Results of conservative surgery and radiation for mammographically detected ductal carcinoma in situ (DCIS)

Int. J. Radiation

Biol.

Phys., Vol. 38, No. 5. pp. 949-957, 1997 Copyright 0 1997 Elsevier Science Inc. Printed in the USA. All rights reserved 0360-3016197 $17.00 + .OO

PI1 SO360-3016( 97)00153-3

ELSEVIER

l

Oncology

Clinical

Investigation

RESULTS OF CONSERVATIVE MAMMOGRAPHICALLY DETECTED B. FOWBLE, M.D.,

* A. L. HANLON, MS.,

E. R. SIGURDSON, M.D., Department

of *Radiation

Oncology,

SURGERY AND RADIATION FOR DUCTAL CARCINOMA IN SITU (DCIS) * D. A. FEIN, M.D.,

t A. PATCHEFSKY, M.D.*

AND

* J. P. HOFFMAN, M.D.,’ H. KESSLER, M.D.”

‘Department of Surgery, *Department of Pathology, and “Department Fox Chase Cancer Center, Philadelphia, PA

of Radiology,

Purpose: The role of conservative surgery and radiation for mammographlcally detected ductal carcinoma in situ (DCIS)i controversial. In particular, there is little data for outcome with radiation in a group of patients comparable to those treated with local excision and surveillance (mammographic calcitlcations &&5cm, negative resection margins, negative postbiopsy mammogram). This study reports outcome of conservative surgery and radiation for mammographically detected DCIS wlth an emphasis on results in patients considered candidates for excision alone. Methods and Materials: From 1983 to 1992, 110 women with mammographically detected DCIS (77% calcifications 2 mass) and no prior history of breast cancer underwent needle localization and biopsy with (55% ) or without a reexcision and radiation. Final margins of resection were negative in 62 % , positive 7 %, close 11% , and unknown 20%. The median patient age was 56 years. The most common histologic subtype was comedo (54% ), followed by cribriform (22% ). The median pathologic tumor size was 8 mm (range 2 mm to 5 cm). Forty-seven percent of patients with calcifications only had a negative postbiopsy mammogram prior to radiation. Radiation consisted of treatment to the entire breast (median 50.00 Gy ) and a boost to the primary site (97% ) for a median total dose of 60.40 Gy. Results: With a median follow-up of 5.3 years, three patients developed a recurrence in the treated breast. The median interval to recurrence was 8.8 years and all were invasive cancers. Two (67%) occurred outside the initial quadrant. Tbe 5- and lo-year actuarial rates of recurrence were 1 and 15%. Cause-specific survival was 100% at 5 and 10 years. Contralateral breast cancer developedin two patients. There were too few failures for statistical significance to be achieved with any of the following factors: patient age, family history, race, mammographic llndlngs, locution primary, pathologic size, histologic subtype, reexcision, or lhtal margin status. However, young age, positive or close marghts, and the presence of a mass without calcilications had a trend for an increased risk of recurrence. There were no recurrences in the subset of 16 patients who would be candidates for surveillance by Lagios’ criteria. Conclusion: For selected patients, conservative surgery and radiation for mammographically detected DCIS results in a low risk of recurrence in the treated breast and 100% 5- and lo-year cause-specific survival. Improved mammographic and pathologic evaluation results in better patient selection and reduces the risk of the subsequent appearance of DCIS in the biopsy site. The identification of risk factors for an ipsilateral invasive breast recurrence is evolving. 0 1997 Elsevier Science Inc. Ductal carcinoma

in situ, Breast conservative

therapy,

Mammographically

INTRODUCTION

detected DCIS.

Ductal carcinoma in situ (DCIS ) represent 10% of all breast cancers and 20-30% of all cancers detected solely by mammography (2). Since 1983, a significant increase in the incidence of DCIS has been noted and attributed to the increased use of screening mammography (2). Despite two recent studies (2, 30) demonstrating that mastectomy remains the most common surgical procedure for DCIS, interest in breast conservation therapy with or without radiation has increased and has followed the success of the conservative approach for early-stage invasive cancer. The initial reports of conservative surgery and radiation for DCIS reflected our then limited understanding of

the natural history and heterogeneity of the disease.Treatment policies for DCIS often paralleled those for invasive cancers, including the use of axillary node dissection, regional node irradiation, and adjuvant systemic therapy ( 25 ) . Limitations of these early studiesincluded the small number of patients treated by individual institutions and relatively short follow-up. In an effort to address these issues,10 institutions in the United States and Europe collaborated to establish a databasethat has provided the basis for a series of publications (24-27). However, the results reported often combined patients with clinically apparent DCIS (presentation asa palpable massor bloody nipple discharge) and those with clinically occult DCIS (i.e., mammographic presentation only) and outcome spe-

Reprint requests to: Barbara Fowble, M.D., Department of Radiation Oncology, Fox Chase Cancer Center, 7701 Burholme Av-

enue, Philadelphia, PA 19 111. Accepted for publication 25 March 949

1997.

9.50

I. J. Radiation

Oncology

l

Biology

0 Physics

Table 1. Patient and tumor-related characteristics marnmographically detected DCIS No. of patients (%) Total Age (yea=)

110

540 41-50 >50

8 (7) 33 (30) 69 (63)

Menopausalstatus Pre Peri Post Race White African American Other Family history Negative Positive 1 relative 2-3 relatives Mammographicfindings Calcificationsonly Mass Massand calcifications Locationprimary Outer Inner Central Subareolar

Path sizeprimary 55 mm 6-10 mm 1.1-2 cm >2 cm Unknown Predominant Comedo

histologic

Cribriform Papillary Micropapillary Solid Mixed Unknown

34 (31) 4 (4) 72 (65) 99 (90) 9 (8)

Volume

38, Number

5, 1997

tected. However, no pathologic correlation was provided other than resection margin status. Nevertheless, the results of these seriesare not infrequently compared to those of conservative surgery alone and claimed to be equal (10-12). There is a paucity of data for outcome with radiation in a group of patients similar to those treated with local excision and surveillance (mammographically detected DCIS presenting as calcifications only, ~2.5 cm, negative resection margins, and negative postbiopsy mammogram for residual calcifications). The purpose of this study is to report the outcome of conservative surgery and radiation for mammographically detected DCIS with emphasis on the results in patients who may be considered candidates for excision alone.

