Search PubMed⌕ Search

SEARCH · Search PubMed

Results for “Software Validation”

Search indexed PubMed citations on genomics, clinical trials, systematic reviews and public health. Explore titles, authors and supplied subject terms, then open the PubMed record.

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

At least 721 records · Page 40Linked to original sources
PubMedDate not supplied

Techniques in evaluating nursing expert systems: A case study.

This study addresses the problems in evaluating nursing diagnostic artificial intelligence (AI) expert systems. Two separate experiments (N = 49) were conducted using a computer expert system. The first experiment, the "white box" experiment (n = 9), compared the diagnostic techniques applied by experience RNs against the programmed techniques used by the expert system. The second experiment, the "black box" experiment (n = 40), compared diagnostic results of beginning nurses against the computer expert systems results. In some cases the computer outperformed the nurses and vice versa. The evaluation techniques, as applied in both experiments, enhance the ability of nurses to evaluate and select AI expert systems to be used in computer-assisted diagnosis of nursing problems.

Clinical Competence↗
PubMed2006

Guided self diagnosis: an innovative approach to triage for emergency dental care.

BACKGROUND: The basic model of triage, developed more than 200 years ago by the French, is based on categorization of patient needs and thus the urgency and time required for care. This model in various forms is used in most hospitals throughout the world. METHODS: In this study, a fully computerized guided self-diagnosis system (based on a neural network design) was designed, prototyped, developed and trialled by front line non-clinically trained personnel in emergency dentistry. RESULTS: A total of 699 patients were seen as a result of the triage assessment within 14 days of their initial phone call and/or first contact. Patients categorized as requiring care on the day (Category 1) were provided more items of care (50 per cent of all items) and the greatest number of extractions (76 per cent) and endodontic treatments (78 per cent). Other categories were found to require less urgent care. CONCLUSIONS: The triage system developed in this study clearly holds significant promise in reducing the impact of emergency dental patients on dental health care systems. However, it is recognized that the system still requires some adjustment to ensure all Category 1 patients are examined before less urgent categories.

Dental Care↗
PubMed2005

Comparison of deformable and elastic foundation finite element simulations for predicting knee replacement mechanics.

Rigid body total knee replacement (TKR) models with tibiofemoral contact based on elastic foundation (EF) theory utilize simple contact pressure-surface overclosure relationships to estimate joint mechanics, and require significantly less computational time than corresponding deformable finite element (FE) methods. However, potential differences in predicted kinematics between these representations are currently not well understood, and it is unclear if the estimates of contact area and pressure are acceptable. Therefore, the objectives of the current study were to develop rigid EF and deformable FE models of tibiofemoral contact, and to compare predicted kinematics and contact mechanics from both representations during gait loading conditions with three different implant designs. Linear and nonlinear contact pressure-surface overclosure relationships based on polyethylene material properties were developed using EF theory. All other variables being equal, rigid body FE models accurately estimated kinematics predicted by fully deformable FE models and required only 2% of the analysis time. As expected, the linear EF contact model sufficiently approximated trends for peak contact pressures, but overestimated the deformable results by up to 30%. The nonlinear EF contact model more accurately reproduced trends and magnitudes of the deformable analysis, with maximum differences of approximately 15% at the peak pressures during the gait cycle. All contact area predictions agreed in trend and magnitude. Using rigid models, edge-loading conditions resulted in substantial overestimation of peak pressure. Optimal nonlinear EF contact relationships were developed for specific TKR designs for use in parametric or repetitive analyses where computational time is paramount. The explicit FE analysis method utilized here provides a unique approach in that both rigid and deformable analyses can be run from the same input file, thus enabling simple selection of the most appropriate representation for the analysis of interest.

Arthroplasty, Replacement, Knee↗
PubMed2000

A finite element method for mechanical response of hair cell ciliary bundles.

This paper describes the development of a methodology for performing a mechanical analysis of hair cell ciliary bundles. The cilia were modeled as shear deformable beams, and interconnections were modeled as two-force members. These models were incorporated into software, which performs a finite element analysis of a user-defined bundle. The algorithm incorporates aspects of the bundle such as geometric realignment and buckling of compressed side links. A sample bundle is introduced and results of modeling it are presented.

Actins↗
PubMed2002

A preliminary comparative study of two treatment planning systems developed for boron neutron capture therapy: MacNCTPlan and SERA.

