Engineering Papers⌕ Search

SEARCH · Engineering Papers

Results for “Randomization”

Search indexed NASA NTRS and DOE OSTI research on propulsion, heat transfer, battery materials and energy systems. Follow report and document links to the original sources.

Quote a phrase for an exact phrase match. Source license links do not imply unrestricted reuse.

At least 289 records · Page 16

Fast generation of sparse random kernel graphs

The development of kernel-based inhomogeneous random graphs has provided models that are flexible enough to capture many observed characteristics of real networks, and that are also mathematically tractable. We specify a class of inhomogeneous random graph models, called random kernel graphs, that produces sparse graphs with tunable graph properties, and we develop an efficient generation algorithm to sample random instances from this model. As real-world networks are usually large, it is essential that the run-time of generation algorithms scales better than quadratically in the number of vertices n. We show that for many practical kernels our algorithm runs in time at most ο(n(logn)²). As an example, we show how to generate samples of power-law degree distribution graphs with tunable assortativity.

97 MATHEMATICS AND COMPUTING↗

HTR-PROTEUS Pebble Bed Experimental Program Core 4: Random Packing With a 1:1 Moderator-to-Fuel Pebble Ratio

In its deployment as a pebble bed reactor (PBR) critical facility from 1992 to 1996, the PROTEUS facility was designated as HTR-PROTEUS. This experimental program was performed as part of an International Atomic Energy Agency (IAEA) Coordinated Research Project (CRP) on the Validation of Safety Related Physics Calculations for Low Enriched HTGRs. Within this project, critical experiments were conducted for graphite moderated LEU systems to determine core reactivity, flux and power profiles, reaction-rate ratios, the worth of control rods, both in-core and reflector based, the worth of burnable poisons, kinetic parameters, and the effects of moisture ingress on these parameters. One benchmark experiment was evaluated in this report: Core 4. Core 4 represents the only configuration with random pebble packing in the HTR-PROTEUS series of experiments, and has a moderator-to-fuel pebble ratio of 1:1. Three random configurations were performed. The initial configuration, Core 4.1, was rejected because the method for pebble loading, separate delivery tubes for the moderator and fuel pebbles, may not have been completely random; this core loading was rejected by the experimenters. Cores 4.2 and 4.3 were loaded using a single delivery tube, eliminating the possibility for systematic ordering effects. The second and third cores differed slightly in the quantity of pebbles loaded (40 each of moderator and fuel pebbles), stacked height of the pebbles in the core cavity (0.02 m), withdrawn distance of the stainless steel control rods (20 mm), and withdrawn distance of the autorod (30 mm). The 34 coolant channels in the upper axial reflector and the 33 coolant channels in the lower axial reflector were open. Additionally, the axial graphite fillers used in all other HTR-PROTEUS configurations to create a 12-sided core cavity were not used in the randomly packed cores. Instead, graphite fillers were placed on the cavity floor, creating a funnel-like base, to discourage ordering effects during pebble loading. Core 4 was determined to be acceptable benchmark experiment.

11 NUCLEAR FUEL CYCLE AND FUEL MATERIALS↗

HTR-proteus pebble bed experimental program core 4: random packing with a 1:1 moderator-to-fuel pebble ratio

