REVIEWS AND COMMENTARY • REVIEW

Next-Generation Hardware Advances in CT:

Cardiac Applications
Alan C. Kwan, MD  •  Amir Pourmorteza, PhD  •  Dan Stutman, PhD  •  David A. Bluemke, MD, PhD  • 
João A. C. Lima, MD, MBA
From the Smidt Heart Institute, Cedars-Sinai Medical Center, 127 S San Vicente Blvd, AHSP, Suite A3600, Los Angeles, CA 90048-0750 (A.C.K.); Department of
Radiology and Imaging Sciences, Emory University, Atlanta, Ga (A.P.); Winship Cancer Institute, Emory University, Atlanta, Ga (A.P.); Department of Biomedical Engi-
neering, Georgia Institute of Technology-Emory University, Atlanta, Ga (A.P.); Department of Physics and Astronomy, Johns Hopkins University, Baltimore, Md (D.S.);
Extreme Light Infrastructure-Nuclear Physics, Bucharest-Magurele, Romania (D.S.); Department of Radiology, University of Wisconsin School of Medicine and Public
Health, Madison, Wis (D.A.B.); and Department of Cardiology, The Johns Hopkins Hospital, Baltimore, Md (J.A.C.L.). Received December 23, 2019; revision requested
February 7, 2020; revision received August 3; accepted August 18. Address correspondence to A.C.K. (e-mail: alan.kwan@cshs.org).
A.C.K. supported by the National Institutes of Health (grant T32HL116273) and Doris Duke Charitable Foundation (grant 2020059). J.A.C.L. supported by Canon
Medical Systems, Astra Zeneca, and MedImmune.

Conflicts of interest are listed at the end of this article.

Radiology 2021; 298:3–17 • https://doi.org/10.1148/radiol.2020192791 • Content codes:

Impending major hardware advances in cardiac CT include three areas: ultra-high-resolution (UHR) CT, photon-counting CT, and
phase-contrast CT. Cardiac CT is a particularly demanding CT application that requires a high degree of temporal resolution, spa-
tial resolution, and soft-tissue contrast in a moving structure. In this review, cardiac CT is used to highlight the strengths of these
technical advances. UHR CT improves visualization of calcified and stented vessels but may result in increased noise and radiation
exposure. Photon-counting CT uses multiple photon energies to reduce artifacts, improve contrast resolution, and perform material
decomposition. Finally, phase-contrast CT uses x-ray refraction properties to improve spatial and soft-tissue contrast. This review
describes these hardware advances in CT and their relevance to cardiovascular imaging.
© RSNA, 2020

Online SA-CME • See www.rsna.org/learning-center-ry

Learning Objectives:
After reading the article and taking the test, the reader will be able to:
n Discuss current challenges to cardiovascular CT imaging related to spatial resolution, contrast resolution, and temporal resolution
n Describe the potential benefits of photon-counting detector-based spectral CT
n Recognize fundamental differences between acquisition of x-ray–based images with attenuation and phase-contrast methods
Accreditation and Designation Statement
The RSNA is accredited by the Accreditation Council for Continuing Medical Education (ACCME) to provide continuing medical education for physicians. The RSNA designates this journal-based SA-CME activity for a maximum of 1.0 AMA
PRA Category 1 Credit™. Physicians should claim only the credit commensurate with the extent of their participation in the activity.
Disclosure Statement
The ACCME requires that the RSNA, as an accredited provider of CME, obtain signed disclosure statements from the authors, editors, and reviewers for this activity. For this journal-based CME activity, author disclosures are listed at the end
of this article.

T here have been significant advances in CT hardware since

its initial development in the early 1970s. Current detec-
tors allow isotropic spatial resolution of approximately 0.25–
To overcome these challenges, we present a narrative
review on three hardware advances within CT that are near-
ing the beginning of clinical application and that we be-
0.5 mm. Gantry rotation speed is now 250 msec with dual- lieve will affect the future of cardiovascular imaging. These
source imaging, enabling high-pitch helical CT scanning of include next-generation ultra-high-resolution (UHR) CT,
the entire torso in 1–2 seconds (1,2). However, challenges spectral imaging with photon-counting (PC) CT, and
remain, particularly for cardiac CT; these challenges include phase-contrast CT (Fig 1). We acknowledge the contribu-
spatial resolution of microstructures, tissue characterization, tion of software advances, including artificial intelligence
and soft-tissue contrast (3). In regard to spatial resolution of (8), deep learning (9–11), fractional flow reserve CT (12),
microstructures, the normal epicardial coronary luminal di- and multiple other advances, but we did not include them
ameter ranges from 1.0 to 4.6 mm (4). Evaluation of cardio- within the scope of the review. Reviews of software ad-
vascular risk now includes coronary plaque quantification vances can be found in the cited sources (13–15). In this
and identification of smaller anatomic features that are at review, we describe hardware advances in CT and focus on
or below the current spatial resolution limits (5,6). In regard their application to cardiac CT as a unique and important
to tissue characterization, quantitative tissue characterization driver of change in the CT field.
is a major diagnostic strength of cardiac MRI that CT has
not yet accomplished (7). Regarding soft-tissue imaging in UHR CT
the heart, iodinated contrast material is required because At present, nearly all conventional CT scanners are mul-
contrast resolution is not yet high enough; however, while tidetector CT units that use arrays of energy-integrating
exogenous contrast yields a high contrast-to-noise ratio of detectors (EIDs) to achieve spatial resolution of approxi-
10–20:1, these agents cannot be used in all patients. mately 0.25–0.5 mm. The terms high-resolution CT and
This copy is for personal use only. To order printed copies, contact reprints@rsna.org
Next-Generation Hardware Advances in CT

Abbreviations
EID = energy-integrating detector, PC = photon counting, PCD = PC
detector, UHR = ultra-high-resolution

Summary
Next-generation CT hardware advances applied to cardiac CT
include ultra-high-resolution methods, photon-counting CT, and
phase-contrast CT, which improve spatial resolution, allow material
decomposition, and improve soft-tissue contrast.

Essentials
n The spatial resolution of next-generation CT has increased to
150–200 µm and allows improved depiction of heavily calcified
vessels and stents.
n Photon-counting CT is a form of spectral CT that uses detectors
to enable high spatial resolution, multicontrast imaging, and im-
proved material decomposition.
n Phase-contrast CT uses refraction instead of attenuation-based
x-ray imaging to provide improved resolution and three times
greater soft-tissue contrast than attenuation-based CT.

thin-section CT usually refer to specific methods for paren-

chymal lung imaging using a slice thickness of 1–2 mm. In
this review, UHR CT is defined as CT with the potential to
improve spatial resolution from approximately 400–450 µm
to nearly 150–200 mm, corresponding to approximately 49.5
line pairs per centimeter (16,17). While one way to achieve
ultrahigh spatial resolution is through reduction in detector
size, spatial resolution is also determined by other factors, in-
cluding x-ray tube focal spot size, magnification by position of
the x-ray source, number of projections recorded per gantry
rotation, and reconstructed kernel shape, algorithm, and voxel
size. Thus, direct comparisons of detector size alone may not
accurately reflect differences in spatial resolution.
Cardiac CT presents a unique set of problems to spatial reso-
lution. Improvements in coronary stent technology have made
intervention of small-vessel (,2.75 mm) stenosis feasible (18)
and an important clinical consideration (19). In addition, coro- Figure 1:  Clinical challenges to cardiac CT, including temporal, spatial, and
nary plaque size and distribution are frequently measured using contrast resolution, and quantitative imaging. Current technologic approaches and
coronary CT angiography because the quantity of plaque affects upcoming approaches with ultra-high-resolution CT, photon-counting CT, and
patient outcome (20,21). Plaque subtypes (lipid core, calcified, phase-contrast CT with potential clinical cardiac applications. FDA = Food and
Drug Administration, FOV = field of view, N/A = not applicable.
and noncalcified components) and certain plaque features (ul-
cerated plaque, napkin sign, areas of low-attenuation plaque, and
spotty calcification) are associated with incident cardiovascular
events (22). However, clear visualization of these vessels and the [Siemens Healthineers, Erlangen, Germany] has a similar focal
small plaque features within coronary walls (ranging from tens to spot size of 0.4 3 0.5 mm; other comparable models range from
hundreds of micrometers) are at or below the spatial resolution 0.6–1.1 3 0.7–1.6 mm) (24). Model-based iterative reconstruc-
limits of cardiac CT (5,20,21,23). tion techniques were also incorporated (25). Research is taking
To address the issue of spatial resolution, the only currently place on cardiovascular applications (Table 1), with US-based
available clinically applicable option for UHR cardiac imaging is systems having been recently installed at Johns Hopkins Hospi-
the Aquilion Precision CT platform (Canon Medical Systems, tal and the University of California at Davis. For this CT plat-
Otawara, Tochigi, Japan). In this form of UHR CT, the mini- form, a spatial resolution of 150 µm at 0% modulation transfer
mum detector element size was reduced to 0.25 3 0.25 mm function (MTF) has been reported (16), which may represent
versus the previous Canon Medical Systems detector element a substantial improvement over that reported with other plat-
size of 0.5 3 0.5 mm (other comparable models range from 0.4 forms. MTF is a continuous function expressing the ability of an
to 0.625 mm). To maximize spatial resolution for this platform, imaging system to resolve objects at different spatial frequencies.
x-ray tubes were designed with adjustable focal spot sizes down While the maximum resolvable spatial frequency is 0% MTF,
to 0.4 3 0.5 mm from the previous 0.9 3 0.8-mm focal spot this does not necessarily imply superior image quality across all
size in Canon Medical Systems models (notably, Somatom Force spatial frequencies. Full MTF profiles may be difficult to find,