Two treatment planning systems have been specifically developed for BNCT: MacNCTPlan by the Harvard-MIT group and the SERA system developed by the INEEL/Montana State University group. In order to compare the dose components computed by the treatment planning systems, and therefore the clinical results available, it is important to first compare the basic features of the codes. This study investigated the effect of the reconstruction techniques and the form of the cross-section libraries used by the two treatment planning systems on the thermal neutron fluence distributions. Six mono-directional neutron sources of 0.0253 eV, 1 keV, 5 keV, 10 keV, 100 keV, and 1 MeV were directed onto simple geometric phantoms filled with water. The general-purpose Monte Carlo radiation transport code MCNP was used as a reference tool in this study. For the examples studied, it is found that the methods used for geometry representation (in MacNCTPlan) and for cross-section representation (in SERA) were found to impact directly on the accuracy of the calculated results.

Boron Neutron Capture Therapy↗
PubMed2003

Commissioning, evaluation, quality assurance and clinical application of a virtual micro MLC technique.

Multileaf collimators (MLCs) are a valuable tool in modern radiation therapy, offering flexible and convenient field shaping. One disadvantage, however, is the undulation of the dose distribution at the edge shaped by the leaves due to the finite leaf width. An attempt to reduce the effect of this undulation is the objective of the commercial linear accelerator package HD270, which incorporates three-dimensional couch translation together with leaf adjustment to emulate finer leaf widths. In this paper we report on the commissioning and evaluation of this feature, together with the development of a process for quality assurance, as well as description of a clinical application of this technique. It is concluded that this technique could be applied reliably to situations currently utilizing MLC for shielding, with little added cost in treatment time, provided that a comprehensive quality assurance program is in place to monitor the performance of this complicated procedure.

Algorithms↗
PubMed2003

Incorporation of a combinatorial geometry package and improved scoring capabilities in the EGSnrc Monte Carlo Code system.

A description is given of a generic EGSnrc Monte Carlo user code, GenUC, which was developed as an attempt to simplify and optimize the geometry and scoring coding of EGSnrc user codes. GenUC was developed using the methodology of combinatorial geometry that allows a straightforward implementation of complicated geometric setups with intersecting boundaries, where subsequent modifications to the geometry are easily performed. Presently, GenUC has five elemental volumes that can be defined in any position in space: spheres, ellipsoids, parallelepipeds, and circular cylinders and cones. The mortran macro-based implementation of the combinatorial geometry package allows an easy definition/extension of any other elemental volume, e.g., elliptical cylinders and cones. The scoring of the relevant parameters and the output of the results in GenUC are performed with two CERN data analysis packages, which permit the generation of nonplanar phase space distribution files and can also be used for geometry verification, among many other capabilities. GenUC has been successfully applied to complex geometric setups, e.g., intracavitary brachytherapy applicators and was also benchmarked against the EGSnrc user code, DOSRZnrc.

Algorithms↗
PubMed2003

ROC analysis for assessment of lesion detection performance in 3D PET: influence of reconstruction algorithms.

Image quality in positron emission tomography (PET) can be assessed with physical parameters, as spatial resolution and signal-to-noise ratio, or using psychophysical approaches, which include the observer performance and the considered task (ROC analysis). For PET in oncology, such a task is the detection of hot lesions. The aim of the present study was to assess the lesion detection performance due to adequate modeling of the scanner and the measurement process in the image reconstruction process. We compared the standard OSEM software of the manufacturer with a sophisticated fully 3D iterative reconstruction technique (USC MAP). A rectangular phantom with 6 oblique line sources in a homogeneous background (2.6 kBq/ml 18F) was imaged dynamically with an ECAT EXACT HR+ scanner in 3D mode. Reconstructed activity contrasts varied between 15 and 0, as the line sources were filled with 11C (3.2 MBq/ml). Measured attenuation and standard randoms, dead time, and scatter corrections of the manufacturer were employed. For the ROC analysis, a software tool presented a cut-out of the phantom (15 x 15 pixels) to two observers. These cut-outs were rated (5 classes) and the area Az under the ROC curve was determined as a measure of detection performance. The improvement for Az with USC MAP compared to the OSEM reconstructions ranged between 0.02 and 0.23 for signal-to-noise ratios of the background between 2.8 and 3.1 and lesion contrast between 2.1 and 4.2. This study demonstrates that adequate modeling of the measurement process in the reconstruction algorithm improves the detection of small hot lesions markedly.

Algorithms↗
PubMed2004

Analysis of the penumbra enlargement in lung versus the quality index of photon beams: a methodology to check the dose calculation algorithm.

