ORIGINAL RESEARCH
MRI Phenotype of Breast Cancer: Kinetic Assessment for Molecular Subtypes Eric Blaschke, MD* and Hiroyuki Abe, MD Purpose: To evaluate the dynamic contrast-enhanced magnetic resonance imaging (MRI) kinetic characteristics of newly diagnosed breast cancer molecular subtypes. Materials and Methods: Breast MRI examinations of 112 patients with newly diagnosed breast cancer were reviewed. Cases of newly diagnosed invasive ductal carcinoma were sorted by molecular subtype (28 TN, 11 HER2 1, 73 Lum A/ B) and MRI field strength, with lesion segmentation and kinetic analyses performed on a dedicated workstation. For kinetic assessment, 50% and 100% thresholds were employed for display of medium and rapid uptake. Kinetic profiles in terms of percent volume for six kinetic types (medium-persistent, medium-plateau, medium-washout, fast-persistent, fast-plateau, fast-washout) relative to the whole volume of the lesion were obtained. Statistical analysis of the kinetic profiles was performed using Welch’s t-test. Results: Percent volume of HER2-positive lesions with >100% uptake at early phase on 3T strength MRI exams was significantly greater compared with luminal A/B (93.8 6 0.92 vs. 77.3 6 7.2; P < 0.01) and triple negative (93.8 6 0.92 vs. 81.3 6 8.2; P < 0.05) subtypes. The >50% early phase uptake for HER21 lesions was also higher than Lum A/B (99.1 6 0.73 vs. 93.6 6 3.05; P < 0.01) at 3T. In the 1.5T subgroup the percent volume of HER21 tumors with >50% and >100% early phase uptake trended higher than Lum A/B lesions without reaching significance. Conclusion: The percent volume of HER2-positive tumors demonstrating rapid early contrast uptake is significantly increased compared to other molecular subtypes. J. MAGN. RESON. IMAGING 2015;42:920–924.
M
agnetic resonance imaging (MRI) is widely used in the preoperative evaluation of newly diagnosed breast cancer. Dynamic contrast-enhanced (DCE) MRI has excellent sensitivity, moderate specificity, and can detect occult disease more accurately than mammography.1–4 Quantitative breast MR analysis has shown promise in complementing traditional MR techniques, with multiple emerging applications including identification of cancer subtype, invasiveness, grade, lymph node status, and predicting tumor response to therapy.5–10 Molecular subclassification of breast cancer forecasts clinical course and is associated with specific patterns of gene expression.11–14 In this study we analyzed quantitative MR kinetic profiles of newly diagnosed invasive ductal carcinoma (IDC) sorted by molecular subtype. Our goal was to evaluate the DCE MR kinetic characteristics of newly diagnosed IDC by quantifying the percent volume of tumor with a given uptake rate and washout type, which we predicted would be associated with specific molecular subtypes.
MATERIALS AND METHODS Patients Patients with newly diagnosed breast cancer who underwent DCEMRI examinations from 2011 through 2014 at our institution were selected as part of this retrospective Institutional Review Board-approved study. Informed consent was waived. In all, 112 cases of patients found to have newly diagnosed IDC were reviewed and sorted by molecular subtype (28 triple negative, 11 HER21, 73 Luminal A/B) and by MRI Tesla strength (51 Luminal A/B, 24 TN, 6 HER21 at 3T and 22 Luminal A/B, 4 TN, 5 HER21 at 1.5T). Patient age and tumor size were similar between subtypes (Table 1). HER21 and triple negative disease tended to be of higher grade (Table 1). The classification of molecular subtypes was based on the results of immunohistochemical (IHC) staining of estrogen receptor (ER), progesterone receptor (PR), human epidermal growth factor receptor 2 (HER2), and the value of HER2/chromosome 17 fluorescence in situ hybridization (FISH).15,16 For ER and PR, it was considered positive when the expression was 10% or greater. For the IHC staining of HER2, which was scored from 0 to 31,
View this article online at wileyonlinelibrary.com. DOI: 10.1002/jmri.24884 Received Sep 12, 2014, Accepted for publication Feb 20, 2015. *Address reprint requests to: E.B., University of Chicago Medicine, Radiology, 5841 S. Maryland Ave., MC 2026, Chicago, IL, 60637. E-mail: EMBlaschke@ gmail.com From the University of Chicago Medicine, Radiology, Chicago, Illinois, USA
C 2015 Wiley Periodicals, Inc. 920 V
Blaschke and Abe: MRI Phenotype of Breast Cancer
TABLE 1. Range of Tumor Diameter, Patient Age, and Distribution by Grade and Tesla Strength for Luminal A/B, Her21 and Triple Negative Tumor Subtypes
Subtype
Grade 1 (3/1.5 T)
Grade 2 (3/1.5 T)
Grade 3 (3/1.5 T)
Age (mean)
Tumor diameter (mean 6 SD)
Lum A/B
10/4
22/12
19/6
35–77 (53.8 yrs.)
0.4–4.5 (1.50 6 0.75 cm)
HER2 1
0/0
1/1
5/4
23–67 (51.5 yrs.)
