MRI evaluation of residual tumor size after neoadjuvant endocrine therapy vs. neoadjuvant chemotherapy

European Journal of Radiology 81 (2012) 2148–2153

Contents lists available at ScienceDirect

European Journal of Radiology journal homepage: www.elsevier.com/locate/ejrad

MRI evaluation of residual tumor size after neoadjuvant endocrine therapy vs. neoadjuvant chemotherapy Kazuna Takeda a , Shotaro Kanao a,∗ , Tomohisa Okada a , Takayuki Ueno b , Masakazu Toi b , Hiroshi Ishiguro c , Yoshiki Mikami d , Shiro Tanaka e , Kaori Togashi a a

Department of Diagnostic Imaging and Nuclear Medicine, Kyoto University Graduate School of Medicine, 54 Shogoin Kawaharacho, Sakyoku, Kyoto 606-8507, Japan Department of Breast Surgery, Kyoto University Graduate School of Medicine, Kyoto 606-8507, Japan c Outpatient Oncology Unit, Kyoto University Graduate School of Medicine, Kyoto 606-8507, Japan d Department of Diagnostic Pathology, Kyoto University Graduate School of Medicine, Kyoto 606-8507, Japan e Translational Research Center, Kyoto University Graduate School of Medicine, Kyoto 606-8507, Japan b

a r t i c l e

i n f o

Article history: Received 28 February 2011 Received in revised form 9 May 2011 Accepted 10 May 2011 Keywords: MRI Breast cancer Neoadjuvant endocrine therapy Neoadjuvant chemotherapy Residual tumor size

a b s t r a c t Aim: To investigate if there is any difference in evaluation of residual tumor size after neoadjuvant chemotherapy (NAC) and neoadjuvant endocrine therapy (NAE). Methods: Seventy-eight tumors in 57 patients were prospectively enrolled. Residual tumor sizes in contrast-enhanced MRI after NAC and NAE were compared with those measured on surgical specimen by using linear regression analyses. The line slope values >1 indicates overestimation by MRI. Differences in types of shrinkage patterns: concentric shrinkage (CS) and dendritic shrinkage (DS) were also investigated. Results: Fifty lesions were treated with NAC and 28 lesions were treated with NAE. Shrinkage patterns were CS in 33 lesions and in 45 lesions. The slopes values were 0.75 (R = 0.92) and 0.70 (R = 0.90) for NAC and NAE, respectively, and no significant difference was observed (p = 0.46). However, they were 1.02 (R = 0.92) and 0.68 (R = 0.92), respectively for CS and DS with significant difference (p < 0.01). The difference between CS and DS was found only in a subgroup with size by MRI >20 mm. Conclusion: Contrast enhanced MRI enabled fairly accurate measurement in NAE as well as in NAC. © 2011 Elsevier Ireland Ltd. All rights reserved.

1. Introduction Neoadjuvant chemotherapy (NAC), which is conducted before surgery, is now commonly adopted to treat breast cancer [1]. Initially, neoadjuvant chemotherapy was indicated in patients with locally advanced or inflammatory breast cancers, then indication has become to include initially operable breast cancer that is not a candidate for breast conserving surgery or that would get some benefit from tumor volume reduction [2]. Advantages of this therapeutic approach are: early treatment of micrometastatic disease, assessment of tumor response to specific chemotherapeutic regimens in vivo and analyses of biological markers that may predict the response. Randomized trials comparing neoadjuvant chemotherapy and postoperative adjuvant chemotherapy have shown similar survival benefit, and preoperative chemotherapy has allowed more patients to have successful breast-conserving treatment [3]. Moreover, complete pathological response of primary breast

∗ Corresponding author. Tel.: +81 75 751 3760; fax: +81 75 771 9709. E-mail address: [email protected] (S. Kanao). 0720-048X/$ – see front matter © 2011 Elsevier Ireland Ltd. All rights reserved. doi:10.1016/j.ejrad.2011.05.013

