ARTICLE IN PRESS

The Breast (2007) 16, 137–145

THE BREAST

0960-9776/$ - sdoi:10.1016/j.b

$Part of thisheld between S�CorrespondiE-mail addr

www.elsevier.com/locate/breast

ORIGINAL ARTICLE

Comparison between breast volume measurementusing 3D surface imaging and classical techniques$

Laszlo Kovacsa,�, Maximilian Edera, Regina Hollweckb,Alexander Zimmermanna, Markus Settlesc, Armin Schneiderd,Matthias Endlicha, Andreas Muellera, Katja Schwenzer-Zimmerere,Nikolaos A. Papadopulosa, Edgar Biemera

aDepartment of Plastic and Reconstructive Surgery, Klinikum rechts der Isar, Technical University Munich,Ismaninger strabe 22, D-81675 Munich, GermanybInstitute of Medical Statistics and Epidemiology, Klinikum rechts der Isar, Technical University Munich,Ismaningerstr. 22, 81675 Munich, GermanycDepartment of Radiology, Klinikum rechts der Isar, Technical University Munich, Ismaningerstr. 22,81675 Munich, GermanydWorkgroup for Minimally Invasive Therapy and Intervention–MITI of Technical University Munich,Department of Surgery, Klinikum rechts der Isar, Technical University Munich, Ismaningerstr. 22,81675 Munich, GermanyeCenter of Advanced Studies in Cranio-Maxillo-Facial Surgery, Department of Reconstructive Surgery,Division of Cranio-Maxillo-Facial Surgery, University of Basel, Spitalstrasse 21, 4031 Basel, Switzerland

Received 18 March 2006; received in revised form 4 June 2006; accepted 9 August 2006

KEYWORDSBreast;Volume;Three-dimensionalimaging;3D scan

ee front matter & 2006reast.2006.08.001

work was presented ateptember 28 and Octong author. Tel.: +49 89ess: l.kovacs@lrz.tum.d

Summary Quantification of the complex breast region can be helpful in breastsurgery, which is shaped by subjective influences. However, there is no generallyrecognized method for breast volume calculation. Three-dimensional (3D) bodysurface imaging represents a new alternative for breast volume computation. Theaim of this work was to compare breast volume calculation with 3D scanning andthree classic methods, focusing on relative advantages, disadvantages, andreproducibility. Repeated breast volume calculations of both breasts in six patients(n ¼ 12) were performed using a 3D laser scanner, nuclear magnetic resonanceimaging (MRI), thermoplastic castings, and anthropomorphic measurements. Meanvolumes (cc) and mean measurement deviations were calculated, and regressionanalyses were performed. MRI showed the highest measurement precision, with amean deviation (expressed as a percentage of mean breast volume) of 1.5670.52%

Elsevier Ltd. All rights reserved.

the Annual Meeting of the German Association of Plastic Surgeons (VDPC) in Munich, Germany,ber 1, 2005.4140 5073; fax: +49 89 4140 7399.e (L. Kovacs).

ARTICLE IN PRESS

L. Kovacs et al.138

compared with 2.2770.99% for the 3D scanner, 7.9773.53% for thermoplasticcastings, and 6.2671.56% for the anthropomorphic measurements. Breast volumecalculations using MRI showed the best agreement with 3D scanning measurement(r ¼ 0.990), followed by anthropomorphic measurement (r ¼ 0.947), and thermo-plastic castings (r ¼ 0.727). Compared with three classical methods of breast volumecalculation, 3D scanning provides acceptable accuracy for breast volume measure-ments, better spatial interpretation of the anatomical area to be operated on (dueto lack of chest deformation), non-invasiveness, and good patient tolerance. Afterthis preliminary study and further development, we believe that 3D body surfacescanning could provide better preoperative planning and postoperative control ineveryday clinical practice.& 2006 Elsevier Ltd. All rights reserved.

Introduction

A primary objective of all breast surgeries is breastsymmetry. Quantification of breast volume may behelpful in obtaining optimal results.1,2 However, ineveryday clinical life breast volume measurementsare not routinely accomplished because thus far nogenerally recognized method of volume calculationexists.3 Generally, these methods of breast volumeassessment fall into one of five different categories.

