Ultrasound in Med. & Biol. Vol. 14, Sup. i, pp. 175-182, 1988 0301-5629/88 $3.00 + .00 Printed In the U.S.A. (c) 1988 Pergamon Press plc

CAN MICROCALCIFICATIONS LOCATED WITHIN BREAST CARCINOMAS BE DETECTED BY ULTRASOUND IMAGING ?

Fujio Kasumi, MD.

Department of Surgery, Chief of Ultrasonic Laboratory Cancer Ins t i tu te Hospital, Tokyo, Japan

ABSTRACT

The author succeeded in detecting microcalc i f icat ions within breast carcinomas, by ultrasound(US) imaging, at the beginning of 1982, and published the results of his c l i n i ca l and experimental researches at the Third Internat ional Congress on the Ultrasonic Examination of the Breast held on June I0 - 12, 1983, in Tokyo, Japan~)

After that , more deta i led c l i n i ca l evidence was acquired and new experiments, using breast phantoms, were carr ied out. This paper introduces an experiment using phantoms which shows that US can detect and iden t i f y even lO0-500M-sized t iny objects in an " idea l " low echoic area s imi lar to microca lc i f icat ions wi th in breast carcinomas. Actual c l i n i ca l evidence is also presented.

Key Words : M ic roca lc i f i ca t ion , US detection, Breast Phantom, Bio-Gel.

INTRODUCTION

The author fabr icated breast cancer phantoms using d i f fe ren t densit ies of agar and d i f fe ren t sizes of minute glass beads o r i g i n a l l y designed to test the resolut ion of xerography. These phantoms were examined to determine whether US could detect minute glass beads less than Imm in diameter. I t was found that beads as small as I00~ were imaged by US as strong and impressives spots within the low echoic areas of the Phantom.

However, i t was suggested that such phantoms were "too ideal" in terms of a c l i n i ca l picture. J. W. Hunt 2~ of the Ontario Cancer Ins t i t u te reported the use of "Bio-Gel" acrylamide, held wi th in an agar gel matrix, in which small scatters were used to mimic parenchyma, adipose t issue and cancer areas of the breast~) The author used such phantoms which were found to be very good with images closely resembling c l i n i ca l US images. Various sizes of glass beads were enveloped in the cancer area before the sol became a sol id gel. Phantoms were examined by US and xerography and then were compared with each other.

MATERIALS AND METHODS

Medical powder agar for a bacter ia l cul ture medium and acrylamide (which is usual ly used for electrophoresis and is avai lab le under the name of "Bio-Gel", manufactured by Bio-Rad laborator ies, catalog No. 150-1030) were used. In p last ic boxes, I% Bio-Gel in 2% boiled agar solut ion was mixed to simulate preglandular (sub-cutaneous) adipose t issue and, a f ter 15 minutes, 5% Bio-Gel in boi led 2% agar solut ion was added to form a second layer to simulate breast glandular t issue. After these layers become sol id , index- f inger - t ip -s ized holes were cut from behind and were f i l l e d with a 0.5% Bio-Gel solut ion containing a small number of glass beads of d i f fe ren t sizes. These beads were embedded in the solut ion in such a way that they did not overlap each other e i ther a x i a l l y or l a t e r a l l y . These beads were designed by XEROX for test ing the resolut ion of xerography. This arrangement was meant to simulate breast cancers with embedded microca lc i f ica t ions. To simulate the retroglandular adipose t issue, more I% Bio-Gel solut ion was added last to form a th i rd layer. Fi 9. 1 Removed from the i r p las t ic boxes, these phantoms were examined by US and xerography. Six phantoms which contained speci f ic sizes of glass beads were imaged by US and xerography.

The US apparatus used was a Hitachi EUB2B mechanical l inear motion system. The transducer was a modified PVOF, 7.5 MHz single focus, one inch un i t . Fig. 2 The phantoms and probe were immersed in water and scanned. The xerography apparatus was a XEROX 125.

