Impact of time-of-flight technology on PET/CT image quality and SUV quantitation

Abstract

Time-of-flight technology (TOF) is at the leading edge of advancements in PET/CT imaging, and is made possible due to the availability of new scanner design, scintillating materials and fast computers. The location of annihilation is estimated according to the time difference between scintillation events produced by two 511KeV gamma photons travelling in opposite directions. The additional information is incorporated into image reconstruction to suppress noise propagation and increase signal-to-noise ratio. As shown in previous studies, these properties can be translated in to clinical usage to obtain improved image quality using the same acquisition time, while also enhancing image quality under shortened acquisition or reduce tracer dosage injected to patients.

Semi-quantification units such as standardized uptake value (SUV) and standard deviation (SD) are measured on static PET images and are widely used clinically to quantify the rate of cell metabolism. Signal-to-noise ratio (SNR) and coefficient of variation (CV) can be calculated to assess image quality. However, these parameters depend on image reconstruction and are therefore expected to change with the use of TOF.

Breath holding (BH) PET/CT imaging protocol has been suggested to help evaluation of lesions affected by respiratory motion. However, the PET image quality is significantly degraded since the acquisition time is shortened (e.g. 30 seconds). Noise on BH PET image may cause inaccuracy in semi-quantitation and difficulties in visual assessment. TOF can be applied to improve the image quality and enhance subtle lesions, thus helping the characterization of small lung nodules (SLN) with potential clinical impact.

This thesis aims to evaluate the impact of TOF on PET image quality and the quantification of SUV in clinical practice. Firstly, the performance of a TOF PET/CT scanner is tested using a phantom mimicking human body trunk with ‘lesions’. Using the same acquisition time (2 minutes/bed), TOF images outperformed conventional images in terms of lesion detectability and background uniformity, especially for small spheres and under low lesion-to-background ratio. Moreover, TOF images acquired using a shorter scan time (1.5 minutes/bed) also maintained acceptable image quality. In the second study, the influence on SUV measurement and image quality (in terms of CV) are evaluated in twelve normal organ structures on whole-body FDG PET/CT images. TOF significantly decreased SUVmax (9/12) and SUVmean (8/12) among the twelve normal organ structures investigated. It also improved the image quality, particularly in solid organs in the abdomen. The combined utility of TOF and BH PET/CT imaging protocol for the detection of SLN is demonstrated in the third study. Lesions showing PET activity, which were misaligned with anatomical location on CT, had higher SUVmax and comparable SNR on BH images. PET/CT fusion was also markedly improved. The additional use of TOF further enhanced lesion detection and improved SNR.

The results demonstrate the benefits of using TOF in clinical PET/CT, including its advantage with breath holding PET/CT imaging. This information not only offers a better understanding of this increasingly popular technology in modern PET/CT scanners, but also highlights its potential applications in clinical practice.