The anthropomorphic method attempts to derivea correlation between breast volume data obtainedby other methods and standardized end-to-endmeasurements of the thorax region.4,5 As analternative, with modified anthropomorphic meth-ods the breast volume is equated to a half-ellipseand the parameters of the mathematical formula ofa half-ellipse are measured directly at the patientor indirectly from a two-dimensional (2D) photo-graph of the breast region.6,7

Volume methods based on the basis of 2D imagessuch as mammograms and ultrasound are somewhatcomparable to modified anthropomorphic measure-ment with the help of 2D photography.8,9 Geo-metric parameters (e.g., partial ellipse) areprojected onto the 2D image or individual ultra-sonic layers and the breast volume is calculatedusing the appropriate mathematical formulas.

Archimedean methods of breast volume mea-surement are based on Archimedes’ principle ofwater displacement.10,11 The female patient bendsover a water-filled vessel, lowering her breast intothe water, and breast volume is calculated based ondisplaced water. Alternatively, modified methodsuse calibrated measurement cylinders placedagainst the thorax wall; the rigid thorax wall formsthe rear demarcation of the breast and the ventraltissue portions are measured as the displaced‘‘breast volume.’’12–14

Another method is the use of plaster andthermoplastic materials to generate a three-

dimensional (3D) negative cast of the breast.15,16

The cast materials are placed on the upright,seated patient and left to harden. The resulting 3Dshell model is filled with water or sand in order todetermine breast volume.

Modern imaging procedures such as computedtomography (CT) and nuclear magnetic resonanceimaging (MRI) offer an alternative means ofmodeling the breast in 3D.17,18 The patient isplaced in the scanner in a prone position, and thebreast volume is calculated by the summation ofsegmented monolayers.

An alternative to these classical methods is 3Dbody surface imaging. With the help of different 3Dimaging devices, a non-invasive recording in astanding position and the creation of a virtual 3Dmodel of the breast region are possible. Further-more, the 3D technology provides the abilityto quantitatively evaluate symmetry, volume,shape, contour, surface, and distance measure-ments.2,19–26 However, no comparisons have beenmade between breast volume calculations using 3Dimaging versus other methods; cross-comparisonsof the classic volume computation methods arethemselves uncommon.3,27

Based on the results of our preparatory stu-dies,2,21,22,28 the focus of this work was topreliminarily assess the potential of 3D surfacescanning in relation to the existing methods ofbreast volume computation by performing a criticalanalysis of the advantages and disadvantages ofeach method with special emphasis on the repro-ducibility and inter-correlation of the individualmethods.

Materials and methods

The breast volumes (n ¼ 12) of a homogeneousgroup of six test subjects (mean age 2771.86years, mean body mass index 2071.17 kg/m2, and

ARTICLE IN PRESS

Comparison between breast volume measurement 139

mean sternal notch to nipple distance 20.4171.82 cm) were measured by one observer with fourdifferent methods and the results compared.

3D laser scanner

3D breast imaging was performed with the testparticipants standing and arms down using aMinolta Vivid 910 3D linear laser scanner (Konica-Minolta, Osaka, Japan) according to a standardized3D scanning protocol.22 Applying the protocoldeveloped in our lab,28 one observer performedten measurements per breast, per participant(n ¼ 120) and calculated breast volumes in cc usingthe Raindrop Geomagic Studio 7 software (RaindropGeomagic, Durham, NC, USA; Fig. 1).

Nuclear magnetic resonance imaging

Test participants were positioned prone in a 1.5TPhilips Intera Upgrade R7 MR scanner (PhilipsMedizin Systeme, Hamburg, Germany) for MRI. Nocontrast medium was used. Analogous to theprocedure described by Mineyev et al.,29 volumecomputation was accomplished with a 3-mm layerthickness using Easy Vision 4.0 software (PhilipsMedizin Systeme, Hamburg, Germany). Imagingand ten volume computations per breast andparticipant (n ¼ 120) were performed by oneobserver under the guidance of an experiencedradiological specialist. Manual segmentation of thetissue portions was performed in each ventralmonolayer along the outside breast form and onthe dorsal aspect of the pectoral muscle (Fig. 2a).

