ct fluoroscopy–guided interventional procedures ...€¦ · a promise of this technology is to...

Erik K. Paulson, MDDouglas H. Sheafor, MDDavid S. Enterline, MDH. Page McAdams, MDTerry T. Yoshizumi, PhD

Index terms:Computed tomography (CT),

guidance, 70.12119Computed tomography (CT),

radiation exposure, 70.12119Fluoroscopy, technologyPhantomsRadiations, exposure to patients and

personnel, 70.12119Radiations, measurement, 70.12119

Radiology 2001; 220:161–167

1 From the Department of Radiology,Duke University Medical Center, ErwinRd, Box 3808, Durham, NC 27710.Received September 11, 2000; revi-sion requested October 27; revisionreceived January 15, 2001; acceptedJanuary 23. Address correspondenceto E.K.P. (e-mail: [email protected]).© RSNA, 2001

Author contributions:Guarantor of integrity of entire study,E.K.P.; study concepts, E.K.P., T.T.Y.;study design, E.K.P.; literature research,E.K.P.; clinical studies, E.K.P., D.H.S.,D.S.E., H.P.M.; experimental studies,T.T.Y.; data acquisition, E.K.P.; dataanalysis/interpretation, T.T.Y., E.K.P.;manuscript preparation and definitionof intellectual content, E.K.P.; manu-script editing and revision/review, allauthors; manuscript final version ap-proval, E.K.P.

CT Fluoroscopy–guidedInterventional Procedures:Techniques and RadiationDose to Radiologists1

PURPOSE: To determine the radiation dose to radiologists who perform computedtomographic (CT) fluoroscopic interventional procedures by using a quick-checkmethod and a low-milliampere technique.

MATERIALS AND METHODS: Two hundred twenty CT fluoroscopy–guided inter-ventional procedures were performed in 189 patients. Procedures included 57spinal injections, 17 spinal biopsies, 24 chest biopsies, 20 abdominal aspirations, 44abdominal biopsies, and 58 abdominal drainages. Procedure details were prospec-tively recorded and included site, depth, target diameter, milliampere value, kilovoltpeak, fluoroscopic time, and CT technique (continuous CT fluoroscopy, quick-checkmethod, or a combination of these techniques). An individual collar and fingerradiation detector were worn by each radiologist during each procedure to deter-mine the dose per procedure.

RESULTS: The quick-check technique was performed in 191 (87%) of 220 proce-dures. Four procedures were performed with continuous CT fluoroscopy, and acombination technique was used for 25 (11%) procedures. The overall mean CTfluoroscopic time was 17.9 seconds (range, 1.2–101.5 seconds). The mean milli-ampere value was 13.2 mA (range, 10–50 mA). The overall mean radiologistradiation dose per procedure was 2.5 mrem (0.025 mSv) (whole body). Individualprocedure doses ranged from 0.66 to 4.75 mrem (0.007–0.048 mSv). The fingerradiation dose was negligible.

CONCLUSION: By using a low-milliampere technique and the quick-check method,CT fluoroscopic time and radiation exposure can be minimized.

Computed tomographic (CT) fluoroscopy is a technical advance resulting from slip-ringtechnology, x-ray tubes with improved heat capacity, high-speed array processors, andpartial reconstruction algorithms (1,2). The images can be reconstructed at a rate ofapproximately 6 frames per second, allowing near real-time visualization similar to that ofultrasonography (US). A promise of this technology is to facilitate interventional proce-dure guidance by means of combining the localizing strengths of CT with the real-timeadvantages of US.

Findings of recent clinical studies (3–6) have shown that CT fluoroscopy is a safe andeffective guidance tool for percutaneous interventional procedures in the chest, spine,abdomen, and pelvis. With this technology, procedures are performed more quickly thanwith traditional CT (7). CT fluoroscopy is particularly useful for procedures involving deepstructures, such as retroperitoneal masses, or for procedures involving organs prone tophysiologic motion, including the liver and lungs (3,6).

One of the concerns with the use of CT fluoroscopy is the high radiation exposure(4,5,8). In contrast with conventional fluoroscopy in which the patient dose is on theorder of centigrays per minute of exposure, with CT fluoroscopy, patient doses may be onthe order of centigrays per second. An additional concern is the scattered exposure to thehands and body of radiologists, since they may be close to the x-ray source during themanipulation of the needle (9).

