Faculty Bibliography

OBJECTIVE:The aim of this study was to evaluate the impact of a noise-optimized virtual monoenergetic imaging (VMI+) reconstruction technique on image quality and lesion delineation in patients with renal cell carcinoma (RCC) undergoing abdominal dual-energy computed tomography (DECT). MATERIALS AND METHODS/METHODS:Fifty-two patients (33 men; 61.5±13.6years) with RCC underwent contrast-enhanced DECT during the corticomedullary and nephrogenic phase of renal enhancement. DECT datasets were reconstructed with standard linearly-blended (M_0.6), as well as traditional virtual monoenergetic (VMI) and VMI+ algorithms in 10-keV increments from 40 to 100 keV. Contrast-to-noise (CNR) and tumor-to-cortex ratios for corticomedullary- and nephrogenic-phase images were objectively measured by a radiologist with 3 years of experience. Subjective image quality and RCC delineation were evaluated by three independent radiologists. RESULTS:Greatest CNR values were found for 40-keV VMI+ series in both corticomedullary- (8.9±4.9) and nephrogenic-phase (7.1±4.6) images and were significantly higher compared to all other reconstructions (P<0.001). Furthermore, tumor-to-cortex ratios were highest for 40-keV nephrogenic-phase VMI+ (2.1±3.5; P≤0.016), followed by 50-keV and 60-keV VMI+ (2.0±3.2 and 1.8±2.8, respectively). Qualitative image quality scored highest for 50-keV VMI+ series in corticomedullary-phase reconstructions and 60-keV in nephrogenic-phase reconstructions (P≤0.031). Highest scores for lesion delineation were assigned for 40-keV VMI+ reconstructions (P≤0.074). CONCLUSION/CONCLUSIONS:Low-keV VMI+ reconstructions lead to improved image quality and lesion delineation of corticomedullary- and nephrogenic-phase DECT datasets in patients with RCC.

OBJECTIVES:The aim of this study was to investigate the impact of virtual monoenergetic imaging (VMI+) and dual-energy computed tomography perfusion maps (DECT-PMs) on reader confidence and diagnostic accuracy in dual-energy computed tomography pulmonary angiography (DE-CTPA) studies with suboptimal contrast attenuation, compared with standard linearly blended reconstruction series. MATERIALS AND METHODS:Dual-energy computed tomography pulmonary angiography examinations with suboptimal contrast attenuation of 68 patients with suspected pulmonary embolism (PE) were included in this institutional review board-approved retrospective study. Virtual monoenergetic imaging series at 40 keV, DECT-PM, and linearly blended images (M_0.6, 60% 90-kV spectrum) were reconstructed. Contrast-to-noise ratio and signal-to-noise ratio within the pulmonary trunk were calculated. Four independent radiologists assessed the presence of PE and their diagnostic confidence using 3 DE-CTPA reconstruction protocols: protocol 1, M_0.6 images; protocol 2, M_0.6 series and DECT-PM; and protocol 3, M_0.6, DECT-PM, and VMI+ series. Receiver operating characteristic (ROC) analysis was performed. RESULTS:Fourteen patients showed central and 29 segmental PE. Greater contrast-to-noise ratio and signal-to-noise ratio values were measured in VMI+ series at 40 keV in comparison to M_0.6 images (P < 0.001). Diagnostic accuracy for segmental PE detection was as follows: protocol 1 (69.1%); protocol 2 (86.8%); and protocol 3 (92.6%). Protocol 3 resulted in a significantly greater area under the curve for diagnosing segmental PE (0.991, P ≤ 0.033), compared with protocol 1 and 2 (0.897 and 0.951, respectively), and provided the highest diagnostic confidence (P < 0.001). CONCLUSIONS:A reconstruction protocol including 40-keV VMI+ series and DECT-PM improves reader confidence and diagnostic accuracy for segmental PE detection compared with standard M_0.6 images in DE-CTPA with suboptimal contrast attenuation.

