I

LD

a

ARA

KRPsI

1

ftiafthssgbio3td

1s

it

0d

European Journal of Radiology 78 (2011) 210–224

Contents lists available at ScienceDirect

European Journal of Radiology

journa l homepage: www.e lsev ier .com/ locate /e j rad

ntegrating image processing in PACS

orenzo Faggioni, Emanuele Neri ∗, Francesca Cerri, Francesca Turini, Carlo Bartolozziiagnostic and Interventional Radiology, University of Pisa, Via Paradisa, 2, 56100 Pisa, Italy

r t i c l e i n f o

rticle history:eceived 21 June 2009ccepted 22 June 2009

eywords:adiology information system

a b s t r a c t

Integration of RIS and PACS services into a single solution has become a widespread reality in daily radio-logical practice, allowing substantial acceleration of workflow with greater ease of work compared witholder generation film-based radiological activity. In particular, the fast and spectacular recent evolutionof digital radiology (with special reference to cross-sectional imaging modalities, such as CT and MRI)has been paralleled by the development of integrated RIS–PACS systems with advanced image processing

icture archiving and communicationystemmage processing

tools (either two- and/or three-dimensional) that were an exclusive task of costly dedicated workstationsuntil a few years ago. This new scenario is likely to further improve productivity in the radiology depart-ment with reduction of the time needed for image interpretation and reporting, as well as to cut costs forthe purchase of dedicated standalone image processing workstations. In this paper, a general descriptionof typical integrated RIS–PACS architecture with image processing capabilities will be provided, and themain available image processing tools will be illustrated.

. Introduction

The increasing availability of integrated RIS–PACS solutionsor image reading and reporting of medical imaging examina-ions has fuelled the development of integrated tools for digitalmage processing on RIS–PACS systems. This evolution representsfurther step towards total integration of all instruments needed

or interpretation and reporting of diagnostic imaging examina-ions in a health service environment, and technical problemsave been faced, especially in early times, due to the relativelytringent hardware and software requirements for accomplishinguch task. In the following sections, a general description of inte-rated RIS–PACS systems with image processing capabilities wille provided, and the main available image processing tools will be

llustrated. Examples of the various image processing options werebtained by means of the open-source software OsiriX (version.5.1, http://www.osirix-viewer.com) integrated with our Institu-ion PACS network (Fig. 1), and are illustrated in the followingedicated sections.

.1. From dedicated workstations to server-based RIS–PACSolutions

As known, a strong impulse toward evolution and diffusion ofmage processing tools has been represented, on one side, by theransition from analogue to digital image acquisition. In partic-

∗ Corresponding author. Tel.: +39 050997313, fax: +39 050997313.E-mail address: neri@med.unipi.it (E. Neri).

720-048X/$ – see front matter © 2009 Elsevier Ireland Ltd. All rights reserved.oi:10.1016/j.ejrad.2009.06.022

© 2009 Elsevier Ireland Ltd. All rights reserved.

ular, the evolution of digital imaging (for instance, conventionalradiography) into a filmless direction has introduced the need forimage visualisation software in order for images to be evaluated[1,2]. In addition, the possibility for the technician and the report-ing radiologist to modify image settings in the post-processingphase as to emphasise peculiar image features (that might beunderestimated or missed under standard display settings, therebypotentially compromising diagnosis), has made image condition-ing functions on visualisation stations a must-have. This is thecase for elementary tools for adjusting window levels (as pre-viously available on consoles of computed tomography [CT] ormagnetic resonance imaging [MRI] scanners) or applying filters,such as high frequency filters for magnification of bone details orlow frequency filters for noise reduction. Such scenario has beendecisive for the evolution of PACS systems and their integrationwith RIS stations. Another important point is that data transmit-ted to the integrated RIS–PACS station must contain the wholeimage information in order to allow accurate image processing, andare to be shared over a universal digital platform. This has lead toadoption of the DICOM format as a universal standard for medi-cal image data exchange; however, the necessity to preserve dataintegrity of medical images with a high degree of accuracy makesDICOM files relatively large, thus prompting development of fasterinternet and intranet connections for reasonably quick data flow[3,4].

