Introduction

Over the past 13 years, the field of computed tomogra-phy (CT) has been revolutionized by two major techni-cal developments: the introduction of helical CT as asingle detector technique in 1989 [1] and the develop-ment of multislice helical CT. The introduction of heli-cal CT has changed CT from a two-dimensional to athree-dimensional imaging modality, as the helical tech-nique produces volume data sets and slices are recon-structed secondarily. The first dual-slice helical CTscanner, introduced in 1992, consisted of two paralleldetector rows that allowed the simultaneous acquisitionof two interwoven helices [2]. The concept of dual-sliceCT advanced to multislice helical CT in 1998 when ar-rays between 8 and 34 detector rows became available,thus simultaneously generating up to four slices and fourinterwoven helices. The latest technical developmentwas the introduction of a 16-slice helical CT scanner,presented to the radiologic community at the 2001 meet-ing of the Radiological Society of North America.

Multislice helical CT technology

The scanner manufacturing companies have developeddifferent concepts of detector arrays consisting of multi-ple parallel detector rows with varying widths and allow-ing slice collimations of 2×0.625 mm (General Electrics),2×0.5 mm (Siemens/Philips), 4×0.5 mm (Toshiba), 2×10mm (General Electrics/Siemens/Philips) and 4×8 mm(Toshiba) [3]. The rotation time has been reduced to aminimum of 0.5 s per 360° rotation and pitch-factors of0.75 or 1.5 are offered with the General Electrics multi-slice CT scanners or can be chosen freely up to a value of2 with other manufacturers’ scanners [3]. Image recon-struction is performed by z-interpolation and the recon-struction increment to determine the degree of overlapof the reconstructed slices [3].

These advancements have created a new dimensionin temporal and spatial resolution in CT scanning. Thespeed of multislice CT can either be used to reduce thetime needed to cover a given volume or to use thinnercollimations to increase the resolution along the z-axis.For example, the high speed of multislice CT allows ex-amination of the whole abdomen within one breathhold(collimation 4×2.5 mm, pitch 1.5). Motion artifactscaused by breathing or bowel movement are thus re-duced [4] and sedation of young children is less fre-quently needed than in single-slice CT, as reported ar-restingly by Pappas et al. [5]. Only three of 219 childrenrequired sedation (1.4%) in this study. The sedation ratewas 3% (three of 90) for children aged six years oryounger and 8% (three of 37) for those aged one year oryounger. Examinations were of the chest, abdomen, andpelvis in 68 patients, of the abdomen and pelvis in 112,and of the chest alone in 39. Contrast material was in-travenously administered in 186 (85%) examinations.All scans were of diagnostic quality.

Thin collimations result in an improvement of thespatial resolution along the z-axis and in almost isotrop-ic voxels, e.g. collimation 4×1 mm, which is an importantprerequisite for high-quality interactive multiplanar andthree-dimensional (3D) image reconstruction. In addi-tion, there is a reduction of volume averaging and anoverall improvement in the assessment of small struc-tures [6, 7].

Economic issues have often been overlooked in thepast but have gained importance over the last severalyears. The short acquisition time and high diagnostic qual-ity of multislice CT lead to an increase in CT productivityeven though more advanced protocols are performedcompared to single-slice CT. Jhaveri et al. [8] from Massa-chusetts General Hospital in Boston reported the impactof multislice CT technology on scanner productivity in atertiary care medical center. They compared the produc-tivity of two diagnostic CT scanners during the periodsfrom 1 January to 31 August, 1999 (when both scannershad single-slice CT capability) and 1 January to 31 Au-

Eur. Radiol. 12 (Suppl. 2) S5-S10 (2002)© Springer-Verlag 2002 Present and Future Trends in MDCT Applications

Multislice helical CT of the abdomenSabine Fenchel1, Thorsten R. Fleiter1, Elmar M. Merkle2

1Department of Radiology, University of Ulm, Ulm, Germany2Department of Radiology, Case Western Reserve University/University Hospitals of Cleveland, Cleveland, Ohio, USA

S6

gust, 2000 (when one of these scanners was replaced witha multislice CT scanner). The scanners were used primar-ily for outpatients during the day shift and for inpatientsduring the evening shift; the demand for CT services wasstable. The number of CT studies performed on the twoscanners increased by 1772 (13.1%) from 13 548 (beforemultislice CT) to 15 320 (when multislice CT was avail-able). The proportion of examinations enhanced with con-trast medium increased from 52% to 65%. Between 9:00A.M. and 5:00 P.M., the number of CT examinations wassimilar on the single-slice scanners during the two periods(p > 0.05). However, when multislice CT was available,the number of studies performed on the multislice scanner(5919) was 51.9% higher than those performed using thesingle-slice scanner (3896) (p < 0.0006).