In its deployment as a pebble bed reactor (PBR) critical facility from 1992 to 1996, the PROTEUS facility was designated as HTR-PROTEUS. This experimental program was performed as part of an International Atomic Energy Agency (IAEA) Coordinated Research Project (CRP) on the Validation of Safety Related Physics Calculations for Low Enriched HTGRs. Within this project, critical experiments were conducted for graphite moderated LEU systems to determine core reactivity, flux and power profiles, reaction-rate ratios, the worth of control rods, both in-core and reflector based, the worth of burnable poisons, kinetic parameters, and the effects of moisture ingress on these parameters. One benchmark experiment was evaluated in this report: Core 4. Core 4 represents the only configuration with random pebble packing in the HTR-PROTEUS series of experiments, and has a moderator-to-fuel pebble ratio of 1:1. Three random configurations were performed. The initial configuration, Core 4.1, was rejected because the method for pebble loading, separate delivery tubes for the moderator and fuel pebbles, may not have been completely random; this core loading was rejected by the experimenters. Cores 4.2 and 4.3 were loaded using a single delivery tube, eliminating the possibility for systematic ordering effects. The second and third cores differed slightly in the quantity of pebbles loaded (40 each of moderator and fuel pebbles), stacked height of the pebbles in the core cavity (0.02 m), withdrawn distance of the stainless steel control rods (20 mm), and withdrawn distance of the autorod (30 mm). The 34 coolant channels in the upper axial reflector and the 33 coolant channels in the lower axial reflector were open. Additionally, the axial graphite fillers used in all other HTR-PROTEUS configurations to create a 12-sided core cavity were not used in the randomly packed cores. Instead, graphite fillers were placed on the cavity floor, creating a funnel-like base, to discourage ordering effects during pebble loading. Core 4 was determined to be acceptable benchmark experiment.

11 NUCLEAR FUEL CYCLE AND FUEL MATERIALS↗

Machine Learning Inference of Random Medium Properties

Earth materials are heterogeneous across a range of spatial scales, but the resolvability of small structures is limited by sparse data coverage, noise, bandlimitedness, and other difficulties. In practice, heterogeneities below a certain size cannot be recovered from seismic data except through statistical medium descriptions, which even then can be difficult to uniquely determine. To improve the characterization of such heterogeneities, we develop a novel supervised machine learning (ML) model that provides insight about the recoverability of statistical medium properties from elastic waveform data and succeeds despite cycle-skipping and other challenges well known from elastic waveform inversion. We demonstrate the approach using random media generated by superimposing self-affine random variations on homogeneous and layered background structures. After training on sparsely-recorded, high-frequency waveforms from hundreds of different random medium realizations, we show the ability of our ML model to recover correlation lengths and other statistical properties of interest to near-surface and crustal seismology, among other fields. For frequency passbands and spatial offsets encountered in seismology, Gaussian correlation lengths and the amplitude of the random variations relative to the background model are recovered even in challenging scenarios involving unknown medium parameters, complex crustal structures, and low signal-to-noise ratio. In comparison, von Kármán correlation lengths, which are related to larger-wavelength variations of the medium than Gaussian correlation lengths, are not as well recovered. These results provide one of the first and most systematic investigations of the recoverability of statistical properties of heterogeneities below the resolution limit of deterministic seismic tomography, and suggest practical ML strategies for high-frequency waveform seismology.

58 GEOSCIENCES↗

Development of a High Min-Entropy Quantum Random Number Generator Based on Amplified Spontaneous Emission

We present the theory, architecture, and performance characteristics of a quantum random number generator (QRNG) which operates in a PCI express form factor-compatible plug-and-play design. The QRNG relies on a thermal light source (in this case, amplified spontaneous emission), which exhibits photon bunching according to the Bose–Einstein (BE) statistics. We demonstrate that 98.7% of the unprocessed random bit stream min-entropy is traceable to the BE (quantum) signal. The classical component is then removed using a non-reuse shift-XOR protocol, and the final random numbers are generated at a 200 Mbps rate and shown to pass the statistical randomness test suites FIPS 140-2, Alphabit, SmallCrush, DIEHARD, and Rabbit of the TestU01 library.

71 CLASSICAL AND QUANTUM MECHANICS, GENERAL PHYSIC↗

A stochastic analysis of steady and transient heat conduction in random media using a homogenization approach

We present a new stochastic analysis for steady and transient one-dimensional heat conduction problem based on the homogenization approach. Thermal conductivity is assumed to be a random field K consisting of random variables of a total number N. Both steady and transient solutions T are expressed in terms of the homogenized solution (symbol) and its spatial derivatives (equation), where homogenized solution (symbol) is obtained by solving the homogenized equation with effective thermal conductivity. Both mean and variance of stochastic solutions can be obtained analytically for K field consisting of independent identically distributed (i.i.d) random variables. The mean and variance of T are shown to be dependent only on the mean and variance of these i.i.d variables, not the particular form of probability distribution function of i.i.d variables. Variance of temperature field T can be separated into two contributions: the ensemble contribution (through the homogenized temperature (symbol)); and the configurational contribution (through the random variable Ln(x)Ln(x)). The configurational contribution is shown to be proportional to the local gradient of (symbol). Large uncertainty of T field was found at locations with large gradient of (symbol) due to the significant configurational contributions at these locations. Numerical simulations were implemented based on a direct Monte Carlo method and good agreement is obtained between numerical Monte Carlo results and the proposed stochastic analysis.