The problems related to the high amount of generated data have

been boosted by the concurrent evolution of cross-sectional imag-ing techniques, such as CT and MRI. In particular, the advent ofmultidetector CT (MDCT) scanners with an increasing number ofdetector rows has urged the search for new technical solutions

L. Faggioni et al. / European Journal of Radiology 78 (2011) 210–224 211

from

ernigreswefgipoIs

Fig. 1. OsiriX integrated RIS–PACS interface: query list for retrieval

nabling efficient RIS–PACS integration, both in terms of imageeading and reporting times [5,6]. In fact, modern MDCT scan-ers with 64 detector rows and beyond can produce thousands of

mages, each of them having submillimetric thickness in the lon-itudinal axis and a relatively high noise index in order to keepadiation dose down to acceptable levels. This would make analyticvaluation of every single image (as traditionally done with singlelice CT examinations) extremely time consuming – incompatiblyith the routine workflow of a diagnostic imaging centre – and

rror-prone due to the high image noise and increasing reader’satigue. In addition, the advancements of MDCT technology haveained to this latter application fields once pertaining to othermaging modalities: this is the case, for example, of CT colonogra-

hy, CT angiography (with particular reference to CT angiographyf peripheral vessels and CT coronography), and trauma evaluation.n this respect, if such transition to MDCT has, on one hand, broughteveral diagnostic advantages, on the other hand it has implied a

PACS (a) and image list for a given PACS-retrieved examination (b).

further multiplication of the amount of generated MDCT images.The key point is twofold:

(1) many images need a large amount of space to be loaded, bothin terms of hard disk and random access memory (RAM) space;

(2) post-processing of such large datasets – both in 2D, and espe-cially, in 3D mode – requires vast calculating power, both from CPUand graphics card.

Putting things into perspective, for large MDCT datasets tobe processed, a relatively enormous amount of power would beneeded, far superior to that of common RIS–PACS systems, andcomparable to that afforded by dedicated image processing work-stations. Indeed, the former are usually cheap, medium-to-lowpower computers designed for word processing (the RIS part) and

visualisation of digital images retrieved online from the local areanetwork (the PACS part), while high-end image processing con-soles are much more expensive and usually not designed for RISintegration.

212 L. Faggioni et al. / European Journal of Radiology 78 (2011) 210–224

Fig. 2. By clicking onto one of the buttons encircled in red (a), it is possible to instantly create orthogonal MPR views of a stack of cross-sectional images (b), in this example,MDCT images of a cyst of the mesenteric root (yellow asterisk) are reformatted on the three orthogonal planes. (c) Buttons are provided (red circle) that allow export of aspecified set of MPR images in DICOM format as a separate image series.

rnal o

fuwspRaswcfaop

2a

2

2

iMspaaffittmts

Fs

L. Faggioni et al. / European Jou

A widely adopted solution to this problem, permitting to per-orm heavy-duty image processing on RIS–PACS terminals withoutpgrading hardware, has been represented by accessing via net-ork post-processing tools that are physically installed on remote

ervers. More precisely, post-processing algorithms are run onowerful remote servers connected through the PACS to everyIS–PACS station, which therefore acts as a client computer. Thisllows to leave the whole image processing workload onto theerver, which receives input commands from RIS–PACS computers,hile in turn, the server sends the output in real time to RIS–PACS

lients. The output can be saved either in DICOM or other imageormat (such as JPEG, TIFF, PNG, etc.) on RIS–PACS clients or ton external user device if allowed by the administrator settingsf the client system, or even sent to the PACS in DICOM format forermanent online storage.