1.0–3.8 (1.92 6 0.93 cm)
Triple Neg
0/0
4/0
16/8
36–77 (52.1 yrs.)
0.8–4.5 (2.68 6 0.82 cm)
31 was considered positive and 0 and 11 were considered negative. If it was scored 21, greater than 2.2 in the value of FISH was considered positive. Classification for four molecular subtypes was performed as follows: Luminal A: ER-positive and/or PRpositive and HER2-negative, Luminal B: ER-positive and/or PRpositive and HER2-positive, HER2 type: ER-negative, PgR-negative and HER2-positive, triple negative: negative for ER, PgR and HER2.
MRI Technique All patients underwent MRI examination using a dedicated breast coil in prone position, on either a 1.5T or 3T Philips Healthcare Achieva system (Best, Netherlands). The 1.5T imaging protocol used axial fast spin-echo T2-weighted images (TR/ TE 2000/326, echo train length 15, slice thickness 2 mm, matrix 480 3 480, field of view [FOV] 360 mm) as well as DCE images obtained using one pre- and six postcontrastenhanced axial T1-weighted fat suppressed 3D gradient echo sequences (5.5/2.7, flip angle 10 , matrix 480 3 480, slice thickness 2 mm, 200 slices, FOV 360 mm, acquisition time 65 sec). 3T MR examinations obtained axial fast spin-echo T2weighted images (TR/TE 2000/270, echo train length 15, slice thickness 1.6 mm, matrix 448 3 448, FOV 340 mm) as well as DCE pre- and five postcontrast-enhanced axial T1-weighted images with fat-suppressed 3D gradient-echo sequences (5.0/2.5, flip angle 10 , matrix 448 3 448, slice thickness 1.6 mm, 250 slices, FOV 340 mm, acquisition time 75 sec). The second postcontrast phase was used for early phase uptake, starting 65 seconds after contrast injection for 1.5T and 75 seconds at 3T. The fifth postcontrast phase was used for the delayed phase, starting at 325 seconds after contrast injection for 1.5T and 375 seconds at 3T. Contrast material (0.1 mmol/kg of gadodiamide; Omniscan, GE Healthcare, Milwaukee, WI) was injected intravenously followed by a 20-mL saline flush at 2 mL/s. Dynamic postcontrast imaging was begun immediately following contrast injection.
Postprocessing MRI examinations were sent to a dedicated breast MR workstation employing QuantX software (Quantitative Insights, Chicago, IL), which allows color overly representing changes in signal intensity over time as well as 3D volume analysis by segmenting continuous enhancing voxels. A single radiologist with 4 years of experience reviewed and manually adjusted any segmentation errors and assessed the known cancer’s kinetic profile. October 2015
Using 50% and 100% early enhancement thresholds and standard delayed postcontrast series, kinetic profiles in terms of percent volume for six kinetic types (medium-persistent, medium-plateau, medium-washout, fast-persistent, fast-plateau, fast-washout) were obtained relative to the whole lesion volume for each case (Figs. 1, 2). The additional three standard BIRADS categories (slow-persistent, slow-plateau, slow-persistent) were grouped with medium early enhancing types for ease of analysis.17 In order to classify whether a lesion demonstrated significant enhancement, pre- and postcontrast pixel values were compared using 50% and 100% thresholds for medium and fast early uptake. On delayed postcontrast series pixel values decreased by >10% were considered washout and were color-coded red. Pixel values increasing by >10% were classified as persistent enhancement and color-coded blue. Pixel values with less than 10% change in value were classified as plateau and color-coded yellow.
Statistical Analysis Statistical analysis of the kinetic profiles was performed using Welch’s t-test. P < 0.05 were considered significant. Because there were several pairwise tests in these analyses, significant values were adjusted by the Bonferroni method.
RESULTS Analysis of 3T strength MRI exams demonstrated that the percent volume of HER21 lesions demonstrating >100% contrast uptake was significantly greater compared with both Lum A/B lesions (93.8 6 0.92 vs. 77.3 6 7.2; P < 0.01) and TN tumors (93.8 6 0.92 vs. 81.3 6 8.2; P < 0.05) (Table 2). At 3T strength the percent volume of HER21 lesions demonstrating >50% uptake was also significantly greater than Lum A/B lesions (99.1 6 0.73 vs. 93.6 6 3.05; P < 0.01) and trended higher than TN tumors without reaching significance (99.1 6 0.73 vs. 94.4 6 4.12; P > 0.05). Figure 1 demonstrates a representative HER21 cancer compared with a Luminal A cancer at 3T, both appearing similar with washout kinetics but demonstrating markedly different early kinetic enhancement profiles. There was no significant difference in the percent volume of TN and Lum A/B tumors demonstrating >100% or >50% early contrast uptake at 3T (Table 2). Figure 2 shows a 921
Journal of Magnetic Resonance Imaging
FIGURE 1: Dynamic postcontrast MR images of representative HER2-positive (left) and Luminal A (right) breast cancer subtypes showing lesion segmentation and kinetic profiles in terms of percent volume (medium-persistent, medium-plateau, medium-washout, rapid-persistent, rapid-plateau, rapid-washout). Early phase uptake is greater for the HER2-positive lesion.
representative triple negative tumor compared with a Luminal B type at 3T, with both demonstrating washout kinetics as well as similar kinetic profiles.