2 (2) 72 34 25 9

(65) (31) (23) (8)

79 (72) 16 (15) 14 (13) 58 17 32 3

(53) (15) (29) (3)

13 (12) 30 (27) 10 (9)

2 (2)

55 (50) subtype 37 15 3 1

(34) (14) (3) (1)

2 (2) 11 (10) 41 (37)

cifically related to mammographically detected DCIS was often not presented. More recent efforts ( 11, 18,21) have emphasized the importance of mammographic and pathologic evaluation in reducing ipsilateral breast recurrence after conservation therapy for DCIS. Unfortunately, few studies (9, 20, 22, 29) have reported the outcome of conservative surgery and radiation for mammographically detected DCIS in which detailed mammographic and pathologic correlation has been performed. The collaborative group recently reported the largest series ( 110 patients) with the longest median follow-up (9.3 years) (26). However, 46% of these patients had unknown resection margins, and the pathologic size or extent of the DCIS was not available. Hiramatsu et al. (7) reported outcome on 76 patients with DCIS treated with conservative surgery and radiation of whom 80% were mammographically de-

METHODS

AND MATERIALS

Between 1983 and 1992, 110 women with mammographically detected DCIS and no prior history of breast cancer underwent needle localization biopsy with (61 patients) or without reexcision (49 patients) followed by radiation at Fox Chase Cancer Center (61 patients) and the University of Pennsylvania (49 patients). Patients from the University of Pennsylvania were treated and/or followed by one of the authors (B.F.) . None of the patients had Paget’s diseaseof the nipple or physical findings suggestive of DCIS (i.e., palpable massor bloody nipple discharge). Patient and tumor-related characteristics are presented in Table 1. The median age of the patient population was 56 years (range 37-81 years). Sixty-five percent were postmenopausal,3 1% were premenopausal, and 4% were perimenopausal.Eight percent of the patients were African-American, 2% were Hispanic or Asian, and 90% were white. Thirty-four patients had a positive family history of breast cancer ( 16 with a history in a first-degree relative). Twenty-five patients had a single affected relative, 7 had two and 2 had three. Family history was unknown in four patients. The most common mammographic finding was the presence of microcalcifications without an associated mass (72%). Fourteen percent of the patients had a masswithout calcifications. Fifty-three percent were located in the outer quadrant, 15% were inner quadrant, 29% were central (6 oclock or 12 oclock) and 3% were subareolar. The median follow-up was 5.3 years (range 0.3- 14.4 years). A needle localization biopsy was performed in all patients and 55% had a reexcision (Table 2). At the time of reexcision, 29 (48%) of the 61 patients had no residual tumor and 32 had residual DCIS. The final margin of resection was negative ( >2 mm) for 68 patients (62%), positive for 8 patients (7%), close (52 mm) for 12 patients ( 11% ) , and unknown for 22 patients (20% ) . Thirtyone patients had an axillary dissection, all of which were negative. All outside pathology slides were reviewed at the University of Pennsylvania or Fox ChaseCancer Center at the time of the initial radiotherapy consultation. The predom-

Radiation for DCIS 0 B. FOWBLE Table 2. Treatment-related characteristics mammographically detected DCIS

performed significant

using the log rank statistic (13). Statistically differences were conferred by p values of

50.05.

No. of patients (%) Reexcision No Yes Positive Negative Final margin status Negative Positive Close Unknown Radiation median dose (Gy) Whole breast 50.00 10.80 Boost Total dose 60.40 primary

951

et al.

RESULTS 49

(44) 61(56) 32(29) 29 (27) 68 (62)

8 (7) 12 (11) 22 (20) range WY) 46.00-50.40 10.00-18.00 50.00-66.60

inant histologic subtype based on architectural pattern was comedo in 34%, cribriform 14%, papillary 3%, micropapillary l%, solid 2%, mixed lo%, and not reported in 37%. Nuclear grade was assessed in only 13% (5.5% Grade 3, 5.5% Grade 2, 2% Grade 1). Necrosis was recorded as present in 21 patients, absent in 1, and not reported in 88. Five patients had DCIS and lobular carcinoma in situ (LCIS). The pathologic tumor size was assessed in 55 patients. The median tumor size was 8 mm (range 2 mm to 5 cm). A postbiopsy mammogram was obtained prior to the initiation of radiation in 37 patients and was negative in 47% of the patients presenting with microcalcifications only. Sixty-four patients did not have a postbiopsy mammogram and in 9 patients it was uncertain as to whether one was performed. Radiotherapy consisted of treatment to the entire breast with tangential fields to a total dose of 46.00-50.00 Gy delivered in 1.80-2.00 Gy fractions over a period of 4.55 weeks (Table 2). The median whole breast dose was 50.00 Gy (range 46.00-50.40 Gy). A 6 MV linear accelerator was used for the majority of patients. Women with large breasts or a separation greater than 21-22 cm were treated with higher energy photons (10 or 15 MV) and a beam spoiler. Ninety-seven percent had a boost to the primary site (electrons, 1921r,external beam) for an additional lO.OO- 18.00 Gy. Results are presented in terms of ipsilateral breast recurrence, cause-specific, and overall survival. Breast recurrence was defined as first site of failure in the treated breast. A regional node failure included first site of failure in the ipsilateral axillary, internal mammary, or supraclavicular region without a simultaneous breast or distant failure. Events for cause-specific survival included only deaths from breast cancer. Survival curves and actuarial rates of breast recurrence were calculated using the KapIan-Meier (8) method with time beginning at the initiation of radiation. Statistical comparisons between curves were

