Note: This copy is for your personal non-commercial use only. To … · 2019. 4. 11. · Abdominal...

Abdominal Wall CT Angiography:A Detailed Account of a NewlyEstablished Preoperative ImagingTechnique1

Timothy J. Phillips, MBBS, GradDipSurgAnatDamien L. Stella, MBBS, FRANZCRWarren M. Rozen, MBBS, BMedSci,

PGDipSurgAnatMark Ashton, MBBS, MD, FRACSG. Ian Taylor, MBBS, MD, FRCS, FRACS, AO

Institutional review board approval was obtained for thisstudy, and all patients gave written informed consent.Autologous surgical breast reconstruction with use of ab-dominal wall donor flaps based on the deep inferior epigas-tric artery (DIEA) and one or more of its anterior muscu-locutaneous perforating branches (DIEA perforator flap) isbeing used with increasing frequency instead of breastreconstruction with use of traditional transverse rectusabdominus musculocutaneous and modified muscle-spar-ing flaps. Preoperative mapping of the DIEA perforatorswith abdominal wall computed tomographic (CT) angiog-raphy may improve patient care by providing the surgeonwith additional information that will lead to optimization ofthe surgical technique, shorter procedure time, and reductionin the frequency of surgical complications. The branching pat-terns of the DIEA, the segmental anatomy of the anterioradipocutaneous perforating branches of the DIEA, and theimportance of these features in pre- and intraoperativesurgical planning necessitate a different approach to ab-dominal wall CT angiography than that used with otherabdominal CT angiographic techniques. In abdominal wallCT angiography, the common femoral artery is used as thebolus trigger, CT scanning is performed in the caudocra-nial direction, the automatic exposure control feature isdisabled, a scaled grid overlay tool is used to presentinformation to the surgeons, and radiation dose is mini-mized (average dose, 6 mSv). The anatomic accuracy ofabdominal wall CT angiography has been investigated incadaveric and surgical studies, with sensitivity of 96%–100% and specificity of 95%–100%. This detailed descrip-tion will allow other radiologists and surgeons interested infree DIEP flap surgery to incorporate this useful tool intotheir practice.

� RSNA, 2008

Supplemental material: http://radiology.rsnajnls.org/cgi/content/full/249/1/32/DC1

1 From the Department of Radiology, the Royal MelbourneHospital, Grattan Street, Parkville, Melbourne, Victoria3050, Australia. Received November 26, 2007; revisionrequested January 15, 2008; revision received March 21;final version accepted April 3. Address correspondenceto T.J.P. (e-mail: [email protected] ).

� RSNA, 2008

REVI

EWS

AND

COM

MEN

TARY

�HO

WID

OIT

32 Radiology: Volume 249: Number 1—October 2008

Note: This copy is for your personal non-commercial use only. To order presentation-ready copies for distribution to your colleagues or clients, contact us at www.rsna.org/rsnarights.

Autologous surgical breast recon-struction with abdominal wall do-nor flaps has evolved since its in-troduction in the late 1970s and early1980s (1,2). Free flaps based on thedeep inferior epigastric artery (DIEA)and one or more of its anterior muscu-locutaneous perforating branches

(hereafter, perforators) are termedDIEA perforator flaps, and they are be-ing used with increasing frequency toreplace the traditional transverse rectusabdominus musculocutaneous flaps andthe modified muscle-sparing flaps (3–6).The DIEA perforator flap is used to har-vest fat and skin (adipocutaneous tis-sue) without muscle and has resulted infewer abdominal wall complications—including hernia, bulge, and weakness(3,7–10)—and shorter hospital stays,with a cheaper overall cost of treatment(11,12) when compared with alterna-tive surgical techniques. The superficialinferior epigastric artery can also beused for an adipocutaneous muscle-sparing flap, and it is the preferred sur-gical option for an abdominal wall freeflap, as it does not require incision ofthe anterior rectus sheath and thus hasminimal donor site morbidity (13,14).However, this vessel is not uniformlypresent or of sufficient caliber to per-form the operation.

The branching pattern of the DIEAand the location of its perforators vary,and each patient’s anatomy dictates theDIEA and perforator(s) on which theflap will be based (15,16). PreoperativeDoppler ultrasonography (US) is usefulwhen defining the DIEA branching pat-tern; however, locating and choosingthe perforators was previously com-plete only after intraoperative abdomi-nal wall dissection. Identification of op-timal perforators on which to base theflap is highly desirable. There is evi-dence that flaps based on fewer perfora-tors have a lower incidence of fat necro-sis (7). Harvesting a flap supplied by oneperforator requires the identification ofa perforator with optimal characteris-tics (discussed in detail later in this arti-cle).

Abdominal wall computed tomo-graphic (CT) angiography is a new tech-nique tailored to assist in locating andcharacterizing the perforators, with re-ported sensitivity and specificity of96%–100% and 95%–100%, respec-tively, in cadaveric and clinical studies(17–23). In addition, this technique canbe used to define the branching patternof the DIEA (18,23). The intent of thisarticle is to provide those who are inter-

ested in abdominal wall CT angiographywith a step-by-step manual that de-scribes our approach to this modality.Included are considerations regardingappropriate patient selection, contrastmaterial enhancement, image acquisi-tion and reconstruction parameters, im-age display and analysis techniques, andthe radiology report, which includes thepresentation of data to the referringsurgeon. An explanation of the surgicaland radiologic anatomy of the DIEA andthe characteristics that define favorableperforators is also included. These rec-ommendations are based on the au-thors’ experience in performing and in-terpreting the results of 65 abdominalwall CT angiographic examinations andon close collaboration between the de-partments of radiology and plastic sur-gery at our institution.

Anatomy

DIEA ClassificationCadaveric and surgical studies havebeen performed to classify the branch-ing pattern of the DIEA. The branchingclassification first described by Moonand Taylor has been widely adopted andis shown in Figure 1 (15). In addition tothe major branching patterns (types I,II, and III) of the DIEA above the arcu-ate line, early branches also include apubic branch, muscular branches, andan umbilical branch (15).