At 1.5T the percent volume of HER21 lesions with >100% early phase contrast uptake trended higher compared with Lum A/B lesions (74.7 6 11.5 vs. 50.9 6 13.2;
FIGURE 2: Dynamic postcontrast MR images of representative triple negative (left) and Luminal B (right) breast cancer subtypes showing kinetic curve distribution of segmented lesions. There was no significant difference between triple negative and other subtypes.
922
Volume 42, No. 4
Blaschke and Abe: MRI Phenotype of Breast Cancer
TABLE 2. Percent Volume of Tumor Demonstrating 50% and 100% Rapid Uptake Ratios by Molecular Subtype
Subtype
>50% uptake ratio (3T/1.5T)
>100% uptake ratio (3T/1.5T)
HER2 1
99.1 6 0.73 / 94.3 6 5.35
93.8 6 0.92 / 74.7 6 11.5
Luminal A/B
93.6 6 3.05 / 86.5 6 6.88
77.3 6 7.2 / 50.9 6 13.2
Triple negative
94.4 6 4.12 / 97.2 6 2.15
81.3 6 8.2 / 78.9 6 11.0
P > 0.05) and was not significantly different from TN tumors (74.7 6 11.5 vs. 78.96 11.0; P > 0.05) (Table 2). At 1.5T the percent volume of HER21 tumors with >50% early phase contrast uptake was also not significantly different from Lum A/B lesions (94.3 6 5.35 vs. 86.5 6 6.88; P > 0.05) or TN tumors (94.3 6 5.35 vs. 97.2 6 2.15; P > 0.05) (Table 2). Table 3 demonstrates the average percent volume of each tumor subtype associated with washout, persistent, and plateau kinetics. No significant difference was observed in BIRADS assessment of kinetic patterns between tumor subtypes, as all tumor subtypes demonstrated a heterogeneous kinetic pattern consisting of persistent, plateau, and washout kinetics. The percent volume of tumors with persistent, plateau, and washout kinetics overlapped greatly, without significant difference between subtypes.
DISCUSSION In this study we found that the percent volume of HER21 tumors with >100% early contrast uptake was significantly higher than Luminal A/B and TN cancers at 3T and trended higher than Lum A/B subtypes at 1.5T. Using a >50% early contrast uptake threshold was less useful, as HER21 lesions differed significantly only compared with Lum A/B tumors at 3T, and did not significantly differ from any subtype at 1.5T. These findings contribute to a range of recently elucidated quantitative markers that together define imaging phenotypes that have shown promise in the identification of additional cancer subtypes, lesion invasiveness, grade, and lymph node status, with associated treatment and prognostic implications.5–11,18–20 Classification of breast cancer by molecular subtype furthermore has
clinical value, as current treatment guidelines are determined by tumor subtype.16 HER2 expression by cancers in particular needs to be determined, as anti-HER2 therapy has well documented efficacy. Quantitative imaging markers reflect underlying physical characteristics of tumors and as such provide a glimpse into tumor biology and molecular makeup. In the case of HER21 lesions, HER2 expression is known to correlate with the overexpression of vascular endothelial growth factor (VEGF), an avid stimulator of angiogenesis.21 The marked early enhancement we observed is consistent with increased tumor vascularity. Variations in such imaging markers might be predicted to reflect changes in associated biophysical properties or related gene expression. Whole volume evaluation of tumors may also identify regions of tumor heterogeneity, which could portend differing treatment response or tumor behavior. In the case of triple negative tumors, our results show early uptake profiles overlapping with both HER21 and Luminal A/B subtypes, perhaps reflective of the underlying heterogeneity observed with this subtype. Major limitations of this study include that it was a retrospective, single-institution study with potential for population bias. In addition, lesion segmentation was performed by a single radiologist, introducing potential observer bias. This study also includes a relatively small number of HER21 cancers, particularly as examined at 1.5T. As exams were performed on two MR machines of different field strengths, subgroup analysis at 3T and 1.5T was performed, but the small number of HER21 cases examined at 1.5T likely contributed to our findings not reaching significance at this field strength. Indeed, the small number of HER21 cases overall requires further validation of our findings with
TABLE 3. Percent Volume of Tumor Demonstrating Washout, Plateau, or Persistent Contrast Enhancement Kinetics
Enhancement kinetics (% volume)
Lum A/B
HER2 1
Triple negative
Washout (mean 6SD)
25.4 6 4.48
31.3 6 9.99
25.6 6 8.08
Plateau (mean 6SD)
24.7 6 3.19
28.7 6 5.05
26.6 6 7.99
Persistent (mean 6SD)
19.2 6 3.01
25.2 6 6.55
27.46 7.14
October 2015
923
Journal of Magnetic Resonance Imaging
both a greater number of cases and with different MRI protocols. The classification of tumors by molecular subtype was done by conventional methods and was limited by the lack of availability of the Ki-67 marker status of our cases, precluding independent analysis of Luminal A and B subtypes. In conclusion, our study suggests that the percent volume of HER2-positive tumors demonstrating rapid early contrast uptake is significantly increased compared to other molecular subtypes. Continued validation of MRI phenotypes like that which we identify for HER21 tumors may aid in diagnostic and treatment planning, and moreover yield independent markers for prognosis and treatment response, complementing traditional histopathologic and molecular analysis.