175

176 5th International Congress on the Ultrasonic Examination of the Breast

B r e a s t C a n c e r P h a n t o m

" " : : : . i 5%

- 1 %

Glass Beads Cancer 0.5%

of Bio-Gel in 2% Agar

Fig. 1

Fig. 2-1

Fig. 2-2

5th International Congress on the Ultrasonic Examination of the Breast 177

IDENTIFICATION OF GLASS BEADS IN AGAR

BY ULTRASOUND AND XEROMAMMOGRAPHY

MEAN ~AMETER OF BEADS ULTRASOUND XEROMAMMOGRAPHY

I 1100p

2 460

3 300

4 200

5 160

6 110

Fig. 3

178 5th Internatlonal Congress on the Ultrasonlc Examination of the Breast

RESULTS

Results are shown in ~Fig'3 s By xerography, 160~ was the lower l imi t for ident i f icat ion of the glass ; for a size-llO~ bead, detection was almost impossible. On the other hand, even llO~ beads were easily detected by US and adequately imaged separated ax ia l ly and latera l ly in low echoic cancer areas. Fig. 4,5 However, because of differences in resolution power, the size of glass beads was more accurately determined by xerography than by US.

,5 160p Fig. 4

6 110k Fig. 5

When the glass beads were enveloped di f fus ively within the 5% Bio-Gel area,however, even 460~-sized beads could not be detected by US because they exhibited the same ecogenicity as the 5% Bio-Ge]. Fig, 6 This i nab i l i t y to detect is similar to the fact that scattered microcalcifications found in f ibrocystic diseases or noninvasive ductal carcinoma without making remarkable mass can not be detected by US.

Identification of Glass Beads in

5% Bio-Gel,2% Agar

420--500

Fig.6-1 Fig:6-2

5th International Congress on the Ultrasonic Examination of the Breast 179

DISCUSSIONS

The lOdB down resolution of the modified PVDF polymer transducer used in these studies is 800~ in the lateral direction and 400~ in the axial direction. Diameters of breast microcalcif ication are usually under Imm, with the majority between lO0 and 500~. The experiment described here shows that US can detect glass beads smaller than i ts theoretical resolution power, providing they exist in ideal low echoic areas. The images of the glass beads, shown as strong spots, were usually without acoustic shadows.

Cl in ica l ly , when mammography (MMG) shows microcalcifications inside breast cancers, US detects f ine, strong echoes located within the low echoic cancer areas. Fig, 7-11 By c l in ical and pathological findings, these echogenic spots correspond to the microcalcif ications.

Fig.7-1 Fig.7-2

Fig.8-1 Fig.8-2

Fig.9-1 Fig.9-2

Fig.lO F ig . l l

From the ultrasonographical point of view, this experiment connected US and xeromammography. US should detect minute calci f icat ions between l O0 and 500~ in diameter separated from each other, since breast cancers are imaged as low echoic areas, with the exception of scirrhous, shadow-type cancers.

180 5th International Congress o n the Ultrasonic Examination of the Breast

Fig. 12 shows the mutual re lat ionship between MMG and US in one hundred cases of breast cancer pat ients. By US, 25 + 56 cases were correct ly diagnosed, 13 were underdiagnosed, mainly because of the poorer resolut ion of US in comparison to MMG. The fact that MMG is a whole survey, while US is tomographic, must also be considered. Six of the cases were overdiagnosed by US. I t is true that US images strong echogenic spots some of which are not microca lc i f icat ions. For MMG, su f f i c ien t Ca3(P04)4 deposition

may be necessary before the i den t i f i ca t i on of microca lc i f ica t ion can be made. I t is, however, probable that for US, s ign i~can t calcium deposit ion may not be necessary.

