Evaluation with 3.0-T MR imaging: predicting the pathological response of triple-negative breast cancer treated with anthracycline and taxane neoadjuvant chemotherapy

Abstract

Background

Triple-negative breast cancer (TNBC) which expresses neither hormonal receptors nor HER-2 is associated with poor prognosis and shorter survival. Several studies have suggested that TNBC patients attaining pathological complete response (pCR) after neoadjuvant chemotherapy (NAC) show a longer survival than those without pCR.

Purpose

To assess the accuracy of 3.0-T breast magnetic resonance imaging (MRI) in predicting pCR and to evaluate the clinicoradiologic factors affecting the diagnostic accuracy of 3.0-T breast MRI in TNBC patients treated with anthracycline and taxane (ACD).

Material and Methods

This retrospective study was approved by the institutional review board; patient consent was not required. Between 2009 and 2012, 35 TNBC patients with 3.0-T breast MRI prior to (n = 26) or after (n = 35) NAC were included. MRI findings were reviewed according to pCR to chemotherapy. The diagnostic accuracy of 3.0-T breast MRI for predicting pCR and the clinicoradiological factors affecting MRI accuracy and response to NAC were analyzed.

Results

3.0-T MRI following NAC with ACD accurately predicted pCR in 91.4% of TNBC patients. The residual tumor size between pathology and 3.0-T MRI in non-pCR cases showed a higher correlation in the Ki-67-positive TNBC group (r = 0.947) than in the Ki-67 negative group (r = 0.375) with statistical trends (P = 0.069). Pre-treatment MRI in the non-pCR group compared to the pCR group showed a larger tumor size (P = 0.030) and non-mass presentation (P = 0.015).

Conclusion

3.0-T MRI in TNBC patients following NAC with ACD showed a high accuracy for predicting pCR to NAC. Ki-67 can affect the diagnostic accuracy of 3.0-T MRI for pCR to NAC with ACD in TNBC patients.

Keywords

Breast magnetic resonance imaging (MRI)neoplasms – primary staging imaging sequences

Introduction

Triple-negative breast cancer (TNBC) is a subgroup of breast cancers that do not express estrogen receptor (ER), progesterone receptor (PR), or human epidermal growth factor receptor type 2 (HER2). They account for 15–20% of newly diagnosed breast cancers (1) and are associated with aggressive histologic features, poor prognosis, unresponsiveness to endocrine therapy, and shorter survival (1). Because of the lack of targeted therapy, although TNBC tends to respond to chemotherapy, it is reported to show heterogeneous responsiveness and a higher rate of relapse and metastasis than non-TNBC (2). Recently, several studies have suggested that TNBC patients attaining pathological complete response (pCR) after neoadjuvant chemotherapy (NAC) showed a longer survival than those without pCR (3).

Since Martincich et al. reported that dynamic contrast-enhanced magnetic resonance imaging (DCE-MRI) allowed prediction of the effect of NAC in breast cancer (4), there have been several reports about DCE-MRI’s ability to predict the response to chemotherapeutics (5 –9) and the MRI characteristics of TNBC (5,8,10 –14). Some of these showed a tendency for TNBC to present with benign morphology such as smooth mass margins, persistent enhancement pattern, and high intratumoral T2 signal intensity (10,13,14). In addition, MRI was found to predict the pCR of TNBC more accurately than other subtypes of breast cancers (7 –9); however, the majority of studies used 1.5-T MRI with a heterogeneous therapeutic regimen (6 –8,10,15,16). There have been only two reports that used 3.0-T MRI (5,14), in which the number of TNBCs included was fewer than 20. Among these, one discussed the responsiveness to NAC, but the study population included only seven cases.

The purpose of this study was to assess the accuracy of 3.0-T breast MRI in predicting a pCR and to evaluate the clinicoradiological factors affecting the diagnostic accuracy of 3.0-T MRI in TNBC patients who received NAC with anthracycline and taxane.

