Wo m e n ’s I m a g i n g • O r i g i n a l R e s e a r c h Sung et al. Breast MRI
Downloaded from www.ajronline.org by Library Of Medicine on 06/05/14 from IP address 128.103.149.52. Copyright ARRS. For personal use only; all rights reserved
Women’s Imaging Original Research
Preoperative Breast MRI for Early-Stage Breast Cancer: Effect on Surgical and Long-Term Outcomes Janice S. Sung1 Jie Li2 Glenys Da Costa1 Sujata Patil1 Kimberly J. Van Zee1 D. David Dershaw 1 Elizabeth A. Morris1 Sung JS, Li J, Da Costa G, et al.
Keywords: breast cancer, breast MRI, preoperative MRI DOI:10.2214/AJR.13.11355 Received June 8, 2013; accepted after revision August 27, 2013. 1 Department of Radiology, Memorial Sloan Kettering Cancer Center, 300 E 66th St, New York, NY 10065. Address correspondence to J. S. Sung ([email protected]). 2
Key Laboratory of Carcinogenesis and Translational Research, Peking University Cancer Hospital and Institute, Beijing, China.
AJR 2014; 202:1376–1382 0361–803X/14/2026–1376 © American Roentgen Ray Society
1376
OBJECTIVE. The purpose of this study was to evaluate the effect of the use of preoperative breast MRI on surgical and long-term outcomes among women with early-stage breast cancer undergoing breast conservation therapy. MATERIALS AND METHODS. A retrospective review yielded the cases of 174 women with stage 0, I, or II breast cancer who underwent preoperative MRI between 2000 and 2004. A control group of 174 patients who did not undergo preoperative MRI before breast conservation therapy was matched by age, histopathologic finding, stage, and surgeon. Features compared included breast density, presence of mammographically occult disease, margin status, lymph node involvement, lymphovascular invasion, extensive intraductal component, hormone receptor status, and use of adjuvant therapy. Outcomes, including rates of reexcision, ipsilateral recurrence, and disease-free survival, were compared by Kaplan-Meier methods and the log-rank test. RESULTS. Patients referred for preoperative breast MRI were more likely to have extremely dense breasts (28% vs 6%, p < 0.0001) and mammographically occult cancer (24% vs 9%, p = 0.0003). The two groups had identical rates of final negative margins, lymph node involvement, lymphovascular invasion, extensive intraductal component status, positive hormone receptor results, and systemic adjuvant therapy. Fewer patients in the preoperative MRI group needed reexcision (29% vs 45%, p = 0.02). The median follow-up period after treatment was 8 years. There was no significant difference in locoregional recurrence (p = 0.33) or disease-free survival (p = 0.73) rates between the two groups. CONCLUSION. Reexcision rates among patients with early breast cancer undergoing conservation therapy were lower among women who underwent preoperative breast MRI. There was no statistically significant effect of the use of preoperative MRI on rates of locoregional recurrence or disease-free survival.
B
reast conservation therapy, defined as local excision followed by whole-breast radiation therapy, is an alternative to mastectomy for early breast cancer and has equivalent survival rates [1–3]. For patients with newly diagnosed breast cancer, MRI is more accurate for determining true tumor size and extent compared with mammography and ultrasound. MRI depicts additional areas of malignancy that are occult with other imaging techniques [4–7]. In a meta-analysis of 19 studies that included 2610 patients [8], preoperative MRI depicted additional sites of disease in 16% of patients. Because rates of local recurrence are increased among patients with incompletely excised cancer, preoperative MRI has been used to maximize tumor excision within the breast [1, 2, 9–11].