cancer to neoadjuvant chemotherapy is a surrogate marker for patient outcome [3]. Evaluation of tumor before surgery has been conducted by some modalities including mammography, ultrasound and MRI. MRI is recognized as the most reliable method for detection of residual tumor after NAC [4–7]. Recently, neoadjuvant endocrine therapy (NAE) is gradually adopted in addition to NAC [8]. NAE is considered to be another therapeutic option for a tumor that is highly sensitive to endocrine treatment, especially in elderly population [9]. NAE is often selected for cases of less aggressive cancer than NAC [10], and their similarity and difference in MRI evaluation after presurgical systemic therapies have not been much investigated and still remains to be clarified. Discrepancy was sometimes observed after NAC between assumed residual tumor size evaluated by MRI and that confirmed by resected specimen [7,11]. Patterns of shrinkage, i.e., concentric shrinkage (CS) or dendritic shrinkage (DS) [12], and status of human epidermal growth factor receptor related 2 (HER2), i.e., positive or negative [13], were reported to have some roles for correct evaluation by MRI. Confirmation on MRI evaluation of residual tumor size after NAE has large clinical importance for selection of further therapeutic options.

K. Takeda et al. / European Journal of Radiology 81 (2012) 2148–2153 Table 2 Lesion characteristics: CS vs. DS.

Table 1 Lesion characteristics: NAC vs. NAE.

Mean age (range) Clinical tumor stage T1 T2 T3 T4 Nodal status Negative Positive Histological subtype IDC NOS ILC IDC + ILC Metaplastic carcinoma Mucinous carcinoma Histological grade Grade 1 Grade 2 Grade 3 Receptor status ER and/or PR (+) HER2 (+) ER and/or PR (+) HER2 (−) ER and PR (−) HER2 (+) ER and PR (−) HER2 (−) Ki-67 labeling index Low (<15%) Intermediate (16–30%) High (>31%) n.a. Final pathological status Non-CR CR

NAC lesions n = 50

NAE lesions n = 28

51 (34–76)

59 (39–80)

27 (54%) 16 (32%) 2 (4%) 5 (10%)

9 (32%) 17 (61%) 1 (4%) 1 (4%)

31 (62%) 19 (38%)

18 (64%) 10 (36%)

48 (96%) 0 (0%) 1 (2%) 1 (2%) 0 (0%)

24 (86%) 1 (4%) 0 (0%) 0 (0%) 3 (11%)

3 (6%) 31 (62%) 16 (32%)

12 (43%) 14 (50%) 2 (7%)

8 (16%) 25 (50%) 6 (12%) 11 (22%)

0 (0%) 28 (100%) 0 (0%) 0 (0%)

7 (14%) 8 (16%) 10 (20%) 25 (50%)

23 (82%) 2 (7%) 0 (0%) 3 (11%)

p < 0.01 p = 0.10

p = 0.84

p = 0.04

p < 0.01

p < 0.01

p < 0.01

p < 0.01 39 (76%) 12 (24%)

2149

27 (96%) 1 (4%)

Mean age (range) Clinical tumor stage T1 T2 T3 T4 Nodal status Negative Positive Histological subtype IDC NOS ILC IDC + ILC Metaplastic carcinoma Mucinous carcinoma Pathological grade Grade 1 Grade 2 Grade 3 Ki-67 labeling index Low (<15%) Intermediate (16–30%) High (>30%) n.a. Final pathological status Non-CR CR

CS lesions n = 33

DS lesions n = 45

54 (34–80)

55 (34–76)

13 (39%) 18 (55%) 1 (3%) 1 (3%)

23 (51%) 15 (33%) 2 (4%) 5 (11%)

20 (61%) 13 (39%)

28 (65%) 17 (35%)

30 (91%) 0 (0%) 0 (0%) 1 (3%) 2 (6%)

42 (93%) 1 (2%) 1 (2%) 0 (0%) 1 (2%)

9 (27%) 16 (48%) 8 (24%)

6 (13%) 29 (64%) 10 (22%)

12 (36%) 3 (9%) 3 (9%) 15 (45%)

18 (40%) 7 (16%) 7 (16%) 13 (29%)

25 (76%) 8 (24%)

40 (89%) 5 (11%)

p = 0.58 p = 0.10

p = 0.89

p = 0.31

p = 0.28

p = 0.77

p = 0.13

Abbreviations: CS: concentric shrinkage; DS: dendritic shrinkage; ER: estrogen receptor; PR: progesterone receptor; HER2: human epidermal growth factor receptor related 2; IDC: invasive ductal carcinoma; NOS: not otherwise specified; IDC: invasive ductal carcinoma; ILC: invasive lobular carcinoma; CR: complete response.

