| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Breast Imaging |
1 From the Departments of Radiology (C.S.G., D.M.K., R.R.) and Surgery (M.P.O.), New York Presbyterian Hospital–Weill Medical College of Cornell University, Strang Cornell Breast Center, New York. Received June 17, 1998; revision requested August 6; final revision received November 17; accepted March 26, 1999. Address reprint requests to C.S.G., Department of Radiology, Memorial Sloan-Kettering Cancer Center, 1275 York Ave, New York, NY 10021.
 |
Abstract |
|---|
 TOP Abstract IntroductionMATERIALS AND METHODSRESULTSDISCUSSIONReferences |
|---|
MATERIALS AND METHODS: Medical records and mammographic findings were retrospectively reviewed; 26 patients with 27 local recurrences after breast-conservation therapy were identified.
RESULTS: Primary histopathologic findings included six in situ and 20 invasive carcinomas. Of the 27 recurrences, 19 (70%) were at or adjacent to the lumpectomy site and eight (30%) were elsewhere in the breast. All primary ductal carcinoma in situ (DCIS) cases manifested mammographically as microcalcifications and recurred as DCIS with microcalcifications. Eleven primary invasive carcinomas (10 masses, one case of microcalcifications) were detected only mammographically, three were detected only with physical examination, and six (six masses) were detected with both. Among these 20 recurrences, 14 (five masses, nine cases of microcalcifications) were detected only mammographically, one was detected only with physical examination, and five (five masses) were detected with both. Seventeen (85%) of 20 primary invasive carcinomas recurred invasively: 16 (94%) with similar histopathologic findings and eight (47%) with similar mammographic findings.
CONCLUSION: In local recurrence after breast-conservation therapy for DCIS, histopathologic findings, detection method, and mammographic findings are usually similar. Histopathologic findings of primary invasive breast carcinoma and local recurrence are usually similar, but the detection method and mammographic findings vary. This is relevant to the interpretation of new clinical or mammographic findings following lumpectomy.
Index terms: Breast neoplasms, calcification, 00.4535 • Breast neoplasms, diagnosis, 00.32, 00.33 • Breast neoplasms, surgery, 00.455, 00.458 • Breast neoplasms, therapeutic radiology, 00.459
|
Introduction |
|---|
|
TOPAbstractIntroductionMATERIALS AND METHODSRESULTSDISCUSSIONReferences |
|---|
Mammography is less sensitive in the treated breast than in the untreated breast due to surgery- and radiation-induced changes in the parenchymal pattern (8). The sensitivity of mammography in this population ranges from 55% to 68% in various studies (4–7). When detected mammographically, local breast cancer recurrence has a similar appearance as cancer in the untreated breast, which is most commonly a mass, microcalcifications, or both (4–8).
The appearance and histopathologic findings of locally recurrent breast cancer compared with those of the primary lesion have been evaluated in several studies. Liberman et al (9) compared imaging and histopathologic findings of primary ductal carcinoma in situ (DCIS) with those of local recurrences in 20 patients and found that 65% of local recurrences consist of pure DCIS, with the remaining 35% containing invasive and in situ carcinoma. In 14 patients, they compared mammograms at the time of primary diagnosis with those at the time of local recurrence and found the mammographic pattern and calcification morphology to be similar in 79% and 82% of patients, respectively. Philpotts et al (10) compared the mammographic appearance of primary breast cancer with that of local recurrence and found that the lesions had a similar appearance 81% of the time. Those authors suggested that knowledge of the mammographic appearance of a patient's primary breast tumor may influence the decision to perform biopsy of a mammographic finding in the treated breast and may improve the positive predictive value of biopsy in this mammographically challenging patient population.
