Mannudeep K. Kalra, MD,DNB

Michael M. Maher, MD, FFR(RCSI), FRCR

Thomas L. Toth, DScBernhard Schmidt, PhDBryan L. Westerman, PhDHugh T. Morgan, PhDSanjay Saini, MD

Index terms:Computed tomography (CT), image

qualityComputed tomography (CT),

radiation exposureComputed tomography (CT),

technologyRadiations, exposure to patients and

personnelReview

Published online before print10.1148/radiol.2333031150

Radiology 2004; 233:649–657

Abbreviations:ACS � automatic current settingATCM � automatic tube current

modulation

1 From the Division of Abdominal Im-aging and Intervention, Departmentof Radiology, Massachusetts GeneralHospital and Harvard Medical School,Wht-270, 32 Fruit St, Boston, MA02134 (M.K.K., M.M.M.); GE MedicalSystems, Waukesha, Wis (T.L.T.); Sie-mens Medical Solutions, Forchheim,Germany (B.S.); Toshiba AmericaMedical Systems, Tustin, Calif (B.L.W.);Philips Medical Systems, Cleveland,Ohio (H.T.M.); and Department of Ra-diology, Emory University School ofMedicine, Emory University Hospital, At-lanta, Ga (S.S.). Received July 25, 2003;revision requested September 30; finalrevision received and accepted Decem-ber 9. Address correspondence to S.S.(e-mail: mannudeep_k_kalra@yahoo.com).© RSNA, 2004

Techniques and Applicationsof Automatic Tube CurrentModulation for CT1

Introduction of slip-ring technology with subsequent development of single– andmulti–detector row helical computed tomographic (CT) scanners have expandedthe applications of CT, leading to a substantial increase in the number of CTexaminations being performed. Owing to concerns about the resultant increase inassociated radiation dose, many technical innovations have been introduced. Onesuch innovation is automatic tube current modulation. The purpose of automatictube current modulation is to maintain constant image quality regardless of patientattenuation characteristics, thus allowing radiation dose to patients to be reduced.This review discusses the principles, clinical use, and limitations of different auto-matic tube current modulation techniques.© RSNA, 2004

Tube current (measured in milliamperes) is an important determinant of radiation doseand image quality in x-ray–based examinations. Recent advances in computed tomo-graphic (CT) technology, including implementation of automatic tube current modula-tion (ATCM), allow reduction in radiation exposure during CT examinations (1–6). ATCMmay be defined as a set of techniques that enable automatic adjustment of the tube currentin the x-y plane (angular modulation) or along the z-axis (z-axis modulation) according tothe size and attenuation characteristics of the body part being scanned and achieveconstant CT image quality with lower radiation exposure. Hence, ATCM techniques areanalogous to the automatic exposure-control or photograph-timing techniques used inconventional radiography.

Amid growing concerns about CT radiation exposure, the adoption of ATCM techniquesshould permit overall reduction in radiation exposure in CT examinations. Unfortunately,owing to rapid technologic advances, different vendors have developed different ATCMtechniques and use proprietary nomenclature.

Nevertheless, the introduction of ATCM techniques in modern CT scanners representsan important step toward standardization of tube current protocols, with elimination ofarbitrary selection by radiologists and technologists. This article will describe the princi-ples of ATCM techniques, their clinical use, and the advantages and disadvantages ofapplication in diagnostic CT scanning.

PRINCIPLES OF ATCM TECHNIQUES

Unlike in conventional radiography, during CT scanning the x-ray tube continuouslyrotates around the patient, emitting x-rays that traverse through a cross section of thebody to generate attenuation profiles (image raw data) at the detectors. Thus, incidentx-ray beam projections from multiple directions around the region of interest are used toreconstruct each cross-sectional CT image. Scanning parameters such as tube current andtube potential determine photon fluence and incident beam energy, which affect imagequality and absorbed radiation dose. With other scanning parameters held constant, areduction in tube current decreases radiation exposure but increases image noise or mottle,a principle determinant of image quality. Likewise, an increase in tube current leads to anincrease in radiation exposure and a decrease in image noise. On a scale of diagnosticacceptability, image noise and radiation exposure represent conflicting factors, and dis-proportionate emphasis on either factor may have an adverse effect on image quality orradiation dose. Whereas images with lower noise (increased radiation exposure) may not

Review

649

Ra

dio

logy

reveal additional diagnostic information,images with higher noise may obscurelesions that might have been visible inlower-noise images (7–9).

Authors of previous clinical and exper-imental studies have reported that satis-factory image quality can be obtainedwith a reduction of tube current on thebasis of weight and cross-sectional di-mensions of patients undergoing CTscanning (10–12). Results of these studiesmay be explained on the basis of variableattenuation of the incident beam travers-ing a particular cross-sectional dimen-sion at a particular projection angle (1,2).The resultant attenuation determines im-age noise and is affected by scanning pa-rameters, particularly tube current. Forexample, greater beam attenuation in aparticular dimension or projection willresult in greater noise and will requirehigher tube current than that needed bya beam undergoing less attenuation inanother projection.