2 radiology.rsna.org  n  Radiology: Volume 00: Number 0— 2020

 Kwan et al

Table 1: Recent Major Cardiovascular Ultra-High-Resolution CT Studies

Gantry
Slice Image Rotation
Author and No. of Kilovolt Pixel Size Thickness Matrix DLP Time
CT Platform Date Participants Goal Conclusion Peak (kVp) (mm) (mm) Size (mGy·cm) (sec)
Hino T et al 2020 28 Compare standard Visualization is … … … … … …
CT angiography to improved with
UHR CT angiography UHR CT
for identification of angiography
artery of Adamkiewicz and model-based
reconstruction
  UHR CT … … … … 120 0.3125/ 0.25 1024 3 2117 0.5
0.3906 1024
 Multiple … … … … Variable NR 0.5/0.67 512 3 1655 Variable
512
Murayama K 2019 13 Compare standard CT Number of … … … … … …
et al angiography to UHR identifiable
CT angiography for small arteries
number of interpretable is higher in
lenticulostriate arteries UHR CT
  UHR CT … … … … 120 0.19– 0.25 1024 3 741 1
0.23 1024
 Canon … … … … 120 0.35– 0.5 512 3 267 1
Aquilion 0.47 512
ONE
Nagata H 2019 10 Compare standard CT Visual evaluation … … … … … …
et al angiography to UHR scores were
CT for visibility of improved for
small arteries UHR CT
  UHR CT … … … … 135 NR 0.25 1024 3 869 1
1024
 Canon … … … … 120 NR 0.5 387 1
Aquilion
ONE/
VISION
Motoyama S 2018 79 Compare standard CT UHR CT had … … … … … …
et al (30) angiography to UHR improved
CT for assessment of stenosis grading
stent lumen and in both stents
calcified lesions and calcified
lesions
  UHR CT … … … … NR 0.18 0.25 1024 3 NR 0.35
1024
 Canon … … … … NR 0.35 0.5–0.6 512 3 NR 0.275
Aquilion 512
ONE
Takagi H 2018 38 Compare diameter Correlation and … … … … … …
et al (25) stenosis of UHR agreement
CT and invasive between UHR
angiography CT and invasive
angiography was
excellent
  UHR CT … … … … 120 or 400 0.2 0.25 1024 3 388 0.35/
1024 0.375

Table 1 (continues)

Radiology: Volume 00: Number 0— 2020  n  radiology.rsna.org 3

Next-Generation Hardware Advances in CT

Table 1 (continued): Recent Major Cardiovascular Ultra-High-Resolution CT Studies

Gantry
Slice Image Rotation
Author and No. of Kilovolt Pixel Size Thickness Matrix DLP Time
CT Platform Date Participants Goal Conclusion Peak (kVp) (mm) (mm) Size (mGy-cm) (sec)
 Invasive … … … … … … … … … …
angiography
Yoshioka K 2018 24 Use UHR CT with Thinner slices … … … … … …
et al different slice improved
thickness for visualization
identification of artery
of artery of
Adamkiewicz
  UHR CT … … … … 120 0.2 0.25 1024 3 1113 0.5/0.75
1024
  Thick slice … … … … … 0.39 0.5 512 3 … …
reconstruc- 512
tion
Note.—PubMed search criteria ultra-high resolution CT reviewed for all in vivo cardiovascular studies, including human patients in studies performed
on scanners excluding software-based ultra-high-resolution modes, micro-CT, and cone-beam CT. DLP = dose-length product, UHR = ultra
high resolution.

0.625 mm (GE Healthcare) ar-

rays (29). Gantry rotation time
of 350 msec is also slower than
with current lower-spatial-reso-
lution platforms (GE Healthcare
[240 msec, single source], Phil-
ips [270 msec, single source],
and Siemens [250 msec, dual
source]), which limits temporal
resolution (Fig 1) (24).

Cardiac Applications of UHR

CT
Coronary artery imaging with
UHR CT was performed by
Motoyama et al in 2018 (30).
Figure 2:  In vivo CT image shows a 2.5-mm stent depicted with, A, ultra-high-resolution CT (U-HRCT) (0.18-mm resolu-
Intraluminal stent dimensions
tion) and, B, conventional-resolution CT (CRCT) (0.35-mm resolution). (Reprinted, with permission, from reference 30.) and stenosis of calcified le-
sions with UHR CT (spatial
resolution, 0.18 mm) were
making full comparison of resolution between CT platforms compared with conventional-resolution coronary CT angi-
across a spatial spectrum difficult (26). ography (spatial resolution, 0.35 mm) or invasive coronary
Improved spatial resolution is not without drawbacks. Image angiography. A total of 59 patients underwent both UHR CT
noise is partly dependent on the ratio of photons per detector ele- and invasive coronary angiography; UHR CT had a positive
ment; therefore, reduction of EID detector dimensions by 50% predictive value of 80% and a negative predictive value of
theoretically requires quadrupling of x-ray photon flux to main- 100% for stenosis of 70% or more on a per-segment analysis.
tain equivalent signal-to-noise ratio. However, reconstruction Comparisons between UHR CT and standard coronary CT
methods can offset this image noise penalty (27,28). In the future, angiography were available for 17 measurements of in-stent
head-to-head comparisons among different UHR strategies will restenosis. Readers visualized the stent lumen in 16 stents for
define advantages and disadvantages of the different approaches. UHR CT and in 11 stents for conventional-resolution CT.
Smaller detector size also means that with 160 slices, z-axis cover- Quantitative analysis enabled accurate identification of larger
age is only 4 cm, requiring multivolume or helical acquisitions that lumens in both stents and highly calcified lesions using UHR
increase motion artifact susceptibility. In comparison, two current CT, likely representing the improved spatial resolution (Fig
lower-spatial-resolution systems offer 16-cm full-volume cardiac 2). A smaller study of 38 patients with UHR CT compared
coverage with 320 3 0.5 mm (Canon Medical Systems) or 256 3 with diameter stenosis measured with invasive coronary angi-

4 radiology.rsna.org  n  Radiology: Volume 00: Number 0— 2020

 Kwan et al

Figure 3:  Multicontrast photon-counting CT images in a canine abdomen. A, Gray-scale CT image obtained using photons with single-energy bin. B, Multimaterial map
obtained using multienergy reconstruction, labeling iodine, gadolinium, and bismuth with different colors, resulting in different distributions. C, Iodine map alone. D, Gado-
linium (Gd)-enhanced map alone. E, Bismuth (Bi) map alone. F, Calcium map alone. (Reprinted, with permission, from reference 55.)