Material and Methods

Our institutional review board approved this retrospective study, and the requirement for informed consent was waived.

Selection of patients

All patients in the study cohort were histologically confirmed to have primary breast cancer and cytologically confirmed to have axillary lymph node metastasis at initial presentation. All patients received NAC with four cycles of anthracycline/cyclophosphamide (AC × 4) followed by four cycles of docetaxel (D × 4) every 3 weeks. After completion of NAC, all patients underwent curative surgery for the breast and axillary lymph nodes.

Between August 2009 and May 2012, there were 45 consecutive patients in our hospital who received NAC with an anthracycline and taxane regimen due to locally advanced TNBC. Among them, 35 patients who were enrolled in our study population underwent breast MRI following the last cycle of NAC. Among them, 26 patients also underwent breast MRI just before starting NAC. The mean age (35 patients) was 48.5 years (age range, 27–69 years).

MRI acquisition

All of the breast MRI was performed using a 3.0-T scanner (Trio Tim; Siemens Medical Solutions, Erlangen, Germany) equipped with a dedicated four-channel breast array coil. The following images were acquired after obtaining localizer images: T1-weighted (T1W) non-fat-suppressed axial sequence (repetition time/echo time, 280/2.6 ms; flip angle, 65°; bandwidth, 543 Hz/pixel; matrix size, 512 × 512; voxel size, 0.7 × 0.7 × 3.0 mm), T1W non-fat-suppressed precontrast and 2D dynamic contrast-enhanced (DCE) axial (repetition time/echo time, 280/2.6 ms; flip angle, 65°; bandwidth 540 Hz/pixel; matrix size, 512 × 343; voxel size, 1.0 × 0.7 × 3.0 mm) with intravenous injection of 0.2 cc/kg gadolinium-diethylenetriaminepenta acetic acid (Gd-DTPA, Magnevist; Berlex Laboratories Inc., Montville, NJ, USA) in order to obtain one precontrast set and six postcontrast sets, the temporal resolution was 73 s for each frame, and finally T2-weighted (T2W) turbo spin echo axial images (repetition time/echo time, 4360/82 ms; flip angle, 150°; bandwidth, 305 Hz/pixel; matrix size, 512 × 512; voxel size, 0.7 × 0.7 × 3.0 mm). The slice thickness was 3 mm and field of view was 32–34 cm for all of the MRI sequences.

MRI interpretation

Interpretation of breast MRI images was performed by a dedicated breast-imaging radiologist (MJK) with 10 years of experience in breast imaging and 7 years of experience interpreting breast MRI.

Pre-NAC MRI

Tumor size was measured on the second post-contrast subtracted image. With DCE-MR images, the shape and margin of the lesion, the type of lesion, the enhancement pattern, and the time-intensity curve were assessed. The time-intensity curve was evaluated with automated software program (CADstream, Merge Healthcare, Chicago, IL, USA). With T2W images, the presence of intratumoral necrosis, fibrosis, perilesional edema, and the T2W image signal intensity of lesion were evaluated. For areas with no or less-enhancement in the lesion, at the second sequence of DCE-MR images, intratumoral necrosis was considered as areas with bright signal intensity and fibrosis as areas with lower signal intensity in comparison to the surrounding parenchymal signal intensity on T2W imaging (13). The extent of the lesion was also classified as single or multiple. Multiplicity included multifocality and multicentricity.

Post-NAC MRI

The identification of residual tumor and the measurement of tumor size were assessed. If an enhancing area distinct from the background parenchymal enhancement was noted, the maximal dimension was measured, suggestive of the presence of residual tumor. The absence of a distinct enhancing area was considered a complete response to chemotherapy based on the imaging study.

