No abstract

This work focuses on the design of a semiconductor pixelated γ-ray camera with a pixel size of 1 mm2. The cost of semiconductor manufacturing is mainly driven by economies of scale, which makes silicon the cheapest semiconductor material due to its widespread utilization. The energy of γ-photons used in radiation therapy are in a range, in which the dominant interaction mechanism is Compton scattering in every conceivable sensor material. Since the Compton scattering cross section is linearly dependent upon Z, it is less rewarding to utilize high Z sensor materials, than it is in the case of X-ray detectors (X-rays interact also via the photoelectric effect whose cross section scales proportional to Zn, where n is ≈ 4,5). For the stated reasons it was decided to use the low Z material silicon (Z = 14) despite its worse detection efficiency. The proposed detector is designed as a portal detector to be used in radiation cancer therapy. The purpose of the detector is to ensure correct patient alignment, spatial dose monitoring and to provide the feedback necessary for an emergency shutdown should the spatial dose rate profile deviate from the treatment plan. Radiation therapy equipment is complex and thus failure prone and the consequences of malfunction are often life threatening. High spatial resolution and high detection efficiency are not a high design priority. The detector design priorities are focused up on radiation hardness, robustness and the ability to cover a large area cost efficiently. The quintessential idea of the PanterPix detector exploits the relaxed spatial resolution requirement to achieve the stated goals. The detector is composed of submodules, each submodule consisting of a Si sensor with an array of fully depleted detection diodes and 8 miniature custom design readout ASICs collecting and measuring the minuscule charge packets generated due to ionization in the PN junctions.

Hybrid pixel detectors (HPD) are nowadays well known and widely used in fundamental research, e.g. in high energy physics experiments. Over the last decade, segmented semiconductor detectors have also found use in medicine. The total doses received by medical radiation detectors often reach a significant level (up to several hundreds of kGy per decade), especially in applications such as transmission portal in-vivo dosimetry. Such doses might affect detector properties. Therefore, it is necessary to evaluate their performance after absorbing a significant radiation dose. PantherPix is a novel 2D hybrid pixel detector which is designed specifically for use in radiation therapy. As was concluded in earlier studies, it is suitable for radiotherapy quality assurance (QA) and portal dosimetry. In this paper, the PantherPix radiation hardness is investigated using a 60Co source. The dependence on dose of the full depletion voltage, leakage current, detector power consumption and detector response are provided. The PantherPix radiation tolerance has been shown to be adequate for common cumulative doses delivered to radiation detectors in radiotherapy over several decades and its performance has been verified for doses up to 3000 kGy.

Electronic portal imaging device (EPID) is a common part of almost each radiotherapy treatment unit. It is a quick and simple tool for various clinical procedures, e.g. verification of patient positioning or patient specific QA. Currently available EPIDs have usually the spatial resolution below 0.5 mm. As EPIDs are not primarily designed for diagnostics, even lower spatial resolution might be acceptable. This work assesses the effect of MV image spatial resolution on clinical routines which employ the EPID and addresses the minimal technical requirements of such devices. The aim is to determine if EPIDs with poorer spatial resolution are able to reach the same clinical quality as currently used EPIDs. The effects of MV image spatial resolution on the imaging performance, patient positioning and patient specific QA were studied. The lower spatial resolution was simulated by manually downsampling the original images down to 4× poorer spatial resolution. The study includes an analysis of 96 images used for the patient positioning verification and 61 treatment field images used for the patient specific QA.

No abstract