ography had per-segment sensitivity of 100% in the detection Photon-counting CT

of stenosis of more than 30% but a lower specificity of 67%,
with a small bias toward overestimation of stenosis (mean, PC CT is a cutting-edge form of spectral CT that could also
4% 6 7 [standard devation]) (25). Given the relationship represent a transformative technologic breakthrough. Spectral
between spatial resolution and radiation dose in UHR CT, it CT refers to the use of x-ray photon energy–dependent infor-
is noted that the dose reported in these two studies was 7.4 mation to differentiate and quantify material composition in a
mSv and 5.4 mSv, respectively. This is a clinically feasible dose three-dimensional space. The concept of spectral CT was first
in the study patients with reported body mass indexes of 22.9 proposed by Godfrey Hounsfield at the time of his original
and 25 kg/m2, respectively (25,30). These body mass indexes descriptions of x-ray CT, where he addressed the determina-
are lower than those of many patients with cardiovascular dis- tion of atomic number of material, and suggested that the
ease in Western countries who may require higher x-ray tube subtraction of images at 100 kV from 140 kV could enable
current and energy. Small focal spot imaging used in UHR differentiation of calcium from iodine (31). To extract spectral
CT requires lower power output to avoid damaging the an- data, spatially and temporally co-registered data sets of spec-
ode of the x-ray tube. Thus, further studies in obese patients trally distinct photon energies are required, which has not been
are needed. However, overall, these improvements have been feasible historically. However, in recent years, most major CT
substantial and of great clinical relevance for cardiac imaging, vendors have developed clinical dual-energy CT technology.
particularly in patients with coronary stents or heavy calcifi- Importantly, PC CT technologies that allow for more than two
cation due to advanced cardiovascular disease. energies are also in development.

Radiology: Volume 00: Number 0— 2020  n  radiology.rsna.org 5

Next-Generation Hardware Advances in CT

Table 2: Recent Major Photon-counting CT Studies

Author
and CT No. of Kilovolt Mean Gantry
Acquisition No. of Spectral Peak Reconstruc­tion CTDIvol Rotation
Technique Date Participants Goal Conclusion Groups (kVp) Pixel Size (mm) (mGy) Time (sec)
Bratke G et al 2020 10 Compare image Subjective quality and … … … … …
quality for in-stent ability to assess was
restenosis model higher in PC CT, noise
between PC CT was higher in PC CT
and dual-layer CT
 PCD … … … … 2 120 NR NR 1
 Dual-layer … … … … 2 120 NR NR 0.27
EID
Symons R 2019 10 Measure CAC at PC CT is able to quantify … … … … …
et al (49)* different dose CAC with better
scans agreement between low
dose and standard dose
imaging than EID
  PCD normal … … … … 2 120 NR 5.4 NR
dose
  PCD low … … … … 2 120 NR 1.6 NR
dose
  EID normal … … … … 1 120 NR 5.4 NR
dose
  EID low … … … 1 120 NR 1.6 NR
dose
Symons R 2018 16 Compare contrast- Image quality and noise … … … … …
et al (38)* enhanced head and were better on PC CT
neck vascular imaging than EID
with PC CT and EID
 PCD … 2 140 0.5 27.4 0.5
 EID … 1 120 0.5 3 0.6 24.4 0.5
Symons R 2018 18 Test PC CT high PC CT high resolution … … … … …
et al (46) resolution mode acquisition plus high
(0.25 mm) in resolution reconstruction
coronary stents is significantly better
than EID dual energy
in terms of image noise
  PCD high- … … … … 2 140 0.25 33.2 0.5
resolution
mode
  PCD high- … … … … 2 140 0.5 33.2 0.5
resolution
mode
 PCD … … … … 2 140 0.5 33.2 0.5
standard-
resolution
mode
 Dual-source … … … … 2 90/150 0.6 33.2 0.5
EID
Symons R 2017 1 Test simultaneous PC CT is able to
et al (55)* multicontrast accurately measure
imaging with iodine both first pass iodine
and gadolinium in a and late gadolinium
canine infarct model maps
 PCD … … 4 140 NR NR 0.5
Note.—PubMed search criteria PC CT reviewed for all cardiovascular-related studies, including human samples, volunteers, or patients and
major cardiac-relevant studies. CAC = coronary artery calcium score, EID = energy-integrating detector, NR = not reported, PC = photon
counting, PCD = PC detector.
* Studies with in vivo components.

6 radiology.rsna.org  n  Radiology: Volume 00: Number 0— 2020

 Kwan et al

Figure 4:  Images in a 92-year-old woman with chest pain. A, Dual-energy coronary CT angiogram. B, Material decomposition–based
calcium suppression and, C, corresponding invasive coronary angiogram show a patent lumen visualized with calcium suppression imaging.
Arrow indicates the same middle left anterior descending coronary lesion. (Reprinted, with permission, from reference 63.)

Figure 5:  X-ray images in a mouse obtained with, A, standard attenuation-based imaging, B, phase-contrast imaging, and, C, dark-field
imaging. Arrow indicates regions of enhanced contrast in the trachea (B) and lungs (C). (Reprinted, with permission, from reference 69.)

The physical basis for spectral CT has been previously re- simulate images under different conditions (such as subtraction
viewed (32). In brief, the linear attenuation coefficient of any of calcium or contrast) or across different x-ray energies (32).
given voxel can be estimated by imaging with two separate ener- Photon-counting detector (PCD) technology combines
gies and expressing it as the sum of two basis functions. Typically, the benefits of UHR and spectral CT by using a different type
the two functions represent either the photoelectric effect and of detector. This technology has also been recently reviewed
Compton scattering (attenuation basis) or two assumed materi- (33). In brief, traditional EIDs measure the sum (integral)
als (material basis). The combination does not necessarily reflect of the energies of the individual photons projected through
the true components of the object but expresses what combi- an object. Multiple energies are measured with high- and
nation of the two basis materials would accurately generate the low-energy data sets separated either physically (dual-layer,
measured linear attenuation coefficient. Use of more than two twin-beam, or dual-source CT) or temporally (sequential and
basis functions requires more than two energies, but this tech- rapid-switching CT) (Fig 1). Instead of arrays of EIDs, PCDs
nique enables more accurate representation of the linear attenu- are single solid-state detectors. Energy separation is per-
ation coefficient of a voxel and closer identification of materi- formed using user- or manufacturer-specified sequential en-
als. Extraction of the energy-dependent spectral information is ergy thresholds. When a photon hits a PCD, the energy pulse
termed material decomposition. This adds material-specific infor- of the photon is measured, and a photon is “counted” for
mation, such as effective atomic number, as well as the ability to each surpassed threshold. By subtracting counts in sequential