Figure 1 Creation of a closed volume model, whichrepresents the breast volume in cc using a 3D scanner.

Monolayers were totaled to obtain the total breastvolume in cc.

Thermoplastic casts

Analogous to the published method of Edsander-Nordet al.,16 a thermoplastic cast of each test partici-pant’s breast region was taken in an upright, seatedposition and evaluated (Orfit Classic Soft; OrfitIndustries, Wijnegem, Belgium). With this method,a flat dorsal delimitation was created by aligning thefemale (negative) cast form on the horizontal, fillingit up with water, and measuring the water volume inmillimeters (ml ¼ cc); breast volume above thisdelimitation was not included in the calculation(Fig. 3). The examiner performed ten determinationsof breast volume per thermoplastic casting (n ¼ 120).The deformation of the breast shape during themolding process was analyzed.

Anthropomorphic methods

Breast volumes of the six test participants werecomputed using the formula published by Qiaoet al.6: 1/3p�MP2� (MR+LR+IR-MP). MR correspondsto the distance between the nipple and medial breastborder, LR to the distance between the nipple andlateral breast border, IR to the distance between thenipple and inframammary fold, and MP to themammary projection, which was measured by view-ing the test participant from the lateral aspect fromsternum to nipple (Fig. 4). Thus, breast volume wascomputed in cc over a half-ellipse (Fig. 2d).Measurements were made manually ten times perparticipant, per breast (n ¼ 120).

With the formula introduced by Brown et al.,7

the resection weight in cc of five female breastreduction patients (age: 40.20710.59 years, bodymass index: 25.7972.16 kg/m2, average sternalnotch–nipple distance: 28.1372.43 cm) was pre-dicted before surgery. A half-ellipsoid was pro-jected onto a preoperative 2D photograph of afemale patient in order to define the parameters(A, B, and C) needed for computation of theformula (Fig. 4). Next, the postoperative resulttargeted by the surgeon was drawn in and thecorresponding necessary parameters computed(A0, B0, and C0). With the help of the formulaDV ¼ ðp=6ÞaðABC� A0B0C0Þ, the breast volumechanges in cc were predicted by one observer andcompared with the intraoperative resection weight(g, assumed mass density of 1 g/cm3).30 Addition-ally, the predicted breast volume changes werecompared with the results of our study performedwith the same patients in which the difference

ARTICLE IN PRESS

Figure 2 Example of breast volume areas measured in cc using the four different techniques. Measured areasare indicated by black and white lines on the two identical images: (a) MRI, (b) 3D scanning, (c) thermoplastic casts,and (d) modified anthropomorphic measurement with half-ellipse.

Figure 3 Delimitation of breast volume measurement (cc) with thermoplastic cast: (a) inner breast view, (b) caudalview, and (c) resulting dorsal planar boundary.

Figure 4 Breast volume measurement procedures (cc) using anthropomorphic measurement (white lines) and 2Dphotography (black lines). (a) MR ¼ distance between nipple and medial border, IR ¼ distance between nipple andinframammary fold, and (b) LR ¼ distance between nipple and lateral border. (a) A ¼ the ellipsoidal width, front view,(b) B ¼ the ellipsoidal height, and C ¼ the ellipsoidal semiaxis, lateral view; A0, B0, and C0 ¼ the correspondingmeasurements on the predicted postoperative breast.

L. Kovacs et al.140

between pre- and postoperative breast volume (cc)was calculated with the help of the 3D scanningmethod.28

Statistical analysis

Each test participant’s ten replicate breast volumemeasurements were aggregated separately for

each breast as a mean volume in cc. The coefficientof variation (CV ¼ 100� standard deviation/mean)was expressed as a percentage of mean breastvolume to evaluate reproducibility. Dependency onthe volume of the right and left breasts of one testparticipant was not considered in the testingprocedures. To investigate possible relationshipsamong the four breast volume measurementprocedures, the Pearson coefficient of correlation

ARTICLE IN PRESS

Comparison between breast volume measurement 141

was computed and a linear regression analysis wasperformed. All tests were two-tailed using a sig-nificance level of Po0.05 and were performed usingSPSS version 13 software (SPSS, Chicago, IL, USA).