It is well known that considerable underdosage can occur at the edges of a tumor inside the lung because of the degradation of penumbra due to lack of lateral electronic equilibrium. Although present even at smaller energies, this phenomenon is more pronounced for higher energies. Apart from Monte Carlo calculation, most of the existing Treatment Planning Systems (TPSs) cannot deal at all, or with acceptable accuracy, with this effect. A methodology has been developed for assessing the dose calculation algorithms in the lung region where lateral electronic disequilibrium exists, based on the Quality Index (QI) of the incident beam. A phantom, consisting of layers of polystyrene and lung material, has been irradiated using photon beams of 4, 6, 15, and 20 MV. The cross-plane profiles of each beam for 5x5, 10x10, and 25x10 fields have been measured at the middle of the phantom with the use of films. The penumbra (20%-80%) and fringe (50%-90%) enlargement was measured and the ratio of the widths for the lung to that of polystyrene was defined as the Correction Factor (CF). Monte Carlo calculations in the two phantoms have also been performed for energies of 6, 15, and 20 MV. Five commercial TPS's algorithms were tested for their ability to predict the penumbra and fringe enlargement. A linear relationship has been found between the QI of the beams and the CF of the penumbra and fringe enlargement for all the examined fields. Monte Carlo calculations agree very well (less than 1% difference) with the film measurements. The CF values range between 1.1 for 4 MV (QI 0.620) and 2.28 for 20 MV (QI 0.794). Three of the tested TPS's algorithms could not predict any enlargement at all for all energies and all fields and two of them could predict the penumbra enlargement to some extent. The proposed methodology can help any user or developer to check the accuracy of its algorithm for lung cases, based on a simple phantom geometry and the QI of the incident beam. This check is very important especially when higher energies are used, as the inaccuracies in existing algorithms can lead to an incorrect choice of energy for lung treatment and consequently to a failure in tumor control.

Algorithms↗
PubMed2004

Dosimetric evaluation of the clinical implementation of the first commercial IMRT Monte Carlo treatment planning system at 6 MV.

In this work we dosimetrically evaluated the clinical implementation of a commercial Monte Carlo treatment planning software (PEREGRINE, North American Scientific, Cranberry Township, PA) intended for quality assurance (QA) of intensity modulated radiation therapy treatment plans. Dose profiles calculated in homogeneous and heterogeneous phantoms using this system were compared to both measurements and simulations using the EGSnrc Monte Carlo code for the 6 MV beam of a Varian CL21EX linear accelerator. For simple jaw-defined fields, calculations agree within 2% of the dose at d(max) with measurements in homogeneous phantoms with the exception of the buildup region where the calculations overestimate the dose by up to 8%. In heterogeneous lung and bone phantoms the agreement is within 3%, on average, up to 5% for a 1 x 1 cm2 field. We tested two consecutive implementations of the MLC model. After matching the calculated and measured MLC leakage, simulations of static and dynamic MLC-defined fields using the most recent MLC model agreed to within 2% with measurements.

Algorithms↗
PubMed2004

An algorithm for independent verification of Gamma Knife treatment plans.

A formalism for independent treatment verification has been developed for Gamma Knife radiosurgery in analogy to the second checks being performed routinely in the field of external beam radiotherapy. A verification algorithm is presented, and evaluated based on its agreement with treatment planning calculations for the first 40 Canadian Gamma Knife patients. The algorithm is used to calculate the irradiation time for each shot, and the value of the dose at the maximum dose point in each calculation matrix. Data entry consists of information included on the plan printout, and can be streamlined by using an optional plan import feature. Calculated shot times differed from those generated by the treatment planning software by an average of 0.3%, with a standard deviation of 1.4%. The agreement of dose maxima was comparable with an average of -0.2% and a standard deviation of 1.3%. Consistently accurate comparisons were observed for centrally located lesions treated with a small number of shots. Large discrepancies were almost all associated with dose plans utilizing a large number of collimator plugs, for which the simplifying approximations used by the program are known to break down.

Algorithms↗
PubMed2004

Comparison of dose calculation algorithms in phantoms with lung equivalent heterogeneities under conditions of lateral electronic disequilibrium.

