Segmented phosphors: MEMS-based high quantum efficiency detectors for megavoltage x-ray imaging

Current electronic portal imaging devices (EPIDs) based on active matrix flat panel imager (AMFPI) technology use a metal plate+phosphor screen combination for x-ray conversion. As a result, these devices face a severe trade-off between x-ray quantum efficiency (QE) and spatial resolution, thus, significantly limiting their imaging performance. In this work, we present a novel detector design for indirect detection-based AMFPI EPIDs that aims to circumvent this trade-off. The detectors were developed using micro-electro-mechanical system (MEMS)-based fabrication techniques and consist of a grid of up to approximately 2 mm tall, optically isolated cells of a photoresist material, SU-8. The cells are dimensionally matched to the pixels of the AMFPI array, and packed with a scintillating phosphor. In this paper, various design considerations for such detectors are examined. An empirical evaluation of three small-area (approximately 7 x 7 cm2) prototype detectors is performed in order to study the effects of two design parameters--cell height and phosphor packing density, both of which are important determinants of the imaging performance. Measurements of the x-ray sensitivity, modulation transfer function (MTF) and noise power spectrum (NPS) were performed under radiotherapy conditions (6 MV), and the detective quantum efficiency (DQE) was determined for each prototype SU-8 detector. In addition, theoretical calculations using Monte Carlo simulations were performed to determine the QE of each detector, as well as the inherent spatial resolution due to the spread of absorbed energy. The results of the present studies were compared with corresponding measurements published in an earlier study using a Lanex Fast-B phosphor screen coupled to an indirect detection array of the same design. The SU-8 detectors exhibit up to 3 times higher QE, while achieving spatial resolution comparable or superior to Lanex Fast-B. However, the DQE performance of these early prototypes is significantly lower than expected due to high levels of optical Swank noise. Consequently, the SU-8 detectors presently exhibit DQE values comparable to Lanex Fast-B at zero spatial frequency and significantly lower than Fast-B at higher frequencies. Finally, strategies for reducing Swank noise are discussed and theoretical calculations, based on the cascaded systems model, are presented in order to estimate the performance improvement that can be achieved through such noise reduction.     

Segmented crystalline scintillators: empirical and theoretical investigation of a high quantum efficiency EPID based on an initial engineering prototype CsI(TI) detector

Modern-day radiotherapy relies on highly sophisticated forms of image guidance in order to implement increasingly conformal treatment plans and achieve precise dose delivery. One of the most important goals of such image guidance is to delineate the clinical target volume from surrounding normal tissue during patient setup and dose delivery, thereby avoiding dependence on surrogates such as bony landmarks. In order to achieve this goal, it is necessary to integrate highly efficient imaging technology, capable of resolving soft-tissue contrast at very low doses, within the treatment setup. In this paper we report on the development of one such modality, which comprises a nonoptimized, prototype electronic portal imaging device (EPID) based on a 40 mm thick, segmented crystalline CsI(Tl) detector incorporated into an indirect-detection active matrix flat panel imager (AMFPI). The segmented detector consists of a matrix of 160 x 160 optically isolated, crystalline CsI(Tl) elements spaced at 1016 microm pitch. The detector was coupled to an indirect detection-based active matrix array having a pixel pitch of 508 microm, with each detector element registered to 2 x 2 array pixels. The performance of the prototype imager was evaluated under very low-dose radiotherapy conditions and compared to that of a conventional megavoltage AMFPI based on a Lanex Fast-B phosphor screen. Detailed quantitative measurements were performed in order to determine the x-ray sensitivity, modulation transfer function, noise power spectrum, and detective quantum efficiency (DQE). In addition, images of a contrast-detail phantom and an anthropomorphic head phantom were also acquired. The prototype imager exhibited approximately 22 times higher zero-frequency DQE (approximately 22%) compared to that of the conventional AMFPI (approximately 1%). The measured zero-frequency DQE was found to be lower than theoretical upper limits (approximately 27%) calculated from Monte Carlo simulations, which were based solely on the x-ray energy absorbed in the detector-indicating the presence of optical Swank noise. Moreover, due to the nonoptimized nature of this prototype, the spatial resolution was observed to be significantly lower than theoretical expectations. Nevertheless, due to its high quantum efficiency (approximately 55%), the prototype imager exhibited significantly higher DQE than that of the conventional AMFPI across all spatial frequencies. In addition, the frequency-dependent DQE was observed to be relatively invariant with respect to the amount of incident radiation, indicating x-ray quantum limited behavior. Images of the contrast-detail phantom and the head phantom obtained using the prototype system exhibit good visualization of relatively large, low-contrast features, and appear significantly less noisy compared to similar images from a conventional AMFPI. Finally, Monte Carlo-based theoretical calculations indicate that, with proper optimization, further, significant improvements in the DQE performance of such imagers could be achieved. It is strongly anticipated that the realization of optimized versions of such very high-DQE EPIDs would enable megavoltage projection imaging at very low doses, and tomographic imaging from a "beam's eye view" at clinically acceptable doses.     