Immunohistochemistry and pathologic examination

The presence of hormone receptors (ER, PR), Ki-67, and HER2 oncogene expression were based on the pathologic result of an immunohistochemical assay and core-needle biopsy specimen prior to NAC. Tumors with ≥1% nuclear-stained cells were considered positive for ER and PR according to the American Society of Clinical Oncology/College of American Pathologists (ASCO/CAP) guidelines. Ki-67 staining was assessed as the percentage of nuclei showing a positive reaction. An arbitrary cut-off point of 14% was used for determining positive Ki-67 expression based on the result at the diagnosis of TNBC. HER2 immunohistochemical assay using the HercepTest TM (DAKO) was interpreted as 0, 1+, 2+, or 3+ and was defined as positive in cases with 2+ or 3+ according to the ASCO/CAP guidelines. A pCR was defined as the absence of invasive carcinoma at microscopic study. Otherwise, the pathologic size for invasive cancer was determined as the longest dimension.

Statistical analysis

Either Chi-square test or Fisher’s exact test was used for the comparison of MRI findings between pCR lesions and non-pCR lesions. The Mann-Whitney U test or t-test was used to compare continuous variables between the two groups according to the normality based on the Kolmogorov Smirnov test. Pearson’s correlation or Spearman correlation was used for comparing MR image-based residual tumor size and pathologic size in the surgical specimen according to normality. Fisher’s z transformation test was used for correlation comparison. Statistical analysis was performed using the SPSS statistical analysis software (PASW Statistics, version 18.0.0; SPSS Inc., Chicago, IL, USA), with the significance level set at a two-sided P value of 0.05 and with the trend level set at a two-sided P value of 0.20 (17).

Results

Among the 35 patients who received NAC due to locally advanced TNBC, 17 (48.6%) patients show pCR in the breast on surgical pathology (Table 1). The remaining 18 patients showed residual invasive malignancy on surgical pathology (mean size, 14 mm; range, 2–50 mm). Ki-67 data were available in 33 patients. Twenty-seven patients showed positive Ki-67, and the remaining six patients showed negative Ki-67 on the biopsy specimen before starting NAC. There was no statistically significant difference of Ki-67 positivity between pCR and non-pCR group (P = 0.186).

Table 1.

Clinicopathologic and post-chemotherapy MRI findings of 35 patients with triple negative breast cancer.

	N (n = 35)	Pathological complete response (n = 17)	Non-pathologic complete response (n = 18)	P value
Age, mean ± SD (years)	48.49 ± 10.04	47.88 ± 10.87	49.06 ± 9.47	0.735
Ki-67				0.186
positive	27	14 (93.3)*	13 (72.2)*
negative	6	1 (6.7)*	5 (27.8)*
Presence of residual lesion at MR				<0.0001
Yes	19	2 (11.8)*	17 (94.4)*
No	16	15 (88.2)*	1 (5.6)*
Median maximal dimension at MR (range)	6 (0–46)	0 (0–16)	11 (0–46)	<0.0001
Median maximal dimension at surgical pathology (range)	2 (0–50)	0 (0–0)	14 (2–50)	<0.0001

Percentage.

Post-NAC MRI vs. pCR

All 35 patients underwent breast MRI following NAC and prior to surgery. The mean time from the last breast MRI to surgery was 7.8 days (range; 2–20 days). MRI showed residual tumors in 19 patients (54.3%, mean size, 16.0 mm; range, 3–46 mm). Seventeen of these patients (89.5%) turned out to have residual tumors at surgery. MRI depicted no residual tumors in the remaining 16 of 35 patients (45.7%). Among them, only one patient (6.3%) had an invasive residual tumor of 5 mm in size on surgical pathology, and 15 patients (93.7%) showed pCR. Overall, MR imaging accurately predicted pCR in 91.4% of patients (32 of 35), with three incorrect predictions (2 false negatives and 1 false positive for pCR). The residual tumor size between MRI and pathologic examination at surgery showed a high correlation overall (r = 0.912, P < 0.001, Fig. 1). A higher correlation was shown in tumors with positive Ki-67 (r = 0.936, P < 0.001) than in tumors with negative Ki-67 (r = 0.705, P = 0.118) with statistical trends (P = 0.177). Additionally, the statistically significant difference in regard to the positivity of Ki-67 was exaggerated when lesions achieving pCR at surgical pathology were excluded (r = 0.947, P < 0.001 with positive Ki-67; r = 0.375, P = 0.534 with negative Ki-67; difference of correlation, P = 0.069). All three patients with an incorrect prediction on MRI for pCR had tumors with positive Ki-67 without statistical significance (P = 0.226).

