Comparison of lung CT number and airway dimension evaluation capabilities of ultra-high-resolution CT, using different scan modes and reconstruction methods including deep learning reconstruction, with those of multi-detector CT in a QIBA phantom study.

OBJECTIVE: Ultra-high-resolution CT (UHR-CT), which can be applied normal resolution (NR), high-resolution (HR), and super-high-resolution (SHR) modes, has become available as in conjunction with multi-detector CT (MDCT). Moreover, deep learning reconstruction (DLR) method, as well as filtered back projection (FBP), hybrid-type iterative reconstruction (IR), and model-based IR methods, has been clinically used. The purpose of this study was to directly compare lung CT number and airway dimension evaluation capabilities of UHR-CT using different scan modes with those of MDCT with different reconstruction methods as investigated in a lung density and airway phantom design recommended by QIBA. MATERIALS AND METHODS: Lung CT number, inner diameter (ID), inner area (IA), and wall thickness (WT) were measured, and mean differences between measured CT number, ID, IA, WT, and standard reference were compared by means of Tukey's HSD test between all UHR-CT data and MDCT reconstructed with FBP as 1.0-mm section thickness. RESULTS: For each reconstruction method, mean differences in lung CT numbers and all airway parameters on 0.5-mm and 1-mm section thickness CTs obtained with SHR and HR modes showed significant differences with those obtained with the NR mode on UHR-CT and MDCT (p < 0.05). Moreover, the mean differences on all UHR-CTs obtained with SHR, HR, or NR modes were significantly different from those of 1.0-mm section thickness MDCTs reconstructed with FBP (p < 0.05). CONCLUSION: Scan modes and reconstruction methods used for UHR-CT were found to significantly affect lung CT number and airway dimension evaluations as did reconstruction methods used for MDCT. KEY POINTS: • Scan and reconstruction methods used for UHR-CT showed significantly higher CT numbers and smaller airway dimension evaluations as did those for MDCT in a QIBA phantom study (p < 0.05). • Mean differences in lung CT number for 0.25-mm, 0.5-mm, and 1.0-mm section thickness CT images obtained with SHR and HR modes were significantly larger than those for CT images at 1.0-mm section thickness obtained with MDCT and reconstructed with FBP (p < 0.05). • Mean differences in inner diameter (ID), inner area (IA), and wall thickness (WT) measured with SHR and HR modes on 0.5- and 1.0-mm section thickness CT images were significantly smaller than those obtained with NR mode on UHR-CT and MDCT (p < 0.05).

Image Quality and Lesion Detection on Deep Learning Reconstruction and Iterative Reconstruction of Submillisievert Chest and Abdominal CT.

OBJECTIVE. The objective of this study was to compare image quality and clinically significant lesion detection on deep learning reconstruction (DLR) and iterative reconstruction (IR) images of submillisievert chest and abdominopelvic CT. MATERIALS AND METHODS. Our prospective multiinstitutional study included 59 adult patients (33 women, 26 men; mean age ± SD, 65 ± 12 years old; mean body mass index [weight in kilograms divided by the square of height in meters] = 27 ± 5) who underwent routine chest (n = 22; 16 women, six men) and abdominopelvic (n = 37; 17 women, 20 men) CT on a 640-MDCT scanner (Aquilion ONE, Canon Medical Systems). All patients gave written informed consent for the acquisition of low-dose (LD) CT (LDCT) after a clinically indicated standard-dose (SD) CT (SDCT). The SDCT series (120 kVp, 164-644 mA) were reconstructed with interactive reconstruction (IR) (adaptive iterative dose reduction [AIDR] 3D, Canon Medical Systems), and the LDCT (100 kVp, 120 kVp; 30-50 mA) were reconstructed with filtered back-projection (FBP), IR (AIDR 3D and forward-projected model-based iterative reconstruction solution [FIRST], Canon Medical Systems), and deep learning reconstruction (DLR) (Advanced Intelligent Clear-IQ Engine [AiCE], Canon Medical Systems). Four subspecialty-trained radiologists first read all LD image sets and then compared them side-by-side with SD AIDR 3D images in an independent, randomized, and blinded fashion. Subspecialty radiologists assessed image quality of LDCT images on a 3-point scale (1 = unacceptable, 2 = suboptimal, 3 = optimal). Descriptive statistics were obtained, and the Wilcoxon sign rank test was performed. RESULTS. Mean volume CT dose index and dose-length product for LDCT (2.1 ± 0.8 mGy, 49 ± 13mGy·cm) were lower than those for SDCT (13 ± 4.4 mGy, 567 ± 249 mGy·cm) (p < 0.0001). All 31 clinically significant abdominal lesions were seen on SD AIDR 3D and LD DLR images. Twenty-five, 18, and seven lesions were detected on LD AIDR 3D, LD FIRST, and LD FBP images, respectively. All 39 pulmonary nodules detected on SD AIDR 3D images were also noted on LD DLR images. LD DLR images were deemed acceptable for interpretation in 97% (35/37) of abdominal and 95-100% (21-22/22) of chest LDCT studies (p = 0.2-0.99). The LD FIRST, LD AIDR 3D, and LD FBP images had inferior image quality compared with SD AIDR 3D images (p < 0.0001). CONCLUSION. At submillisievert chest and abdominopelvic CT doses, DLR enables image quality and lesion detection superior to commercial IR and FBP images.