We reviewed the clinical, imaging, and histopathologic findings in patients who developed local recurrence following breast-conservation therapy at our institution to correlate the histopathologic findings of primary tumors and local recurrences with the detection method and imaging features.
|
MATERIALS AND METHODS |
|---|
|
TOPAbstractIntroductionMATERIALS AND METHODSRESULTSDISCUSSIONReferences |
|---|
The patients were aged 28–85 years (mean, 57 years). Primary cancer in this population was diagnosed from June 1985 to December 1995; local recurrence was diagnosed from May 1990 to January 1998. The time from primary breast cancer diagnosis to local recurrence was 18–120 months (mean, 46 months). All patients were initially treated with wide local excision with or without axillary dissection. Initial tumor stage was known in 24 patients: Six had stage 0, 12 had stage I, and six had stage II disease. Two patients with invasive carcinomas smaller than 2 cm did not undergo axillary dissection due to concurrent medical problems. Margins were negative at histopathologic analysis in 18 patients, positive in one patient, and unknown in seven patients. Twenty-two patients underwent radiation therapy (one patient had treatment interrupted prior to the boost dose to the tumor bed), three refused radiation therapy, and radiation treatment information was unknown in one patient. Patients were evaluated clinically and mammographically at 6- or 12-month intervals on the basis of physician preference; during the study interval, no uniform protocol for follow-up surveillance existed at our institution. Magnification views of the lumpectomy bed were obtained at the discretion of the initial interpreting radiologist.
All available mammograms were retrospectively reviewed by two radiologists in conference (C.S.G., D.M.K.). All mammograms were considered technically adequate. Mammograms obtained at the time of primary diagnosis were reviewed simultaneously with mammograms obtained at the time of local recurrence, and findings were described by using the American College of Radiology's Breast Imaging Reporting and Data System (11).
Lesions were classified as masses, microcalcifications, or calcified masses. Primary lesions and local recurrences were classified as similar or dissimilar in appearance on the basis of these classifications; calcified masses were considered similar between primary and locally recurrent malignancies if either calcification or a mass was present mammographically. Calcified masses were classified in this manner because both masses and calcifications are primary signs of breast cancer and both may be prominent imaging features of tumors.
The location of local recurrences in the treated breast was classified as at or adjacent to the tumor bed, within the same quadrant, or in a different quadrant. According to the definition by Recht et al (3), recurrences at or adjacent to the tumor bed were considered "true recurrences or marginal misses" and those recurring in the same quadrant but at a distance from the lumpectomy site or in a different quadrant were considered "elsewhere." In this study, recurrences within 2 cm of the lumpectomy bed (marked by surgical clips or scarring) were considered to be true recurrences or marginal misses.
Medical records were reviewed (C.S.G.) to determine the clinical findings at the time of primary diagnosis and local recurrence, and histopathologic records were reviewed to determine the features of primary and locally recurrent cancers. Cancers were classified histopathologically according to the most aggressive component present; lesions containing invasive and in situ carcinoma were considered invasive. Microinvasive DCIS was classified as DCIS, but the presence of microinvasion was noted. Microinvasive DCIS was not classified as invasive carcinoma for the purpose of this retrospective study, since histopathologic definitions of microinvasion vary and since treatment in patients with microinvasive DCIS is similar to that in patients with DCIS but without microinvasion at our institution.
For the majority of tumors, size was determined histopathologically; when unavailable, size was determined mammographically. For lesions containing both invasive and intraductal carcinoma, the size of the invasive component was recorded; however, if an extensive, mammographically evident intraductal component was present, this was noted.
|
RESULTS |
|---|
|
TOPAbstractIntroductionMATERIALS AND METHODSRESULTSDISCUSSIONReferences |
|---|
The mean size of primary tumors was 12.6 mm (range, 3–26 mm), and the mean size of local recurrences was 17.3 mm (range, 3–35 mm). One patient had a 4-mm invasive recurrence with an extensive intraductal component; her recurrence mammographically manifested as a 30-mm area of fine microcalcifications. All 26 patients with mammographically evident local recurrences had lesions considered to be Breast Imaging Reporting and Data System category 4 or 5.
Ductal Carcinoma in Situ
In six patients, the primary tumor histopathologic diagnosis was DCIS. In one of these patients, the tumor was multifocal with foci of microinvasion. One patient had widely dispersed DCIS. The remaining four patients had focal DCIS histopathologically. In all six of these patients, lesions were detected only mammographically due to microcalcifications. The morphology and distribution of the primary tumor calcifications were pleomorphic and clustered in all six patients.