Fig. 1.

Axial maximal intensity projection (a, c) and original subtraction MR image (b, d) in a 45-year-old woman with Ki-67-positive triple-negative breast cancer in left breast. Initial MR images (a, b) obtained before NAC show about 2.1-cm mass with rim enhancement. On MR images obtained after NAC showed approximately 1.4-cm residual. Final pathology at the surgery showed 1.4-cm invasive residual cancer.

Pre-NAC MRI vs. pCR

Twenty-six of the 35 patients underwent breast MRI prior to NAC (Table 2). Among them, 13 patients (50.0%) showed pCR to chemotherapy, and the remaining 13 (50.0%) showed residual invasive cancer at surgery. For analysis of the initial MRI, all 13 patients with pCR showed a mass presentation (100.0%), while seven patients (53.8%) with non-pCR showed a mass presentation and six (46.2%, P = 0.015) showed a non-mass presentation. Although the median tumor size showed no statistically significant difference, large tumors (larger than 5 cm) were more frequent in the non-pCR tumor group (P = 0.030). There were no other statistically significant factors related to pCR in our study.

Table 2.

Pretreatment MRI findings and clinicopathologic factors of 26 patients with triple negative breast cancers.

	Total	Pathological complete response	Non-pathologic complete response	P value
	(n = 26)	(n = 13)	(n = 13)	P value
Age, mean ± SD (years)	46.615 ± 11.385	50.308 ± 10.291	42.923 ± 11.604	0.099
Tumor size, mean ± SD (mm)	36.077 ± 18.119	30.692 ± 10.283	41.462 ± 22.703	0.138
Tumor size (mm)				0.030
<20	4	2 (15.4)	2 (15.4)
20–50	17	11 (84.6)	6 (46.2)
>50	5	0 (0.0)	5 (38.5)
Presentation				0.015
Mass	20	13 (100.0)	7 (53.9)
Non-mass	6	0 (0.0)	6 (46.2)
Shape				0.291
Round	12	6 (46.2)	6 (46.2)
Oval	5	4 (30.8)	1 (7.7)
Lobular	5	1 (7.7)	4 (30.8)
Irregular	4	2 (15.4)	2 (15.4)
Margin				0.116
Smooth	12	8 (61.5)	4 (30.8)
Irregular	9	4 (30.8)	6 (46.2)
Spiculated	4	1 (7.7)	3 (23.1)
Enhancement				0.584
Rim	13	8 (61.5)	5 (38.5)
Homogeneous	9	4 (30.8)	5 (38.5)
Heterogeneous	4	1 (7.7)	3 (23.0)
Time-intensity curve				>0.999
Washout	15	7 (53.8)	8 (61.5)
Plateau	11	6 (46.2)	5 (38.5)
Persistent	0	0 (0.0)	0 (0.0)
Intra-tumoral necrosis				>0.999
Yes	9	4 (30.8)	5 (38.5)
No	17	9 (69.2)	8 (61.5)
Fibrosis				0.645
Yes	6	2 (15.4)	4 (30.8)
No	20	11 (84.6)	9 (69.2)
T2 SI				0.827
Low	2	1 (7.7)	1 (7.7)
Iso	16	7 (53.9)	9 (69.2)
High	8	5 (38.5)	3 (23.1)
Perilesional edema				0.420
Yes	16	7 (53.9)	9 (69.2)
No	10	6 (46.1)	4 (30.8)
Multiplicity				0.107
Single	16	10 (76.9)	6 (46.1)
Multiple	10	3 (23.1)	7 (53.9)
Ki-67*				0.220
>14	22	12 (100.0)	10 (76.9)
≤14	3	0.0	3 (23.1)

Numbers in parentheses are percentages.