Correction to: Deep learning reconstruction improves image quality of abdominal ultra-high-resolution CT.

The original version of this article, published on 11 April 2019, unfortunately, contained a mistake. The following correction has therefore been made in the original: The image in Fig. 3c was wrong. The corrected figure is given below. The original article has been corrected.

Deep learning reconstruction improves image quality of abdominal ultra-high-resolution CT.

OBJECTIVES: Deep learning reconstruction (DLR) is a new reconstruction method; it introduces deep convolutional neural networks into the reconstruction flow. This study was conducted in order to examine the clinical applicability of abdominal ultra-high-resolution CT (U-HRCT) exams reconstructed with a new DLR in comparison to hybrid and model-based iterative reconstruction (hybrid-IR, MBIR). METHODS: Our retrospective study included 46 patients seen between December 2017 and April 2018. A radiologist recorded the standard deviation of attenuation in the paraspinal muscle as the image noise and calculated the contrast-to-noise ratio (CNR) for the aorta, portal vein, and liver. The overall image quality was assessed by two other radiologists and graded on a 5-point confidence scale ranging from 1 (unacceptable) to 5 (excellent). The difference between CT images subjected to hybrid-IR, MBIR, and DLR was compared. RESULTS: The image noise was significantly lower and the CNR was significantly higher on DLR than hybrid-IR and MBIR images (p < 0.01). DLR images received the highest and MBIR images the lowest scores for overall image quality. CONCLUSIONS: DLR improved the quality of abdominal U-HRCT images. KEY POINTS: • The potential degradation due to increased noise may prevent implementation of ultra-high-resolution CT in the abdomen. • Image noise and overall image quality for hepatic ultra-high-resolution CT images improved with deep learning reconstruction as compared to hybrid- and model-based iterative reconstruction.

Energy-integrating detector based ultra-high-resolution CT with deep learning reconstruction for the assessment of calcified lesions in coronary artery disease.

BACKGROUND: The aim of this study to compare of the image quality of calcified lesions in coronary artery disease between deep learning reconstruction (DLR) and model-based iterative reconstruction (MBIR) on energy-integrating detector (EID) based ultra-high-resolution CT (UHRCT). METHODS: We performed a phantom study on EID-based UHRCT using a dedicated insert for calcifications and obtained the derivative values for DLR and MBIR. In the clinical study, the derivative values were compared between DLR and MBIR across 73 calcified lesions in 62 patients. Edge sharpness of calcifications and contrast resolution at the coronary lumen side were quantified by the maximum and minimum derivative values. Two radiologists independently analyzed image quality of the calcified lesions using a 5-point Likert scale. RESULTS: In the phantom study, the edge sharpness of the 3-mm calcifications on DLR (median, 924 HU/mm; IQR, 580-1741 HU/mm) was significantly higher than on MBIR (median, 835 HU/mm; IQR, 484-1552; p â€‹< â€‹0.001). In the clinical study, the image quality of the calcified lesions was significantly better on DLR with significantly reduced reconstruction time (p â€‹< â€‹0.001). The contrast resolution at the coronary lumen side on DLR (median, -99.1 HU/mm; IQR, -209 to -34.3 HU/mm) was significantly higher than on MBIR (median, -41.8 HU/mm; IQR, -121 to 22.3 HU/mm, p â€‹< â€‹0.001) although the edge sharpness of calcifications was similar between DLR and MBIR (p â€‹= â€‹0.794) in the clinical setting. CONCLUSION: EID-based UHRCT reconstructed using DLR represents better image quality of calcified lesions in coronary artery disease compared with MBIR, with significantly reduced reconstruction time.