Liver

The high speed of multislice CT allows liver scanningwithin one breathhold and a better differentiation ofthe arterial and portal venous perfusion phases. Withan acquisition time for the liver of about 10 s (collima-tion 4×2.5 mm, pitch 1.5), especially the arterial phase(with a duration of about 7-8 s) can be more clearly dif-ferentiated and the detection and characterization ofparticularly hypervascularized liver lesions are im-proved [3, 7]. Even double arterial phase examinationswithin a single breathhold are now possible for the firsttime [9]. The high spatial resolution along the z-axiswith nearly isotropic voxels reduces the volume averag-ing effects and is the basis for multiplanar reformationsof high image quality; this is the prerequisite for opti-mal detection, characterization, and localization of

small liver lesions. Coronal and sagittal reformations ofhigh image quality are enormously helpful for surgeryplanning since localization of malignant liver lesions, at-tribution of lesions to liver segments, and localization oflesions in relation to hepatic vessels, are distinctivelyfacilitated (Fig. 1) [7].

Pancreas

Examination of the pancreas is possible with a collima-tion of 4×1 mm and a pitch of 1.5 during the arterial andvenous (portal) phases. The pancreas, pancreatic duct,and bile duct can thus be clearly delineated, and the de-tection and characterization of small hyper- or hypovas-cularized lesions are improved. Thin collimation leads tomultiplanar reformations and 3D reconstructions ofhigh image quality, and the complex anatomic relation-ship of pancreatic tumors to the surrounding vessels andorgans can be more easily assessed (Fig. 2) [6]. In partic-ular, the assessment of vessel infiltration by pancreatictumors is improved [6].

Gastrointestinal tract

The high scanning speed of multislice helical CT allowsexamination of the whole abdomen within one breath-hold and with a spatial resolution along the z-axis ofabout 3 mm. Motion artifacts by breathing or bowelmovement are reduced by the high scanning speed. Af-ter oral or rectal contrast medium administration andadditional intravenous contrast medium, an exact analy-sis of the wall of the gastrointestinal tract is possible

a

Fig. 1a, b. Multiplanar reconstruction of several hypointense colorectal liver metastases. a Axial image. b Coronal image

b

S7

(Fig. 3). The possibility to examine large volumes withhigh spatial resolution and reduced motion artifacts isthe basis for virtual endoscopy leading to promising re-sults (Fig. 4). Hara et al. [10] compared respiratory arti-facts, colonic distention, and polyp detection on CTcolonography using single- and multidetector row heli-cal CT systems. A total of 237 consecutive patients re-ceived subcutaneously administered glucagon and un-derwent prone and supine CT colonography with single-detector row CT (n=77) and multi-detector row CT(n=160) followed by colonoscopy. Suboptimal colonicdistention was significantly more common using single-detector row CT and was present in at least one segmentin 52% (40 of 77 patients) of the examinations versusonly 19% (30 of 160 patients) of the multidetector row

CT exams (p < 0.001). Mild respiratory artifacts werepresent in 61% (47 of 77 patients) of single-detector rowCT examinations versus only 16% (26 of 160 patients) ofmultidetector row CT examinations (p < 0.001). Depic-tion of polyps larger than 10 mm was 89% (8 of 9polyps) on single-detector row CT and 80% (8 of 10polyps) on multidetector row CT (p > 0.05).