CONFLICT OF INTEREST
8.
Abramson RG, Arlinghaus LR, Weis JA, et al. Current and emerging quantitative magnetic resonance imaging methods for assessing and predicting the response of breast cancer to neoadjuvant therapy. Breast Cancer (London) 2012;2012:139–154.
9.
Bhooshan N, Giger ML, Jansen SA, Li H, Lan L, Newstead GM. Cancerous breast lesions on dynamic contrast-enhanced MR images: computerized characterization for image-based prognostic markers. Radiology 2010;254:680–690.
10.
Golden DI, Lipson JA, Telli ML, Ford JM, Rubin DL. Qualitative and quantitative image-based biomarkers of therapeutic response in triple-negative breast cancer. AMIA Summits Transl Sci Proc 2013; 2013:62.
11.
Ashraf AB, Daye D, Gavenonis S, et al. Identification of intrinsic imaging phenotypes for breast cancer tumors: preliminary associations with gene expression profiles. Radiology 2014:131375.
12.
Brenton JD, Carey LA, Ahmed AA, Caldas C. Molecular classification and molecular forecasting of breast cancer: ready for clinical application? J Clin Oncol 2005;23:7350–7360.
13.
Network CGA. Comprehensive molecular portraits of human breast tumours. Nature 2012;490:61–70.
14.
Sørlie T, Perou CM, Tibshirani R, et al. Gene expression patterns of breast carcinomas distinguish tumor subclasses with clinical implications. Proc Natl Acad Sci U S A 2001;98:10869–10874.
15.
Tamimi RM, Baer HJ, Marotti J, et al. Comparison of molecular phenotypes of ductal carcinoma in situ and invasive breast cancer. Breast Cancer Res 2008;10:R67.
16.
Goldhirsch A, Wood WC, Coates AS, et al. Strategies for subtypes– dealing with the diversity of breast cancer: highlights of the St. Gallen International Expert Consensus on the Primary Therapy of Early Breast Cancer 2011. Ann Oncol 2011;22:1736–1747.
17.
Morris EA, Comstock CE, Lee CH, et al. ACR BI-RADS magnetic resonance imaging. In: ACR BI-RADS Atlas, Breast Imaging Reporting and Data System. Reston, VA: American College of Radiology; 2013.
18.
Giger ML, Karssemeijer N, Schnabel JA. Breast image analysis for risk assessment, detection, diagnosis, and treatment of cancer. Annu Rev Biomed Eng 2013;15:327–357.
19.
Hsiang DJ, Yamamoto M, Mehta RS, et al. Predicting nodal status using dynamic contrast-enhanced magnetic resonance imaging in patients with locally advanced breast cancer undergoing neoadjuvant chemotherapy with and without sequential trastuzumab. Arch Surg 2007;142:855–861; discussion 860–851.
20.
Yamaguchi K, Abe H, Newstead GM, et al. Intratumoral heterogeneity of the distribution of kinetic parameters in breast cancer: comparison based on the molecular subtypes of invasive breast cancer. Breast Cancer 2014.
21.
Brown LF, Berse B, Jackman RW, et al. Expression of vascular permeability factor (vascular endothelial growth factor) and its receptors in breast cancer. Hum Pathol 1995;26:86–91.
Consultant, Seno Medical Instruments Inc.
REFERENCES 1.
Bluemke DA, Gatsonis CA, Chen MH, et al. Magnetic resonance imaging of the breast prior to biopsy. JAMA 2004;292:2735–2742.
2.
Esserman L, Hylton N, Yassa L, Barclay J, Frankel S, Sickles E. Utility of magnetic resonance imaging in the management of breast cancer: evidence for improved preoperative staging. J Clin Oncol 1999;17: 110–119.
3.
4.
Liberman L, Morris EA, Dershaw DD, Abramson AF, Tan LK. MR imaging of the ipsilateral breast in women with percutaneously proven breast cancer. AJR Am J Roentgenol 2003;180:901–910. Peters NH, Borel Rinkes IH, Zuithoff NP, Mali WP, Moons KG, Peeters PH. Meta-analysis of MR imaging in the diagnosis of breast lesions. Radiology 2008;246:116–124.
5.
El Khouli RH, Macura KJ, Jacobs MA, et al. Dynamic contrastenhanced MRI of the breast: quantitative method for kinetic curve type assessment. AJR Am J Roentgenol 2009;193:W295–300.
6.
Li SP, Padhani AR, Taylor NJ, et al. Vascular characterisation of triple negative breast carcinomas using dynamic MRI. Eur Radiol 2011;21: 1364–1373.
7.
Abramson RG, Li X, Hoyt TL, et al. Early assessment of breast cancer response to neoadjuvant chemotherapy by semi-quantitative analysis of high-temporal resolution DCE-MRI: preliminary results. Magn Reson Imaging 2013;31:1457–1464.
