Zurich Open Repository and

Archive
University of Zurich
University Library
Strickhofstrasse 39
CH-8057 Zurich
www.zora.uzh.ch

Year: 2016

Cinematic rendering - an alternative to volume rendering for 3D computed

tomography imaging

Dappa, Evelyn ; Higashigaito, Kai ; Fornaro, Jürgen ; Leschka, Sebastian ; Wildermuth, Simon ; Alkadhi, Hatem

DOI: https://doi.org/10.1007/s13244-016-0518-1

Posted at the Zurich Open Repository and Archive, University of Zurich

ZORA URL: https://doi.org/10.5167/uzh-126348
Journal Article
Published Version

The following work is licensed under a Creative Commons: Attribution 4.0 International (CC BY 4.0) License.

Originally published at:

Dappa, Evelyn; Higashigaito, Kai; Fornaro, Jürgen; Leschka, Sebastian; Wildermuth, Simon; Alkadhi, Hatem
(2016). Cinematic rendering - an alternative to volume rendering for 3D computed tomography imaging. Insights
into Imaging, 7(6):849-856.
DOI: https://doi.org/10.1007/s13244-016-0518-1
Insights Imaging
DOI 10.1007/s13244-016-0518-1

PICTORIAL REVIEW

Cinematic rendering – an alternative to volume rendering for 3D

computed tomography imaging
Evelyn Dappa 1 & Kai Higashigaito 1 & Jürgen Fornaro 2 & Sebastian Leschka 2 &
Simon Wildermuth 2 & Hatem Alkadhi 1

Received: 6 July 2016 / Revised: 23 August 2016 / Accepted: 29 August 2016

# The Author(s) 2016. This article is published with open access at Springerlink.com

Abstract Keywords Computed tomography . Image processing .

Volume rendering (VR) represents today’s standard three- Three-dimensional . Volume rendering . Cinematic rendering
dimensional (3-D) image post-processing technique, and often
is used to visualize complex anatomical information.
Recently, a novel 3-D technique for post-processing of com- Introduction
puted tomography (CT) image data has been introduced,
which is called cinematic rendering (CR). The objective of Current computed tomography (CT) scanners provide isotro-
this review is to illustrate the image appearance and potential pic submillimeter resolution enabling the utilization of various
value of CR in comparison with conventional VR in a number 2-D and 3-D post-processing techniques with excellent image
of various applications and different anatomical regions. quality. Whereas 2-D techniques such as planar or curved
Similar to VR, CR best visualizes high density and high con- multiplanar reformations (MPR) as well as maximum intensi-
trast structures such as bones and contrast-enhanced vessels, ty projection (MIP) images can be considered part of daily
but at the same time provides a more natural and photo- radiological routine, 3-D imaging post-processing techniques
realistic illumination of the rendered data. Further research such as surface shaded display (SSD) and volume rendering
will be necessary for determining possible advantages of CR (VR) are mainly used to convey complex anatomy. Here, 3-D
over conventional VR and over two-dimensional (2-D) image visualization of CT image data can offer important informa-
post-processing for CT image data. tion compared with 2-D axial images alone both for pre-
operative planning and post-treatment follow-up.
Teaching Points While the generation of VR images in the early days was
• Cinematic rendering is a novel post-processing technique for hampered by high computational demands and often was as-
3D visualization of CT image data. sociated with long post-processing times with considerable
• Compared to volume rendering, CR results in a more photo- user interaction, today’s powerful computers allow for a
realistic representation of anatomy. near-real time and automatic rendering of 3-D images in a fast,
• Similar to volume rendering, CR provides best image quality clinically compatible manner. Thus, VR is today used as the
of high density structures. standard technique for the 3-D visualization of CT image data.
The main advantages of VR lie in a relatively simple evalua-
tion of different anatomical structures of a larger body region
at the same time. Moreover, VR enables the display of CT
* Hatem Alkadhi
hatem.alkadhi@usz.ch
image data in a coloured fashion [1, 2].
However, the VR technique also suffers from some disad-
1
Institute of Diagnostic and Interventional Radiology, University
vantages. While VR can be helpful in evaluating complex
Hospital Zurich, University of Zurich, Raemistrasse 100, CH- anatomy, it can also mask anatomical information and, hence,
8091 Zurich, Switzerland pathology. This has been recently illustrated for the visualiza-
2
Division of Radiology and Nuclear Medicine, Kantonsspital, St tion of intracranial aneurysms with VR, where 3-D CT angi-
Gallen, Switzerland ography with bone subtraction provided a significantly higher
 Insights Imaging