Abbreviations: NAC: neoadjuvant chemotherapy; NAE: neoadjuvant endocrine therapy; n.a.: not accessed. ER: estrogen receptor; PR: progesterone receptor; HER2: human epidermal growth factor receptor related 2; IDC: invasive ductal carcinoma; NOS: not otherwise specified; ILC: invasive lobular carcinoma; CR: complete response.

Therefore, we made a prospective study for size evaluation of residual tumor by MRI at the end of presurgical systemic therapies of NAC and NAE, and compared the results with pathological specimen. They were analyzed in terms of differences between NAC and NAE, and also between shrinkage patterns of CS and DS.

2. Materials and methods 2.1. Patients and therapeutics This study was approved by institutional review board and informed consent was acquired by all subjects enrolled. Seventynine pathologically confirmed breast cancer in 58 female patients (34–80 years old, average 54 years) were prospectively enrolled from June 2008 to February 2010, however, one tumor of progressive disease in one patient was excluded, because our study also included analysis of shrinkage cases. Therefore, 78 carcinomas in 57 patients were included for further analysis. They were treated with NAC or NAE and surgically resected thereafter at our hospital. Therapy was decided by the patient with informed consent after discussion with breast surgeons and medical oncologist. Many were based on St. Gallen Consensus Guideline 2007 or 2009 [10,14]. NAC was defined as usage of standard chemotherapeutic agents of taxane or anthracycline and their derivatives without any endocrine therapy. NAE was defined as application of endocrine therapy without chemotherapy except cyclophosphamide. The basic regimen of NAC was a combination of docetaxel (75 mg/m2 ) and cyclophosphamide (600 mg/m2 ) for 3–6 cycles in every 3 weeks. For NAE, letrozole was administered with addition of some

Fig. 1. Linear relationship between MRI measurement compared with that on pathological specimen. By therapeutic regimens, no significant interaction between NAC and NAE was observed (p = 0.46). NAC: neoadjuvant chemotherapy; NAE: neoadjuvant endocrine therapy.

2150

K. Takeda et al. / European Journal of Radiology 81 (2012) 2148–2153

the abnormally enhancing lesion or lesions corresponding to the known malignancy was measured using contrast-enhanced 3Dvolume images reconstructed in multiple planes by using Aquarius NET Server (TeraRecon Inc., Tokyo, Japan). This maximum dimension was chosen to represent the extent of disease present in the breast. Thus, if the lesion consisted of multiple adjacent abnormalities, the maximum dimension was not the sum of their diameters, but a single measurement encompassing the lesions farthest apart [6]. The shrinkage patterns were divided into two types [11,12]. One type was concentric shrinkage (CS), which indicates the tumor shrunk from all directions toward a center. The other type was dendritic shrinkage (DS), which indicates fragmentation into multiple smaller tumor foci or uneven shrinkage of the tumor resulting in dendritic shape. 2.4. Pathological analysis

Fig. 2. Linear relationship between MRI measurement compared with that on pathological specimen. By shrinkage patterns, significant interaction between shrinkage patterns of Concentric and Dendritic was observed (p < 0.01).

other agents (R2) (cyclophosphamide, LH–RH analogue, etc.), when considered necessary. 2.2. MRI protocol Using the same protocol, MRI evaluation was conducted twice before and after neoadjuvant therapies. MRI was acquired with a 3 T scanner (Magnetom Trio Tim; Siemens Medical Solutions, Erlangen, Germany) with a breast-dedicated 4-channel coil (Invivo, FL, USA) at the end of NAC or NAE. After getting pre-contrast axial T2-weighted, T1-wighted and diffusion-weighted images, fat-suppressed T1-weighted dynamic images were acquired once before and three times after Gadolinium infusion. At 0–1, 1–2 and 5–6 min after injection. The whole breasts were scanned in high temporal resolution of 1 min (3D-VIBE: TR/TE 3.8/1.48 ms, FA 15, FOV 330 mm × 330 mm, matrix 448 × 461, 2.5 mm thickness, 60 slices) in axial orientation. At 2–5 min, a scan was conducted with high spatial resolution (3D-VIBE: TR/TE 4.2/1.5 ms, FA 15, FOV 330 mm × 330 mm, matrix 448 × 412, 0.8 mm thick, 176 slices) in coronal orientation. Infused Gadolinium contrast materials were either Gadoteridol (ProHance, Eisai Inc., Tokyo, Japan) (R3) for patients under or equal to 60 kg, or Gadodiamide (Omniscan, Daiichi-Sankyo Inc., Tokyo, Japan) (R3) for patients over 60 kg, with a dose of 0.2 ml/kg power injected at the speed of 2.0 ml/s and flashed with 20 ml of saline at the same rate (R3). These two contrast agents was reported to have similar contrast effect [15]. 2.3. Image analysis Residual tumor was defined by their degree, pattern and distribution of dynamic enhancement. Enhancements at the area where a tumor existed before therapy were considered to indicate a residual tumor. Enhancements in other area were considered nonspecific except when they had definite malignant morphology or dynamic enhancement pattern. Evaluation was conducted by an experienced radiologist (Shotaro Kanao) with 10-year experience in breast MRI examination (R4). He was blinded to pathological results and other imaging results. The maximum dimension of

