You are on page 1of 5

Tarja J.

Rissanen, MD Hanna
#{149} P. M#{227}k#{228}r#{228}inen,
MD #{149}Seija I. Mattila, MD
Eija-Leena Lindholm, MD Miija
#{149} I. Heikkinen, MD #{149}Heikki 0. Kiviniemi, MD

Breast Cancer Recurrence after Mastectomy:


Diagnosis with Mammography and US’

The authors reviewed mammograms T HE reported local failure rate after Helsinki,
0.1-mm
Finland)
focal spots. Mammograms
with 0.4-mm and
of the
and/or ultrasound (US) scans of 76 treatment of breast cancer with
lesions (67 patients) suspected of be- mastectomy ranges from 5% to 27% soft tissue surrounding the scar and the
ing recurrent breast cancer. All pa- (1,2). Although the prognosis for a greater pectoral muscle were obtained
with compression whenever there was
tients had previously undergone recurrent cancer in the chest wall af-
sufficient subcutaneous tissue. Mammog-
mastectomy. Sixty-one cases were ter mastectomy is not as favorable as raphy was performed with both mediolat-
malignant disease and 15 were be- that after lumpectomy (3-5), and de- eral oblique and craniocaudal views (Fig 1)
nign. Mammography and US were spite the fact that most patients even- in 120 women and with a mediolateral
complementary to clinical examina- tually develop distant metastases, oblique view (Fig 2) in 463. Magnified spot
lion in evaluating palpable lesions at there are reports of patients being compression views were obtained if a le-
the mastectomy site. Both imaging alive and free of disease 10-15 years sion was suspicious for carcinoma. The
methods revealed nonpalpable recur- after the treatment of such a recur- examinations were recorded on Kodak
rences. The sensitivity of US was rence (3,6). Mammography is an im- Min-R film with Kodak Min-R screens
(Eastman Kodak, Rochester, NY). Kodak
91%, whereas the sensitivities of din- portant adjunct to physical examina-
Lanex medium screens were used for
ical examination and mammography lion for the follow-up of patients with
magnified views.
were 79% and 45%, respectively. US breast cancer who have been treated US was performed by using the direct
was the best imaging method for with lumpectomy and radiation ther- contact method with real-time equipment
evaluating tumors in the chest wall apy (7). Ultrasound (US) has been (model SSA 77 or 100; Toshiba, Tokyo, Ja-
far from the scar and in the axilla, as found to be sensitive in the diagnosis pan) and a 7.5-MHz transducer. In cases
these could not be visualized on of postoperative complications and with an equivocal finding at US, US-
mammograms. Most recurrent can- benign abnormalities but inferior to guided fine-needle aspiration biopsy
cers were seen as circumscribed mammography in depicting local re- (FNAB) was performed freehand with a

masses at mammography and as hy- currence (8). According to earlier re- Cameco syringe holder (London, England)
connected to a 10-mL syringe and a 0.6-
poechoic lesions at US. Hyperechoic ports (8,9), neither mammography
mm-diameter needle.
recurrences could not be differenti- nor US has been found to be of im- Recurrent cancer in the mastectomy scar
ated from benign lesions. The value portance after mastectomy. The pur- or the soft tissues of the ipsilateral thoracic
of early detection of recurrences with pose of this study was to evaluate the wall or lymph node metastasis in the ipsi-
imaging methods remains to be seen. usefulness of mammography and US lateral axilla was suspected at the clinical
for the diagnosis of local and regional or imaging examination in 154 of the 833
breast cancer recurrences after mas- patients. The diagnosis was verified by
Index terms: Breast neoplasms, diagnosis,
00.12981, 00.33 Breast neoplasms,
#{149} postopera- tectomy. means of cytologic or histologic examina-
tive, 00.45 tion in 91 patients; mammography and/or
breast US was performed within 4 months
Radiology 1993; 188:463-467 before the biopsy. The mammograms
MATERIALS AND METHODS and/or US scans of 76 lesions (67 patients)
were available for review, and these cases
From January 1989 through December
comprise the study. Mammograms and US
1991, 833 women underwent mammogra-
scans were reviewed by two radiologists
phy and/or US of the mastectomy site at
(SI.M. and E.L.L.) trained in mammogra-
least once at the Department of Diagnostic
phy and breast US independently and
Radiology as part of the routine follow-up
without knowledge of the clinical findings
protocol for breast cancer surgery. These
and the original interpretations. At mam-
patients had been referred from the De-
mography, a new or growing mass and/or
partment of Radiotherapy and Oncology,
microcalcifications were classified as suspi-
and the clinical examination was per-
formed by both the oncologist and the ra- cious for recurrence. Skin thickening was
1 From the Departments of Diagnostic Radiol- diologist. The median time from diagnosis considered a posttreatment change unless
ogy (T.J.R., H.P.M., S.I.M., E.L.L.), Radiotherapy to the last follow-up examination was 86
a progressive thickening was noted. The
and Oncology (MLH.), and Surgery (H.O.K), US criterion for recurrence was a focal hy-
months (range, 1-290 months). During
Oulu University Central Hospital, SF-90220 poechoic lesion. Anechoic and hyper-
this time, each patient underwent one to
Oulu, Finland. From the 1992 RSNA scientific
assembly. Received November 17, 1992; revision
12 (median, five) clinical and imaging
requested January 19, 1993; revision received breast examinations.
March 8; accepted March 29. Address reprint Mammograms were obtained with a
requests to T.J.R. dedicated microfocus tube mammo- Abbreviation: FNAB = fine-needle aspiration
0 RSNA, 1993 graphic unit (Mamex DC Mag Soredex, biopsy.