Data were available in 25 patients.

Discussion

NAC is administered to patients with breast cancer with the aim of inducing tumor regression before therapeutic surgery, which can render the tumors operable or facilitate improved outcomes with breast conserving surgery. Moreover, it has been reported that the prognosis will be favorable if a pCR after NAC is achieved (18). Thus, the clinical impact of accurately predicting responsiveness to NAC cannot be overestimated. Dynamic contrast-enhanced MRI is well reported as the most promising method for the prediction of pCR in several imaging studies (19).

Several potential factors affecting the diagnostic accuracy of DCE-MRI in predicting responsiveness have been suggested. First, the accuracy of DCE-MRI can depend on the subtype of breast cancer defined by receptor status (7 –9). Recently, Moon et al. reported that the accuracy of MRI in predicting residual tumor extent was the lowest in ER-positive tumors treated with NAC (9). TNBC has been reported as one of the subtypes showing a good accuracy in predicting responsiveness with MRI (7,9,20). Moreover, TNBC tends to show a mass presentation (10), and in breast cancers presenting with a mass type, MRI is reported to predict the residual tumor size better than in cancers presenting as a non-mass type (5). Therefore, MRI can be expected to accurately predict pCR and show a higher size correlation between MR images and pathology in TNBC than in breast cancers overall or other subtypes of breast cancer, which is consistent with several previous studies (Table 3) (8,9,20). In our study, breast cancer presenting as a non-mass type showed statistically significant non-complete responsiveness.

Table 3.

Review of the literature for DCE-MRI accuracy in predicting residual tumor size with neoadjuvant chemotherapy for TNBC.

Author	Publication year	Patient no.	TNBC patient no.	Tesla of MRI	Timing of MRI	Regimen (n)	Suggested factors affecting the accuracy of MRI	The size correlation between MR vs. pathology (overall)	The size correlation between MR vs. pathology (TNBC)	Sensitivity* of MRI for pCR	Specificity^† of MRI for pCR
Chen	2007	15	15	1.5	Before/ during NAC/ after NAC	4AC + 3 P Cb	N-D	NA	0.998	88.9% (8/9)	83.3% (5/6)
Moon	2009	258	53	1.5	After NAC	A or T	Age, HER negativity	0.584	0.781	37.9%^‡ (11/29)	96.4%^‡
Straver	2010	208	47	1.5/3	Before and after NAC	6D × C, 6D × 6 CpD 3AD +3 CpD	Type of lesion at baseline Pattern of decline Histologic subtype	N-D	N-D	66.7%^‡ (28/42)	78.3%^‡ (130/166)^‡
Nakahara	2011	83	18	1.5	Before, during, and after NAC	CE (8) CE +T (3) DEC (3) Other (4)	–	0.21^§	0.92	7 (pCR)	11 (non-pCR)
De Los Santos	2011	81	25	1.5	Before and after NAC	4D×+ 4 P+ 4 C	Neither chemo regimen nor subtypes	N-D	N-D	50.0% (4/8)	100.0% (17/17)
McGuire	2011	592	44	0.5 or 1.5	Before and after NAC	N-D	Subtype as defined by receptor status	N-D	N-D	20 (pCR) 78%^‡ (40/51)	24 (non-pCR)
Chen	2011	50	9	3	Before and after NAC	D × C + T (16) PCa (34) +Bevacizumab	Ki-67	0.30**	N-D	100.0% (3/3) 93%^‡	66.7% (4/6) 81%^‡
Moon	2013	463	114	1.5	After NAC	A (24) A + T (383)	Age, initial tumor size, ER positivity	0.644	0.754	N-D	N-D
This study	2014	35	35	3	Before and after NAC	4AC + 4D	Ki-67	NA	0.912	88.2% (15/17)	94.4% (17/18)

Sensitivity was estimated as the proportion of those with pCR who were determined to have complete response by MRI.