During the next few years, the number of procedures performed with CT fluoroscopy

161

may increase. Because a greater numberof physicians use CT fluoroscopy, it iscritical to determine the radiation dose topatients and radiologists and to exploremethods to limit the dose. Our purposewas to determine the radiation dose toradiologists during CT fluoroscopic pro-cedures by using a low-milliampere tech-nique and short CT fluoroscopic expo-sures.

MATERIALS AND METHODS

From May 1, 1999, to September 20,1999, we prospectively monitored the ra-diation dose to radiologists performinginterventional procedures with CT fluo-roscopy. Our institutional review boarddeemed this project exempt from its re-view.

During the study period, 220 consecu-tive CT fluoroscopy–guided interven-tional procedures were performed in 189patients. One hundred sixty patients un-derwent one procedure, 27 patients un-derwent two procedures, and two patientsunderwent three procedures. Interven-tional procedures included spinal, chest,and abdominal and pelvic procedures asfollows: 40 lumbar, 13 sacroiliac, andfour cervical injections for neurolysis; 17biopsies of spinal lesions; 24 chest biop-sies; 44 biopsies in the abdomen and pel-vis; 20 fluid aspirations in the abdomenor pelvis; and 58 catheter drainages offluid collections in the abdomen or pel-vis. There were 98 women and 91 men.Their mean age was 58 years (age range,24–80 years).

The procedures were performed orclosely supervised by one of 12 attendingradiologists (including E.K.P., D.H.S.,D.S.E., H.P.M.) who were stationed at theside of the CT table during CT fluoros-copy. These 12 radiologists had a mean of9 years (range, 3–15 years) of experienceas attending staff. For 207 (94%) of 220procedures, a senior resident or fellowprovided assistance, but the individualradiation dose from CT fluoroscopy tothese individuals was not specificallymonitored. Individual doses were notmonitored in residents and fellows be-cause we estimated that they would per-form too few procedures to justify mon-itoring.

CT fluoroscopic images were acquiredduring continuous scanning with a ma-chine (CT/i equipped with SmartView;GE Medical Systems, Milwaukee, Wis),which was controlled by an integratedfoot switch and hand-held controller. Itsreconstruction duration was 6 frames per

second, which provided real-time guid-ance to facilitate patient positioning andneedle placement. In addition, individ-ual CT fluoroscopic spot images werepossible. The radiologist was able to re-lease the table to facilitate patient posi-tioning; the scanning plane was high-lighted with a laser marker light. Gantrytilt was possible. The monitor was conve-niently suspended from the ceiling. Im-age reconstruction was performed with a256 3 256 matrix, with images displayedon a 768 3 768 matrix.

Patients provided routine written in-formed consent for each interventionalprocedure. For all procedures, a prelimi-nary scan limited to the region of thetarget was obtained, with the patient po-sitioned appropriately for the interven-tional procedure. The target in questionwas localized on the preliminary scans,and the anticipated path and depth ofthe target were determined by using elec-tronic calipers on the technologist’s con-sole. The anticipated skin entry site wasmarked, prepared with providone iodinesolution (Betadine; Purdue Frederick,Norwalk, Conn), and draped in routinesterile fashion. The soft tissues were anes-thetized with lidocaine hydrochloride(Abbott Laboratories, North Chicago, Ill).Many patients received conscious seda-tion with appropriate monitoring. Themilliampere value was set at 10 mA un-less the target was small or the lesion wassubtle, as in some liver lesions. The deci-sion to increase the milliampere valuewas made at the discretion of the radiol-ogist.

After the site was prepared and draped,one of the three CT fluoroscopic tech-niques was used: the quick-check tech-nique, continuous CT fluoroscopy, or acombination of these techniques (5).One of these three techniques was cho-sen on the basis of radiologist preference.

The quick-check technique is analo-gous to conventional CT and is most fre-quently used in our practice. This tech-nique used single-section CT fluoroscopicspot images. The biopsy or the aspirationneedle was advanced through the skin tothe target. By moving the table, the nee-dle was aligned with the plane of imagingby using the laser marker. Single CT flu-oroscopic spot images were then ac-quired to check needle location and toconfirm appropriate alignment. The nee-dle position was adjusted accordingly.Once the single CT fluoroscopic spot im-ages helped to confirm that the needlewas at the appropriate position and tra-jectory, the needle was advanced to thetarget.

When necessary, repeat CT fluoro-scopic spot images were acquired to con-firm needle location, as determined bythe radiologist. In cases in which needlepositioning was difficult, such as those inwhich the target was deep, small, or pre-cariously located, several single CT fluo-roscopic spot images may have been re-quired. Once the needle tip was positionedin the target, an additional CT fluoro-scopic spot image was obtained to docu-ment the needle tip location. For thistechnique, continuous CT fluoroscopywas not used. Thus, this technique isanalogous to guidance with conven-tional CT, except reconstruction timesare faster and the table may be manuallypositioned by the radiologist.