Surgical resections were conducted within 2 weeks after MRI scans. The specimen were dissected and evaluated by pathologists. They were sectioned at 5-mm intervals perpendicular to the longest axis of the specimen. Partial resection specimens were completely embedded (and in cases of mastectomy grossly visible tumor beds were also thoroughly sampled) for optimal evaluation of tumor response. Pathological complete response (pCR) was defined as no residual invasive tumor on the specimen. The pathological size was defined as longest diameter of invasive tumor. 2.5. Statistical analysis General linear model analyses with no intercept term were conducted to examine the linear relationship between size measurements by MRI and those of pathological specimen. Regression coefficient for the slope greater than 1 meant overestimation in MRI measurements. Discrepancies between NAC vs. NAE were analyzed by fitting regression lines separately and by tests for interaction terms. Difference in CS vs. DS was also analyzed in the same manner [7]. Statistical analysis was conducted by using SAS version 9.2 (SAS Institute Inc., Cary, NC, USA). p Values less than 0.05 were considered significant. 3. Results 3.1. Pathological findings Characteristics of patients and lesions were analyzed for differences between NAC vs. NAE, as well as CS vs. DS in terms of mean age, clinical tumor stage, nodal status, histological subtype, histological grade, receptor status, Ki-67 labeling index, and final pathological status (see Tables 1 and 2). Significant difference was observed between NAC vs. NAE in age, histological subtype, histological grade, receptor status, Ki-67 labeling index, and final pathological status. These differences were unavoidable because selection of neoadjuvant therapy linked to those parameters. No significant difference was observed between CS vs. DS. 3.2. Size evaluation with MRI vs. microscopic examination of specimen A linear relationship was observed between size evaluation by MRI and pathological examination. The regression coefficient for slope of all tumors was 0.73 (95% confidence interval (CI): 0.65–0.80, R2 = 0.83) and the relationship was statistically significant (p < 0.01). The regression coefficients of slopes were 0.75 (95%

K. Takeda et al. / European Journal of Radiology 81 (2012) 2148–2153

2151

Fig. 3. A concentric shrinkage (CS) case. (A) Dynamic contrast-enhanced T1WI at early phase before NAC. Lobulated mass with rim enhancement was observed at upper-inner area of the right breast. (B) Dynamic contrast-enhanced T1WI at early phase after NAC. The early phase image showed rapid enhanced focus about 4 mm in diameter which was identical to the residual tumor at the pathological specimen. (C) Dynamic contrast-enhanced T1WI at delayed phase after NAC. Delayed image showed heterogeneous non-mass-like enhancement about 23 mm in diameter which was identical to the scar area at the pathological specimen.

CI: 0.66–0.84, R = 0.92) and 0.70 (95% CI: 0.56–0.83, R = 0.90) for NAC and NAE, respectively (Fig. 1), and no significant interaction was observed between them (p = 0.46). In contrast, the analysis of interaction between shrinkage patterns (Fig. 2) revealed regression coefficients of slopes as 1.02 (95% CI: 0.87–1.17, R = 0.92) and 0.68 (95% CI: 0.60–0.77, R = 0.92), respectively for CS and DS with significant difference (p < 0.01). In subgroup analyses by residual tumor size, this result was confirmed when lesion size by MRI was larger than 20 mm (regression coefficients (95% CI): 0.99 (0.47–1.51, R = 0.91) and 0.67 (0.51–0.83, R = 0.94)), but the difference between CS and DS was not found, when residual tumor size by MRI was equal to or less than 20 mm (1.08 (0.94–1.21, R = 0.95) and 1.11 (0.86–1.36, R = 0.84)). 4. Discussion MRI enabled fairly accurate measurement of the largest diameter after NAE as well as NAC, with slight underestimation in both cases, and no significant difference was observed between NAE and NAC in this study (Fig. 1) despite differences in grades,

