Jacob Geleijns

Development and validation of segmentation and interpolation techniques in sinograms for metal artifact suppression in CT.

PURPOSE Metal prostheses cause artifacts in computed tomography (CT) images. The purpose of this work was to design an efficient and accurate metal segmentation in raw data to achieve artifact suppression and to improve CT image quality for patients with metal hip or shoulder prostheses. METHODS The artifact suppression technique incorporates two steps: metal object segmentation in raw data and replacement of the segmented region by new values using an interpolation scheme, followed by addition of the scaled metal signal intensity. Segmentation of metal is performed directly in sinograms, making it efficient and different from current methods that perform segmentation in reconstructed images in combination with Radon transformations. Metal signal segmentation is achieved by using a Markov random field model (MRF). Three interpolation methods are applied and investigated. To provide a proof of concept, CT data of five patients with metal implants were included in the study, as well as CT data of a PMMA phantom with Teflon, PVC, and titanium inserts. Accuracy was determined quantitatively by comparing mean Hounsfield (HU) values and standard deviation (SD) as a measure of distortion in phantom images with titanium (original and suppressed) and without titanium insert. Qualitative improvement was assessed by comparing uncorrected clinical images with artifact suppressed images. RESULTS Artifacts in CT data of a phantom and five patients were automatically suppressed. The general visibility of structures clearly improved. In phantom images, the technique showed reduced SD close to the SD for the case where titanium was not inserted, indicating improved image quality. HU values in corrected images were different from expected values for all interpolation methods. Subtle differences between interpolation methods were found. CONCLUSIONS The new artifact suppression design is efficient, for instance, in terms of preserving spatial resolution, as it is applied directly to original raw data. It successfully reduced artifacts in CT images of five patients and in phantom images. Sophisticated interpolation methods are needed to obtain reliable HU values close to the prosthesis.

Computed tomography dose assessment for a 160 mm wide, 320 detector row, cone beam CT scanner

Computed tomography (CT) dosimetry should be adapted to the rapid developments in CT technology. Recently a 160 mm wide, 320 detector row, cone beam CT scanner that challenges the existing Computed Tomography Dose Index (CTDI) dosimetry paradigm was introduced. The purpose of this study was to assess dosimetric characteristics of this cone beam scanner, to study the appropriateness of existing CT dose metrics and to suggest a pragmatic approach for CT dosimetry for cone beam scanners. Dose measurements with a small Farmer-type ionization chamber and with 100 mm and 300 mm long pencil ionization chambers were performed free in air to characterize the cone beam. According to the most common dose metric in CT, namely CTDI, measurements were also performed in 150 mm and 350 mm long CT head and CT body dose phantoms with 100 mm and 300 mm long pencil ionization chambers, respectively. To explore effects that cannot be measured with ionization chambers, Monte Carlo (MC) simulations of the dose distribution in 150 mm, 350 mm and 700 mm long CT head and CT body phantoms were performed. To overcome inconsistencies in the definition of CTDI100 for the 160 mm wide cone beam CT scanner, doses were also expressed as the average absorbed dose within the pencil chamber (D100). Measurements free in air revealed excellent correspondence between CTDI300air and D100air, while CTDI100air substantially underestimates CTDI300air. Results of measurements in CT dose phantoms and corresponding MC simulations at centre and peripheral positions were weighted and revealed good agreement between CTDI300w, D100w and CTDI600w, while CTDI100w substantially underestimates CTDI300w. D100w provides a pragmatic metric for characterizing the dose of the 160 mm wide cone beam CT scanner. This quantity can be measured with the widely available 100 mm pencil ionization chamber within 150 mm long CT dose phantoms. CTDI300w measured in 350 mm long CT dose phantoms serves as an appropriate standard of reference for characterizing the dose of this CT scanner. A CT dose descriptor that is based on an integration length smaller than the actual beam width is preferably expressed as an (average) dose, such as D100 for the 160 mm wide cone beam CT scanner, and not as CTDI100.