. Image processing tools on integrated RIS–PACS systems:n overview

.1. 2D post-processing tools

.1.1. Multiplanar reformat (MPR)As known, the MPR technique allows to create a bidimensional

mage on an arbitrary plane from a dataset of complanar images.ore precisely, the information about position and signal inten-

ity of voxels belonging to a stack of images acquired on parallellanes with a given spatial orientation (for example, a series ofxial images) is used to produce an output image where voxelsre geometrically projected on a user-defined plane with a dif-erent inclination (e.g. a coronal, sagittal or oblique plane), as aunction of position and signal intensity of voxels from sourcemages [7,8]. In order to achieve good quality MPR reconstruc-ions, it is essential that image spatial resolution is high enough

o approach voxel isotropy [9], so that no substantial loss of infor-

ation occurs when image data are displayed on planes other thanhe native one. In this sense, the widespread availability of MDCTcanners and the progressive diffusion of high-field MRI equipment

ig. 3. CPR of renal arteries (lower image) can be performed by manually tracing the vesselource images (upper image).

f Radiology 78 (2011) 210–224 213

(especially for vascular and neuroimaging applications) has beenvery beneficial to the generation of excellent MPR reconstructionswith submillimetric voxel size. The MPR algorithm has the advan-tage of preserving the whole contrast resolution of native images,thus allowing reliable measurement of tissue density and contrastenhancement; indeed, on most integrated RIS–PACS systems it ispossible to take density measurements on regions of interest (ROI)from MPR images, as well as geometric measurements, such as dis-tances, areas, or angles. Moreover, almost all current integratedRIS–PACS systems with image processing options allow instantMPR of CT and MRI datasets; in addition, the majority of themallow to automatically perform MPR on standard planes (axial,coronal, sagittal: orthogonal MPR) by generating stacks of refor-matted images spanning from a given initial to a given final limit ofthe source dataset, that can be saved as independent image series inDICOM format or even as a movie (Fig. 2). It is also possible to viewcurrent and previous PACS-retrieved studies of the same patient inMPR mode with image synchronisation, allowing for more accu-rate cross-study image comparison. More generally, the user isfree to define oblique reconstruction planes arbitrarily by orientingand moving an electronic caliper on the three standard orthogonalplanes, on which structures (e.g. blood vessels) are identified basedon their anatomical course. In this way it is possible, for example,to create images on planes perpendicular to the longitudinal axisof a vessel in order to measure its diameter or section area, therebyavoiding the measurement error due to the geometric distortionthat occurs whenever such measures are taken on native imagesfrom vessels coursing non-perpendicularly to the acquisition plane.

A variant of the MPR algorithm is CPR (Curved Planar Refor-mation), that allows to project native images from a succession ofplanes lying onto a user-defined polygonal trajectory [8]. CPR isparticularly useful, for example, to display vessels with a tortuouscourse in the three spatial directions, such as coronary arteries. This

operation can be carried out by selecting a series of points alignedalong the centre of the lumen (centreline) on one or more MPR pro-jections, thus obtaining a continuous rectified representation of thevessel course (Fig. 3). With CPR it may also be easier to perform

centreline (green dotted line) on MIP-reformatted magnetic resonance angiography

214 L. Faggioni et al. / European Journal of Radiology 78 (2011) 210–224

Fig. 4. Usage of slice averaging allows to reduce image noise on MPR images (a, source images; b, orthogonal MPR with slice averaging).

Fig. 5. MIP of the left renal artery (right image) can be performed through the MPR interface by selecting slab thickness and orientation (blue parallel lines) on MPR views(upper and lower left images).

L. Faggioni et al. / European Journal of Radiology 78 (2011) 210–224 215

f bron

mcstt

i[ewr

Fig. 6. minIP can be effective as to enhance visualisation o

easurements of vessel size, as the selected vessel is displayedontinuously without distortions of its lumen contour. Options areometimes offered that allow automatic or semi-automatic detec-ion of the centreline, as described more thoroughly in Section 3,hus simplifying and/or quickening the radiologist’s work.

Alongside ‘pure’ MPR viewing, it is possible to modify visual-sation settings of MPR views by applying slice averaging tools

8] (Fig. 4), which are helpful for reducing image noise (thereforenabling radiation dose reduction in MDCT data acquisition), asell as effective image number, thus accelerating the whole image

eading process. Of course, the degree of averaging must be set at

chial fistula (a, source MDCT axial images; b, MPR views).

a level that allows relatively expedite image evaluation withoutmissing findings of diagnostic importance due to partial volumeeffect; as a rule of thumb, at our Institution an average reconstruc-tion thickness of 2–3 mm is considered to be a good compromise,given a native slice thickness of 0.625 mm for source MDCT axialimages.

