Proton beam range verification with a dedicated in-situ partial-ring time-of-flight PET scanner

Objectives Proton beam therapy promises to deliver higher radiation dose localized to the tumor region relative to photon radiation therapy. Proton beam range (PR) uncertainties however necessitate employing large safety margins, reducing its efficacy. It has been shown that PET imaging of the short-lived isotopes (primarily ¹⁵O & ¹¹C) generated within the patient along the proton beam path can be used to verify the range & hence reduce these large safety margins. This work experimentally assesses the feasibility of a prototype in-situ PET scanner to image these isotopes & then utilize the information to perform range-verification using a clinically relevant dose & scan time.

Methods PET measurements were performed with a prototype time-of-flight (TOF) scanner using two detector modules. Each module measures ~12 x 25 cm² & is comprised of an array of 4 x 4 x 30 mm³ LaBr₃ crystals readout by an array of PMTs. The LaBr₃ detectors offer very good coincidence timing resolution of ~400 ps. Use of TOF assists in maintaining image quality & minimizing artifacts with limited angle coverage for the dual-detector design that enables in-situ measurements. Standard off-the-shelf lab data acquisition was used to acquire and write list-mode data to the disk. An offline iterative TOF list-mode reconstruction & analytic attenuation correction was used to generate tomographic images. Three different phantoms (a) carbon-rich polyethylene (PE), (b) oxygen-rich gel-water (GW), & (c) adipose-tissue equivalent (ATE) were separately irradiated with either 54 or 92 MeV mono-energetic protons. Each irradiation lasted ~3 seconds & deposited 2 Gy dose over 4 x 4 cm² area. Since the prototype PET scanner could not be mounted with the proton gantry, it was placed close to the proton treatment table in the gantry room, thereby minimizing the delay (~1 min) between end of proton irradiation & start of data acquisition. The PR is estimated from the 30% pick-off averaged over multiple depth-profiles (along the proton beam direction) extracted from the image over a region corresponding to the proton dose deposition.

Results The proton beam generated: ¹¹C in PE, ¹⁵O in GW, & a mixture of ¹¹C & ¹⁵O isotopes in the ATE phantoms respectively. From the depth-profiles in the ATE phantom, the precision (std. dev.) in determining the PR was ~1.1 mm for a 10 min scan with 1 min delay. In contrast, a 10 min scan after 20 min delay missed the ¹⁵O signal, collected fewer total counts & only achieved ~4 mm precision in determining PR. These results were validated with studies obtained from PE & GW phantoms that measure ¹¹C and ¹⁵O separately.

Conclusions An in-situ PET scanner allows ¹⁵O as well as ¹¹C istotope imaging to allow PR determination with ~1 mm precision. This is especially important for imaging irradiated tumors located in skeletal/soft tissue, which primarily produce ¹⁵O after proton irradiation, and also to reduce biological washout effects due to short scan times. The current proto-type scanner is suitable for imaging small animals, & studies are underway to demonstrate the ability to utilize the PR verification in research studies. The field-of-view of the scanner can be increased for human use by making use of additional detectors to assist in treatment planning. Research Support: NIH Grant R01-CA113941 & internal funding from the Dept. of Rad. Onc. at the Univ. of Pennsylvania.