receptor status, Ki-67 index and others between NAE and NAC [10]. Correct visualization of residual tumor after presurgical therapy is highly important, because breast-conserving surgery can be adopted when residual tumor is small. It significantly lessens the burden on patients [16] and is important for prognosis, because overall and disease free survival is higher in cases of pathological CR [3]. Compared to palpation, mammography and ultrasound, MRI is more accurate and reliable for evaluation of residual tumor after NAC [4]. MRI provided good correlation of residual tumor size in general compared with pathological evaluation, however, there was a difference dependent on shrinkage patterns, i.e., CS vs. DS. Tumors showing CS were overestimated (Fig. 2), but only to a very slight degree (Fig. 3). It has been reported that overestimation of residual tumor size by MRI compared with pathology was caused by inclusion of ductal carcinoma in situ (DCIS)[7] and reactive changes [17]. Differential diagnosis of DCIS and invasive ductal carcinoma (IDC) can be challenging with MRI [7]. For pathological TNM staging, DCIS is not considered as a residual lesion, and consequently the overestimation on MRI is considered inevitable.

2152

K. Takeda et al. / European Journal of Radiology 81 (2012) 2148–2153

be caused by disappearance of tumor vessels by systemic therapies [19]. Major factors that affect precision of residual tumor size measurement appear to include differences in breast form and measurement planes between MR images and specimen. The resection specimen is generally soft and may spread flat [7], which were considered inevitable and may hold true for any case in general (R5). Interobserver or intraobserver variation may seem to affect precision of measurement. But interobserver variation in measuring residual tumor size in contrast-enhanced MRI was reported to be 11%, and high accuracy is expected [5]. Accuracy of MRI evaluation would be improved by including other method such as 3D volume MRI data and computer assisted detection system [20]. Some limitations may exist in this study. There were some variations in regimens of NAC and NAE. They were customized for each individual and were another confounding factor. The other issue is the selection bias for NAE patients. They were selected based on the St. Gallen Consensus Guideline 2007 or 2009, and were strongly positive for hormonal receptor, low to intermediate tumor grade and had less than 30% for Ki-67 index. Not all selection of NAE or NAC was conducted in accordance with the guideline, because some patients selected their own therapy otherwise. There also exist some variances in tumor characteristics (see Tables 1 and 2). In conclusion, whether NAC or NAE was applied, contrast enhanced MRI enables fairly accurate measurement of the largest diameter. However, it tends to underestimate the residual tumor size, especially when the tumor shrank in DS pattern. Caution has to be paid when evaluating residual tumor size with MRI, taking differences not in kinds of therapeutic methods but in shrinkage patterns in mind.

References

Fig. 4. Underestimation in a dendritic shrinkage (DS) case. (A) Dynamic contrastenhanced T1WI at early phase before NAE. There were an irregular shaped mass near the nipple and multiple non-mass-like enhancements peripherally. (B) Dynamic contrast-enhanced T1WI at early phase after NAE. The tumor shrank as a whole. However, there remained many small tumor foci scattered at periphery in pathological evaluation. Small lesions were not well detected on MRI.

Underestimation of residual tumor size compared with pathology was observed, when little or no enhancement was present in the residual tumor where tumor cell density was low. Such cases were more frequent in DS cases (Fig. 4) as was previously reported [7,18], and a caution has to be paid for resection. Most of the largesize residual tumors had the DS pattern, and no difference in MR estimation was found between CS and DS when residual tumor size was equal to or less than 2 cm on MRI. It may suggest that residual tumor size could be another factor for underestimation as well as the shrinkage patterns. A tendency for underestimation in the DS cases was reported in a CT study [18]. They named DS pattern as “shrinkage with residual multi-nodular lesions,” because “these lesions were characterized histologically by the presence of numerous viable cells with fibrous tissue spread throughout the original tumor bed,” which was the case in our histopathological results. Another study called DS pattern as “nest or rim pattern,” and found that it had lower agreement in size evaluation between MRI and histopathology [7]. Reduction in enhancement and washout may