Results

The mean breast volumes for each individual testparticipant by measurement method are presentedin Table 1, which also shows the mean breastvolume (cc) across test participants and the meanmeasurement deviation (expressed as a percentageof mean volume) for each method. Figure 2 depictsa comparative representation of breast areascomputed with the different methods, overlaid ontwo identical MRI tomography images.

3D laser scanner

The 3D scanning method yielded a mean breastvolume across all test participants of 452.517141.88 cc, and a measurement deviation of2.2770.99% of the measured breast volumes (Table1). The breast volume determined by the 3Dscanner is represented in Fig. 2b.

Nuclear magnetic resonance imaging

Our comparison of measurement techniquesshowed that in contrast to 3D scanning, MRImeasurement determined the entire breast volumeaccording to the real anatomical structures(Fig. 2a). With the MRI method, the mean breastvolume increased to 582.277184.71 cc and ahigher measurement precision with a deviation of

Table 1 Test participants’ breast volume assessmentechniques.

Test persons 3D scan MRI T

Mean volume+S.D. (cc)1 311.5072.40 420.7071.56 362 570.8074.81 758.6079.33 453 231.90721.35 292.30729.27 184 460.2074.24 560.6073.11 425 566.4079.05 709.30728.14 466 574.2576.15 75210710.04 51Mean 452.517141.88 582.277184.71 44

Mean deviation (S.D.) in percentage of volume+S.D. (%)Mean 2.2770.99 1.5670.52

Mean breast volumes (cc) within and across test participants andvolume (%)) are shown.

1.5670.52% of the mean breast volume wasobtained (Table 1).

Thermoplastic casts

Breast volumes determined by means of thermo-plastic casts differed substantially from thosedetermined by 3D scanning (Fig. 2c, Table 1). Themean breast volume across test participants was440.187112.56 cc, smaller than that reported for3D scanning. The measurement deviation,7.9773.53% of the mean volume, was higher thanthat for 3D scanning, indicative of lower measure-ment precision (Table 1).

Anthropomorphic measurement

The mean breast volumes within participants andacross participants determined using the formuladescribed by Qiao et al.6 are presented in Fig. 4 andTable 1. The mean breast volume across all testparticipants was 445.207192.08 cc. A mean mea-surement deviation of 6.2671.56% of mean breastvolume was calculated (Table 1).

The predicted average breast volume changeusing the formula by Brown et al.7 of the fivefemale breast reduction patients was 559.067259.44 cc with a mean measurement deviation of5.5471.50% of the mean breast volume. Theaverage resection weight was 600.537441.67 cc,and the calculated volume difference using 3Dsurface imaging was 600.487460.55 cc with a meandeviation of 2.4770.52% preoperatively and1.3470.23% postoperatively (Po0.001).28

ts (n ¼ 12) with the four different measurement

hermoplastic casts Anthropomorphic measurement

6.50716.26 267.60767.467.0074.24 503.90746.536.50732.47 133.80725.461.00779.20 529.6072.834.00741.01 605.20767.467.50760.10 631.10725.460.187112.56 445.207192.08

7.9773.53 6.2671.56

mean deviations (expressed as a percentage of mean breast

ARTICLE IN PRESS

Figure 5 Regression lines between breast volume mea-surements obtained from 3D scanning and three othertechniques (broken line indicates identity line [resultsfrom linear regression through origin with slope ¼ 1],

L. Kovacs et al.142

Correlations among the four breast volumemeasurement techniques

In spite of using four different recording methods(Fig. 2), the mean breast volumes showed agenerally similar pattern across participants (Table1).

A comparison of the coefficients of correlation (r)for the four methods (Table 2) revealed that MRIshowed the best agreement with 3D scanning(r ¼ 0.990, Po0.001), followed by the anthropo-morphic method (r ¼ 0.947, Po0.001), and ther-moplastic casts (r ¼ 0.727, P ¼ 0.017).