†

Specificity was estimated as the proportion of those without pCR who did not have a complete response (CR) as determined by MRI.

‡

The percent of sensitivity or specificity based on overall breast cancer.

Not statistically significant.

0.76 when four cases with non-mass presentation were excluded.

A, anthracycline; C, cyclophosphamide; Cb, carboplatin; Cp, capecitabine; D, docetaxel; Dx, doxorubicin; E, epirubicin; NA, not applicable; NAC, neoadjuvant chemotherapy; N-D, not documented; P, paclitaxel; pCR, pathologic complete response; T, taxane.

Second, the diagnostic accuracy of MRI can be affected by the NAC regimen. For example, anti-angiogenic drugs or the anti-vascular effects of taxane-containing drugs can lead to an underestimation of the residual tumor size (9,21). Denis et al. showed that MRI frequently underestimated residual tumor sites in treatments, including taxane, because the enhancement on contrast-enhanced MRI seems to be less prominent with the anti-vascular effects of these drugs (21). In this study anthracycline and taxane, nevertheless, accurate prediction of pCR after NAC was achieved using 3.0-T MRI in the present study, which was comparable to previous reports (Table 3).

Another factor that may affect the diagnostic accuracy of DCE-MRI in predicting responsiveness is that, until now, the majority of studies of MRI for the prediction of response to NAC used a 1.5-T MRI scanner. Although there has recently been increasing clinical use of 3.0-T MRI imaging systems, in a meta-analysis of 35 studies on MRI with NAC, Lobbes et al. reported only three studies performed on a 3.0-T MRI scanner and four on both 1.5 - and 3.0-T scanners. The remaining 28 studies were performed on 1.5-T MRI scanners (22). Among the three studies using a 3-T MRI scanner, two included just one and seven TNBC cases, respectively. The study with one TNBC was not included in our review. Table 3 shows the review of studies on MRI with TNBC after NAC (5,6,8,9,15,20,23). There was only a single study other than ours performed on a 3.0-T MRI scanner and one other performed on either a 1.5 - or 3-T scanner (5). The study using a 3.0-T MRI scanner alone included the above-mentioned seven cases of TNBC. Compared with the 1.5-T MRI scanner, the 3.0-T MRI system is known to provide a higher signal-to-noise ratio and spatial resolution, which may be helpful in improving the diagnostic accuracy for small lesions (24). In this study, the correlation coefficient of tumor size between MRI and pathology was 0.912. Although it is within the overall range (0.754–0.998) of the previous literature, it is the second highest value among them. This result may be due to the high spatial resolution of the 3.0-T MRI system that depicts small residual tumor foci well. However, Chen et al. found that the 3.0-T MRI system still had the same limitation for detecting small tumor foci and scattered tumor cell clusters, which might be the cause of false negative diagnoses of MRI for pCR (5). In this study, there were also two false negative cases among the 17 for predicting complete response with MRI, similar to the results for TNBC with the 1.5-T MRI system by Chen et al. (one of eight TNBC) (15). However, because the evaluation by MRI in breast cancer patients receiving NAC is used not only for pCR prediction, but also for the accurate assessment of residual disease for breast conservation surgery, the 3.0-T MRI system may support the decision for breast conservation surgery with better accuracy, which was not evaluated in this study (25).

The MRI characteristics of TNBC tended to show a mass presentation, smooth margins, rim enhancement, intratumoral necrosis, peritumoral edema, and were less likely to show fibrosis, which is similar to previous literature (10,14); however, there was no specific finding that revealed a statistically significant difference between pCR and non-pCR TNBC except for mass presentation. TNBC with mass presentation and tumor size less than 5 cm at pre-NAC MRI tended to show a pCR to NAC in this study, which is consistent with a previous study (23). Breast cancers with non-mass presentation on MRI were likely to have less of a response to NAC in this study. Such a presentation was also reported to lower the diagnostic accuracy of MR in predicting responsiveness (23).