Diagnosing pulmonary embolism in pregnancy: rationalizing fetal radiation exposure in radiological procedures

In pregnancy, the diagnosis of pulmonary embolism (PE) is problematic. There is doubt as to whether objective diagnostic tests are needed and confusion as to what objective test is the safest with respect to fetal radiation exposure. A recent study has reported a very low (1.8%) prevalence of high-probability ventilation-perfusion (VQ) lung scans in pregnant women suspected of PE [1]. From this study it is apparent that the clinical diagnosis of PE is inaccurate and therefore objective diagnostic tests are mandatory, in order to avoid treatment of women that do not have PE. Currently, helical computerized tomography (CT) and VQ scintigraphy are themost common diagnostic tests used in nonpregnant patients with suspected PE. Physicians are reluctant to perform helical CT in pregnant women because of potential adverse effects of radiation exposure to the fetus. VQ scintigraphy has been assumed to be associated with less radiation exposure than helical CT. To compare the relative amounts of radiation exposure to the fetus, we calculated fetal radiation exposure when singleand multidetector row helical CT and VQ scintigraphy were performed using our local hospital protocols. Further, we compared our data with data of the literature. Since there are no established methods for calculating fetal radiation exposure in diagnostic radiological procedures, we used a pragmatic approach. The amount of radiation absorbed by the fetus was assumed to be equal to that absorbed by the uterus of a non-pregnant woman. Assessment of the uterus dose was achieved by measurement of the computed tomography dose index and the application of organ dose conversion factors [2]. The following CT protocols were used for fetal dose assessment: 120 kV, 250 mAs, slice thickness 3 mm and pitch factor 1.7 for single-detector row helical CT (Philips AVE, Best, the Netherlands) 2 and 120 kV, 85 mAs, slice thickness 16 · 0.5 mm and pitch factor 1.4 for multidetector row helical CT (Toshiba Aquilion 16, Shimoishigami, Otawara-shi, Tochigi, Japan) 3 . The scanned range extends from the dorsal lung sinus to the top of the lung. Since fetal radiation exposure was calculated, physical measures (e.g. abdominal shielding with lead) to reduce radiation exposure were not taken into account. For the perfusion scintigraphy protocol we used 40 MBq of Technetium-labeled albumin aggregates. In our institution, ventilation scintigraphy is performed with Krypton-81 m, which is inhaled for two minutes per image. The RubidiumKrypton generator generates 450–750 MBq per min. Our calculated data of CT radiation exposure were compared with doses in nuclear medicine and doses calculated by the International Commission on Radiological Protection (ICRP) and the National Radiological Protection Board (NRPB) of the UK [3,4]. The calculated dose of radiation absorbed by the fetus for a single-detector row helical CT was 0.026 mSv. An even lower dose (0.013 mSv) was calculated for the multidetector row helical CT. In comparison, the calculated dose of fetal radiation with perfusion scintigraphy was 0.11–0.20 mSv. In comparison with doses given by the ICRP and NRPB (Table 1), our calculated doses of helical CT were low. Our study suggests that performing a helical CT according to our local protocol, whether singleor multidetector row, exposes the fetus to less radiation than perfusion scintigraphy. Our findings are clearly contradictive to the general idea that helical CT is more hazardous to the fetus than perfusion scintigraphy. Regarding the generalizability of our data, it is apparent that our calculated fetal radiation dose for CT is well within the range of that found by others. It has been documented that radiation exposure to the patient for a given radiological procedure can vary considerably between different institutions and even within the same institution. There are several factors that affect radiation dose from CT, e.g. (beam energy, tubecurrent time product, pitch, collimation, patient size and dose reduction options). Each institution should therefore carefully scrutinize its protocol for performing helical CT and define the optimal balance between minimal patient radiation exposure and maximal diagnostic CT image quality. Correspondence: Dr M.V. Huisman, Department of General Internal Medicine, Room C1 R 43, Leiden University Medical Center, PO Box 960

Artifacts in ECG-Synchronized MDCT Coronary Angiography

OBJECTIVE In MDCT coronary angiography, image artifacts are the major cause of false-positive and false-negative interpretations regarding the presence of coronary artery stenoses. Hence, it is important that observers reporting these investigations are aware of the potential presence of image artifacts and that these artifacts are recognized. CONCLUSION The article explores the technical causes for various artifacts in MDCT coronary angiography imaging and clinical examples are given.