Type I (single trunk) and type II (bi-furcation into two trunks) branchingpatterns occur more commonly thandoes the type III pattern (division intomore than two trunks). The other rele-vant characteristics are the caliber ofthe trunk(s) and any intramuscular

Published online10.1148/radiol.2483072054

Radiology 2008; 249:32–44

Abbreviations:CLUT � color look-up tableDIEA � deep inferior epigastric arteryMIP � maximum intensity projectionVRT � volume-rendered technique

Authors stated no financial relationship to disclose.

Essentials

� We have detailed the use of radio-logic classification of the segmen-tal anatomy of the anterior mus-culocutaneous perforatingbranches (hereafter, perforators)of the deep inferior epigastric ar-tery (DIEA) and the surgical rele-vance of this classification for freeadipocutaneous flap harvest.

� We have developed a CT angiog-raphy acquisition technique opti-mized for imaging the DIEA andthe anterior musculocutaneousbranches (perforators) supplyingthe abdominal wall.

� We have described CT angiogra-phy postprocessing techniquesoptimized for display of the surgi-cally relevant findings.

� We have demonstrated a volume-rendering color look-up table opti-mized for imaging of anteriormusculocutaneous branches ofthe DIEA in the subcutaneous fatof the abdominal wall.

� We have demonstrated a real-time volume-rendered image ma-nipulation reading technique thatenables preoperative mapping ofDIEA perforators.

� Autologous breast reconstructionwith use of adipocutaneous freeflaps (DIEA perforator flap) is theoptimal standard of care aftermastectomy.

� Preoperatively mapping DIEAperforators with this abdominalwall CT angiography techniquemay improve patient care by pro-viding the surgeon with additionalinformation that leads to optimi-zation of the surgical technique,shorter operative time, and a re-duction in the frequency of opera-tive complications.

HOW I DO IT: Abdominal Wall CT Angiography Phillips et al

Radiology: Volume 249: Number 1—October 2008 33

course. Usually, the DIEA enters therectus sheath superficial to the arcuateline and courses superiorly between theposterior layer of the rectus sheath andthe rectus muscle. It is not uncommonfor one or both of the DIEAs to pene-trate and course within the rectus mus-cle for some part of their length(15,24,25).

Segmental Arterial Anatomy of DIEAPerforators

The descriptions and nomenclature ap-plied to the variations in perforatoranatomy in cadaveric studies and surgi-cal observations are diverse, but the keyfindings are similar (24–26). The coursea perforator takes from its origin at theDIEA through four anatomic spaces toits end supply of the subcutaneous fatand skin of the anterior abdominalwall is shown in Figure 2. We refer tothe portions of each perforator de-fined by the anatomic spaces as seg-ments (Figs 2, 3). Perforators originatefrom the DIEA, which is usually locatedbetween the posterior layer of the rec-tus sheath and the posterior border ofthe rectus muscle. The perforator pene-trates the posterior surface of the rectusabdominus and has a variable intramus-cular segment. In cases where the ar-tery penetrates a musculotendinous in-tersection, the intramuscular segmentmay be absent. Often, this segment pen-etrates the anterior surface of the rec-tus sheath and the anterior layer of thefascial sheath simultaneously; however,it can alternatively have a varying sub-fascial segment between the anteriorborder of the rectus muscle and the an-terior layer of the rectus sheath. Thecourse and branching of the subcutane-ous segment within the subcutaneousfat is variable and involves varying anas-tomoses with the superficial inferiorepigastric artery. Perforators originat-ing from type II DIEAs are often ar-ranged in two rows along the course ofeach trunk. Some authors refer to theseas medial and lateral row perforators.In addition, some authors refer to per-forators that penetrate the anterior rec-tus sheath near the umbilicus and sup-ply the adjacent subcutaneous fat as

Figure 1

Figure 1: Three-dimensional graphic illustrations show the classification of DIEA branching described byMoon and Taylor (15). Type I (single trunk) and type II (bifurcation into two trunks) branching are more com-mon than type III branching (division into more than two trunks). The umbilical and descending pubicbranches are seen regularly. The level of the umbilicus (�) is indicated on each illustration. EIA � externaliliac artery.

Figure 2

Figure 2: Three-dimensional graphic illustration of the course of a perforator (anterior musculocutaneousbranch) from its origin at the DIEA between the posterior layer of the rectus sheath and the rectus muscle, itsintramuscular segment within the rectus muscle, its subfascial segment (not always present) between theanterior aspect of the rectus muscle and the anterior layer of the rectus sheath, and its end branching (subcuta-neous segment) within the subcutaneous fat of the anterior abdominal wall. (Image courtesy of FrancescoGaillard, MBBS, Royal Melbourne Hospital, Australia.)



Figures 3, 4

Figure 3: Axial 50-mm section thickness maximum intensity projection (MIP)image obtained with abdominal wall CT angiography shows the labeled segmentsof a perforator originating from the right DIEA.

Figure 4: Axial 50-mm section thickness MIP image shows a perforator with along intramuscular segment coursing medially toward the umbilicus. The subfas-cial and subcutaneous segments course away from the umbilicus. Some authorsrefer to these as paraumbilical perforators or paramedian perforators.

Table 1

Relevant Characteristics of the Arteries of the Anterior Abdominal Wall and Their Importance to DIEA Perforator Flap Surgery

Abdominal Wall Artery Relevant Characteristic Importance to DIEA Perforator Surgery

DIEA Caliber A large-caliber trunk is a more favorable pedicle.DIEA Branching pattern (type I, II, or III) The DIEA branch forms the pedicle for the anastomosis. Type I and

type II arteries have perforators with shorter intramuscularsegments. Type II arteries often have perforators arranged in rowsalong each trunk; two or more of these perforators may bedissected to a single trunk. Type II and type III arteries have a lateralbranch that is in the vicinity of the motor nerves to the rectusabdominus; dissection of this branch raises the risk of muscledenervation.

DIEA Any intramuscular course Any intramuscular course makes dissection of the DIEA more difficult.DIEA perforator

Intramuscular perforator Direction and length Oblique intramuscular segments require more dissection of the rectusabdominus. Perforators that penetrate a musculotendinousintersection may not have an intramuscular segment and arepreferred by some surgeons.