Ultra-high-resolution CT vs. invasive angiography for detecting hemodynamically significant coronary artery disease: Rationale and methods of the CORE-PRECISION multicenter study.

BACKGROUND: Direct coronary arterial evaluation via computed tomography (CT) angiography is the most accurate noninvasive test for the diagnosis of coronary artery disease (CAD). However, diagnostic accuracy is limited in the setting of severe coronary calcification or stents. Ultra-high-resolution CT (UHR-CT) may overcome this limitation, but no rigorous study has tested this hypothesis. METHODS: The CORE-PRECISION is an international, multicenter, prospective diagnostic accuracy study testing the non-inferiority of UHR-CT compared to invasive coronary angiography (ICA) for identifying patients with hemodynamically significant CAD. The study will enroll 150 patients with history of CAD, defined as prior documentation of lumen obstruction, stenting, or a calcium score ≥400, who will undergo UHR-CT before clinically prompted ICA. Assessment of hemodynamically significant CAD by UHR-CT and ICA will follow clinical standards. The reference standard will be the quantitative flow ratio (QFR) with <0.8 defined as abnormal. All data will be analyzed in independent core laboratories. RESULTS: The primary outcome will be the comparative diagnostic accuracy of UHR-CT vs. ICA for detecting hemodynamically significant CAD on a patient level. Secondary analyses will focus on vessel level diagnostic accuracy, quantitative stenosis analysis, automated contour detection, in-depth plaque analysis, and others. CONCLUSION: CORE-PRECISION aims to investigate if UHR-CT is non-inferior to ICA for detecting hemodynamically significant CAD in high-risk patients, including those with severe coronary calcification or stents. We anticipate this study to provide valuable insights into the utility of UHR-CT in this challenging population and for its potential to establish a new standard for CAD assessment.

Capability for dose reduction while maintaining nodule detection: Comparison of silver and copper X-ray spectrum modulation filters for chest CT using a phantom study with different reconstruction methods.

PURPOSE: To compare the capability of CTs obtained with a silver or copper x-ray beam spectral modulation filter (Ag filter and Cu filter) and reconstructed with FBP, hybrid-type IR and deep learning reconstruction (DLR) for radiation dose reduction for lung nodule detection using a chest phantom study. MATERIALS AND METHODS: A chest CT phantom was scanned with a 320-detector row CT with Ag filter at 0.6, 1.6 and 2.5 mGy and Cu filters at 0.6, 1.6, 2.5 and 9.6 mGy, and reconstructed with the aforementioned methods. To compare image quality of all the CT data, SNRs and CNRs for any nodule were calculated for all protocols. To compare nodule detection capability among all protocols, the probability of detection of any nodule was assessed with a 5-point visual scoring system. Then, ROC analyses were performed to compare nodule detection capability of Ag and Cu filters for each radiation dose data with the same method and of the three methods for any radiation dose data and obtained with either filter. RESULTS: At any of the doses, SNR, CNR and area under the curve for the Ag filter were significantly higher or larger than those for the Cu filter (p < 0.05). Moreover, with DLR, those values were significantly higher or larger than all the others for CTs obtained with any of the radiation doses and either filter (p < 0.05). CONCLUSION: The Ag filter and DLR can significantly improve image quality and nodule detection capability compared with the Cu filter and other reconstruction methods at each of radiation doses used.

Effectiveness of deep learning reconstruction on standard to ultra-low-dose high-definition chest CT images.