924
Volume 42, No. 4
MRI Phenotype of Breast Cancer: Kinetic Assessment for Molecular Subtypes Eric Blaschke, MD* and Hiroyuki Abe, MD Purpose: To evaluate the dynamic contrast-enhanced magnetic resonance imaging (MRI) kinetic characteristics of newly diagnosed breast cancer molecular subtypes. Materials and Methods: Breast MRI examinations of 112 patients with newly diagnosed breast cancer were reviewed. Cases of newly diagnosed invasive ductal carcinoma were sorted by molecular subtype (28 TN, 11 HER2 1, 73 Lum A/ B) and MRI field strength, with lesion segmentation and kinetic analyses performed on a dedicated workstation. For kinetic assessment, 50% and 100% thresholds were employed for display of medium and rapid uptake. Kinetic profiles in terms of percent volume for six kinetic types (medium-persistent, medium-plateau, medium-washout, fast-persistent, fast-plateau, fast-washout) relative to the whole volume of the lesion were obtained. Statistical analysis of the kinetic profiles was performed using Welch’s t-test. Results: Percent volume of HER2-positive lesions with >100% uptake at early phase on 3T strength MRI exams was significantly greater compared with luminal A/B (93.8 6 0.92 vs. 77.3 6 7.2; P < 0.01) and triple negative (93.8 6 0.92 vs. 81.3 6 8.2; P < 0.05) subtypes. The >50% early phase uptake for HER21 lesions was also higher than Lum A/B (99.1 6 0.73 vs. 93.6 6 3.05; P < 0.01) at 3T. In the 1.5T subgroup the percent volume of HER21 tumors with >50% and >100% early phase uptake trended higher than Lum A/B lesions without reaching significance. Conclusion: The percent volume of HER2-positive tumors demonstrating rapid early contrast uptake is significantly increased compared to other molecular subtypes. J. MAGN. RESON. IMAGING 2015;42:920–924.
M
agnetic resonance imaging (MRI) is widely used in the preoperative evaluation of newly diagnosed breast cancer. Dynamic contrast-enhanced (DCE) MRI has excellent sensitivity, moderate specificity, and can detect occult disease more accurately than mammography.1–4 Quantitative breast MR analysis has shown promise in complementing traditional MR techniques, with multiple emerging applications including identification of cancer subtype, invasiveness, grade, lymph node status, and predicting tumor response to therapy.5–10 Molecular subclassification of breast cancer forecasts clinical course and is associated with specific patterns of gene expression.11–14 In this study we analyzed quantitative MR kinetic profiles of newly diagnosed invasive ductal carcinoma (IDC) sorted by molecular subtype. Our goal was to evaluate the DCE MR kinetic characteristics of newly diagnosed IDC by quantifying the percent volume of tumor with a given uptake rate and washout type, which we predicted would be associated with specific molecular subtypes.
MATERIALS AND METHODS Patients Patients with newly diagnosed breast cancer who underwent DCEMRI examinations from 2011 through 2014 at our institution were selected as part of this retrospective Institutional Review Board-approved study. Informed consent was waived. In all, 112 cases of patients found to have newly diagnosed IDC were reviewed and sorted by molecular subtype (28 triple negative, 11 HER21, 73 Luminal A/B) and by MRI Tesla strength (51 Luminal A/B, 24 TN, 6 HER21 at 3T and 22 Luminal A/B, 4 TN, 5 HER21 at 1.5T). Patient age and tumor size were similar between subtypes (Table 1). HER21 and triple negative disease tended to be of higher grade (Table 1). The classification of molecular subtypes was based on the results of immunohistochemical (IHC) staining of estrogen receptor (ER), progesterone receptor (PR), human epidermal growth factor receptor 2 (HER2), and the value of HER2/chromosome 17 fluorescence in situ hybridization (FISH).15,16 For ER and PR, it was considered positive when the expression was 10% or greater. For the IHC staining of HER2, which was scored from 0 to 31,
View this article online at wileyonlinelibrary.com. DOI: 10.1002/jmri.24884 Received Sep 12, 2014, Accepted for publication Feb 20, 2015. *Address reprint requests to: E.B., University of Chicago Medicine, Radiology, 5841 S. Maryland Ave., MC 2026, Chicago, IL, 60637. E-mail: EMBlaschke@ gmail.com From the University of Chicago Medicine, Radiology, Chicago, Illinois, USA
C 2015 Wiley Periodicals, Inc. 920 V
Blaschke and Abe: MRI Phenotype of Breast Cancer
TABLE 1. Range of Tumor Diameter, Patient Age, and Distribution by Grade and Tesla Strength for Luminal A/B, Her21 and Triple Negative Tumor Subtypes
Subtype
Grade 1 (3/1.5 T)
Grade 2 (3/1.5 T)
Grade 3 (3/1.5 T)
Age (mean)
Tumor diameter (mean 6 SD)
Lum A/B
10/4
22/12
19/6
35–77 (53.8 yrs.)
0.4–4.5 (1.50 6 0.75 cm)
HER2 1
0/0
1/1
5/4
23–67 (51.5 yrs.)