Subfascial perforator Length and direction Radiologically, the subfascial segment origin can be confused with thesubcutaneous segment origin. Long subfascial segments requirecareful dissection to preserve the perforator when incising therectus sheath.

Subcutaneous perforator Location and caliber Large-caliber (�1 mm in diameter) arteries supply a greater volume ofadipocutaneous tissue and represent better vascular pedicles.Preoperative knowledge of the subcutaneous segment originfacilitates accurate planning of dissection.

Superficial inferior epigastric artery Caliber and location Use of a large-caliber superficial inferior epigastric artery as a vascularpedicle for a free flap is favored by some surgeons.



paraumbilical or paramedian perfora-tors, an example of which is shown inFigure 4.

Important Anatomic Features for DIEAPerforator Flap SurgeryTable 1 details the relevant character-istics of the DIEA, the segments of theperforators, and the superficial infe-rior epigastric artery. Both type I andtype II branching patterns have beenshown to have perforators that have ashorter intramuscular course than doperforators in patients with type IIIbranching patterns (20). Individualperforators can be traced to individualbranches, and the branch forms thepedicle for the anastomosis. A type IIor type III branching pattern has alateral branch that is in the vicinity ofthe motor nerves to the rectus abdo-minus. Leaving this lateral branch insitu by removing a medial branch onlyleaves those nerves undisturbed. De-nervation of the rectus abdominus de-feats the purpose of a muscle-sparingadipocutaneous flap and is highly un-desirable. Preoperative knowledge ofthe DIEA branching pattern maytherefore aid the surgeon’s choice ofwhich perforator to use and thehemiabdomen of choice.

Perforator anatomy is used to de-termine the surgeon’s choice of perfo-rator(s) on which to base a DIEA per-

forator flap, and it influences surgicalplanning and technique (20,21,26). Alarge-caliber (�1 mm in diameter)perforator with a large subcutaneoussegment will supply a larger volume ofadipocutaneous tissue. Perforatorswith a subcutaneous segment origin

that is at a musculotendinous intersec-tion are preferred by some surgeonsand anecdotally have shorter (or ab-sent) intramuscular and subfascialsegments. Knowledge of the subfascialsegment (if any) is desirable, as thesesegments require careful dissection. A

Figure 5

Figure 5: Coronal MIP image with voxels cut away from the three-dimensional dataset to optimize displayof the branching pattern of the DIEAs from their origins at the external iliac arteries (EIAs). The right DIEA has atype II pattern, and the left DIEA has a type I pattern. The umbilicus is visible. The deep circumflex iliac artery(DCIA) can also be seen branching from the external iliac arteries.

Table 2

Acquisition Parameters

ParameterSomatomSensation 64

SomatomSensation 16

Collimation (mm) 64 � 0.6 16 � 0.75Helical beam

pitch 0.9 1.13Tube voltage (kV) 120 120Tube current

(mAs) 180–200 180–200Rotation time

(sec) 0.37 0.5Reconstruction

sectionthickness (mm) 0.75 1

Reconstructionsection spacing(mm) 0.4 0.7

Table 3

Reconstruction and Image Display Methods Used in Abdominal Wall CT Angiography

Reconstruction Method Image Display Method

Reconstruction generated automatically atCT scanner workstation

Thin overlapping 0.75/0.4 mm axial images, thick contiguous20/20 mm MIP images, or thick overlapping 50/20 mm MIPimages*

Reconstruction generated at CT scannerworkstation by imaging technologist

Coronal MIP image with voxels cut away to demonstrate theDIEA branching pattern

Reconstruction manipulated in real time atreporting workstation by readingradiologist

VRT images or multiplanar reformation images

Reconstructions created at reportingworkstation by the radiologist forpresentation to the surgeon

VRT reconstructed abdominal wall images with perforatorsmarked or coronal MIP images with perforators marked

Note.—VRT � volume-rendered technique.

* To ensure the entire course of each perforator is captured on one image, both thick contiguous MIPs and thick overlappingMIPs are created. The thick overlapping MIPs are used only when the entire course of a perforator is not evident on the thickcontiguous MIPs, as these images contain many overlapping vessels.



long intramuscular segment with asubstantial lateral or medial compo-nent to its course requires more dis-section of the rectus muscle. Extensiverectus muscle dissection is not desiredand theoretically increases abdominalwall morbidity. A more medial perfo-rator means that dissection throughthe rectus muscle is less likely to de-nervate the muscle that receives itsnerve supply laterally. Denervation ofthe rectus muscle negates the abdom-inal wall morbidity benefits of theDIEA perforator flap technique and ishighly undesirable. A row of perfora-tors in a sagittal plane originating fromthe same DIEA trunk is favorable forsurgeons who prefer to base DIEAperforator flaps on more than one per-forator, as the rectus muscle can besplit in a sagittal plane to dissect thesevessels to their origin.

Patient Selection

Planned autologous breast reconstructionwith use of an abdominal wall donor flap isan indication for abdominal wall CT angiog-raphy. The technique may also be per-formed for other planned abdominal wallflap–based reconstructions, such as thosethat use the superficial inferior epigastricartery, deep circumflex iliac artery, or su-perficial circumflex iliac artery. Theonly ab-solute contraindication is a previous ana-phylactic reaction to intravenous iodinatedcontrast material. Severe renal impairmentmay also preclude the use of intravenousiodinated contrast material.

Image Acquisition

Patient Positioning and PreparationIdeally, patients undergo imaging inthe supine position, as this mimics the

intraoperative position. It is importantfor patients to not wear clothing thatdeforms the contour of the abdominalwall. This includes the patient’s ownclothing, as well as any strings at-tached to hospital gowns. The Velcrostraps (which may compress the ab-dominal wall and force the patient’sclothing to lie close to her skin) ensurestable patient position and prevent thepatient from falling from the CT scan-ner table. These straps are placed out-side the scanning range to optimizesubsequent image reconstruction.

Acquisition Volume and RangeThe superior extent of the tissue volumeto be characterized is the superior limitof the planned (donor site) surgicalfield. This is a point 3–4 cm above theumbilicus. The inferior extent is the or-igin of the superficial inferior epigastricartery from the common femoral ar-tery.