Radiology: Volume 00: Number 0— 2020  n  radiology.rsna.org 7

Next-Generation Hardware Advances in CT

tector element size is not deter-

mined by physical separation
between elements. Therefore,
PCD element size can be quite
small (range, 0.11 3 0.11 mm
to 0.5 3 0.5 mm physical de-
tector pixel size). In contrast,
EIDs require light-reflecting
septa between elements of scin-
tillating crystals, which makes
manufacturing of small ele-
ments difficult. Initial studies
in UHR PC CT using a detec-
tor mode with 0.25-mm appar-
ent pixel size at the isocenter
compared with a 0.5-mm de-
tector mode (standard resolu-
Figure 6:  Attenuation-based radiograph and phase-contrast refraction radiograph of a poly(methyl methacrylate) tion) showed improvements in
(PMMA) rod and a 0.3-mm nylon filament immersed in a 5-cm-thick water bath. The images were obtained at 60 kVp with spatial resolution in humans
a glancing angle interferometer. The computed radiation dose is 2 mGy absorbed dose. The thin filament is invisible on the
conventional image (left) but has good contrast with refraction (right).
(42,44) and in phantoms with
coronary stents (46–48), albeit
at the cost of increased noise
thresholds from each other, photon counts within each en- when using high-resolution reconstruction (37).
ergy bin are obtained (33). Additionally, in PCDs, all photons are weighted equally in-
Current challenges include accurately registering and measur- stead of being weighted by their energies in EID. This results in
ing photon energies. Specifically, cross-talk effects (“k-escape,” accurate weighting of photons in the low-energy range, where
“charge-sharing,” or fluorescence), where multiple detector pixels there is greater contrast between soft tissues, even without us-
register a single higher energy charge due to scattering, can cause ing spectral information. Setting minimum energy thresholds
misregistration of high-energy photons as multiple lower-energy filters out electronic noise, which improves reduced-dose (20 vs
photons. Pulse pile-up effects occur if multiple photons interact 80 mAs) performance of the scanner, especially in a screening
with the detector within too brief a timeframe for the detector task, such as coronary calcium scoring. These factors reduced
to resolve the separate events, resulting in multiple lower-energy the radiation dose of PC CT for coronary artery calcium scoring
charges incorrectly registered as a single higher-energy photon CT by 75% compared with EID, while maintaining image qual-
(33). Cross talk can be decreased by summing adjacent pixels ity in human studies (49). Thus, PC CT may improve contrast
during coincident events to identify the true energy and by as- resolution and decrease noise simultaneously.
signing the count to the area with the highest energy (34,35). Finally, PC CT theoretically has superior spectral imaging
However, this results in increasing susceptibility to pulse pile-up, compared with spectral CT using EID detectors. PCD spectral
which has required multiple approaches to address. These ap- information is inherently co-registered, and spectral separation
proaches have included beam filtering to avoid excessive flux, may be more effective with the adjustable energy thresholds
electronics improvements to decrease overall deadtime by layer- (50,51). PCD CT can also use more than two energy bins,
ing detector elements, and post hoc compensation algorithms improving the ability to quantify materials within a given space
and reconstruction methods (36). (52) and making multicontrast imaging possible (Fig 3) (53–
PC CT is currently only available on experimental scan- 57). However, using more energy bins does result in decreased
ners, with Siemens prototypes being used to perform imag- photon count and increased noise in each bin, although re-
ing in human volunteers at three locations (33) and a Phil- construction and postprocessing denoising algorithms have
ips prototype being tested in Lyons, France. Multiple small been proposed (58,59). In summary, PCDs may revolutionize
human studies totaling more than 100 patients have already CT by offering increased contrast, lower noise, and spectral
been performed and published by the National Institutes of capabilities, while improving spatial resolution, with multiple
Health and the Mayo Clinic. These studies include varying cardiovascular-related studies underway (Table 2).
scenarios of diagnostic quality PC CT, including abdominal,
chest, cardiovascular, and brain imaging to demonstrate tech- Cardiac Applications of PC CT
nical feasibility (37–45). Because of the superior spectral abilities already cited, PC CT
is expected to exploit current dual-energy techniques for car-
Potential Advantages of PC CT Compared with EIDs diovascular applications within cardiology. For example, vir-
There are multiple theoretical advantages of CT examinations tual monochromatic images optimize the visibility of coronary
that use PCDs instead of EIDs. The direct conversion of x-ray stenosis (60) and infarcted myocardium (61,62); virtual non-
photons to electric charges in solid-state PCDs means that de- calcium images improve visualization of the coronary artery

8 radiology.rsna.org  n  Radiology: Volume 00: Number 0— 2020

 Kwan et al

Figure 7:  A, Talbot-Lau interferometer set-up with, B, glancing angle grids. G =

grating. (Reprinted, with permission, from reference 75.)

lumen (Fig 4) (63); and quantification of iodine content may

improve myocardial perfusion imaging (64) and extracellular
volume measurement (65). Beyond dual-energy capabilities,
the spectral potential of PC cardiac CT has not been fully
demonstrated. An exception is a proof-of-concept study with
PC CT that evaluated iodine first-pass images and delayed gad-
olinium-enhanced images simultaneously in a canine model of
Figure 8:  High-energy region-of-interest phase-contrast tomosynthesis
myocardial infarction (53). Further improvement in basis ma- system for the internal organs based on the glancing angle interferometer
terial decomposition is needed (66). A dual-source PC CT de- design. G = grating.
tector developed by Tao et al may improve temporal resolution
while maintaining UHR and spectral advantages (67). Refine-
ment of spectral applications, radiation reduction, and 250- the object (d), related to electron density, and the wavelength
mm high-spatial-resolution capabilities of PC CT are likely to of the x-ray. Dark-field imaging uses disorganized scatter from
be initial applications of this technology. microscopic density gradients in the medium to create contrast.
This reveals tissue-specific characteristics, like microstructural
Phase-Contrast CT inhomogeneity (72). More detailed technical explanations can
Current CT technologies achieve contrast resolution due to the be found in the referenced literature and fall outside the scope of
attenuation of x-ray photons or the use of intravenous iodinated this review (73,74).
contrast media to detect small differences in attenuation between
soft tissues. Instead of attenuation, an alternative contrast mecha- Benefits of Phase-Contrast CT
nism is to exploit x-ray wave properties of refraction, phase shift, Phase-contrast and dark-field imaging have a strong theoretical
or ultrasmall angle scattering. These modes of contrast in CT are advantage over attenuation-based CT for soft-tissue contrast.
known as phase-contrast CT and dark-field CT, which reveal dif- This may be highly applicable to cardiac imaging. The refractive
ferent tissue characteristics and improve soft-tissue contrast resolu- index (d) of soft tissue is three orders of magnitude greater than
tion versus attenuation-based CT (68). Phase-contrast and attenu- the absorptive index (b) at standard x-ray energy and diminishes
ation imaging should be conceptualized as complementary forms less with increased photon energy than b. This means greater
of imaging that provide different information (Fig 5) (69). soft-tissue contrast with higher x-ray energy (75). Phase-contrast
Phase-contrast imaging uses either the organized changes in imaging may show improved edge contrast due to high sensitiv-
x-ray phase (F) or refraction (a) to generate contrast (70,71). ity to density gradients, which could improve high-resolution
The degree of change is determined by the refractive index of imaging of stents or catheters. This capability is illustrated in

Radiology: Volume 00: Number 0— 2020  n  radiology.rsna.org 9

Next-Generation Hardware Advances in CT

Figure 9:  Top: Full-field and region of interest (ROI) phase-contrast CT (PC-CT) scans of fresh pig soft tissues in water at diagnostic energy. The imaged section is 25
mm in diameter. Bottom: Full-field and full-scan phase-contrast CT images and simultaneous region-of-interest and limited-angle phase-contrast tomosynthesis of soft tis-
sues in water. Also shown is the conventional attenuation–based tomosynthesis image, indicating much less soft-tissue contrast than in the phase-contrast image. PMMA =
poly(methyl methacrylate).

Figure 6 (75). The use of higher energy x-rays without relying on clinically feasible. Some limitations in stability, field of view, and
attenuation could reduce radiation absorption (70,71). need for x-ray monochromacy and spatial coherence have been
Preclinical studies including in vivo imaging show the po- addressed in part by using a Talbot-Lau shearing interferometer
tential of phase-contrast CT and dark-field imaging. A large (84). This design places gratings between the x-ray source and the
body of these studies are from the Munich Biomedical Phys- imaged object, which creates spatial coherence and decodes the
ics group under the purview of Dr Franz Pfeiffer. In vivo im- phase shifts from the object. Glancing angle interferometry (Fig
aging has been performed in swine (76,77) and mouse (78) 7) angles the Talbot-Lau grating, which increases the interferom-
models. Pulmonary and mammography applications appear eter fringe contrast. This interferometer set-up enabled the first
particularly promising (79,80). Efforts are being made to in- demonstration of high-energy phase-contrast CT with clinically
crease the field of view and improve temporal resolution and compatible absorbed dose exposure (8 mGy) (75,85). The current
mechanical stability (81,82). designs maintain beam spatial coherence using a standard poly-
chromatic x-ray tube, with sufficient mechanical stability to ap-
Development of Phase-Contrast CT toward Clinical ply to CT, and can also simultaneously acquire dark-field imaging
Cardiovascular Imaging and attenuation images (72). With these principles, the Munich
Phase-contrast CT has not been developed for in vivo human group has recently created a dark-field chest x-ray prototype. It
application. Phase-contrast radiography has existed since the has received approval from the German Federal Office for Radia-
1960s, but the first published phase-contrast CT study was per- tion Protection to begin in vivo patient imaging to detect incipient
formed by Momose et al in 1994 using monochromatic x-rays alveolar inflammation and edema from COVID-19 infection. It
from a synchrotron (83). Clinical cardiac application of phase- allows enhanced soft-tissue resolution and lower radiation doses
contrast CT requires multiple items for success: sufficient me- than does attenuation-based x-ray imaging (72,86).
chanical stability to both rotate and measure small changes in High-intensity laser-driven x-ray sources may be the
x-rays, adequate field of view, and clinically feasible nonsynchro- next step for cardiac phase-contrast imaging. While the
tron x-ray sources with energies sufficient for clinical imaging. Talbot-Lau approach has addressed issues with coherence,
Synchrotron radiation has been used, as it provides spec- intensity is also important, as the technology moves from
trally and spatially coherent high-flux photons, where refraction small-object studies to studies in larger, more clinically ap-
or phase changes can be easily measured; however, this is not plicable subjects. Conventional x-ray tubes are suboptimally