Fig. 1 Contrast-enhanced 3-D images of a malignant type coronary side. The volume rendered (VR) image (b) depicts the same anomaly
anomaly with origin of the right coronary artery (RCA) from the left from a similar perspective. Note that vessel contour is smoother on CR
coronary sinus and interarterial course (arrows). Cinematic rendering compared to VR images
(CR) (a) shows the RCA coursing ventrally to the aortic root to the right

diagnostic accuracy compared to 3-D CT angiography without knowledge—no publication describing and discussing the
bone subtraction [3]. Thus, in order to avoid missing critical new technique in more detail.
information, VR images should always be correlated and Thus, the purpose of this pictorial review is to illustrate the
compared with corresponding MPR images. potential value of CR in comparison with conventional VR in
Finally, it should be kept in mind that visualization of im- various applications and body regions. For doing so, we (i)
aging data with VR – as with all other post-processing tech- summarized the technical background and potential clinical
niques such as MIP or SSD—never adds information to the applications of the conventional VR technique, (ii) introduced
reader beyond what is present in the original source images. It the technical concept of CR, and (iii) selected a number of
always remains an option for a different representation of im- common and some rare pathology and disease for visualizing
aging data, which might be helpful for anatomically complex the potential of the new 3-D post-processing technique for CT
structures and disease, and for easier and comprehensible il- image post-processing.
lustration of imaging findings to clinicians.
Recently, a new technique for 3-D visualization of cross-
sectional image data from CT has been introduced, which is Volume rendering – technical background
entitled cinematic rendering (CR). CR works with random
sampling computational algorithms and uses different light The technical background of the VR technique has been re-
maps to generate a realistic depiction of medical data. While petitively explained in detail in a number of previous publica-
first studies have started using the CR technique for visualiz- tions [2, 6]. VR represents a computer algorithm used to trans-
ing anatomy from CT data [4, 5], there is so far—to our form cross-sectional image datasets (from, e.g., CT or

Fig. 2 Contrast-enhanced 3-D

CT images of a basilar artery
aneurysm. CR (a) and
conventional VR (b) depict the
base of the skull with the circle of
Willis, as well as the fusiform
dilation of the basilar artery. Note
the more natural representation of
the bones and vessels in the CR
image
Insights Imaging

Fig. 3 Contrast-enhanced 3-D

CT images after endovascular
stent graft treatment of the
abdominal aorta with visceral
debranching in a patient with
infrarenal aortic and splenic artery
aneurysm, shown with CR (a) and
with VR (b)

magnetic resonance (MR) imaging) into 3-D images. The VR modulates the colour of the light depending on the assigned
technique consists of the following two steps: classification of colour and transparency and contributes to the final projection
each voxel and image projection. Classification determines and final image.
how each point on the artificial rays that pass through the data Finally, 3-D perception in the projected image can be en-
contributes to the pixel value on the picture [7, 8]. For sepa- hanced by implementing additional effects, such as reflections
ration of different tissues (e.g., bone, soft tissue), a trapezoid is and shadows on the surface of the rendered image. The final
used for each tissue type. The tissue represented in each voxel image can be manipulated by increasing or decreasing the
of the volume dataset is determined by using predefined at- slope of the trapezoid, which can be observed nearly in real
tenuation threshold levels and is assigned to a specific colour time [8]. When sliding the trapezoid towards lower
and opacity. Then, the weighted sum of the percentage of each Hounsfield unit values, more structures with lower attenuation
tissue type represented in the voxel is calculated to determine will be included (e.g., smaller vessels), and vice versa.
the overall colour and transparency of each voxel. This step is
performed for each voxel and for the whole volume dataset.
Then, the 3-D volume is displayed by using a projection Volume rendering – potential clinical applications
technique. This is done by simulating rays of light, which are
projected through the 3-D volume that contains the classified Complex anatomy Useful clinical applications of the VR
voxels. Each voxel, which is passed by the simulated rays, technique for visualizing complex anatomy and pathology