Contrast Material Enhancement andScanning DirectionThe DIEA and superficial inferior epi-gastric artery fill from the external il-iac and common femoral arteries, re-spectively, in a caudocranial direction.Thus, contrast material enhancementof the inferior epigastric vessels is op-timized by timing the CT acquisitionwith use of bolus tracking at the com-mon femoral artery rather than at theaorta and scanning in a caudocranialdirection. A total of 100 mL of an in-travenous, low-osmolar, nonionic io-dinated high-concentration (350–370mg/mL) contrast agent is adminis-tered with a dual-chamber injectionpump at a rate of 4 mL/sec and imme-diately followed by a 50-mL saline bo-lus flush administered at a rate of 4mL/sec. For the suggested injectionrate, the choice of high-concentration(�350 mg/mL) monomeric iodinatedcontrast material may aid in the iden-tification of tiny perforator branches,particularly intramuscular segments.The evidence for this possibility comesfrom literature in which researchersevaluated visualization of small vesselssupplying the brain, heart, pancreas,liver, and lungs (27–31). Alternately,

Figure 6

Figure 6: Oblique VRT image obtained with use of our customized CLUT to optimize visualization of theanterior rectus sheath, the subfascial and subcutaneous segments of the perforators, the musculotendinousinsertions, and the superficial inferior epigastric artery (SIEA). Small blue arrows � perforators that emergefrom the sheath without a substantial subfascial segment, black arrows � subfascial segments, green ar-rows � subcutaneous segments of perforators with subfascial segments.



use of a monomeric iodinated contrastagent with a concentration of 300mg/mL and an injection rate of 5 mL/sec should produce equivalent vascularattenuation results. The trigger for acqui-sition is attenuation of the common fem-oral artery to a level of more than 100 HUat the axial level of the pubic symphysis.This level also provides a bone landmarkfor the inferior limit of the scan volume in

the z-axis. Acquisition begins with theminimum delay achievable, typically 4seconds, on modern multidetector CTscanners.

Acquisition ParametersContiguous isometric acquisition dataobtained with thin-collimation helicalCT are ideal for quality CT angiography.On the multidetector CT scanners (So-

matom Sensation 16 and Somatom Sen-sation 64; Siemens Medical Solutions,Erlangen, Germany) used to performabdominal wall CT angiographic exami-nations, we used settings that balancedlow radiation dose with good imagequality; this is known as the ALARA (aslow as reasonably achievable) principle(Table 2). Diagnostic-quality images wereobtained with both scanners. Use of

Figure 7

Figure 7: Oblique close-up VRT image shows the subfascial (white arrows) and subcutaneous (yellow-green arrows) segments of a right-sided perforator. The proximal subcutaneous segment (blue arrows) of aleft-sided perforator is also visible.

Figure 8

Figure 8: Screen-capture image obtained dur-ing real-time manipulation of a VRT image, asseen from the right side of the body. The umbili-cus, several perforators (short arrows), and a largeleft-sided superior-inferior epigastric artery (longarrow) are visible.

Figure 9

Figure 9: Screen-capture images obtained during real-time manipulation of a VRT image, with volume cropping in a coronal-frontal orientation. The umbilicus (whitearrow) and three large perforators (blue, yellow, and green arrows) are visible. Three perspectives (left anterior inferior oblique, anteroposterior [AP], and right anteriorinferior oblique) of the same cropped volume with the same perforators marked are shown to demonstrate the benefit of real-time image manipulation. Note the subfascialcourse of the perforator indicated by the blue arrow is more apparent on the anteroposterior and right anterior inferior oblique images.



120-kV tube voltage and 180–200-mAstube current results in high-quality im-ages and acceptable effective dose esti-mates for the limited scanning range ofabdominal wall CT angiography. Withthis technique, the effective dose is 6mSv for an average-sized patient with ascanning range of 30 cm (ImPACT CTPatient Dosimetry Calculator, version0.99w; St George’s Hospital, London,England). Automatic exposure controlwas notable for its absence from ourscanning protocol.

Modern multidetector CT scannershave software that can be used to mod-ulate tube current in two axes. Modula-tion in the z-axis based on the varyingshape of the patient and angular modu-lation is adjusted in real time with asingle gantry rotation delay (32). Whenwe compared abdominal wall CT an-giography performed with dose modula-tion enabled with abdominal wall CT

angiography performed with dose mod-ulation disabled, we found that disablingdose modulation resulted in a similardose estimate and, subjectively, less im-age noise in the abdominal wall (there-fore enabling better visualization of thetiny perforator branches, particularlythe intramuscular segments). The doseestimate was determined by evaluatingeffective dose in an average-sized pa-tient who underwent imaging with dosemodulation enabled and subsequentlycalculating the effective dose had dosemodulation been disabled. Both tech-niques resulted in a dose of 6 mSv. Thesimilar dose estimate is not surprising,as the scanning range almost entirelycovered the bony pelvis, where dosemodulation techniques had little effectin reducing dose. Image quality wasjudged subjectively by comparing side-by-side studies of patients with similarbody habitus and scanning protocols

but with dose modulation enabled ordisabled. It is assumed that the per-ceived improvement in image qualitycould have been caused by the angularmodulation algorithms, which weredesigned to optimize signal-to-noiseratio in the intraabdominal region ofthe images rather than in the abdomi-nal wall (32).

Image Reconstruction

In our clinical practice of abdominal wallCT angiography, all image reconstruc-tion is performed at a Leonardo (Sie-mens Medical Solutions) workstationequipped with imaging software (Syngo3D VRT and Syngo InSpace 4D, version2006A; Siemens Medical Solutions).The images in this article (except whenstated otherwise) were created at thissame workstation with the same imag-ing software. None of the tools used inthis study—including the grid overlaytool—is vendor specific, and all areavailable on all CT workstations en-dorsed by major vendors. Table 3 liststhe image reconstruction and displaymethods used. Thin overlapping axialimages and thick MIP axial images arepreprogrammed into the CT scanningprotocol for automatic reconstruction atthe scanner console. The thin overlap-ping axial images are used as image datafor multiplanar reformations, as well asto generate the MIP and VRT imagescreated by the technologist and radiolo-gist. Experienced technologists spendan average of 10 minutes per patient onimage reconstruction. The reporting ra-diologist (D.L.S.) spends an average of7 minutes per patient on image recon-struction. Given the usual course ofDIEAs, the image data effectively provideimages in the transaxial plane in most pa-tients and are therefore used to measurethe caliber of the DIEAs, as various invivo and in vitro studies have shown theuse of MIPs and multiplanar reforma-tions alone is prone to inaccuracy invascular imaging (33–36).