MMG

1 +

2 +

3

4

DE'I'I~CTICI~] OF MICROCALCIFICATIONS

by MMG & US

US

+ 25 correctly detected

- 13 under diagnosed

- 56 correctly denied

+ 6 over diagnosed

* Detection of MicrocalcJfications by MMG 38%

* Sensitivity of US 25/25+13 63

* Specificity of US 56/56+6 90.3

* Accuracy of US 25+56/100 81

Fig.12

Unti l the end of 1980, nothing had been published world-wi le about US detection of microcalc i f icat ions in breast cancers, and i t had been thought to be absolutely impossible to detact them by PZT 3-5MHz transducers. Under th is circumstance, in 1978 Kossoff, Je l l i ns and Reeve 4)reported a case of the v isua l i za t ion of macrocalci f icat ion as a br ight spot in a s tage- l l scirrhous carcinoma, but mentioned that the resolut ion of the System-I echoscope did not allow the display of the microca lc i f ica t ions.

In 1980, Harper and Kelly-FryS) wrote that (1)US was not as capable as of c lear ly imaging microca lc i f icat ions of 0.2 - O.5mm in diameter, (2)the scanning device with the 3.7MHz single focus transducer was not designed to detect microca lc i f ica t ions, and (3)US could detect ca l c i f i ca t i ons larger than O.5mm; they reported a case of the v isua l iza t ion of macrocalci f icat ions in fibroadenomas.

In 19806)and 1982~) Cole-Beuglet described a case of non-palpable cancer with microcalc i f icat ions v isual ized by System-I as a mass containing br ight echoes with poster ior shadowing.

In 1982, Kopans, Meyer and SteinbockS)reported that two cases of clusters of microca lc i f icat ions seen on xeromammography were visual ized on the ultrasonograms by a real- t ime apparatus with a 3.2 MHz sector transducer as echogenic foci without evidence of shadowing.

In 1979, the author adopted a PVDF 7.5MHz transducer as a subst i tute for the PZT 5oOMHz transducer in the EUB 2B mechanical l inear motion system; in 1981, a modified PVDF 7.5MHz convex 1 inch transducer was used with far better resolut ions.

By 1982 the author became convinced of the poss i b i l i t y of c lear ly catching microca lc i f icat ions in breast carcinomas by US, but when viewed ret rospect ive ly , i t had already succeeded in 1980 without his notice. The resul t was reported at the 41st Japan Society of Ultrasonics in Medicine in December, 19829) and - in de ta i l - at the Third Internat ional Congress on the Ultrasonic Examination of the Breast~ )I°)

As the PVDF transducer prevai led, the v isua l i za t ion of microcalc i f icat ions in cancers was recognized by almost a l l invest igators in Japan i f the cancers were described in low echoic areas. In the United States, Jackson and £. Kelly-Fry11)reported the same results with the PVDF 7.5MHz transducer.

5th International Congress on the Ultrasonic Examination of the Breast 181

In 1983, however, Lambie et al~ 2) mentioned that microcalcification could be seen on the sonogram only when they exceeded the resolution of the transducer.

In 1986, Fil ipczynski, Kujawska and Lypacewicz~3) in a theoretical analysis of the detectabi l i ty of microcalcifications by US imaging combined with some experimental studies to determine the interference background of a normal breast, concluded that microcalcifications were not detectable by the ultrasonic echo method.

These reports stand against the author's results and above a l l , against his actual c l in ica l experience. His experiment proved that minute glass beads that had high acoustic impedance were clearly detected separately by US imaging in circumscribed low echoic areas. From the c l in ica l experience, microcalcifications in breast carcinomas are thought to have at least the equal acoustic impedance of minute glass beads because their irregular shapes and abundant calcium deposition are very suitable as scatterers of US.

Becouse of these facts and i ts rather poor resolution, US displays microcalcif i- cations as bigger than actual size.

Acknowledgements: The author grateful ly acknowledges the advice of Dr. J. W. Hunt, in performing the experiment and the kind help of Dr. E. Kelly-Fry in composing this paper.

CONCLUSIONS

l )

2)

3)

4)

5)

6)

Microcalcifications in cancers are shown on echograms as very fine spots usually without acoustic shadows.

These spots are seen in the ideal low echoic cancer areas. This re la t ionsh ip is s imi lar to "stone echoes" in a gal lb ladder.