2.1.2. Maximum Intensity Projection (MIP) and MinimumIntensity Projection (minIP)

Besides slice averaging, MPR-processed data can be displayedby using the Maximum Intensity Projection (MIP) or the Minimum

216 L. Faggioni et al. / European Journal of Radiology 78 (2011) 210–224

F obtain MIP view of CPR of the superior mesenteric artery (right): see accurate depictiono ascular territories.

IwgasSeitstaawttabnaMpesMMa(

2

epSpawl

ig. 7. By tracing the vessel centreline on sagittal MIP image (left), it is possible tof the artery and of its branches without substantial overlap of vessels from other v

ntensity Projection (minIP) visualisation algorithms. As known,ith the MIP technique a slab with a defined thickness and a

iven spatial orientation is selected, and for each set of voxelsligned along the chosen direction, only that with the highest inten-ity is represented on the resulting reconstructed image (Fig. 5).imilarly, with the minIP algorithm only the voxel with the low-st intensity is shown on the final image among those includedn the slab [8] (Fig. 6). As a consequence, the MIP and minIPechniques are of primary importance for displaying anatomicaltructures with homogenous intensity that are not comprised intoheir morphological continuity on a single plane, such as bonesnd contrast-enhanced blood vessels and urinary system with MIP,nd air-containing structures (e.g. airways or the large bowel)ith minIP. As for MPR with slice averaging, it is possible for

he user to select slab width and orientation, in order to adjusthe degree of voxel superimposition. Some software packages alsollow to perform MIP or minIP onto curved MPR views, which cane beneficial e.g. to get a panoramical view of branch vessels origi-ating from a main trunk with a tortuous course (Fig. 7). Moreover,s unwanted anatomical structures may overlap with thick-slabIP/minIP, some integrated RIS/PACS solutions offer the option of

re-editing the original dataset as to remove undesired structures,ither through intensity threshold- and/or region growing-basedegmentation algorithms, or by manual cutting on native and/orPR images [10,11]. Other applications allow direct editing onIP/minIP views, while others offer editing options on 3D views

nd permit MIP/minIP on datasets previously modified in 3D modeFig. 8).

.2. 3D post-processing tools

The server-based approach for remote data handling has beenspecially beneficial as for the possibility to perform 3D imagerocessing on RIS–PACS stations, such as Volume Rendering (VR),haded Surface Display (SSD), and Virtual Endoscopy (VE). This is

articularly true for VR algorithms, that require a relatively largemount of computational power as their output must retain thehole information content of the input dataset, in contrast with

ess hardware-demanding 2D tools (for instance, MIP uses approx-

Fig. 8. Coronal thick-slab MIP view generated after removal of the aorta and renalarteries allows optimal depiction of the hepatic arterial vasculature and the superiormesenteric artery system. In this case of a patient candidate to liver transplantation,an accessory artery to the right hepatic lobe stems from the superior mesentericartery (red arrow).

L. Faggioni et al. / European Journal of Radiology 78 (2011) 210–224 217

Fig. 9. By choosing different transfer curves for VR, the user can emphasise different features of the image dataset: in this case of Leriche syndrome, a logarithmic inversecurve with a relatively low intensity threshold may be used to display collateral vessels (a), while one with a higher threshold allows better depiction of the aorto-iliacaxes and splanchnic arteries (b). A linear curve with a very low threshold would obscure vessels but enable visualisation of soft tissues, such as the skin (c). In addition, thepossibility for the user to create customised transfer curves is included (d).

218 L. Faggioni et al. / European Journal of Radiology 78 (2011) 210–224

Fig. 9. (Continued ).

L. Faggioni et al. / European Journal of Radiology 78 (2011) 210–224 219

Fig. 10. In this case of sacral metastasis from thyroid tumour vascularised from branches of the right internal iliac artery (a: MIP semi-axial image), it is possible to usedifferent CLUT as to obtain optimal VR depiction of the relationship between the lesion and the host bone, together with its feeding vessels (b and c).

