Conservative surgery and radiotherapy for early-stage breast cancer using a lung density correction: The university of michigan experience

Conservative surgery and radiotherapy for early-stage breast cancer using a lung density correction: The university of michigan experience

lnt. J. Radiation Oncology BioI. Phys., Vol. 39, No.4, pp. 921-928, 1997 Copyright © 1997 Elsevier Science Inc. Printed in the USA. All rights reserve...

760KB Sizes 0 Downloads 1 Views

lnt. J. Radiation Oncology BioI. Phys., Vol. 39, No.4, pp. 921-928, 1997 Copyright © 1997 Elsevier Science Inc. Printed in the USA. All rights reserved 0360-3016/97 $17.00 + .00

PII S0360-3016(97)00464-1

ELSEVIER

CONSERVATIVE SURGERY AND RADIOTHERAPY FOR EARLY-STAGE BREAST CANCER USING A LUNG DENSITY CORRECTION: THE UNIVERSITY OF MICHIGAN EXPERIENCE LORI

J.

PIERCE, M.D., * MYLA

H.

STRAWDERMAN, M.S.,t KATHYE

R.

DOUGLAS* AND

ALLEN S. LICHTER, M.D.*

* Department of Radiation Oncology and tUniversity of Michigan Cancer Center Biostatistics Core, Ann Arbor, MI48109 Purpose: Although an abundance of reports detail the successful use of definitive radiotherapy of the breast in the treatment in Stage I or II breast cancer, little data have been published concerning the use of lung density correction and its effect upon long-term outcome. As it has been the practice at the University of Michigan to routinely use lung density correction in the dose calculations to the breast, we retrospectively analyzed our results for local control, relapse-free, and overall survival. Methods and Materials: Clinical records were reviewed of 429 women with Stage I or II breast cancer treated with lumpectomy, axillary dissection, and breast irradiation with or without systemic chemolhormonal therapy. Tangential radiotherapy fields delivering 45 to 50 Gy were used to treat the entire breast. A boost was delivered in 95% of cases for a total tumor bed dose of 60 to 66 Gy. All treatment plans were calculated using a lung density correction. Results: With a median follow up of 4.4 years, the 5-year actuarial rate of local control with local failure as the only site of first failure was 96% (95% CI 94-98%). Univariate analysis for local failure as only first failure found the following factors to statistically predict for increased risk of breast recurrence: young age (:535 years old), premenopausal status, tumor size >2 em, positive family history, and positive microscopic margins. Multivariate analysis revealed young age and margin status to be the only factors remaining significant for local failure. The 5-year actuarial relapse-free survival was 85% (95% CI 81-89%); overall survival at 5 years was 90% (95% CI 87-94%). Conclusions: Lung density correction results in rates of local control, disease-free, and overall survival at 5 years that compare favorably with series using noncorrected unit density calculations. While we will continue to update our results with increasing follow-up, our 5-year data indicate that the use of lung-density correction for dosimetric accuracy does not compromise local control. © 1997 Elsevier Science Inc. Breast cancer, Radiation therapy, Lung density correction, Local control.

INTRODUCTION

lung density information. However, concern arises that by changing the dosimetric assumptions that produced the excellent treatment outcomes cited above, one might adversely influence local control in the breast. Without large series of patients treated and followed after breast conservation that used 'lung density correction, one could speculate whether density corrected plans might change dose distributions in unanticipated ways, leading to a higher rate of in-breast failure. At the University of Michigan, we have been using lung density correction in all of our breast treatment plans for over a decade. In this report, we summarize the results on 429 patients treated in this fashion following local excision of their primary breast tumor.

Through the use of multiple randomized trials and large single-institution series, conservative surgery and radiotherapy has become an accepted alternative to mastectomy for women with Stage I or II breast cancer (5, 7, 12, 19, 24). Definitive irradiation to the breast has resulted in excellent rates of local control with survival equivalent to that achieved with mastectomy, while affording a woman the psychological advantage of breast preservation. However, in virtually all of these reports, the treatment planning for each case was done without taking into account the influence that low-density lung tissue had on the dose distribution. Rather, the entire irradiated volume was assumed to be unit density and the dosimetry was carried out accordingly. In the early 1980s, computed tomography (CT) scans allowed for a full description of the location and density of the lung in breast tangent fields. Several research groups showed that the homogeneity of a breast treatment plan could be improved by recognizing and accounting for this

METHODS AND MATERIALS Patient eligibility From 1984 through 1995, 396 women with either Stage I or II breast cancer were treated with breast-conserving Acknowledgments-Dr. Pierce is the recipient of the American Cancer Society Career Development Award #95-76. Accepted for publication 2 February 1997.