We prefer to use thick axial MIPimages to demonstrate the intramuscu-lar and subcutaneous segments of theperforators. Figure 3 is a 50-mm axialMIP image obtained with abdominal

Figure 10

Figure 10: Step1ofsurgicalmapcreation.VRT imageseen fromthecoronal-frontalperspectivewith theskincroppedtorevealsubcutaneoussegmentsof theperforators.ThissameimageisusedforFigures11and12.Theradiologistmarksperforators(arrowheads)at thepoint that theypenetrate theanterior rectussheath.Bluearrowheadsindicatea large(�1.0mmdiameter)perforator;yellowarrowheads,asmallperforator (�1.0mmdiameter);andgreenarrowheads,perforatorssuperior to theumbilicus thatarenotsurgically relevant.A largesuperficial inferiorepigastricartery isalsoseen(arrow).Theumbilicusskin isvisible in themiddleof the image.



wall CT angiography that shows the in-tramuscular, subfascial, and subcutane-ous segments of a perforator and itsorigin from the right DIEA.

Certain subtypes of perforators de-scribed by some authors are well dem-onstrated on thick axial MIP images.Figure 4 shows a perforator with a longintramuscular segment coursing medi-ally toward the umbilicus. The subfas-cial and subcutaneous segments courseaway from the umbilicus. Various au-thors refer to these as paraumbilical orparamedian perforators (24,26). Whenpresent, this perforator is often a con-tinuation or branch of the umbilicalbranch of the DIEA. As discussed previ-ously, a long intramuscular segment isan unfavorable characteristic.

Coronal MIP images are obtained todetermine the branching pattern ofDIEAs. Figure 5 is an MIP image thatshows a type II DIEA branching patternon the left side of the image and a type Ipattern on the right side of the image.These images were created by the med-ical imaging technologist at a CT work-station. The cut-away feature of theworkstation can be used to removeparts of the volume anterior (skin) andposterior (omental and/or mesentericarteries and enhancing bowel wall) tothe region of interest to better depictthe DIEA course.

VRT is an image display techniquethat is used to assign color and opacityvalues to voxels and display a two-di-mensional representation of the three-dimensional dataset. Additional fea-tures and graphics-processing methodsvary between software vendors. Com-mon to all VRT software is the colorlook-up table (CLUT), which is used toassign voxel color and opacity on thebasis of CT numbers. We have devel-oped a CLUT that optimally displays thesubfascial segments of the perforators,the location at which perforators passthrough the anterior rectus sheath,and the branching and caliber of thesubcutaneous segments. This CLUT isdesigned to make the subcutaneous fatcompletely transparent and maximizethe surface contours of the anteriorrectus sheath and the subfascial andsubcutaneous segments of the perfora-

tors. The skin has a range of attenua-tions that overlaps with that of the rec-tus sheath and the perforators. TheCLUT is designed to color the skin pink,which allows these structures to be dis-tinguished from each other, and addsrealistic surface rendering of the ab-dominal wall skin for the images pre-sented to the surgeons. We have devel-oped working CLUTs for both SyngoInSpace 4D software and OsiriX open-source software (http://www.osirix-viewer.com/) distributed under the GNU generalpublic license.

VRT images obtained with ourCLUT are optimized for visualizationof the subcutaneous and subfascialsegments of the perforators. The shinygreen surface of the anterior rectussheath demonstrates the contour ofthe subfascial segments, and the red-green surface shading of the subcuta-neous segments is highlighted againstthe transparent fat (Fig 6). The mus-culotendinous insertions appear red

against the green surface of the re-mainder of the anterior rectus sheath.This is the only image format in whichthe musculotendinous insertions aredelineated as being separate from themuscular rectus. Figure 7 is a close-upimage acquired with VRT that high-lights the subfascial and subcutaneoussegments of a perforator. Figure 8demonstrates the ability of the OsiriXsoftware to perform this function witha similarly customized VRT CLUT.

Real-time manipulation of the VRTdisplay involves use of volume croppingand clip planes by the radiologist to fa-cilitate mapping of the locations of thepoints at which the perforators pene-trate the anterior rectus sheath and anysubfascial segments. The radiologistuses this method to ultimately constructa scaled map that the surgeon uses in-traoperatively.

With use of a VRT image viewedfrom an anterior anatomic coronalperspective (effectively viewing the

Figure 11

Figure 11: Step 2 of surgical map creation. A grid is overlaid to facilitate measurement of the subcutaneousperforator origins in orthogonal axes from the umbilicus.



abdominal wall from in front), an an-terior coronal clip plane, volume crop-ping, or both are used to cut away theskin and varying amounts of subcuta-neous fat to survey the anterior rectussheath and locate the point at whichthe perforators penetrate the anteriorrectus sheath. Identification of thecritical anatomy is also aided by rotat-ing the VRT image into the perspectivemost suitable (usually much closer to aview perpendicular to the surface ofthe abdominal wall) to verify that asuspected perforator is a real finding(Fig 9).