To enable more direct comparison of thebreast volumes determined with the individualmethods, regression equations were calculated,which enabled conversion of all the measurementunits into those used for 3D scanning (Fig. 5). Theregression equations revealed the following rela-tionships: (a) 3D scan ¼ 9.83+0.75�MRI, (b) 3Dscan ¼ �47.69+1.22� thermoplastic casts, and (c)3D scan ¼ 141.03+0.70� anthropomorphic mea-surement (all in cc).

continuous line indicates regression line).

Discussion

From the viewpoint of the operating surgeon,determination of breast volume could be helpfuland desirable to potentially facilitate the complexplanning and difficult execution of many surgicalbreast interventions, including correction of breastasymmetry, restoration of an ablated breast, andvolume-changing esthetic intervention.1,2,6,7,20–26

The desire for improved breast volume calculationmethods is reflected by over 50 publications in thelast four decades on the topic. Unfortunately, thevarious breast volume measurement techniques

Table 2 Pearson correlation coefficients (r) and Pvalues for correlation analyses of the four differentmeasurement techniques.

3D scan MRI Thermoplastic casts

MRIr 0.990P 0.000

Thermoplastic castsr 0.727 0.762P 0.017 0.010

Anthropomorphic measurementr 0.947 0.914 0.669P 0.000 0.000 0.035

that have been proposed in such articles exhibitvariable reliability. Moreover, these techniquesinvolve a level of detail that can be difficult toexecute, they are of limited practicability, areoften cost-intensive, and are not always acceptedby the patient.3 For these reasons, these techni-ques have found only limited application in every-day plastic and reconstructive surgery, and in onlyexceptional cases does breast volume measure-ment occur before surgery. Most methods of breastvolume assessment can be categorized into one ofthe five groups described. Our own analysis showsthat these different methods include differentbreast tissue components in volume computation(Fig. 2). This may partially explain the fact that weoften found differences in the measured absolutevolume values across the four methods analyzed.

With anthropomorphic methods, volumes arecomputed based only on individually measuredvalues (end-to-end measurements), and a prede-fined geometrical shape is imposed on the breastform, which does not necessarily correspond to theindividual anatomical conditions of the respectivebreasts. In light of these disadvantages, the meanmeasurement deviation in the literature, 3.61% ofthe volume measured reported by Westreich5 and3.89% by Qiao et al.,6 is surprisingly small. Incontrast, we determined a mean deviation of 6.26%of the mean volume, applying the formula de-scribed by Qiao et al.6 With the help of a modified

ARTICLE IN PRESS

Comparison between breast volume measurement 143

anthropomorphic method, Brown et al.7 tried toanticipate the size of the resection weight before abreast reduction operation. The method involvesprojection of a half-ellipsoid onto a preoperative2D photograph of a female patient (Fig. 4). We haveapplied this formula to five breast reductionpatients and the mean error of our measurementswith the applied formula was 5.5471.50% ofthe mean breast volume, which is comparable tothe results of Brown et al.7 Although these methodsare relatively feasible, they have very low costs(tape measure, calculator, and working time) andcan be accomplished with a standing patient, onemust get used to the calculation formula in manycases and the methods require some subjectivedeterminations because the limits are extremelyarbitrary. This perception is assured by the factthat the mean internal variation of the manualmeasurements, numerically expressed by the stan-dard deviation (S.D.), is the highest of all themethods analyzed (Table 1). In our preparatorywork we analyzed the precision of manual mea-surements for specific anatomical distances of thebreast region.22 We found that the measuringvariability increases with less anatomically well-defined landmarks or those that were located in thesubmammary region. In our opinion, the above-mentioned specific disadvantages of the anthro-pomorphic methods affect the reproducibility andshow clearly that the measurements are dependenton the individual observer.