These six patients had seven local recurrences. One patient had two separate recurrences; the first recurred 53 months after the primary diagnosis and the second recurred 57 months after repeat lumpectomy. All local recurrences were DCIS histopathologically and again were detected mammographically due to microcalcifications (Table 1). In four patients (five local recurrences), microcalcification morphology and distribution in local recurrences were similar to those of the primary tumor. In one patient, the primary tumor appearance was pleomorphic and clustered and the local recurrence appearance was linear and branching with a linear distribution. In one patient, the primary tumor appearance was pleomorphic and clustered and the local recurrence appearance was pleomorphic with a linear distribution (Fig 1).
|
|
|
|
Among the 17 mammographically evident cancers, imaging findings indicated six spiculated masses; five ill-defined masses; three spiculated masses with microcalcifications; one lobulated, ill-defined mass with microcalcifications; one case of clustered, pleomorphic microcalcifications; and one mass with unknown characteristics.
Local recurrence in these 20 patients with primary invasive malignancies was detected only mammographically in 14, only at physical examination in one, and with both methods in five. Of the 19 recurrences detected mammographically, eight manifested as noncalcified masses: Three had spiculated margins, three had ill-defined margins, and two had obscured margins. Two recurrences manifested as calcified masses, one with ill-defined, lobulated margins and one with spiculated margins. Nine recurrences manifested as microcalcifications alone, which were linear in five and pleomorphic in four cases. The distribution in these nine microcalcifications was clustered in five and segmental in four.
Seventeen (85%) of 20 patients with primary invasive carcinomas had invasive local recurrences, 16 (94%) of these 17 patients had primary cancers and local recurrences with similar histopathologic findings. The remaining three (15%) of the 20 patients had cancer that locally recurred as DCIS. Of the 16 patients sharing similar invasive histopathologic findings between primary and locally recurrent carcinoma, 11 had recurrence at or adjacent to the lumpectomy site, and five had recurrence elsewhere in the breast. The degree of histopathologic differentiation in these 16 patients was the same in nine, less well-differentiated in two, better in one, and unknown in four. The 17th patient with an invasive recurrence had primary invasive lobular carcinoma, which recurred as invasive ductal carcinoma at the lumpectomy site. In the remaining three patients with primary invasive carcinoma, cancer locally recurred as DCIS at or adjacent to the lumpectomy site in two and elsewhere in one. All three of these recurrences were detected on the basis of pleomorphic microcalcifications. Correlation of histopathologic findings with location of recurrence is detailed in Table 2.
Imaging features in 17 of the 20 patients with invasive carcinoma in both primary and locally recurrent tumors were similar in eight (47%) and dissimilar in nine (53%) patients. In the group with similar imaging findings between primary and locally recurrent cancers, four primary masses recurred as masses (Fig 2), one calcified mass recurred as a calcified mass, two calcified masses recurred as masses, and one case of microcalcifications recurred as a calcified mass. In the patients with dissimilar imaging findings, five primary tumors manifesting as masses recurred as microcalcifications (Fig 3), three mammographically occult primary tumors had mammographically evident recurrences (two masses, one case of microcalcifications), and one primary tumor manifesting as a mass had a mammographically occult local recurrence. These data are outlined in Table 1.
|
|
|
|
|
DISCUSSION |
|---|
|
TOPAbstractIntroductionMATERIALS AND METHODSRESULTSDISCUSSIONReferences |
|---|
When tumor recurs in a conservatively treated breast, it may recur at the lumpectomy bed (true recurrence), adjacent to it (marginal miss), or elsewhere in the treated breast (3). It is theorized that tumor recurring at or near the primary tumor site represents a true treatment failure, while tumor recurring at some distance from the primary tumor site may represent either an initially unrecognized separate tumor focus or a new primary tumor (3). This theory is supported by the similar histopathologic findings found between primary and locally recurrent tumor at or adjacent to the primary tumor bed in a study by the same group of investigators (12).
Most of our local recurrences developing at or adjacent to the primary tumor site also shared similar histopathologic findings with the primary tumor, although similar histopathologic findings were also found between most primary tumors and local recurrences developing elsewhere in the breast in our series. It is possible that these latter cases may have represented multicentric or multifocal tumor. The time to recurrence at or adjacent to the primary tumor bed has been reported to be earlier than that of tumor recurring elsewhere in the breast. In the series by Recht et al (3), the median time to recurrence for true recurrences or marginal misses was 38.5 months compared to 64.5 months for local recurrences elsewhere in the treated breast. In the series by Fowble et al (13), the median interval to recurrence for true recurrences or marginal misses was 38 months, and that for recurrences elsewhere in the treated breast was 46 months. In a different series (2), while overall 79% of local recurrences were at or within the vicinity of the primary tumor bed, over time an increasing percentage of tumors were located at a distance from the lumpectomy site. The majority of local recurrences developing after 10 years in that study were located elsewhere in the breast.