Kidneys

The high spatial resolution of multislice CT leads to animprovement in the detection and characterization ofsmall kidney lesions [3]. Organ-sparing surgical tech-niques requiring an exact delineation of the tumor and

Fig. 2a-c. Carcinoma of the pancreatic tail with encasement of the celiac trunk and the superior mesenteric artery. a Axial slice of the ar-terial phase. b Axial slice of the portal-venous phase. c Sagittal reconstruction. Encasement of the celiac trunk and the superior mesen-teric artery can be easily assessed in the sagittal reconstruction

a b c

Fig. 3. Axial slice of a patient with diverticulosis of the sigma.The diverticuli are clearly delineated. There is no inflammationof the surrounding fatty tissue

Fig. 4. Virtual endoscopy of the colon ascendens. The high spa-tial resolution of multislice helical CT leads to reconstructions ofhigh image quality

S8

a

Fig 5a, b. Aneurysm of the infrarenal aorta after aorto-biiliacal stent placement. a Axial slice. b Coronal reconstruction. There is regulararterial perfusion in the stent lumen and thrombosis of the aneurysmal sack

b

a

Fig. 6a, b. 3D image reconstruction. Higher image quality and better delineation of the large and small arteries with an iodine concen-tration of 400 mg/ml (a) compared to 300 mg/ml (b)

b

the segmental arteries can be performed based on mul-tislice helical CT data [3]. Low dose CT examinations ofthe urinary tract for detection of calculi are possiblewith collimations of about 1 mm. Multislice CT urogra-phy, a new imaging method for visualization of the up-per urinary tract, is now possible for the first time [11].

CT angiography of the aorta and renal arteries

Multislice helical CT permits examination of the abdomi-nal aorta as well as the iliac and proximal femoral arterieswith a slice thickness of 2-3 mm within one breathhold.The amount of contrast medium can be reduced to about80 ml compared to the amount required with single-slicehelical CT (130 ml) because of the higher scanning speed

and since enhancement of the arteries themselves is suffi-cient for diagnosis [6]. Correct timing of the scan delay isvery important in order to avoid an overtake of the con-trast bolus by the detector array. Data acquisition in mul-tislice CT should be started a few seconds later than in sin-gle slice CT [6]. Standard protocols with a slice thicknessof 2-3 mm allow visualization of the aorta, the iliac andfemoral arteries, and the proximal aortal branches to as-sess stenosis, occlusion, dissection or aneurysm (Fig. 5).

Assessment of the renal arteries is possible if a thincollimation of 4×1 mm and a pitch ≤ 1.5 are used (Fig. 6)[6]. Thus data sets with a high spatial resolution alongthe z-axis are obtained, and multiplanar reconstructionsand maximum intensity projections of high quality thatfacilitate the detection of vessel occlusion, stenosis ordissection can be created.

S9

Contrast media

Another important factor in multislice CT is the con-trast medium dynamics. Especially in imaging the ab-domen and pelvis, where intravenous contrast agentsare administered in approximately 90% of all studies[12], contrast media protocols must be adapted to thehigher scanning speed. Currently, application of con-trast medium by means of a mechanical injector has be-come standard; vascular access is usually through an 18to 20-gauge indwelling catheter placed in a cubital vein.Nonionic contrast media with an iodine concentrationof 300 mg/ml are most commonly used and are appliedat a flow rate of 1-5 ml/s depending on the kind of ex-amination. The recommended maximum amount of ap-plied iodine is 35-45 g; this will probably not change sig-nificantly despite the introduction of multislice helicalCT scanners.

Contrast medium concentration and injection rate ac-tually become more crucial in the case of multislice heli-cal CT studies as the high scanning speed allows a sepa-rate acquisition of the arterial, parenchymal, and portalvenous perfusion phases that are of short duration butmust reach a maximum degree of contrast. A higher ear-ly contrast medium uptake can be achieved either by anincreased contrast medium concentration or by an in-creased flow rate. As increases in the flow rate of thecontrast medium injection are subject to certain limits,concentration of the contrast medium solution (the totalamount of iodine applied being kept constant) wouldappear to be an important parameter [13-15].

