37

ORIGINAL ARTICLE

AN APPROACH FOR ASSESSMENT OF TUMOR VOLUME FROMMAMMOGRAPHY IN LOCALLY ADVANCED BREAST CANCER

Gupreet Singh

Medical Physics Unit, Department of Radiodiagnosis Institute Rotary Cancer Hospital, All India Institute of Medical Sciences,

New Delhi-110029, India.

Tumor size is an important independent indicator in patients with carcinoma of thebreast. Repeated size measurements during primary systemic therapy produce detailedinformation about response that could be used to select the most effective treatmentregimen and to estimate the patient’s prognosis. Measurement of tumor burden withultrasonography and computed tomography is being used with increasing frequencyto assess the effectiveness of cytotoxic anticancer drugs. Standardization of assessmentand results reporting are important steps that aim at increasing the amount of usabletherapeutic information at the physician’s disposal. The purpose of our study is tocalculate the tumor volume by mammography after demagnification and comparethe tumor volume measured from this method with ultrasonography volumes.

Key words : Mammography, tumor volume, magnification, locally advanced breast cancer, prognosis.

Introduction

Breast cancer is reported to be commonest

cancer with an annual age adjusted incidence rateof 22-28 per 100,000 women per year in urban areasand 6 per 100,000 women per year in rural areas.

Submitted-20-02-2007, Accepted-03-12-07

Malaysian Journal of Medical Sciences, Vol. 15, No. 1, January 2008 (37-41)

Table 1 : Comparison of tumor volume measurement by ultrasound (US), Mammagraphy, with verniercalipers and Water Displacement methods

Tumor

mimicking

objects

Vol. by

water

displace

ment (c.c)

Vol. by US

(c.c)

Vol.

obtained

using

vernier

calipers

(c.c)

Vol. by

Mammog

raphy

(c.c)

% diff.

water

displaceme

nt & US

% diff.

water

displaceme

nt &

Mammo

% diff.

water

displaceme

nt & using

vernier

calipers

Potatol

Potato 2

Potato 3

Pumpkin

50

70

40

65

57.47

69.73

39.45

65.78

48.05

67.62

44.15

61.62

52.08

72.89

44.53

67.51

2.98

0.38

1.375

1.2

4.16

4.12

8.37

3.86

3.9

3.4

10.27

5.2

38

Over 75,000 new cases of breast cancer are reportedto occur in India per year. Majority of breast cancersin India (50-70%) present with locally advanceddisease. In the coming year at current growth it willsurpass the cervix cancer [1]. Currently neoadjuvantchemotherapy is a common approach which givesan opportunity to asses the tumor response wherethe benefits are mostly concentrated in a smallnumber of patients who achieved optimal tumorburden reduction. A change in organ size is apredominant feature of a disease process or amanifestation of pathology elsewhere. Variousradiological and clinical techniques have beenattempted in the past to estimate tumor or organvolumes, but have been of limited utility due tounacceptable accuracy, poor reproducibility ordifficulty in obtaining suitable images for volumemeasurements [2]-[4]. Quantification of tumorburden with ultrasonography and computedtomography is being used with increasing frequencyto assess the effectiveness of cytotoxic anticancerdrugs. Standardization of assessment and resultsreporting is an important step that aims at increasingthe amount of usable therapeutic information at thedisposal of the physician. Therefore it is necessaryto develop a ‘common language’ to describe theresults of cancer treatment and to agree uponinternationally acceptable general principles for

reporting and assessing data [5]-[7]. Since 1979,World Health Organization (WHO) used theradiological tumor response evaluation criteria.Whereby, the tumor is measured in two dimensionsits maximum diameter (width) in the translationplane and its largest perpendicular diameter on sameimage (thickness) [8]. The purpose of our study isto potentially calculate the tumor volume bymammography after demagnifying the images andcompare the tumor volume measured from thismethod with ultrasonography volumes. Measuringa mass is a tedious task for most radiologists. Forthe three dimensional measurement approaches, thecriteria for partial treatment response is defined asgreater than 65% reduction in tumor volume, stabledisease being the size between that for partialresponse and that for progressive disease and diseaseprogression is greater than 44% increase in volume[8]. The criterion for complete response was totaltumor disappearance. Tumor size is an importantindependent indicator of prognosis in the patientswith carcinoma of breast. Repeated sizemeasurements during primary systemic therapyproduce detailed information about response that canbe used to select the most effective treatmentregimen and to better estimate patient’s prognosis.