463
echoic lesions were classified as benign.
The clinical and histologic data were ob-
tamed by reviewing patient charts.

RESULTS

Among the 76 reviewed cases, 61


malignant lesions were diagnosed in
52 patients and 15 benign lesions
were diagnosed in 15 patients. These
76 cases were reviewed for mammo-
graphic (n = 7), US (n = 31), or both
mammographic and US (n = 38)
findings.
Of the 61 recurrent cancers, six
were found at clinical examination
alone and 42 at clinical examination
and mammography and/or US. Thir-
teen lesions were nonpalpable and
were found at US and mammography
(n = 3), mammography alone (n = 3),
or US alone (n = 7) (Table 1). Of the
48 palpable recurrences, US was con-
sidered the primary method of diag- a. b.
nosis in four, and mammography
Figure 1. (a) Mediolateral oblique and (b) craniocaudal mammograms of a right mastectomy
helped diagnose two recurrences scar. A nonpalpable circumscribed tumor 8 mm in diameter is seen in the subcutaneous tissue.
with an indeterminate finding at pal- US revealed a well-defined hypoechoic lesion (cf Fig 4b). Invasive ductal cancer was diag-
pation (Fig 2). Seven of the 61 recur- nosed at histologic examination.
rences were situated in the skin at the
mastectomy site, 46 in the subcutane-
ous tissue, and one deep in the chest (Figs 3b, 4a-4c); 29 had a sharp con-
wall. Seven recurrences were re- tour and 22 had an irregular contour.
gional. At US, the median size of the lesions
Accuracies of clinical, mammo- was 10 mm (range, 5-90 mm), and the
graphic, and US examination are pre- anteroposterior diameter was greater
sented in Table 2. Of the 45 cases in than the transverse diameter in five
which mammograms were available, lesions (Figs 3b, 4a). In 41 cases, no
38 were malignant and seven were acoustic attenuation could be seen
benign. Fourteen recurrences were posterior to the tumor. An acoustic
seen as a circumscribed tumor (Fig 1) shadow and acoustic enhancement
and three as a stellate tumor (Fig 2). were seen in five cases each. All hy-
All 17 of these tumors were true- poechoic lesions were true-positive at
positive at review. Mammography review except for one small, oval,
yielded false-negative findings in 21 well-defined lesion that was consid-
cases. At retrospective analysis, no ered to be a fat lobule. Four recurrent
abnormality was seen in 18 cases (Fig cancers had the appearance of a hy-
3a), and the only sign of a recurrence perechoic lesion (Fig 5a) and were
was skin thickening in three cases. considered benign scar tissue at re-
Mammography results were normal view and prospective US reports.
in six of the seven cases of benign dis- Findings at US were true-negative
ease. Recurrence was suspected in in nine of the 14 benign lesions. Four
one case with a stellate opacity and of these nine lesions were hyper-
microcalcifications. Fibrosis and vas- echoic, and the final diagnosis was fat
cular calcification were diagnosed at necrosis in two cases (Fig 5b) and fi- Figure 2. Mediolateral oblique mammo-
histologic examination. Clinical exam- brosis in two. In the remaining five gram of a left mastectomy scar with an inde-
ination had a sensitivity of 79%, a cases, US showed no abnormality terminate finding at palpation. A stellate tu-
specificity of 13%, and an accuracy of (n = 3) or revealed an anechoic lesion mor is in the subcutaneous tissue. Invasive
(n = 2). Recurrence was suspected on ductal cancer was diagnosed at histologic
66%. Mammography had a sensitivity
examination.
of 45%, a specificity of 86%, and an review of five hypoechoic lesions,
accuracy of 51 % . US had a sensitivity with foreign body reaction as the final
of 91%, a specificity of 64%, and an diagnosis in two cases, hematoma in
accuracy of 86%. one, and fibrosis (Fig 4d) in two. The diagnosed by means of surgical bi-
Of the 69 cases in which US scans median size of the benign lesions was opsy and histologic examination and
were available, 55 were malignant 20 mm (range, 5-46 mm). 17 by means of FNAB and cytologic
and 14 were benign. Fifty of the 55 US-guided FNAB was performed in examination. All lesions diagnosed as
recurrences (91 % ) were true-positive 57 lesions, yielding a sample sufficient benign at cytologic examination were
at review. Fifty-one recurrent cancers for cytologic analysis in 55 (Table 3). confirmed with histologic examina-
were seen as hypoechoic lesions at US Of the 61 recurrent cancers, 44 were tion. The sensitivity of FNAB was

