The prediction of brain metastasis (BM) response to stereotactic radiosurgery could assist clinicians when choosing BM treatments. This study investigates the prediction of in-field progression, out-of-field progression, and 1-year overall survival (OS) endpoints using a machine learning classifier. It also investigates the effect of feature type and magnetic resonance imaging (MRI) scanner variability on classifier performance. The study data set consisted of n = 110 BMs across 91 patients for which endpoints, seven clinical features, and MRI scans were available. 635 radiomic features were extracted from the MRI for the BM region-of-interest (ROI) and a 5mm BM ROI dilation. A 1000-iteration bootstrap experimental design was used with a random forest classifier to provide area under the receiver operating characteristic curve (AUC) estimates. This experimental design was used for multiple endpoints, groups of features, and data partitioning by scanner model. In-field progression, out-of-field progression, and 1-year OS were predicted with respective AUC estimates of 0.70, 0.57 and 0.66. For all endpoints, clinical and/or radiomic features from the BM ROI provided optimal performance. MR scanner variability was found to decrease classifier AUC in general, though pre-processing methods were found to counteract this effect for some scanner models. This study shows that in-field progression, out-of-field progression, and 1-year OS can all be predicted to some degree, with in-field progression being predicted most accurately. The effects of scanner variability indicate that more diverse data sets and robust methods to account for scanner variability are required before clinical translation.
KEYWORDS: Ultrasonography, Cancer, Visualization, Real time imaging, Rectum, 3D image processing, Computed tomography, Transducers, 3D metrology, 3D scanning
High-dose-rate (HDR) interstitial brachytherapy is often included in standard-of-care for gynaecological cancers. Needles are currently inserted through a perineal template without any standard real-time imaging modality to assist needle guidance, causing physicians to rely on pre-operative imaging, clinical examination, and experience. While two-dimensional (2D) ultrasound (US) is sometimes used for real-time guidance, visualization of needle placement and depth is difficult and subject to variability and inaccuracy in 2D images. The close proximity to critical organs, in particular the rectum and bladder, can lead to serious complications. We have developed a three-dimensional (3D) transrectal US system and are investigating its use for intra-operative visualization of needle positions used in HDR gynaecological brachytherapy. As a proof-of-concept, four patients were imaged with post-insertion 3D US and x-ray CT. Using software developed in our laboratory, manual rigid registration of the two modalities was performed based on the perineal template’s vaginal cylinder. The needle tip and a second point along the needle path were identified for each needle visible in US. The difference between modalities in the needle trajectory and needle tip position was calculated for each identified needle. For the 60 needles placed, the mean trajectory difference was 3.23 ± 1.65° across the 53 visible needle paths and the mean difference in needle tip position was 3.89 ± 1.92 mm across the 48 visible needles tips. Based on the preliminary results, 3D transrectal US shows potential for the development of a 3D US-based needle guidance system for interstitial gynaecological brachytherapy.
The emergence of helical tomotherapy has provided a unique opportunity to combine aspects of diagnostic computed tomography and radiation treatment. Daily megavoltage computed tomography (MVCT) scans of a patient in the treatment position provide an ideal input for adaptive radiation therapy, whereby the quantitative CT knowledge of a patient from a treatment fraction combined with the knowledge of the therapy dose distribution can be used to alter and correct for the dose delivery in subsequent fractions. In order for adaptive radiotherapy to be successful, the quantitative information from the CT scan must be as accurate as possible in geometric and dosimetric information. One potential impediment to the accuracy of the CT data values is x-ray scatter. In our study, we quantify the magnitude of x-ray scatter in the tomotherapy (fan-beam) MVCT system, based on Monte Carlo simulations of the scatter-to-primary ratio (SPR) as a function of incident x-ray energy, fan-beam slice thickness, patient size, and air gap distance. Furthermore, based on these SPR values, the impact on CT number accuracy is shown, and the implications for adaptive radiotherapy (i.e. dose reconstruction) are discussed. Under
conditions common to tomotherapy MVCT scanning, SPR values range from 0.02 to 0.16 (depending on the size of the phantom), and are generally lower than those encountered in diagnostic cone-beam CT and megavoltage portal imaging. These SPR values are sufficient enough to introduce CT number errors as high as 5 HU in soft-tissue and 100 HU in bone. The implication of this inaccuracy for adaptive radiotherapy would be to cause potential dose calculation errors during dose reconstruction and treatment re-planning.
