Interventional nuclear imaging for radioembolistation guidance

Abstract

The aim of this thesis is to investigate the development and application of a device which can acquire both fluoroscopic and nuclear images in the intervention room for guidance of liver radioembolisation procedures. The first part describes the development of the hardware required for such a hybrid imaging device. A hybrid imaging modality, consisting of an x-ray c-arm combined with a four gamma cameras placed adjacent to the x-ray tube, was previously developed and found feasible. We showed that nuclear image quality of this device can be improved by using parallel hole collimators, instead of the previously used pinhole collimators. However, the required intermediate reconstruction step and the added weight to the already heavy x-ray tube were not ideal. Therefore, a dual layer detector, consisting of an x-ray flat panel placed in front of a gamma camera with cone beam collimator focussed at the x-ray focal spot, was designed. This design relies on the x-ray detector absorbing the majority of the x-rays, while being more transparent to the higher energy gamma photons. No intermediate reconstruction step is needed and fluoroscopic and nuclear images are intrinsically registered to each other. The proposed dual layer detector was found to be capable of acquiring real-time intrinsically registered hybrid images. This could aid interventional procedures involving radionuclides, such as radioembolisation. The efficiency of radioembolisation procedures could be greatly enhanced if results of the 99mTc-MAA pre-treatment procedure would already be available in the intervention room, enabling single-session procedures thanks to direct intrahepatic dose feedback, assessment of extrahepatic depositions and lung shunt estimation. The second part of this thesis concerned the improvement of radioembolisation procedures by interventional hybrid imaging and other nuclear imaging techniques. Depending on the lung shunt fraction (LSF), treatment may be altered or abandoned. We showed that LSF can be estimated with a constant minor error with our hybrid imaging device within a few seconds in an interventional setting. To evaluate the intrahepatic distribution and the presence of any extrahepatic depositions, a SPECT scan is needed. A clinical SPECT scan, however, takes approximately 30 minutes, which is too long for interventional purposes. We have investigated whether a fast SPECT protocol is feasible and if non-uniform projection-time acquisition protocols improve accuracy. In was concluded that both uniform and non-uniform projection-time acquisition liver SPECT protocols yield accurate dosimetric metrics for radioembolisation treatment planning in the interventional suite within 10 minutes, without compromising image quality. Consequently, fast interventional SPECT of the liver in the intervention room is feasible. Radioembolisation is increasingly used as a bridge to resection (i.e. radiation lobectomy). It combines ipsilateral tumor control with the induction of contralateral hypertrophy to facilitate lobar resection. The potential complementary value of hepatobiliary scintigraphy (HBS) before and after radioembolisation in the assessment of the future remnant liver was investigated.