464 Radiology
#{149} August 1993
Table 1
Mode of Detection and Location of 61 Cases of Malignant Disease Recurrences

Location

Mastectomy Scar

Mode of Detection Skin Subcutaneous Deep Axilla Total

Clinical examination 3 3 0 0 6
clinical examination and mammography 0 2 0 0 2
Clinical examination and US 4 22 0 5 31
Clinical examination, mammography,
andUS 0 8 0 1 9
US and mammography 0 3 0 0 3
Mammographyalone 0 3 0 0 3
USalone 0 5 1 1 7

Total 7 46 1 7 61

Table 2
Accuracy of clinical Examination, Mammography, and US in Detecting Local
Recurrences

Clinical Examination Mammography US


(ii = 76) (n = 45) (n = 69)

FinalDiagnosis - + - + - +

Benign 2 13 6 1 9 5
Malignant 13 48 21 17 5 50

Note.-- = no recurrence, + = recurrence.

a. for 2-59 (median, 17) months after and for the follow-up of patients with
treatment of recurrence. Of 39 pa- silicone gel prostheses.
tients with a palpable lesion, five Modified radical mastectomy leaves
(13%) have died of breast cancer and more subcutaneous tissue than does
34 (87%) are alive, 15 (38%) with no radical mastectomy, which makes
evidence of disease after 10-59 (medi- compression mammography of the
an, 25)-month follow-up, five (13%) mastectomy scar technically possible
with recurrence, and 14 (36%) with in most cases. Most (70%) of our pa-
metastases. All 13 women with non- tients underwent mammography of
palpable lesions are alive, eight (62%) the mastectomy site with at least a
with no evidence of disease after 2-57 mediolateral oblique view. Because of
(median, 24)-month follow-up, one the thick subcutaneous tissue, it is not
(7%) with recurrence, and four (31%) always easy to differentiate between
with metastases. benign and malignant lesions with
clinical examination alone. Although
b.
most (79%) of the recurrences in this
Figure 3. (a) Mediolateral oblique mammo- DISCUSSION
series were found at palpation, some
gram of a right mastectomy scar shows no
abnormality. (b) US scan shows a small, hy- Mammography is an important abnormality was identified at clinical
poechoic tumor that was palpated near the complementary method to clinical examination in 87% of the benign le-
lateral end of the scar. A needle tip is inside examination for the follow-up of sions as well. In six of the 48 cases of
the tumor. Ductal cancer was diagnosed at breast cancer patients treated with palpable recurrence, the finding at
cytologic examination. lumpectomy and radiation therapy palpation was considered to be inde-
(7). Imaging modalities have not been terminate.
widely used for examining the scar After mastectomy, fewer false-posi-
94%. US-guided FNAB yielded malig- and axilla after mastectomy and axil- tive mammographic findings are
nant cells in all 10 cases of nonpalpa- lary dissection because recurrent
hi- caused by posttreatment changes
ble recurrence seen at US. The me- mors involving the chest wall or skin than after lumpectomy. In an uncom-
dian diameter of the lesions sampled are frequently detected at clinical ex- plicated postoperative situation only
for biopsy was 12 mm (range, 5-90 amination. Mendelson (9) obtained skin, subcutaneous fat, and the
mm), and true-positive findings were lateral views of the axilla without greater pectoral muscle are visualized.
obtained from lesions as small as 5 compression, and the only abnormal- In our study, only one mammogram
mm in diameter at cytologic examina- ity he found was bone metastasis. yielded a false-positive finding. The
tion. Balu-Maestro et al (8) found US valu- disadvantage of mammography is the
The 52 patients with recurrence of able only for evaluating the depth of poor ability to visualize lesions situ-
malignant disease have been followed penetration of cutaneous recurrences ated deep in the muscle layer, some