99 GENERAL AND MISCELLANEOUS↗

Detector Calibration to Quantify Random Summing in HPGe Spectroscopy

While calibrating a High Purity Germanium (HPGe) Radiation Detector for Xenon-133 (133Xe) the process of random summing occurs which leads to excess counts in the 160.6keV peak from random summing of the 81keV photons at the 162keV energy. Multiple samples of 133Xe with varying levels of radioactivity were assayed to determine a relationship between the occurrence of random summing and the net area of the 160.6keV peak. When the calculated efficiencies in the 160.6keV peak were compared to the theoretical efficiency obtained from a multi-energy calibration at 160.6keV, the amount of random summing could be estimated as excess observed efficiency and correlated to the sample activity. While a trend is apparent in the data, the mechanisms that contribute to the trend are still being studied so the data can be accurately correlated.

07 ISOTOPE AND RADIATION SOURCES↗

A dynamically adaptive wavelet approach to stochastic computations based on polynomial chaos - capturing all scales of random modes on independent grids

Highlights: {yields} New approach for stochastic computations based on polynomial chaos. {yields} Development of dynamically adaptive wavelet multiscale solver using space refinement. {yields} Accurate capture of steep gradients and multiscale features in stochastic problems. {yields} All scales of each random mode are captured on independent grids. {yields} Numerical examples demonstrate the need for different space resolutions per mode. - Abstract: In stochastic computations, or uncertainty quantification methods, the spectral approach based on the polynomial chaos expansion in random space leads to a coupled system of deterministic equations for the coefficients of the expansion. The size of this system increases drastically when the number of independent random variables and/or order of polynomial chaos expansions increases. This is invariably the case for large scale simulations and/or problems involving steep gradients and other multiscale features; such features are variously reflected on each solution component or random/uncertainty mode requiring the development of adaptive methods for their accurate resolution. In this paper we propose a new approach for treating such problems based on a dynamically adaptive wavelet methodology involving space-refinement on physical space that allows all scales of each solution component to be refined independently of the rest. We exemplify this using the convection-diffusion model with random input data and present three numerical examples demonstrating the salient features of the proposed method. Thus we establish a new, elegant and flexible approach for stochastic problems with steep gradients and multiscale features based on polynomial chaos expansions.

97 MATHEMATICS AND COMPUTING↗

Randomized Algorithms for Low-Rank Matrix and Tensor Decompositions

This paper surveys randomized algorithms in numerical linear algebra for low-rank decompositions of matrices and tensors. The survey begins with a review of classical matrix algorithms that can be accelerated by randomized dimensionality reduction, such as the singular value decomposition (SVD) or interpolative (ID) and CUR decompositions. Recent advances in randomized dimensionality reduction are discussed, including new methods of fast matrix sketching and sampling techniques, which are incorporated into classical matrix algorithms for fast low-rank matrix approximations. The extension of randomized matrix algorithms to tensors is then explored for several low-rank tensor decompositions in the CP and Tucker formats, including the higher-order SVD, ID, and CUR decomposition.