KEYWORDS: Sensors, Modulation transfer functions, Monte Carlo methods, Fiber optics, Signal detection, Interference (communication), Digital imaging, Fiber optics sensors, Imaging systems, Medical imaging
Cascaded-systems analyses have been used successfully by many investigators to describe signal and noise transfer in quantum-based x-ray detectors in medical imaging. However, the Fourier-based linear-systems approach is only valid when assumptions of linearity and shift invariance are satisfied. Digital detectors, in which a bounded image signal is spatially integrated in discrete detector elements, are not shift invariant in their response. In addition, many detectors make use of fiber optics or structured phosphors such as CsI to pass light to a photodetector-both of which have a shift-variant response. These issues raise serious concerns regarding the validity of Fourier-based approaches for describing the signal and noise performance of these detectors.
We have used a Monte Carlo approach to compare the image Wiener noise power spectrum (NPS) with that predicted using a Fourier-based approach when these assumptions fail. It is shown that excellent agreement is obtained between Monte Carlo results and those obtained using a Fourier-based wide-sense cyclostationary analysis, including the description of noise aliasing. A simple model of a digital detector coupled to a fiber optic bundle is described using a novel cascaded cyclostationary approach in which image quanta are integrated over fiber elements and then randomly re-distributed at the fiber output. While the image signal sometimes contains significant non-stationary beating artifacts, the Monte Carlo results generally show good agreement with Fourier models of the NPS when noise measurements are made over a sufficiently large region of interest.
KEYWORDS: X-rays, Modulation transfer functions, Monte Carlo methods, Sensors, X-ray detectors, Mammography, Quantum efficiency, X-ray imaging, Selenium, Point spread functions
An often neglected assumption related to detector performance metrics such as the modulation transfer function (MTF), noise power spectrum (NPS), and detective quantum efficiency (DQE) is that they only apply to a small region around the centre of an x-ray image. In the periphery of an image, image formation is from obliquely incident x rays. These off-axis x rays will introduce an additional degrading effect on the above detector performance metrics. In our study, we use Monte Carlo simulations to quantify the effects of off-axis radiation on the MTF, NPS, and DQE on common diagnostic x-ray
detectors. In our simulations, we vary the incident angle of x rays between 0° and 12°, which is a typical range of divergence in
diagnostic x-ray imaging. In the case of amorphous selenium, our results show that off-axis incident x rays degrade the MTF above 5 cycles/mm with increasing severity at higher incident angles and x-ray energy, and more importantly has very little effect on the NPS. Hence, the impact is more severe on the DQE due to the MTF squared dependency. For an incident x-ray angle of 12° (~13 cm from central axis or chest wall in mammography), the DQE falls to 50% of its initial value at 10 and 7 cycles/mm for x-ray energies of 20 and 40 keV, respectively. This loss of signal-to-noise ratio may be most significant near the skin line in mammography studies.
KEYWORDS: X-ray detectors, X-rays, Modulation transfer functions, Sensors, Monte Carlo methods, Spatial resolution, Point spread functions, Medical imaging, Diagnostics, Selenium
The development of new detectors for diagnostic x-ray imaging is a complex and expensive endeavour. An understanding of fundamental performance potential and limitations is therefore critical to the wise allocation of research resources. We present a Monte Carlo study in which the fundamental spatial resolution limitations imposed by x-ray interactions were determined for both direct conversion amorphous selenium (a-Se) and indirect conversion cesium iodide (CsI) detectors. Using a simulated infinitesimal x-ray beam, the absorbed energy point spread function (PSF) in each detector material was scored within rectilinear bin sizes of 5 mm for incident x-ray energies between 10 and 100 keV. The modulation transfer function (MTF) was determined from each simulated PSF and characterized in terms of the 50% MTF frequency, f50, and the equivalent passband, Ne. Both materials demonstrated: (i) a drop in f50 (a-Se: 25%, CsI: 85%) and Ne (a-Se: 45%, CsI: 75%) immediately above the K-edge energy due to re-absorption of characteristic radiation, and (ii) a moderate recovery of f50 and Ne levels with further increase in energy. In addition, within the diagnostic energy range and spatial frequency range of 0 -- 20 cycles/mm, the values of the fundamental MTF due to x-ray interactions remain above 50%. In general, we conclude that existing amorphous selenium and cesium iodide detectors operate far from fundamental spatial resolution limits in both mammography and radiography applications. Further reduction in detector element size will potentially improve spatial resolution in these detectors.
Access to the requested content is limited to institutions that have purchased or subscribe to SPIE eBooks.
You are receiving this notice because your organization may not have SPIE eBooks access.*
*Shibboleth/Open Athens users─please
sign in
to access your institution's subscriptions.
To obtain this item, you may purchase the complete book in print or electronic format on
SPIE.org.
INSTITUTIONAL Select your institution to access the SPIE Digital Library.
PERSONAL Sign in with your SPIE account to access your personal subscriptions or to use specific features such as save to my library, sign up for alerts, save searches, etc.