Reduction of Radiation Exposure in the Cardiac Electrophysiology Laboratory

WITTKAMPF, F.H.M., et al. Reduction of Radiation Exposure in the Cardiac Electrophysiology Laboratory. The purpose of this study was to determine the effects of various protective measures on patient and operator radiation dose levels in catheter ablation procedures. Catheter ablation procedures are associated with significant radiation levels. The patients skin and operator radiation levels were measured (1) at baseline, (2) after primary beam filtration by 0.3‐mm copper sheet and 2‐mm aluminium plate and implementation of the LocaLisa system, and (3) after reduction of the left anterior oblique fluoroscopic pulse rate and installation of a lead glass screen. Additionally, a comparative analysis of radiation exposure levels was performed in the seven Dutch catheter ablation centers. Filtration of both primary beams resulted in a more than two‐fold reduction in patient skin dose. Together with the LocaLisa system, this resulted in a six‐fold reduction in patient and operator dose. As expected, lowering of the left anterior oblique pulse rate from 25 to 12.5 Hz reduced the corresponding patient skin dose with a factor 2 while the leadglass protection caused an extra factor 2 reduction for the operator. Large differences were observed between fluoroscopy systems used for catheter ablation in the Netherlands. Depending on patient body mass and fluoroscopy system, patient skin dose varied between 0.2 and 8.4 Gy/hour. Proper measures may allow for a significant reduction of patient and operator radiation exposure in catheter ablation procedures. The large influence of body mass and equipment on patients skin dose requires a more direct monitoring of skin dose than total fluoroscopy time.

Dose and perceived image quality in chest radiography.

Chest radiography is the most commonly performed diagnostic X-ray examination. The radiation dose to the patient for this examination is relatively low but because of its frequent use, the contribution to the collective dose is considerable. Consequently, optimization of dose and image quality offers a challenging area of research. In this article studies on dose reduction, different detector technologies, optimization of image acquisition and new technical developments in image acquisition and post processing will be reviewed. Studies indicate that dose reduction in PA chest images to at least 50% of commonly applied dose levels does not affect diagnosis in the lung fields; however, dose reduction in the mediastinum, upper abdomen and retrocardiac areas appears to directly deteriorate diagnosis. In addition to patient dose, also the design of the various digital detectors seems to have an effect on image quality. With respect to image acquisition, studies showed that using a lower tube voltage improves visibility of anatomical structures and lesions in digital chest radiographs but also increases the disturbing appearance of ribs. New techniques that are currently being evaluated are dual energy, tomosynthesis, temporal subtraction and rib suppression. These technologies may improve diagnostic chest X-ray further. They may for example reduce the negative influence of over projection of ribs, referred to as anatomic noise. In chest X-ray this type of noise may be the dominating factor in the detection of nodules. In conclusion, optimization and new developments will enlarge the value of chest X-ray as a mainstay in the diagnosis of chest diseases.

Meta-Analysis of 40- and 64-MDCT Angiography for Assessing Coronary Artery Stenosis

OBJECTIVE The purpose of our study was to assess the diagnostic performance of thin-slice (< or = 0.625 mm) MDCT coronary angiography compared with invasive coronary angiography for the detection of significant (> or = 50%) stenosis. MATERIALS AND METHODS Twenty-two articles on 40- and 64-MDCT coronary angiography were included. Sensitivity and specificity were calculated on a per-patient and per-segment basis; in addition, proximal versus distal segments were evaluated. The effect of nonevaluable patients, nonevaluable segments, and disease prevalence on diagnostic performance was assessed. RESULTS Pooled sensitivity on a patient level was 97.7% ([95% CI] 96.2-98.7%) and specificity 91.0% (88.5-93.1%). Pooled sensitivity on a segmental level was 90.8% (89.0-92.4%) and specificity 95.7% (95.2-96.1%); for proximal segments, respectively, 94.2% (92.3-95.7%) and 94.1% (93.4-94.8%), and for distal segments 84.8% (81.1-88.0%) and 96.9% (96.4-97.4%). If nonevaluable MDCT investigations were included, the per-patient specificity was reduced from 91.0% to 89.1% (p > 0.05) when allocating excluded patients as having significant coronary artery stenosis, and the sensitivity was reduced from 97.7% to 96.2% (p > 0.05) when allocating excluded patients as not having significant stenosis. The per-patient prevalence of coronary artery stenosis had no significant influence on the sensitivity for detecting significant stenosis. CONCLUSION Forty- and 64-MDCT provide good-to-excellent performance in detecting or ruling out significant coronary artery stenosis, with better results for proximal than for distal coronary artery segments.

