Improving lesion detectability in PET imaging with a penalized likelihood reconstruction algorithm

Abstract

Ordered Subset Expectation Maximization (OSEM) is currently the most widely used image reconstruction algorithm for clinical PET. However, OSEM does not necessarily provide optimal image quality, and a number of alternative algorithms have been explored. We have recently shown that a penalized likelihood image reconstruction algorithm using the relative difference penalty, block sequential regularized expectation maximization (BSREM), achieves more accurate lesion quantitation than OSEM, and importantly, maintains acceptable visual image quality in clinical wholebody PET. The goal of this work was to evaluate lesion detectability with BSREM versus OSEM. We performed a twoalternative forced choice study using 81 patient datasets with lesions of varying contrast inserted into the liver and lung. At matched imaging noise, BSREM and OSEM showed equivalent detectability in the lungs, and BSREM outperformed OSEM in the liver. These results suggest that BSREM provides not only improved quantitation and clinically acceptable visual image quality as previously shown but also improved lesion detectability compared to OSEM. We then modeled this detectability study, applying both nonprewhitening (NPW) and channelized Hotelling (CHO) model observers to the reconstructed images. The CHO model observer showed good agreement with the human observers, suggesting that we can apply this model to future studies with varying simulation and reconstruction parameters.

A phantom design for assessment of detectability in PET imaging

“…The primary clinical role of PET imaging is the detection of anomalous regions of 18 F-FDG uptake, which are often indicative of malignant lesions. Consequently there have been multiple studies trying to objectively evaluate the impact on detection performance of the PET acquisition method, 6,26-32 the image reconstruction method, 7,[33][34][35] and the combined impact of PET acquisition and reconstruction methods. [36][37][38] In addition, there have been several studies on the optimal choice of model observers.…”

Reconstruction of uptake patterns in PET: The influence of regularizing prior

“…11,14 In 11 it was the best lesion detectability, defined through a FOM taking into account both qualitative and quantitative contents. Here, the idea was to fix the noise texture in uniform areas with about the activity measured inside liver in 18 F-FDG clinical studies. Liver was taken as target since it is generally considered the most critical organ to be evaluated because of the high noise level which reduces lesion detectability.…”

“…The clinical interest toward 18 F-FluoroDeoxyGlucose (FDG) Positron Emission Tomography (PET) image quantification has recently moved from the simple extraction of Standard Uptake Values (SUV) to a more comprehensive assessment of the uptake spatial distribution inside lesions. Volume, total glycolysis, heterogeneity, and morphology of lesions have shown to correlate with prognosis and therapy outcome in many kinds of tumors and thus begin to be taken into account even in radiotherapy planning.…”

“…11 The assessment of a prior and the comparison with un-regularized reconstructions are generally based on contrast to noise ratio, 13 accuracy in SUV quantification, 10 and detectability of lesions/objects of interest. 17,18 A prior leads however to a characteristic and object-dependent image covariance; 19 noise texture and smoothness can dramatically vary by simply changing the prior weight. 13 The choice of a prior and of its parameters may therefore largely affect image quality, uniformity (or heterogeneity), and boundary sharpness of lesions/organs.…”

“…Accuracy in reconstructing activity values and homogeneous/heterogeneous activity patterns, easiness of lesion boundary extraction, together with lesion detectability and image quality were evaluated. A new strategy to optimize and compare priors is proposed, where prior parameters are set by fixing the noise texture in uniform areas with about the liver radioactivity concentration in 18 F-FDG studies. Two priors are compared: the one proposed by Rapisarda et al 11 and the relative difference prior, proposed by Nuyts et al 16 and largely validated.…”