Pearce, Katherine J. [The University of Texas at A↗

A Meta-analysis of Randomized Trials Comparing Surgery versus Endovascular Therapy for Thrombosed Arteriovenous Fistulas and Grafts in Hemodialysis

Purpose. To carry out a systematic review of randomized trials comparing surgery vs. endovascular therapy for occluded fistulas and grafts. Methods. All randomized trials which compared surgery and endovascular therapy for occluded fistulas and grafts were retrieved from 1990 onwards. The following search terms were used: 'haemodialysis,' 'thrombosis,' 'arteriovenous fistula,' 'arteriovenous shunt,' 'end stage renal failure' on Medline and PubMed. The results of the pooled data were analysed by a fixed-effect model. Results. There were no randomized trials comparing surgery vs. endovascular therapy for native fistulas and vein grafts. Six randomized studies reporting on 573 occluded grafts were identified. Technical success, need for access line and primary patency at 30 days were similar between the two groups (odds ratio [OR] 1.40, 95 % confidence interval [CI] 0.91-2.14; OR 0.77, 95 % CI 0.44-1.34; and OR 1.15, 95 % CI 0.79-1.68, respectively). There was no significant difference in morbidity at 30 days between groups (OR 1.12, 95 % CI 0.67-1.86). There were no statistical difference between the two groups for 1 year primary patency (OR 2.08, 95 % CI 0.97-4.45). Primary assisted patency at 1 year was better with surgery (OR 3.03, 95 % CI 1.12-8.18) in a single study. Conclusion. Comparable results to surgery have been achieved with endovascular techniques for occluded prosthetic grafts for dialysis access. Long-term data comparing the two groups were lacking. Further trials designed to encompass variation in methods are warranted in order to obtain the best available evidence particularly for native fistulas.

62 RADIOLOGY AND NUCLEAR MEDICINE↗

Induction Chemotherapy and Continuous Hyperfractionated Accelerated Radiotherapy (CHART) for Patients With Locally Advanced Inoperable Non-Small-Cell Lung Cancer: The MRC INCH Randomized Trial

Purpose: Recent clinical trials and meta-analyses have shown that both CHART (continuous hyperfractionated accelerated radiation therapy) and induction chemotherapy offer a survival advantage over conventional radical radiotherapy for patients with inoperable non-small cell-lung cancer (NSCLC). This multicenter randomized controlled trial (INCH) was set up to assess the value of giving induction chemotherapy before CHART. Methods and Materials: Patients with histologically confirmed, inoperable, Stage I-III NSCLC were randomized to induction chemotherapy (ICT) (three cycles of cisplatin-based chemotherapy followed by CHART) or CHART alone. Results: Forty-six patients were randomized (23 in each treatment arm) from 9 UK centers. As a result of poor accrual, the trial was closed in December 2007. Twenty-eight patients were male, 28 had squamous cell histology, 34 were Stage IIIA or IIIB, and all baseline characteristics were well balanced between the two treatment arms. Seventeen (74%) of the 23 ICT patients completed the three cycles of chemotherapy. All 42 (22 CHART + 20 ICT) patients who received CHART completed the prescribed treatment. Median survival was 17 months in the CHART arm and 25 months in the ICT arm (hazard ratio of 0.60 [95% CI 0.31-1.16], p = 0.127). Grade 3 or 4 adverse events (mainly fatigue, dysphagia, breathlessness, and anorexia) were reported for 13 (57%) CHART and 13 (65%) ICT patients. Conclusions: This small randomized trial indicates that ICT followed by CHART is feasible and well tolerated. Despite closing early because of poor accrual, and so failing to show clear evidence of a survival benefit for the additional chemotherapy, the results suggest that CHART, and ICT before CHART, remain important options for the treatment of inoperable NSCLC and deserve further study.

62 RADIOLOGY AND NUCLEAR MEDICINE↗

Randomized Controlled Trial of Forward-Planned Intensity Modulated Radiotherapy for Early Breast Cancer: Interim Results at 2 Years