1.0–3.8 (1.92 6 0.93 cm)
Triple Neg
0/0
4/0
16/8
36–77 (52.1 yrs.)
0.8–4.5 (2.68 6 0.82 cm)
31 was considered positive and 0 and 11 were considered negative. If it was scored 21, greater than 2.2 in the value of FISH was considered positive. Classification for four molecular subtypes was performed as follows: Luminal A: ER-positive and/or PRpositive and HER2-negative, Luminal B: ER-positive and/or PRpositive and HER2-positive, HER2 type: ER-negative, PgR-negative and HER2-positive, triple negative: negative for ER, PgR and HER2.
MRI Technique All patients underwent MRI examination using a dedicated breast coil in prone position, on either a 1.5T or 3T Philips Healthcare Achieva system (Best, Netherlands). The 1.5T imaging protocol used axial fast spin-echo T2-weighted images (TR/ TE 2000/326, echo train length 15, slice thickness 2 mm, matrix 480 3 480, field of view [FOV] 360 mm) as well as DCE images obtained using one pre- and six postcontrastenhanced axial T1-weighted fat suppressed 3D gradient echo sequences (5.5/2.7, flip angle 10 , matrix 480 3 480, slice thickness 2 mm, 200 slices, FOV 360 mm, acquisition time 65 sec). 3T MR examinations obtained axial fast spin-echo T2weighted images (TR/TE 2000/270, echo train length 15, slice thickness 1.6 mm, matrix 448 3 448, FOV 340 mm) as well as DCE pre- and five postcontrast-enhanced axial T1-weighted images with fat-suppressed 3D gradient-echo sequences (5.0/2.5, flip angle 10 , matrix 448 3 448, slice thickness 1.6 mm, 250 slices, FOV 340 mm, acquisition time 75 sec). The second postcontrast phase was used for early phase uptake, starting 65 seconds after contrast injection for 1.5T and 75 seconds at 3T. The fifth postcontrast phase was used for the delayed phase, starting at 325 seconds after contrast injection for 1.5T and 375 seconds at 3T. Contrast material (0.1 mmol/kg of gadodiamide; Omniscan, GE Healthcare, Milwaukee, WI) was injected intravenously followed by a 20-mL saline flush at 2 mL/s. Dynamic postcontrast imaging was begun immediately following contrast injection.
Postprocessing MRI examinations were sent to a dedicated breast MR workstation employing QuantX software (Quantitative Insights, Chicago, IL), which allows color overly representing changes in signal intensity over time as well as 3D volume analysis by segmenting continuous enhancing voxels. A single radiologist with 4 years of experience reviewed and manually adjusted any segmentation errors and assessed the known cancer’s kinetic profile. October 2015
Using 50% and 100% early enhancement thresholds and standard delayed postcontrast series, kinetic profiles in terms of percent volume for six kinetic types (medium-persistent, medium-plateau, medium-washout, fast-persistent, fast-plateau, fast-washout) were obtained relative to the whole lesion volume for each case (Figs. 1, 2). The additional three standard BIRADS categories (slow-persistent, slow-plateau, slow-persistent) were grouped with medium early enhancing types for ease of analysis.17 In order to classify whether a lesion demonstrated significant enhancement, pre- and postcontrast pixel values were compared using 50% and 100% thresholds for medium and fast early uptake. On delayed postcontrast series pixel values decreased by >10% were considered washout and were color-coded red. Pixel values increasing by >10% were classified as persistent enhancement and color-coded blue. Pixel values with less than 10% change in value were classified as plateau and color-coded yellow.
Statistical Analysis Statistical analysis of the kinetic profiles was performed using Welch’s t-test. P < 0.05 were considered significant. Because there were several pairwise tests in these analyses, significant values were adjusted by the Bonferroni method.
RESULTS Analysis of 3T strength MRI exams demonstrated that the percent volume of HER21 lesions demonstrating >100% contrast uptake was significantly greater compared with both Lum A/B lesions (93.8 6 0.92 vs. 77.3 6 7.2; P < 0.01) and TN tumors (93.8 6 0.92 vs. 81.3 6 8.2; P < 0.05) (Table 2). At 3T strength the percent volume of HER21 lesions demonstrating >50% uptake was also significantly greater than Lum A/B lesions (99.1 6 0.73 vs. 93.6 6 3.05; P < 0.01) and trended higher than TN tumors without reaching significance (99.1 6 0.73 vs. 94.4 6 4.12; P > 0.05). Figure 1 demonstrates a representative HER21 cancer compared with a Luminal A cancer at 3T, both appearing similar with washout kinetics but demonstrating markedly different early kinetic enhancement profiles. There was no significant difference in the percent volume of TN and Lum A/B tumors demonstrating >100% or >50% early contrast uptake at 3T (Table 2). Figure 2 shows a 921
Journal of Magnetic Resonance Imaging
FIGURE 1: Dynamic postcontrast MR images of representative HER2-positive (left) and Luminal A (right) breast cancer subtypes showing lesion segmentation and kinetic profiles in terms of percent volume (medium-persistent, medium-plateau, medium-washout, rapid-persistent, rapid-plateau, rapid-washout). Early phase uptake is greater for the HER2-positive lesion.
representative triple negative tumor compared with a Luminal B type at 3T, with both demonstrating washout kinetics as well as similar kinetic profiles.