Given the high sensitivity of MRI, clinical outcomes potentially improved by use of preoperative MRI include reductions in the rates of reexcision, local recurrence, and the subsequent development of new contralateral cancer or distant metastasis. However, the use of preoperative MRI remains controversial because of conflicting reports on its effect on both surgical and long-term outcomes. Several retrospective studies have concluded that preoperative MRI does not reduce reexcision rates, whereas others have shown a statistically significant reduction [12–14]. Three retrospective studies have also shown conflicting results with respect to long-term outcomes [15–17]. Fischer et al. [15] reported lower recurrence and contralateral new cancer rates among patients who underwent preoperative MRI. Conversely, two other studies [16, 17]
AJR:202, June 2014
Downloaded from www.ajronline.org by Library Of Medicine on 06/05/14 from IP address 128.103.149.52. Copyright ARRS. For personal use only; all rights reserved
Breast MRI showed no difference in local failure, contralateral breast cancer, or distant metastasis rates between patients with and those without MRI staging. One factor possibly contributing to these conflicting results is that the treatment and control groups in those studies were not equivalent in tumor characteristics, including tumor size and nodal status, or the use of systemic chemotherapy. The Comparative Effectiveness of MRI in Breast Cancer (Comice) trial was the first prospective multicenter randomized trial to assess whether use of preoperative MRI in early-stage breast cancer could decrease reoperation rates [18]. Although use of preoperative MRI did not affect reoperation rates, the Comice trial was compromised by a low reexcision rate of approximately 10% in each group, low patient recruitment at participating centers, lack of MRI-guided biopsy and localization capability, and failure to link findings at breast MRI to changes in surgical management [19]. We undertook our study to compare the effect of the use of preoperative breast MRI on short- and long-term outcomes among patients with early-stage breast cancer matched by surgeon and clinical and histologic features. Materials and Methods Study Design and Patient Characteristics In this retrospective, institutional review board–approved study, the cases of 538 patients who underwent breast MRI for preoperative staging between January 2000 and December 2004 were identified. Patients were included if they underwent breast-conserving surgery without neoadjuvant chemotherapy for American Joint Committee on Cancer (AJCC) stage 0, I, or II breast cancer followed by breast irradiation. According to these criteria, 174 patients with breast cancer treated by 11 breast surgeons were defined as the MRI group. Of the initial 538 patients who underwent preoperative breast MRI, 364 patients were excluded for the following reasons: mastectomy (275 patients), neoadjuvant chemotherapy (25 patients), distant metastases (5 patients), excision performed elsewhere (33 patients), no radiation therapy (21 patients), and lost to follow-up (5 patients). The control group included 174 patients with the diagnosis of breast cancer during the same time period as the MRI group who fulfilled the same inclusion criteria but who did not undergo preoperative breast MRI. Control group patients were randomly selected from a database that included 3645 women with newly diagnosed breast
cancer treated at our institution. The control patients were randomly selected and matched one to one by age at diagnosis (5-year increments starting from age group 20–25 years), tumor histologic features (ductal carcinoma in situ [DCIS], invasive ductal carcinoma, invasive lobular carcinoma, or invasive mammary carcinoma), AJCC stage, and breast surgeon. Eighty-five percent of the cases were matched by use of these criteria. Further matches were made by increasing age increments and eliminating the surgeonmatching criterion. There were no significant differences in final margins status, presence of extensive intraductal component (EIC), lymphovascular invasion, and receptor status. Rates of adjuvant chemotherapy, hormonal therapy, and radiation boost were also similar between the two groups. Electronic medical records were reviewed to determine initial presentation, breast density, mammographically occult disease, presence of a synchronous contralateral cancer, and number of operations performed.
Preoperative Imaging Technique and Interpretation MRI examinations were performed with the patient prone in a 1.5-T commercially available system (Sigma, GE Healthcare) with a dedicated surface breast coil. The breast MRI protocol included a localizing sequence followed by a sagittal fat-suppressed T2-weighted sequence (TR/TE, 4000/85). A T1-weighted 3D fat-suppressed fast spoiled gradient-echo sequence (TR/ TE 17/2.4; flip angle, 35°; bandwidth, 31.25 Hz) was performed before and three times after a rapid bolus injection of 0.1 mmol/L of gadopentetate dimeglumine (Magnevist, Berlex) per kilogram of body weight. Image acquisition started after contrast injection and saline bolus administration. Sagittal images were obtained for an acquisition time per volumetric acquisition of less than 3 minutes each. Section thickness was 2–3 mm with no gap with a matrix of 256 × 192 and an FOV of 18– 22 cm. Frequency was in the anteroposterior direction. Image subtraction was performed.