Purpose: This single-center randomized trial was designed to investigate whether intensity-modulated radiotherapy (IMRT) reduces late toxicity in patients with early-stage breast cancer. Methods and Materials: The standard tangential plans of 1,145 nonselected patients were analyzed. The patients with inhomogeneous plans were randomized to a simple method of forward-planned IMRT or standard radiotherapy (RT). The primary endpoint was serial photographic assessment of breast shrinkage. Results: At 2 years, no significant difference was found in the development of any photographically assessed breast shrinkage between the patients randomized to the interventional or control group (odds ratio, 1.51; 95% confidence interval, 0.83-1.58; p = .41). The patients in the control group were more likely to develop telangiectasia than those in the IMRT group (odds ratio, 1.68; 95% confidence interval 1.13-2.40; p = .009). Poor baseline surgical cosmesis resulted in poor overall cosmesis at 2 years after RT. In patients who had good surgical cosmesis, those randomized to IMRT were less likely to deteriorate to a moderate or poor overall cosmesis than those in the control group (odds ratio, 0.63; 95% confidence interval, 0.39-1.03, p = .061). Conclusions: IMRT can lead to a significant reduction in telangiectasia at comparatively early follow-up of only 2 years after RT completion. An important component of breast induration and shrinkage will actually result from the surgery and not from the RT. Surgical cosmesis is an important determinant of overall cosmesis and could partially mask the longer term benefits of IMRT at this early stage.

62 RADIOLOGY AND NUCLEAR MEDICINE↗

Conjunctival Autograft Alone or Combined With Adjuvant Beta-Radiation? A Randomized Clinical Trial

Purpose: To evaluate the effectiveness and safety of postoperative low single-dose of beta-irradiation ({beta}-RT) in pterygium comparing conjunctival autograft (CAG) surgery with CAG plus adjuvant {beta}-RT in a randomized clinical trial. Methods: This trial was designed as a prospective, randomized, single-center study. Surgery was performed in all cases according to the CAG technique. One hundred and eight pterygia were postoperatively randomized to CAG + {beta}-RT or CAG alone. In the case of {beta}-RT, a (90) Sr eye applicator was used to deliver 10 Gy to the sclera surface at a dose rate of between 200 and 250 cGy/min. After treatment, both an ophthalmologist and a radiation oncologist performed the follow-up examinations. The accumulated data were analyzed using a group sequential test. Results: Between February 2008 and September 2008, 116 eyes with primary pterygium were operated on according to the trial protocol. Adjuvant treatment was performed within 24 h postoperatively. Eight patients were lost to follow-up, resulting in 108 patients who could be analyzed. At a mean follow-up of 18 months (range, 8-33), in the 54 eyes randomized to receive CAG + {beta}-RT, 5 relapses occurred compared with 12 recurrences in the 54 eyes in CAG, for a crude control rate of 90.8 % vs. 78%; p = 0.032, respectively. The treatment complications as hyperemia, total dehiscence of the autograft and dellen were significantly more frequent in the CAG (p < 0.05). The arm of {beta}-RT resulted in better cosmetic results and improves of symptoms than CAG. Conclusions: A low single-dose of {beta}-RT of 10 Gy after CAG surgery was a simple, effective, and safe treatment that reduced the risk of primary pterygium recurrence, improved symptoms after surgery, resulting in a better cosmetic effect than only CAG.

62 RADIOLOGY AND NUCLEAR MEDICINE↗

A Phase 3 Trial of 2 Years of Androgen Suppression and Radiation Therapy With or Without Adjuvant Chemotherapy for High-Risk Prostate Cancer: Final Results of Radiation Therapy Oncology Group Phase 3 Randomized Trial NRG Oncology RTOG 9902