Development of high quantum efficiency, flat panel, thick detectors for megavoltage x-ray imaging: an experimental study of a single-pixel prototype

Our overall goal is to develop a new generation of electronic portal imaging devices (EPIDs) with a quantum efficiency (QE) more than an order of magnitude higher and a spatial resolution equivalent to that of EPIDs currently used for portal imaging. A novel design of such a high QE flat-panel based EPID was introduced recently and its feasibility was investigated theoretically [see Pang and Rowlands, Med. Phys. 31, 3004 (2004)]. In this work, we constructed a prototype single-pixel detector based on the novel design. Some fundamental imaging properties including the QE, spatial resolution, and sensitivity of the prototype detector were measured with a 6 MV beam. It has been shown that the experimental results agree well with theoretical predictions and further development based on the novel design including the construction of a prototype area detector is warranted.     

Development of high quantum efficiency, flat panel, thick detectors for megavoltage x-ray imaging: a novel direct-conversion design and its feasibility

Most electronic portal imaging devices (EPIDs) developed to date, including recently developed flat panel systems, have low x-ray absorption, i.e., low quantum efficiency (QE) of 2%-4% as compared to the theoretical limit of 100%. A significant increase of QE is desirable for applications such as a megavoltage cone-beam computed tomography (MVCT) and megavoltage fluoroscopy. However, the spatial resolution of an imaging system usually decreases significantly with an increase of QE. The key to the success in the design of a high QE detector is therefore to maintain the spatial resolution. Recently, we demonstrated theoretically that it is possible to design a portal imaging detector with both high QE and high resolution [see Pang and Rowlands, Med. Phys. 29, 2274 (2002)]. In this paper, we introduce such a novel design consisting of a large number of microstructured plates (made by, e.g., photolithographic patterning of evaporated or electroplated layers) packed together and aligned with the incident x rays. On each plate, microstrip charge collectors are focused toward the x-ray source to collect charges generated in the ionization medium (e.g., air or gas) surrounded by high-density materials that act as x-ray converters. The collected charges represent the x-ray image and can be read out by various means, including a two-dimensional (2-D) active readout matrix. The QE, spatial resolution, and sensitivity of the detector have been calculated. It has been shown that the new design will have a QE of more than an order of magnitude higher and a spatial resolution equivalent to that of flat panel systems currently used for portal imaging. The new design is also quantum noise limited down to very low doses (approximately 1-2 radiation pulses of the linear accelerator).     

Development of high quantum efficiency flat panel detectors for portal imaging: intrinsic spatial resolution

Recently developed flat panel detectors have been proven to have a much better image quality than conventional electronic portal imaging devices (EPIDs). They are, however, not yet the ideal systems for portal imaging application due to the low x-ray absorption, i.e., low quantum efficiency (QE), which is typically on the order of 2-4% as compared to the theoretical limit of 100%. The QE of current flat panel systems can be improved by significantly increasing the thickness of the energy conversion layer (i.e., amorphous selenium or phosphor screen). This, however, will be at the expense of a decrease in spatial resolution mainly due to x-ray scatter in the conversion layer (and also the spread of optical photons in the case of phosphor screen). In this paper, we investigate theoretically the intrinsic spatial resolution of a high QE flat panel detector with a new energy conversion layer that is much denser and thicker than that of current flat panel systems. The modulation transfer function (MTF) of the system is calculated based on a theoretical model using a novel approach, which uses an analytical expression for absorbed dose. It is found that if appropriate materials are used for the conversion layer, then the intrinsic MTF of the high QE flat panel is better than that of current EPIDs, and in addition they have a high QE (e.g., approximately 60%). Some general rules for the design of the conversion layer to achieve both high QE and high resolution as well as high DQE are also discussed.     