The steps the radiologist takes tocreate the final images for presentationto the surgeon are outlined sequentiallyin Figures 10–13. The radiologist usesarrows to mark perforators (Fig 10),with color coding used to denote large(�1.0 mm diameter) and small (�1.0mm diameter) perforators. It is ensuredthat the VRT image is restored to theoriginal scanning orientation when thearrowheads are placed to avoid misreg-istration artifacts that may result from

the multiple VRT rotations that occurduring the mapping process. These ar-rowheads are located outside the VRTvolume, so they do not inadvertentlymove during image manipulation. Ascaled grid is applied to the image withthe 0,0 point of the axis at the umbilicus(Fig 11). The clip plane is removed,leaving a surface-rendered view of theanterior abdominal wall, with the ar-rowheads still marking the location ofthe perforators (Fig 12). This is one ofthe images the surgeon uses intraoper-atively. The same grid and arrowheadsare overlaid onto the coronal MIP imagepreviously described (Fig 13). This sin-gle image facilitates a spatial apprecia-tion of the relationship between the per-forators and their parent DIEA trunks,allowing the surgeon to easily identifylarge perforators that have long obliqueintramuscular or subfascial segments.Two movies demonstrating this processin real time are available online as AVIfiles (Movies 1, 2; http://radiology.rsnajnls.org/cgi/content/full/249/1/32/DC1).

Radiology Report

The key information obtained with ab-dominal wall CT angiography is docu-mented in the radiology report anddistributed to the referring surgeon.The important findings are listed inTable 4. Reconstructed standard axialabdominal images, which are essen-tially late arterial phase images ob-tained through the lower abdomen,are also reviewed to exclude gross in-traabdominal abnormalities detectedwithin this scanning range. The sur-geons at our institution review all im-ages presented to them on the picturearchiving and communication system,and they use these images to plan thesurgical procedure. Intraoperatively,they use the picture archiving andcommunication system to view theVRT perforator map to mark the per-forators on the patient’s abdominalwall with the aid of a ruler and perma-nent marker pen.

Discussion

Abdominal wall CT angiography is anewly developed technique used forpreoperative imaging that may im-prove the standard of care for autolo-gous breast reconstruction with a freeDIEA perforator flap. To our knowl-edge, Masia et al (17) are the onlyauthors to have related the use of ab-dominal wall CT angiography to thesurgical outcome. In two cohorts of 30patients in whom the same surgeonsperformed surgical breast reconstruc-tion with DIEA perforator flaps and(a) with preoperative abdominal wallCT angiography in one cohort and (b)without preoperative abdominal wallCT angiography in the other cohort,Masia et al (17) reported that the pa-tients who underwent preoperative CTangiography had an average per-patientreduction in surgical time of 100 min-utes, one less total flap failure, and oneless partial flap necrosis when com-pared with the patients who did not un-dergo preoperative CT angiography.They failed to report whether thesefindings were significant. The anecdotal

Figure 12

Figure 12: Step 3 of surgical map creation. Anterior volume cropping is disabled, allowing a realistic ren-dering of the skin surface.



experience of our surgeons supportsthese findings.

In this article, we have described anovel approach to abdominal wall CTangiography that differs substantiallyfrom typical abdominal CT angiographictechniques, in that the common femoralartery is used as the site for the bolustrigger, scanning is performed in a cau-docranial direction, the automatic expo-sure control feature is disabled, and ascaled grid overlay tool is used tooptimally present information to thesurgeons. In our experience with 65patients, we have had no technical fail-ures, and we have obtained diagnostic-quality images in all patients. We havenot observed venous contaminationwith this technique. Venous contamina-tion potentially would result in a promi-nent vein being mistaken for a large ar-terial perforator.

We have not encountered any com-plications. Use of a common femoral ar-tery bolus trigger site theoretically min-imizes the risk of substantial stenoticatherosclerotic aortic disease affectingbolus arrival if an upper abdominal aor-tic bolus trigger site has been used in-stead. This bolus trigger site may notprevent substantial stenotic atheroscle-rotic disease from affecting unilateralcommon iliac or external iliac arteries,as this might theoretically result in dif-fering arrival times of contrast materialwithin the common femoral arteries.We have not encountered substantialatherosclerotic disease in small or largearteries in any patient. This likely re-flects the fact that most patients whoundergo breast reconstruction are rela-tively young and healthy. The oldestpatient we examined was 64 years old.This technique has proved accurate inthe assessment of three patients whounderwent previous extensive abdomi-nal wall surgery and consequently hadgross distortion of the normal vascularanatomy, including absence of one orboth of the DIEAs. Surgical findings en-abled us to confirm the abdominal CTangiography findings and allowed confi-dent planning of abdominal wall flapsurgery, with an excellent surgical out-come in these three patients.

We created from scratch a VRT

Figure 13

Figure 13: Step 4 of surgical map creation. This image was obtained in a different patient than the onewhose image was used for Figures 10–12. The same grid and arrows are overlaid onto the coronal MIP image.For the surgeon, this facilitates a spatial appreciation of the relationship between the perforators and theirparent DIEA trunks on a single image. The 0,0 point of the grid marks the umbilicus.

Table 4

Information Obtained with Abdominal Wall CT Angiography and the Image Display,Reconstruction, and Reformation Technique That Best Demonstrates These Findings

Information Obtained with Abdominal Wall CTAngiography Images That Best Demonstrate These Features*

Branching pattern of the DIEAs (type I, II, or III) Coronal MIP imageNumber and location of large perforators Manipulated VRT image display, grid maps

created from manipulated VRT image displayCaliber of large (�1.0-mm diameter) perforators Multiplanar reformations, thick axial MIPsIntramuscular segments of large perforators Thick axial MIPsSubfascial segments of large perforators Thick axial MIPs, manipulated VRT image displaySubcutaneous segments of perforators Thick axial MIPs, manipulated VRT image displayAnterior rectus sheath and musculotendinous

insertionsManipulated VRT image display

Caliber and course of the superficial inferiorepigastric artery

Multiplanar reformations, thick axial MIPs,manipulated VRT image display

* Data are taken from our anecdotal experience.