Purpose: Long-term (LT) androgen suppression (AS) with radiation therapy (RT) is a standard treatment of high-risk, localized prostate cancer (PCa). Radiation Therapy Oncology Group 9902 was a randomized trial testing the hypothesis that adjuvant combination chemotherapy (CT) with paclitaxel, estramustine, and oral etoposide plus LT AS plus RT would improve overall survival (OS). Methods and Materials: Patients with high-risk PCa (prostate-specific antigen 20-100 ng/mL and Gleason score [GS] ≥7 or clinical stage ≥T2 and GS ≥8) were randomized to RT and AS (AS + RT) alone or with adjuvant CT (AS + RT + CT). CT was given as four 21-day cycles, delivered beginning 28 days after 70.2 Gy of RT. AS was given as luteinizing hormone-releasing hormone for 24 months, beginning 2 months before RT plus an oral antiandrogen for 4 months before and during RT. The study was designed based on a 6% improvement in OS from 79% to 85% at 5 years, with 90% power and a 2-sided alpha of 0.05. Results: A total of 397 patients (380 eligible) were randomized. The patients had high-risk PCa, 68% with GS 8 to 10 and 34% T3 to T4 tumors, and median prostate-specific antigen of 22.6 ng/mL. The median follow-up period was 9.2 years. The trial closed early because of excess thromboembolic toxicity in the CT arm. The 10-year results for all randomized patients revealed no significant difference between the AS + RT and AS + RT + CT arms in OS (65% vs 63%; P=.81), biochemical failure (58% vs 54%; P=.82), local progression (11% vs 7%; P=.09), distant metastases (16% vs 14%; P=.42), or disease-free survival (22% vs 26%; P=.61). Conclusions: NRG Oncology RTOG 9902 showed no significant differences in OS, biochemical failure, local progression, distant metastases, or disease-free survival with the addition of adjuvant CT to LT AS + RT. The trial results provide valuable data regarding the natural history of high-risk PCa treated with LT AS + RT and have implications for the feasibility of clinical trial accrual and tolerability using CT for PCa.

62 RADIOLOGY AND NUCLEAR MEDICINE↗

Radiation Therapy Intensification for Solid Tumors: A Systematic Review of Randomized Trials

Purpose: To systematically review the outcomes of randomized trials testing radiation therapy (RT) intensification, including both dose escalation and/or the use of altered fractionation, as a strategy to improve disease control for a number of malignancies. Methods and Materials: We performed a literature search to identify randomized trials testing RT intensification for cancers of the central nervous system, head and neck, breast, lung, esophagus, rectum, and prostate. Findings were described qualitatively. Where adequate data were available, pooled estimates for the effect of RT intensification on local control (LC) or overall survival (OS) were obtained using the inverse variance method. Results: In primary central nervous system tumors, esophageal cancer, and rectal cancer, randomized trials have not demonstrated that RT intensification improves clinical outcomes. In breast cancer and prostate cancer, dose escalation has been shown to improve LC or biochemical disease control but not OS. Radiation therapy intensification may improve LC and OS in head and neck and lung cancers, but these benefits have generally been limited to studies that did not incorporate concurrent chemotherapy. Conclusions: In randomized trials, the benefits of RT intensification have largely been restricted to trials in which concurrent chemotherapy was not used. Novel strategies to optimize the incorporation of RT in the multimodality treatment of solid tumors should be explored.

62 RADIOLOGY AND NUCLEAR MEDICINE↗

Randomized Clinical Trial Comparing Proton Beam Radiation Therapy with Transarterial Chemoembolization for Hepatocellular Carcinoma: Results of an Interim Analysis

Purpose: To describe results of a planned interim analysis of a prospective, randomized clinical trial developed to compare treatment outcomes among patients with newly diagnosed hepatocellular carcinoma (HCC). Methods and Materials: Eligible subjects had either clinical or pathologic diagnosis of HCC and met either Milan or San Francisco transplant criteria. Patients were randomly assigned to transarterial chemoembolization (TACE) or to proton beam radiation therapy. Patients randomized to TACE received at least 1 TACE with additional TACE for persistent disease. Proton beam radiation therapy was delivered to all areas of gross disease to a total dose of 70.2 Gy in 15 daily fractions over 3 weeks. The primary endpoint was progression-free survival, with secondary endpoints of overall survival, local tumor control, and treatment-related toxicities as represented by posttreatment days of hospitalization. Results: At the time of this analysis 69 subjects were available for analysis. Of these, 36 were randomized to TACE and 33 to proton. Total days of hospitalization within 30 days of TACE/proton was 166 and 24 days, respectively (P<.001). Ten TACE and 12 proton patients underwent liver transplantation after treatment. Viable tumor identified in the explanted livers after TACE/proton averaged 2.4 and 0.9 cm, respectively. Pathologic complete response after TACE/proton was 10%/25% (P=.38). The 2-year overall survival for all patients was 59%, with no difference between treatment groups. Median survival time was 30 months (95% confidence interval 20.7-39.3 months). There was a trend toward improved 2-year local tumor control (88% vs 45%, P=.06) and progression-free survival (48% vs 31%, P=.06) favoring the proton beam treatment group. Conclusions: This interim analysis indicates similar overall survival rates for proton beam radiation therapy and TACE. There is a trend toward improved local tumor control and progression-free survival with proton beam. There are significantly fewer hospitalization days after proton treatment, which may indicate reduced toxicity with proton beam therapy.