OBJECTIVE:The aim of this study was to investigate the impact of noise-optimized virtual monoenergetic imaging (VMI+) reconstructions on quantitative and qualitative image parameters in patients with cutaneous malignant melanoma at thoracoabdominal dual-energy computed tomography (DECT). MATERIALS AND METHODS/METHODS:Seventy-six patients (48 men; 66.6±13.8years) with metastatic cutaneous malignant melanoma underwent DECT of the thorax and abdomen. Images were post-processed with standard linear blending (M_0.6), traditional virtual monoenergetic (VMI), and VMI+ technique. VMI and VMI+ images were reconstructed in 10-keV intervals from 40 to 100keV. Attenuation measurements were performed in cutaneous melanoma lesions, as well as in regional lymph node, subcutaneous and in-transit metastases to calculate objective signal-to-noise (SNR) and contrast-to-noise (CNR) ratios. Five-point scales were used to evaluate overall image quality and lesion delineation by three radiologists with different levels of experience. RESULTS:Objective indices SNR and CNR were highest at 40-keV VMI+ series (5.6±2.6 and 12.4±3.4), significantly superior to all other reconstructions (all P<0.001). Qualitative image parameters showed highest values for 50-keV and 60-keV VMI+ reconstructions (median 5, respectively; P≤0.019) regarding overall image quality. Moreover, qualitative assessment of lesion delineation peaked in 40-keV VMI+ (median 5) and 50-keV VMI+ (median 4; P=0.055), significantly superior to all other reconstructions (all P<0.001). CONCLUSION/CONCLUSIONS:Low-keV noise-optimized VMI+ reconstructions substantially increase quantitative and qualitative image parameters, as well as subjective lesion delineation compared to standard image reconstruction and traditional VMI in patients with cutaneous malignant melanoma at thoracoabdominal DECT.

The American College of Radiology (ACR) Breast Imaging-Reporting and Data System (BI-RADS) lexicon, which is used ubiquitously to standardize reporting of breast magnetic resonance imaging (MRI), provides 7 BI-RADS assessment categories to indicate the level of suspicion of malignancy and guide further management. A BI-RADS category 4 assessment is assigned when an imaging abnormality does not fulfill the typical criteria for malignancy, but is suspicious enough to warrant a recommendation for biopsy. The BI-RADS category 4 assessment covers a wide range of probability of malignancy, from >2 to <95%. MRI is an essential noninvasive technique in breast imaging and the role of MRI in the assessment of ACR BI-RADS 4 lesions is manifold. In lesions classified as suspicious on imaging with mammography, digital breast tomosynthesis, and sonography, MRI can aid in the noninvasive differentiation of benign and malignant lesions and obviate unnecessary breast biopsies. When the suspicion of cancer is confirmed with MRI, concurrent staging of disease for treatment planning can be accomplished. This article will provide a comprehensive overview of the role of breast MRI in the assessment of ACR BI-RADS 4 lesions. In addition, we will discuss strategies to decrease false positives and avoid false negative results when reporting MRI of the breast.

PURPOSE/OBJECTIVE:To evaluate diagnostic accuracy of a noise-optimized virtual monoenergetic imaging (VMI+) reconstruction technique for detection of active arterial abdominal bleeding on dual-energy (DE) CT angiography compared with standard image reconstruction. MATERIALS AND METHODS/METHODS:DE CT angiography data sets of 71 patients (46 men; age 63.6 y ± 13.3) with suspected arterial bleeding of the abdomen or pelvis were reconstructed with standard linearly blended (F_0.5), VMI+, and traditional virtual monoenergetic imaging (VMI) algorithms in 10-keV increments from 40 to 100 keV. Attenuation measurements were performed in the descending aorta, area of hemorrhage, and feeding artery to calculate contrast-to-noise ratios (CNRs) in patients with active arterial bleeding. Based on quantitative image quality results, the best series for each reconstruction technique were chosen to analyze the diagnostic performance of 3 blinded radiologists. RESULTS:DE CT angiography showed acute arterial bleeding in 36 patients. Mean CNR was superior in 40-keV VMI+ compared with VMI series (all P < .001), which showed highest CNRs in 70-keV VMI and F_0.5 (21.6 ± 7.9, 12.9 ± 4.7, and 10.4 ± 3.6) images. Area under the curve analysis for detection of arterial bleeding showed significantly superior (P < .001) results for 40-keV VMI+ (0.963) compared with 70-keV VMI (0.775) and F_0.5 (0.817) series. CONCLUSIONS:Diagnostic accuracy in patients with active arterial bleeding of the abdomen can be significantly improved using VMI+ reconstructions at 40 keV compared with standard linearly blended and traditional VMI series in DE CT angiography.