Assessment of Agatston Coronary Artery Calcium Score Using Contrast-Enhanced CT Coronary Angiography

OBJECTIVE The purpose of this article is to evaluate to what extent Agatston scores may be derived from CT coronary angiography (CTA) examinations, compared with traditional unenhanced CT calcium scores. MATERIALS AND METHODS Fifty patients with a CT calcium score-Agatston score of zero and 50 patients with a CT calcium score-Agatston score of 1 or greater whose CT calcium scores had been calculated and who had undergone CTA using volumetric 320-MDCT were included. Agatston scores were obtained at 3.0-mm slices for CT calcium score and CTA. Method agreement, interobserver agreement, and diagnostic performance of CTA for detecting coronary calcium were evaluated. RESULTS Of 50 patients with a positive CT calcium score-Agatston score, coronary artery calcium was detected with CTA in 43 patients by observer 1 (mean CTA score, 102 ± 202; mean CT calcium score, 254 ± 501) and in 46 patients by observer 2 (mean CTA score, 94 ± 147; mean CT calcium score, 272 ± 531). Of the 50 patients with a CT calcium score-Agatston score of zero, 49 (98%, observer 1) and 50 (100%, observer 2) had a zero score with CTA as well. An intraclass correlation of 0.78 and 0.62 was found between CT calcium score and CTA (p < 0.01), whereas higher Agatston scores were underestimated with CTA. For observer 1, the sensitivity, specificity, positive predictive value, negative predictive value, and diagnostic accuracy for detection of coronary calcium with CTA were 86%, 98%, 98%, 88%, and 92%, respectively, and the corresponding values for observer 2 were 92%, 100%, 100%, 93%, and 96%, respectively. Interobserver agreement was 0.996 for CT calcium score and 0.93 for CTA. CONCLUSION Coronary artery calcium can be detected on CTA images with high accuracy. The Agatston calcium score derived from CTA images shows good correlation with unenhanced CT calcium score and is highly reproducible. However, higher Agatston scores are systematically underestimated when derived from CTA images.

Metal artifact reduction for CT: development, implementation, and clinical comparison of a generic and a scanner-specific technique.

PURPOSE To develop, implement, and compare two metal artifact reduction methods for CT. METHODS Two methods for metal artifact reduction were developed. The first is based on applying corrections in a Radon transformation of the CT images. The second method is based on a forward projection of the CT images and applying corrections in the scanners original raw data. The first method is generic since it does not depend on the scanner specifications. For the second method, detailed information on the design of the CT scanner and the raw data of the study is required. Clinical implementation and evaluation were performed using pre- and post-operative CT scans of four patients with shoulder prosthesis. For comparison of these methods, the authors developed a quantitative technique that compares improvement in image quality for the two metal artifact reduction techniques with the image quality of the uncorrected images. RESULTS Metal artifact reduction using either of the two methods yields a decrease of noise and artifacts in CT scans of patients with shoulder prostheses. Artifacts that appeared as bright and dark streaks were reduced or eliminated and as a result image quality improved. Quantitative assessment of clinical images showed improved image quality for both techniques of metal artifact reduction, but the method based on correction in original raw data performed better in all comparisons. CONCLUSION Both methods are effective for metal artifact reduction, but better performance was observed for the method that is based on correcting the original raw data. The used evaluation technique provides an objective way of evaluating the metal artifacts in clinical CT images.

Radiation exposure to patients in a multicenter coronary angiography trial (CORE 64).

OBJECTIVE The objective of this study was to assess the exposure of patients to radiation for the cardiac CT acquisition protocol of the multicenter Coronary Artery Evaluation Using 64-Row Multidetector Computed Tomography Angiography (CORE 64) trial. MATERIALS AND METHODS An algorithm for patient dose assessment with Monte Carlo dosimetry was developed for the Aquilion 64-MDCT scanner. During the CORE 64 study, different acquisition protocols were used depending on patient size and sex; therefore, six patient models were constructed representing three men and three women in the categories of small, normal size, and obese. Organ dose and effective dose resulting from the cardiac CT protocol were assessed for these six patient models. RESULTS The average effective dose for coronary CT angiography (CTA) calculated according to Report 103 of the International Commission on Radiological Protection (ICRP) is 19 mSv (range, 16-26 mSv). The average effective dose for the whole cardiac CT protocol including CT scanograms, bolus tracking, and calcium scoring is slightly higher-22 mSv (range, 18-30 mSv). An average conversion factor for the calculation of effective dose from dose-length product of 0.030 mSv/mGy · cm was derived for coronary CTA. CONCLUSION The current methods of assessing patient dose are not well suited for cardiac CT acquisitions, and published effective dose values tend to underestimate effective dose. The effective dose of cardiac CT is approximately 25% higher when assessed according to the preferred ICRP Report 103 compared with ICRP Report 60. Underestimation of effective dose by 43% or 53% occurs in coronary CTA according to ICRP Report 103 when a conversion factor (E / DLP, where E is effective dose and DLP is dose-length product) for general chest CT of 0.017 or 0.014 mSv/mGy · cm, respectively, is used instead of 0.030 mSv/mGy · cm.