When the catheter drainage proce-dures were performed by using the quick-check technique, guide-wire placement,stepwise dilation, and catheter place-ment were performed by using the CTfluoroscopic spot technique. For these pro-cedures, multiple CT fluoroscopic spot im-ages were often obtained to place, con-firm, and document the appropriatelocation of the needle, guide wire, andcatheter.

Continuous CT fluoroscopy denotesthe use of a continuous CT fluoroscopicexposure during needle advancement orneedle manipulation. For this technique,patients were moved into the CT gantryso that the lesion or needle was in theplane of imaging. For these procedures, a24-cm metallic sponge forceps was usedas a needle holder to prevent direct handexposure to the primary beam. Needleswere advanced by using real-time CT flu-oroscopic monitoring.

The combination technique was usedin cases in which the quick-check tech-nique failed to depict the needle tip. Inthese cases, patients were either manu-ally or mechanically moved through theplane of imaging until the needle or tar-get became visible. With this technique,there was no real-time needle manipula-tion. The video playback function wasused to move through the acquisition toconfirm needle placement.

A lead shield was not used during ma-nipulation of the needles during CT flu-oroscopy. In no case was there direct ex-posure of the radiologist’s hand to theprimary beam during needle manipula-tion.

For each CT fluoroscopic procedure, adata sheet was prospectively completedby the radiologist; the sheet includedqueries in regard to the procedure site,target depth, type of procedure, targetdiameter, CT fluoroscopic section thick-

162 z Radiology z July 2001 Paulson et al

ness, kilovolt peak, milliampere value,and fluoroscopic time. For each proce-dure, details of the use of CT fluoroscopywere collected; these included whetherthe quick-check technique, continuousreal-time CT fluoroscopy, or a combina-tion of both techniques was used.

A set of individualized commerciallyavailable radiation dosimeters (Landauer,Glenwood, Ill) was made available toeach attending radiologist who partici-pated in CT fluoroscopy. The detectorswere housed in a cabinet convenientlylocated in the CT fluoroscopic controlarea. Each radiologist was instructed towear his or her radiation detectors duringeach CT fluoroscopic procedure. At the

end of each procedure, the detectors werereturned to the cabinet to be retrieved forthe next procedure. For whole-body–dosemonitoring, a dosimeter (Luxel; Land-auer) was placed on the collar outside alead apron. This dosimeter allows estima-tion of the doses to the whole body, skin,and lens of the eye. For hand-dose mon-itoring, a thermoluminescence ring do-simeter was used.

Minimal detectable radiation doseswere 1 mrem (0.01 mSv) for dosimetersand 30 mrem (0.3 mSv) for the thermolu-minescence ring monitors, respectively.The cumulative dose from these dedi-cated detectors represented the dose fromCT fluoroscopy alone and did not reflectthe dose received from other exposuresin the radiology department. The meandose per procedure was determined bydividing the cumulative dose by thenumber of procedures. Specific case-by-case procedure doses were not measured;only the dose accumulated during thestudy was measured.

The whole-body dose refers to the ex-ternal whole-body exposure and was de-fined as the dose equivalent at a tissuedepth of 1 cm (1,000 mg/cm2). The skindose referred to the external exposure ofthe skin and was defined as the doseequivalent at a tissue depth of 0.007 cm,averaged over an area of 1 cm2.

To estimate direct radiation doses topatients, a body phantom (Rando; Phan-tom Laboratory, Salem, NY) was used tomeasure the surface dose rate and thescatter dose rate. The body phantom waschosen because it was equivalent to thetissue. Dose rates were determined by us-ing our typical clinical technique with140 kV and 10 mA. The value of 140 kV

was chosen on the basis of the manufac-turer recommendations. Surface dose rateswere measured at the surface by usingthermoluminescence dosimeter chips.The chips were 3 3 3 3 1 mm and werearranged in a linear pattern with no gapsbetween them. Two consecutive mea-surements were obtained. A commercialvendor (Landauer) was used for dosime-ter chip analysis.

Exposures were estimated from scat-tered radiation from the phantom. Anion chamber (model 450P; Victoreen, So-lon, Ohio) was used to measure scatteredradiation levels at 25 and 60 cm from thesection plane at the edge of the patienttable (Fig 1). These distances were chosento simulate the position of the radiolo-gist’s hands and body. Dose rates werecalculated by dividing the total cumula-tive exposure by the exposure time (10seconds).