CLUT that was specifically tailored toabdominal wall CT angiography on thebasis of the attenuation of (a) contrastmaterial in the DIEA and its branches,(b) muscle of the rectus sheath, and(c) subcutaneous fat. Contrasting colors(red and green) were chosen. ThisCLUT enables one to quickly identifythe location where the perforators pen-etrate the anterior rectus sheath whenclip planes are used (average reportingtime, 7 minutes). Clip planes are neces-sary, as the skin is not transparent withthis VRT CLUT. We have no data toshow that the VRT CLUT is more or lessaccurate than vendor-provided CLUTs;however, we noted that reporting timeswere at least halved in several cases inwhich the experience with this newlycreated CLUT was comparable withthat with several vendor-providedCLUTs that were compared side by sidein the same patient data sets. We havenot used other VRT CLUTs since. Whileuse of this VRT technique is straightfor-ward, we have become aware of severalpotential pitfalls. Heavy abdominal wallscarring can be both distracting andmisleading. In these situations, frequentreferral to the axial MIP images is re-quired. This is also required to discernvessels that run within scars. Anotherpotential pitfall is differentiating a per-forator with a subfascial segment fromone that perforates the anterior rectussheath and subsequently courses super-ficially along it. This is important to dif-ferentiate, as it could lead to incorrectlocalization of rectus sheath perfora-tion. Again, this is overcome by refer-ring to axial MIPs, image data, or both.

The most serious drawback of ab-dominal wall CT angiography is theuse of ionizing radiation, particularlyin a population that is already at risk.Doppler US and color duplex US havebeen used with some success in plan-ning perforator flap surgery (37–39).However, this is a technically challeng-ing and time-consuming process thatrequires experienced sonographerswith a good understanding of DIEAperforator flap surgery (17). This wasthe experience at our institution, andour surgeons abandoned preoperativeUS in favor of abdominal wall CT an-

giography early in our experience, de-spite misgivings about the use of ion-izing radiation. Giunta et al (39) re-ported a high rate of false-positivefindings with use of Doppler US (17).To our knowledge, there have been nolarge head-to-head comparisons of USand abdominal wall CT angiographyfor preoperative perforator flap plan-ning. In a recent report of eight pa-tients who underwent DIEA perfora-tor surgery, researchers comparedpreoperative Doppler US with abdom-inal wall CT angiography and foundthat US could not be used to detectperforators or to identify DIEA ana-tomic variations, whereas CT angiog-raphy could be used to identify 22 per-forators with 100% specificity (23).

High-spatial-resolution magnetic reso-nance (MR) angiography has the poten-tial to provide results similar to thoseprovided by abdominal wall CT angiog-raphy; however, MR angiography hasthe advantage of not exposing patientsto ionizing radiation. Kelly et al (40)report high sensitivity for use of MRangiography in planning free fibula flapreconstructive surgery. To our knowl-edge, there have been no reports of ab-dominal wall MR angiography. Our ownpreliminary experience with abdominalwall MR angiography (three patientswere examined with a 1.5-T system andthree were examined with a 3-T system)has been mixed. Reproducibility hasbeen poor, and the depiction of smallerperforators has been disappointing.

In this article, we have defined thesegmental anatomy of the anterior mus-culocutaneous perforating branches ofthe DIEA and discussed the importanceof these features in pre- and intraoper-ative surgical planning and technique.The acquisition protocol and postpro-cessing, image reconstruction, andreading techniques we have formulatedare optimized to obtain comprehensiveanatomic information while minimizingradiation dose to the patient. Its ana-tomic accuracy has been reported inseveral preliminary cadaveric and sur-gical studies (17–23). To our knowl-edge, no large prospective study hasbeen performed. The detailed descrip-tion provided will allow other radiolo-

gists and surgeons interested in freeDIEA perforator flap surgery to incor-porate this useful tool into their prac-tice.

Acknowledgments: We thank FrancescoGaillard, MBBS, PG Dip Surg Anat, Royal Mel-bourne Hospital, for the three-dimensional il-lustration of perforator anatomy; FrancescaLangenberg, DCR, The Royal Melbourne Hos-pital, for assistance with acquisition and post-processing CT protocols; and PaulEinsiedel BSc (Hons), The Royal MelbourneHospital, for performing CT dosimetry.

References1. Holmström H. The free abdominoplasty flap

and its use in breast reconstruction: an ex-perimental study and clinical case report.Scand J Plast Reconstr Surg 1979;13:423–427.

2. Granzow JW, Levine JL, Chiu ES, et al.Breast reconstruction with perforator flaps.Plast Reconstr Surg 2007;120:1–12.

3. Nahabedian MY, Momen B, Galdino G, et al.Breast reconstruction with the free TRAM orDIEP flap: patient selection, choice of flap,and outcome. Plast Reconstr Surg 2002;110:466–475.

4. Garvey PB, Buchel EW, Pockaj BA, et al.DIEP and pedicled TRAM flaps: a compari-son of outcomes. Plast Reconstr Surg 2006;117:1711–1719.

5. Bajaj AK, Chevray PM, Chang DW. Compar-ison of donor-site complications and func-tional outcomes in free muscle-sparingTRAM flap and free DIEP flap breast recon-struction. Plast Reconstr Surg 2006;117:737–746.

6. Schaverien MV, Perks AG, McCulleySJ. Comparison of outcomes and donor-sitemorbidity in unilateral free TRAM versusDIEP flap breast reconstruction. J Plast Re-constr Aesthet Surg 2007;60:1219–1224.

7. Gill PS, Hunt JP, Guerra AB, et al. A 10-yearretrospective review of 758 DIEP flaps forbreast reconstruction. Plast Reconstr Surg2004;113:1153–1160.

8. Blondeel PN. One hundred free DIEP flapbreast reconstructions: a personal experi-ence. Br J Plast Surg 1999;52:104–111.

9. Blondeel N, Vanderstraeten GG, MonstreySJ, et al. The donor site morbidity of freeDIEP flaps and free TRAM flaps for breastreconstruction. Br J Plast Surg 1997;50:322–330.

10. Allen RJ. DIEP versus TRAM for breast re-construction. Plast Reconstr Surg 2003;111:2478.



11. Kaplan JL, Allen RJ. Cost-based comparisonbetween perforator flaps and TRAM flaps forbreast reconstruction. Plast Reconstr Surg2000;105:943–948.

12. Allen RJ. Comparison of the costs of DIEPand TRAM flaps. Plast Reconstr Surg 2001;108:2165.

13. Spiegel AJ, Khan FN. An intraoperative algo-rithm for use of the SIEA flap for breastreconstruction. Plast Reconstr Surg 2007;120:1450–1459.