Volume 188 Number


#{149} 2 Radiology #{149} 465
distance away from the scar, or in the
axilla. Mammography offered no ad-
ditional information to the clinical
findings in examining skin lesions
because it was technically impossible
or because the skin thickening repre-
senting the recurrent tumor could not
be differentiated from posttreatment
changes. For these reasons, the sensi-
tivity of mammography in detecting
recurrent cancer was inferior to that
of both clinical examination and US.
In the cases in which the lesion was a. b.
visualized, mammography was the
first method to reveal the abnormal-
ity; US was used as a complementary
method to locate the tumor for FNAB.
The sensitivity of US for detecting
local recurrences was superior to that
of palpation and mammography.
Most difficulties were encountered in
analyzing the findings seen at US.
Most recurrent cancers appeared as
hypoechoic lesions at US, which con-
forms to the pattern described in pre-
viously published reports (8,10). In
10% of the hypoechoic lesions, the
anteroposterior diameter was greater
than the transverse diameter, a find- C. d.
ing not seen in benign lesions. No Figure 4. US scans of four hypoechoic lesions. (a) A palpable hypoechoic tumor with acous-
additional features suggestive of a tic enhancement is seen near the mastectomy scar. Ductal cancer was diagnosed at cytologic
examination. (b) A nonpalpable hypoechoic lesion is seen at the mastectomy site. Invasive
malignant lesion could be found. As-
ductal cancer was diagnosed at histologic examination. r = ribs. (c) A palpable hypoechoic
sessment of the attenuation of the
tumor is seen between the skin and a silicone gel prosthesis (*). A serous fluid with malignant
ultrasound beam after it has passed cells was aspirated under US guidance. Surgical biopsy revealed a cystic tumor. Invasive duc-
through the lesion was of no help in tal cancer was diagnosed at histologic examination. (d) A palpable hypoechoic lesion is seen in
differentiating a recurrent cancer the axilla, with acoustic shadowing posteriorly. Fibrosis was diagnosed at a histologic exami-
from a benign lesion. The attenuation nation.
shadowing exhibited by malignant
breast masses appears to be related to
the amount of fibrosis in the tumor
(11), and circumscribed types of can-
cer with less fibrosis may resemble
benign lesions with a sharp contour
and posterior acoustic enhancement
(12). This was the case with most re-
currences in this study; they ap-
peared as well-defined lesions with-
out posterior shadowing at US.
Hypoechoic fibrotic lesions or foreign
body granulomas, some of which
were also associated with a posterior
a. b.
acoustic shadow, were classified as
Figure 5. US scans of two hyperechoic lesions (arrows) in mastectomy scars. Histologic ex-
suspicious for recurrence. Four re- amination revealed invasive recurrent cancer (a) and fat necrosis (1,). * in b indicates a cyst.
currences manifested as hyperechoic
lesions, a finding that, to our knowl-
edge, has not been reported previ-
ously. These findings were identical aspiration and were not included in ing recurrences when only radiation
to those of the cases diagnosed as fat the study because there was no need therapy, chemotherapy, or hormonal
necrosis at histologic examination. No for surgical biopsy and histologic veri- therapy was indicated.
recurrences were found at biopsy of fication. Among the 48 recurrent Thirteen (21%) of the 61 malignant
palpable lesions with normal US cancers examined with US-guided recurrent cancers in our study were
scans. FNAB, only three (6%) had acellular found at imaging examination alone.
US is superior to mammography for or false-negative findings at cytologic The question of whether the detec-
the detection of postoperative compli- examination. FNAB was useful for tion of occult recurrences with rou-
cations and is the method of choice obtaining an accurate preoperative tine use of imaging methods changes
for guiding fluid aspirations (8). In diagnosis of both palpable and non- the overall outcome of the patients
our practice, fluid collections are diag- palpable recurrences and was also remains unanswered at this time. The
nosed and treated with US-guided used as the only method for diagnos- reported overall survival rate 5 years