10 radiology.rsna.org  n  Radiology: Volume 00: Number 0— 2020

 Kwan et al

Table 3: Recent Major Phase-Contrast CT Studies

No. of
Author Date Imaging Sample Subjects Goal Conclusion PXI Method X-Ray Source
Oda H et al 2020 Rabbit heart 1 Compare PXI CT to Both PXI CT and x-ray Crystal based Synchrotron
(100) absorption microfocus micro CT could accurately
x-ray CT for cardiac track fibers, but micro CT
fiber analysis had lower contrast and
artifacts
Reichardt M 2020 Mouse heart 5 Test different sample PXI CT can visualize Propagation Liquid metal jet
et al (98) preparations for individual heart based with metal
visualization of fibers at resolution filter
fiber structure less than 10 µm
Braig E et al 2018 Chicken heart 1 Quantitative material Material decomposition Grating based Laser
(103) decomposition of by effective atomic synchrotron
phantom and number is possible
chicken heart
Romell J et al 2018 Human 1 Visualize blood PXI CT can visualize Propagation Microfocus
mummy vessels, tendons, tissues with resolution based x-ray source
and anatomy less than 10 µm
Vagberg W 2018 Human 5 Compare x-ray “virtual Visualization of coronary Propagation Liquid metal jet
et al coronary histology” with pathology including based
artery histology in coronary cholesterol crystals,
samples arteries foam cells, microscopic
plaque is possible
Ferraro M et al 2018 Mouse aorta 1 Visualize aortic strain Microtears and early Propagation Synchrotron
(102) in aneurism model vascular damage are seen based
in susceptible areas
Hetterich H 2017 Human 15 Visualize coronary Dark field imaging can Grating based Polychromatic
et al (94) coronary microcalcifications and assess coronary plaque x-ray tube
artery test improvement with microcalcifications
samples dark field imaging
Gonzalez- 2017 Rabbit fetus 2/1 Visualize in vitro PXI CT can visualize Propagation Synchrotron
Tendero A and rat heart cardiac structure microanatomy including based
et al (99) full coronary tree and
myocardial fiber
orientation
Shinohara G 2016 Human heart 4 Visualize conduction Conduction system Propagation Synchrotron
et al (101) sample system in human is accurately visualized based
heart and compare by PXI CT
with histology
Zamir A et al 2016 Phantoms and 1 Improve artifacts in PXI Performance removes Grating based Polychromatic
rat heart CT by compensating major artifacts and x-ray tube
for system instability improves image quality
Allner S et al 2016 Human coronary 1/1 Reduce noise using Noise reduction and Grating based Polychromatic
and cerebellum bilateral filter improved edge x-ray tube
sample detection is seen
Hetterich H 2016 Human coronary 127 Classify coronary and PXI CT is reliable in Grating based Polychromatic
et al (95) and carotid carotid plaque into noninvasive plaque x-ray tube
samples AHA histologic classification versus
categories by PXI CT histology
Winklhofer S 2015 Human coronary 40 Compare PXI CT PXI CT has higher Grating based Synchrotron
et al (97) artery samples to absorption CT quality and
for quantitative better diagnostic
assessment of plaque accuracy
versus histology
Hetterich H 2014 Human carotid 56 Compare PXI CT to PXI CT is well correlated Grating based Polychromatic
et al artery samples histology for to histology for multiple x-ray tube
identification of high-risk plaque
plaque characteristics characteristics

Table 3 (continues)

Radiology: Volume 00: Number 0— 2020  n  radiology.rsna.org 11

Next-Generation Hardware Advances in CT

Table 3 (continued): Recent Major Phase-contrast CT Studies

No. of Phase-contrast
Author Date Imaging Sample Subjects Goal Conclusion Method X-Ray Source
Saam T et al 2013 Human carotid 5 Compare synchrotron PXI CT overestimated Grating based Synchrotron/
artery samples versus polychromatic vessel sizes. Synchrotron Polychromatic
x-ray with grating imaging had higher SNR x-ray tube
imaging methods for but both methods feasible
carotid vessel analysis
versus histology
Hetterich H 2013 Human carotid 5 Compare 23 versus 53 PXI CT was feasible Grating based Synchrotron
et al (93) artery samples keV energies for PXI and accurate at 53 keV
CT of carotid arteries
versus histology
Appel AA 2013 Human carotid 20 Compare PXI CT PXI CT was accurate Analyzer based Synchrotron
et al artery plaque absorption/refraction/ and was able to
samples scatter images to visualize small plaque
histology details and structure
Takeda M et al 2012 Mouse 50 Assess plaque content PXI CT can identify Crystal based Synchrotron
brachiocephalic for in mice treated decreased lesion volume
arteries with antiplatelet and increased stability in
therapies antiplatelet treatment
Shinohara M 2008 Mouse arteries NR Compare atherosclerotic PXI CT can visualize Crystal based Synchrotron
et al (101) plaque quantity by PXI plaque and identify
CT versus histology plaque components
Note.—PubMed search criteria phase-contrast CT reviewed for all phase-contrast studies with cardiovascular imaging. All studies were
performed ex vivo. AHA = American Heart Association, OCT = optical coherence tomography, PXI = phase contrast, SNR = signal-
to-noise ratio, NR = not reported.

phase-contrast CT of medi-
cally relevant samples (89).
Field-of-view limitations
may make phase-contrast CT
more applicable for cardiovascu-
lar imaging. Phase-contrast CT
is well-suited to limited-angle,
region-of-interest (ROI), or in-
terior tomography. A practical
option for cardiac CT would
be to use a scanner with a small
field of view (Fig 8) with ROI
Figure 10:  Ex vivo images of complex fibrous material (*) within plaque in the right coronary artery obtained, A, with and phase-contrast tomosynthe-
phase-contrast CT and, B, at histology. (Reprinted, with permission, from reference 95.) sis (limited-angle CT) to target
a small organ. Theoretical and
experimental results show that phase-contrast projections are
suited for phase-contrast imaging. This is due to the need better suited for ROI CT than for conventional attenuation-
for spatial coherence and long propagation distances for based projections (90,91). Figure 9 shows simultaneous ROI and
the phase effects to develop (order of meters). The photon tomosynthesis of a fresh tissue phantom at high energy, obtained
concentration from conventional x-rays at these distances with a glancing angle interferometer operated at 80 kVp. ROI
is low. The next step for cardiac phase-contrast imaging phase-contrast CT using smaller gratings may reduce both the
may include high-intensity lasers (peak power in the 100- effective dose and the scan time, even at high spatial resolution.
TW range and femtosecond pulse duration). These lasers
produce an ultrabright, directional, and spatially coherent Cardiac Applications of Phase-Contrast CT
x-ray source. By using ultrashort x-ray pulses with a high Cardiovascular focused phase-contrast CT studies are reviewed
repetition rate, these lasers may address temporal resolu- in Table 3. The bulk of cardiovascular studies have performed
tion for cardiac phase-contrast imaging. The 100-TW class “virtual histology,” comparing phase-contrast CT of coronary
lasers are becoming commercially affordable, and world- and carotid atherosclerosis with histology (92–97). These studies
wide efforts to develop laser-based phase-contrast imaging identify high-risk plaque features and quantify plaque contents
are ongoing (87,88), as demonstrated by laser-driven x-ray with excellent correlation with histology (Fig 10) and spatial res-