With a median follow-up of 5.3 years, three patients (3%) developed a recurrence in the treated breast. All of the recurrences were invasive cancers. The interval to recurrence was 4.3 years, 8.8 years, and 8.9 years. All three recurrences were detected mammographically and two of the three patients also had a palpable mass. Two of the recurrences were in a separate quadrant of the breast from the original primary and one was in the same quadrant. The histologic subtype of the primary tumor was comedo in one patient and not recorded in two patients. None of the patients developed a regional node recurrence. One patient developed distant metastases 3 years after salvage mastectomy for an invasive recurrence. Actuarial outcome for breast recurrence is presented in Table 3. The 5- and lo-year actuarial rates of breast recurrence for all patients is 1 and 15%, respectively (Fig. 1). With only three failures, none of the factors assessed for their impact on breast recurrence achieved statistical significance. However, three factors had a trend for an increased risk of breast recurrence: young age (540 years), positive or close final resection margin, and mammographic presentation of a mass only. One of the eight women 540 years of age developed a recurrence compared to none of the 33 women age 41-50 years and 2 of the 69 patients >50 years. The single young woman also had a final positive margin. One of 16 women with a mass only on mammogram developed a recurrence compared to 2 of the 79 women with calcifications only and none of the 14 women with calcifications and a mass. There was one recurrence in the 68 patients with negative margins, one in 20 patients with a close or positive margin, and one in 22 patients with unknown margins. The interval to recurrence was 4.3 years in the patient with the positive margin compared to 8.8 and 8.9 years for others. There were no breast recurrences in the nonwhite population. Two breast recurrences were observed after 5 years in the 72 patients with a negative family history, and there was one breast recurrence at 4.3 years in the 25 patients with a single affected relative. There were no recurrences in the nine patients with two or more affected relatives. One of the 16 patients with a positive family history in a firstdegree relative recurred. The median follow-up was 5.3 years for the family history positive patients and 5.1 years for the family history negative patients. As previously noted, all three recurrences were invasive cancers and, therefore, family history did not impact on the type of recurrence or the location of the recurrence. The patient with the positive family history developed a recurrence in the same quadrant as the primary compared with one recurrence in the same quadrant and one in a separate quadrant for the family history negative patients. Two of the

952

I. J. Radiation Oncology l Biology 0 Physics Table 3. Actuarial rates of breast recurrence conservative surgery and radiation DCIS No. of patients

All patients Age (yeas) 540 41-50 >50 Race White African-American Other Family history Negative Positive 1 relative 2-3 relatives Mammographic findings Calcifications only Mass and calcifications Mass Location Outer Inner Central Subareolar Pathologic size 5.5 mm 6-10 mm 1.1-2 cm >2 cm Unknown Re-excision No Yes Negative Positive Final margin status Negative Positive/close Unknown Histology Comedo Noncomedo Unknown

5-Year actuarial rate breast recurrence %

110

1

8 33 69

25 0 0

99 9 2

2 0 0

72 34 25 9

0 5 8 0

79 14 16

0 0 14

58 17 32 3

3 0 0 0

13 30 10 2 55

0 0 0 0 3

49 61 29 32

0 3 0 5

68 20 22

0 8 0

37 32 41

0 0 3

Volume 38, Number 5, 1997

tients with nuclear grade or necrosis recorded for the analysis. One of the five patients with DCIS and LCIS recurred. All three breast recurrences occurred in patients who had a reexcision. One breast recurrence occurred in a patient whose reexcision demonstrated no residual tumor, and there were two breast recurrences in the 32 patients with a positive reexcision. One of these patients had a focally positive reexcision margin, and the other had a negative margin. There were no recurrences in the 37 patients with a negative postbiopsy mammogram prior to radiation. All three recurrences occurred in patients in whom a postbiopsy mammogram was not performed prior to radiation. There were only 16 patients who would have met the current criteria for surveillance (mammographic calcifications ~2.5 cm, negative margins, and negative postbiopsy mammogram for residual calcifications). None of these patients has recurred with a median follow-up of 4.9 years. The histology of these 16 patients was comedo (7 patients), cribriform (3 patients), papillary ( 1 patient), and unknown (5 patients). There were no recurrences in the 29 patients with a negative postbiopsy mammogram and negative resection margins. The median follow-up was 5 years. The histology of these 29 patients was comedo (16 patients), cribriform (3 patients), papillary (2 patients), and unknown (8 patients). All three patients who developed an invasive recurrence underwent salvage mastectomy. The histology of the recurrence was colloid carcinoma (one patient), and invasive ductal and ductal carcinoma in situ (two patients). The size of the recurrence was 1.2 cm, 1.5 cm, and unknown in one patient. All of the recurrences were detected mammographically, and two also had a palpable mass. The median interval to recurrence was 8.8 years (range 4.3-8.9 years). Two of the three recurrences were in a separate quadrant from the primary. At the time of mas-

ACTUARIAL BREAST RECURRENCE DCIS PATIENTS WITH MAMYO DETECTION

ONLY

58 patients with outer quadrant tumors had a recurrence, and 1 of the 32 with central tumors recurred. None of the three patients with subareolar tumors or the 17 patients with

inner quadrant

tumors

recurred.

Pathologic size did not appear to correlate with the risk of recurrence in these patients. There was one recurrence in a patient whose primary size was 2 mm, and two recurrences in those whose pathologic size was unknown. There was also no correlation with histology by architectural pattern and the risk of a breast recurrence. One of the 37 patients with predominantly comedo histology recurred compared to none of 32 patients with noncomedo histology. Two recurrences were in patients whose histologic subtype was not recorded. There were too few pa-

Fig. 1. Actuarial curve for ipsilateral breast recurrence following conservative surgery and radiation mammographically DCIS.

