Impact of Time-of-Flight PET on Whole-Body Oncologic Studies: A Human Observer Lesion Detection and Localization Study

Suleman Surti, Joshua Scheuermann, Georges El Fakhri, Margaret E. Daube-Witherspoon, Ruth Lim, Nathalie Abi-Hatem, Elie Moussallem, François Bénard, David A. Mankoff, Joel S. KarpJournal of Nuclear Medicine 2011 · DOI

Abstract

A combination of longer scan time (3 min in this study) and Time-of-Flight (TOF) imaging provides the best performance for imaging large patients or low-uptake lesions. This imaging protocol provides similar performance across all patient sizes for lesions in the same organ type with similar relative uptake, indicating an ability to provide a uniform clinical diagnosis in most oncologic lesion detection tasks.

Introduction: The "Needle in a Haystack" Problem

In Positron Emission Tomography (PET), detecting small lesions is often a challenge due to background noise caused by photon attenuation and scatter within the patient's body. Historically, larger patients (those with a high Body Mass Index) have suffered from poorer image quality because their bodies create more signal interference, making lesions harder to distinguish.

While theoretical models suggested that Time-of-Flight (TOF) technology—which measures the tiny difference in arrival time between two photons to better localize annihilation events—could improve the signal-to-noise ratio (SNR), its impact on actual human observer performance in clinical environments required validation. This study evaluates whether TOF PET acts as a "diagnostic equalizer" for human readers.

Methods: A Hybrid Data Approach

To evaluate performance with a known "ground truth," researchers used a hybrid data approach. They merged clinical $^{18}$F-FDG PET data from 100 healthy patient studies with simulated 10-mm spherical lesions. This process included rigorous corrections for attenuation and scatter to ensure the simulated tumors appeared authentic within the real patient backgrounds.

Hybrid data generation process for creating realistic PET images with known lesion locations.

Images were reconstructed using both TOF and non-TOF algorithms at simulated scan durations of 1 minute and 3 minutes per bed position. Five human observers (physicians and physicists) evaluated 720 images each, identifying lesion locations and providing confidence scores. Performance was quantified using the Area under the Localized Receiver Operating Characteristic (AROC) curve.

Key Results

The study demonstrated that TOF imaging significantly improves detection and localization, particularly in challenging scenarios.

Comparison of observer performance (AROC) across different imaging conditions, organs, and patient sizes.
  • The "Equalizer" Effect: For heavy patients (BMI ≥ 26, labeled 'H' in results), TOF led to a statistically significant improvement ($P < 0.01$) in AROC for both 1-minute and 3-minute scans. TOF effectively brought the diagnostic certainty for large patients much closer to that of smaller patients ('L').
  • Scan Time vs. Technology: While increasing scan time from 1 to 3 minutes improved results across the board, TOF imaging at 1 minute often outperformed non-TOF imaging at 3 minutes for certain tasks, suggesting TOF can potentially allow for shorter scan times without sacrificing quality.
  • Organ-Specific Performance: In the liver, TOF significantly improved AROC for heavy patients. In the lungs, where the background is more non-uniform due to anatomical structures, TOF improved performance for all patient sizes.

Discussion and Conclusion

TOF PET imaging acts as a critical technological advancement for clinical oncology. By reducing the performance gap between small and large patients, it allows for a more uniform clinical diagnosis. Regardless of patient size, if a lesion has a similar relative uptake, a physician's ability to locate it remains relatively stable when using TOF.

The highest observer performance was consistently achieved using a combination of 3-minute scan times and TOF reconstruction. While this study used static simulated lesions, future work will need to incorporate respiratory motion to further refine these findings for lesions in the lungs and upper liver.

Summary updated

Abstract

A combination of longer scan time (3 min in this study) and TOF imaging provides the best performance for imaging large patients or a low-uptake lesion in small or large patients. This imaging protocol also provides similar performance for all patient sizes for lesions in the same organ type with similar relative uptake, indicating an ability to provide a uniform clinical diagnosis in most oncologic lesion detection tasks.
BibTeX
@article{surti2011impact,
  author = {Suleman Surti and Joshua Scheuermann and Georges El Fakhri and Margaret E. Daube-Witherspoon and Ruth Lim and Nathalie Abi-Hatem and Elie Moussallem and François Bénard and David A. Mankoff and Joel S. Karp},
  title = {{Impact of Time-of-Flight PET on Whole-Body Oncologic Studies: A Human Observer Lesion Detection and Localization Study}},
  journal = {Journal of Nuclear Medicine},
  year = {2011},
  doi = {10.2967/jnumed.110.086678},
}
Figure