RESULTS

The mean target depth (measured fromthe skin to the leading edge of the le-sion), target diameter, and technical as-pects are detailed in Table 1.

One hundred ninety-one (87%) of 220procedures were performed by using thequick-check method alone. Twenty-five(11%) procedures were performed by us-ing a combination of the quick-checkmethod and continuous CT fluoroscopy.Only four (2%) of the procedures wereperformed with needle manipulation byusing continuous CT fluoroscopy.

Of the 12 radiologists who performedCT fluoroscopy, six used the quick-checktechnique exclusively, four used eitherthe quick-check technique or the quick-

TABLE 1Mean Values of Technical Aspects of Procedures Performed with CT Fluoroscopy

Procedure Target Diameter (cm) Target Depth (cm) Section Thickness (mm)Value(mA) Fluoroscopic Time (sec)

Spinal injectionLumbar (n 5 40) 0.7 (0.2–1.0) 8.2 (4.0–13.1) 3.0 (3.0–3.0) 16.7 (10.0–50.0) 17.6 (3.6–65.0)Cervical (n 5 4) 0.3 (0.2–0.4) 5.4 (4.0–7.6) 3.0 (3.0–3.0) 16.4 (10.0–30.0) 18.4 (12.0–27.6)Sacroiliac joint (n 5 13) 0.2 (0.2–0.3) 6.4 (3.3–10.0) 3.0 (3.0–3.0) 16.4 (10.0–30.0) 11.0 (6.0–19.2)

Spinal biopsyLumbar (n 5 7) 1.8 (0.3–4.0) 7.8 (3.0–11.9) 3.6 (3.0–5.0) 12.9 (10.0–20.0) 20.1 (4.8–31.2)Thoracic (n 5 9) 1.9 (1.0–3.5) 6.1 (1.0–8.6) 3.0 (3.0–3.0) 15.0 (10.0–20.0) 26.4 (2.4–52.0)Sacroiliac joint (n 5 1) 2.0 6.5 3.0 10.0 19.2

Chest biopsy (n 5 24) 3.1 (1.0–8.0) 6.2 (2.0–11.0) 9.1 (5.0–10.0) 10.8 (10.0–20.0) 17.2 (1.2–43.0)Abdominal

Aspiration (n 5 20) 4.4 (3.0–8.5) 7.9 (1.0–16.0) 8.7 (5.0–10.0) 12.6 (10.0–50.0) 13.7 (1.2–54.1)Biopsy (n 5 44) 3.8 (1.0–10.2) 8.5 (2.2–15.0) 8.3 (5.0–10.0) 16.2 (10.0–20.0) 21.0 (1.2–101.5)Drainage (n 5 58) 5.9 (2.0–20.0) 6.7 (2.0–15.0) 8.7 (5.0–10.0) 11.2 (10.0–30.0) 16.9 (1.2–93.0)

All (n 5 220) 3.4 (0.2–20.0) 7.4 (1.0–16.0) 6.7 (3.0–10.0) 13.2 (10.0–50.0) 17.9 (1.2–101.5)

Note.—All procedures were performed with 140 kVp. Numbers in parentheses are ranges.

Figure 1. Diagram shows detector placementfor measurements of scattered radiation fromthe phantom. A and B denote the placement ofthe ion chamber detectors 25 and 60 cm fromthe section plane. These distances were chosento simulate the position of the radiologist’shand and body. Exposures were measuredfrom scattered radiation from the phantom(Rando phantom, Phantom Laboratory).

Volume 220 z Number 1 CT Fluoroscopy-guided Procedures: Radiation Dose z 163

check technique with continuous CT flu-oroscopy, and two used all three tech-niques at least once.

Not surprisingly, the frequent use ofthe quick-check method resulted in shortCT fluoroscopic times (Fig 2). Overall, themean CT fluoroscopic time was 17.9 sec-onds per procedure (range, 1.2–101.5 sec-onds). Eighty-six (39%) of 220 procedureshad a CT fluoroscopic time of shorter than10 seconds. Seventy-two (33%) of the pro-cedures had a CT fluoroscopic time of10–20 seconds. The remaining 62 (28%)procedures had a CT fluoroscopic timelonger than 20 seconds. Mean CT fluoro-scopic times for individual radiologistsranged from 4.8 to 37.6 seconds per proce-dure. The longest CT fluoroscopic time was101.5 seconds, which occurred during adifficult biopsy of a 2-cm lesion in thedome of a liver.