hybridMANTIS: a CPU-GPU Monte Carlo method for modeling indirect x-ray detectors with columnar scintillators

The computational modeling of medical imaging systems often requires obtaining a large number of simulated images with low statistical uncertainty which translates into prohibitive computing times. We describe a novel hybrid approach for Monte Carlo simulations that maximizes utilization of CPUs and GPUs in modern workstations. We apply the method to the modeling of indirect x-ray detectors using a new and improved version of the code MANTIS, an open source software tool used for the Monte Carlo simulations of indirect x-ray imagers. We first describe a GPU implementation of the physics and geometry models in fastDETECT2 (the optical transport model) and a serial CPU version of the same code. We discuss its new features like on-the-fly column geometry and columnar crosstalk in relation to the MANTIS code, and point out areas where our model provides more flexibility for the modeling of realistic columnar structures in large area detectors. Second, we modify PENELOPE (the open source software package that handles the x-ray and electron transport in MANTIS) to allow direct output of location and energy deposited during x-ray and electron interactions occurring within the scintillator. This information is then handled by optical transport routines in fastDETECT2. A load balancer dynamically allocates optical transport showers to the GPU and CPU computing cores. Our hybridMANTIS approach achieves a significant speed-up factor of 627 when compared to MANTIS and of 35 when compared to the same code running only in a CPU instead of a GPU. Using hybridMANTIS, we successfully hide hours of optical transport time by running it in parallel with the x-ray and electron transport, thus shifting the computational bottleneck from optical tox-ray transport. The new code requires much less memory than MANTIS and, asa result, allows us to efficiently simulate large area detectors.     

Detective quantum efficiency of an amorphous selenium detector to megavoltage radiation.

The spatial frequency dependent detective quantum efficiency (DQE(f)) of a high-resolution selenium-based imaging system has been measured at megavoltage energies. These results have been compared with theoretical calculations. The imaging system was a video tube with a 5 microm amorphous selenium (a-Se) target which was irradiated by 1.25 MeV gamma-rays. The modulation transfer function (MTF) decreased rapidly with spatial frequency (determined by spread of electrons in the build-up material) while the noise power spectrum was constant as a function of spatial frequency. The DQE obtained from these MTF and noise power measurements was compared with a Monte Carlo model of the pulse height spectrum of the detector. The DQE(0) model accounted for the interaction of x rays with the detector as well as the energy-dependent gain (charge generated/energy deposition). Good agreement between the calculated and measured DQE(0) was found. The model was also used to estimate the DQE(f) of a metal plate + a-Se detector which was compared with a metal plate + phosphor system of the same mass thickness. The DQE(f) s of both detectors are very similar, indicating that the choice of which detector is better will be based upon criteria other than DQE(f), such as read-out approach, ease of manufacture or sensitivity.     

A comparison of x-ray detectors for mouse CT imaging

There is significant interest in using computed tomography (CT) for in vivo imaging applications in mouse models of disease. Most commercially available mouse x-ray CT scanners utilize a charge-coupled device (CCD) detector coupled via fibre optic taper to a phosphor screen. However, there has been little research to determine if this is the optimum detector for the specific task of in vivo mouse imaging. To investigate this issue, we have evaluated four detectors, including an amorphous selenium (a-Se) detector, an amorphous silicon (a-Si) detector with a gadolinium oxysulphide (GOS) screen, a CCD with a 3:1 fibre taper and a GOS screen, and a CCD with a 2:1 fibre taper and both GOS and thallium-doped caesium iodide (CsI:Tl) screens. The detectors were evaluated by measuring the modulation transfer function (MTF), noise power spectrum (NPS), detective quantum efficiency (DQE), stability over multiple exposures, and noise in reconstructed CT images. The a-Se detector had the best MTF and the highest DQE (0.6 at 0 lp mm(-1)) but had the worst stability (45% reduction after 2000 exposure frames). The a-Si detector and the CCD with the 3:1 fibre, both of which used the GOS screen, had very similar performance with a DQE of approximately 0.30 at 0 lp mm(-1). For the CCD with the 2:1 fibre, the CsI:Tl screen resulted in a nearly two-fold improvement in DQE over the GOS screen (0.4 versus 0.24 at 0 lp mm(-1)). The CCDs both had the best stability, with less than a 1% change in pixel values over multiple exposures. The pixel values of the a-Si detector increased 5% over multiple exposures due to the effects of image lag. Despite the higher DQE of the a-Se detector, the reconstructed CT images acquired with the a-Si detector had lower noise levels, likely due to the blurring effects from the phosphor screen.     