Reprint requests to: Lori J. Pierce, M.D., Department of Radiation Oncology, University of Michigan Medical School, URB2C490/001O, 1500 E. Medical Center Drive, Ann Arbor, MI 48109-0010. 921

922

I. 1. Radiation Oncology • Biology • Physics

therapy at the University of Michigan Hospital using excisional biopsy, axillary dissection, and definitive radiotherapy. Thirty-three additional patients with clinically negative axillae were treated with radiotherapy following excisional biopsy in the absence of an axillary dissection for a total patient cohort of 429. Patients seen in referral who received radiotherapy at an outside institution were excluded from review. Patient selection for breast-conserving therapy in this early-stage series followed commonly accepted guidelines. Briefly, women presenting with a single, nonfixed palpable and/or mammographic ally detected lesion measuring up to 5 cm in size were candidates for definitive radiotherapy, assuming that resection of the primary lesion resulted in an acceptable cosmetic outcome. Patients excluded from this series included women presenting with gross multifocal or multicentric disease, diffuse microca1cifications present on mammogram, the diagnosis of active systemic lupus, discoid lupus, or scleroderma, or patients presenting with clinical Stage III disease. Surgical therapy Surgical treatment to the primary lesion consisted of an excisional biopsy of the tumor bed. Resected specimens were routinely inked to assess margins. Sixty-five percent of the patients in this series also underwent a reexcision. Axillary surgical staging generally consisted of a level I and II dissection. A median of 16 nodes was dissected (range 1-52). Radiotherapy Radiotherapy to th~ jntact breast consisted of 45 to 50 Gy in 1.8 to 2.0 Gy fractions delivered over 4.5 to 5 weeks through tangential fields using 6 MV photons. All plans were optimized such that, in general, dose inhomogeneity did not exceed 10%. A boost dose was delivered to 95% of patients using electrons generally resulting in a total dose of 60 to 66 Gy to the tumor bed. Specifically, 24% of patients received a total dose up to 60 Gy, 37% received between 60 and 64 Gy, 28% receiv5'Q between 64 and 66 Gy, and 11 % received in excess of 66 Gy. Pathologically node negative patients did not receive regional radiotherapy. All nodepositive patients received a supraclavicular field delivering 45 to 46 Gy to a depth of 3 cm. The majority of patients with a dissected axilla were not treated with a posterior axillary . boost unless the dissection was felt to be inadequate or . macroscopic extracapsular extension was identified upon pathologic review. All patients who did not undergo a dissection received radiotherapy to the supraclavicular and axillary nodes, delivering 45 to 46 Gy to the supraclavicular fossa and the midplane of the axilla. Lung density correction was used in all patients using a density of 0.20 g/cm3 (18, 23). The dose calculation model and the treatment-planning system used were both three dimensional. The equivalent path length method of inhomogeneity correction was utilized. Contour data were used to generate a surface description that was then used for all C

Volume 39, Number 4, 1997

depth and distance measurements for the planning system. The amount of lung tissue in the tangent fields was restricted to a 3-cm maximum central lung distance measured at the central axis. All plans were normalized at the isocenter and, generally, dose was prescribed to the isocenter. To optimize the plans such that inhomogeneity at the medial edge, lateral edge, and apex of the breast did not exceed 10%, 15 or 30° wedges were commonly used at the lateral tangent. Adjuvant therapy Adjuvant chemotherapy was administered to approximately 38% of the women in this report. Approximately 25% received cyclophosphamide, methotrexate, and 5-fluorouracil (CMF) and 11 % received a doxorubicin-containing regimen. The sequencing of chemotherapy and radiotherapy was not standardized. Radiotherapy was given either preceding or concurrently with CMF, at the discretion of the treating radiation oncologist. If a doxorubicin-containing regimen was administered, chemotherapy was given initially with radiotherapy to follow. Forty percent received adjuvant hormonal therapy, which consisted of tamoxifen in 39.5% and megace in 0.5%. Follow-up evaluation Following the completion of therapy, patients were followed at 4-month intervals for the first 2 years with physical examination. A base!ine postradiation mammogram was obtained at the first 4-month visit. Chest x-rays, liver function! tests, and bilateral mammograms were performed yearly. Bone scans with plain film correlation were performed if clinically indicated. Follow-up intervals increased to 6 months after 2 years if patients were free of disease at 24 months. Annual follow-up started at year 5. Data analysis Patient records were reviewed for the following characteristics: age, race, menopausal status, family history (defined as any first or second degree relative with a history of breast cancer), tumor size, pathologic nodal status, pathologic stage, number of positive nodes, estrogen and progesterone status, tumor grade, pathologic margin status, and extensive intraductal component (EIC). A local only failure as first site of failure was defined as a breast-only failure. A local failure as a component of first failure included all breast failures whether isolated or concurrent with regional or distant failure. A regional failure consisted of a recurrence in the axillary, supraclavicular, or infraclavicular regions. A patient was considered relapse-free if continuously free of disease from the completion of radiotherapy to the last follow-up visit. Patients considered without evidence of disease at last observation included all relapse-free patients and those with local and/or regional failures successfully Salvaged by the last observation period. Local control, relapse-free, and overall survival curves were generated using the Kaplan-Meier method, with time