Overall, the mean milliampere valuewas 13.2 mA (range, 10–50 mA) (Fig 3).The low mean milliampere value resultedin adequate image quality for the effi-cient performance of most procedures.One hundred seventy-two (78%) of 220procedures were performed with 10 mA,38 (17%) were performed with 20 mA,and 10 (5%) were performed with morethan 20 mA. While in this study we didnot measure lesion conspicuity as a func-tion of the technique, the cases in whichthe milliampere value increased morethan 10 mA were in targets that weresubtle or small, requiring a higher-expo-sure technique to achieve adequate im-age quality. For example, the mean mil-liampere value of spinal interventionalprocedures was 16.5 mA, considerablyhigher than the mean milliampere value ofcatheter drainages, which was 11.2 mA.This likely reflects the differences in targetsize; spinal facets are a considerably smallertarget than fluid collections, which had amean diameter of 5.9 cm. Obese patientsalso required a higher milliampere valuefor adequate image quality.

Overall, the mean radiation dose perprocedure (by radiologist) were as fol-lows: 0.66–4.75 mrem (0.007–0.048 mSv),whole body; 0.7–4.75 mrem (0.007–0.048mSv), lens of the eye; and 0.79–5.38mrem (0.008–0.054 mSv), skin (Table 2).The overall mean dose per procedure av-eraged for all radiologists was 1.03 mrem(0.010 mSv), 1.03 mrem (0.010 mSv), and1.19 mrem (0.012 mSv) for the wholebody, lens of the eye, and skin, respec-tively. There was no radiologist in whomthe cumulative radiation dose to thehand was greater than the minimal de-tectable value of 30 mrem (0.3 mSv) orless for the ring monitor.

The maximum dose rate on the surfaceof the phantom determined from thethermoluminescence detector profiles was0.18 cGy/sec 6 .003 and 0.177 cGy/sec 6.003. Figure 4 shows the dose- rate profileas a function of the thermoluminescencedetector position on the basis of onemeasurement of dose rate.

The scattered radiation exposure rate(peak value) at 25 and 60 cm from thecenter of the section was 23 mR/h (5.9 31026 C/kg/h) and 12 mR/h (3.0 3 1026

C/kg/h), respectively. These dose rateswere obtained from the phantom.

DISCUSSION

Conventional CT has been widely usedand has been proven to be safe and effec-

tive in the performance of percutaneousinterventional procedures (10–13). Incontrast with US or fluoroscopy, conven-tional CT-guided procedures were limitedby the lack of real-time capability. Theimaging steps required to monitor anddocument needle placement are oftentime-consuming. In many cases, severalpasses are required to satisfactorily posi-tion needles to the lesion. CT-guided pro-cedures are particularly challenging inthe uncooperative patient or in organsthat are prone to respiratory motion,such as the lung and liver.

Katada et al (1) described CT fluoros-copy for clinical use in 1994. This tech-nology combines the advantages of CTwith needle depiction with the speed, ac-curacy, and elegance of real-time US or

Figure 2. Transverse CT fluoroscopic (140 kV, 10 mA, 5-mm section thickness) images obtainedin a 75-year-old man. (a) Image shows a fluid collection (arrows) following appendectomy.CT-guided drainage was requested. Following localization of the fluid collection, the quick-checktechnique was used to guide the placement of an 18-gauge needle (arrowhead) into the collec-tion. At this point in the procedure, cumulative CT fluoroscopic time was 2.4 seconds. (b) Imageobtained after the placement of the needle (arrowhead) into the purulent fluid. At this point inthe procedure, cumulative CT fluoroscopic time was approximately 4 seconds. (c) Image depictsthe appropriate location of the guide wire (arrowheads). Cumulative CT fluoroscopic time at thispoint was 6 seconds. (d) Following stepwise dilation, a 14-F pigtail catheter (arrow) was placedinto the abscess. Total CT fluoroscopic time for this procedure was 8 seconds.


conventional fluoroscopic guidance. Oneof the concerns of performing CT fluo-roscopy is the radiation to patients andoperators (4,5,8,9). While radiologists are

knowledgeable regarding the radiationdoses encountered during conventionalfluoroscopy, the doses from CT or CTfluoroscopy are considerably higher due

to the higher-exposure techniques re-quired to achieve acceptable image qual-ity. In the worst-case scenario of selectingthe highest-milliampere technique (90mA, 120 kVp, 10-mm section thickness)permissible by one manufacturer, Nawfelet al (8) have shown that patients re-ceived 830 mGy during an 80-second CTfluoroscopic exposure. Such a dose iscomparable with the patient dose re-ceived during 20–30 minutes of conven-tional fluoroscopy during cardiac cathe-terization (8). While the 830 mGy dose isless than the often-quoted 2,000 mGythreshold for inducing radiation derma-titis, it approaches 1,000 mGy, which is alevel at which radiation-induced skinchanges have been reported (14–16).