With regard to MRI, the breast volume calcula-tion is based on actual anatomical structures.Fowler et al.18 reported a mean deviation of 4.3%for MRI-based volume measurements. In our studywe determined a mean deviation of 1.5670.52%.Because of the easily recognizable landmarks wefound that the manual segmentation was to beperformed very precisely and the discrepancy withthe above-mentioned results can only be explained

Figure 6 Color-coded image demonstrating areas of breast(a) colors are superimposed on the breast region and (b)millimeters. Blue regions indicate compressed areas and yellosqueezed out from the thorax wall.

by the improvement in the MRI imaging. MRIprovides the most precise volume assessmentmethod, although the most costly. Caruso et al.31

analyzed that for a single volume measurement,the cost of the time and materials was $1400 forthe MRI, which is too expensive for a routinepreoperative breast volume measurement. In addi-tion to the high costs and the time-consumingassessment of the results, another disadvantage ofMRI is that deformation of the chest region isapparent on images taken in the supine position;the resulting 3D models do not correspond to thereal anatomical shape of the breast and are of littlehelp to the clinicians.

In the production of thermoplastic casts, thebreast is deformed and pressed against the thoraxwall. The extent of breast compression duringproduction of a thermoplastic cast is shown inFig. 6. Clearly apparent are areas of compressedtissue from the thermoplastic material, as well ascast areas shaped by material squeezed out fromthe thorax wall. The relatively inflexible thermo-plastic material cannot perfectly model the breastform and is distorted under the manual pressure ofthe examiner. In addition, by filling up the castswith water or sand, the rear demarcation of thechest is defined as a flat level that connects theedges of the chest (Fig. 3). The breast portionsabove the flat rear wall are not included in thevolume calculation, and the actual curvature of thebreast wall is not considered; consequently, smallervolumes are computed. Thus, while an actual 3Dbreast volume is created (unlike with anthropo-morphic methods), because the shape of the breastregion is changed, the resulting volume calculationis problematic. Manual evaluation in particular issomewhat subjective. It is not clearly definedwhich line constitutes the rear delimitation of thecast. According to Edsander-Nord et al.,16 the castshould be filled with water ‘‘until it reached two

compression during generation of thermoplastic casts;sagittal slice with 2D deviation; scale is calibrated inw to red areas indicate cast areas resulting from material

ARTICLE IN PRESS

L. Kovacs et al.144

opposite points of the delineated breast bound-ary’’. As shown in Fig. 6, the breast is compressedand the breast borders are thus not obvious andarbitrary. Therefore, we found it extremely diffi-cult to perform the manual evaluation. In theliterature, the precision of the breast volumecalculation has been reported as being relativelyconstant with the thermoplastic cast method, witha deviation of 6%.16 Our results show a deviation of7.97% of the measured volume value. In addition tothe high material costs of approximately $150 perbreast, we found that the relatively inflexiblethermoplastic material with the resulting deforma-tion of the breast tissue and the difficult manualevaluation of the thermoplastic casts affect repro-ducibility and should be used with caution forbreast volume assessment.

The advantages of breast volume computationusing 3D scanning are objectively based on the fastand extensive recording of the breast region’ssurface geometry. A ‘‘virtual casting’’ is developedof the breast region. Unlike conventional methodssuch as gypsum or thermoplastic casting, the datarecording takes place without body contact. Thus,data can be collected without deformation of thebreast and with the patient in an upright standingposition.22 However, as is the case with conven-tional castings of the breast region, the reardemarcation of the volume area that can bemeasured is not obtained. In contrast to theclassical methods, where the curvature ofthe thorax wall can be approximately modeledonly by exerting pressure against the thorax wall, itis possible to correctly compute the thorax wallcurvature with the help of special software.21 Thiscomputed rear demarcation plane runs parallel tothe anatomical breast wall curvature and suffi-ciently accurate breast volume measurements arepossible.2,22,24,28 Besides the advantages there arestill the technical obstacles, the sometimes time-consuming analysis, and the high costs of this newtechnology to overcome in the future.32 There is awide variety of available 3D surface imagingsystems and appropriate software applications,which differ enormously in cost and quality. There-fore, it is difficult to make a universally validstatement concerning the cost effectiveness of 3Dsurface imaging technology. But we believe thatthe development of 3D scanning systems in the lastfew years and the improvement in clinical applica-tions will make this technology more user-friendly,more affordable with regard to compact applica-tion packages, and will one day be established inthe everyday life of plastic surgeons, similar in costand applicability to the currently used 2D digitalphotography systems.33