Radiation

for DCIS

tectomy, two patients had negative axillary nodes and one patient had no axillary dissection. None of the patients received adjuvant systemic therapy. With a median follow-up of 1 year (range O-2.2 years) after salvage mastectomy, two of the three patients remain alive without evidence of disease. The 40-year-old woman with a positive family history and focally positive reexcision margin whose recurrence at 4.3 years was a colloid carcinoma with negative axillary nodes developed bone metastases 3 years after her salvage mastectomy. She is alive with disease 8.5 years after her initial diagnosis. The 5- and lo-year actuarial overall survival rates are 96 and 94%, respectively. The 5- and IO-year actuarial cause-specific survival are 100 and 100%. Five patients died of other causes at a median of 2.6 years after their treatment (range 2.1-6 years). One of these patients died from pancreatic cancer at 2.1 years. Two patients developed contralateral breast cancer at 4 and 5.8 years. The histology of the contralateral primary was invasive ductal cancer and DCIS. Both of these patients are alive without disease. There were no contralatera1 breast cancers in the five patients with DCIS and LCIS. One contralateral breast cancer occurred in a patient with a positive family history and one in a patient with a negative family history. DISCUSSION This study presents the results of conservative surgery and radiation for mammographically detected DCIS. Unlike other retrospective series (7, 9, 22, 26, 29 ) , the degree of mammographic and pathologic correlation has been relatively complete, with 80% of the patients having microscopic margin assessment, 63% having classification of the DCIS by architectural pattern, 50% having pathologic determination of the tumor size, and 47% of those presenting with mammographic calcifications only having a postbiopsy mammogram prior to the initiation of radiation. Sneige et al. (22) have reported the only other retrospective series in which a postbiopsy mammogram was stated to have been routinely performed, although it is uncertain as to whether it preceded radiation in all cases. In their series, pathologic assessment of histologic subtype, nuclear grade, and necrosis was available for all patients, but margin assessment was determined microscopically in only 55%, with the remaining determined by gross inspection of the specimen or from the specimen x-ray. The extent of the tumor-free margin was not assessed. In the collaborative group study of mammographically detected DCIS (26)) pathologic review was performed for 73% of the cases and 54% had microscopic margin assessment. Silverstein er al. (20) reported the results of conservative surgery and radiation for 133 patients with mammographic and pathologic correlation in virtually all. Pathologic size was determined by serially sectioning the entire specimen, and margin assessment

0 B. FOWBLE

et al.

953

included a measurement of the closest distance of the tumor from the edge of the specimen. Unfortunately, this detailed evaluation did not always translate into better patient selection for the conservative approach. In a recent analysis by the authors (21)) patients whose tumor size was 24.1 cm with margins < 1 mm and with high nuclear-grade or low-grade tumors with necrosis had only a 5-year disease-free survival of 36% when treated with excision and radiation. Hiramatsu et al. (7 ) reported a decreased 6.5-year actuarial recurrence rate in patients treated between the years 1986-1990 when compared to those treated during the years 1976- 1985 (2 vs. 12%, p = 0.3). The more recent cohort was characterized by larger surgical resections, an increased use of reexcision, fewer unknown margins, and a greater reliance on specimen radiography and a postbiopsy mammogram to assess the completeness of the excision ( 18). The authors suggest that improved mammographic and pathologic correlation will result in better patient selection and improved results in terms of breast recurrence. The present series provides further evidence that detailed mammographic and pathologic evaluation can diminish breast recurrence rates following conservative surgery and radiation for DCIS. The results of conservative surgery and radiation for mammographically detected DCIS are presented in Tables 4, 5, and 6. Breast recurrence rates range from l-10% at 5 years and 8-23% at 10 years. In the present series, there were only three recurrences in the 110 patients treated resulting in a 5-year actuarial rate of breast recurrence of 1% and a IO-year rate of 15%. As previously noted, these variations most likely reflect differences in patient selection, the degree of the mammographic and pathologic evaluation, the extent of the surgical resection, and margin assessment. The patterns of breast recurrence following conservative surgery and radiation for mammographically detected DCIS are presented in Table 5. This series differs from others (3,4,7,20,22,26,28) in that all of the recurrences were invasive cancers, and two-thirds occurred in a quadrant separate from the primary. The median interval to recurrence was 8.8 years (range 4.3-8.9 years). It appears from our data that increased attention to efforts that assure a more complete excision of DCIS prior to radiation (reexcision, negative margins r2 mm, negative postbiopsy mammogram) diminishes the risk of a noninvasive recurrence in the original quadrant. As this pattern of recurrence is eliminated, late invasive recurrences may predominate. In the NSABP B17 randomized trial, 72% of the recurrences after excision and radiation were DCIS and 28% were invasive cancers (3,4). Fifty-eight percent of all recurrences (including patients who did not receive radiation) occurred within 2 years, and the majority were in the vicinity of the primary tumor (4). Postbiopsy mammograms were not required and negative margins of resection were defined as no DCIS at the inked margin. This early pattern of failure with predominantly noninvasive

954

I. J. Radiation Oncology l Biology 0 Physics

Volume 38, Number 5, 1997

Table4. Resultsof conservative surgeryandradiationfor mammographicallydetectedDCIS Actuarial breast recurrence%

NSABP B17 (3,4) Kuskeet al. (9) White et al. (29) Hiramatsuet al. (7) Sneigeet al. (22) Silversteinet al. (20) CollaborativeGroup (26) Presentseries

Total no. pts.