PURPOSE: Deep learning reconstruction (DLR) has been introduced by major vendors, tested for CT examinations of a variety of organs, and compared with other reconstruction methods. The purpose of this study was to compare the capabilities of DLR for image quality improvement and lung texture evaluation with those of hybrid-type iterative reconstruction (IR) for standard-, reduced- and ultra-low-dose CTs (SDCT, RDCT and ULDCT) obtained with high-definition CT (HDCT) and reconstructed at 0.25-mm, 0.5-mm and 1-mm section thicknesses with 512 × 512 or 1024 × 1024 matrixes for patients with various pulmonary diseases. MATERIALS AND METHODS: Forty age-, gender- and body mass index-matched patients with various pulmonary diseases underwent SDCT (CT dose index volume : mean ± standard deviation, 9.0 ± 1.8 mGy), RDCT (CTDIvol: 1.7 ± 0.2 mGy) and ULDCT (CTDIvol: 0.8 ± 0.1 mGy) at a HDCT. All CT data set were then reconstructed with 512 × 512 or 1024 × 1024 matrixes by means of hybrid-type IR and DLR. SNR of lung parenchyma and probabilities of all lung textures were assessed for each CT data set. SNR and detection performance of each lung texture reconstructed with DLR and hybrid-type IR were then compared by means of paired t tests and ROC analyses for all CT data at each section thickness. RESULTS: Data for each radiation dose showed DLR attained significantly higher SNR than hybrid-type IR for each of the CT data (p < 0.0001). On assessments of all findings except consolidation and nodules or masses, areas under the curve (AUCs) for ULDCT with hybrid-type IR for each section thickness (0.91 ≤ AUC ≤ 0.97) were significantly smaller than those with DLR (0.97 ≤ AUC ≤ 1, p < 0.05) and the standard protocol (0.98 ≤ AUC ≤ 1, p < 0.05). CONCLUSION: DLR is potentially more effective for image quality improvement and lung texture evaluation than hybrid-type IR on all radiation dose CTs obtained at HDCT and reconstructed with each section thickness with both matrixes for patients with a variety of pulmonary diseases.

Validation of synthesized normal-resolution image data generated from high-resolution acquisitions on a commercial CT scanner.

PURPOSE: To validate a normal-resolution (NR) simulation (NRsim) algorithm that uses high-resolution (HR) or super-high resolution (SHR) acquisitions on a commercial HR computed tomography (CT) scanner by comparing image quality between NRsim-generated images and actual NR images. NRsim is intended to allow direct comparison between normal-resolution CT and HR/SHR reconstructions in clinical investigations, without repeating exams. METHODS: The Aquilion Precision CT (Canon Medical Systems Corporation) HR CT scanner has three resolution modes resulting from detector binning in the channel (x-y) and row (z) directions. For NR, each detector element is 0.5 mm × 0.5 mm along the channel and row directions, 0.25 mm × 0.5 mm for HR, and 0.25 mm × 0.25 mm for SHR. The NRsim algorithm simulates NR acquisitions from HR or SHR acquisitions (termed NRHR and NRSHR , respectively) by downsampling the pre-log raw data in the channel direction for the HR acquisitions and in the channel and row direction for the SHR acquisition. The downsampled data are then reconstructed using the same process as NR. The axial modulation transfer function (MTF), slice sensitivity profile (SSP), and CT number accuracy were measured using the Catphan 600 phantom, and the three-dimensional noise power spectrum (NPS) was measured in water-equivalent phantoms for standard protocols across a range of size-specific dose estimates (SSDE): head (6.2-29.8 mGy), lung (2.2-18.2 mGy), and body (5.6-19.4 mGy). The MTF and NPS measurements were combined to estimate low-contrast detectability (LCD) using a non-prewhitening model observer with an eye filter for a 5-mm disk with 10 HU contrast. All metrics were compared for NR, NRHR , and NRSHR images reconstructed using filtered back projection (FBP) and an iterative reconstruction algorithm (AIDR3D). We chose a 15% error threshold as a reasonable definition of success for NRsim when compared against actual NR based on published studies showing that a just-noticeable difference in image noise level for human observers is typically <15%. RESULTS: The axial MTF and SSPs for NRsim were in good agreement with NR demonstrated by a maximum difference of 5.1% for the MTF at 10% and 50% across materials (air, Teflon, LDPE, and polystyrene) and a maximum SSP difference of 2.2%. Noise magnitude differences were within 15% across the SSDE levels with the exception of below 4.5 mGy for the lung protocol with FBP. The relative RMSE of normalized NPS comparisons were all <15%. Differences in CT numbers for NRsim reconstructions were within 2 HU of NR. LCD for NRsim was within 15% of NR with the exception of NRSHR for the lung protocol SSDE levels below 3.7 mGy with FBP. CONCLUSIONS: NRsim, an algorithm for simulating NR acquisitions using HR and SHR raw data, was introduced and shown to generate images with spatial resolution, noise, HU accuracy, and LCD largely equivalent to scans acquired using an actual NR acquisition. At SSDE levels below ~5 mGy for the lung protocol, differences in noise magnitude and LCD for NRSHR were >15% which defines a region where NRsim degrades due to contributions from electronic noise.