Lung density correction. L. J.

PIERCE

et al.

Table 1. Clinical characteristics

Table 2. Tumor characteristics

Factor

Level

n (%)

Menopausal status

:0;35 36-50 >50 Black White Other/unknown Pre Post Peri Unknown No Yes Unknown

20 (5) 135 (32) 274 (64) 30 (7) 391 (91) 8 (2) 138 (32) 237 (55) 31 (7) 23 (5) 196 (46) 93 (22) 140 (33)

Age

Race

Family history

923

Factor Primary size

Histology

Pathologic nodal status

beginning at the completion of radiotherapy; comparisons between the curves were calculated using the log rank test (4, 13). Multivariate analyses for local-regional control, relapse-free survival, and overall survival were done using a step-wise Cox regression model. These models were used to assess independent prognostic factors. Missing data were assumed to be missing at random and were excluded from all analyses. Median follow-up for surviving patients was 4.4 years (range 1.0-11.5 years). All patients had a minimum of I-year follow-up at the time of the analysis.

RESULTS

Clinical and tumor characteristics Tables 1 and 2 provide the clinical and tumor characteristics of the series, respectively. The median age at diagnosis was 55 years, ranging from 22 to 89 years. The majority of women were postmenopausal at the time of diagnosis, and 22% had a family history of breast cancer, defined as at least one first or second degree family member diagnosed with breast cancer. Three hundred and ninety-six women (92%) underwent an axillary dissection and could be pathologically staged. Of those surgically staged, 59% were diagnosed with pathologic Stage I disease and 41 % with path Stage II disease. Sixty-five percent had a reexcision at the primary site such that the final margin status was microscopically negative in 85% of the cases. The presence of an extensive intraductal component was reviewed in the pathology data base. EIC was defined using the Joint Center for Radiation Therapy definition of either intraductal carcinoma composing greater than 25 % of the tumor mass and extending into surrounding normal breast parenchyma, or predominantly intraductal carcinoma with areas of microinvasion (1). Only 52% of the specimens [222] had been pathologically reviewed for the presence of EIC; 85% (189 of 222) were EIC negative. Due to the small number of EIC positive cases, EIC could not be analyzed as a predictive factor for local recurrence.

Pathologic stage

Grade

Estrogen receptor status

Progesterone receptor status

Final microscopic margin

Level

n (%)

:0;1 cm 1.1-2.0 2.1-3.0 3.1-4.0 4.1-5.0 >5.0 Unknown Ductal Lobular Ductal and lobular Medullary Tubular Colloid Other/unknown Negative Positive 1-3 2:4 No dissection I II Unknown 1 2 3 Unknown Positive Negative Unknown Positive Negative Unknown Positive Negative Unknown

140 (33) 178 (41) 61 (14) 19 (4) 6 (1) 2 (1) 23 (5) 339 (79) 25 (6) 11 (3) 20 (5) 16 (4) 4 (1) 14 (3) 281 (66) 115 (26) 92 (21) 23 (5) 33 (8) 232 (54) 164 (38) 33 (8) 53 (12) 95 (22) 80 (19) 201 (47) 231 (54) 99 (23) 99 (23) 189 (44) 128 (30) 112 (26) 33 (8) 363 (85) 33 (8)