In a recent study presented at the 2001 annual meet-ing of the Radiological Society of North America, we re-ported our experience in an abdominal study using anonionic contrast medium with an iodine concentrationof 400 mg/ml [16, 17]. Following Institutional ReviewBoard (IRB) approval, 50 patients with suspected orknown pancreatic tumors were examined in this study.The patients in group 1 received 130 ml nonionic con-trast medium with an iodine concentration of 300 mg/mlat a flow rate of 5 ml/s. The patients in group 2 received98 ml nonionic contrast medium with an iodine concen-tration of 400 mg/ml at a flow rate of 5 ml/s (Tables 1, 2).The following parameters were evaluated quantitiative-ly: arterial and venous phase contrast enhancement ofthe aorta, superior mesenteric artery, celiac trunk, portalvein, liver, pancreas, spleen, and both kidneys. In addi-tion, contribution of the contrast agent to the diagnosticvalue, technical quality, tumor delineation, evaluabilityof organ infiltration, and evaluability of vessels in thevarious phases were graded by two independent blindedreaders (Table 3). The patients in group 2 (iodine con-centration, 400 mg/ml) demonstrated a significantlyhigher arterial phase enhancement of the aorta, superiormesenteric artery, celiac trunk, pancreas, spleen, andkidneys. A significantly higher portal venous phase en-

Table 1. Injection conditions for dual-phase multidetector CT us-ing a noniomic contrast medium (iomeprol) containing iodine at300 or 400 mg/ml, as used in [16, 17]

Iomeprol Iomeprol300 400

Volume, ml 130 98Duration, s 26 20Iodine dose rate, g/s 1.5 1.96Total iodine dose, g 39.0 39.2

Table 2. Technical parameters for dual-phase multidetector CT ac-cording to the iomeprol 300 and 400 protocols described in Table 1.Values are for a 4-slice helical CT scanner operating at a maximumof 0.5 s gantry rotation time, 120 rV and 165 mAs, depending onthe individual patient and study (modified from [16, 17])

Arterial Venousphase phase

Scan delay Bolus-triggered 60 sChannel width, mm 1 × 2.5 4 × 5Table speed (mm/rotation) 0.75 0.75Pitch 1.25 1.25Reconstruction slice thickness, mm 3.2 6.5Reconstruction interval, mm 1.6 3.2

hancement was noted in the pancreas and in the portalvein in group 2 patients. Semiquantitative analysis re-vealed a significantly greater contribution of the 400 mgiodine/ml contrast agent to the diagnostic value withreader 1, a significantly improved technical quality withreader 2, and a significantly greater evaluability of thevessels in the arterial phase with reader 2 (Table 3).

Radiation exposure

Phantom studies have revealed a possible increase in ra-diation exposure when imaging protocols of multisliceCT scanners have been compared to those of single-sliceCT scanners [18]. In general, energy dose values of theabdomen increased 2.6-fold with multislice CT com-pared to single-slice CT. To take full advantage of thepotential of multislice helical CT, imaging protocolsmust be adapted and optimized but the radiation dosemust also be taken into consideration.

Data explosion

Without a doubt, a great challenge of multislice CT isdealing with the “data explosion” [19]. For example, atypical abdominal examination yields about 300-500 ax-ial images (collimation 4×2.5 mm, reconstruction incre-ment 1.6 mm). If radiologists are to take full advantageof multislice CT scanners, they will need to change theway they interpret, transfer, and store CT data. Film isno longer an option. Instead, workstation-based review

S10

of multiplanar reconstructions for interpretation is a ne-cessity. In addition, alternative visualization and analysisusing volumetric tools, such as three-dimensional op-tions including maximum intensity projections, surfaceshaded displays, and volume rendering, must evolvefrom a luxury to daily practice.

References

1. Kalender WA, Seissler W, Klotz E, Vock P (1990) Spiralvolumetric CT with single-breath-hold technique, continu-ous transport, and continuous scanner rotation. Radiology176:181-183

2. Silverman PM, Kalender WA, Hazle JD (2001) Commonterminology for single and multislice helical CT. AJR Am JRoentgenol 176:1135-1136

3. Shin H, Heiland S (2001) Multi-Detektor-Computertomo-graphie (German). Radiologie up2date 4:403-414

4. Aschoff AJ, Gorich J, Sokiranski R, Rieber A, Brambs HJ,Kramer SC (1999) Pancreas: does hyoscyamine butylbro-mide increase the diagnostic value of helical CT? Radiology210: 861-864