5 year

399* 44 46 54 31 133+ 110 110

10 7 5 2 0 7 7 1

10 year

Cause-specific survival % 5 year

10 year

100 23 8 19 14 15

96 100 100

97 96 100

Median followup years 3.6 mean 4 5.7 6.2 7.2 7.8 9.3 5.3

* 81% mammodetected. ’ 89% mammodetected.

cancer in the biopsy site suggeststhe presence of residual DCIS that was not controlled with moderate dosesof radiation. The longer interval to an invasive recurrence following conservative surgery and radiation has been reported by two other series (7,26). In the collaborative group study (26)) the median interval to an invasive recurrence was 5 years (range 4.3- 15.2 years) and 4 years (range 2.1-8.7 years) for a noninvasive recurrence. Hiramatsu et al. (7 ) reported a single invasive recurrence in 54 patients with mammographically detected DCIS treated with excision and radiation. This recurrence was observed at 6.9 years and occurred in a separatequadrant. The median interval for recurrence for the three noninvasive cancers was 5.2 years (range 3.4-8.7 years), and all three of these were in the vicinity of the primary tumor. White et al. (29) and Kuske et al. (9), however, have reported a shorter interval for invasive recurrences. The median interval to the three invasive recurrences reported by Kuske et al. was 2.6 years. The location was not stated. The single invasive recurrence reported by White et al. (29) occurred at 2.1 years and was in the vicinity of the primary. It is possible that these recurrences represent the appearance of an occult invasive cancer that was undetected at the time of initial diagnosis. Salvage of an ipsilateral breast recurrence following conservative surgery and radiation for mammographitally detected DCIS is presented in Table 6. Virtually all noninvasive recurrences have been salvaged primarily with mastectomy. Five of the 30 patients with invasive recurrences have died of disease and 2 are alive with distant metastases, including one patient from the present series. Cause-specific survival following conservative surgery and radiation for mammographically detected DCIS ranges from 96- 100% at 10 years (Table 4). Death due to nonbreast cancer-related causesmay be more common. In the present series, five patients died of other causesat a median of 2.6 years after treatment. One patient died from pancreatic cancer and the remaining four died of noncancer-related causes. The overall survival was 96%

at 5 years and 94% at 10 years. In the collaborative group study (26), the lo-year overall survival was 93% and cause-specific survival was 96%, suggestingthat the likelihood or death from breast cancer was only slightly greater than that from other causes. In the NSABP B17 trial (3), 8 of the 399 (0.6%) patients treated with conservative surgery and radiation developed a second nonbreast-related cancer. Contralateral breast cancer occurred in two patients in the present series. Both were invasive ductal cancers that were diagnosed at 4 and 5.8 years after the initial treatment for DCIS. Neither of these occurred in the five patients whose original DCIS was associated with LCIS. Both patients are alive without disease. The collaborative group reported a 6% lo-year actuarial rate of contralateral breast cancer (26). Six of the seven cancers were invasive and the other was unknown. The median interval to the contralateral cancer was 4.9 years (range 1.3- 13.4 years). Hiramatsu et al. (7) reported two contralateral breast cancers in their entire series of 76 patients, of which 54 were mammographically detected. Ten of the 399 (0.7%) patients treated with radiation in the NSABP B17 trial developed a contralatera1breast cancer (3 ) . A number of series have evaluated various clinical, pathologic, and treatment-related factors for their ability to predict for a breast recurrence in patients undergoing conservative surgery and radiation for DCIS (4, 5, 7, 9, 14, 17, 22, 23, 25, 26, 28, 29). However, only one series (26) has confined the analysis to patients with mamtnographically detected DCIS. In the collaborative group study (26)) there was no correlation between histologic subtype, nuclear grade, necrosis, or the combination of comedo histology and nuclear Grade 3 and the 5-year actuarial rate of breast recurrence. However, with a median follow-up of 9.3 years, the crude breast recurrence rate was 25% for women less than 50 years of age compared to 2% for those greater than 50 years. Fourteen of the fifteen recurrences reported occurred in women ~50 years of age compared to 2% for those >50 years. The median

955

Radiation for DCIS 0 B. FOWBLE et al. Table 5. Patterns of breast recurrence following

conservative

Total no. recurrences

Location recurrence % true/marginal

1 2 3 4 15 16 28 3

100 33 75 73 100 33

Sneige et al. (22) White et al. (29) Kuske et al. (9) Hiramatsu et al. (7) Collaborative Group (26) Silverstein et al. (20) NSABP B17 (3) Present series

to recurrence was 4.9 years for the younger women compared to 8.7 years for the single recurrence in the women >50 years. Van Zee et al. (28) also reported an increased 6-year actuarial breast recurrence rate in women <40 years treated with conservative surgery and radiation (40% age ~40 years, 9% age 40-69 years, and 0% age 270 years). White et al. (29) and Hiramatsu et al. (7) found no difference in the breast recurrence rates when comparing women <50 years vs. those >50 years; however, a later pattern of recurrence in older women was also noted by Hiramatsu et al. (7). Sneige et al. (22) found no correlation between age (as a continuous variable) and the risk of a breast recurrence. In the present series,young age ( <40 years) was associatedwith a trend towards an increased .5-year actuarial breast recurrence rate; however, because of the small number of patients, this was not statistically significant. A second clinical factor that has been associated with an increased risk of breast recurrence is a positive family history of breast cancer. Hiramatsu et al. (7) reported a IO-year actuarial breast recurrence rate of 37% for 17 patients with a positive family history compared to 9% for those with a negative family history (p = 0.008). The authors suggested that patients with a positive family history were at increased risk for both an

interval

Table 6. Results of salvage treatment for a breast recurrence following conservative surgery and radiation DCIS Number salvaged/total Noninvasive recurrence Sneige et al. (22) White et al. (29) Kuske et al. (9) Hiramatsu et al. (7) Collaborative Group (26) Silverstein er al. (26) NSABP B17 (3) Present series Total