Materials and Methods

Table 2 : Analysis of breast tumor volumes (c.c.) by ultrasound and mammography

Figure 1 : Plot of volume determined by observer 1 and observer2 in Mammography (n=25)

Gupreet Singh

N Mean Std. Devation Std. Error Mean

1

2

3

25

25

25

17.75

18.09

18.16

21.73

21.76

21.67

4.34

4.35

4.33

Lower Bound

8.78

9.10

9.21

Upper Bound

26.72

27.07

27.10

39

The oncology data base was reviewed for allfemale patients with biopsy proven breastcarcinomas that were treated with standardchemotherapeutic protocols. Inclusion criteriarequired that the mass lesion should be clearly visiblewith that of mammography and on ultrasound.Delineation of tumor sizes from the mammogramsas well as from the ultrasound was performed bysenior radiologists. Tumors ranged from the clearlymarginated to poorly marginated. Those that werejudged by senior radiologists as not beingmeasurable lesions were not included in the study.The sizes were measured with mammography andultrasound. Tumor volumes were calculated usingthe formula of an ellipsoid. In mammography twoviews were used to measure tumor diameter whichwere in the craniocaudal view and the mediolateralview. In ultrasonography a 5 MHz linear array probewas used. The transducer was placed in transversedirection on the breast tumor, the width and depthare measured and then the transducer was placed inlongitudinal direction which gives the anotherdimension of the breast tumor. The three greatestdimensions (referred to as a, b and c) were then usedto calculate several estimates of tumor volume usingthe equations for the volume of a sphere ([4 /3]r3),Here r is average of a, b and c the volume of anellipsoid V= (abc )/6. Where a and b are themaximum dimensions of tumor in cm from thecraniocaudal view as well as mediolateral viewrespectively and c was maximum dimensionmeasured from the mediolateral view. Sphericalvolume calculations were based on the averageradius across all three axes. The comparisons weredone among the volumes measured bymammography and ultrasound. The percentagevariations between the two modalities werecalculated. Demagnifying volume was obtained bydividing each dimension of mammography image

by magnification factor. The volume measurementmethod by vernier calipers, ultrasound (US) andmammography method calibrated using tumormimicking substances (like Potato, Pumpkin). Theseobjects dipped in water were subjected to volumeestimation by the above methods. Standard wastaken as volume estimated by water displacementmethod.

Result

Figure 1 show the tumor volume determinedby observer 1 plotted against the volume measuredby observer 2. Regressing on linear scale shows thehigh correlation (r = 0.99) with the slope of 1.018.Figure 2 shows the tumor volume calculated bymammography (Mean volume measured byobserver 1 and observer 2) plotted against thevolume determined by ultrasonography. In this plotthere is high correlation of tumor volumes (r = 0.99)and slope of 0.99.

Estimates based on volume equations ofan ellipse, a sphere of mammography differed fromthe ultrasound volume by a mean of -33.58% (95%confidence interval (CI) from -22.78 to 22.11), -40.68% (95% CI from -22.85 to 22.04) respectively.

Discussion

Pharmaceutical companies have developed anapproval process for anticancer drugs on the basisof tumor shrinkage. Volumetric measurementswould overcome difficulty estimating the size oflesions that are irregular in shapes and sizes. Thetumor volumes measured in breast cancer cases byUS may be taken as standard since calibrationexperiment (Table I) showed that US is very accuratein measuring the volume of an irregular shaped

Figure 2 : Plot of volume calculated by mammography (Meanof volume measured by observer 1 and observer 2)and ultrasound (n=25)

AN APPROACH FOR ASSESSMENT OF TUMOR VOLUME FROM MAMMOGRAPHY IN LOCALLY ADVANCED BREAST CANCER

40

subject (less than 3% error). The percentagedifference between ultrasound and mammographyvolumes decreases after demagnification as shownin Table III. So mammography can also be used foraccurate measurement of the volume and can helpthe radiologists screen the breast cancer patients usedwith volume. The maximum tumor volume was88.71 cm3 and minimum tumor volume was 2.84cm3. According to response evaluation criteria inbreast tumors 4% patients shows the completeresponse, 25% shows the partial response, 63 %patients the stable disease and 8% shows the diseaseprogression after chemotherapy in the present study.