Manual adjustment of tube currentbased on patient weight or dimensionscan aid in establishing an appropriatebalance between image noise and radia-tion exposure (10,12). However, theseadjustments do not guarantee constantimage quality throughout the entire ex-amination. For example, in CT scanningof chest, the choice of a fixed tube cur-rent does not account for differences inbeam attenuation between the shoulderregion and midchest region or between an-teroposterior and lateral cross-sectional di-mensions. ATCM techniques allow main-tenance of constant image quality at arequired radiation exposure level becauseATCM rapidly responds to large variationsin beam attenuation. ATCM is based on

the fact that image noise is determinedby x-ray quantum noise in the transmit-ted beam projections. This techniqueaims to modulate tube current on thebasis of regional body anatomy for ad-justment of x-ray quantum noise tomaintain constant image noise with im-proved dose efficiency.

Currently, two distinct techniques areavailable for ATCM: angular (x-y) modu-lation and z-axis modulation. Both tech-niques modulate tube current in an effortto maintain constant image quality atthe lowest dose while simultaneously re-ducing tube loading (heating) and mini-mizing streak artifacts caused by a mini-mal number of photons.

Angular Modulation

The angular-modulation techniquewas introduced in 1994 for a single–de-tector row helical CT scanner (Smart-Scan; GE Medical Systems, Waukesha,Wis) (13–15). This software-based tech-nique modulated tube current on the ba-sis of the measured density of regionalstructures and the absorption values ofthe object of interest. This informationwas obtained by measuring local x-raybeam absorption in 100 central channelson two localizer radiographs (lateral andanteroposterior views). Preprogrammedsinusoidal modulation of tube currentwas achieved during 360° rotation forequalization of local differences in beamabsorption to obtain relatively constantnoise content and reduce radiation expo-sure. A radiation dose reduction of up to20%, depending on patient geometry(asymmetry), has been reported (14). Arecent refinement of the angular-modu-lation approach is an online, real-time,anatomy-adapted, attenuation-based tubecurrent modulation technique (CAREDose; Siemens Medical Solutions, Forch-heim, Germany) that does not need theinformation of radiographic localizer im-ages to achieve ATCM (1–3,16–20).

Angular-modulation techniques auto-matically adjust the tube current for eachprojection angle to the attenuation of thepatient to minimize x-rays in projectionangles (eg, anteroposterior or posteroan-terior angles are less important than arelateral projections because the formercause less beam attenuation and henceare associated with less noise) that areless important with regard to reducingthe overall noise content (Fig 1). In phan-tom studies in which an online angular-modulation technique was comparedwith the previous angular-modulationtechnique, a substantially greater radia-

tion reduction (up to 50%) with the on-line modulation technique was docu-mented (1).

In patients with circular cross-sectionalgeometry, beam attenuation is constantin all projections (x-ray beam projectionangles). However, in a noncircular cross-sectional geometry, attenuation variesstrongly in different projections, some-times by more than three orders of mag-nitude. In these settings, image noise inhigh-attenuation projection angles deter-mines the overall image noise content.Thus, at angular projections with a smallpatient diameter or body region beingimaged (eg, greater tube current reduc-tion occurs in chest than in abdomen),corresponding to a relatively lower atten-uation, the tube current can be reducedsubstantially without a measurable in-crease in the overall noise content of theimage. Without angular modulation, tubecurrent is held constant over the 360° ro-tation, regardless of the patient attenua-tion profile. The angular-modulation tech-nique reduces tube current as a function ofprojection angles for low-attenuationprojections (anteroposterior vs lateralprojections). This technique calculatesthe modulation function (an objectiveimage quality parameter) from the onlineattenuation profile of the patient. Themodulation function data are processedand sent to the generator control for tubecurrent modulation with a delay of 180°from the x-ray generation angle. In re-gions with marked asymmetry, such asthe shoulders in CT scanning of chest,where attenuation is substantially less inthe anteroposterior direction than in thelateral direction, a reduction in radiationdose of up to 90% can be achieved in theanteroposterior or posteroanterior direc-tion by using the angular-modulationtechnique (20). In summary, the tech-nique of angular modulation aids in im-proving dose efficiency in the x-y axis byreducing radiation exposure in a particu-lar scanning plane.

The Dose-Right Dose Modulation, orDOM, technique (Philips Medical Sys-tems, Eindhoven, the Netherlands) of an-gular modulation also adjusts tube cur-rent in asymmetric anatomic regions.This technique is based on the premisethat tube current should be modulatedaccording to the square root of the mea-sured attenuation for that projection (1).The DOM technique modulates the cur-rent within a 360° tube rotation accord-ing to the square root of the attenuationmeasured from the similar and previous180° or 360° views. In other words, theattenuation measured at an angular view

ESSENTIALS● Automatic tube current modulation

techniques are based on the fact thatimage noise is determined by x-rayquantum noise in the transmittedbeam projections.

● These techniques adjust tube current inan effort to maintain constant imagequality at the lowest dose.

● Automatic tube current modulationrepresents an exciting recent techno-logic innovation to enable radiationdose optimization.