Deep Learning-based CT Image Reconstruction: Initial Evaluation Targeting Hypovascular Hepatic Metastases.

PURPOSE: To evaluate the effect of a deep learning-based reconstruction (DLR) method on the conspicuity of hypovascular hepatic metastases on abdominal CT images. MATERIALS AND METHODS: This retrospective study with institutional review board approval included 58 patients with hypovascular hepatic metastases. A radiologist recorded the standard deviation of attenuation in the paraspinal muscle as the image noise and the contrast-to-noise ratio (CNR). CNR was calculated as region of interest ([ROI]L - ROIT)/N, where ROIL is the mean liver parenchyma attenuation, ROIT, the mean tumor attenuation, and N, the noise. Two other radiologists graded the conspicuity of the liver lesion on a five-point scale where 1 is unidentifiable and 5 is detected without diagnostic compromise. Only the smallest liver lesion in each patient, classified as smaller or larger than 10 mm, was evaluated. The difference between hybrid iterative reconstruction (IR) and DLR images was determined by using a two-sided Wilcoxon signed-rank test. RESULTS: The image noise was significantly lower, and the CNR was significantly higher on DLR images than hybrid IR images (median image noise: 19.2 vs 12.8 HU, P < .001; median CNR: tumors < 10 mm: 1.9 vs 2.5; tumors > 10 mm: 1.7 vs 2.2, both P < .001). The scores for liver lesions were significantly higher for DLR images than hybrid IR images (P < .01 for both in tumors smaller or larger than 10 mm). CONCLUSION: DLR improved the quality of abdominal CT images for the evaluation of hypovascular hepatic metastases.© RSNA, 2019Supplemental material is available for this article.

[Evaluation of time resolution in cardiac synchronized image reconstruction using multi-slice CT].

One of the newest CT application technologies is cardiac synchronized image reconstruction. In this technology, evaluation of time-resolution is very important. We developed a method of measuring time-resolution in cardiac synchronized reconstruction, and evaluated various scanning protocols. In our experiment, ECG-gated scanning was done by multi-slice CT (Aquilion16 Super Heart Edition, Toshiba Medical Systems Co., Ltd., Japan). The nominal slice thickness was 0.5 mm, and rotation time was 0.5 sec. Input heart rate was set at 40, 45, 50, 55, 60, 70, 75, 80, and 90 bpm, and helical pitch at 3.2, 4.0, and 4.8 (beam-pitch: 0.200, 0.250 and 0.300). We measured FWTM of the obtained sensitivity distribution and compared at each scanning protocol. Time resolution improved as helical pitch decreased and heart rate increased. However, phase-time resolution deteriorated as heart rate increased. The results of our experiment indicated that a segment center was determined by X-ray tube rotation time and heart rate, and the number of segments was determined by heart rate, helical pitch, and reconstruction position. Time resolution changed with X-ray tube rotation time, heart rate, helical pitch, and reconstruction position. In this report, we provide a reference for an optimal scanning protocol in cardiac synchronized image reconstruction.