TABLE 1: Matched Variables in Matched Control and Preoperative MRI Groups Clinical Feature
Matched Control (n = 174)
Preoperative MRI (n = 174)
1
25 (14)
25 (14)
Surgeon 2
15 (9)
14 (8)
3
4 (2)
3 (2)
4
35 (20)
35 (20)
5
2 (1)
3 (2)
6
15 (9)
15 (9)
7
28 (16)
28 (16)
8
12 (7)
13 (7)
9
7 (4)
8 (5)
10
7 (4)
6 (3)
11
24 (14)
24 (14)
Histologic result Ductal carcinoma in situ
22 (13)
22 (13)
Invasive ductal carcinoma
117 (67)
116 (67)
Invasive lobular carcinoma
21 (12)
22 (13)
Invasive mammary carcinoma
14 (8)
14 (8)
American Joint Committee on Cancer stage 0
22 (13)
22 (13)
I
101 (58)
101 (58)
IIA
39 (22)
40 (23)
IIB
12 (7)
11 (6)
55
52
Mean age (y)a
Note—Values are numbers of patients with percentages in parentheses. Owing to rounding, some percentages do not total 100. aMatched in 5-year increments.
AJR:202, June 2014 1377
Downloaded from www.ajronline.org by Library Of Medicine on 06/05/14 from IP address 128.103.149.52. Copyright ARRS. For personal use only; all rights reserved
Sung et al. Any suspicious lesion remote from the index lesion potentially representing multifocal, multicentric, or contralateral disease detected at mammography or MRI was routinely sampled by either percutaneous or surgical biopsy. Radiologic-pathologic concordance was determined by the breast imaging radiologist performing the biopsy.
Breast Conservation Surgery and Histopathology Needle localization was performed before surgical excision for all nonpalpable lesions. Positive
or close margins routinely resulted in reexcision to obtain negative margins. Margin status, EIC, lymphovascular invasion, estrogen receptor, progesterone receptor, and ERBB2 receptor (formerly HER2 or HER2/neu) status were routinely reported by pathologists. Final margin status was recorded as negative (all invasive or in situ carcinoma ≥ 2 mm from the inked margin of the final excision specimen), close (invasive or in situ tumor ≤ 2 mm from the inked margin) or positive (invasive or in situ tumor extending to the inked
margin). EIC was present if DCIS constituted more than 25% of the tumor or if tumor was reported as DCIS with microinvasion. Estrogen and progesterone receptor status was defined as positive if immunohistologic staining results were reported as positive or greater than 5%. ERBB2 expression was identified as positive if the immunohistochemical staining result was reported as 3+ or if fluorescence in situ hybridization showed twofold or greater amplification.
Radiation Therapy and Adjuvant Therapy TABLE 2: Unmatched Variables in Matched Control and Preoperative MRI Groups Clinical Feature
Matched Control (n = 174)
Preoperative MRI (n = 174)
Palpable
0.28
Yes
76 (44)
87 (50)
No
98 (56)
87 (50)
Yes
16 (9)
41 (24)
No
157 (90)
132 (76)
NA
1 (< 1)
1 (< 1)
Mammographically occult
0.0003
Breast density Predominantly fatty
p
< 0.0001 6 (3)
3 (2)
Scattered fibroglandular densities
52 (30)
29 (17)
Heterogeneously dense
104 (60)
94 (54)
Extremely dense
11 (6)
48 (28)
NA
1 (< 1)
Risk factors for breast cancer First-degree relative with breast cancer before age 40 y
8
5
0.57
First-degree relative with breast cancer after age 40 y
29
37
0.34
Second-degree relative with breast cancer
59
67
0.43
Previous high-risk lesion
3
2
1.0
Final margin Negative
0.29 156 (90)
154 (89)
Positive
2 (1)
0 (0)
Close
16 (9)
20 (11)
Present
26 (15)
27 (16)
Absent
126 (72)
125 (72)
NA
22 (13)
22 (13)
Present
22 (13)
23 (13)
Absent
129 (74)
131 (75)
NA
23 (13)
20 (12)
Extensive intraductal component
0.88
Vascular invasion
0.93
(Table 2 continues on next page)
1378
Patients underwent either whole-breast radiation with or without a boost to the primary tumor site or intraoperative partial-breast radiation. The 10 patients and their matched controls who underwent partial-breast irradiation were excluded from the analysis of long-term outcomes. Adjuvant systemic therapy was prescribed on the basis of patient age, tumor stage, and receptor status.
