak Radiology