650 � Radiology � December 2004 Kalra et al

Ra

dio

logy

(projection angle) is used to optimize thetube current later for a similar angularview.

Z-Axis Modulation

The z-axis–modulation (AutomA, GEMedical Systems; Real E.C., Toshiba Med-ical, Tokyo, Japan) technique functionssomewhat differently than does angularmodulation (21). The AutomA techniqueadjusts the tube current automatically tomaintain a user-specified quantum noiselevel in the image data. It provides anoise index to allow users to select theamount of x-ray noise that will bepresent in the reconstructed images. Us-ing a localizer radiograph, the scannercomputes the tube current needed to ob-tain images with a selected noise level.Hence, z-axis modulation attempts tomake all images have a similar noise irre-spective of patient size and anatomy. Thenoise index value is approximately equalto the image noise (standard deviation)in the central region of an image of auniform phantom.

In the z-axis–modulation technique,the system determines the tube currentby using the patient’s localizer radio-graph projection data and a set of empir-ically determined noise prediction coeffi-cients by using the reference technique(Fig 2). The reference technique com-prises an arbitrary 2.5-mm-thick sectionobtained at the selected peak voltage and100 mAs for transverse reconstructionwith a standard reconstruction algorithm.The projection data from a single localizerradiograph can be used to determine thedensity, size, and shape information of thepatient (4,5). The total projection attenua-

tion data of a single localizer radiographcontain the patient’s density and size in-formation about the projection area,whereas the amplitude and area of the pro-jection contain the patient’s shape infor-mation, which gives an estimate of the pa-tient’s elliptic asymmetry expressed as anoval ratio at a given z-axis position. Theoval ratio is the ratio of the a and b param-eters (lengths of the long and short axes) ofan ellipse. The ellipse parameters can bedetermined for the patient by using theequation for the area of an ellipse. Themeasured projection area and amplitudefrom the localizer radiograph give the areaand the length of one axis, a, of the ellipse,allowing the length of the other axis, b, tobe calculated. These characteristics of thelocalizer radiograph predict the amount ofx-rays that will reach the detector for aspecified technique and, hence, determinethe image standard deviation due to x-raynoise for a given reconstruction algorithm.The predicted x-ray noise at a given z-axisposition for the reference technique (ie,reference noise) is calculated from the pro-jection area and oval ratio from the local-izer radiograph by using the polynomialcoefficients that were determined from thenoise measurements in a set of phantomsrepresenting a wide range of patient sizesand shapes.

With knowledge of the reference noiseand the difference between the referencetechnique and the selected techniquedata, the tube current required to obtaina prescribed noise index is easily calcu-lated by using known x-ray physics equa-tions, which state that noise is inverselyrelated to the square root of the numberof photons and that the number of pho-

tons is proportional to section thickness,section acquisition time, and tube cur-rent. In the currently available version ofz-axis modulation of tube current, an ad-justment factor for different helicalpitches is incorporated in the calculationto account for noise differences betweenhelical selections and the transverse ref-erence technique data. Recent data (4–6)from clinical evaluations of z-axis modu-lation suggest that the radiation dose re-duction with this technique is expectedto be greater than that with fixed-tube-current methods, since the tube currentis automatically reduced for smaller pa-tients and specific anatomic regions.

Often, the actual noise measured onthe image by drawing a region of interestwill differ from the noise index selectedfor scanning. This is due to the fact thatnoise index settings only adjust the tubecurrent, whereas the standard deviationis also affected by other parameters, in-cluding the reconstruction algorithm,the reconstructed section thickness (ifdifferent from the prospective thickness),the use of image space filters, variationsin patient anatomy and patient motion,and the presence of beam-hardening ar-tifacts. Substantial differences betweenthe selected noise index and the standarddeviation can also occur in very largepatients owing to insufficient signalstrength at the detector and superimpo-sition of electronic noise, which can beminimized by using a higher peak volt-age. Likewise, improper centering of pa-tients in the scan field of view can resultin noisier images owing to inappropri-ate beam attenuation by the bow-tiefilter (5). As bow-tie filter attenuation

Figure 1. Angular modulation with CARE Dose (Siemens Medical Solutions) ATCM technique. (a) Online modulation of tube current is performedat different projections in the x-y plane within each 360° x-ray tube rotation. Thin arrows indicate reduction of tube current relative to higher tubecurrent (thick arrows). (b) Users specify an effective tube current value in milliamperes (circled) to perform scanning with this technique.

Volume 233 � Number 3 Automatic Tube Current Modulation � 651

Ra

dio

logy

increases with distance from isocenter,the thickest part of the patient shouldbe approximately centered in the scanfield of view to prevent inappropriateattenuation compensation by the bow-tie filter due to incorrect patient posi-tioning.