Fig. 4 Contrast-enhanced 3-D

CT images of an arteriovenous
malformation (AVM) in the upper
pole of the left kidney. CR (a) and
VR (b) images and close-up view
(c, d) images allow for a more
detailed depiction of the AVM.
The tortuous vessels at the upper
kidney pole are visualized with a
more realistic representation on
CR than on VR images
 Insights Imaging

Fig. 5 Contrast-enhanced 3-D

CT images of a cervical spine
injury with luxation of C3/4. Both
CR (a) and VR (b) images show
the anterior subluxation of the
cervical vertebrae (arrows). The
enhanced lightning of CR creates
a high contrast between vessels
and bone, making it easy to
follow the course of the vertebral
artery

have been described in many radiological fields and include application of the VR technique is the visualization of
otorhinolaryngiology [9], neurosurgery and cranio-facial sur- the complex post-operative anatomy after coronary ar-
gery [10], female imaging [1], vascular anomalies and variants tery bypass graft surgery [23]. Here, the variable course
[11, 12], and coronary artery anatomy and anomalies [13–15]. of bypass grafts often crossing the axial image plane
several times in different directions might be difficult
Pre-operative planning In trauma patients, VR yields infor- to appreciate using 2-D post-processing tools alone. In
mation about fragment size and position in relation to adjacent regard to arteries and veins, VR provides visual infor-
structures [16–19], and can be used as a preoperative and pre- mation, which normally is obtained with conventional
radiotherapy planning tool leading to high satisfaction and catheter angiography. Here, VR is not only noninvasive,
acceptability from referring surgeons [20]. Also, imaging in but also fast and cost-effective [6].
oncology may profit from the application of the 3-D VR tech-
nique. VR can help evaluate tumour origin in difficult ana-
tomical settings such as, e.g., in the pelvis [1], and can be used Cinematic rendering – technical background
to determine tumour resectability and for preoperative plan-
ning [6]. The development of CR was inspired by the entertainment
industry based on the quality of computer animation programs
Post-operative setting The VR technique can also pro- in cinema, and had the aim to generate more photo-realistic
vide useful information in patients after surgery, such representations of the human body from CT and MR image
as, for example, for visualizing the results from data sets.
endovascular aortic repair and the relationship between CR is currently available as a research tool in an open
stent grafts and arterial vessels [21, 22]. An additional research environment containing various prototype software

Fig. 6 Non-enhanced 3-D CT

images of multiple traumatic
fractures of the spine and pelvis.
Both CR (a) and VR (b)
demonstrate multiple fractures of
the ribs, lumbal transverse
processes, sacrum, and pubic
bones
Insights Imaging

Fig. 7 Non-enhanced 3-D CT

images of a head shot injury. Both
CR (a) and VR (b) show the
bullets and extensive damage of
the skull base. The more photo-
realistic representation of
anatomy with CR allows for an
easy identification of the complex
fracture course

(Frontier, version 1.0.0, syngo.Via, version VB10A, Siemens representing radiance at random positions with light scattered
Healthcare, Forchheim, Germany) and has so far no approval in random directions.
for clinical use. High dynamic range (HDR) rendering light maps are used
The CR technique [24] introduces a new paradigm, enabled for illumination, which lead to a natural illumination of the
by recent advances in computer graphics, to render volumetric rendered data, in contrast to synthetic light sources used in
medical image data by using a physically based real-time conventional rendering methods such as VR [27].
technique [25]. Conventional VR, such as ray casting [26] As a result, the physically based VR method called CR
only considers emission and absorption of energy along a light computes in real-time the complex physics of lighting effects.
ray to calculate 3-D images. Scattering effects are modelled It models shadows, ambient occlusion, multi-scattering, and
using a local gradient shading model. Although simple to color transmittance as well as sophisticated camera properties,
compute using a Riemann integral, such conventional ap- which include concepts such as aperture, exposure and shutter
proaches neglect complex light paths with multiple scattering speed. This approach leads to a natural and physically accurate
patterns as well as light extinction, which might lead to a less presentation of the medical data, with a focus on an enhanced
artificial and potentially less accurate VR images. depth and shape perception [28].
As opposed to conventional VR methods, CR solves the The handling of transparency in CR does not differ from
multi-dimensional and non-continuous rendering equation the conventional VR technique. The transparency is computed
[24] to integrate the light scattered from all possible directions based on the transfer-function, which assigns colour and opac-
along a ray. Thus, path tracing used in CT integrates a huge ity to each attenuation value. Potential differences in transpar-
number of light rays, each with different paths to form each ency approaches can be easily adjusted by fine tuning of the
pixel of the rendered image. Since the number of light paths transfer-function.
which can be traced is in theory infinite, and tracing of light Representative image examples of CR in comparison with
paths is computationally expensive, Monte-Carlo simulations VR are provided in Figs. 1, 2, 3, and 4 for vascular anatomy
are used to generate a randomized subset of light paths with an and pathology, in Figs. 5, 6, and 7 for bones and skeletal
adequate distribution. The final image is obtained iteratively disease, and in Figs. 8 and 9 for visualization of soft tissue
by progressively averaging numerous Monte Carlo samples anatomy. All CT source images were reconstructed with our