Volumetric measurement can helpsignificantly for evaluating therapeutic responsecompared to unidimensional and bidimensionalmeasurements. The ability to accurately calculatethe in vivo volume of tumors also has clinicalpotential especially in oncology. Tumor volumescould be used to assess the progression of diseaseand success of therapy, such as response toirradiation or chemotherapy. Tumor volume usingellipsoid formula is preferred over other cubical andspherical methods. As most of the tumors are ofirregular shapes so spherical formulae are not usedand the cubical formulae overestimate the volume.For treatment evaluation stress is on reproducibilityand accuracy of serial measurements. The rule forobtaining bidimensional measurement of a tumor isto find its longest diameter (length) and thedimension which is perpendicular to length. Thismethod provides a fixed set of values that are notaltered so producing reproducibility as right anglelaw is followed. Here is the first study to calculatethe tumor volume using mammograms. However,manual measurement of tumor volume is timeconsuming and estimates based on simple geometricshapes are inaccurate. The errors, whenever theyoccur, are due to error in marking the tumor on themammography film, which become significant incase of speculated, vague and diffused tumor massinstead of solid and well defined one. In routinereporting of regression or progression of tumor fromthe imaging film radiologists either use scale or tryto infer the tumor size by visual assessment of theirregular tumor part appearing in the images. In sucha situation a change of tumor sizes up to 20% mayallude by visual perception [9-11]. A carefulstatistical analysis of the measurements of 16experienced oncologists showed the false positiveresponse rate in unchanged masses as measured bysame investigator at 25% when 25% regressioncriteria (product of two diameters) was employed

[12]. Even volumes by diameters measurement bycalipers are reported to have large errors andmeasured the three mutually perpendicular diametersin mouse tumors by calipers and calculated tumorvolume by the product of these diameters [13].Tumors were excised and volumes were measuredby Archimedes’s principle. Their correlation showedan error of 50% on mean diameter measurement ata given volume. Area (product of two perpendiculardiameters), volume and diameter of tumor mass arefrequently used in practice to indicate the size,growth or regression.

CONCLUSION

Tumor volume is a more representativequantity of three dimensional tumor mass which isgenerally of irregular shape and size. Thus, it isdesirable to use mammography based tumor volumeas response indicator which is not only accurate butbetter representative of tumor mass. This approachof assessment of tumor volume can be used fordevelopment of mammography computedtomography 3-D scans.

Corresponding Author :

Dr. Pratik Kumar (Ph.D)Medical Physics Unit,Department of RadiodiagnosisInstitute Rotary Cancer Hospital,All India Institute of Medical Sciences,New Delhi-110029, India.Tel: + 01126594448Fax: + 91-11-26588663Email: guraiims@gmail.com

References

1. Chopra R. The Indian scene Journal of clinicaloncology 2001; 15(suppl 8): 106-111.

2. Richard S, Breiman, John W. Beck, Melvyn Korobkinet al. Volume determinations using computedtomography AJR 1982; 138: 329-333.

3. Koga T, Morikawa Y. Ultrasonographic determinationof the splenic size and its clinical usefulness in variousliver diseases Radiology 1975; 115: 157-161.

4. Riemenschneider RA, Ehalen JA. The relativeaccuracy of estimation of enlargement of the liver andspleen by radiologic and clinical methods AJR 1965;94: 462-468.

5. A.B. Miller, B. Hoogstraten et al. reporting results ofcancer treatment Cancer 1981; 47: 207-214.

Gupreet Singh

41

6. Hayward Jl, Carbone PP, Heuson JC et al. Assessmentof response to therapy in advanced breast cancerCancer 1977; 39: 1289-1294.

7. Karnofsky DA, Abelman WH et al. The use of nitrogenmustards in the palliative treatment of carcinomaCancer 1948; 1: 634-656.

8. Therasse P, Arbuk SG, Eisenhauer EA, et al. Newguidelines to evaluate response to treatment in solidtumors J Natl Cancer Inst 2000; 92: 205-16.

9. Moertel CG, Hanley JA. The effect of measuring erroron the results of therapeutic trials in advanced cancer.Cancer 1976; 38: 388-394.

10. Aydin H, Richter E, Feyerbend T, Bohndorf W. In vivodetermination of tumor volume: what does a CTcalculation mean in recurrences of rectal carcinomaStrahlenther onkol 1990; 3: 204-210.

11. Kumar P, Rehani MM, Anand, Raina V, Rao K.Assessment of lymphnode tumor mass from CT scans:How good is diameter and visual assessment J MedPhy 1995; 3: 20-24.

12. Watson JV. The cell proliferation kinetics of the EMT6/M/AC mouse tumors at four volumes duringunperturbed growth in vivo Cell Tissue Kinetic 1976,9: 147-154.

13. Spears CP. Volume doubling measurement of sphericaland ellipsoidal tumors Med Pedai Oncol 1984; 4: 212-220.

AN APPROACH FOR ASSESSMENT OF TUMOR VOLUME FROM MAMMOGRAPHY IN LOCALLY ADVANCED BREAST CANCER