The Real E.C. technique implementedon Toshiba CT scanners is a z-axis–mod-ulation technique that is also based onpatient attenuation measurements ac-quired from a single localizer radio-graph. Evolving from its original form,which provided operators with severaloptions in the choice of nominal tubecurrent to modulate on the basis of pa-tient attenuation data, Real E.C. nowoffers four levels of image noise to matchthe diagnostic needs of the examination.The scanner software enables this by calcu-lating the “water-equivalent” thickness ofeach section from the localizer radiograph(Figs 3, 4). The appropriate tube current isapplied at the thickest section in the z-axis direction to achieve the selected

standard deviation (noise level). Tubecurrent is then modulated to maintain

the selected noise level throughout theentire scan.

Figure 2. Z-axis modulation with AutomA (GE Medical Systems) ACTM technique. (a) Auto-matic modulation of tube current from one section location to others in the scanning directionis performed. (b) User selects noise index or enters value for desired noise index and setsminimum and maximum current values. (c) Tube current values (in milliamperes) can bepreviewed prior to scanning.

Figure 3. Z-axis modulation with Real E.C. technique (Toshiba Medical). (a) Attenuation ismeasured on a digital radiograph (left) and is converted to water-equivalent thickness (right),allowing user to specify image quality by choosing different noise levels. (b) After user selects tubecurrent or, more appropriately, desired noise level for the examination (left), the software displaysthe automatic modulation of tube current that will be used to achieve selected image quality.

652 � Radiology � December 2004 Kalra et al

Ra

dio

logy

Automatic current setting (ACS) is acomponent of the Dose-Right technique(Philips Medical Systems). ACS suggestsa tube current–time product level for in-dividual patient examinations to achievea preselected quantitative image noiselevel measured in the central region of astored reference image from a prior ex-amination performed to generate this im-age with a desired noise content for aspecific protocol. ACS is calibrated on aprotocol-by-protocol basis at the clinicalsite by storing reference images thatachieve the clinical site’s desired imagequality (ie, signal-to-noise ratio) for thatprotocol. In practice, a protocol is se-lected, the ACS is chosen, the localizerradiograph associated with the protocolis obtained, and the system host com-puter compares the localizer radiographdata with the protocol’s reference local-izer radiograph data and calculates pa-tient-specific calibration factor and tubecurrent–time product for the CT scan.The suggested tube current–time productfor the study will produce CT imageswith image noise (image quality) similarto that in the site’s prestored referenceCT images. The system displays the vol-ume CT dose index (radiation dose re-duction with ACS prior to scanning) bycalculating the dose difference betweenscans obtained with and scans obtainedwithout the ACS technique. The volumeCT dose index represents the averagedose within a scan volume (relative tothat of a standardized CT phantom) andis now required to be displayed on theuser interface of the CT scanner. The CTdose index is given in milligrays. While itis not the dose to any specific patient, it is

a standardized index of the average dosedelivered from the scan series. Thus, theuser can select the suggested tube cur-rent–time product or individual settingsat his or her discretion.

CLINICAL USE OF ATCMTECHNIQUES

From a practical standpoint, bringingATCM techniques into everyday CT scan-ning means that most currently usedscanning protocols, which involve man-ual selection of tube current for radiationdose optimization and image quality,need to be modified. Basic differences inATCM techniques necessitate separateprotocols for each technique. In this sec-tion, we shall discuss the protocols forroutine CT scanning with individualATCM techniques.

Z-Axis Modulation

The AutomA z-axis–modulation tech-nique is offered with recently availablemulti–detector row CT scanners (Light-Speed; GE Medical Systems). Althoughangular modulation has been comparedwith fixed-tube-current techniques, todate there are no studies of which we areaware in which angular modulation hasbeen compared with z-axis modulationin terms of image quality or radiationdose efficiency (18–20).

In scanners with the AutomA feature,the operator can use either a fixed-tube-current technique or a z-axis–modulationtechnique. In fixed-tube-current scan-ning protocols, the technologist selects asuitable tube current for all examinations

on the basis of his or her judgment ordepartmental guidelines. With this z-axis–modulation technique, instead ofselecting a fixed tube current the tech-nologist selects a noise index and a rangeof acceptable tube current (minimumand maximum milliampere) settings. Asdescribed in the preceding section, thenoise index determines image quality.No unit of measure has been assigned tothe noise index, but because it approxi-mates the standard deviation (in Houns-field units) on CT images of a phantom,the noise index can be expressed inHounsfield units. From a practical stand-point, the technologist selects the noiseindex and the acceptable range of tubecurrent settings for the scan (Fig 2b). A5% decrease in noise index (in Houns-field units) typically increases radiationexposure by 10% and vice versa.

Determination of the range of tubecurrent values can sometimes be influ-enced by patient habitus. Because noiseindex and tube current range affect im-age quality and radiation exposure to thepatient, these two parameters must bejudiciously selected. A very low noise in-dex may provide higher image qualitybut will also result in higher than neces-sary radiation exposure to the patient.Conversely, a higher noise index will re-sult in radiation dose reduction at theprice of noisier images. Thus, with theAutomA technique, radiation exposuredepends on the selected noise index andpatient size. Higher radiation exposurecan be avoided by selecting a highernoise index or by setting the maximumtube current parameter to the same levelused with fixed-tube-current protocols.