5. Pappas JN, Donnelly LF, Frush DP (2000) Reduced fre-quency of sedation of young children with multisection heli-cal CT. Radiology 215:897-899

6. Schoepf UJ, Becker C, Bruening R, Hong C, Rust GF,Helmberger T, Leimeister P, Stadie A, Niethammer M,Klingemann B, Reiser MF (1999) Computed tomography ofthe abdomen with multidetector-array CT. Radiologe39:625-661

7. Kulinna C, Helmberger T, Kessler M, Reiser M (2001) Im-proved detectability of liver metastases with multislice CT.Radiologe 41:16-23

8. Jhaveri KS, Saini S, Levine LA, Piazzo DJ, Doncaster RJ,Halpern EF, Jordan PF, Thrall JH (2001) Effect of multi-slice CT technology on scanner productivity. AJR Am JRoentgenol 177:769-772

9. Takahashi S, Murakami T, Takamura M, Kim T, Hori M,Narumi Y, Nakamura H, Kudo M (2002) Multi-detector rowhelical CT angiography of hepatic vessels: depiction withdual-arterial phase acquisition during single breath hold. Ra-diology 222:81-88

10. Hara AK, Johnson CD, MacCarty RL, Welch TJ, McCol-lough CH, Harmsen WS (2001) CT colonography: single-versus multi-detector row imaging. Radiology 219:461-465

11. Nolte-Ernsting CC, Wildberger JE, Borchers H, Schmitz-Rode T, Gunther RW (2001) Multi-slice CT urography afterdiuretic injection: initial results. Rofo Fortschr Geb Ront-genstr Neuen Bildgeb Verfahr 173:176-180

12. O’Malley ME, Halpern E, Mueller PR, Gazelle GS (2000)Helical CT protocols for the abdomen and pelvis: a survey.AJR Am J Roentgenol 175:109-113

13. Engeroff B, Kopka L, Harz C, Grabbe E (2001) Impact ofdifferent iodine concentrations on abdominal enhancementin biphasic multislice helical CT (MS-CT). Rofo FortschrGeb Rontgenstr Neuen Bildgeb Verfahr 173:938-941

14. Tublin ME, Tessler FN, Cheng SL, Peters TL, McGovernPC (1999) Effect of injection rate of contrast medium onpancreatic and hepatic helical CT. Radiology 210:97-101

15. Kim T, Murakami T, Takahashi S, Okada A, Hori M, Naru-mi Y, Nakamura H (1999) Pancreatic CT imaging: effects ofdifferent injection rates and doses of contrast material. Ra-diology 212:219-225

16. Fenchel S, Fleiter TR, vanGessel R, Brambs HJ, Merkle EM(2001) Effect of iodine concentration on contrast enhance-ment in multislice helical CT of the abdomen. Radiology.221(P):119 (abstract)

17. Fenchel S, Fleiter TR, Aschoff AJ, vanGessel R, BrambsHJ, Merkle EM (2001) Multislice helical CT of the pancreas:effect of iodine concentration on diagnostic value. Radiology221(P):119 (abstract)

18. Giacomuzzi SM, Torbica P, Rieger M, Lottersberger C, PeerS, Peer R, Perkmann R, Buchberger W, Bale R, Mallouhi A,Jaschke W (2001) Radiation exposure in single slice andmulti-slice spiral CT (a phantom study). Rofo Fortschr GebRontgenstr Neuen Bildgeb Verfahr 173:643-649

19. Rubin GD (2000) Data explosion: the challenge of multide-tector-row CT. Eur J Radiol 36:74-80

Table 3. Comparison of outcomes of dual-phase pancreatic multidetector CT according to the iomeprol 300 and 400 protocolsdescribed in Table 1 (modified from [16, 17]

Reader 1 Reader 2

Contribution of the contrast agent Iomeprol 400 superior* Iomeprol 400 superiorto the diagnostic value (10-point scale)

Technical quality Iomeprol 400 superior Iomeprol 400 superior*

Tumor delineation No difference No difference

Evaluability of organ infiltration No difference No difference

Evaluability of vessels in the arterial phase Iomeprol 400 superior Iomeprol 400 superior*

Evaluability of vessels in the portal-venous phase No difference No difference

* statistically significant (p < 0.05)