Recently Silverman et al (5) haveshown that CT fluoroscopy is at least aseffective as conventional CT in the per-formance of abdominal and pelvic biop-sies and catheter drainages. Comparedwith conventional CT, CT fluoroscopywas shown to be an efficient technique,with reductions in needle placementtime (29 vs 36 minutes, P , .005). In thework of Silverman et al, the most com-monly used CT technical parameterswere 120 kVp and 50 mA, but in selectedcases, a milliampere value as high as 90mA was used. The mean CT fluoroscopictime was 79 seconds per procedure andranged from 8 to 546 seconds. Silvermanet al provided an estimate of radiationexposure to operators by multiplying themean CT fluoroscopic times by the expo-sure rates of scattered radiation on thebasis of phantom measurements. Theseestimates suggest mean hand exposuresof 305 mR (7.9 3 1025 C/kg) per proce-dure and mean neck exposures of 10 mR(2.5 3 1026 C/kg) per procedure. A limi-tation of the Silverman et al study is that

TABLE 2Mean Scattered Radiation Dose per Procedure

AttendingRadiologist Procedure

Procedure (mrem)*

WholeBody†

Lensof Eye Skin‡

1 Facet injection (n 5 2), spine biopsy(n 5 4)

0.66 1.0 1.5


2.8 2.8 2.8


1.875 1.875 2.125

4 Chest biopsy (n 5 9) 1.0 1.0 0.895 Chest biopsy (n 5 15) 1.6 1.6 1.76 Abdominal biopsy (n 5 1), abdominal

drainage (n 5 11)2.0 2.0 2.1

7 Abdominal aspiration-biopsy (n 5 3),abdominal drainage (n 5 5)

4.75 4.75 5.38


4.2 4.2 4.5


1.35 1.35 1.47


0.9 0.9 1.0


0.74 0.74 0.81


0.7 0.7 0.8

All Not applicable 1.03 1.03 1.19

* To convert from milliroentgen-equivalent-man units to millisieverts, multiply by 0.01.† Whole-body dose refers to the external whole-body exposure and is defined as the dose

equivalent at a tissue depth of 1 cm.‡ Skin dose refers to the external exposure of the skin and is defined as the dose equivalent at a

tissue depth of 0.007 cm averaged over an area of 1 cm2.

Figure 3. Transverse images obtained in a 69-year-old man with back pain after Bilroth IIgastrojejunostomy. (a) Diagnostic CT image (140 kV, 210 mA, 5-mm section thickness, pitch of1.5, 0.8-second gantry rotation) shows a 2.0 3 1.8-cm low-attenuating lesion (arrows) in theuncinate process of the pancreas, with encasement (arrowhead) of the superior mesenteric arteryconsistent with pancreatic adenocarcinoma. (b) Quick-check CT fluoroscopic image (140 kV, 10mA, 5-mm section thickness) is inferior to a, but it is of sufficient quality to depict the pancreaticlesion (arrow), needle (arrowhead), and intervening structures. Only a portion of the needle isseen because the needle is oblique to the transverse plane. Pathologic analysis revealed pancreaticadenocarcinoma.

Figure 4. Graph shows the dose-rate profilefrom the phantom as a function of thermolu-minescence detector position. The maximumdose rate on the surface of the phantom was0.18 cGy/sec. The technical parameters were140 kV, 10 mA, and 10-mm section thickness.


scattered radiation to personnel was notmeasured directly.

In a similar study, Daly et al (6) exam-ined 97 patients who underwent 119 ab-dominal or pelvic procedures by usingeither continuous real-time or intermit-tent CT fluoroscopy during needle inser-tion. Technical parameters were 120–135kV and 30–80 mA. The mean CT fluoro-scopic time for biopsy was 136 seconds(range, 33–336 seconds). Each operatorwore standard dosimeters on the neck,waist, and finger, and cumulative doseswere measured monthly. The monthlydoses were 10–30 mrem (0.1–0.3 mSv) inthe neck, 0–10 mrem (0–0.1 mSv) for thedetectors worn beneath a 0.5-mm leadapron, 10–970 mrem (0.1–9.7 mSv)(mean, 176 mrem [1.76 mSv]) for thehands during the 1st month of opera-tion, and 10–130 mrem (0.1–1.3 mSv) forthe hands thereafter. A limitation of theDaly et al study is that individual proce-dure doses per physician were not mea-sured.