An extensive set of benchmark measurement of PDDs and beam profiles was performed in a heterogeneous layer phantom, including a lung equivalent heterogeneity, by means of several detectors and compared against the predicted dose values by different calculation algorithms in two treatment planning systems. PDDs were measured with TLDs, plane parallel and cylindrical ionization chambers and beam profiles with films. Additionally, Monte Carlo simulations by means of the PENELOPE code were performed. Four different field sizes (10 x 10, 5 x 5, 2 x 2, and 1 x 1 cm2) and two lung equivalent materials (CIRS, p(w)e=0.195 and St. Bartholomew Hospital, London, p(w)e=0.244-0.322) were studied. The performance of four correction-based algorithms and one based on convolution-superposition was analyzed. The correction-based algorithms were the Batho, the Modified Batho, and the Equivalent TAR implemented in the Cadplan (Varian) treatment planning system and the TMS Pencil Beam from the Helax-TMS (Nucletron) treatment planning system. The convolution-superposition algorithm was the Collapsed Cone implemented in the Helax-TMS. The only studied calculation methods that correlated successfully with the measured values with a 2% average inside all media were the Collapsed Cone and the Monte Carlo simulation. The biggest difference between the predicted and the delivered dose in the beam axis was found for the EqTAR algorithm inside the CIRS lung equivalent material in a 2 x 2 cm2 18 MV x-ray beam. In these conditions, average and maximum difference against the TLD measurements were 32% and 39%, respectively. In the water equivalent part of the phantom every algorithm correctly predicted the dose (within 2%) everywhere except very close to the interfaces where differences up to 24% were found for 2 x 2 cm2 18 MV photon beams. Consistent values were found between the reference detector (ionization chamber in water and TLD in lung) and Monte Carlo simulations, yielding minimal differences (0.4%+/-1.2%). The penumbra broadening effect in low density media was not predicted by any of the correction-based algorithms, and the only one that matched the experimental values and the Monte Carlo simulations within the estimated uncertainties was the Collapsed Cone Algorithm.

Algorithms↗
PubMed2004

CSnrc: correlated sampling Monte Carlo calculations using EGSnrc.

CSnrc, a new user-code for the EGSnrc Monte Carlo system is described. This user-code improves the efficiency when calculating ratios of doses from similar geometries. It uses a correlated sampling variance reduction technique. CSnrc is developed from an existing EGSnrc user-code CAVRZnrc and improves upon the correlated sampling algorithm used in an earlier version of the code written for the EGS4 Monte Carlo system. Improvements over the EGS4 version of the algorithm avoid repetition of sections of particle tracks. The new code includes a rectangular phantom geometry not available in other EGSnrc cylindrical codes. Comparison to CAVRZnrc shows gains in efficiency of up to a factor of 64 for a variety of test geometries when computing the ratio of doses to the cavity for two geometries. CSnrc is well suited to in-phantom calculations and is used to calculate the central electrode correction factor Pcel in high-energy photon and electron beams. Current dosimetry protocols base the value of Pcel on earlier Monte Carlo calculations. The current CSnrc calculations achieve 0.02% statistical uncertainties on Pcel, much lower than those previously published. The current values of Pcel compare well with the values used in dosimetry protocols for photon beams. For electrons beams, CSnrc calculations are reported at the reference depth used in recent protocols and show up to a 0.2% correction for a graphite electrode, a correction currently ignored by dosimetry protocols. The calculations show that for a 1 mm diameter aluminum central electrode, the correction factor differs somewhat from the values used in both the IAEA TRS-398 code of practice and the AAPM's TG-51 protocol.

Algorithms↗
PubMed2006

Efficient photon beam dose calculations using DOSXYZnrc with BEAMnrc.

This study examines the efficiencies of doses calculated using DOSXYZnrc for 18 MV (10 X 10 cm2 field size) and 6 MV (10 X 10 cm2 and 20 X 20 cm2 field sizes) photon beams simulated using BEAMnrc. Both phase-space sources and full BEAMnrc simulation sources are used in the DOSXYZnrc calculations. BEAMnrc simulation sources consist of a BEAMnrc accelerator simulation compiled as a shared library and run by the user code (DOSXYZnrc in this case) to generate source particles. Their main advantage is in eliminating the need to store intermediate phase-space files. In addition, the efficiency improvements due to photon splitting and particle recycling in the DOSXYZnrc simulation are examined. It is found that photon splitting increases dose calculation efficiency by a factor of up to 6.5, depending on beam energy, field size, voxel size, and the type of secondary collimation used in the BEAMnrc simulation (multileaf collimator vs photon jaws). The optimum efficiency with photon splitting is approximately 55% higher than that with particle recycling, indicating that, while most of the gain is due to time saved by reusing source particle data, there is significant gain due to the uniform distribution of interaction sites and faster DOSXYZnrc simulation time when photon splitting is employed. Use of optimized directional bremsstrahlung splitting in the BEAMnrc simulation sources increases the efficiency of photon beam simulations sufficiently that the peak efficiencies (i.e., with optimum setting of the photon splitting number) of DOSXYZnrc simulations using these sources are only 3-13% lower than those with phase-space file sources. This points towards eliminating the need for storing intermediate phase-space files.