466 Radiology
#{149} August 1993
3. Chu FCH, Lin FJ, Kim JH, Huh SH, Garma-
tis CJ. Locally recurrent carcinoma of the
breast: results of radiation therapy. Cancer
1976; 37:2677-2681.
4. Bedwinek JM, Lee J, Fineberg B, Ocwieza
M. Prognostic indicators in patients with
isolated local-regional recurrence of breast
cancer. Cancer 1981; 47:2232-2235.
5. Kurtz KM, Amalric R, Brandone H, et al.
Local recurrence after breast-conserving
surgery and radiotherapy: frequency, time
course, and prognosis. Cancer 1989; 63:
1912-1917.
6. Probstfeld MR, O’Connell TX. Treatment
of locally recurrent breast carcinoma. Arch
Surg 1989; 124:1127-1130.
7. Dershaw DD, McCormick B, Osborne MP.
Detection of local recurrence after conser-
vative therapy for breast carcinoma. Can-
cer 1992; 70:493-496.
8. Balu-Maestro C, Bruneton JN, Geoffray A,
Chauvel C, Rogopoulos A, Bittman 0. UI-
trasonographic posttreatment follow-up of
breast cancer patients. J Ultrasound Med
1991; 10:1-7.
9. Mendelson EB. Evaluation of the postop-
erative breast. Radiol Clin North Am 1992;
30: 107-138.
10. Leucht WJ, Rabe DR. Sonographic find-
ings following conservative surgery and
irradiation for breast carcinoma. Ultra-
from diagnosis of a local recurrence group with nonpalpable recurrences sound Med Biol 1988; 14:27-41.
11. Cole-Beuglet C, Soriano RZ, Kurtz AB,
after mastectomy ranges from 21 % to compared with those with palpable
Goldberg BB. Ultrasound analysis of 104
36% (3,4,13), and the prevalence of recurrences, no definitive conclusions primary breast carcinomas classified ac-
distant metastases ranges from 45% to can be made because of the limited cording to histopathologic type. Radiology
more than 90% (2,6). A recurrence is number of patients and the short fol- 1983; 147:191-196.
12. Kasumi J, Fukami A, Kuno K, Kajitani T.
not necessarily a sign of a systemic low-up period. Further studies with
Characteristic echographic features of cir-
disease with poor prognosis, and pa- larger groups of patients are needed cumscribed cancer. Ultrasound Med Biol
tients with one small recurrent cancer to resolve this question. U 1982; 8:369-375.
seem to survive longer than those 13. Magno L, Bignardi M, Micheletti E, Bardelli
D, Plebani F. Analysis of prognostic fac-
with large and/or multiple recur- References
tors in patients with isolated chest wall re-
rences (3,4,6,13). However, to our 1. Zimmerman KW, Montague ED, Fletcher
currence of breast cancer. Cancer 1987; 60:
knowledge, the effect of imaging fol- GH. Frequency, anatomical distribution
240-244.
and management of local recurrences after
low-up on survival is not known. Al- definitive therapy for breast cancer. Cancer
though there is a trend toward a bet- 1966; 19:67-74.
ter outcome among the patients in the 2. Valagussa P, Bonadonna G, Veronesi U.
Patterns of relapse and survival following
radical mastectomy: analysis of 716 consec-
utive patients. Cancer 1978; 41:1170-1178.

Volume 188 Number


#{149} 2 Radiology 467
#{149}

You might also like