Local and regional control The 5-year actuarial rate o~ local control with local failure as the only site of first failure was 96%, with a 95% CI of 94-98% (The time to local failure is presented in Figure 1). Sixteen patients had a localf~lure as the only site of first failure. Eleven of the ~f1169%) were successfully salvaged by mastectomy at tJte time of their isolated breast recurrence, such that the overall 5-year actuarial rate of local control following salvage was 98.6%. One additional patient had a simultaneous local and distant failure, resulting in a 5-year actuarial local control rate with local failure as any component of first failure of 96% (95% CI 94-98%). Six patients experienced a regional only failure. Four of the six failures were in the supraclavicular fossa, one recurrence was in the axilla only, and one patient experienced an isolated infraclavicular failure. Four of these patients have expired with disease; two remain alive with disease. Univariate analyses for both local failure as only first failure and for local failure as component of first failure were performed using the following factors: age, menopausal status, family history, pathological stage, primary size, margin status, estrogen and progesterone receptor status, and number of positive axillary nodes. As shown

924

1. J. Radiation Oncology. Biology. Physics

Volume 39, Number 4, 1997

prediction of local failure after considering age and margin status at the 5% significance level.

0

«')

0 on

"! 0

Local Only First Failure Local Component of FIrst Failure

5 Yrs. 4% 4%

10 Yrs. 9% 11 %

Patterns of failure Table 4 shows the patterns of first failure by clinical and tumor characteristics. As previously discussed, young age, premenopausal status, positive family history, T2 lesions, and positive microscopic margins were correlated with increased breast failure. Distant failure as isolated first failure was associated with pathologic Stage II disease (p = 0.0009) and with positive axillary nodes (p < 0.0001) by Fisher's exact test.

~

0

i

£

~

0

r---------

;::

I

I

0 on

0

0

Relapse-free survival As demonstrated in Fig. 2, the 5-year actuarial relapsefree survival was 85% (95% CI 81-89%). Using the same clinical and tumor characteristics used in the local control univariate analysis, a univariate analysis for relapse-free survival was performed. Factors that significantly predicted for decreased relapse-free survival included young age, pathologic Stage II disease, T2 primary, and positive axillary nodes. Specifically, the 5-year actuarial rate of relapsefree survival was 72% for women :::;35 years old, compared to 82% for women 36-50, and 88% for patients over 50 years of age, p = 0.002. The 5-year relapse-free survival was 78% for women with pathologic Stage II disease vs. 90% for path Stage I, p = 0.0002. Patients wIth primary tumors greater than 2/cm experienced a 5-year relapse-free survival of 80 vs. 88% for primaries :::;2 cm, p = 0.02. Positive nodes significantly predicted for increased relapse, with patients with zero positive nodes having a 88% 5-year relapse-free survival compared to 77% for one to three positive nodes, and 76% for four or more positive nodes, p = 0.004. Multivariate analysis for relapse-free survival using the Cox model revealed the factors that independently predicted relapse were pathologic stage and age. Young age was the most significant predictor for relapse, p = 0.0005. Pathologic Stage II (vs. Stage I) also independently predicted relapse, with a hazard ratio of 2.3 (95% CI 1.3-4.0), P = 0.003. The number of positive nodes did not add to the

C! 0

0

2

4

6

8

10

12

Years

Fig. 1. Five- and lO-year actuarial rates of local failure using lung density correction.

in Table 3, the factors found to statistically predict for increased risk of breast recurrence as an isolated first failure were young age (:::;35 years old), premenopausal status, positive family history, tumor size >2 cm, and positive microscopic margins. In the univariate analysis for breast recurrence as a component of first failure, the same factors were statistically significant with the exception of family history (p = 0.08) and primary size, which was of borderline significance (p = 0.06). The best predictive models for time to local only first failure and local component of first failure included age and final margin status. After controlling for margin status, younger women were at greater risk for both local only failure and local component of first failure (p = 0.0001 and p = 0.0003, respectively). Independent of age, patients with positive margin status had an estimated fivefold (95% CI 1.3-18.6; p = 0.02) hig~5)r risk oflocal only first failure and 4.7 times (95% CI 1.3':"'17.3; p = 0.02) higher risk oflocal component of first failure. No other factor added to the

Table 3. Univariate analyses of breast failure as isolated first failure Factor Age Menopausal status Family history Primary size Margin status

Level :0;35 36-50 >50 -Pre Post Peri No Yes :o;2cm >2cm Positive Negative

5-Year rate (%)

95% CI

Log rank p-value

85 94 98 92 98 100 98 94

71-100 89-98 96-100 87-97 96-100

<0.0001

96-100 87-100 96-100 90-100 79-100 95-99

0.04

~8

95, 90 97

0.0009

0.03 0.04

Lung density correction. L. J.