Influence of metal screens on contrast in megavoltage x-ray imaging

The radiographic contrast of metal screen-film detectors was investigated in order to determine the contrast capabilities of these detectors applied to megavoltage x-ray imaging. The film contrast gamma was found to be independent of the metal screen composition. Measurement of the scatter-to-primary film dose ratio in contact geometry demonstrated that a thick front screen of either 1.5 g/cm2 copper of 2.5 g/cm2 lead provides optimum contrast for the photon energies studied (60Co and 4- and 8-MV x rays). The same thickness were also found to be suitable in an air gap geometry which significantly improved the contrast compared to the contact geometry. Rear lead screens were found to provide no contrast improvement.     

Self-normalizing method to measure the detective quantum efficiency of a wide range of x-ray detectors.

The detective quantum efficiency (DQE) is widely accepted as the most relevant parameter to characterize the image quality of medical x-ray systems. In this article we describe a solid method to measure the DQE. The strength of the method lies in the fact that it is self-normalizing so measurements at very low spatial frequencies are not needed. Furthermore, it works on any system with a response function which is linear in the small-signal approximation. We decompose the DQE into several easily accessible quantities and discuss in detail how they can be measured. At the end we lead the interested reader through an example. Noise equivalent quanta and normalized contrast values are tabulated for standard radiation qualities.     

Countering beam divergence effects with focused segmented scintillators for high DQE megavoltage active matrix imagers

The imaging performance of active matrix flat-panel imagers designed for megavoltage imaging (MV AMFPIs) is severely constrained by relatively low x-ray detection efficiency, which leads to a detective quantum efficiency (DQE) of only ～1%. Previous theoretical and empirical studies by our group have demonstrated the potential for addressing this constraint through the utilization of thick, two-dimensional, segmented scintillators with optically isolated crystals. However, this strategy is constrained by the degradation of high-frequency DQE resulting from spatial resolution loss at locations away from the central beam axis due to oblique incidence of radiation. To address this challenge, segmented scintillators constructed so that the crystals are individually focused toward the radiation source are proposed and theoretically investigated. The study was performed using Monte Carlo simulations of radiation transport to examine the modulation transfer function and DQE of focused segmented scintillators with thicknesses ranging from 5 to 60?mm. The results demonstrate that, independent of scintillator thickness, the introduction of focusing largely restores spatial resolution and DQE performance otherwise lost in thick, unfocused segmented scintillators. For the case of a 60?mm thick BGO scintillator and at a location 20?cm off the central beam axis, use of focusing improves DQE by up to a factor of ～130 at non-zero spatial frequencies. The results also indicate relatively robust tolerance of such scintillators to positional displacements, of up to 10?cm in the source-to-detector direction and 2?cm in the lateral direction, from their optimal focusing position, which could potentially enhance practical clinical use of focused segmented scintillators in MV AMFPIs.     

A preliminary investigation of local tomography for megavoltage CT imaging

We investigate the problem of reconstructing a two-dimensional (2-D) cross-sectional image of a tumor volume from a set of truncated MV projections that are produced by radiation therapy treatment beams. Our proposed approach is conceptually distinct from previously investigated approaches in that it utilizes a noniterative local tomography reconstruction algorithm. A local tomography reconstruction algorithm is implemented and systematically investigated using several sets of simulated and experimental MV projection data. We demonstrate that the conventional (nonlocal) filtered backprojection reconstruction algorithm cannot, in general, accurately reconstruct the edges and boundaries of low-contrast features from truncated MV projection data. We demonstrate that the local tomography algorithm is not adversely affected by projection truncation and can reconstruct accurately the boundaries of low-contrast structures within the region of interest from truncated MV projections.     