12 radiology.rsna.org  n  Radiology: Volume 00: Number 0— 2020

 Kwan et al

olution to around 10 mm (95). In 2017, Habbel et al compared State-of-the-Art Review. J Am Coll Cardiol 2019;74(12):1608–1617.
6. Kwan AC, Cater G, Vargas J, Bluemke DA. Beyond coronary stenosis:
ex vivo human coronary phase-contrast CT with optical coher- coronary computed tomographic angiography for the assessment of athero-
ence tomography, with histopathology as the reference standard. sclerotic plaque burden. Curr Cardiovasc Imaging Rep 2013;6(2):89–101.
In a sample of 241 matched cross sections, phase-contrast CT 7. Seraphim A, Knott KD, Augusto J, Bhuva AN, Manisty C, Moon JC.
Quantitative cardiac MRI. J Magn Reson Imaging 2020;51(3):693–711.
images had superior correlation with histology findings between 8. Siegersma KR, Leiner T, Chew DP, Appelman Y, Hofstra L, Verjans JW.
volumetric measures of vessel, lumen, and plaque versus optical Artificial intelligence in cardiovascular imaging: state of the art and implica-
coherence tomography. Phase-contrast CT also yielded supe- tions for the imaging cardiologist. Neth Heart J 2019;27(9):403–413.
9. Litjens G, Ciompi F, Wolterink JM, et al. State-of-the-art deep learning
rior identification of lipid-rich, fibrous, or calcified plaque (92). in cardiovascular image analysis. JACC Cardiovasc Imaging 2019;12(8 Pt
Other applications have leveraged the 10-mm spatial resolution 1):1549–1565.
and high soft-tissue contrast. These have included visualizing 10. Boedeker K. AiCE Deep Learning Reconstruction: Bringing the power of
Ultra-High Resolution CT to routine imaging. Tustin, Calif: Canon Medical
myocardial fiber orientation (98–100), identifying the cardiac Systems Corporation, 2019.
conduction system (101), visualizing microscopic vascular dam- 11. Benz DC, Benetos G, Rampidis G, et al. Validation of deep-learning image
age (102), and performing cardiac material decomposition reconstruction for coronary computed tomography angiography: Impact on
noise, image quality and diagnostic accuracy. J Cardiovasc Comput Tomogr
(103). These experimental techniques may enable noninvasive 2020. 10.1016/j.jcct.2020.01.002. Published online January 13, 2020.
high-resolution imaging of high-risk plaque and identification of 12. Nørgaard BL, Terkelsen CJ, Mathiassen ON, et al. Coronary CT angiographic
subtle myocardial abnormalities in patients with cardiac disease. and flow reserve-guided management of patients with stable ischemic heart
disease. J Am Coll Cardiol 2018;72(18):2123–2134.
13. Weir-McCall JR, Nicol E, Abbara S, et al. Highlights of the fourteenth annual
Conclusion scientific meeting of the Society of Cardiovascular Computed Tomography.
J Cardiovasc Comput Tomogr 2020;14(2):118–123.
Hardware advances, including ultra-high-resolution (UHR) CT, 14. Chandrashekhar Y; JACC: Cardiovascular Imaging Editors. What Is of Recent
Interest in Cardiac CT: Insights From the JACC Family of Journals. J Am
photon-counting CT, and phase-contrast CT, are opening new Coll Cardiol 2019;73(25):3352–3355.
aspects to the field of cardiac CT. UHR CT improves imaging 15. Nicol ED, Norgaard BL, Blanke P, et al. The future of cardiovascular com-
of heavily calcified arteries, cardiac valves, and stents. Functional puted tomography: advanced analytics and clinical insights. JACC Cardiovasc
Imaging 2019;12(6):1058–1072.
applications for further exploration include atherosclerotic 16. Food and Drug Administration 510(k) Clearances. Aquilion Precision
plaque characterization, stent imaging, fractional flow reserve, (TSX-304A/2) V8.6 with FIRST 3.0 501(k) Clearance. Department of
myocardial perfusion imaging, and microvessel imaging. Pho- Health and Human Services, Food and Drug Administration. https://
www.accessdata.fda.gov/cdrh_docs/pdf17/K173468.pdf. Published 2018.
ton-counting CT uses multiple photon energies to reduce arti- Accessed November 9, 2018.
facts, improve contrast, and perform material decomposition. Its 17. Hata A, Yanagawa M, Honda O, et al. Effect of Matrix Size on the Image
potential could revolutionize medical imaging, but the detector Quality of Ultra-high-resolution CT of the Lung: Comparison of 512 3
512, 1024 3 1024, and 2048 3 2048. Acad Radiol 2018;25(7):869–876.
technology needs further refinement for clinical applications. 18. Biondi-Zoccai G, Moretti C, Abbate A, Sheiban I. Percutaneous coronary
Finally, phase-contrast CT uses x-ray refraction properties to im- intervention for small vessel coronary artery disease. Cardiovasc Revasc Med
prove spatial and soft-tissue contrast. Potential applications of 2010;11(3):189–198.
19. Silverio A, Buccheri S, Venetsanos D, et al. Percutaneous treatment and
phase-contrast CT to breast and lung cancer in addition to pul- outcomes of small coronary vessels: a SCAAR report. JACC Cardiovasc
monary parenchymal disease processes complement its potential Interv 2020;13(7):793–804.
in cardiovascular medicine. Despite technical hurdles, hardware 20. Motoyama S, Ito H, Sarai M, et al. Plaque characterization by coronary
computed tomography angiography and the likelihood of acute coronary
advances in cardiac CT continue to expand and have the poten- events in mid-term follow-up. J Am Coll Cardiol 2015;66(4):337–346.
tial to transform cardiac imaging. 21. Williams MC, Moss AJ, Dweck M, et al. Coronary artery plaque charac-
teristics associated with adverse outcomes in the SCOT-HEART study. J
Disclosures of Conflicts of Interest: A.C.K. disclosed no relevant relationships. Am Coll Cardiol 2019;73(3):291–301.
A.P. Activities related to the present article: disclosed no relevant relationships. Ac- 22. Virmani R, Kolodgie FD, Burke AP, Farb A, Schwartz SM. Lessons from sud-
tivities not related to the present article: is a consultant for Cardiowise, institution den coronary death: a comprehensive morphological classification scheme for
received grants from Siemens Healthcare and Canon Medical Research, holds stock atherosclerotic lesions. Arterioscler Thromb Vasc Biol 2000;20(5):1262–1275.
options in Cardiowise. Other relationships: disclosed no relevant relationships. D.S. 23. Arbab-Zadeh A, Fuster V. From Detecting the Vulnerable Plaque to Man-
Activities related to the present article: disclosed no relevant relationships. Activities aging the Vulnerable Patient: JACC State-of-the-Art Review. J Am Coll
not related to the present article: disclosed no relevant relationships. Other rela- Cardiol 2019;74(12):1582–1593.
tionships: Johns Hopkins University was assigned U.S. patent 9,329,141 for the 24. DAIC: Diagnostic and Interventional Cardiology. Cardiac CT Systems
Glancing Angle Interferometer. D.A.B. disclosed no relevant relationships. J.A.C.L. Comparison Chart. https://www.dicardiology.com/content/cardiac-ct-
disclosed no relevant relationships. A.C.K. disclosed no relevant relationships. systems. Published 2019. Accessed April 26, 2020.
25. Takagi H, Tanaka R, Nagata K, et al. Diagnostic performance of coronary
CT angiography with ultra-high-resolution CT: Comparison with invasive
References coronary angiography. Eur J Radiol 2018;101:30–37.
1. Schmidt B, Grant K, Flohr TG, Allmendinger T. Cardiac CT Platforms: 26. Friedman SN, Fung GS, Siewerdsen JH, Tsui BM. A simple approach to
State of the Art. In: Schoepf UJ, ed. CT of the Heart. 2nd ed. New York, measure computed tomography (CT) modulation transfer function (MTF)
NY: Humana/Springer, 2019; 51–67. and noise-power spectrum (NPS) using the American College of Radiology
2. Flohr TG, Allmendinger T, Bruder H, Schwemmer C, Kappler S, Schmidt (ACR) accreditation phantom. Med Phys 2013;40(5):051907.
B. Future Technological Advances in Cardiac CT. In: Schoepf UJ, ed. CT 27. 10. Spatial Resolution in CT. J ICRU 2012;12(1):107–120.
of the Heart. 2nd ed. New York, NY: Humana/Springer, 2019; 873–892. 28. Lin E, Alessio A. What are the basic concepts of temporal, contrast,
3. Schoepf UJ, ed. CT of the Heart. 2nd ed. New York, NY: Humana/Springer, and spatial resolution in cardiac CT? J Cardiovasc Comput Tomogr
2019. 2009;3(6):403–408.
4. Dodge JT Jr, Brown BG, Bolson EL, Dodge HT. Lumen diameter of normal 29. Lewis MA, Pascoal A, Keevil SF, Lewis CA. Selecting a CT scanner for cardiac
human coronary arteries. Influence of age, sex, anatomic variation, and left imaging: the heart of the matter. Br J Radiol 2016;89(1065):20160376.
ventricular hypertrophy or dilation. Circulation 1992;86(1):232–246. 30. Motoyama S, Ito H, Sarai M, et al. Ultra-High-Resolution Computed To-
5. Ahmadi A, Argulian E, Leipsic J, Newby DE, Narula J. From Subclinical mography Angiography for Assessment of Coronary Artery Stenosis. Circ J
Atherosclerosis to Plaque Progression and Acute Coronary Events: JACC 2018;82(7):1844–1851.