PIERCE

et al.

925

Table 4. Patterns of first failure by clinical and tumor characteristics [n (%)] Characteristic Age :535 36-50 >50 Race Black White Other Menopausal status Pre Post Peri Family history No Yes Primary size :52cm >2cm Pathological stage I II

Node positivity 0 1-3 2::4 Margin status Positive Negative

Local only

Regional only

Distant only

Local and distant

4 (20) 9 (7) 3 (1)

0 6 (4) 0

1 (5) 11 (9) 21 (7)

0 0 1 «1)

1 (5) 2 (1) 0

14 (70) 107 (80) 249 (90)

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

0 6 (2)· 0

3 (10) 30 (8) 0

0 1 «1) 0

0 3 «1) 0

26 (87) 337 (86) 6 (86)

12 (9) 3 (1) 0

5 (3) 1 (1) 0

8 (6) 20 (8) 3 (10)

0 1 «1) 0

2 (1) 1 «1) 0

111 (81) 211 (89) 28 (90)

2 (1) 4 (4)

3 (1) 2 (2)

12 (7) 5 (5)

1 «1) 0

0 1 (1)

178 (91) 81 (87)

7 (2) 6(7)

4 (1) 2 (2)

23 (7) 10 (11)

1 «1) 0

2 (1) 1 (1)

281 (88) 69 (79)

6 (3) 10 (6)

3 (1) 3 (2)

9 (4) 22 (12)

1 «1) 0

1 «1) 2 (1)

212 (91) 127 (77)

14 (5) 2 (2) 0

4 (1) 1 (1) 1 (4)

10 (4) 17 (18) 4 (17)

1 «1) 0 0

1 «1) 1 (1) 1 (4)

251 (89) 71 (77) 17 (74)

3 (9) 11 (3)

0 5 (1)

2 (6) 27 (7)

0 1 «1)

0 2 «1)

28 (85) 317 (87)

predictive ability of the multivariate Cox model after considering age and pathologic stage. Overall survival The 5-year actuarial overall survival for the series was 90%, (95% CI 87-94%). There were 46 deaths in the cohort overall. The only factors that significantly predicted for death in the univariate analysis were pathologic Stage II disease compared to Stage I, and number of positive nodes. Specifically, the 5-year actuarial overall survival in patients with path Stage II disease was 87% compared to 92% for path Stage I, p = 0.03. Patients who were pathologically node negative had a 5-year overall survival of 93% compared to 85% for women with one to three positive nodes vs. 76% for four or more positive nodes, p < 0.0001. After adjusting for the number of positive axillary nodes, the strongest predictor for overall survival by univariate analysis, no other factors were independently predictive for survival using the Cox model. With respect to number of positive nodes, patients with one to three positive nodes were estimated to die at 2.4 times the rate of patients with negative nodes (i.e., hazard ratio 2.4, 95% CI 1.2-4.7), p = 0.01, and patients with four or more positive nodes had a hazard ratio of death of 5.2 relative to patients with negative nodes (95% CI 2.2-12.4), p = 0.0002. Complications The complications experienced by the patients in this series were minimal. Overall, 8% developed various de-

Regional and distant

No failure

grees of arm edema. Five percent developed arm edema with increased arm circumference measuring :52 cm greater than the contralateral arm, and 3% developed arm edema exceeding a 2 cm increased circumference. Persistent chest wall pain occured in 1.4% of patients. Complications re-

:3 . . . -..."....--00

-..,

ci

L. _ _ _ _ _ _ _ _

10

.~

ci

]

'"

..c

£

~

ci

N

85%

10 Yrs. 78% 74%

8

10

5Yrs.

ci

Overall Survival Relapse-free Survival

90%

0

ci

0

2

4

6

12

Years

Fig. 2. Overall survival and relapse-free survival at 5- and 10years.