Follow-Up Follow-up information was acquired by review of the electronic medical record. Time of follow-up was defined as years from the last surgical procedure to the most recent mammographic or breast MRI examination. Locoregional recurrence was defined as any new invasive or in situ breast cancer occurring in the treated breast or ipsilateral axilla after radiation therapy. Contralateral new breast cancer was defined as any breast cancer diagnosed in the untreated breast during follow-up. Distant metastasis was considered an event after histologic or cytologic examination showed metastasis in distant organs.
Statistical Analysis Patient characteristics of the preoperative MRI and control groups were compared by chisquare or Fisher exact test. Disease-free survival and recurrence-free survival were analyzed with Kaplan-Meier methods and tested with the logrank test. All statistical analyses were two sided, and significance was assigned at p < 0.05.
Results The median follow-up time for survivors after treatment was 8 years for both the control (range, 0–13 years) and MRI (range, 0.03–12 years) groups. Matched features of the two populations included age, tumor type, and tumor stage (Table 1). The groups were also matched for treating surgeon. Unmatched features of the groups are summarized in Table 2. Patients in the preoperative MRI group were more likely to have extremely dense breast parenchyma (p
Downloaded from www.ajronline.org by Library Of Medicine on 06/05/14 from IP address 128.103.149.52. Copyright ARRS. For personal use only; all rights reserved
Women’s Imaging Original Research
Preoperative Breast MRI for Early-Stage Breast Cancer: Effect on Surgical and Long-Term Outcomes Janice S. Sung1 Jie Li2 Glenys Da Costa1 Sujata Patil1 Kimberly J. Van Zee1 D. David Dershaw 1 Elizabeth A. Morris1 Sung JS, Li J, Da Costa G, et al.
Keywords: breast cancer, breast MRI, preoperative MRI DOI:10.2214/AJR.13.11355 Received June 8, 2013; accepted after revision August 27, 2013. 1 Department of Radiology, Memorial Sloan Kettering Cancer Center, 300 E 66th St, New York, NY 10065. Address correspondence to J. S. Sung ([email protected]). 2
Key Laboratory of Carcinogenesis and Translational Research, Peking University Cancer Hospital and Institute, Beijing, China.
AJR 2014; 202:1376–1382 0361–803X/14/2026–1376 © American Roentgen Ray Society
1376
OBJECTIVE. The purpose of this study was to evaluate the effect of the use of preoperative breast MRI on surgical and long-term outcomes among women with early-stage breast cancer undergoing breast conservation therapy. MATERIALS AND METHODS. A retrospective review yielded the cases of 174 women with stage 0, I, or II breast cancer who underwent preoperative MRI between 2000 and 2004. A control group of 174 patients who did not undergo preoperative MRI before breast conservation therapy was matched by age, histopathologic finding, stage, and surgeon. Features compared included breast density, presence of mammographically occult disease, margin status, lymph node involvement, lymphovascular invasion, extensive intraductal component, hormone receptor status, and use of adjuvant therapy. Outcomes, including rates of reexcision, ipsilateral recurrence, and disease-free survival, were compared by Kaplan-Meier methods and the log-rank test. RESULTS. Patients referred for preoperative breast MRI were more likely to have extremely dense breasts (28% vs 6%, p < 0.0001) and mammographically occult cancer (24% vs 9%, p = 0.0003). The two groups had identical rates of final negative margins, lymph node involvement, lymphovascular invasion, extensive intraductal component status, positive hormone receptor results, and systemic adjuvant therapy. Fewer patients in the preoperative MRI group needed reexcision (29% vs 45%, p = 0.02). The median follow-up period after treatment was 8 years. There was no significant difference in locoregional recurrence (p = 0.33) or disease-free survival (p = 0.73) rates between the two groups. CONCLUSION. Reexcision rates among patients with early breast cancer undergoing conservation therapy were lower among women who underwent preoperative breast MRI. There was no statistically significant effect of the use of preoperative MRI on rates of locoregional recurrence or disease-free survival.