Radiology: Volume 00: Number 0— 2020  n  radiology.rsna.org 13

Next-Generation Hardware Advances in CT

31. Hounsfield GN. Computerized transverse axial scanning (tomography). 1. 56. Cormode DP, Si-Mohamed S, Bar-Ness D, et al. Multicolor spectral photon-
Description of system. Br J Radiol 1973;46(552):1016–1022. counting computed tomography: in vivo dual contrast imaging with a high
32. McCollough CH, Leng S, Yu L, Fletcher JG. Dual- and Multi-Energy CT: count rate scanner. Sci Rep 2017;7(1):4784.
Principles, Technical Approaches, and Clinical Applications. Radiology 57. Dangelmaier J, Bar-Ness D, Daerr H, et al. Experimental feasibility of spec-
2015;276(3):637–653. tral photon-counting computed tomography with two contrast agents for
33. Willemink MJ, Persson M, Pourmorteza A, Pelc NJ, Fleischmann D. the detection of endoleaks following endovascular aortic repair. Eur Radiol
Photon-counting CT: technical principles and clinical prospects. Radiology 2018;28(8):3318–3325.
2018;289(2):293–312. 58. Harrison AP, Xu Z, Pourmorteza A, Bluemke DA, Mollura DJ. A multichan-
34. Taguchi K, Polster C, Lee O, Stierstorfer K, Kappler S. Spatio-energetic cross nel block-matching denoising algorithm for spectral photon-counting CT
talk in photon counting detectors: Detector model and correlated Poisson images. Med Phys 2017;44(6):2447–2452.
data generator. Med Phys 2016;43(12):6386–6404. 59. Yu Z, Leng S, Li Z, McCollough CH. Spectral prior image constrained
35. Koenig T, Hamann E, Procz S, et al. Charge summing in spectroscopic x-ray compressed sensing (spectral PICCS) for photon-counting computed to-
detectors with high-Z sensors. IEEE Trans Nucl Sci 2013;60(6):4713–4718. mography. Phys Med Biol 2016;61(18):6707–6732.
36. Taguchi K. Energy-sensitive photon counting detector-based X-ray computed 60. Stehli J, Clerc OF, Fuchs TA, et al. Impact of monochromatic coronary
tomography. Radiological Phys Technol 2017;10(1):8–22. computed tomography angiography from single-source dual-energy
37. Pourmorteza A, Symons R, Henning A, Ulzheimer S, Bluemke DA. Dose CT on coronary stenosis quantification. J Cardiovasc Comput Tomogr
Efficiency of Quarter-Millimeter Photon-Counting Computed Tomography: 2016;10(2):135–140.
First-in-Human Results. Invest Radiol 2018;53(6):365–372. 61. Sandfort V, Palanisamy S, Symons R, et al. Optimized energy of spectral
38. Symons R, Reich DS, Bagheri M, et al. Photon-Counting Computed To- CT for infarct imaging: Experimental validation with human validation. J
mography for Vascular Imaging of the Head and Neck: First In Vivo Human Cardiovasc Comput Tomogr 2017;11(3):171–178.
Results. Invest Radiol 2018;53(3):135–142. 62. Sandfort V, Kwan AC, Elumogo C, et al. Automatic high-resolution infarct
39. Pourmorteza A, Symons R, Reich DS, et al. Photon-counting CT of the detection using volumetric multiphase dual-energy CT. J Cardiovasc Comput
brain: in vivo human results and image-quality assessment. AJNR Am J Tomogr 2017;11(4):288–294.
Neuroradiol 2017;38(12):2257–2263. 63. Yunaga H, Ohta Y, Kaetsu Y, et al. Diagnostic performance of calcification-
40. Symons R, Pourmorteza A, Sandfort V, et al. Feasibility of dose-reduced chest suppressed coronary CT angiography using rapid kilovolt-switching dual-
CT with photon-counting detectors: initial results in humans. Radiology energy CT. Eur Radiol 2017;27(7):2794–2801.
2017;285(3):980–989. 64. Jin KN, De Cecco CN, Caruso D, et al. Myocardial perfusion imaging with
41. Pourmorteza A, Symons R, Sandfort V, et al. Abdominal imaging with dual energy CT. Eur J Radiol 2016;85(10):1914–1921.
contrast-enhanced photon-counting CT: first human experience. Radiology 65. Lee HJ, Im DJ, Youn JC, et al. Myocardial Extracellular Volume Fraction with
2016;279(1):239–245. Dual-Energy Equilibrium Contrast-enhanced Cardiac CT in Nonischemic
42. Bartlett DJ, Koo CW, Bartholmai BJ, et al. High-Resolution Chest Computed Cardiomyopathy: A Prospective Comparison with Cardiac MR Imaging.
Tomography Imaging of the Lungs: Impact of 1024 Matrix Reconstruction Radiology 2016;280(1):49–57.
and Photon-Counting Detector Computed Tomography. Invest Radiol 66. Leng S, Zhou W, Yu Z, et al. Spectral performance of a whole-body research
2019;54(3):129–137. photon counting detector CT: quantitative accuracy in derived image sets.
43. Marcus RP, Fletcher JG, Ferrero A, et al. Detection and Characteriza- Phys Med Biol 2017;62(17):7216–7232.
tion of Renal Stones by Using Photon-Counting-based CT. Radiology 67. Tao S, Rajendran K, McCollough CH, Leng S. Multi-contrast imaging on
2018;289(2):436–442. dual-source photon-counting-detector (PCD) CT. In: Gilat Schmidt T, Chen
44. Leng S, Rajendran K, Gong H, et al. 150-mm spatial resolution using photon- GH, Bosmans H, eds. Proceedings of SPIE: medical imaging 2019—physics
counting detector computed tomography technology: technical performance of medical imaging. Vol 10948. Bellingham, Wash: International Society
and first patient images. Invest Radiol 2018;53(11):655–662. for Optics and Photonics, 2019; 109484L.
45. Zhou W, Bartlett DJ, Diehn FE, et al. Reduction of metal artifacts and 68. Bonse U, Hart M. An X-ray interferometer. Appl Phys Lett 1965;6(8):155–
improvement in dose efficiency using photon-counting detector computed 156.
tomography and tin filtration. Invest Radiol 2019;54(4):204–211. 69. Bech M, Tapfer A, Velroyen A, et al. In-vivo dark-field and phase-contrast
46. Symons R, De Bruecker Y, Roosen J, et al. Quarter-millimeter spectral x-ray imaging. Sci Rep 2013;3(1):3209.
coronary stent imaging with photon-counting CT: Initial experience. J 70. Stutman D, Beck TJ, Carrino JA, Bingham CO. Talbot phase-contrast x-ray
Cardiovasc Comput Tomogr 2018;12(6):509–515. imaging for the small joints of the hand. Phys Med Biol 2011;56(17):5697–
47. Mannil M, Hickethier T, von Spiczak J, et al. Photon-counting CT: high- 5720.
resolution imaging of coronary stents. Invest Radiol 2018;53(3):143–149. 71. Lewis RA. Medical phase contrast x-ray imaging: current status and future
48. von Spiczak J, Mannil M, Peters B, et al. Photon counting computed to- prospects. Phys Med Biol 2004;49(16):3573–3583.
mography with dedicated sharp convolution kernels: tapping the potential 72. Pfeiffer F, Bech M, Bunk O, et al. Hard-X-ray dark-field imaging using a
of a new Technology for Stent Imaging. Invest Radiol 2018;53(8):486–494. grating interferometer. Nat Mater 2008;7(2):134–137.
49. Symons R, Sandfort V, Mallek M, Ulzheimer S, Pourmorteza A. Coronary 73. Fitzgerald R. Phase-Sensitive X-Ray Imaging. Phys Today 2000;53(7):23–26.
artery calcium scoring with photon-counting CT: first in vivo human experi- 74. Momose A, Takeda T, Itai Y, Hirano K. Phase-contrast X-ray computed
ence. Int J Cardiovasc Imaging 2019;35(4):733–739. tomography for observing biological soft tissues. Nat Med 1996;2(4):473–475
50. Atak H, Shikhaliev PM. Dual energy CT with photon counting and dual source [Published correction appears in Nat Med 1996;2(5):596.].
systems: comparative evaluation. Phys Med Biol 2015;60(23):8949–8975. 75. Sarapata A, Stayman JW, Finkenthal M, Siewerdsen JH, Pfeiffer F, Stut-
51. Tao S, Wu Y, Rajendran K, McCollough CH, Leng S. Optimal acquisition man D. High energy x-ray phase contrast CT using glancing-angle grating
setting for dual-contrast imaging with gadolinium and iodine on a research interferometers. Med Phys 2014;41(2):021904.
whole-body photon-counting-detector (PCD) CT. In: Gilat Schmidt T, Chen 76. De Marco F, Willer K, Gromann LB, et al. Contrast-to-noise ratios and
GH, Bosmans H, eds. Proceedings of SPIE: medical imaging 2019—physics thickness-normalized, ventilation-dependent signal levels in dark-field
of medical imaging. Vol 10948. Bellingham, Wash: International Society and conventional in vivo thorax radiographs of two pigs. PLoS One
for Optics and Photonics, 2019; 109484I. 2019;14(6):e0217858.
52. Yveborg M, Danielsson M, Bornefalk H. Theoretical comparison of 77. Gromann LB, De Marco F, Willer K, et al. In-vivo x-ray dark-field chest
a dual energy system and photon counting silicon detector used for radiography of a pig. Sci Rep 2017;7(1):4807.
material quantification in spectral CT. IEEE Trans Med Imaging 78. Gradl R, Morgan KS, Dierolf M, et al. Dynamic in vivo chest x-ray dark-field
2015;34(3):796–806. imaging in mice. IEEE Trans Med Imaging 2019;38(2):649–656.
53. Symons R, Cork TE, Lakshmanan MN, et al. Dual-contrast agent photon- 79. Grandl S, Scherer K, Sztrókay-Gaul A, et al. Improved visualization of breast
counting computed tomography of the heart: initial experience. Int J Car- cancer features in multifocal carcinoma using phase-contrast and dark-field
diovasc Imaging 2017;33(8):1253–1261. mammography: an ex vivo study. Eur Radiol 2015;25(12):3659–3668.
54. Si-Mohamed S, Tatard-Leitman V, Laugerette A, et al. Spectral Photon- 80. Scherer K, Yaroshenko A, Bölükbas DA, et al. X-ray Dark-field Radiography -
Counting Computed Tomography (SPC CT): in-vivo single-acquisition In-Vivo Diagnosis of Lung Cancer in Mice. Sci Rep 2017;7(1):402.
multi-phase liver imaging with a dual contrast agent protocol. Sci Rep 81. Meiser J, Willner M, Schröter T, et al. Increasing the field of view in grat-
2019;9(1):8458. ing based X-ray phase contrast imaging using stitched gratings. J XRay Sci
55. Symons R, Krauss B, Sahbaee P, et al. Photon-counting CT for simultaneous Technol 2016;24(3):379–388.
imaging of multiple contrast agents in the abdomen: An in vivo study. Med 82. Marschner M, Willner M, Potdevin G, et al. Helical X-ray phase-contrast
Phys 2017;44(10):5120–5127. computed tomography without phase stepping. Sci Rep 2016;6(1):23953.