926

1. 1. Radiation Oncology. Biology. Physics

Volume 39, Number 4, 1997

ported in less than 1% of patients included rib fracture [1], chronic cellulitis of the breast [1], clinical pneumonitis [1], decreased shoulder mobility [4], and matchline fibrosis [2]. DISCUSSION

The use of lung density correction has obvious implications in the optimization of the breast plans. As previously described by Fraass et. at. (8), correction for the decreased attenuation of the photon beam by lung tissue in the deep part of the tangents compensates for the extra thickness of the tissue in that part of the field. This offsets the need for increased wedging as used in the noncorrected unit density plans and removes the necessity of a medial wedge and often a lateral wedge. As shown in Fig. 3A, to optimize the dose distribution on a unit-density plan, 30° wedges are required for both the medial and lateral tangents. On the lung density-corrected plan shown in Fig. 3B, optimization is achieved simply with a 15° wedge in the lateral tangent beam only. These changes result in the reduced dose to the opposite breast secondary to lack of scatter from a medial wedge and a reduction in the required monitor units through the medial tangent (9). Although the clinical significance of reduced dose to the opposite breast is largely unknown as randomized trials have not demonstrated an increased rate of contralateral breast cancers in irradiated women relative to mastectomy controls, if lung density corrections are as easy to calculate, dosimetrically more accurate, and result in control rates comparable to noncorrected plans, the benefits appear to warrant routine lung density correction. Comparing local control rates between different series can be difficult because differences in selection criteria, explicit or implicit, can influence the local control rates. Nonetheless, the 4% rate of local failure at 5 years demonstrated in our series using lung density correction compares favorably to those reported using nonlung corrected plans (2, 3, 6, 10, 17, 22). Recht and colleagues from the Joint Center for Radiation Therapy reported a 5-year actuarial local recurrence rate of 9% in their series overall, with a median follow-up of 52 months (17). However, when patients who had less than/an excisional biopsy were excluded and the analysis was restricted to those having had an excisional biopsy, the 5-year rate of recurrence was 7%. Solin et at. (22) reported the 5 year results from the University of Pennsylvania, where the overall 5-year actuarial local failure rate was 6%, with the rate of local only failure as first failure 3%. Similarly, results from the Institut Gustave-Roussy demonstrated a 7% local failure rate at 5 years, and the randomized trial from the National Sllrgical Adjuvant Breast Project (B-06)-reported a 7.7% breast recurrence rate at 5 years (2, 6). While these studies varied in specific technical details such as microscopic margin status and requirement of a boost, none employed the use of lung correction. Similarly, when the 5-year actuarial survivals are compared between series, the 90% overall survival reported in our series is again comparable to the reported range of 83 to 93% in the literature. Therefore, the use of lung

o -5 R

-10 -15

-20

SCALE = 0.sf\;:.'f5........,..,_..,lrO,.....,............._,..5~~-r0<'.-.~--r5~

109.0 105.0 100.0 98.0 95.0 50.0 10.0

ISO-DOSE:

1""1,,,,1

P (a)

o -5

R

-10 -15

-20

SCALE = O. S.f'f_Ff.'9=;=;""'_,...1rO.......~_'T'5~~-rOL,....,~,...,5r ISO-DOSE:

5

107.0 105.0 100.0 95.0 50.0 10.0

1""1,,,,1

P

(b) Fig. 3. (A) Optimization of a dose distribution in a noncorrected unit-density plan. (B) Optimization of a dose distribution in a lung density-corrected plan.

correction in our series has certainly not adversely impacted local control or overall survival. Results similar to our series were also recently reported by Heimann, et at. (11) from the University of Chicago, where the 5-year rate of local failure was 3%, and the overall survival rate was 89% in patients also treated with lung density correction. Factors that were found to significantly impact upon local control in the present report were young age and margin status. Both of these factors have been 'sl;1own to affect breast recurrence rates in other series (14'""l6, 21). Many reasons have been suggested to ac, count for the increased rate of recurrence .following

Lung density correction. L. J.

breast-conserving therapy in young women. In women less than 40 years of age, Kurtz and colleagues found a greater prevalence of certain morphologic features was associated with increased breast recurrence, specifically very extensive intraductal carcinoma and a major lymphocytic stromal reaction (14). Others have attributed the increased failure rate to a higher rate of EIC positive lesions in young women and inadequate microscopic margins (1, 20, 25). Our data do not support that adequacy of margins alone could account for the increased rates of recurrence as both factors were shown to be independent predictors for failure in the Cox regression model. Recurrence rates have also been shown to be increased in young women following mastectomy, which may support a causal role of morphologic factors and increased rates of local failure (15). Eighty-five percent of the cases in our series had negative margins microscopically. Despite the low percentage of specimens with positive margins, margin status was still a significant predictor of breast recurrence,

PIERCE

et al.