Body Burden↗
PubMed2006

Monte Carlo dose voxel kernel calculations of beta-emitting and Auger-emitting radionuclides for internal dosimetry: A comparison between EGSnrcMP and EGS4.

Dose-point kernels (DPKs) can be widely applied to therapeutic nuclear medicine to obtain more accurate absorbed dose assessments in internal dosimetry assuming a spherical geometry. Recently, EGSnrc-the latest in the family of EGS Monte Carlo codes--has been tested for isotropic monoenergetic electrons and Y-90 beta spectrum in spherical geometry. The availability of SPECT images allows one to take into account heterogeneities in activity distribution within tumors, and to perform dose calculations using voxel dosimetry based on Monte Carlo simulations in a Cartesian geometry. The purpose of this study is to evaluate the differences of dose distributions scored in Cartesian voxels also known as Dose Voxel Kernels (DVKs) for five beta-emitting (131I, 89Sr, 153Sm, 186Re, and 90Y) and Auger-emitting (111In) radionuclides, when their computation is made using these two Monte Carlo codes from the same family to check if the new physics in EGSnrc simulation system produces DVK very different from those calculated with EGS4. We have calculated the DVKs for point and voxel sources in Cartesian scoring grids of different spatial resolutions. Our results for the point source, scored in the finer spatial resolution, show a poor agreement between EGSnrc and EGS4 (up to about 20%) for voxels closer to the origin, and a better agreement (below 5%) for longer distances for all radionuclides. For the voxel source, where doses were scored in the coarser spatial resolution, dose deposition in the central voxel is in good agreement for all the radionuclides; while surrounding voxels exhibit a slightly worse agreement.

Algorithms↗
PubMedDate not supplied

Automatic on-line inspection of patient setup in radiation therapy using digital portal images.

A new method is presented for inspection of patient setup in radiation therapy by automatic comparison of the patient position relative to the beam position in portal and simulator images. Quantification of patient-setup errors in terms of translation, rotation, and magnification is achieved by chamfer matching, a robust technique to match drawings and images, which is applied to both anatomy outlines and field edges. Applied to field edges, chamfer matching detects and visualizes deviations in field shape. Applied to anatomy outlines, the matching procedure quantifies and visualizes deviations in patient position relative to the radiation field. To test the method and to judge its feasibility, its behavior for four hundred different patient-setup deviations, which were simulated in four clinical images, was examined. These images show a top view of the pelvic region. The performance was measured in terms of accuracy and success rate for numerous cost functions and distance codings associated with the chamfer matching procedure. An average accuracy of 1.8 mm was found, a success rate of 90%, and an average overall computation time of 3 s on a 486 microcomputer. The whole analysis procedure is fast enough to allow on-line application.

Humans↗
PubMedDate not supplied

Verification of radiotherapy treatments: computerized analysis of the size and shape of radiation fields.

An automated technique has been developed for the verification of treatment field size and shape in external beam radiation therapy. Portal images from film or digital on-line imaging system are analyzed, and basic parameters are derived to describe the field size and shape from the contour points on the field boundary. The initial set of parameters included length of the perimeter, area, aspect ratio, and orientation angle. The parameters found for the actual field in the portal image are compared against those calculated for the prescribed field and any discrepancies indicated to the operator. The accuracy of the field parameterization scheme has been tested on a number of on-line portal images with varying fields. The relative error did not exceed a few percent in perimeter and area or 2 degrees in the angle, which should be sufficiently low for the detection of major errors in field shaping.

Computer Simulation↗
PubMedDate not supplied

Uncertainty analysis of field placement error measurements using digital portal and simulation image correlations.

An uncertainty analysis of a field placement error measurement method using correlations between digital portal and simulation images has been performed. This analysis used the Monte Carlo method to simulate random uncertainties arising in image registration and portal field edge detection. The uncertainty analysis used data from a field placement error study to predict the dependences of the measurement method upon image matchpoint number and spatial distribution, image registration, and uncertainties arising in the detection and description of the field edge in a portal image. It was shown that the uncertainty in the field registration had the most significant effect upon the measurement uncertainty, whereas the number of matchpoints had an extremely limited effect (beyond the minimum required for the mathematical procedure used in registration). In contrast, it was shown that an accurate field placement error measurement required that the number of pixels describing a field edge used for comparisons should exceed 20, thus automatically setting a minimum field length necessary for field placement comparisons. It was also shown that the spatial uniformity of the matchpoints in an image pair had a limited effect upon the field placement error measurement, provided that the points are distributed over about 15% of the image area.

Humans↗