At 1.5T the percent volume of HER21 lesions with >100% early phase contrast uptake trended higher compared with Lum A/B lesions (74.7 6 11.5 vs. 50.9 6 13.2;
FIGURE 2: Dynamic postcontrast MR images of representative triple negative (left) and Luminal B (right) breast cancer subtypes showing kinetic curve distribution of segmented lesions. There was no significant difference between triple negative and other subtypes.
922
Volume 42, No. 4
Blaschke and Abe: MRI Phenotype of Breast Cancer
TABLE 2. Percent Volume of Tumor Demonstrating 50% and 100% Rapid Uptake Ratios by Molecular Subtype
Subtype
>50% uptake ratio (3T/1.5T)
>100% uptake ratio (3T/1.5T)
HER2 1
99.1 6 0.73 / 94.3 6 5.35
93.8 6 0.92 / 74.7 6 11.5
Luminal A/B
93.6 6 3.05 / 86.5 6 6.88
77.3 6 7.2 / 50.9 6 13.2
Triple negative
94.4 6 4.12 / 97.2 6 2.15
81.3 6 8.2 / 78.9 6 11.0
P > 0.05) and was not significantly different from TN tumors (74.7 6 11.5 vs. 78.96 11.0; P > 0.05) (Table 2). At 1.5T the percent volume of HER21 tumors with >50% early phase contrast uptake was also not significantly different from Lum A/B lesions (94.3 6 5.35 vs. 86.5 6 6.88; P > 0.05) or TN tumors (94.3 6 5.35 vs. 97.2 6 2.15; P > 0.05) (Table 2). Table 3 demonstrates the average percent volume of each tumor subtype associated with washout, persistent, and plateau kinetics. No significant difference was observed in BIRADS assessment of kinetic patterns between tumor subtypes, as all tumor subtypes demonstrated a heterogeneous kinetic pattern consisting of persistent, plateau, and washout kinetics. The percent volume of tumors with persistent, plateau, and washout kinetics overlapped greatly, without significant difference between subtypes.
DISCUSSION In this study we found that the percent volume of HER21 tumors with >100% early contrast uptake was significantly higher than Luminal A/B and TN cancers at 3T and trended higher than Lum A/B subtypes at 1.5T. Using a >50% early contrast uptake threshold was less useful, as HER21 lesions differed significantly only compared with Lum A/B tumors at 3T, and did not significantly differ from any subtype at 1.5T. These findings contribute to a range of recently elucidated quantitative markers that together define imaging phenotypes that have shown promise in the identification of additional cancer subtypes, lesion invasiveness, grade, and lymph node status, with associated treatment and prognostic implications.5–11,18–20 Classification of breast cancer by molecular subtype furthermore has
clinical value, as current treatment guidelines are determined by tumor subtype.16 HER2 expression by cancers in particular needs to be determined, as anti-HER2 therapy has well documented efficacy. Quantitative imaging markers reflect underlying physical characteristics of tumors and as such provide a glimpse into tumor biology and molecular makeup. In the case of HER21 lesions, HER2 expression is known to correlate with the overexpression of vascular endothelial growth factor (VEGF), an avid stimulator of angiogenesis.21 The marked early enhancement we observed is consistent with increased tumor vascularity. Variations in such imaging markers might be predicted to reflect changes in associated biophysical properties or related gene expression. Whole volume evaluation of tumors may also identify regions of tumor heterogeneity, which could portend differing treatment response or tumor behavior. In the case of triple negative tumors, our results show early uptake profiles overlapping with both HER21 and Luminal A/B subtypes, perhaps reflective of the underlying heterogeneity observed with this subtype. Major limitations of this study include that it was a retrospective, single-institution study with potential for population bias. In addition, lesion segmentation was performed by a single radiologist, introducing potential observer bias. This study also includes a relatively small number of HER21 cancers, particularly as examined at 1.5T. As exams were performed on two MR machines of different field strengths, subgroup analysis at 3T and 1.5T was performed, but the small number of HER21 cases examined at 1.5T likely contributed to our findings not reaching significance at this field strength. Indeed, the small number of HER21 cases overall requires further validation of our findings with
TABLE 3. Percent Volume of Tumor Demonstrating Washout, Plateau, or Persistent Contrast Enhancement Kinetics
Enhancement kinetics (% volume)
Lum A/B
HER2 1
Triple negative
Washout (mean 6SD)
25.4 6 4.48
31.3 6 9.99
25.6 6 8.08
Plateau (mean 6SD)
24.7 6 3.19
28.7 6 5.05
26.6 6 7.99
Persistent (mean 6SD)
19.2 6 3.01
25.2 6 6.55
27.46 7.14
October 2015
923
Journal of Magnetic Resonance Imaging
both a greater number of cases and with different MRI protocols. The classification of tumors by molecular subtype was done by conventional methods and was limited by the lack of availability of the Ki-67 marker status of our cases, precluding independent analysis of Luminal A and B subtypes. In conclusion, our study suggests that the percent volume of HER2-positive tumors demonstrating rapid early contrast uptake is significantly increased compared to other molecular subtypes. Continued validation of MRI phenotypes like that which we identify for HER21 tumors may aid in diagnostic and treatment planning, and moreover yield independent markers for prognosis and treatment response, complementing traditional histopathologic and molecular analysis.