927

predicting local failure at an estimated rate fivefold that of margin negative tumors. We were unable to determine whether the presence of EIC resulted in an adverse effect upon breast recurrence because so few of the specimens in our series reviewed for EIC were found to contain an extensive intraductal component. We routinely now assess all tumors excised for the presence of EIC and will analyze its effect in predicting breast recurrence in subsequent reports. In conclusion, we have demonstrated the successful use of lung-density correction in the definitive treatment of Stages I and II breast cancer. Local control was comparable to published rates using unit-density correction, with the advantage of decreased dose to the contralateral breast. While we will continue to update our patient series because in-breast recurrences increase with time, our 5-year data support the use of lung density correction for dosimetric accuracy and optimal local control.

REFERENCES 1. Boyages, J.; Recht, A; Connolly, J. L.; Schnitt, S. J.; Gelman, R; Kooy, H.; Love, S.; Osteen, R T.; Cady, B.; Silver, B.; Harris, J. R. Early breast cancer: Predictors of breast recurrence for patients treated with conservative surgery and radiation therapy. Radiother. Oncoi. 19:29-41; 1990. 2. Calle, R; Vilcoq, J. R; Zafrani, B.; Vielh, P.; Fourquet, A Local control and survival of breast cancer treated by limited surgery followed by irradiation. Int. J. Radiat. Oncoi. BioI. Phys. 12:873-878; 1986. 3. Clarke, D. H.; Le, M. G.; Sarrazin, D.; Lacombe, M.; Fontaine, F.; Travagli, J.; May-Levin, F.; Contesso, G.; Arriagada, R Analysis of local-regional relapses in patients with early breast cancers treated by excision and radiotherapy: Experience of the Institut Gustave-Roussy. Int. J. Radiat. Oncoi. BioI. Phys. 11:137-145; 1985. 4. Cox, D. R; Analysis of binary data. London: Chapman and Hall; 1980. 5. Fisher, B.; Anderson, S.; Redmond, C. K; Wolmark, N.; Wickerham, D. L.; Cronin, W. M. Reanalysis and results after 12 years of follow-up in a randomized clinical trial comparing total mastectomy with lumpectomy with or without irradiation in the treatment of breast cancer. N. Engi. J. Med. 333:14561461; 1995. 6. Fisher, B.; Bauer, M.; Margolese, R; Poisson, R; Pilch, Y.; Redmond, c.; Fisher, E.; Wolmark, N.; Deutsch, M.; Montague, E.; Saffer, E.; Wickerham, L.; Lerner, H.; Glass, A; Shibata, H.; Deckers, P.; Ketcham, A; Oishi, R.; Russell, I. Five-year results of a randomized clinical trial comparing total mastectomy and segmental mastectomy with or without radiation in the treatment of breast cancer. N. Engi. J. Med. 312:665-673; 1985. 7. Fowble, B. L.; Solin, L. J.; Schultz, D. J.; Goodman, R L. Ten-year results of conservative surgery and irradiation for stage I and II breast cancer. Int. J. Radiat. Oncoi. BioI. Phys. 21:269-277; 1991. 8. Fraass, B. A.; Lichter, A. S.; McShan, D. L.; Yanke, R R; Diaz, R F.; Yeakel, K S.; van de Geijn, J. The influence of lung density corrections on treatment planning for primary breast cancer. Int. J. Radiat. Oncoi. BioI. Phys. 14:179-190; 1988. 9. Fraass, B. A; Roberson, P. L.; Lichter, A S. Dose to the

10.

11.

12.

13.

14.

15.

16.

17.

18.