220 L. Faggioni et al. / European Journal of Radiology 78 (2011) 210–224

F ogramo

ihctm

2

i

ig. 11. VR with manual sculpting of patient table (a) and abdominal VR MDCT angif Riolan were manually cropped for better image clarity.

mately 10% of the original information, because only voxels withighest intensity are represented on the output view). This way, aomplete set of 2D and 3D options integrated on a RIS–PACS sta-ion is available to the user, virtually eliminating the need for him to

ove onto high-end standalone workstations for image processing.

.2.1. Volume Rendering (VR)As known, while MIP is based on the extraction of maximum

ntensity voxels inside a user-defined slab among those crossed by

obtained after automatic bone removal (b); in this latter case, vessels of the arcade

an array of parallel lines, VR algorithms harness the entire spa-tial and intensity information of the dataset as to represent onthe reconstructed image a weighted average of the intensity ofall voxels included in the slab. Each voxel is assigned a value ofopacity, transparency, and colour depending on its intensity, posi-

tion, and the prospective direction in which the image volumeis observed according to a predetermined transfer function, thatallows to generate the final image by weighing the intensity ofthe various voxels encountered by each ray. VR plug-ins let the

L. Faggioni et al. / European Journal of Radiology 78 (2011) 210–224 221

Fig. 12. (a) SSD view of lower limb arteries in patient with obstruction of right superficial femoral artery and aneurysm of the left superficial femoral artery; the dialogue boxf owera

utovdts(ptrecR

or colour and threshold settings is visible on the left half of the screen. (b) MIP of lre shown as a reference.

ser define his own transfer function by choosing a colour and aransparency value and setting at least two voxel intensity thresh-lds, corresponding to the minimum and maximum luminance ofisualisable points on the reconstructed image. So, it is possible toefine linear transfer functions (implying that the ratio betweenhe intensity of two voxels on native images and those of the corre-ponding pixels on the VR image is constant), as well as nonlineare.g. logarithmic or logarithmic inverse) ones, or even more com-lex functions with multiple thresholds, as to enhance or attenuatehe contribution to the VR image of voxels within a given intensity

ange [7,8,12,13] (Fig. 9). It is also possible to choose among differ-nt colour look-up tables (CLUT) in order to enhance depiction ofomplex anatomical structures; preset CLUT are available on mostIS–PACS software, and can be manually customised by the user for

limb arteries (left) and magnified VR of popliteal and infrapopliteal arteries (right)

his convenience (Fig. 10). Furthermore, VR plug-ins often provideeither automatic or manual image editing tools: automatic editingallows to remove unwanted voxels (usually bone or patient table)from the final image either by automatic software recognition orby mouse-clicking on the structures to be sculpted away, whilemanual editing can be done by the user by mouse-tracing the partsof the VR image to be preserved or deleted (Fig. 11). This enablesaccurate selection of the anatomy to be represented, thereby reduc-ing or eliminating superimposition of surrounding structures thatmight hamper the diagnostic quality of the reconstructed VR cap-

ture. It is also possible to on-the-fly change the rendering mode foredited data (for example, from VR to MIP) as to optimise depiction ofcertain image features with the most appropriate post-processingtechnique [13]. VR images can be exported either in DICOM or non-

2 rnal o

Drlav

2

tteif(f[pdrabtdvhoa(

2

ssVghalteiwc

22 L. Faggioni et al. / European Jou

ICOM exchange format, or dynamically assembled in a movieotating around a user-defined axis for offline file export. To thisatter purpose, several file formats (such as AVI, MOV, or MPEG)re available, along with the possibility to use standard codecs forideo encoding.

.2.2. Shaded surface display (SSD)Some integrated RIS–PACS solutions also include SSD capabili-

ies. Briefly, with this algorithm the user fixes a threshold defininghe minimum intensity of the voxels to be imaged, while vox-ls with lower intensity will be discarded, and the reconstructedmage will be composed by surfaces corresponding to the inter-ace between voxels with intensity above and below the thresholdi.e. void voxels), obtained by simulating the illumination of sur-aces from an external light source in a given user-defined direction8,11,14]. In the past, given the availability of relatively low-owered workstations, SSD was usually advantageous over VRue to its relatively modest computational requirements, becauseepresentation of dataset voxels on the output image is binary,llowing only a fraction of them to be processed. However, theinary nature of the SSD algorithm is also its main weakness, due tohe difficulty to find an appropriate threshold for generation of non-istorted 3D images, causing hole-like artefacts to appear on outputiews in case of bad tissue separation. For this reason, and becauseeavy data processing is no longer a problem (as it is performedn remote servers), SSD has been superseded by the more robustnd powerful VR algorithms, and few applications still support itFig. 12).