PURPOSE/OBJECTIVE:To assess image quality and diagnostic performance of a noise-optimized algorithm to reconstruct virtual monoenergetic images (VMI+) for the detection of endoleaks after endovascular abdominal aortic aneurysm repair (EVAR) using dual-energy CT angiography (DE-CTA). MATERIALS AND METHODS/METHODS:Seventy-five patients (42 men; 66.2±11.7years) underwent DE-CTA following EVAR. Arterial phase images were acquired in dual-energy mode for the reconstruction of standard linearly-blended M_0.5, VMI+ and traditional monoenergetic images (VMI) at 40-100keV in 10-keV intervals. Contrast-to-noise ratios (CNR) were calculated for the area of leakage in patients with endoleaks. Diagnostic accuracy for endoleak detection was evaluated by three blinded radiologists using the objectively best series for each reconstruction technique. RESULTS:Thirty-four out of 75 patients showed endoleaks. Quantitative image parameters were highest at 40-keV VMI+ (CNR, 21.3±11.1), compared to M_0.5 (CNR, 10.9±5.5) and all VMI series that showed highest values at 70keV (CNR, 13.5±6.6; all P<0.001). ROC analysis for endoleak detection revealed an area under the curve (AUC) of 0.992 for 40-keV VMI+ series, which was significantly higher (P≤0.039) compared to 70-keV VMI (0.914) and M_0.5 series (0.916). CONCLUSIONS:Noise-optimized VMI+ series at 40keV improve diagnostic accuracy for the detection and rule-out of endoleaks after EVAR.

OBJECTIVE:The purpose of our study was to determine diagnostic accuracy, image quality, and radiation dose of low-dose single- and dual-energy unenhanced third-generation dual-source head CT for detection of intracranial hemorrhage (ICH). MATERIALS AND METHODS/METHODS:A total of 123 patients with suspected ICH were examined using a dual-source 192-MDCT scanner. Standard-dose 120-kVp single-energy CT (SECT; n = 36) and 80-kVp and 150-kVp dual-energy CT (DECT; n = 30) images were compared with low-dose SECT (n = 32) and DECT (n = 25) images obtained using automated tube current modulation (ATCM). Advanced modeled iterative reconstruction (ADMIRE) was used for all protocols. Detection of ICH was performed by three readers who were blinded to the image acquisition parameters of each image series. Image quality was assessed both quantitatively and qualitatively. Interobserver agreement was calculated using the Fleiss kappa. Radiation dose was measured as dose-length product (DLP). RESULTS:Detection of ICH was excellent (sensitivity, 94.9-100%; specificity, 94.7-100%) in all protocols (p = 1.00) with perfect interobserver agreement (0.83-0.96). Qualitative ratings showed significantly better ratings for both standard-dose protocols regarding gray matter-to-white matter contrast (p ≤ 0.014), whereas highest gray matter-to-white matter contrast-to-noise ratio was observed with low-dose DECT images (p ≥ 0.057). The lowest posterior fossa artifact index was measured for standard-dose DECT, which showed significantly lower values compared with low-dose protocols (p ≤ 0.034). Delineation of ventricular margins and sharpness of subarachnoidal spaces were rated excellent in all protocols (p ≥ 0.096). Low-dose techniques lowered radiation dose by 26% for SECT images (DLP, 575.0 ± 72.3 mGy · cm vs 771.5 ± 146.8 mGy · cm; p < 0.001) and by 24% in DECT images (DLP, 587.0 ± 103.2 mGy · cm vs 770.6 ± 90.2 mGy · cm; p < 0.001). No significant difference was observed between the low-dose protocols (p = 1.00). CONCLUSION/CONCLUSIONS:Low-dose unenhanced head SECT and DECT using ATCM and ADMIRE provide excellent diagnostic accuracy for detection of ICH with good quantitative and qualitative image quality in third-generation dual-source CT while allowing significant radiation dose reduction.