The calculated regression equations can assist incomparing the absolute values of the individualvolume calculation methods. We used regressionanalysis to explore the optimal methods of volumemeasurement that can be accomplished simply andfavorably—the gold standard of 3D measurement.A straight line was used as an interpolation functionthat showed a good approximation of the data.The regression lines showed in many cases cleardeviations from the reference lines; the data ofthermoplastic casting, in particular, revealed astrong dependence on the actual volume comparedwith the other methods. In this context, we have toadmit that due to the limited sample size in thisstudy our results have to be described as prelimin-ary and further investigations are needed.

Conclusions

Our analysis has shown that the four breast volumecalculation methods considered take into accountdifferent areas of breast tissue (Fig. 2). Since thesedifferent methods measure different breast volumeareas, a comparison is possible only at the level ofmeasurement precision.

In summary, the 3D scanner represents a simpleand promising method. Based on the determinedmeasurement precision,28 and the characteristicrelationships to the other recognized methods wehave shown here, there is reason to believe thatthis technology—in spite of the existing technicalobstacles and only preliminary results in compar-ison to other breast volume measurement techni-ques—could contribute to the establishment ofroutine breast volume measurements in everydayplastic and reconstructive surgery after furtherclinical investigations and technical development.

Acknowledgments

The authors are grateful to Prof. Dr. E.J. Rummeny,Director of the Department of Radiology and Prof.Dr. H. Feussner, Clinical Head of the Workgroup forMinimally Invasive Therapy and Intervention (MITI),Department of Surgery, both Technical University ofMunich, Germany, for their cooperation and infra-structural support, which contributed enormouslyto the realization and success of this study.

The authors also thank Prof. Dr. H.F. Zeilhofer,Director of the Department of ReconstructiveSurgery, Division of Cranio-Maxillo-Facial Surgery,University of Basel, Switzerland and Prof. Dr. R. Sader,Department of Cranio-Maxillo-Facial and Facial Plastic

ARTICLE IN PRESS

Comparison between breast volume measurement 145

Surgery, J.W. Goethe University, Frankfurt am Main,Germany for the continuous support of our projectsand for the provision of the infrastructure that hasenabled this study to be carried out. The authorswould also like to thank Mr. Marco Zajac, Germanrepresentative of Minolta Co., Ltd., Osaka, Japan, forhis long-standing support of our research projects.

References

1. Hudson DA. Factors determining shape and symmetry inimmediate breast reconstruction. Ann Plast Surg2004;52:15–21.

2. Kovacs L, Zimmermann A, Papadopulos NA, et al. Re: factorsdetermining shape and symmetry in immediate breastreconstruction. Ann Plast Surg 2004;53:192–4.

3. Bulstrode N, Bellamy E, Shrotria S. Breast volume assess-ment: comparing five different techniques. Breast 2001;10:117–23.

4. Smith Jr. DJ, Palin Jr. WE, Katch VL, et al. Breast volume andanthropomorphic measurements: normal values. Plast Re-constr Surg 1986;78:331–5.

5. Westreich M. Anthropomorphic breast measurement: proto-col and results in 50 women with aesthetically perfectbreasts and clinical application. Plast Reconstr Surg 1997;100:468–79.

6. Qiao Q, Zhou G, Ling Y. Breast volume measurement in youngChinese women and clinical applications. Aesthetic PlastSurg 1997;21:362–8.

7. Brown RW, Cheng YC, Kurtay M. A formula for surgicalmodifications of the breast. Plast Reconstr Surg 2000;106:1342–5.

8. Kalbhen CL, McGill JJ, Fendley PM, et al. Mammographicdetermination of breast volume: comparing differentmethods. Am J Roentgenol 1999;173:1643–9.

9. Malini S, Smith EO, Goldzieher JW. Measurement of breastvolume by ultrasound during normal menstrual cyclesand with oral contraceptive use. Obstet Gynecol 1985;66:538–41.

