Henri A. Vrooman

Fast and accurate automated measurements in digitized stereophotogrammetric radiographs.

Until recently, Roentgen Stereophotogrammetric Analysis (RSA) required the manual definition of all markers using a high-resolution measurement table. To automate this tedious and time-consuming process and to eliminate observer variabilities, an analytical software package has been developed and validated for the detection, identification, and matching of markers in RSA radiographs. The digital analysis procedure consisted of the following steps: (1) the detection of markers using a variant of the Hough circle-finder technique; (2) the identification and labeling of the detected markers; (3) the reconstruction of the three-dimensional position of the bone markers and the prosthetic markers; and (4) the computation of micromotion. To assess the influence of film digitization, the measurements obtained from nine phantom radiographs using two different film scanners were compared with the results obtained by manual processing. All markers in the phantom radiographs were automatically detected and correctly labeled. The best results were obtained with a Vidar VXR-12 CCD scanner, for which the measurement errors were comparable to the errors associated with the manual approach. To assess the in vivo reproducibility, 30 patient radiographs were analyzed twice with the manual as well as with the automated procedure. Approximately, 85% of all calibration markers and bone markers were automatically detected and correctly matched. The calibration errors and the rigid-body errors show that the accuracy of the automated procedure is comparable to the accuracy of the manual procedure. The rigid-body errors had comparable mean values for both techniques: 0.05 mm for the tibia and 0.06 mm for the prosthesis. The reproducibility of the automated procedure showed to be slightly better than that of the manual procedure. The maximum errors in the computed translation and rotation of the tibial component were 0.11 mm and 0.24, compared to 0.13 mm and 0.27 for the manual RSA procedure. The total processing time is less than 10 min per radiograph, including interactive corrections, compared to approximately 1 h for the manual approach. In conclusion, a new and widely applicable, computer-assisted technique has become available to detect, identify, and match markers in RSA radiographs and to assess the micromotion of endoprostheses. This new technique will be used in our clinic for our hip, knee, and elbow studies.

Model-based Roentgen stereophotogrammetry of orthopaedic implants

Attaching tantalum markers to prostheses for Roentgen stereophotogrammetry (RSA) may be difficult and is sometimes even impossible. In this study, a model-based RSA method that avoids the attachment of markers to prostheses is presented and validated. This model-based RSA method uses a triangulated surface model of the implant. A projected contour of this model is calculated and this calculated model contour is matched onto the detected contour of the actual implant in the RSA radiograph. The difference between the two contours is minimized by variation of the position and orientation of the model. When a minimal difference between the contours is found, an optimal position and orientation of the model has been obtained. The method was validated by means of a phantom experiment. Three prosthesis components were used in this experiment: the femoral and tibial component of an Interax total knee prosthesis (Stryker Howmedica Osteonics Corp., Rutherfort, USA) and the femoral component of a Profix total knee prosthesis (Smith & Nephew, Memphis, USA). For the prosthesis components used in this study, the accuracy of the model-based method is lower than the accuracy of traditional RSA. For the Interax femoral and tibial components, significant dimensional tolerances were found that were probably caused by the casting process and manual polishing of the components surfaces. The largest standard deviation for any translation was 0.19mm and for any rotation it was 0.52 degrees. For the Profix femoral component that had no large dimensional tolerances, the largest standard deviation for any translation was 0.22mm and for any rotation it was 0.22 degrees. From this study we may conclude that the accuracy of the current model-based RSA method is sensitive to dimensional tolerances of the implant. Research is now being conducted to make model-based RSA less sensitive to dimensional tolerances and thereby improving its accuracy.

Digital automated RSA compared to manually operated RSA

The accuracy of digital Roentgen stereophotogrammetric analysis (RSA) was compared to the accuracy of a manually operated RSA system. For this purpose, we used radiographs of a phantom and radiographs of patients. The radiographs of the patients consisted of double examinations of 12 patients that had a tibial osteotomy and of double examinations of 12 patients that received a total hip prosthesis. First, the radiographs were measured manually with an accurate measurement table. Subsequently, the images were digitized by a film scanner at 150 DPI and 300 DPI resolutions and analyzed with the RSA-CMS software. In the phantom experiment, the manually operated system produced significantly better results than the digital system, although the maximum difference between the median values of the manually operated system and the digital system was as low as 0.013mm for translations and 0.033 degrees for rotations. In the radiographs of the patients, the manually operated system and the digital system produced equally accurate results: no significant differences in translations and rotations were found. We conclude that digital RSA is an accurate, fast, and user friendly alternative for manually operated RSA. Currently, digital RSA systems are being used in a growing number of clinical RSA-studies.

Sources of error in lung densitometry with CT.

RATIONALE AND OBJECTIVES To determine and analyze the most important error sources in lung CT densitometry in vivo. METHODS The authors examined the influences of CT acquisition errors, physiologic changes, and image segmentation errors on lung densitometry. Among others, spatial dependency and long-term reproducibility of the density measurements of blood and air were examined over a period of 4 years in a group of 28 patients with pulmonary emphysema. These results were related to the measured lung densities in this group. RESULTS The density measurement of blood and air is strongly dependent on the position in the thorax. Despite full-scanner calibrations, x-ray tube replacement can induce a significant increase in measured blood density. CONCLUSIONS A change in a lung density parameter over time can actually be the result of tube replacement or changing blood density. A simple postprocessing technique can correct for these changes.

Detection of areas with viable remnant tumor in postchemotherapy patients with Ewing's sarcoma by dynamic contrast-enhanced MRI using pharmacokinetic modeling.