The results of these different studies are in agreement with our data, in which the median time to recurrence at or adjacent to the lumpectomy bed was 28 months (range, 18–91 months) compared with 44.5 months (range, 36–120 months) for tumors recurring elsewhere. In all of these studies, as well as our own, the majority of local tumor recurrences were at or adjacent to the lumpectomy bed (true recurrences or marginal misses). Therefore, the tumor bed must receive particularly close attention following breast-conservation therapy, with careful assessment of any new mammographic findings. Careful physical examination of the lumpectomy bed is also important, since mammographic sensitivity has been reported to be most impaired in the lumpectomy quadrant (4).
Prior to the widespread use of mammography, DCIS manifested most often as a palpable mass; currently, it most commonly manifests as microcalcifications at mammography (14,15). A number of series (16–20) have demonstrated high rates of local control and overall survival for patients with DCIS treated with breast-conservation therapy and irradiation. However, several of these studies also reported protracted cases of local recurrence in DCIS (16,17,19,20), indicating the need for long-term mammographic surveillance. When DCIS locally recurs, it usually recurs in the lumpectomy quadrant (9,15). Primary DCIS may recur as pure DCIS or invasive ductal carcinoma. In the largest of these series (20), approximately half of local recurrences were invasive and half were in situ carcinoma. Authors of several studies (6,12,21) have reported a trend toward noninvasive histopathologic findings when local recurrence is detected with mammography alone. The majority of mammographically detected recurrences in these series contained microcalcifications.
The imaging appearance of local recurrence following breast-conservation therapy for DCIS has been assessed by Liberman et al (9). In that series, primary DCIS most often recurred as DCIS and usually manifested as microcalcifications within the lumpectomy quadrant. Although the number of primary DCIS cases in our series was small (n = 6), all recurrences were also DCIS that manifested mammographically as microcalcifications.
Invasive carcinoma is the most frequent breast carcinoma (22). Mammographically, invasive carcinoma most often manifests as a mass with or without microcalcifications; less commonly, invasive carcinoma may manifest as microcalcifications alone (23). Treatment of early stage invasive breast cancer with lumpectomy and radiation therapy has been shown in several studies to result in good local control and overall survival rates (12,13,24). When invasive carcinoma locally recurs, the histopathologic findings of the recurrence are usually invasive as well (10,12,13,24), although at times the tumor may recur as pure DCIS.
In our series, the majority (85%) of invasive carcinomas recurred invasively, but when they recurred as DCIS, they manifested as microcalcifications. Philpotts et al (10) compared the imaging findings of primary and locally recurrent breast cancers; they found most recurrences to be similar in imaging appearance to the primary tumor, particularly if the recurrence was within the lumpectomy quadrant. Most of the tumors in that study were invasive and recurred invasively most of the time. The imaging findings in that study contrast with our own, in which only 16 (59%) of 27 local recurrences had similarity between the mammographic appearance of the local recurrence and that of the primary tumor. Dissimilarity of appearance occurred when primary histopathologic findings were invasive, not in situ. Although most invasive tumors in our series recurred invasively, only 47% of these had a similar mammographic appearance as the primary tumor. Of note, six of the nine invasive local recurrences having a different appearance from the primary lesion manifested as only microcalcifications. In addition, the three cases of primary invasive carcinoma that recurred noninvasively contained only microcalcifications; two of these cases had an appearance dissimilar to that of the primary tumor. As microcalcifications were the only mammographic sign of local recurrence (three in situ, six invasive) in nine of 17 primary invasive carcinomas in our series, we believe that developing indeterminate microcalcifications at the lumpectomy site should be viewed with a high index of suspicion, regardless of whether the primary tumor was invasive or in situ and regardless of the appearance of the primary lesion. In this series, six (38%) of the 16 local recurrences that manifested as suspicious microcalcifications alone represented recurrent invasive carcinoma.