Fundamental x-ray interaction limits in diagnostic imaging detectors: spatial resolution

The practice of diagnostic x-ray imaging has been transformed with the emergence of digital detector technology. Although digital systems offer many practical advantages over conventional film-based systems, their spatial resolution performance can be a limitation. The authors present a Monte Carlo study to determine fundamental resolution limits caused by x-ray interactions in four converter materials: Amorphous silicon (a-Si), amorphous selenium, cesium iodide, and lead iodide. The "x-ray interaction" modulation transfer function (MTF) was determined for each material and compared in terms of the 50% MTF spatial frequency and Wagner's effective aperture for incident photon energies between 10 and 150 keV and various converter thicknesses. Several conclusions can be drawn from their Monte Carlo study. (i) In low-Z (a-Si) converters, reabsorption of Compton scatter x rays limits spatial resolution with a sharp MTF drop at very low spatial frequencies (< 0.3 cycles/mm), especially above 60 keV; while in high-Z materials, reabsorption of characteristic x rays plays a dominant role, resulting in a mid-frequency (1-5 cycles/mm) MTF drop. (ii) Coherent scatter plays a minor role in the x-ray interaction MTF. (iii) The spread of energy due to secondary electron (e.g., photoelectrons) transport is significant only at very high spatial frequencies. (iv) Unlike the spread of optical light in phosphors, the spread of absorbed energy from x-ray interactions does not significantly degrade spatial resolution as converter thickness is increased. (v) The effective aperture results reported here represent fundamental spatial resolution limits of the materials tested and serve as target benchmarks for the design and development of future digital x-ray detectors.     

Effects of characteristic x rays on the noise power spectra and detective quantum efficiency of photoconductive x-ray detectors.

The effects of K fluorescence on the imaging performance of photoconductor-based x-ray imaging systems are investigated. A cascaded linear systems model was developed, where a parallel cascaded process was implemented to take into account the effect of K-fluorescence reabsorption on the modulation transfer function (MTF), noise power spectrum (NPS), and the spatial frequency dependent detective quantum efficiency [DQE(f)] of an imaging system. The investigation was focused on amorphous selenium (a-Se), which is the most highly developed photoconductor material for x-ray imaging. The results were compared to those obtained with Monte Carlo simulation using the same imaging condition and detector parameters, so that the validity of the cascaded linear system model could be confirmed. Our results revealed that K-fluorescence reabsorption in a-Se is responsible for a 18% drop in NPS at high spatial frequencies with an incident x-ray photon energy of E=20 keV (which is just above the K edge of 12.5 keV). When E increases to 60 keV, the effects of K-fluorescence reabsorption on NPS decrease to approximately 12% at high spatial frequencies. Because the high frequency drop is present in both MTF and NPS, the effect of K fluorescence on DQE(f) is minimal, especially for E that is much higher than the K edge. We also applied the cascaded linear system model to a newly developed compound photoconductor, lead iodide (PbI2), and found that at 60 keV there is a high frequency drop in NPS of 19%. The calculated NPS were compared to previously published measurements of PbI2 detectors.     

Demonstration of megavoltage and diagnostic x-ray imaging with hydrogenated amorphous silicon arrays.

Flat-panel imagers consisting of the first large area, self-scanning, pixelated, solid-state arrays made with hydrogenated amorphous silicon (a-Si:H) are under development by the authors for applications in diagnostic x-ray and megavoltage radiotherapy imaging. The arrays, designated by the acronym MASDA for multi-element amorphous silicon detector array, consist of a two-dimensional array of a-Si:H photodiodes and thin-film transistors and are used in conjunction with scintillating materials. Imagers utilizing MASDA arrays offer a variety of advantages over existing technologies. This article presents initial megavoltage and diagnostic-quality x-ray images taken with several such arrays including the first examples of anatomical-phantom images. The external readout electronics and imaging techniques required to obtain such images are outlined, the construction, operation, and advantages of the arrays briefly reviewed, and the future potential of this new technology discussed.     

Fundamental x-ray interaction limits in diagnostic imaging detectors: frequency-dependent Swank noise