.2.3. Virtual endoscopy (VE)By applying SSD or VR algorithms in order to explore hollow

tructures from an intraluminal, rather than extraluminal lightource, it is possible to achieve endoscopy-like views. SSD-basedE techniques require the definition of an intensity threshold andenerate surfaces formed by the interfaces between lower andigher intensity voxels. VR-based VE techniques are more robustnd require the definition of a central intensity value (windowevel) and an intensity range (window width), corresponding tohe signal intensity spectrum of voxels to be displayed with differ-

nt opacities on the VE view. Most common integrated RIS–PACSmage processing tools include VE capabilities, allowing to define

indow levels, opacity, and colour settings, together with a set ofustomisable CLUT, resembling the user interface of VR tools. On

Fig. 13. VE image of colonic polyp with synchronised orth

f Radiology 78 (2011) 210–224

some software packages it is even possible to toggle between VRand VE modes from the same dialogue panel, reflecting the concep-tual and mechanistic similarity between those two techniques. Theposition and orientation of the virtual endoscope, together with themagnification degree of the VE view, can be defined by means ofan electronic cursor onto synchronised orthogonal or oblique MPRprojections (Fig. 13). It is also possible to create a centreline (‘fly-through’) along which VE views are taken, either automatically bysoftware-assisted image segmentation, or manually by user trac-ing. This is particularly useful for exploration of anatomical cavitiesof homogenous intensity, such as airways, the colon, or contrast-enhanced vessels or cardiac cavities. In addition, some programsallow to make geometric measurements from VE views, that canbe useful e.g. for the assessment of vascular ostia from inside bloodvessels or cardiac cavities [15]. Finally, software implementationsexist that enable export of VE images either as standalone capturesor in conjunction with MPR views, thus indicating the exact posi-tion from which VE is performed. As with VR reconstructions, VEscreenshots can be saved either in DICOM mode for PACS storage,or in non-DICOM image format for offline file export, or even asa movie of consecutive VE images. This latter option is especiallyuseful in case of fly-through VE, having the purpose of showing theinternal structure of a hollow viscus or cavity through its entirelength.

3. Dedicated image analysis plug-ins

With the growing diffusion of cross-sectional imaging and theincreasing size of image datasets over time (with particular refer-ence to MDCT datasets), the need for image processing tools witha higher degree of automation has occurred. This holds especiallytrue for cardiovascular imaging, in which the enormous amount ofobtainable data has justified the development of automated imageprocessing tools assisting the radiologist into the interpretationof findings from thousands of images in a reliable and reasonablyfast fashion, compatibly with the routine workflow of a health ser-vice. Dedicated plug-ins for automated and semi-automated imageanalysis are currently available on most RIS–PACS software pack-ages, ranging from CT colonography applications to solutions forcardiovascular imaging. Such tools integrate standard 2D and 3D

post-processing algorithms with filtering and segmentation func-tions dedicated to one particular organ or anatomical territory, as toenable anatomy-specific image analysis. For example, CT colonog-raphy tools perform segmentation of air-containing voxels from the

ogonal MPR views for virtual endoscope positioning.

L. Faggioni et al. / European Journal of Radiology 78 (2011) 210–224 223

F tic vesL

obbtutievpcltvowep

4

r

ig. 14. Segmentation of the right coronary artery with CPR view (a) and automainköping University, Sweden).

riginal dataset (e.g. through region-growing algorithms), allowingoth panoramic or targeted 2D and 3D visualisation of the largeowel, together with more organ-tailored functions such as vir-ual dissection, virtual biopsy, or virtual navigation; this latter cansually be accomplished with optimised window and CLUT set-ings through fly-through centreline-oriented VE. Another examples represented by vessel and cardiac analysis plug-ins, that allowxtraction of voxels from contrast-enhanced vessels (automaticessel tracking) by prior user definition of vessel start and endoints (Fig. 14); subsequently, the software can derive geometri-al information (such as vessel lumen diameter and section area,uminal stenosis or dilation degree, or aneurysm volume), leadingo potential reduction of reader’s error due to inter- and intra-raterariability of manually taken measurements, as well as to lowerverall image interpretation time. Of course, vessel tracking – asell as segmentation algorithms in general – can be controlled and

ventually modified by the user in case of unsuccessful or erroneouserformance, for example due to artefacts on native images.