14 radiology.rsna.org  n  Radiology: Volume 00: Number 0— 2020

 Kwan et al

83. Momose A. Demonstration of phase-contrast X-ray computed tomogra- 94. Hetterich H, Webber N, Willner M, et al. Dark-field imaging in coronary
phy using an X-ray interferometer. Nucl Instrum Methods Phys Res A atherosclerosis. Eur J Radiol 2017;94:38–45.
1995;352(3):622–628. 95. Hetterich H, Webber N, Willner M, et al. AHA classification of coronary
84. Weitkamp T, David C, Kottler C, Bunk O, Pfeiffer F. Tomography with and carotid atherosclerotic plaques by grating-based phase-contrast computed
grating interferometers at low-brilliance sources. In: Bonse U, ed. Proceedings tomography. Eur Radiol 2016;26(9):3223–3233.
of SPIE: SPIE optics + photonics 2006—developments in x-ray tomography 96. Vågberg W, Persson J, Szekely L, Hertz HM. Cellular-resolution 3D virtual
V. Vol 6318. Bellingham, Wash: International Society for Optical Engineer- histology of human coronary arteries using x-ray phase tomography. Sci Rep
ing, 2006; 63180S. 2018;8(1):11014.
85. Pfeiffer F, Weitkamp T, Bunk O, David C. Phase retrieval and differen- 97. Winklhofer S, Peter S, Tischler V, et al. Diagnostic accuracy of quantitative
tial phase-contrast imaging with low-brilliance X-ray sources. Nat Phys and qualitative phase-contrast imaging for the ex vivo characterization of
2006;2(4):258–261. human coronary atherosclerotic plaques. Radiology 2015;277(1):64–72.
86. Piwnicki P. New x-ray method for Corona diagnosis ready for patient testing. 98. Reichardt M, Töpperwien M, Khan A, Alves F, Salditt T. Fiber orientation
Technical University of Munich. https://www.tum.de/nc/en/about-tum/ in a whole mouse heart reconstructed by laboratory phase-contrast micro-CT.
news/press-releases/details/36031/. Published 2000. Updated May 7, 2020. J Med Imaging (Bellingham) 2020;7(2):023501.
Accessed May 16, 2020. 99. Gonzalez-Tendero A, Zhang C, Balicevic V, et al. Whole heart detailed
87. Götzfried J, Döpp A, Gilljohann M, et al. Research towards high-repetition and quantitative anatomy, myofibre structure and vasculature from X-ray
rate laser-driven X-ray sources for imaging applications. Nucl Instrum Methods phase-contrast synchrotron radiation-based micro computed tomography.
Phys Res A 2018;909:286–289. Eur Heart J Cardiovasc Imaging 2017;18(7):732–741.
88. Gambari M, Clady R, Stolidi A, Utéza O, Sentis M, Ferré A. Exploring 100. Oda H, Roth HR, Sugino T, et al. Cardiac fiber tracking on super high-
phase contrast imaging with a laser-based K a x-ray source up to relativistic resolution CT images: a comparative study. J Med Imaging (Bellingham)
laser intensity. Sci Rep 2020;10(1):6766. 2020;7(2):026001.
89. Döpp A, Hehn L, Götzfried J, et al. Quick x-ray microtomography using a 101. Shinohara G, Morita K, Hoshino M, et al. Three Dimensional visualization of
laser-driven betatron source. Optica 2018;5(2):199–203. human cardiac conduction tissue in whole heart specimens by high-resolution
90. Pfeiffer F, David C, Bunk O, et al. Region-of-interest tomography for phase-contrast CT imaging using synchrotron radiation. World J Pediatr
grating-based X-ray differential phase-contrast imaging. Phys Rev Lett Congenit Heart Surg 2016;7(6):700–705.
2008;101(16):168101. 102. Ferraro M, Trachet B, Aslanidou L, Fehervary H, Segers P, Stergiopulos
91. Cong W, Yang J, Wang G. Differential phase-contrast interior tomography. N. Should we ignore what we cannot measure? How non-uniform stretch,
Phys Med Biol 2012;57(10):2905–2914. non-uniform wall thickness and minor side branches affect computational
92. Habbel C, Hetterich H, Willner M, et al. Ex Vivo Assessment of Coronary aortic biomechanics in mice. Ann Biomed Eng 2018;46(1):159–170.
Atherosclerotic Plaque by Grating-Based Phase-Contrast Computed To- 103. Braig E, Böhm J, Dierolf M, et al. Direct quantitative material decomposition
mography: Correlation With Optical Coherence Tomography. Invest Radiol employing grating-based X-ray phase-contrast CT. Sci Rep 2018;8(1):16394
2017;52(4):223–231. [Published correction appears in Sci Rep 2019;9(1):11076.].
93. Hetterich H, Fill S, Herzen J, et al. Grating-based X-ray phase-contrast
tomography of atherosclerotic plaque at high photon energies. Z Med Phys
2013;23(3):194–203.

Radiology: Volume 00: Number 0— 2020  n  radiology.rsna.org 15