Two of the invasive primary carcinomas in our series were occult at initial imaging but locally recurred as a mass or as microcalcifications at mammography. Thus, mammographic surveillance was important even in patients in whom the primary tumor was occult. This has also been demonstrated by Samuels et al (25); in their study, four of six local recurrences following breast-conservation therapy in patients with occult primary tumors were mammographically visible. In the series by Schnitt et al (12), 89% of local recurrences manifesting as a new abnormality at physical examination were invasive carcinoma, while all of the local recurrences detected only as a new mammographic abnormality (all microcalcifications) contained predominantly or exclusively intraductal carcinoma. In the study by Liberman et al (9), only three patients with primary DCIS had a palpable local recurrence; however, in two of the three, the lesion contained invasive as well as in situ histopathologic findings. From these data as well as from our own, it is evident that when local recurrence is palpable, it is usually invasive.
It is surprising that the majority of our local recurrences were detected mammographically, given the decreased sensitivity of mammography shown by others in this population (4–7). The decreased sensitivity in detecting local recurrences compared with the sensitivity in detecting primary tumor has been ascribed to postoperative scarring and radiation changes (8), findings particularly pronounced in the lumpectomy quadrant.
It has been suggested that the mammographic detection of local recurrence manifesting as a mass may be more difficult than that of local recurrence manifesting as microcalcifications (10). This is supported by the series by Dershaw et al (6), in which 19 of 29 mammographically detected recurrences contained microcalcifications, and that by Orel et al (4), in which 10 of 21 mammographically detected recurrences contained microcalcifications. In the study by Orel et al (4), the majority of mammographically detected recurrences at the lumpectomy bed were detected because of the presence of microcalcifications. Similarly, 16 of 27 local recurrences in our series were mammographically detected as microcalcifications, although masses were also mammographically detectable.
There are several possible explanations for the high sensitivity of mammography in detecting local recurrence in our series. Local recurrences in our series were diagnosed more recently than those in other published series, and it is possible that the higher rate of mammographic detection is due to intervening improvements in mammographic technique since the 1970s and 1980s. Improvements in surgical and radiation therapy techniques could account for less severe surgical and radiation changes in the treated breast. Finally, heightened awareness in recent years of the need for close mammographic surveillance has allowed for the increased detection of subtle mammographic changes, which can herald local recurrence. At our institution, the current protocol for mammographic surveillance following adequate breast-conservation therapy includes biannual mammography for 2 years and annual mammography thereafter. Magnification views are obtained at the discretion of the interpreting mammographer.
A limitation of this retrospective study is the lack of information on surgical margins in seven primary tumors and information regarding the degree of histopathologic differentiation in four invasive recurrences. Surgical inking of margins was not routinely performed on specimens in prior years (3,12,16,17,19,20). In addition, as our medical center is a major referral center for the greater metropolitan area, some patients return to local centers for treatment after consultation at our institution. Therefore, some of the histopathologic material is no longer available for additional review.
In conclusion, in our series of local breast cancer recurrences following breast-conservation therapy, primary DCIS uniformly recurred as in situ carcinoma. Both primary and locally recurrent DCIS were detected solely mammographically as microcalcifications. Invasive carcinoma usually recurred invasively; less often, it recurred as in situ lesion. However, imaging features varied between primary invasive cancers and local recurrences. Although primary invasive carcinoma usually manifested clinically and/or mammographically as a mass, local recurrences in these patients were detected 45% of the time as microcalcifications alone; the majority of which represented invasive disease. This has important implications for the evaluation of new mammographic findings in the treated breast.
|
Footnotes |
|---|
Author contributions: Guarantor of integrity of entire study, C.S.G.; study concepts and design, C.S.G., D.M.K.; definition of intellectual content, C.S.G.; literature research, C.S.G.; data acquisition, C.S.G., D.M.K.; data analysis, C.S.G.; manuscript preparation, C.S.G.; manuscript editing and review, all authors
|
References |
|---|
|
TOPAbstractIntroductionMATERIALS AND METHODSRESULTSDISCUSSIONReferences |
|---|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| RADIOLOGY | RADIOGRAPHICS | RSNA JOURNALS ONLINE |
