Question

75. How does time of flight improve the reconstruction process?

Please be specific and clear! Thank you.

Answer 1

The method for measuring the distance between a sensor and an object, based on the time difference between the emission of a signal and its return to the sensor, after being reflected by an object is known as Time-of-Flight principle (ToF).

The combination of fast photomultipliers, scintillators with high density, modern electronics, and faster computing power for image reconstruction have made it possible to introduce this principle in clinical TOF-PET systems.

•According to investigation and research Correction of the measured timing differences between crystals , timing calibration methods using either a rotating line source or a stationary, shielded point source at the center of the scanner .

• For example in case of the Gemini, a 68Ge rod source is used to measure the timing offsets . The range of offset factors for the entire scanner is found to be approx., 61 ns. single-scatter simulation (SSS) should be modified to estimate the scatter distribution as a function of radial and time domains for scatter correction.

•SSS refers to single scatter simulation.

•Some how use of fine timing bins has a greater effect over crossed timing bins as maintanence of intrinsic time resolution of the system is very important.

•Use of fine timing bins will allow accurate modeling of scatter estimation simultaneously maintaining the intrinsic timing resolution of the system , and impractical demands on storage requirements or reconstruction time sould be avoided.

•The reconstruction approach implemented for TOF is a list-mode version of the MULTIPLE HIDDEN LAYERS EXTREME LEARNING MACHINE algorithm with a TOF kernel applied in both the forward- and the back projection operations

TOF KERNEL is used in reconstrustion is taken to be the measured coincidence timing resolution of the system,generally moleded as a Gaussian function of known width.

•MLEM REFERS TO MULTIPLE HIDDEN LAYERS EXTREME LEARNING MACHINE .

•The Poisson statistics of the data are preserved by using the data in their original form (i.e., without binning) and incorporating the physical effects of PET in the system model.

• The TOF response function should be modeled as a 1-D Gaussian function along the LOR.

• The MULTIPLE HIDDEN LAYERS EXTREME LEARNING MACHINE algorithm is accelerated by dividing the data into chronologically ordered subsets .

•A fully 3D list-mode iterative reconstruction algorithm is computationally intensive and has proven to be challenging to implement for clinical purposes .

•Most importantly a 10-node dual CPU is must needed.

• Using a 10-node (dual central-processing-unit) computer cluster, the image processing with this algorithm proceeds in parallel with data acquisition and is typically completed for clinical whole-body (multi bed) studies after the end of the acquisition, depending on the number of counts collected.

•A minimum study of 10-30 mins is needed for better imaging .

Hence , the time of flight or TOF plays a great role to improve the reconstruction process of respective image.