14. Holm C, Mayr M, Höfter E, et al. The versa-tility of the SIEA flap: a clinical assessment ofthe vascular territory of the superficial epi-gastric inferior artery. J Plast Reconstr Aes-thet Surg 2007;60:946–951.

15. Moon HK, Taylor GI. The vascular anatomyof rectus abdominus musculocutaneous flapsbased on the deep superior epigastric sys-tem. Plast Reconstr Surg 1988;82:815–832.

16. Lindsey JT. Integrating the DIEP and muscle-sparing (MS-2) free TRAM techniques opti-mizes surgical outcomes: presentation of analgorithm for microsurgical breast recon-struction based on perforator anatomy. PlastReconstr Surg 2007;119:18–27.

17. Masia J, Clavero JA, Larranaga JR, et al.Multidetector-row computed tomography inthe planning of abdominal perforator flaps. JPlast Reconstr Aesthet Surg 2006;59:594–599.

18. Alonso-Burgos A, Garcı́a-Tutor E, Bastarrika G,et al. Preoperative planning of deep inferior epi-gastric artery perforator flap reconstruction withmultislice-CT angiography: imaging findings andinitial experience. J Plast Reconstr Aesthet Surg2006;59:585–593.

19. Rosson GD, Williams CG, Fishman EK, et al.3D CT angiography of abdominal wall vascu-lar perforators to plan DIEAP flaps. Micro-surgery 2007;27:641–646.

20. Rozen WM, Phillips TJ, Ashton MW, et al.Pr15 The branching pattern of the DIEA forperforator flaps: the importance of preoper-ative CT angiography. ANZ J Surg 2007;77:A65.

21. Rozen WM, Phillips TJ, Ashton MW, et al.Pr11 preoperative imaging for DIEA perfora-

tor flaps: a comparative study of CT angiog-raphy and Doppler ultrasound. ANZ J Surg2007;77:A64.

22. Rozen WM, Stella DL, Ashton MW, et al.Three-dimensional CT angiography: a newtechnique for imaging microvascular anat-omy. Clin Anat 2007;20:1001–1003.

23. Rozen WM, Phillips TJ, Ashton MW, et al.Preoperative imaging for DIEA perforatorflaps: a comparative study of computed to-mographic angiography and Doppler ultra-sound. Plast Reconstr Surg 2008;121:9–16.

24. Tregaskiss AP, Goodwin AN, Acland RD.The cutaneous arteries of the anterior ab-dominal wall: a three-dimensional study.Plast Reconstr Surg 2007;120:442–450.

25. Boyd JB, Taylor GI, Corlett R. The vascularterritories of the superior epigastric and thedeep inferior epigastric systems. Plast Re-constr Surg 1984;73:1–16.

26. Vandevoort M, Vranckx JJ, Fabre G. Perfo-rator topography of the deep inferior epigas-tric perforator flap in 100 cases of breastreconstruction. Plast Reconstr Surg 2002;109:1912–1918.

27. Awai K, Takada K, Onishi H, et al. Aorticand hepatic enhancement and tumor-to-livercontrast: analysis of the effect of differentconcentrations of contrast material at multi-detector row helical CT. Radiology 2002;224:757–763.

28. Cademartiri F, Mollet NR, van der Lugt A,et al. Intravenous contrast material adminis-tration at helical 16-detector row CT coro-nary angiography: effect of iodine concentra-tion on vascular attenuation. Radiology2005;236:661–665.

29. Chawla S. Advances in multidetector com-puted tomography: applications in neurora-diology. J Comput Assist Tomogr 2004;28(suppl 1):S12–S16.

30. Fenchel S, Fleiter TR, Aschoff AJ, et al. Ef-fect of iodine concentration of contrast me-dia on contrast enhancement in multisliceCT of the pancreas. Br J Radiol 2004;77:821–830.

31. Schoellnast H, Deutschmann HA, Fritz GA,et al. MDCT angiography of the pulmonary

arteries: influence of iodine flow concentra-tion on vessel attenuation and visualization.AJR Am J Roentgenol 2005;184:1935–1939.

32. Mulkens TH, Bellinck P, Baeyaert M, et al.Use of an automatic exposure control mech-anism for dose optimization in multi-detec-tor row CT examinations: clinical evaluation.Radiology 2005;237:213–223.

33. Berg MH, Manninen HI, Vanninen RL, et al.Assessment of renal artery stenosis with CTangiography: usefulness of multiplanar refor-mation, quantitative stenosis measurements,and densitometric analysis of renal paren-chymal enhancement as adjuncts to MIP filmreading. J Comput Assist Tomogr 1998;22:533–540.

34. Brink JA, Lim JT, Wang G, et al. Technicaloptimization of spiral CT for depiction ofrenal artery stenosis: in vitro analysis. Radi-ology 1995;194:157–163.

35. Hyde DE, Habets DF, Fox AJ, et al. Compar-ison of maximum intensity projection anddigitally reconstructed radiographic projec-tion for carotid artery stenosis measure-ment. Med Phys 2007;34:2968–2974.

36. Van Hoe L, Vandermeulen D, Gryspeerdt S,et al. Assessment of accuracy of renal arterystenosis grading in helical CT angiographyusing maximum intensity projections. EurRadiol 1996;6:658–664.

37. Mah E, Temple F, Morrison W. Value ofpre-operative Doppler ultrasound assess-ment of deep inferior epigastric perforatorsin free flap breast reconstruction. ANZJ Surg 2005;75(S1):A89.

38. Hallock GG. Doppler sonography and colorduplex imaging for planning a perforatorflap. Clin Plast Surg 2003;30:347–357.

39. Giunta RE, Geisweid A, Feller AM. The valueof preoperative Doppler sonography forplanning free perforator flaps. Plast Recon-str Surg 2000;105:2381–2386.

40. Kelly AM, Cronin P, Hussain HK, et al. Preoper-ative MR angiography in free fibula flap transferfor head and neck cancer: clinical applicationand influence on surgical decision making. AJRAm J Roentgenol 2007;188:268–274.



Note: This copy is for your personal non-commercial use only. To … · 2019. 4. 11. · Abdominal...

Documents

Transcript of Note: This copy is for your personal non-commercial use only. To … · 2019. 4. 11. · Abdominal...