Our work is distinct from others’ inthat the milliampere value (mean, 13.2mA) was intentionally set as low as pos-sible for each procedure. In contrast, oth-ers report (4–6) using higher milliamperevalues, ranging from 30 to 90 mA, with50 mA being the most frequently usedvalue. Our use of a milliampere valuehigher than 10 mA was restricted to thoseprocedures in which the target was smallor the lesion was subtle, as in some liverlesions. We suggest that to minimize ra-diation exposure, the operators select thelowest milliampere value possible and in-crease it only when it proves inadequate.In theory, one could use a low milliam-pere value for gross needle positioningand increase the milliampere value forfinal needle placement. This compromisetechnique was not used in this study butcould potentially decrease the radiationdose. It should be noted that many fac-tors underlie the creation of a high-qual-ity CT fluoroscopic image. It can be an-ticipated that there will be differencesamong vendors in the radiation tech-nique required to produce adequate im-age quality.

Our results differ from those of others(3–6) in that our CT technique consistedprimarily of the quick-check method. Inthe minority of cases, the quick-checkmethod was combined with short expo-sures of continuous CT fluoroscopy toconfirm needle placement. In contrast,others (3–6) report using primarily con-tinuous CT fluoroscopy during needlemanipulation, which allows real-time vi-sualization but also necessarily increases

the CT fluoroscopic time. The quick-checktechnique is analogous to a conventionalCT technique in which the needle is ad-vanced in the absence of real-time visual-ization. Single-section CT fluoroscopic spotimages (1.2-second exposure) are used sim-ply to check needle placement.

The use of the quick-check method re-sults in short CT fluoroscopic times perprocedure. Our mean CT fluoroscopic timeof 17.9 seconds is considerably shorterthan the 80–143 seconds reported by oth-ers (4–6). One could argue that the quick-check technique is simplistic and fails toexploit the full benefits of continuous CTfluoroscopy. However, even this methodhas considerable advantages than the con-ventional CT technique, including use ofthe floating table to precisely position thepatient at the level of the needle tip, use ofhand and foot controls, and direct com-munication with patients.

When the quick-check technique isused, scenarios in which the needle tip isdifficult to localize, such as when theneedle is oblique to the transverse planeor when the patient cannot suspend res-piration are possible. In these scenarios,we identify the needle tip by acquiringcontinuous CT fluoroscopic images inthe region of the needle. Such acquisi-tions are usually only 5–10 seconds. Ingeneral, with careful attention to placingthe needle in the transverse plane, such acombined technique is used relatively in-frequently. In scenarios in which theneedles must be angled, tilting the CTgantry facilitates efficient visualization.

Our project included the estimation ofthe patient dose with a standard anthro-pomorphic phantom by using the tech-nical parameters most frequently used inour practice (140 kV, 10 mA). The maxi-mum dose rate on the surface of thephantom reported here (0.18 cGy/sec) isconsiderably lower than the 0.52 cGy/secreported by Nawfel et al (8) using 120kVp, 50 mA, and a 10-mm thick image.Given our mean CT fluoroscopic time of18 seconds per procedure, it can be esti-mated that patients receive approxi-mately 3.2 cGy per procedure.

The scattered radiation dose, as mea-sured from the phantom by using our rou-tine clinical parameters, was extremelylow: 23 mrem/h (0.23 mSv/h) at 25 cmfrom the section and 12 mrem/h (0.12mSv/h) at 60 cm from the section. Withthis dose rate, it would be necessary to per-form CT fluoroscopy for hundreds of hoursto exceed the occupational whole-bodylimit of 5,000 mrem (50 mSv) per year (17).

The physician dose per procedure wasalso low, with whole-body doses ranging

from 0.66 to 4.75 mrem (0.007–0.048mSv) per procedure. With this dose rate,to exceed the 5,000 mrem (50 mSv) bodydose per year, a radiologist would have toperform more than 2,000 CT fluoro-scopic procedures (on the basis of themean dose of 2.5 mrem [0.025 mSv] perprocedure).

The explanation of our low radiationdose rests in the following equation:dose ' kV2 3 mA 3 sec, where kV2 is thekilovolt value squared, mA is the milli-ampere value, and sec is seconds (17).Our CT fluoroscopic times were one-fourth to one-seventh those reported byothers (4–6). Additionally, our milliam-pere value of close to 10 mA was consid-erably less than the 30–90 mA reportedby others (4–6). Any decrease in the mil-liampere value or fluoroscopic time willdecrease the dose.