B
reast conservation therapy, defined as local excision followed by whole-breast radiation therapy, is an alternative to mastectomy for early breast cancer and has equivalent survival rates [1–3]. For patients with newly diagnosed breast cancer, MRI is more accurate for determining true tumor size and extent compared with mammography and ultrasound. MRI depicts additional areas of malignancy that are occult with other imaging techniques [4–7]. In a meta-analysis of 19 studies that included 2610 patients [8], preoperative MRI depicted additional sites of disease in 16% of patients. Because rates of local recurrence are increased among patients with incompletely excised cancer, preoperative MRI has been used to maximize tumor excision within the breast [1, 2, 9–11].
Given the high sensitivity of MRI, clinical outcomes potentially improved by use of preoperative MRI include reductions in the rates of reexcision, local recurrence, and the subsequent development of new contralateral cancer or distant metastasis. However, the use of preoperative MRI remains controversial because of conflicting reports on its effect on both surgical and long-term outcomes. Several retrospective studies have concluded that preoperative MRI does not reduce reexcision rates, whereas others have shown a statistically significant reduction [12–14]. Three retrospective studies have also shown conflicting results with respect to long-term outcomes [15–17]. Fischer et al. [15] reported lower recurrence and contralateral new cancer rates among patients who underwent preoperative MRI. Conversely, two other studies [16, 17]
AJR:202, June 2014
Downloaded from www.ajronline.org by Library Of Medicine on 06/05/14 from IP address 128.103.149.52. Copyright ARRS. For personal use only; all rights reserved
Breast MRI showed no difference in local failure, contralateral breast cancer, or distant metastasis rates between patients with and those without MRI staging. One factor possibly contributing to these conflicting results is that the treatment and control groups in those studies were not equivalent in tumor characteristics, including tumor size and nodal status, or the use of systemic chemotherapy. The Comparative Effectiveness of MRI in Breast Cancer (Comice) trial was the first prospective multicenter randomized trial to assess whether use of preoperative MRI in early-stage breast cancer could decrease reoperation rates [18]. Although use of preoperative MRI did not affect reoperation rates, the Comice trial was compromised by a low reexcision rate of approximately 10% in each group, low patient recruitment at participating centers, lack of MRI-guided biopsy and localization capability, and failure to link findings at breast MRI to changes in surgical management [19]. We undertook our study to compare the effect of the use of preoperative breast MRI on short- and long-term outcomes among patients with early-stage breast cancer matched by surgeon and clinical and histologic features. Materials and Methods Study Design and Patient Characteristics In this retrospective, institutional review board–approved study, the cases of 538 patients who underwent breast MRI for preoperative staging between January 2000 and December 2004 were identified. Patients were included if they underwent breast-conserving surgery without neoadjuvant chemotherapy for American Joint Committee on Cancer (AJCC) stage 0, I, or II breast cancer followed by breast irradiation. According to these criteria, 174 patients with breast cancer treated by 11 breast surgeons were defined as the MRI group. Of the initial 538 patients who underwent preoperative breast MRI, 364 patients were excluded for the following reasons: mastectomy (275 patients), neoadjuvant chemotherapy (25 patients), distant metastases (5 patients), excision performed elsewhere (33 patients), no radiation therapy (21 patients), and lost to follow-up (5 patients). The control group included 174 patients with the diagnosis of breast cancer during the same time period as the MRI group who fulfilled the same inclusion criteria but who did not undergo preoperative breast MRI. Control group patients were randomly selected from a database that included 3645 women with newly diagnosed breast