Figure 4. Transverse CT scans obtained with AutomA technique (noise index, 15 HU; 75–380 mA; 140 kVp; gantry rotation time, 0.5 second) ina 32-year-old woman with treated lymphoma. (a) Chest image obtained at 136 mA shows satisfactory image quality with soft-tissue algorithm.(b) Pelvic image obtained at 295 mA shows that noise in soft tissue is similar to that in a.

Volume 233 � Number 3 Automatic Tube Current Modulation � 653

Ra

dio

logy

Although image noise will increase in re-gions where the required tube current islimited by the maximum tube current,overall image quality would be similar tothat with fixed-tube-current scanning.The vendor prescribes a noise index of11–12 HU for routine abdominal and pel-vic examinations and 10–11 HU for chestexaminations, with a minimum tube cur-rent of 60–80 mA.

The protocol used at our institution forroutine chest, abdominal, and pelvic CTexaminations with a 16–detector row CTscanner and the AutomA technique ofz-axis modulation is summarized in Ta-ble 1. Once the user has determined thedesired parameter adjustments (includ-ing noise index), the parameters can bestored in the computer memory and re-called without need for modification insimilar clinical cases.

Although the scanner automaticallymodulates tube current for a selectednoise index, the user should be aware ofsome important implications of the tech-nique. In patients with large cross-sec-tional dimensions, the scanner may notprovide images with the selected noiseindex because the maximum tube cur-rent parameter may be less than the tubecurrent required to generate images withthat noise index. In this situation, theuser should prospectively preview theamperage table on the console monitorbefore initiating scanning to determinewhether the maximum tube current limithas been reached (Fig 2). If the maximumtube current limit is reached with a 0.5-second rotation, the user can increase thegantry rotation time or use higher peakkilovoltage (140 instead of 120 kVp).

Conversely, with small patients, partic-ularly when not correctly positioned,ATCM techniques can result in excessivereduction in tube current and very noisyimages. Instructions to technologists aboutthe necessity of proper centering of pa-tients in the gantry isocenter and increas-ing the limit for the minimum tube cur-rent parameter usually help limit theprobability that noisy images will be ob-tained in small patients. Selecting a lowernoise index for small patients while cap-ping the maximum tube current parameterto avoid higher than necessary radiationexposure can also help improve imagequality (5).

Authors of a recent study (4) foundthat the AutomA technique for pelvicand abdominal CT resulted in a meanreduction in tube current–time productof 33% (range, 1%–91%), compared withmanual selection of tube current, in 87%(54 of 62) of patients. Compared with

manual selection of a fixed tube current,CT examinations of abdomen and pelvisperformed with the z-axis–modulationtechnique provided images with similarnoise, diagnostic acceptability, and le-sion detectability. In a study of 153 pa-tients who underwent abdominal-pelvicCT with the AutomA technique (5), agreater reduction in mean radiation dosewas noted in smaller patients (meanweight � 72 kg � 17 [standard devia-tion], mean anteroposterior abdominaldiameter � 23.2 cm � 3.5, mean trans-verse diameter � 30.9 cm � 3.8) than inlarger patients (weight � 82 kg � 16,anteroposterior diameter � 26.9 cm �3.9, transverse diameter � 34 cm � 3.4).Findings from a recent phantom study(6) showed that kidney stones smallerthan 5 mm can be adequately evaluatedby using AutomA technique, with 56%–77% reduction in radiation dose relativeto the dose from a fixed-tube-currenttechnique. In addition, the authors ofthat study reported that patients withkidney stones can be adequately scannedwith a noise index of 20 HU.

With the Real E.C. technique of z-axismodulation, the noise level can be incor-porated into the examination protocol,

thus relieving the operator of any need toapply the technique to individual exam-inations. Should any protocol need ad-justment to suit the needs of individualpatients, the Real E.C. technique auto-matically evaluates the effects on thenoise level of parameters such as peakvoltage, section thickness, helical pitch,and reconstruction algorithm and selectsthe required tube current to deliver thedesired standard deviation.

The availability of multiple choices ofimage noise is consistent with recent at-tempts to reduce patient dose by carefulmatching of technique not only to pa-tient habitus but also to diagnostic needs.By incorporating the choices of imagequality and dose within the stored proto-col, the operator can make modificationsat the time of scanning to ensure thatradiation dose reduction is achievedwhile a specific image quality is gener-ated.

Angular Modulation

In many studies, angular modulation(eg, CARE Dose) has been reported tohelp reduce radiation exposure in phan-tom experiments and clinical studies in-

TABLE 2Protocol for CARE Dose Technique of Angular Modulation for RoutineAbdominal and Pelvic CT

Parameter Value

Tube current (effective mA) 200Gantry rotation time (sec) 0.5Voltage (kVp) 140CT pitch factor 1Table speed (mm per rotation) 24Detector configuration 16 rows, 1.25-mm-thick sections*Reconstructed section thickness (mm) 5

* Senstation 16; Siemens Medical Solutions.

