Model observer design for multi-signal detection in the presence of anatomical noise

Abstract

As psychophysical studies are resource-intensive to conduct, model observers are commonly used to assess and optimize medical imaging quality. Model observers are typically designed to detect at most one signal. However, in clinical practice, there may be multiple abnormalities in a single image set (e.g. multifocal multicentric (MFMC) breast cancer), which can impact treatment planning. Prevalence of signals can be different across anatomical regions, and human observers do not know the number or location of signals a priori. As new imaging techniques have the potential to improve multiple-signal detection (e.g. digital breast tomosynthesis may be more effective for diagnosis of MFMC than mammography), image quality assessment approaches addressing such tasks are needed. In this study, we present a model observer to detect multiple signals in an image dataset. A novel implementation of partial least squares (PLS) was developed to estimate different sets of efficient channels directly from the images. The PLS channels are adaptive to the characteristics of signals and the background, and they capture the interactions among signal locations. Corresponding linear decision templates are employed to generate both image-level and location-specific scores on the presence of signals. Our results show that: (1) the model observer can achieve high performance with a reasonably small number of channels; (2) the model observer with PLS channels outperforms that with benchmark modified Laguerre-Gauss channels, especially when realistic signal shapes and complex background statistics are involved; (3) the tasks of clinical interest, and other constraints such as sample size would alter the optimal design of the model observer.

Virtual assessment of stereoscopic viewing of digital breast tomosynthesis projection images

“…This shortcoming may be critical for patients with multifocal multicentric breast cancer (i.e., presence of more than two tumors in the same breast). [64][65][66] As locally varying anatomical backgrounds may notably impact lesion detectability, 41,[67][68][69] it would be interesting to consider, the impact of multiple lesions, and additional factors such as the distances between lesions. Moreover, similar model observer studies could be conducted to study other questions such as: (1) the efficacy of stereo viewing for detecting other types of breast abnormalities such as clusters of microcalcifications and architectural distortion, (2) the effective use of radiation dose in the context of stereo viewing, and (3) other feasible DBT configurations (not limited to current geometry designs in commercial DBT scanners) and visualization methods (e.g., cine loop of stereo pairs).…”

Synthetic breast phantoms from patient based eigenbreasts

“…In previous work, we presented a cohort of approximately 200 patient‐based phantoms . Although that dataset is the largest of its type, VCTs may require cases numbering in the thousands or even tens of thousands . It would be prohibitively difficult to acquire and process images from such large numbers of actual human subjects.…”

“…However, clinical bCT data is still not widely available and the segmentation process is computationally expensive. Thus, it is difficult to directly scale‐up these patient‐based breast phantoms for VCTs, where large quantities of phantoms would be necessary to tune the model observers and account for nonstationary of the data statistics over different locations with‐in the breast …”

Model observer for assessing digital breast tomosynthesis for multi-lesion detection in the presence of anatomical noise

“…for each possible lesion location, a binary decision regarding lesion presence is made). In our previous studies with synthetic 2D mammograms (Wen et al 2016b(Wen et al , 2017a, we developed a 2D ML-CHO model for accurately detecting multiple lesions (e.g. circular Gaussian lesions, irregularly-shaped masses) in the presence of anatomical noise.…”

“…The main idea is to consider 3D DBT reconstructed image data (varying the number of slices N s around the central slice that lesions may appear in) as a single object to be examined, and that detection decisions regarding the presence of lesions are made upon the overall information from both individual slices and their correlations (Platiša et al 2011. There are three main steps for the 3D ML-CHO to generate the detection variables: (1) concatenate the stack of relevant reconstructed image slices together to represent the 3D DBT image data X; (2) compute the location-level 3D channels T loc (Wen et al 2016a(Wen et al , 2016b(Wen et al , 2017a; and (3) use the estimated T loc channels in the 3D ML-CHO for generating decision variables (Platiša et al 2011).…”