Monoenergetic imaging is an increasingly used reconstruction technique in postprocessing of dual-energy computed tomography (DECT). The main advantage of this technique is the ability to substantially increase image contrast of structures with uptake of iodinated contrast material. Although monoenergetic imaging was mainly used in oncological DECT applications, recent research has further demonstrated its role in vascular imaging. Using this dedicated postprocessing algorithm, image contrast of vascular structures in the thorax can be increased, a drastic reduction of contrast material is feasible, and even beam-hardening artifacts can be reduced. The aim of this review article is to explain the technical background of this technique, showcase its relevance in cardiothoracic DECT, and provide an outlook on the clinical impact of this technique beyond solely improvements in image quality.

PURPOSE:The aim of this study was to evaluate the impact of a noise-optimized virtual monoenergetic imaging (VMI+) reconstruction technique on quantitative and qualitative image analysis in patients with gastrointestinal stromal tumors (GISTs) at dual-energy computed tomography (DECT) of the abdomen. METHODS:Forty-five DECT datasets of 21 patients (14 men; 63.7 ± 9.2 years) with GISTs were reconstructed with the standard linearly blended (M_0.6) and VMI+ and traditional virtual monoenergetic (VMI) algorithm in 10-keV increments from 40 to 100 keV. Attenuation measurements were performed in GIST lesions and abdominal metastases to calculate objective signal-to-noise (SNR) and contrast-to-noise ratios (CNR). Five-point scales were used to evaluate overall image quality, lesion delineation, image sharpness, and image noise. RESULTS:Quantitative image parameters peaked at 40-keV VMI+ series (SNR 27.8 ± 13.0; CNR 26.3 ± 12.7), significantly superior to linearly blended (SNR 16.8 ± 7.3; CNR 13.6 ± 6.9) and all VMI series (all P < 0.001). Qualitative image parameters were highest for 60-keV VMI+ reconstructions regarding overall image quality and image sharpness (median 5, respectively; P ≤ 0.023). Qualitative assessment of lesion delineation peaked in 40 and 50-keV VMI+ series (median 5, respectively). Image noise was superior in 90 and 100-keV VMI and VMI+ reconstructions (all medians 5). CONCLUSIONS:Low-keV VMI+ reconstructions significantly increase SNR and CNR of GISTs and improve quantitative and qualitative image quality of abdominal DECT datasets compared to traditional VMI and standard linearly blended image series.

BACKGROUND:Current techniques for evaluation of bone mineral density (BMD) commonly require phantom calibration. The purpose of this study was to evaluate a novel algorithm for phantomless in vivo dual-energy computed tomography (DECT)-based assessment of BMD of the lumbar spine in comparison with dual-energy X-ray absorptiometry (DEXA). METHODS:Data from clinically indicated DECT and DEXA examinations within two months comprising the lumbar spine of 47 patients were retrospectively evaluated. By using a novel automated dedicated post-processing algorithm for DECT, the trabecular bone of lumbar vertebrae L1-L4 was selected and analysed. Linear correlation was analysed using Pearson's product-moment correlation coefficient for the comparison of the results from DECT and DEXA. RESULTS:. Regression analysis demonstrated a lack of correlation between DECT- and DEXA-based BMD values with a Pearson's product-moment correlation coefficient r = 0.4205. CONCLUSIONS:Dedicated post-processing of DECT data using a novel algorithm for retrospective phantomless BMD assessment of the trabecular bone of lumbar vertebrae from clinically indicated DECT examinations is feasible.