TABLE 1Protocol for AutomA Technique of Z-Axis Modulation for Routine Chest,Abdominal, and Pelvic CT

Parameter Value

Noise index 15*Tube current (mA) 75–380Gantry rotation time (sec) 0.5†

Voltage (kVp) 140Beam pitch 0.94Table speed (mm per rotation) 18.75Detector configuration 16 rows, 1.25-mm-thick sections‡

Reconstructed section thickness (mm) 5

* Noise index of 20 used for unenhanced phase of CT and for assessment of kidney stones; indexof 55, for CT colonography; and index of 12, for liver transplants.

† Time increased in 0.1-second steps to prevent maximum tube current limit.‡ LightSpeed 4.X; GE Medical Systems.

654 � Radiology � December 2004 Kalra et al

Ra

dio

logy

volving both adults and children (18–20). With angular modulation, theeffective tube current–time product is thekey difference. Scanners with the angularmodulation technique offer users the op-tion of scanning with angular modula-tion or fixed tube current (Fig 1b). Afterselecting other scanning parameters suchas peak voltage, table feed, and detectorconfiguration, the user enters a value forthe effective tube current–time productfor the scan. The effective tube current–time product (mAseff) is defined as fol-lows: mAseff � (TC � GR)/PF, where TC isthe tube current (in milliamperes), GR is

the gantry rotation time (in seconds),and PF is the CT pitch factor.

The scanner maintains a constant ef-fective tube current–time product irre-spective of pitch value so that radiationdose does not vary as pitch is changed.This helps to maintain a constant imagequality as the pitch factor (table feed)changes. Consequently, an increase inpitch leads to an increase in tube current,and a decrease in pitch leads to a reduc-tion in tube current. After selection of theeffective tube current–time product, thescanner performs real-time modulationof tube current in the x-y plane. The se-

lected effective tube current–time prod-uct represents the maximum effectivevalue that will be used for scanning. Theangular-modulation technique reducesthe effective tube current–time productto a value less than that selected for pro-jection angles with a lower attenuationprofile. The final displayed effective tubecurrent–time product represents an aver-age of various effective values used indifferent projections (Fig 5).

Real-time online selection of the effec-tive tube current–time product with an-gular modulation enhances the radiationdose efficiency of the CT scanner. Theuser must remember that, like z-axismodulation, accurate centering of pa-tients in the scanning field of view iscritical with angular modulation, sinceshifts from the center can lead to errone-ous estimation of the projection anglearea and tube current modulation, result-ing in noisier images. The scanning pro-tocol used at our institution for routineabdominal and pelvic CT with angularmodulation is summarized in Table 2.

In initial investigations in patients inwhich online angular modulation wasused (3), authors have reported an aver-age reduction in tube current–time prod-uct of up to 38%, with less image noise in75% of CT studies, better low-contrastdetectability in 51%, and superior overallimage quality in 71%, compared with im-ages obtained with fixed-tube-currentprotocols. Authors of a subsequent study(17) in which angular modulation wasused for scanning six anatomic regions,including head, shoulder, thorax, abdo-men, pelvis, and extremities (knee), re-ported a 15%–50%-mAs reduction. Mas-tora et al (19) used an angular-modulationtechnique and reported dose reductionwith improved image quality for CT angio-grams of the thoracic outlet (which is oftenaffected by noise and artifacts when afixed-tube-current protocol is used). A10%–60% reduction in the tube current–time product, depending on patient ge-ometry and anatomic region, has beenreported in children scanned with angu-lar modulation with a mean effectivetube current–time product reduction of22.3% (20% for neck, 23% for thorax,23% for abdomen) (20).

The Dose-Right Dose Modulation tech-nique of ATCM uses the milliamperevalue of the default protocol as a near-upper (within 10%) bound and modu-lates the tube current during gantry rota-tion on the basis of the system’s priorattenuation measurements in cross-sec-tional anatomy. The Dose-Right DoseModulation technique produces an im-

Figure 5. Transverse CT images obtained with CARE Dose technique(140 kVp, 0.5-second rotation time, 1:1 pitch) in a 65-year-oldwoman evaluated for metastases from colon cancer. (a) Chest imageobtained at 177 mAs (effective) and (b) abdominal image obtained at188 mAs (effective) show satisfactory image quality.

Volume 233 � Number 3 Automatic Tube Current Modulation � 655

Ra

dio

logy

age signal-to-noise ratio and image qual-ity that are very close to those of the fullnonmodulated milliampere image qual-ity of the system but with marked reduc-tion in radiation dose. Within a particu-lar protocol, the user can select both ACSand Dose-Right Dose Modulation inde-pendently or together. As described in aprevious section, the ACS icon on theuser interface automatically suggests apatient-tailored tube current–time prod-uct setting; Dose-Right Dose Modulationthen modulates this setting downwardduring rotation to reduce dose withoutaffecting the desired image quality. Thevolume CT dose index and dose lengthproduct are displayed according to theprotocol tube current–time product andthe selected protocol parameters. Radia-tion dose reduction with Dose-RightDose Modulation ranges between 15%and 40%, depending on the patient’s sizeand the anatomy being scanned. If Dose-Right Dose Modulation is selected, theactual volume CT dose index and doselength product are displayed after thescan on the basis of the actual averagetube current–time product used for thescan, which aids the operator in monitor-ing reductions in the volume CT doseindex and dose length product by com-paring the prescan (programmed) andpostscan (actual) dose values.