It should be emphasized that the scat-tered radiation dose to the operators re-ported in this study represents the doseoutside the protective lead gown. Withthe use of an appropriate lead apron, thy-roid shield, and protective eyewear, ac-tual doses should be negligible. Furthermeasures to reduce the dose include step-ping away from the gantry when possi-ble, preventing direct eye contact withthe gantry during exposure, and standingat the head side of the table. Recent work(8) indicates that placing a lead drapeover the patient further reduces the scat-tered dose.

In summary, we have shown that thepatient and physician radiation doses atCT fluoroscopy can be reduced to an ac-ceptably low level. To minimize the dose,we suggest limiting CT fluoroscopic timeby using the quick-check method and bydecreasing the milliampere value asmuch as possible.

References1. Katada K, Anno H, Ogura Y, et al. Early

clinical experience with real-time CT flu-oroscopy. Nippon Acta Radiol 1994; 54:1172–1174.

2. Katada K, Kato R, Anno H, et al. Guidancewith real-time CT fluoroscopy: early clin-ical experience. Radiology 1996; 200:851–856.

3. Meyer CA, White CS, Wu J, Futterer SF,Templeton PA. Real-time CT fluoroscopy:usefulness in thoracic drainage. AJR Am JRoentgenol 1998; 171:1097–1101.

4. Froelich JJ, Saar B, Hoppe M, et al. Real-time CT-fluoroscopy for guidance of per-cutaneous drainage procedures. J Vasc In-terv Radiol 1998; 9:735–740.

5. Silverman SG, Tuncali K, Adams DF, Naw-fel RD, Zou KH, Judy PF. CT fluoroscopy-guided abdominal interventions: tech-niques, results, and radiation exposure.Radiology 1999; 212:673–681.

6. Daly B, Krebs TL, Wong-You-Cheong JJ,


Wang SS. Percutaneous abdominal andpelvic interventional procedures usingCT fluoroscopy guidance. AJR Am JRoentgenol 1999; 173:637–644.

7. Sheafor DH, Paulson EK, Kliewer MA, De-Long DM, Nelson RC. Comparison ofsonographic and CT guidance tech-niques: does CT fluoroscopy decrease pro-cedure time? AJR Am J Roentgenol 2000;174:939–942.

8. Nawfel RD, Judy PF, Silverman SG,Hooton S, Tuncali K, Adams DF. Patientand personnel exposure during CT fluo-roscopy-guided interventional procedures.Radiology 2000; 216:180–184.

9. Kato R, Katada K, Anno H, Suzuki S, Ida Y,Koga S. Radiation dosimetry at CT fluo-roscopy: physician’s hand dose and de-velopment of needle holders. Radiology1996; 201:576–578.

10. Haaga JR, Alfidi RJ. Precise biopsy local-ization by computed tomography. Radi-ology 1976; 118:603–607.

11. Moran CJ, Naidich TP, Gado MH, Mar-chosky JA. Central nervous system le-sions biopsied or treated by CT-guidedneedle placement. Radiology 1979; 131:681–686.

12. Harter LP, Moss AA, Goldberg HI, GrossBH. CT-guided fine-needle aspirations fordiagnosis of benign and malignant dis-ease. AJR Am J Roentgenol 1983;140:363–367.

13. Mueller PR, vanSonnenberg E. Interven-tional radiology in the chest and abdo-men. N Engl J Med 1990; 322:1364–1374.

14. Shope TB. Regulations and recommenda-tions relevant to interventional radiol-ogy. In: Balter S, Shope T, eds. Syllabus: acategorical course in physics—physical

and technical aspects of angiography andinterventional radiology. Oak Brook, Ill:Radiological Society of North America,1995; 195–205.

15. Wagner LK, Eifel PJ, Geise RA. Potentialbiological effects following high x-raydose interventional procedures. J Vasc In-terv Radiol 1994; 5:71–84.

16. Hwang E, Gaxiola E, Vlietstra RE, BrennerA, Ebersole D, Browne K. Real-time mea-surement of skin radiation during cardiaccatheterization. Cathet Cardiovasc Diagn1998; 43:367–370.

17. Curry TS, Dowdey JE, Murry RC Jr, eds.Christensen’s physics of diagnostic radi-ology. 4th ed. Philadelphia, Pa: Lea & Fe-biger, 1990; 372–391.


ct fluoroscopy–guided interventional procedures ...€¦ · a promise of this technology is to...

Documents