cancer treated at our institution. The control patients were randomly selected and matched one to one by age at diagnosis (5-year increments starting from age group 20–25 years), tumor histologic features (ductal carcinoma in situ [DCIS], invasive ductal carcinoma, invasive lobular carcinoma, or invasive mammary carcinoma), AJCC stage, and breast surgeon. Eighty-five percent of the cases were matched by use of these criteria. Further matches were made by increasing age increments and eliminating the surgeonmatching criterion. There were no significant differences in final margins status, presence of extensive intraductal component (EIC), lymphovascular invasion, and receptor status. Rates of adjuvant chemotherapy, hormonal therapy, and radiation boost were also similar between the two groups. Electronic medical records were reviewed to determine initial presentation, breast density, mammographically occult disease, presence of a synchronous contralateral cancer, and number of operations performed.
Preoperative Imaging Technique and Interpretation MRI examinations were performed with the patient prone in a 1.5-T commercially available system (Sigma, GE Healthcare) with a dedicated surface breast coil. The breast MRI protocol included a localizing sequence followed by a sagittal fat-suppressed T2-weighted sequence (TR/TE, 4000/85). A T1-weighted 3D fat-suppressed fast spoiled gradient-echo sequence (TR/ TE 17/2.4; flip angle, 35°; bandwidth, 31.25 Hz) was performed before and three times after a rapid bolus injection of 0.1 mmol/L of gadopentetate dimeglumine (Magnevist, Berlex) per kilogram of body weight. Image acquisition started after contrast injection and saline bolus administration. Sagittal images were obtained for an acquisition time per volumetric acquisition of less than 3 minutes each. Section thickness was 2–3 mm with no gap with a matrix of 256 × 192 and an FOV of 18– 22 cm. Frequency was in the anteroposterior direction. Image subtraction was performed.
TABLE 1: Matched Variables in Matched Control and Preoperative MRI Groups Clinical Feature
Matched Control (n = 174)
Preoperative MRI (n = 174)
1
25 (14)
25 (14)
Surgeon 2
15 (9)
14 (8)
3
4 (2)
3 (2)
4
35 (20)
35 (20)
5
2 (1)
3 (2)
6
15 (9)
15 (9)
7
28 (16)
28 (16)
8
12 (7)
13 (7)
9
7 (4)
8 (5)
10
7 (4)
6 (3)
11
24 (14)
24 (14)
Histologic result Ductal carcinoma in situ
22 (13)
22 (13)
Invasive ductal carcinoma
117 (67)
116 (67)
Invasive lobular carcinoma
21 (12)
22 (13)
Invasive mammary carcinoma
14 (8)
14 (8)
American Joint Committee on Cancer stage 0
22 (13)
22 (13)
I
101 (58)
101 (58)
IIA
39 (22)
40 (23)
IIB
12 (7)
11 (6)
55
52
Mean age (y)a
Note—Values are numbers of patients with percentages in parentheses. Owing to rounding, some percentages do not total 100. aMatched in 5-year increments.
AJR:202, June 2014 1377
Downloaded from www.ajronline.org by Library Of Medicine on 06/05/14 from IP address 128.103.149.52. Copyright ARRS. For personal use only; all rights reserved
Sung et al. Any suspicious lesion remote from the index lesion potentially representing multifocal, multicentric, or contralateral disease detected at mammography or MRI was routinely sampled by either percutaneous or surgical biopsy. Radiologic-pathologic concordance was determined by the breast imaging radiologist performing the biopsy.