Fig. 8 Non-enhanced 3-D CT

images of traumatic herniation of
small bowel through the anterior
abdominal wall. CR (a) and VR
(b) show the soft tissues of the
abdominal wall and upper thigh
along with herniating of the
bowel loops (arrows). CR allows
for a natural representation of
anatomy
 Insights Imaging

Fig. 9 Contrast-enhanced 3-D

CT images of a patient with
polytrauma. CR (a) and VR (b)
visualize the soft tissues of the
head, neck, chest and abdomen.
Note the differences in photo-
realistic representation of 3-D
anatomy in CR as compared to
VR images

default settings using a slice thickness of 2 mm (increment and soft tissue structures providing a more photo-realistic de-
1.5 mm), except for coronary CT angiography, which was piction of human anatomy and disease.
reconstructed with a slice thickness of 0.75 mm (increment In general, lighting in 3-D rendered images such as VR and
0.5 mm). All CT angiography data were acquired after the CR is a function of ambient, diffuse, and specular light prop-
intravenous injection of 50–80 ml non-ionic, iodinated con- erties. In both techniques, the brightness of each voxel is de-
trast media (concentration 300–370 mg J/ml, depending on fined by the distribution of these light properties, which results
the body region). in different lighting of different body parts relative to the arti-
ficial light source introduced into the volume, giving rise to 3-
D impressions of the images. In VR, the differences in light
emitted to the voxels are rather small. In contrast, CR uses a
Discussion more complex lighting model taking into account the effect of
lighting for other voxels and subsequent reflections as well.
Both, VR and CR have the same general rendering concept in Also, the effect of body parts blocking the trace from the
common: segmentation of data based on voxel attenuation and artificial light source to other structures introduces shadowing
use of colour look-up tables taking into account opacity and into the images. As a result of the differences in lighting func-
brightness. As a consequence, both VR and CR share also the tions—as being illustrated in the representative image exam-
same problem of potentially masking findings in datasets by ples of this pictorial review—CR images go along with a more
inappropriate use of rendering parameters and by adjacent natural image impression as compared to conventional VR.
structures [3]. Albeit CR currently is only provided by a single Despite the potential benefits of CR compared to VR in the
vendor in a pre-clinical research tool with still limited display visualization of volume datasets, there is a higher computa-
options, CR can provide the same advanced display functions tional power demand required for the CR technique because
such as VR including flying-through, flying-around and mul- of the more complex lighting model. Therefore, real-time dis-
tiple views of the volume data with independent parameters in play of, for instance, rotating the CR image is currently
equal segments of the displayed window [2, 6]. interrupted by repetitive recalculation processes. Rendering
Due to the abovementioned similarities between CR and of the final image needs some seconds, ranging from 5 to
VR, there are no major differences in the visualization of CT 30 s depending on the quality of the resulting image [28].
image data when regarding the diagnostic value of the presen- One major limitation of this pictorial review is that we
tation. Similar to VR, we found that particularly high-contrast focused our comparison of VR and CR on some clinical cases
structures such as contrast-enhanced vessels and bones can be to highlight the differences in image appearance between both
depicted with high quality also with CR. At the same time, we post-processing methods. However, we did not perform an
found a major improvement of CR for the perception of depth analysis of both methods regarding the diagnostic
Insights Imaging