Experiments detect minute glass beads including those lOOu in diameter, despite poor resolution, provided they are scattered in an ideal background.

Microca lc i f i ca t ions in benign lesions, especia l ly in f i b rocys t i c diseases, are not detectable by US, wi th some exceptions, because they are not present in circumscribed low echoic areas.

From actual c l i n i c a l images and on the basis of the phantom experiment, these "spots" w i th in ultrasound images are thought to correspond to spec i f ic m ic roca lc i f i ca t ions .

S e n s i t i v i t y of the c l i n i c a l detection of m ic roca lc i f i ca t ions by US in MMG-proven cancers can be demonstrated to be 63%, Specif ic i ty-90.3%, and Accuracy-81%.

l )

2)

3)

4)

5)

REFERENCES

Kasumi, F. and Tanaka, H. (1983) Detection of Microcalcifications in Breast Carcinoma by Ultrasound. Ultrasonic Examination of the Breast, Edited by J. Jel l ins and T. Kobayashi : pp. 89 - 97 John Wiley & Sons

Hunt, J. W., Personal Communication, Physics Div is ion, Ontario Cancer I n s t i t u t e , 500 Sherbune Street, Toronto, Ontario, Canada M4XIKg.

Madsen E. L. , Zagsebski, J. A., Banjavic, R.A. and J u t i l a , R.E. (1978) Tissue Mimicking Materials fo r Ultrasound Phantoms, Medical Physics 5(5) : 391-394

Kossoff, G., J e l l i n s , J. and Reeve T.S. (1978) Ultrasound in the Detection of Early Breast Cancer, Early Diagnosis of Breast Cancer, Edited by Grundmann, E. and Beck, L. pp. 149-158 Gustar Fischer Verlag. S tu t tgar t , New York

Harper, P. and Kel ly -Fry , E. (1980) Ultrasound V isua l iza t ion of the Radiology 137 (2) : 465 - 469

Breast in Symptomatic Pat ients , -

182 5 t h I n t e r n a t i o n a l C o n g r e s s on t h e U l t r a s o n i c E x a m i n a t i o n o f t h e B r e a s t

6) Cole-Beuglet, C., Kurtz, A.B., Rubin, C.S. and Goldberg, B.B. (1980) Ultrasound Mammography Radiologic C l in ics of North America, 18 ( I ) 133-143

7) Cole-Beuglet, C. V. (1982) Sonographic Manifestations of Malignant Breast Disease. Semin. Ultrasound 3 ( I ) : 51

8) Kopans, D.B., Meyer, J.E. and Steinbock, R.T. (1982) Breast Cancer : The Appearance as Delineated by Whole Breast Water-path Ultrasound Scanning, J. Cl in . Ultrasound 10 : 313

9) Kasumi, F., Tanaka, H. and Gondo, M. (1982) Detection of Mic roca lc i f - icat ions in Breast Carcinoma by U.S. ( in Japanese) Japanese J. Cl in . Ultrasonics 41 : 155 - 156

I0) Kasumi, F. (1983) Ultrasound of the Breast Lesions ( in Japanese) pp. 138-156 Shinohara Publishes Co., Tokyo

I I ) Jackson, V. P., Kelly-Fz~/, E., Rothschild, P.A., Robert, W.H. and Clark, S.A. (1986) Automated Breast Sonography Using a 7.5MHz PVDF Transducer : Prel iminary C l in i ca l Evaluation Radiology 159 (3) : 679 - 684

12) Lambie, R. W., Hodgden, D. Herman, E. W. and Kopperman, M. (1983) Sonomammographic Manifestat ions of Mommographically Detectable Breast M ic roca lc i f i ca t ions , J. Ultrasound Med. 2 : 509 - 514

13) F i l i pczynsk i , L., Kujawska, T. and Lypacewicz, G. (1986) Ultrasonic Echo Method in Detection of Breast Ca lc i f i ca t ions-Trans ien t Analysis Arch. Acoust. 3 : 287 - 298