10. Bouman FG. Volumetric measurement of the human breastand breast tissue before and during mammaplasty. Br J PlastSurg 1970;23:263–4.

11. Schultz RC, Dolezal RF, Nolan J. Further applications ofArchimedes’ principle in the correction of asymmetricalbreasts. Ann Plast Surg 1986;16:98–101.

12. Kirianoff TG. Volume measurements of unequal breasts.Plast Reconstr Surg 1974;54:616.

13. Tegtmeier RE. A quick, accurate mammometer. Ann PlastSurg 1978;1:625–6.

14. Wilkie T. Volumetric breast measurement during surgery.Aesthetic Plast Surg 1977;1:301–5.

15. Campaigne BN, Katch VL, Freedson P, et al. Measurement ofbreast volume in females: description of a reliable method.Ann Hum Biol 1979;6:363–7.

16. Edsander-Nord A, Wickman M, Jurell G. Measurement ofbreast volume with thermoplastic casts. Scand J PlastReconstr Surg Hand Surg 1996;30:129–32.

17. Neal AJ, Torr M, Helyer S, et al. Correlation of breast doseheterogeneity with breast size using 3D CT planning anddose–volume histograms. Radiother Oncol 1995;34:210–8.

18. Fowler PA, Casey CE, Cameron GG, et al. Cyclic changes incomposition and volume of the breast during the menstrualcycle, measured by magnetic resonance imaging. Br JObstet Gynaecol 1990;97:595–602.

19. Nahabedian MY, Galdino G. Symmetrical breast reconstruc-tion: is there a role for three-dimensional digital photo-graphy? Plast Reconstr Surg 2003;112:1582–90.

20. Galdino GM, Nahabedian M, Chiaramonte M, et al. Clinicalapplications of three-dimensional photography in breastsurgery. Plast Reconstr Surg 2002;110:58–70.

21. Kovacs L, Eder M, Papadopoulos NA, et al. Re: validatingthree-dimensional imaging of the breast. Ann Plast Surg2005;55:695–6.

22. Kovacs L, Yassouridis A, Zimmermann A, et al. Optimisationof the three-dimensional imaging of the breast region with3D Laser Scanners. Ann Plast Surg 2006;56:229–36.

23. Lee HY, Hong K, Kim EA. Measurement protocol of women’snude breasts using a 3D scanning technique. Appl Ergon2004;35:353–9.

24. Losken A, Seify H, Denson DD, et al. Validating three-dimensional imaging of the breast. Ann Plast Surg2005;54:471–6 (Discussion 477–478).

25. Losken A, Fishman I, Denson DD, et al. An objectiveevaluation of breast symmetry and shape differences using3-dimensional images. Ann Plast Surg 2005;55:571–5.

26. Isogai N, Sai K, Kamiishi H, et al. Quantitative analysis of thereconstructed breast using a 3-dimensional laser lightscanner. Ann Plast Surg 2006;56:237–42.

27. Palin Jr. WE, von Fraunhofer JA, Smith Jr. DJ. Measurementof breast volume: comparison of techniques. Plast ReconstrSurg 1986;77:253–5.

28. Kovacs L, Eder M, Hollweck R, et al. New aspects of breastvolume measurement using 3D surface imaging. Ann PlastSurg., 2006;57: in press.

29. Mineyev M, Kramer D, Kaufman L, et al. Measurement ofbreast implant volume with magnetic resonance imaging.Ann Plast Surg 1995;34:348–51.

30. Katch VL, Campaigne B, Freedson P, et al. Contribution ofbreast volume and weight to body fat distribution infemales. Am J Phys Anthropol 1980;53:93–100.

31. Caruso MK, Guillot TS, Nguyen T, et al. The cost effective-ness of three different measures of breast volume. AestheticPlast Surg 2006;30:16–20.

32. Nahabedian MY. Invited discussion: validating three-dimen-sional imaging of the breast. Ann Plast Surg 2005;54:477–8.

33. Jacobs RA. Three-dimensional photography. Plast ReconstrSurg 2001;107:276–7.