62 RADIOLOGY AND NUCLEAR MEDICINE↗

Reduced Mortality With Partial-Breast Irradiation for Early Breast Cancer: A Meta-Analysis of Randomized Trials

Purpose: With earlier detection and more effective treatment, mortality from breast cancer continues to fall and it has become increasingly important to reduce the toxicity of treatments. Partial-breast radiation therapy, which focuses radiation to the tumor bed, may achieve this aim. We analyzed mortality differences in randomized trials of partial-breast irradiation (PBI). Methods and Materials: We included data from published randomized trials of PBI (alone or as part of a risk-adapted approach) versus whole-breast irradiation (WBI) for invasive breast cancer suitable for breast-conserving therapy. We identified trials using PubMed and Google searches with the terms “partial breast irradiation” OR “intraoperative radiotherapy” OR “IMRT” OR (“accelerated” AND “radiation”) AND “randomised/randomized,” as well as through discussion with colleagues in the field. We calculated the proportion of patients who had events in each randomized arm at 5 years' follow-up and created a forest plot using Stata, version 14.1. Results: We identified 9 randomized trials of PBI versus WBI in invasive breast cancer; 5-year outcomes were available for non–breast cancer mortality in 5 trials (n=4489) and for breast cancer mortality in 4 trials (n=4231). The overall mortality was 4.9%. There was no detectable heterogeneity between the trials for any of the outcomes. There was no difference in the proportion of patients dying of breast cancer (difference, 0.000% [95% confidence interval (CI), −0.7 to +0.7]; P=.999). Non–breast cancer mortality with PBI was lower than with WBI (difference, 1.1% [95% CI, −2.1% to −0.2%]; P=.023). Total mortality with PBI was also lower than with WBI (difference, 1.3% [95% CI, −2.5% to 0.0%]; P=.05). Conclusions: Use of PBI instead of WBI in selected patients results in a lower 5-year non–breast cancer and overall mortality, amounting to a 25% reduction in relative terms. This information should be included when breast-conserving therapy is proposed to a patient.

62 RADIOLOGY AND NUCLEAR MEDICINE↗

Bayesian estimation of Karhunen–Loève expansions; A random subspace approach

One of the most widely-used statistical procedures for dimensionality reduction of high dimensional random fields is Principal Component Analysis (PCA), which is based on the Karhunen-Lo eve expansion (KLE) of a stochastic process with finite variance. The KLE is analogous to a Fourier series expansion for a random process, where the goal is to find an orthogonal transformation for the data such that the projection of the data onto this orthogonal subspace is optimal in the L 2 sense, i.e, which minimizes the mean square error. In practice, this orthogonal transformation is determined by performing an SVD (Singular Value Decomposition) on the sample covariance matrix or on the data matrix itself. Sampling error is typically ignored when quantifying the principal components, or, equivalently, basis functions of the KLE. Furthermore, it is exacerbated when the sample size is much smaller than the dimension of the random field. In this paper, we introduce a Bayesian KLE procedure, allowing one to obtain a probabilistic model on the principal components, which can account for inaccuracies due to limited sample size. The probabilistic model is built via Bayesian inference, from which the posterior becomes the matrix Bingham density over the space of orthonormal matrices. We use a modified Gibbs sampling procedure to sample on this space and then build a probabilistic Karhunen-Lo eve expansions over random subspaces to obtain a set of low-dimensional surrogates of the stochastic process. We illustrate this probabilistic procedure with a finite dimensional stochastic process inspired by Brownian motion.

97 MATHEMATICS AND COMPUTING↗