[1] Tardivon A, Ollivier L, El Khoury C, Thibault F. Monitoring therapeutic efficacy in breast carcinomas. Eur Radiol 2006;16(11):2549–58. [2] van der Hage JA, van de Velde CJ, Julien JP, Tubiana-Hulin M, Vandervelden C, Duchateau L. Preoperative chemotherapy in primary operable breast cancer: results from the European Organization for Research and Treatment of Cancer trial 10902. J Clin Oncol 2001;19(22):4224–37. [3] Wolmark N, Wang J, Mamounas E, Bryant J, Fisher B. Preoperative chemotherapy in patients with operable breast cancer: nine-year results from National Surgical Adjuvant Breast and Bowel Project B-18. J Natl Cancer Inst Monogr 2001;30:96–102. [4] Balu-Maestro C, Chapellier C, Bleuse A, Chanalet I, Chauvel C, Largillier R. Imaging in evaluation of response to neoadjuvant breast cancer treatment benefits of MRI. Breast Cancer Res Treat 2002;72(2):145–52. [5] Partridge S, Gibbs J, Lu Y, Esserman L, Sudilovsky D, Hylton N. Accuracy of MR imaging for revealing residual breast cancer in patients who have undergone neoadjuvant chemotherapy. Am J Roentgenol 2002;179(5):1193–9. [6] Rosen E, Blackwell K, Baker J, et al. Accuracy of MRI in the detection of residual breast cancer after neoadjuvant chemotherapy. Am J Roentgenol 2003;181(5):1275–82. [7] Kim HJ, Im YH, Han BK, et al. Accuracy of MRI for estimating residual tumor size after neoadjuvant chemotherapy in locally advanced breast cancer: relation to response patterns on MRI. Acta Oncol 2007;46(7):996–1003. [8] Connolly R, Stearns V. A multidisciplinary approach to neoadjuvant therapy for primary operable breast cancer. Challenges and opportunities. Oncology (Williston Park) 2010;24(2):135–43. [9] Sachelarie I, Grossbard ML, Chadha M, Feldman S, Ghesani M, Blum RH. Primary systemic therapy of breast cancer. Oncologist 2006;11(6):574–89. [10] Goldhirsch A, Ingle J, Gelber R, et al. Thresholds for therapies: highlights of the St. Gallen International Expert Consensus on the primary therapy of early breast cancer 2009. Ann Oncol 2009;20(8):1319–29. [11] Thibault F, Nos C, Meunier M, et al. MRI for surgical planning in patients with breast cancer who undergo preoperative chemotherapy. Am J Roentgenol 2004;183(4):1159–68. [12] Nakamura S, Kenjo H, Nishio T, Kazama T, Doi O, Suzuki K. Efficacy of 3D-MR mammography for breast conserving surgery after neoadjuvant chemotherapy. Breast Cancer 2002;9(1):15–9. [13] Chen JH, Feig B, Agrawal G, et al. MRI evaluation of pathologically complete response and residual tumors in breast cancer after neoadjuvant chemotherapy. Cancer 2008;112(1):17–26.

K. Takeda et al. / European Journal of Radiology 81 (2012) 2148–2153 [14] Goldhirsch A, Wood W, Gelber R, et al. Progress and promise: highlights of the international expert consensus on the primary therapy of early breast cancer 2007. Ann Oncol 2007;18(7):1133–44. [15] Rinck PA, Muller RN. Field strength and dose dependence of contrast enhancement by gadolinium-based MR contrast agents. Eur Radiol 1999;9(5):998–1004. [16] Deo SV, Bhutani M, Shukla NK, Raina V, Rath GK, Purkayasth J. Randomized trial comparing neo-adjuvant versus adjuvant chemotherapy in operable locally advanced breast cancer (T4b N0-2 M0). J Surg Oncol 2003;84(4):192–7. [17] Liberman L, Morris E, Lee M, et al. Breast lesions detected on MR imaging: features and positive predictive value. Am J Roentgenol 2002;179(1):171–8.

2153

[18] Tozaki M, Kobayashi T, Uno S, et al. Breast-conserving surgery after chemotherapy: value of MDCT for determining tumor distribution and shrinkage pattern. Am J Roentgenol 2006;186(2):431–9. [19] Wasser K, Sinn H, Fink C, et al. Accuracy of tumor size measurement in breast cancer using MRI is influenced by histological regression induced by neoadjuvant chemotherapy. Eur Radiol 2003;13(6):1213–23. [20] Partridge SC, Gibbs JE, Lu Y, et al. MRI measurements of breast tumor volume predict response to neoadjuvant chemotherapy and recurrence-free survival. Am J Roentgenol 2005;184(6):1774–81.