CONFLICT OF INTEREST
8.
Abramson RG, Arlinghaus LR, Weis JA, et al. Current and emerging quantitative magnetic resonance imaging methods for assessing and predicting the response of breast cancer to neoadjuvant therapy. Breast Cancer (London) 2012;2012:139–154.
9.
Bhooshan N, Giger ML, Jansen SA, Li H, Lan L, Newstead GM. Cancerous breast lesions on dynamic contrast-enhanced MR images: computerized characterization for image-based prognostic markers. Radiology 2010;254:680–690.
10.
Golden DI, Lipson JA, Telli ML, Ford JM, Rubin DL. Qualitative and quantitative image-based biomarkers of therapeutic response in triple-negative breast cancer. AMIA Summits Transl Sci Proc 2013; 2013:62.
11.
Ashraf AB, Daye D, Gavenonis S, et al. Identification of intrinsic imaging phenotypes for breast cancer tumors: preliminary associations with gene expression profiles. Radiology 2014:131375.
12.
Brenton JD, Carey LA, Ahmed AA, Caldas C. Molecular classification and molecular forecasting of breast cancer: ready for clinical application? J Clin Oncol 2005;23:7350–7360.
13.
Network CGA. Comprehensive molecular portraits of human breast tumours. Nature 2012;490:61–70.
14.
Sørlie T, Perou CM, Tibshirani R, et al. Gene expression patterns of breast carcinomas distinguish tumor subclasses with clinical implications. Proc Natl Acad Sci U S A 2001;98:10869–10874.
15.
Tamimi RM, Baer HJ, Marotti J, et al. Comparison of molecular phenotypes of ductal carcinoma in situ and invasive breast cancer. Breast Cancer Res 2008;10:R67.
16.
Goldhirsch A, Wood WC, Coates AS, et al. Strategies for subtypes– dealing with the diversity of breast cancer: highlights of the St. Gallen International Expert Consensus on the Primary Therapy of Early Breast Cancer 2011. Ann Oncol 2011;22:1736–1747.
17.
Morris EA, Comstock CE, Lee CH, et al. ACR BI-RADS magnetic resonance imaging. In: ACR BI-RADS Atlas, Breast Imaging Reporting and Data System. Reston, VA: American College of Radiology; 2013.
18.
Giger ML, Karssemeijer N, Schnabel JA. Breast image analysis for risk assessment, detection, diagnosis, and treatment of cancer. Annu Rev Biomed Eng 2013;15:327–357.
19.
Hsiang DJ, Yamamoto M, Mehta RS, et al. Predicting nodal status using dynamic contrast-enhanced magnetic resonance imaging in patients with locally advanced breast cancer undergoing neoadjuvant chemotherapy with and without sequential trastuzumab. Arch Surg 2007;142:855–861; discussion 860–851.
20.
Yamaguchi K, Abe H, Newstead GM, et al. Intratumoral heterogeneity of the distribution of kinetic parameters in breast cancer: comparison based on the molecular subtypes of invasive breast cancer. Breast Cancer 2014.
21.
Brown LF, Berse B, Jackman RW, et al. Expression of vascular permeability factor (vascular endothelial growth factor) and its receptors in breast cancer. Hum Pathol 1995;26:86–91.
Consultant, Seno Medical Instruments Inc.
REFERENCES 1.
Bluemke DA, Gatsonis CA, Chen MH, et al. Magnetic resonance imaging of the breast prior to biopsy. JAMA 2004;292:2735–2742.
2.
Esserman L, Hylton N, Yassa L, Barclay J, Frankel S, Sickles E. Utility of magnetic resonance imaging in the management of breast cancer: evidence for improved preoperative staging. J Clin Oncol 1999;17: 110–119.
3.
4.
Liberman L, Morris EA, Dershaw DD, Abramson AF, Tan LK. MR imaging of the ipsilateral breast in women with percutaneously proven breast cancer. AJR Am J Roentgenol 2003;180:901–910. Peters NH, Borel Rinkes IH, Zuithoff NP, Mali WP, Moons KG, Peeters PH. Meta-analysis of MR imaging in the diagnosis of breast lesions. Radiology 2008;246:116–124.
5.
El Khouli RH, Macura KJ, Jacobs MA, et al. Dynamic contrastenhanced MRI of the breast: quantitative method for kinetic curve type assessment. AJR Am J Roentgenol 2009;193:W295–300.
6.
Li SP, Padhani AR, Taylor NJ, et al. Vascular characterisation of triple negative breast carcinomas using dynamic MRI. Eur Radiol 2011;21: 1364–1373.
7.
Abramson RG, Li X, Hoyt TL, et al. Early assessment of breast cancer response to neoadjuvant chemotherapy by semi-quantitative analysis of high-temporal resolution DCE-MRI: preliminary results. Magn Reson Imaging 2013;31:1457–1464.
924
Volume 42, No. 4
