European Journal of Radiology 83 (2014) 2144–2150

Contents lists available at ScienceDirect

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

Correlation between 3 T apparent diffusion coefficient values and grading of invasive breast carcinoma Valentina Cipolla a,∗ , Domiziana Santucci a , Daniele Guerrieri a , Francesco Maria Drudi a , Maria Letizia Meggiorini b,1 , Carlo de Felice a a b

Department of Radiological Sciences, University of Rome “Sapienza”, Viale del Policlinico 155, 00161 Rome, Italy Department of Gynaecological Sciences, University of Rome “Sapienza”, Viale del Policlinico 155, 00161 Rome, Italy

a r t i c l e

i n f o

Article history: Received 12 July 2014 Received in revised form 18 September 2014 Accepted 22 September 2014 Keywords: Breast magnetic resonance Diffusion-weighted imaging Apparent diffusion coefficient Invasive breast carcinoma Grading

a b s t r a c t Purpose: The aim of this study was to evaluate whether the apparent diffusion coefficient (ADC) provided by 3.0 T (3 T) magnetic resonance diffusion-weighted imaging (DWI) varied according to the grading of invasive breast carcinoma. Materials and methods: A total of 92 patients with 96 invasive breast cancer lesions were enrolled; all had undergone 3 T magnetic resonance imaging (MRI) for local staging. All lesions were confirmed by histological analysis, and tumor grade was established according to the Nottingham Grading System (NGS). MRI included both dynamic contrast-enhanced and DWI sequences, and ADC value was calculated for each lesion. ADC values were compared with NGS classification using the Mann–Whitney U and the Kruskal–Wallis H tests. Grading was considered as a comprehensive prognostic factor, and Rho Spearman test was performed to determine correlation between grading and tumor size, hormonal receptor status, HER2 expression and Ki67 index. Pearson’s Chi square test was carried out to compare grading with the other prognostic factors. Results: ADC values were significantly higher in G1 than in G3 tumors. No significant difference was observed when G1 and G3 were compared with G2. Tumor size, hormonal receptor status, HER2 expression and Ki67 index correlated significantly with grading but there was a significant difference only between G1 and G3 related to the ER and PR status, HER2 expression and Ki67 index. There was no statistically significant difference in lesion size between the two groups. Conclusion: ADC values obtained on 3 T DWI correlated with low-grade (G1) and high-grade (G3) invasive breast carcinoma. 3 T ADC may be a helpful tool for identifying high-grade invasive breast carcinoma. © 2014 Elsevier Ireland Ltd. All rights reserved.

1. Introduction Dynamic contrast-enhanced magnetic resonance imaging (DCEMRI) is considered the most accurate imaging technique for diagnosis and local staging of breast cancer [1], however, several authors have reported a limited specificity [2]. In addition to this, DCE-MRI requires longer examination time and increased financial resources and may furthermore lead to various reactions to contrast medium [3]. Diffusion weighted imaging (DWI) has recently been integrated into the standard MRI protocol to increase the specificity of

∗ Corresponding author. Tel.: +39 064455602; fax: +39 064456695. E-mail addresses: [email protected], [email protected] (V. Cipolla). 1 Tel.: +39 3470385597; fax: +39 068554119. http://dx.doi.org/10.1016/j.ejrad.2014.09.015 0720-048X/© 2014 Elsevier Ireland Ltd. All rights reserved.

DCE-MRI in the differentiation between benign and malignant lesions. DWI is a fast and non-invasive technique, which reveals the biological character of the lesion, particularly the Brownian motion of protons in bulk water molecules. Apparent Diffusion Coefficient (ADC) is a quantitative parameter of DWI, and it reflects the mean diffusivity of water molecules in tissues [4,5]. It is influenced by certain microstructural characteristics such as tissue cellularity, fluid viscosity, membrane permeability and blood flow [6,7]. Several authors have reported substantially lower ADC values in malignant breast tumors than in benign breast lesions and normal tissues [4,5,7,8]. ADC has thus the ability to improve the accuracy of DCE-MRI in the characterization of breast lesions. The most significant advantage of DWI over DCE-MRI is its high sensitivity to changes in the microscopic environment without the need for contrast injection, which should be avoided in pregnant women and in patients with impaired renal function. However, DWI has a low spatial resolution, and small cancer foci or non-mass-like lesions may