Breast Conservation Surgery and Histopathology Needle localization was performed before surgical excision for all nonpalpable lesions. Positive
or close margins routinely resulted in reexcision to obtain negative margins. Margin status, EIC, lymphovascular invasion, estrogen receptor, progesterone receptor, and ERBB2 receptor (formerly HER2 or HER2/neu) status were routinely reported by pathologists. Final margin status was recorded as negative (all invasive or in situ carcinoma ≥ 2 mm from the inked margin of the final excision specimen), close (invasive or in situ tumor ≤ 2 mm from the inked margin) or positive (invasive or in situ tumor extending to the inked
margin). EIC was present if DCIS constituted more than 25% of the tumor or if tumor was reported as DCIS with microinvasion. Estrogen and progesterone receptor status was defined as positive if immunohistologic staining results were reported as positive or greater than 5%. ERBB2 expression was identified as positive if the immunohistochemical staining result was reported as 3+ or if fluorescence in situ hybridization showed twofold or greater amplification.
Radiation Therapy and Adjuvant Therapy TABLE 2: Unmatched Variables in Matched Control and Preoperative MRI Groups Clinical Feature
Matched Control (n = 174)
Preoperative MRI (n = 174)
Palpable
0.28
Yes
76 (44)
87 (50)
No
98 (56)
87 (50)
Yes
16 (9)
41 (24)
No
157 (90)
132 (76)
NA
1 (< 1)
1 (< 1)
Mammographically occult
0.0003
Breast density Predominantly fatty
p
< 0.0001 6 (3)
3 (2)
Scattered fibroglandular densities
52 (30)
29 (17)
Heterogeneously dense
104 (60)
94 (54)
Extremely dense
11 (6)
48 (28)
NA
1 (< 1)
Risk factors for breast cancer First-degree relative with breast cancer before age 40 y
8
5
0.57
First-degree relative with breast cancer after age 40 y
29
37
0.34
Second-degree relative with breast cancer
59
67
0.43
Previous high-risk lesion
3
2
1.0
Final margin Negative
0.29 156 (90)
154 (89)
Positive
2 (1)
0 (0)
Close
16 (9)
20 (11)
Present
26 (15)
27 (16)
Absent
126 (72)
125 (72)
NA
22 (13)
22 (13)
Present
22 (13)
23 (13)
Absent
129 (74)
131 (75)
NA
23 (13)
20 (12)
Extensive intraductal component
0.88
Vascular invasion
0.93
(Table 2 continues on next page)
1378
Patients underwent either whole-breast radiation with or without a boost to the primary tumor site or intraoperative partial-breast radiation. The 10 patients and their matched controls who underwent partial-breast irradiation were excluded from the analysis of long-term outcomes. Adjuvant systemic therapy was prescribed on the basis of patient age, tumor stage, and receptor status.
Follow-Up Follow-up information was acquired by review of the electronic medical record. Time of follow-up was defined as years from the last surgical procedure to the most recent mammographic or breast MRI examination. Locoregional recurrence was defined as any new invasive or in situ breast cancer occurring in the treated breast or ipsilateral axilla after radiation therapy. Contralateral new breast cancer was defined as any breast cancer diagnosed in the untreated breast during follow-up. Distant metastasis was considered an event after histologic or cytologic examination showed metastasis in distant organs.
Statistical Analysis Patient characteristics of the preoperative MRI and control groups were compared by chisquare or Fisher exact test. Disease-free survival and recurrence-free survival were analyzed with Kaplan-Meier methods and tested with the logrank test. All statistical analyses were two sided, and significance was assigned at p < 0.05.
Results The median follow-up time for survivors after treatment was 8 years for both the control (range, 0–13 years) and MRI (range, 0.03–12 years) groups. Matched features of the two populations included age, tumor type, and tumor stage (Table 1). The groups were also matched for treating surgeon. Unmatched features of the groups are summarized in Table 2. Patients in the preoperative MRI group were more likely to have extremely dense breast parenchyma (p