performance for the diagnosis of specific pathological entities. evaluation of individualized low-volume contrast media protocols.
Radiology 280(3):960–968
Here, future studies should determine whether or not there is
5. Higashigaito K, Husarik DB, Barthelmes J, Plass AR, Manka R,
an improvement of the new 3-D imaging technique described Maisano F et al (2016) Computed tomography angiography of cor-
herein. onary artery bypass grafts: low contrast media volume protocols
adapted to tube voltage. Invest Radiol 51(4):241–8
6. Calhoun PS, Kuszyk BS, Heath DG, Carley JC, Fishman EK
(1999) Three-dimensional volume rendering of spiral CT data: the-
Conclusion ory and method. Radiog : Rev Publ Radiol Soc North Am, Inc
19(3):745–764
This pictorial review aimed at a first, preliminary demonstra- 7. Luccichenti G, Cademartiri F, Pezzella FR, Runza G, Belgrano M,
tion of a new post-processing technique for the 3-D visualiza- Midiri M et al (2005) 3D reconstruction techniques made easy:
know-how and pictures. Eur Radiol 15(10):2146–2156
tion of CT image data. Our initial experience indicates that CR 8. Lell MM, Anders K, Uder M, Klotz E, Ditt H, Vega-Higuera F et al
of CT images is particularly impressive when high density and (2006) New techniques in CT angiography. Radiog : Rev Publ
high contrast structures such as bones and contrast-enhanced Radiol Soc North Am, Inc 26(Suppl 1):S45–S62
vessels are to be visualized. The main innovation as compared 9. Guo Y, Liu Y, Lu QH, Zheng KH, Shi LJ, Wang QJ (2013) CT two-
dimensional reformation versus three-dimensional volume render-
to conventional VR appear to be the more natural and photo-
ing with regard to surgical findings in the preoperative assessment
realistic representation of the CT image data, with an en- of the ossicular chain in chronic suppurative otitis media. Eur J
hanced and more natural depth and shape perception. Future Radiol 82(9):1519–1524
studies should aim at determining whether the CR technique 10. Sun W, Jia L, Dong Y, Zhao H, Liu H, Yang K et al (2014)
offers advantages over cross-sectional and 3-D CT images in Morphological study of surgical approach by superior temporal
sulcus-temporal horn of lateral ventricle approach using volume
terms of a diagnostic gain. rendering. J Craniofacial Surg 25(2):611–613
11. Li HO, Huo R, Wang XM, Xu GQ, Duan YH, Nie P et al (2015)
Acknowledgments We thank Dr. Razvan Ionasec and Dr. Max High-pitch spiral CT with 3D reformation: an alternative choice for
Schoebinger (both Siemens Healthineers) for their valuable input regard- imaging vascular anomalies with affluent blood flow in the head
ing the technical background of cinematic rendering. and neck of infants and children. Br J Radiol 88(1052):20150005
12. Abdullah A, Williamson K, Lewis T, Elsamaloty H (2011) Variant
ventral intrahepatic course of inferior vena cava: volume-rendering
Compliance with ethical standards and maximum intensity projection CT findings. Br J Radiol
84(1003):e135–e137
Conflict of interest None. 13. Mlynarski R, Mlynarska A, Sosnowski M (2013) Anatomical var-
iants of left circumflex artery, coronary sinus and mitral valve can
determine safety of percutaneous mitral annuloplasty. Cardiol J
Funding None. 20(3):235–240
14. Graidis C, Dimitriadis D, Karasavvidis V, Dimitriadis G,
Argyropoulou E, Economou F et al (2015) Prevalence and charac-
Open Access This article is distributed under the terms of the Creative
teristics of coronary artery anomalies in an adult population under-
Commons Attribution 4.0 International License (http://
going multidetector-row computed tomography for the evaluation
creativecommons.org/licenses/by/4.0/), which permits unrestricted use,
of coronary artery disease. BMC Cardiovasc Disord 15:112
distribution, and reproduction in any medium, provided you give appro-
15. Karlo CA, Leschka S, Stolzmann P, Glaser-Gallion N, Wildermuth
priate credit to the original author(s) and the source, provide a link to the
S, Alkadhi H (2012) A systematic approach for analysis, interpre-
Creative Commons license, and indicate if changes were made.
tation, and reporting of coronary CTA studies. Insights Imag 3(3):
215–228
16. Falchi M, Rollandi GA (2004) CT of pelvic fractures. Eur J Radiol
50(1):96–105
References 17. Song SW, Jun BC, Chae SR, Kim BG (2013) Clinical utility of
three-dimensional facial computed tomography in the treatment of
nasal bone fractures: a new modality involving an air-bone view
1. Li Y, Zheng Y, Chen J, Chen X, Lin J, Cai A et al (2015) with a volume rendering technique. Indian J Otolaryngol Head
Determining the organ of origin of large pelvic masses in females Neck Surg : Off Publ Assoc| Otolaryngol India 65(Suppl 2):210–
using multidetector CT angiography and three-dimensional volume 215
rendering CT angiography. Eur Radiol 25(4):1032–1039 18. Schmidle G, Rieger M, Klauser AS, Thauerer M, Hoermann R,
2. Fishman EK, Ney DR, Heath DG, Corl FM, Horton KM, Johnson Gabl M (2014) Intraosseous rotation of the scaphoid: assessment
PT (2006) Volume rendering versus maximum intensity projection by using a 3D CT model–an anatomic study. Eur Radiol 24(6):
in CT angiography: what works best, when, and why. Radiog : Rev 1357–1365
Publ Radiol Soc North Am, Inc 26(3):905–922 19. Mirka H, Ferda J, Baxa J (2012) Multidetector computed tomogra-
3. Hwang SB, Kwak HS, Han YM, Chung GH (2011) Detection of phy of chest trauma: indications, technique and interpretation.
intracranial aneurysms using three-dimensional multidetector-row Insights Imag 3(5):433–449
CT angiography: is bone subtraction necessary? Eur J Radiol 79(2): 20. Wang H, Wang F, Newman S, Lin Y, Chen X, Xu L et al. (2016)
e18–e23 Application of an innovative computerized virtual planning system
4. Higashigaito K, Schmid T, Puippe G, Morsbach F, Lachat M, in acetabular fracture surgery: a feasibility study. Injury 47(8):
Seifert B, et al. (2016) CT Angiography of the aorta: prospective 1698–1701
 Insights Imaging