. Conclusions

Integration of image processing tools with RIS–PACS systemsepresents an evolutionary step toward centralisation of all instru-

sel analysis (b) by means of the open-source CMIV CTA Plug-in for OsiriX (CMIV,

ments for medical image reading and reporting onto one singlemachine. This has the potential, on one hand, to save money andspace by avoiding purchase of dedicated high-end workstations,and, on the other hand, to reduce overall reading time, thus result-ing into increased productivity. We believe that with the increasingavailability of adequate RIS–PACS systems with image processingcapabilities, usage of standalone workstations should be reservedto selected cases of particular complexity, in which the added valueof highly advanced and costly dedicated software might be relevantto diagnosis.

References

[1] Carrino JA. Digital imaging overview. Semin Roentgenol 2003;38(3):200–15.[2] Tellis WM, Andriole KP, Jovais CS, Avrin DE. RIS minus PACS equals film. J Digit

Imaging 2002;15(Suppl. 1):20–6.[3] Mildenberger P, Eichelberg M, Martin E. Introduction to the DICOM standard.

Eur Radiol 2002;12(4):920–7.[4] Khorasani R. Image compression in your PACS: should you do it? What are the

issues? J Am Coll Radiol 2004;1(10):780–1.

[5] Kotter E, Baumann T, Jäger D, Langer M. Technologies for image distribution in

hospitals. Eur Radiol 2006;16(6):1270–9.[6] Lee KH, Lee HJ, Kim JH, et al. Managing the CT data explosion: initial expe-

riences of archiving volumetric datasets in a mini-PACS. J Digit Imaging2005;18(3):188–95.

2 rnal o

[

[

[

[

[

24 L. Faggioni et al. / European Jou

[7] Lell MM, Anders K, Uder M, et al. New techniques in CT angiography. Radio-graphics 2006;26(Suppl. 1):S45–62.

[8] Cody DD. AAPM/RSNA physics tutorial for residents: topics in CT. Image pro-cessing in CT. Radiographics 2002;22(5):1255–68.

[9] Dalrymple NC, Prasad SR, El-Merhi FM, Chintapalli KN. Price of isotropy inmultidetector CT. Radiographics 2007;27(1):49–62.

10] Glockner JF. Three-dimensional gadolinium-enhanced MR angiogra-

phy: applications for abdominal imaging. Radiographics 2001;21(2):357–70.

11] Prokop M, Shin HO, Schanz A, Schaefer-Prokop CM. Use of maximumintensity projections in CT angiography: a basic review. Radiographics1997;17(2):433–51.

[

f Radiology 78 (2011) 210–224

12] Calhoun PS, Kuszyk BS, Heath DG, Carley JC, Fishman EK. Three-dimensionalvolume rendering of spiral CT data: theory and method. Radiographics1999;19(3):745–64.

13] Fishman EK, Ney DR, Heath DG, Corl FM, Horton KM, Johnson PT. Volume ren-dering versus maximum intensity projection in CT angiography: what worksbest, when, and why. Radiographics 2006;26(3):905–22.

14] Davis CP, Hany TF, Wildermuth S, Schmidt M, Debatin JF. Postprocess-

ing techniques for gadolinium-enhanced three-dimensional MR angiography.Radiographics 1997;17(5):1061–77.

15] Cirillo S, Tosetti I, Gaita F, Bianchi F, Gandini G, Regge D. Magnetic resonanceangiography of the pulmonary veins before and after radiofrequency ablationfor atrial fibrillation. Radiol Med 2005;109(5–6):488–99.