surgery and radiation

l/l l/l 313 919 8/8 20120 42142100%)

number

Invasive recurrence O/l 313 O/l 516 618 718 213 23130 (77%)

mammographically

% Recurrence invasive 0 50 100 25 40 50 28 100

detected DCIS Median interval to recurrence years (range) 3.2 (283-4.2) 2.6 (1.3-4.4) 6.1 (3.4-8.7) 5 (2.1-15.2) 4.9 8.8 (4.3-8.9)

ipsilateral and contralateral invasive cancer. McCormick et al. (14) reported a 44% breast recurrence rate for 9 patients with a positive family history compared to 13% for the 45 patients with a negative history. The median follow-up was 3 years. In the present series, the 5-year actuarial breast recurrence rate was 5% for the 34 patients with a positive family history and 0% for those with a negative history. Similar to our findings with invasive cancers (16), the risk of a breast recurrence following radiation for DCIS decreased as the number of affected relatives increased. We did not observe an increased risk of ipsilateral or contralateral invasive cancer in the family history positive women. The influence of pathologic factors on breast recurrence rates is controversial. As previously noted none of the pathologic factors (necrosis, nuclear grade, architectural pattern) evaluated in the collaborative study predicted for an increased 5-year actuarial breast recurrence rate (26). In an analysis that was not limited to mammographically detected DCIS, Sneige et al. (22) noted an increased loyear actuarial breast recurrence rate in patients with necrosis, nuclear Grade 3, periductal fibrosis, and periductal lymphoid infiltrate. Silverstein et at. (21) , employing the Van Nuys prognostic index, correlated the risk of a breast recurrence in patients receiving radiation with nuclear grade and/or necrosis in conjunction with margin status and tumor size. In the NSABP B17 randomized trial necrosis correlated with an increased risk of breast recurrence in all patients. However, for patients receiving radiation, the crude breast recurrence rate was 5% in the presenceor absenceof necrosis (4). In contrast, the crude breast recurrence rate for nuclear Grade 3 tumors was 7% compared to 3% for nuclear Grade 1. In the present series, there was no correlation with architectural subtype and breast recurrence rates and there were too few patients whose pathologic evaluation included necrosis or nuclear grade to perform an analysis. One limitation of the present study is the relatively short median follow-up (5.3 years) compared to the long natural history of DCIS, especially the low-grade or noncomedo lesions. Several studies suggest that the interval to recurrence may be prolonged for

956

I. J. Radiation

Oncology

l Biology

0 Physics

these tumors regardless of whether radiation is given or not (15, 25). One of five patients with DCIS and LCIS recurred in this series. Fisher ef al. (4) reported no breast recurrences in 12 patients with DCIS and LCIS treated with radiation in the NSABP B17 trial. Recent studies have indicated that microscopic margin status is an important predictor for ipsilateral breast recurrence in patients with mammographically detected DCIS. In the collaborative study with a median follow-up of 9.3 years, the crude breast recurrence rate was 29% for patients with a positive or close margin, 14% for unknown margins and 7% for negative margins (26). In the NSABP B 17 trial with a mean follow-up of 4 years, the crude breast recurrence rate was 10% for patients with uncertain/involved margins compared to 4% for negative margins. Sneige et al. (22) reported a lo-year actuarial breast recurrence rate of 17% for patients with unknown margins compared to 4% for those with negative margins. White et al. (29) also reported an increased 5-year actuarial breast recurrence rate for patients with unknown margins (33%) when compared to those with negative margins (5%). However, Hiramatsu et al. (7) found no correlation with margin status and breast recurrence rates. The loyear actuarial breast recurrence rate for all patients was 0% for positive margins, 9% for close, 20% for unknown, and 11% for negative. The authors did find a correlation with breast recurrence rates and the use of reexcision or larger surgical resections. The lo-year actuarial breast recurrence rate was 17% for patients who did not have a reexcision compared to 7% for those who did. Patients whose surgical excision exceeded 60 cm3 had a 0% IO-year actuarial breast recurrence rate compared to 25% for those with small resected volumes. White et al. (29) reported a lower 5-year actuarial breast recurrence rate in patients who had a reexcision (9 vs. 4%, p = 0.80). In the present series, the 5-year actuarial breast recurrence rate was 0% for patients with negative or unknown margins and 8% for those with positive or close margins. The 5-year actuarial breast recurrence rate was 0% for patients who did not have a reexcision, 0% for those whose reexcision revealed no residual tumor, and 5% for those whose reexcision had residual DCIS. Patients with positive margins had a shorter interval to recurrence than those with negative margins. The earliest recurrence (4.3 years) was in the patient with a focally positive margin. The collaborative study reported a median interval to recurrence of 3.6 years for positive margin patients compared to 4.3 years for negative margins and 5.2 years for unknown margins (26). The presence of residual malignant appearing calcifications that are not removed prior to radiation has been associated with a 100% risk of breast recurrence. Two series ( 14, 22) have reported a total of five patients and all have recurred. In this series, there were no recurrences in the 37 patients presenting with only microcalcifications who had a negative postbiopsy mammogram prior to radiation. However, it is also unlikely that any of the patients who