Furthermore, there are other factors suggested to affect the accuracy of MRI, including age, the pattern of decline after NAC, histologic subtype, Ki-67, and other subtypes as defined by receptor status, which are under debate in other studies (Table 3). In this study, according to Ki-67 level, the size correlation between MRI and pathology showed a statistically marginal difference, which validates the results by Chen et al. with a larger study population of TNBC. In their study, only seven TNBC cases were included in the analysis (5). Even though the residual size of TNBC is well-known to be accurately predicted by MRI, our results suggest that the residual size may not be well predicted by MRI in TNBC cases with low Ki-67 (P = 0.534). However, the small number of Ki-67 negative cases limits any solid conclusion and further study with a large study population is needed. Ki-67 is one of the proliferation markers and has been reported to be an independent prognostic and predictive marker in breast cancer patients. We used pretreatment Ki-67 in this study, which is known to be associated with more aggressive clinical features despite a higher pCR rate (26). In the interpretation of responsiveness to chemotherapy, physicians should take the level of Ki-67 into account even when considering TNBC.

There are several limitations in this study. First, the number of subjects in the study population was not sufficient to generalize our results. TNBC accounts for 15–20% of overall breast cancers diagnosed. Although we did not present the percentage of TNBC in patients with NAC, the small sample size would be a limitation. In our review of the literature, and among the patients with TNBC, only 74% of the included patients (26 of 35) underwent MRI prior to NAC. Another limitation is that, as mentioned above, there are several possible factors affecting the diagnostic accuracy of DCE-MRI; however, we did not compare the groups with different conditions. Thus, we could not evaluate the relative diagnostic accuracy in predicting a pCR to chemotherapy of TNBC in comparison to other subtypes of breast cancer, other MRI protocols, or other chemotherapeutic regimens. Finally, the radiologist who analyzed the MRI findings knew that all of the included cancers were TNBC but did not know the final pathologic outcome, which could affect the radiologist’s description of the MRI findings or responsiveness.

In conclusion, 3.0-T MRI in TNBC patients following NAC with anthracycline and taxane showed a high accuracy for predicting pCR to chemotherapy, but TNBCs with mass presentation and tumor size less than 5 cm at pre-treatment MRI tended to show a pCR to NAC. In addition, Ki-67 can affect the diagnostic accuracy of 3.0-T MRI for pCR to anthracycline and taxane.

Footnotes

Conflict of interest

None declared.

Funding

This research received no specific grant from any funding agency in the public, commercial, or not-for-profit sectors.

References

Bauer

Brown

Cress

. Descriptive analysis of estrogen receptor (ER)-negative, progesterone receptor (PR)-negative, and HER2-negative invasive breast cancer, the so-called triple-negative phenotype: a population-based study from the California cancer Registry. Cancer 2007; 109: 1721–1728.

Liedtke

Mazouni

Hess

. Response to neoadjuvant therapy and long-term survival in patients with triple-negative breast cancer. J Clin Oncol 2008; 26: 1275–1281.

von Minckwitz

Untch

Blohmer

. Definition and impact of pathologic complete response on prognosis after neoadjuvant chemotherapy in various intrinsic breast cancer subtypes. J Clin Oncol 2012; 30: 1796–1804.

Martincich

Montemurro

De Rosa

. Monitoring response to primary chemotherapy in breast cancer using dynamic contrast-enhanced magnetic resonance imaging. Breast Cancer Res Treat 2004; 83: 67–76.

Chen

Bahri

Mehta

. Breast cancer: evaluation of response to neoadjuvant chemotherapy with 3.0-T MR imaging. Radiology 2011; 261: 735–743.

De Los Santos

Bernreuter

Keene

. Accuracy of breast magnetic resonance imaging in predicting pathologic response in patients treated with neoadjuvant chemotherapy. Clin Breast Cancer 2011; 11: 312–319.

McGuire

Toro-Burguete

Dang

. MRI staging after neoadjuvant chemotherapy for breast cancer: does tumor biology affect accuracy? Ann Surg Oncol 2011; 18: 3149–3154.