contralateral breast due to primary breast irradiation. Int. J. Radiat. Oncoi. BioI. Phys. 11:609-616; 1985. Haffty, B. G.; Goldberg, N. B.; Fischer, D.; McKhann, c.; Beinfield, M.; Weissberg, J. B.; Carter, D.; Gerald, W. Conservative surgery and radiation therapy in breast carcinoma: Local recurrence and prognostic implications. Int. J. Radiat. Oncoi. BioI. Phys. 17:727-732; 1989. Heimann, R.; Powers, c.; Halpern, H. J.; Michel, A G.; Ewing, C. A; Wyman, B.; Recant, W.; Weichselbaum, R. R Breast preservation in stage I and II carcinoma of the breast. Cancer 78:1722-1730; 1996. Jacobson, J. A; Danforth, D. N.; Cowan, K; d'Angeleo, T.; Steinberg, S.; Pierce, L. J.; Lichter, AS.; Glatstein, E.; Okunieff, P. Ten-year results of the National Cancer Institute's randomized trial of breast conservation versus mastectomy for Stages I and II breast cancer. N. Engi. J. Med. 332:907-911; 1995. Kaplan, E. L.; Meier, P. Nonparametric estimation from incomplete observations. J. Am. Stat. Assoc. 53:457-481; 1958. Kurtz, J. M.; Jacquemier, J.; Amalric, R.; Brandone, H.; Ayme, Y.; Hans, D.; Bressac, C.; Spitalier, J. Why are local recurrences after breast-conserving therapy more frequent in younger patients? J. Clin. Oncoi. 8:591-598; 1990. Matthews, R H.; McNeese, M. D.; Montague, E. D.; Oswald, M. J. Prognostic implications of age in breast cancer patients treated with tumorectomy and irradiation or with mastectomy. Int. J. Radiat. Oncoi. BioI. Phys. 14:659-663; 1988. Recht, A.; Connolly, J. L.; Schnitt, S. J.; Silver, B.; Rose, M. A; Love, S.; Harris, J. R. The effect of young age on tumor recurrence in the treated breast after conservative surgery and radiotherapy. Int. J. Radiat. Oncoi. BioI. Phys. 14:3-10; 1988. Recht, A.; Silver, B.; Schnitt, S.; Connolly, J.; Hellman, S.; Harris, J. R. Breast relapse following primary radiation therapy for early breast cancer. I. Classification, frequency and salvage. Int. J. Radiat. Oncoi. BioI. Phys. 11:1271-1276; 1985. Rosenblum, L. J.; Mauceri, R A; Wellenstein, D. E.; Thomas, F. D.; Bassano, D. A; Raasch, B. N.; Chamberlain, C. C.; Heitzman, E. R Density patterns in the normal lung as deter-

928

1. J. Radiation Oncology • Biology • Physics

mined by computerized tomography. Radiology 137:409416; 1980. 19. Sarrazin, D.; Le, M. G.; Arriagada, R.; Contesso, G.; Fontaine, F.; Spielmann, M.; Rochard, F.; Le Chevalier, T.; Lacour, J. Ten-year results of a randomized trial comparing a conservative treatment to mastectomy in early breast cancer. Radiother. Oncol. 14:177-184; 1989. 20. Schnitt, S. J.; Abner, A.; Gelman, R.; Connolly, J. L.; Recht, A.; Duda, R. B.; Eberlein, T. J.; Mayzel, K.; Silver, B.; Harris, J. R. The relationship between microscopic margins of resection and the risk of local recurrence in patients with breast cancer treated with breast-conserving surgery and radiation therapy. Cancer 74:1746-1751; 1994. 21. Smitt, M. C.; Nowels, K. W.; Zdeblick, M. 1.; Jeffrey, S.; Carlson, R. W.; Stockdale, F. E.; Goffinet, D. R. The importance of the lumpectomy surgical margin status in long term results of breast conservation. Cancer 76:259267; 1995.

Volume 39, Number 4, 1997

22. Solin, L. J.; Fowble, B.; Martz, K. L.; Goodman, R. L. Definitive irradiation for early breast cancer: The University of Pennsylvania experience. Int. 1. Radiat. Oncol. BioI. Phys. 14:235-242; 1988. 23. Van Dyk, J.; Keane, T. J.; Rider, W. D. Lung density as measured by computerized tomography: Implications for radiotherapy. Int. J. Radiat. Oncol. BioI. Phys. 8:1361-1372; 1982. 24. Veronesi, U.; Bonadonna, G.; Zurrida, S.; Galimberti, V.; Greco, M.; Brambilla, C.; Luini, A.; Andreola, S.; Rilke, F.; Raselli, R. Conservation surgery after primary chemotherapy in large carcinomas of the breast. Ann. Surg. 222:612-618; 1995. 25. Vicini, F. A.; Eberlein, T. J.; Connolly, J. L.; Recht, A.; Abner, A.; Schnitt, S. J.; Silen, W.; Harris, J. R. The optimal extent of resection for patients with stages I or II breast cancer treated with conservative surgery and radiotherapy. Ann. Surg. 214:200-204; 1991.