An approach is presented for monitoring the effects of neoadjuvant chemotherapy in patients with Ewings sarcoma using dynamic contrast-enhanced perfusion magnetic resonance (MR) images. For that purpose, we modify the three-compartment pharmacokinetic permeability model introduced by Tofts et al. (Magn Reson Med 1991;17:357-67) to a two-compartment model. Perfusion MR images acquired using an intravenous injection with Gadolinium (Gd-DTPA) are analyzed with this two-compartment pharmacokinetic model as well as the with an extended pharmacokinetic model that includes the (local) arrival time t(0) of the tracer as an endogenous (estimated) parameter. For each MR section, a wash-in parameter associated with each voxel is estimated twice by fitting each of the two pharmacokinetic models to the dynamic MR signal. A comparison of the two wash-in parametric images (global versus local arrival time) with matched histologic macroslices demonstrates a good correspondence between areas with viable remnant tumor and a high wash-in rate. This can be explained by the high number and permeability of the (leaking) capillaries in viable tumor tissue. The novel pharmacokinetic model based on a local arrival time of tracer results in the best fit of the wash-in rate, the most important factor discerning viable from nonviable tumor components. However, parameter estimates obtained with this model are also more sensitive to noise in the MR signal. The novel pharmacokinetic model resulted in a sensitivity between 0.22 and 0.60 and a specificity between 0.61 and 1. The model based on a global arrival time gave sensitivities between 0.33 and 0.77 and specificities between 0.58 and 0.99. Both statistics are computed as the fraction of correctly labeled voxels (viable or nonviable tumor) within a specified ROI, which delineates the tumor. We conclude that the added value of estimating the local arrival time of tracer first manifests itself for moderate noise levels in the MR signal. The novel pharmacokinetic model should moreover be preferred when pharmacokinetic modeling is applied on the average signal intensity within a ROI, where noise has less effect on the fitted parameters.

ADVANCED MULTIPLE BEAM EQUALIZATION RADIOGRAPHY (AMBER) COMBINED WITH COMPUTED RADIOGRAPHY Preliminary evaluation

The combined use of AMBER (Advanced Multiple Beam Equalization Radiography) and a digital storage phosphor (SP) radiography system was evaluated for chest radiography in a pilot study with 4 patients. Four image modes with different dose levels were compared: the SP in combination with an AMBER equalized exposure (SP/AMBER) and 3 nonequalized exposures with dose levels corresponding to the respective calculated AMBER lung dose (SP/lung field dose), the calculated AMBER mediastinal dose (SP/mediastinal dose) and the calculated AMBER average dose (SP/average dose). All image modes were matched for Hurter and Driffield characteristics and subjectively rated according to visibility of details. The improved signal-to-noise (S/N) ratio of SP/AMBER resulted in a better visualization of structures in the mediastinum and the basal lung where SP/lung field dose scored lowest. For the central lung no quality differences were seen between techniques. The compressed dynamic range of the SP/AMBER images was more easily displayed on the hard-copy film. The combination of AMBER with SP radiography promises to overcome the dynamic range limitations of digital displays while, at moderate doses, giving better S/N and image quality than standard SP technique.

Segmentation of Dynamic Contrast-Enhanced MR-Images of Post Chemotherapy Ewing’s Sarcoma with a Pharmacokinetic Model and a Neural Network

Most patients with Ewing’s sarcoma undergo neoadjuvant (preoperative) chemotherapy before surgery is performed. Generally, chemotherapy reduces the size of the tumor which makes the subsequent treatment more successful. MR-imaging aims at monitoring the effect of chemotherapy by identifying areas of vital remnant tumor. An MR-examination includes static T1- and T2-weighted MR-images as well as dynamic, contrast-enhanced T1-weighted MR-images. Whereas the static MR-images are used to estimate the volume of intra- and extra-osseous bone tumor, the dynamic contrast-enhanced MR-sequence indicates which parts of the tumor are highly perfused by blood. In general, malignant bone tumors are highly perfused. Moreover, these lesions are heterogenuous (sometimes multifocal) containing viable as well as nonviable (necrotic) parts. The only way to reliably distinguish viable from nonviable tumor tissue is by performing a perfusion study by dynamic contrast-enhanced MRI [1].

Computer-simulated multiple beam equalization: image processing algorithm and research tool for Amber

Advanced multiple beam equalization radiography (Amber) has been successfully applied to chest radiography. More recently, the applications have been extended to mammography. The Amber chest unit (Oldelft, Delft, The Netherlands) controls the local X-ray exposure to the patient by means of a feedback loop consisting of a number of detectors in front of the film cassette and the same number of absorbers in front of the X-ray tube. The detector readouts and a predefined compression curve determine the position of the absorbers, while the patient is being scanned by means of a horizontally oriented fan beam. As a consequence, the multiple beam equalization technology has introduced new concepts such as beam profile, compression curve, number of absorbers, and detector weighting function to projection imaging. In order to optimize these different parameters we have developed a computer program, which simulates the multiple beam equalization techniques. Conventionally exposed films are laser scanned resulting in a matrix of optical density values. The program calculates for each pixel the X-ray transmission. These X-ray transmission values are the basis for the simulations with varying beam profile characteristics (i.e. the intensity distribution of the X-ray beam of a channel in horizontal and vertical direction), compression curves, number of channels, detector weighting functions and H&D film curves In order to accurately simulate a particular exposure, the program can be calibrated using optical density and X-ray dose measurements on a conventional X-ray unit or on the Amber unit.