LIMITATIONS OF AVAILABLEATCM TECHNIQUES

At our institution, both angular- and z-axis–modulation techniques have beenin use for over 18 months for routinescanning of patients. While each tech-nique has strengths, there are some lim-itations with each technique. With angu-lar modulation, the user has to prescribea specific tube current value, which intro-duces subjectivity in the selection of aneffective tube current–time product. Onoccasion in scanners with an angular-modulation technique, selection of highereffective tube current–time product valuesfor scanning of larger patients may requirean increase in the scanning time (espe-cially if the operator believes that selectedor allowed effective tube current–timeproduct will not result in optimum imagequality).

Because z-axis modulation is a recentinnovation, there is a noticeable lack ofscientific documentation in the medicalliterature regarding appropriate noise in-dexes for specific sizes or ages of patientsand specific clinical indications. In addi-tion to selecting manufacturer-recom-

mended noise indexes, however, userscan restrict the range of tube currentsthat can be used for scanning in order tocap the maximum dose to ensure a min-imum image quality. However, use of anarbitrary range of tube currents intro-duces subjective constraints to the tech-nique. A lower minimum tube currentmay result in reduced patient exposure,which occasionally results in noisier im-ages in small patients who are scanned ata substantially reduced tube current.Conversely, large patients occasionallyreceive higher tube current with z-axismodulation than they would receive if afixed-tube-current technique was used inorder to maintain the selected imagenoise. While ATCM results in better im-age quality in this setting, it can result ingreater radiation exposure than would afixed-tube-current protocol in large pa-tients.

The foremost limitation of ATCM isthe lack of uniformity between tech-niques developed by different vendors.With the addition of ATCM techniquesto the technologic revolution in CT scan-ners, user comprehension is confoundedby the separate protocols for systemsfrom different vendors. Industry stan-dards organizations need to build a con-sensus so that a more uniform ATCMtechnique is offered to the user in orderto minimize confusion and ensure appro-priate use of the technique. Further expe-rience and research with ATCM tech-niques, regardless of type, will aid inoptimization these techniques. In addi-tion, the introduction of ATCM to aninstitution requires close communica-tion between radiologists, medical phys-icists, and technologists, because a defi-nite learning curve exists.

Presently, no vendor offers a combinedATCM technique that uses both angularand z-axis modulation as complementaryapproaches for maximum radiation ben-efits, although industry expects to releasea combination technique by the end of2003. A work-in-progress installation (CareDose 4D; Siemens Medical Solutions),which combines the angular- and z-axis–modulation techniques, has been re-ported (22,23). This technique modulatesthe tube current within the section (an-gular modulation) and also effects achange of attenuation in different ana-tomic regions (z-axis modulation). Toadapt the tube current in the z-axis, thistechnique uses a database that correlatesimage quality at a specific effective tubecurrent–time product for different ana-tomic regions and a “standard” body withthe effective tube current–time product

that is necessary for individual patientbody habitus to achieve similar imagequality. For the examination, the user en-ters an effective milliampere-seconds set-ting for a “normal-sized” patient to ob-tain a specific expected image quality.The CARE Dose 4D protocol adapts thetube current to the patient’s individualanatomy and modulates the tube currentin the section and the z-axis to obtain thedesired image quality for all images atthe lowest dose levels. Initial results(22,23) have shown a 20%–60% dose re-duction, depending on the anatomic re-gion and patient habitus, with improvedimage quality.

Authors of another study (24) in whichcombined angular and z-axis modulation(3D Auto mA; GE Yokogowa Medical Sys-tems, Tokyo, Japan) was used have alsoreported dose reductions of 60% in ab-dominal-pelvic CT examinations. Thistechnique uses a single localizer radio-graph to determine patient asymmetryand appropriate angular and z-axis mod-ulation for the patient. The investigatorsadded noise (computer modification oforiginal raw scan data to simulate lowertube current noise levels) to patients’scan data to produce images and calcu-late the radiation dose reduction.

CONCLUSION

Currently available ATCM techniquescan be used to maintain acceptable imagequality while reducing radiation expo-sure on the basis of patient geometry andclinical indications. ATCM techniquesrepresent an exciting recent technologicinnovation toward radiation dose opti-mization. Further research is needed tostandardize these techniques and defineappropriate protocols for different pa-tient sizes and indications.

References1. Gies M, Kalender WA, Wolf H, Suess C.

Dose reduction in CT by anatomicallyadapted tube current modulation. I. Sim-ulation studies. Med Phys 1999; 26:2235–2247.

2. Kalender WA, Wolf H, Suess C. Dose re-duction in CT by anatomically adaptedtube current modulation. II. Phantommeasurements. Med Phys 1999; 26:2248–2253.