Volume

38, Number

5, 1997

did not have a postbiopsy mammogram prior to radiation had residual malignant appearing calcifications since the lyear follow-up mammogram failed to reveal suspicious calcifications in any of these patients. Lagios ( 11) has proposed the following criteria for observation of mammographically detected DCIS: presentation as calcifications only, pathologic size ~2.5 cm, negative resection margins, negative postbiopsy mammogram, and patient capable of understanding the risks of surveillance. He recently reported a 17% breast recurrence rate in 78 patients with a follow-up of 10.3 years ( 11). Other series have reported similar results ( 13-22% breast recurrence rates) employing some or all of these criteria ( 1, 6, 19, 21), with median follow-up periods ranging from 3.58.3 years. It has been suggested that the risk of a breast recurrence in these series is similar to the 14% IO-year actuarial rate reported by the collaborative group (26). However, only 21 patients in the collective data base would have met all of Lagios’ criteria and none of these have recurred. In the present series, only 16 patients would meet these criteria, and none have recurred, with a median follow-up of 4.9 years. Seven of these 16 patients had comedo DCIS, 4 noncomedo, and 5 unknown histology. There were a total of 29 patients who had negative margins and a negative postbiopsy mammogram. None of these have recurred and 16 of these patients had comedo DCIS, 6 noncomedo, and 7 unknown. In Lagios’ series, the crude breast recurrence rate was 33% for the 36 high-grade/corned0 lesions compared to the 2% for the 42 low-grade/noncomedo lesions ( 11) . While prospective randomized trials provide the best evidence to support treatment recommendations for DCIS, retrospective comparisons may provide limited data until the trials mature. It is important for these comparisons to be made among similar patients for any meaningful conclusions to be drawn. In summary, the present series provides further evidence that detailed mammographic and pathologic evaluation will result in diminished breast recurrence rates for mammographically detected DCIS treated with conservative surgery and radiation. This approach minimizes the subsequent appearance of recurrent DCIS in the biopsy site. Its ultimate effect on late invasive recurrences or new cancers remains to be determined. Salvage of a breast recurrence to date has been excellent with IO-year cause-specific survivals ranging from 96-100%. The determination of factors that predict for an increased risk of breast recurrence in mammographically detected DCIS treated with radiation is evolving. These factors may change with improved mammographic and pathologic correlation resulting in better patient selection. At the present time, resection margin status and the presence of residual malignant appearing calcifications prior to radiation appear to be important factors. The role of patient age and family history of breast cancer requires further evaluation. Patients with adequately evaluated and treated DCIS may have a greater likelihood of dying of nonbreast cancer-related causes than their breast cancer.

Radiation for DCIS

0

B.

FOWBLE

et nl.

957

REFERENCES 1. Arnesson, L. G.; Smeds, S.; Fagerberg, G.; Grontoft, 0. Follow-up of two treatment modalities for ductal carcinoma in situ of the breast. Br. J. Surg. 76:672; 1989. 2. Emster, V. L.; Barclay, J.; Kerlikowske, K.; Grady, D.; Henderson, C. Incidence and treatment for ductal carcinoma in situ of the breast. J. Am. Med. Assoc. 275:913-918: 1996. 3. Fisher, B.; Constantino, J.; Redmond, C.; Fisher, E.; Margolese, R.; DiMitrov. N.; Wolmark, N.; Wickerham, D. L.; Deutsch, M.; Ore, L.; Mamounas, E.; Poller, W.; Kavanah, M. Lumpectomy compared with lumpectomy and radiation therapy for the treatment of intraductal breast cancer. N. Engl. J. Med. 328:1581-1586; 1993. 4. Fisher, E. R.; Constantino, J.; Fisher, B.; Palekar, A. S.; Redmond, C.; Mamounas, E. Pathologic findings from the National Surgical Adjuvant Breast Project (NSABP) protocol B-17. Intraductal carcinoma (ductal carcinoma in situ). Cancer 75:1310-1319; 1995. 5. Fourquet, A.; Zafrani, B.; Campana, F. Breast conserving treatment of ductal carcinoma in situ. Semin. Radiat. Onto. 2:116-124; 1992. 6. Gallagher, W. J.; Koemer, F. C.; Wood, W. C. Treatment of intraductal carcinoma with limited surgery: Long term follow-up. J. Clin. 0~01. 7:376; 1989. 7. Hiramatsu, H.; Bomstein, B. A.; Recht, A.; Schnitt, S. J.; Baum, J. K.; Connolly, J. L.; Duda, R. B.; Guidi, A. J.; Kaelin, C. M.; Silver, B.; Harris, J. R. Local recurrence after conservative surgery and radiation therapy for ductal carcinoma in situ. Possible importance of family history. Cancer J. Sci. Am. 1:55-61; 1995. 8. Kaplan, E. L.; Meier, P. Non-parametric estimation from incomplete observation. J. Am. Stat. Assoc. 53:457-481; 1958. 9. Kuske, R. R.; Bean, J. M.; Garcia, D. M.; Perez, C. A.; Andriole, D.; Philpott, G.; Fineberg, B. Breast conservation therapy for intraductal carcinoma of the breast. Int. J. Rad. One. Biol. Phys. 26:391-396; 1993. 10. Lagios, M. D. Duct carcinoma in situ: Biological implications for clinical practice. Sem. Oncol. 23:6- 11; 1996. 11. Lagios, M. D. Ductal carcinoma in situ: Controversies in diagnosis, biology, and treatment. Breast J. 1:68-78; 1995. 12. Lagios, M. D.; Page, D. L. Radiation therapy for in situ or localized breast cancer (letter). New Engl. J. Med. 21: 15771578; 1993. 13. Mantel, M. Evaluation of survival data and two new rank order statistics arising in its considerations. Cancer Chemother. Rep. 50:163-170; 1966. 14. McCormick, B.; Rosen, P. P.; Kinne, D.; Cox, L.; Yahalom, J. Duct carcinoma in situ of the breast: An analysis of local control after conservative surgery and radiotherapy. Int. J. Radiat. Oncol. Biol. Phys. 21:289-292; 1991. 15. Page, D. L.; Dupont, W. D.; Rogers, L. W.; Jensen, R. A.; Schuyler, P. A. Continued local recurrence of carcinoma 15-25 years after a diagnosis of low grade ductal carcinoma in situ of the breast treated only by biopsy. Cancer 76: 1197- 1200; 1995. 16. Peterson, M.; Fowble, B.; Solin, L. J.; Schultz, D. J. Family history status as a prognostic factor for women with early stage breast cancer treated with conservative surgery and radiation. Breast J. 1:202-209; 1995. 17. Ray, G. R.; Adelson, J.; Hayhurst, E.; Marzoni, A.; Gregg, D.; Bronk, M.; McClenathan, J.; Bitar, N.; Macio, I. Ductal carcinoma in situ of the breast: Results of treatment by conservative surgery and definitive irradiation. Int. J. Radiat. Oncol. Biol. Phys. 28:105-111; 1993. 18. Schnitt, S. J.; Harris, J. R.; Smith, B. L. Developing a prognostic index for ductal carcinoma in situ of the breast. Cancer 77:2189-2192: 1996. 19. Schwartz, G. F.; Schwarting, R.; Cornfield, D. B.; Shen, R.; McDade, T. Sub-clinical duct carcinoma in situ of the breast