A frequency-dependent x-ray Swank factor based on the "x-ray interaction" modulation transfer function and normalized noise power spectrum is determined from a Monte Carlo analysis. This factor was calculated in four converter materials: amorphous silicon (a-Si), amorphous selenium (a-Se), cesium iodide (CsI), and lead iodide (PbI2) for incident photon energies between 10 and 150 keV and various converter thicknesses. When scaled by the quantum efficiency, the x-ray Swank factor describes the best possible detective quantum efficiency (DQE) a detector can have. As such, this x-ray interaction DQE provides a target performance benchmark. It is expressed as a function of (Fourier-based) spatial frequency and takes into consideration signal and noise correlations introduced by reabsorption of Compton scatter and photoelectric characteristic emissions. It is shown that the x-ray Swank factor is largely insensitive to converter thickness for quantum efficiency values greater than 0.5. Thus, while most of the tabulated values correspond to thick converters with a quantum efficiency of 0.99, they are appropriate to use for many detectors in current use. A simple expression for the x-ray interaction DQE of digital detectors (including noise aliasing) is derived in terms of the quantum efficiency, x-ray Swank factor, detector element size, and fill factor. Good agreement is shown with DQE curves published by other investigators for each converter material, and the conditions required to achieve this ideal performance are discussed. For high-resolution imaging applications, the x-ray Swank factor indicates: (i) a-Si should only be used at low-energy (e.g., mammography); (ii) a-Se has the most promise for any application below 100 keV; and (iii) while quantum efficiency may be increased at energies just above the K edge in CsI and PbI2, this benefit is offset by a substantial drop in the x-ray Swank factor, particularly at high spatial frequencies.     

Coregistered tomographic x-ray and optical breast imaging: initial results

We describe what is, to the best of our knowledge, the first pilot study of coregistered tomographic x-ray and optical breast imaging. The purpose of this pilot study is to develop both hardware and data processing algorithms for a multimodality imaging method that provides information that neither x-ray nor diffuse optical tomography (DOT) can provide alone. We present in detail the instrumentation and algorithms developed for this multimodality imaging. We also present results from our initial pilot clinical tests. These results demonstrate that strictly coregistered x-ray and optical images enable a detailed comparison of the two images. This comparison will ultimately lead to a better understanding of the relationship between the functional contrast afforded by optical imaging and the structural contrast provided by x-ray imaging.     

Choice of material for HVL measurements in megavoltage x-ray beams

The relative sensitivity of the half-value layer (HVL) method as a quality index for megavoltage x-ray beams is examined by theoretical calculation and experimental measurements for 4-, 6-, 10-, and 25-MV x-ray beams. It is shown that lower atomic number materials are more sensitive to beam quality changes than higher atomic number materials, and that aluminum is a reasonable choice of material for HVL measurements in megavoltage x-ray beams. Further, it was found that the HVL in aluminum or polystyrene is a more sensitive index of spectral quality than the ionization ratio method, recommended by recent dosimetry protocols.     

Detective quantum efficiency of a direct-detection active matrix flat panel imager at megavoltage energies

The use of an amorphous selenium (a-Se) based direct-detection active matrix flat-panel imager (AMFPI) is studied for megavoltage imaging. The detector consists of a 1.2 mm copper front plate and 200 microm a-Se layer, and has a 85 microm pixel pitch. The Modulation Transfer Function (MTF), Noise Power Spectrum (NPS), and Detective Quantum Efficiency (DQE) are measured for 6 and 15 MV photon beams. A theoretical expression for the DQE is derived using a recently developed formalism for nonelementary cascade stages. A comparison of theory with experiment is good for the 6 and 15 MV beams. The model is used to explore the DQE for more typical pixel sizes. The results indicate that with proper modifications, such as a larger a-Se thickness, a direct flat-panel AMFPI is a very promising detector for megavoltage imaging.     

Direct conversion detectors: the effect of incomplete charge collection on detective quantum efficiency

Direct conversion detectors offer the potential for very high resolution and high quantum efficiency for x-ray imaging, however, variations in signal can arise due to incomplete charge collection. A charge transport model was developed to describe the signal and noise resulting from incomplete charge collection. This signal to noise ratio (SNR) reduction was incorporated into the cascaded systems model for a simple x-ray detector. A new excess noise factor, A(c) (termed the collection noise factor) is introduced to describe the reduction in detective quantum efficiency (DQE). The DQE is proportional to the product of the quantum efficiency and the collection noise factor. If the trapping cross sections for electrons and holes are very different, and if the detector is biased improperly, the collection noise factor can drop to as low as 50%. In addition, the signal loss due to incomplete charge collection will reduce the DQE in the presence of added noise. Because of this, the DQE generally does not continue to improve with greater detector thickness. The collection noise factor and DQE are predicted for CdZnTe, PbI2, and Se. The optimization of detector thickness should be based not only on quantum efficiency but also on the charge collection statistics, which are influenced by bias field and polarity.