3. Greess H, Wolf H, Baum U, Kalender WA,Bautz W. Dosage reduction in computedtomography by anatomy-oriented atten-uation-based tube-current modulation:the first clinical results. Rofo Fortschr GebRontgenstr Neuen Bildgeb Verfahr 1999;170:246–250.

4. Kalra MK, Maher MM, Toth TL, KamathRS, Halpern EF, Saini S. Comparison ofz-axis automatic tube current modulation

656 � Radiology � December 2004 Kalra et al

Ra

dio

logy

technique with fixed tube current CTscanning of abdomen and pelvis. Radiol-ogy 2004; 232:347–353.

5. Kalra MK, Maher MM, Kamath RS, et al.Sixteen detector–row CT of abdomen andpelvis: study for optimization of z-axismodulation technique performed in 153patients (abstr). Radiology 2004; 233:241–249.

6. Kalra MK, Maher MM, D’Souza RV, et al.CT detection of urinary tract stones usingz-axis automatic tube current modulationtechnique with low radiation dose: phan-tom and clinical studies. Radiology (inpress).

7. Rehani MM, Bongartz G, Kalender W, etal. Managing x-ray dose in computed to-mography: ICRP special task force report.Ann ICRP 2000; 30:7–45.

8. Li F, Sone S, Abe H, MacMahon H, ArmatoSG 3rd, Doi K. Lung cancers missed atlow-dose helical CT screening in a generalpopulation: comparison of clinical, his-topathologic, and imaging findings. Radi-ology 2002; 225:673–683.

9. Kalra MK, Blake MA, Maher MM, et al.Does low dose abdominal-pelvic CT com-promise the lesion detection and mor-phological appearance? (abstr). AJR Am JRoentgenol 2003; 180(suppl):8.

10. Donnelly LF, Emery KH, Brody AS, et al.Minimizing radiation dose for pediatricbody applications of single-detector heli-cal CT: strategies at a large children’s hos-pital. AJR Am J Roentgenol 2001; 176:303–306.

11. Kalra MK, Maher MM, Toth TL, et al.Strategies for CT radiation dose optimiza-tion. Radiology 2004; 230:619–628.

12. Kalra MK, Prasad S, Saini S, et al. Clinicalcomparison of standard-dose and 50% re-duced-dose abdominal CT: effect on im-age quality. AJR Am J Roentgenol 2002;179:1101–1106.

13. Kopka L, Funke M, Breiter N, HermannKP, Vosshenrich R, Grabbe E. An anatom-ically adapted variation of the tube cur-rent in CT: studies on radiation dosagereduction and image quality. RofoFortschr Geb Rontgenstr Neuen BildgebVerfahr 1995; 163:383–387.

14. Lehmann KJ, Wild J, Georgi M. Clinicaluse of software-controlled x-ray tubemodulation with “Smart-Scan” in spiralCT. Aktuelle Radiol 1997; 7:156–158.

15. Giacomuzzi SM, Erckert B, Schopf T, et al.The smart-scan procedure of spiral com-puted tomography: a new method fordose reduction. Rofo Fortschr Geb Ront-genstr Neuen Bildgeb Verfahr 1996; 165:10–16.

16. Kalender WA, Wolf H, Suess C, Gies M,Greess H, Bautz WA. Dose reduction inCT by on-line tube current control: prin-ciples and validation on phantoms andcadavers. Eur Radiol 1999; 9:323–328.

17. Greess H, Wolf H, Baum U, et al. Dosereduction in computed tomography byattenuation-based on-line modulation oftube current: evaluation of six anatomi-cal regions. Eur Radiol 2000; 10:391–394.

18. Greess H, Baum U, Wolf H, et al. Dose

reduction in spiral-CT: detection of pulmo-nary coin lesions with and without ana-tomically adjusted modulation of tube cur-rent. Rofo Fortschr Geb Rontgenstr NeuenBildgeb Verfahr 2001; 173:466–470.

19. Mastora I, Remy-Jardin M, Suess C, ScherfC, Guillot JP, Remy J. Dose reduction inspiral CT angiography of thoracic outletsyndrome by anatomically adapted tubecurrent modulation. Eur Radiol 2001; 11:590–596.

20. Greess H, Nomayr A, Wolf H, et al. Dosereduction in CT examination of childrenby an attenuation-based on-line modula-tion of tube current (CARE dose). Eur Ra-diol 2002; 12:1571–1576.

21. Westerman BR. Radiation dose fromToshiba CT scanners. Pediatr Radiol2002; 32:735–737.

22. Greess H, Wolf H, Suess C, Lutze J,Kalender WA, Bautz W. Automatic expo-sure control to reduce dose in subsecondmultislice spiral CT: phantom measure-ments and clinical results (abstr). Radiol-ogy 2002; 225(P):593.

23. Suess C, Chen XY. Dose optimization inpediatric CT: current technology and fu-ture innovations. Pediatr Radiol 2002; 32:729–734.

24. Horiuchi T. Study on 3D modulationAuto mA (abstr). Jpn Soc Radiol Technol2002; 78(P):166.

Volume 233 � Number 3 Automatic Tube Current Modulation � 657

Ra

dio

logy