20.

21.

22.

23.

24.

25.

26.

27.

28.

29.

30.

(DCIS) : Treatment by local excision and surveillance alone. Proc. Am. Sot. Clin. One. 15:lOl; 1996. Silverstein, M. J.; Barth, A.; Poller, D. N.; et al. Ten year results comparing mastectomy to excision and radiation therapy for ductal carcinoma in situ of the breast. Eur. J. Cancer 31:1425-1427; 1995. Silverstein, M. J.; Lagios, M. D.; Craig, P. H.; Waisman, J. R.; Lewinsky, B. S.; Colbum, W. J.; Poller, D. N. A prognostic index for ductal carcinoma in situ of the breast. Cancer 7712267-2274; 1996. Sneige, N.; McNeese, M. D.; Atkinson, E. N.; Ames, F. C.; Kemp, B.; Sahin, A.; Ayala, A. G. Ductal sarcinoma in situ treated with lumpectomy and irradiation: Histopathologic analysis of 49 specimens with emphasis on risk factors and long term results. Hum. Pathol. 26:642-649; 1995. Solin, L. F.; Yeh, I.-T.; Kurtz, J.; Fourquet, A.; Recht, A.; Kuske, R.; McCormick, B.; Cross, M. A.; Schultz, D. J.; Amalric, R.; Livolsi, V. A.; Kowalyshyn, M. J.; Torhorst, J.; Jacquemeier, J.; Westermann, C. D.; Mazeujian, G.; Zafrani, B.; Rosen, P. P.; Goodman, R. L.; Fowble, B. L. Ductal carcinoma in situ (intraductal carcinoma) of the breast treated with breast-conserving surgery and definitive irradiation: Correlation of the pathologic parameters with outcome of treatment. Cancer 71:2532-2542; 1993. Solin, L. J.; Fourquet, A.; McCormick, B.; Haffty, B.; Recht, A.; Schultz, D. J.; Barrett, W.; Fowble, B. L.; Kuske, R.; Taylor, M.; McNeese, M.; Amalric, R.; Kurtz, J. Salvage treatment for local recurrence following breast-conserving surgery and definitive irradiation for ductal carcinoma in situ (intraductal carcinoma) of the breast. Int. J. Radiat. Oncol. Biol. Phys. 30:3-9; 1994. Solin, L. J.; Kurtz, J.; Fourquet, A.; Amalric. R.; Recht, A.; Bomstein, B. A.; Kuske, R.; Taylor, M.; Barrett. W.; Fowble, B. L.; Haffty, B.; Schultz, D. J.; Yeh, I.-T.; McCormick, B.; McNeese, M. Fifteen year results of breast conserving surgery and definitive breast irradiation for the treatment of ductal carcinoma in situ (intraductal carcinoma of the breast). J. Clin. Oncol. 14:754-763; 1996. Solin, L. J.; McCormick, B.; Recht, A.; Haffty, B.: Taylor. M.; Kuske, R.; Bomstein, B.; McNeese, M.; Schultz, D. J.; Fowble, B. L.; Barrett, W.; Yeh, I.-T.; Kurtz, J.; Amalric, R.; Fourquet, A. Mammography detected clinically occult ductal carcinoma in situ (intraductal carcinoma) treated with breast-conserving surgery and definitive breast irradiation. Cancer J. Sci. Am. 2:158-165; 1996. Solin, L. J.; Recht, A.; Fourquet, A.; Kurtz, J.; Kuske, R.; McNeese, M.; McCormick, B.; Cross, M. A.; Schultz, D. J.; Bomstein, B.; Spitalier, J.-M.; Vilcoq, J. R.; Fowble, B. L.; Harris, J. R.; Goodman, R. L. Ten-year results of breastconserving surgery and definitive irradiation for intraductal carcinoma (ductal carcinoma in situ) of the breast. Cancer 68:2337-2344; 1990. Van Zee, P. J.; Liberman, L.; McCormick, B.; Petrek, J. A.; Bergen, P. I. Long term follow-up of DCIS treated with breast conservation: Effect of age and radiation. Abstract Book Society of Surgical Oncology, March 21-24. 1996 Atlantic: 1996. White, J.; Levine, A.; Gustafson, G.; Wimbish, K.; Ingold, J.; Pettinga, J.; Matter, R.; Martinez, A.; Vicini, F. Outcone and prognostic factors for local recurrence in mammographically detected ductal carcinoma in-situ of the breast treated with conservative surgery and radiation therapy. Int. J. Radiat. Oncol. Biol. Phys. 31:791-797; 1995. Winchester, D. P.; Menck, H. R.; Osteen, R. T.; Kraybill, W. Treatment trends for ductal carcinoma in situ of the breast. Ann. Surg. Oncol. 2:207-213; 1995.