Nakahara

Yasuda

Machida

. MR and US imaging for breast cancer patients who underwent conservation surgery after neoadjuvant chemotherapy: comparison of triple negative breast cancer and other intrinsic subtypes. Breast Cancer 2011; 18: 152–160.

Moon

Han

Ahn

. Breast Cancer Molecular Phenotype and the Use of HER2-Targeted Agents Influence the Accuracy of Breast MRI After Neoadjuvant Chemotherapy. Ann Surg 2013; 257: 133–137.

10.

Chen

Agrawal

Feig

. Triple-negative breast cancer: MRI features in 29 patients. Ann Oncol 2007; 18: 2042–2043.

11.

Koo

Cho

Song

. Correlation of perfusion parameters on dynamic contrast-enhanced MRI with prognostic factors and subtypes of breast cancers. J Magn Reson Imaging 2012; 36: 145–151.

12.

Padhani

Taylor

. Vascular characterisation of triple negative breast carcinomas using dynamic MRI. Eur Radiol 2011; 21: 1364–1373.

13.

Uematsu

Kasami

Yuen

. Triple-negative breast cancer: correlation between MR imaging and pathologic findings. Radiology 2009; 250: 638–647.

14.

Youk

Son

Chung

. Triple-negative invasive breast cancer on dynamic contrast-enhanced and diffusion-weighted MR imaging: comparison with other breast cancer subtypes. Eur Radiol 2012; 22: 1724–1734.

15.

Chen

Mehta

Carpenter

. Magnetic resonance imaging in predicting pathological response of triple negative breast cancer following neoadjuvant chemotherapy. J Clin Oncol 2007; 25: 5667–5669.

16.

Costantini

Belli

Distefano

. Magnetic resonance imaging features in triple-negative breast cancer: comparison with luminal and HER2-overexpressing tumors. Clin Breast Cancer 2012; 12: 331–339.

17.

Diem

Lentner

. Scientific tables , 7th (enl.) edn. Basel: JR Geigy, 1970.

18.

Symmans

Peintinger

Hatzis

. Measurement of residual breast cancer burden to predict survival after neoadjuvant chemotherapy. J Clin Oncol 2007; 25: 4414–4422.

19.

de Ronde

Hannemann

Halfwerk

. Concordance of clinical and molecular breast cancer subtyping in the context of preoperative chemotherapy response. Breast Cancer Res Treat 2010; 119: 119–126.

20.

Moon

Han

Lee

. Age and HER2 expression status affect MRI accuracy in predicting residual tumor extent after neo-adjuvant systemic treatment. Ann Oncol 2009; 20: 636–641.

21.

Denis

Desbiez-Bourcier

Chapiron

. Contrast enhanced magnetic resonance imaging underestimates residual disease following neoadjuvant docetaxel based chemotherapy for breast cancer. Eur J Surg Oncol 2004; 30: 1069–1076.

22.

Lobbes

Prevos

Smidt

. The role of magnetic resonance imaging in assessing residual disease and pathologic complete response in breast cancer patients receiving neoadjuvant chemotherapy: a systematic review. Insights Imaging 2013; 4: 163–175.

23.

Straver

Loo

Rutgers

. MRI-model to guide the surgical treatment in breast cancer patients after neoadjuvant chemotherapy. Ann Surg 2010; 251: 701–707.

24.

Kuhl

Jost

Morakkabati

. Contrast-enhanced MR imaging of the breast at 3.0 and 1.5 T in the same patients: initial experience. Radiology 2006; 239: 666–676.

25.

Lobbes

. Treatment response evaluation by MRI in breast cancer patients receiving neoadjuvant chemotherapy: there is more than just pathologic complete response prediction. Breast Cancer Res Treat 2012; 136: 313–314.

26.

Keam

Lee

. Ki-67 can be used for further classification of triple negative breast cancer into two subtypes with different response and prognosis. Breast Cancer Res 2011; 13: R22–R22.