21. Sun Z (2007) Multislice CT, angiography in abdominal aortic an- medicine. Medical Image Analysis, Elsevier. 2016; 20th
eurysm treated with endovascular stent grafts: evaluation of 2D and Anniversary Special Issue
3D visualisations. Biomed Imag Interven J 3(4):e20 25. Kroes T, Post FH, Botha CP (2012) Exposure render: an interactive
22. Iezzi R, Santoro M, Dattesi R, Pirro F, Nestola M, Spigonardo F photo-realistic volume rendering framework. PLoS One 7(7):
et al (2012) Multi-detector CT angiographic imaging in the follow- e38586
up of patients after endovascular abdominal aortic aneurysm repair 26. Engel K, Hadwiger M, Kniss J, Rezk-Salama C, Weiskopf D.
(EVAR). Insights Imag 3(4):313–321 (2006) Real-time volume graphics: AK Peters
23. Frazier AA, Qureshi F, Read KM, Gilkeson RC, Poston RS, 27. Debevec ZH (1998) Rendering synthetic objects into real scenes:
White CS (2005) Coronary artery bypass grafts: assessment bridging traditional and image-based graphics with global illumina-
with multidetector CT in the early and late postoperative tion and high dynamic range photography. Proc SIGGRAPH ’98:
settings. Radiog : Rev Publ Radiol Soc North Am, Inc 189-98.
25(4):881–896 28. Fellner FA (2016) Introducing cinematic rendering: a novel tech-
24. Comaniciu D, Engel K, Georgescu B, Mansi T. (2016) Shaping the nique for post-processing medical imaging data. J Biomed Sci Eng
FUTURE through innovations: from medical imaging to precision 9:170–175