V. Cipolla et al. / European Journal of Radiology 83 (2014) 2144–2150

not be depicted. Higher magnetic field strength such as 3 T affords a higher signal to noise ratio (SNR) that can be used to both increase spatial resolution and reduce acquisition time and a higher contrast to noise ratio (CNR) that can be used to improve the detection of smaller lesions [6]. Histological grade, established according to the Nottingham Grading System (NGS), is one of the main prognostic factors in invasive breast carcinoma. NGS provides a simple, inexpensive and easily applicable picture of the biological characteristics of the tumor. Histological grade can accurately predict evolution and response to therapy as well as prognosis [9]. Numerous studies have demonstrated that the prognostic value of NGS is equivalent to that of lymph node (LN) status [10] and greater than that of tumor size [11]. As the prognostic value of NGS is elevated, this parameter has been included in the prognostic indices, e.g., Nottingham Prognostic Index (NPI) [12] and the Kalmar Prognostic Index [11] and in algorithms (e.g., Adjuvant! Online [13]) and guidelines (e.g., the St. Gallen guidelines) [14]. Other biological markers that are used to estimate prognosis and predict response to treatment in breast cancer are hormone receptor (HR) expression, human epidermal growth factor receptor 2 (HER2) status and Ki67 labeling index. Several authors have already demonstrated the relationship between ADC values and prognostic markers [4,5]. The aim of this study was to evaluate whether ADC values provided by 3 T DWI vary according to the grading in invasive breast carcinoma. Correlation between grading and other prognostic markers including tumor size, estrogen receptor (ER), progesterone receptor (PR), human growth factor receptor 2 (HER2) and Ki-67 labeling index was also evaluated in order to demonstrate the ability of grading to reveal intrinsic biological aggressiveness of tumors and therefore to predict tumor evolution. 2. Materials and methods In this retrospective study, all breast MRI examinations performed at the Department of Radiological Sciences for local staging of breast cancer between April 2011 and January 2014 were reviewed. Patients were enrolled in the study only if the diagnostic procedure included the following: (a) 3 T MRI; (b) both DCE-MRI and DWI sequences; (c) ADC evaluation of the main lesion; (d) confirmation of MRI diagnosis by pathological analysis after surgery or core biopsy; (e) histopathological diagnosis demonstrating invasive breast cancer; (f) histological analysis including molecular receptor assessment (estrogen receptor ER, progesterone receptor PR; epidermal growth factor receptor HER2) and Ki-67 index calculation. Patients in neo-adjuvant chemotherapy and patients whose images were not of a good diagnostic quality were excluded from the study. Patients with breast implants were also excluded, as silicone breast implants do not permit successful dual suppression of fat and silicone using the short-tau inversion recovery (STIR) technique. Unsuppressed silicone signals from implants result in ghosting and chemical shift artifacts at DWI, which can obscure lesions and confound interpretation [15]. All MRI examinations were performed on a 3 T magnet (Discovery 750; GE Healthcare, Milwaukee, WI, USA) using a dedicated eight-channel breast coil (8US TORSOPA) with the patient in the prone position. After localizer sequences taken in three orthogonal planes, the following sequences were acquired: (1) Axial T2-weighted single shot fast spin echo sequence using a modified Dixon technique (IDEAL) for intravoxel fat-water separation (TR/TE 3500–5200/120–135 ms, matrix 352 × 224, FoV 370 × 370, NEX 1, slice thickness 3.5 mm).

2145

(2) Axial single shot fat suppressed echo-planar diffusion weighted sequence (TR/TE 2700/58 ms, matrix 100 × 120, FOV 360 × 360, NEX 6, slice thickness 5 mm) with diffusion-sensitizing gradient applied along the x, y, z axes and with a b-value of 0 and 1000 s/mm2 . (3) Axial T1-weighted 3D dynamic gradient echo fat suppressed sequence (VIBRANT) (TR/TE 6.6/4.3 ms, flip angle 10◦ , matrix 512 × 256, NEX 1, slice thickness 2.4 mm), before and five times after contrast administration. Contrast medium, gadobenate-dimeglumine (Multihance® ; Bracco Imaging, Milan, Italy), was administered in a concentration of 0.2 mmol/kg injected through a 20 G intravenous cannula at the rate of 2 ml/s using an automatic injector and followed by infusion of 15 ml saline solution at the same speed. Subtracted images were automatically derived from DCE-MRI. Images were transferred to a workstation (Advantage Windows Workstation 4.4; GE Medical System, Milwaukee) for postprocessing. For a quantitative analysis of the data obtained using DWI, parametric ADC maps were generated. ADC value was calculated according to the following equation: ADC = −(1/b)ln (S0 /S1 ) where b is the diffusion factor, S1 is the attenuated signal (b-value of 1000 s/mm2 ) and S0 is the full spin-echo signal without diffusion gradient (b-value of 0 s/mm2 ) [16]. MRI images were reviewed in consensus by two radiologists with 9 and 4 years’ experience in breast MRI; both were blinded to clinical and histopathological findings except the presence of invasive breast cancer. In order to standardize image analysis as much as possible, the radiologists first reviewed the DCE images, which were considered reference images for tumor detection. Subtracted images were used to detect the largest breast lesion, i.e., the index lesion, and the greatest diameter was submitted to statistical analysis. Additional lesions were considered only if >5 mm. Multifocality was diagnosed in the presence of multiple foci of malignancy in the same breast quadrant. Multicentricity was diagnosed when two or more foci of malignancy were found in more than one quadrant. Bilaterality was diagnosed if neoplastic lesions were found in both breasts (synchronous bilateral breast cancer). Tumors were divided into two groups according to the size (≤2 cm and >2 cm). Dynamic signal intensity-to-time curves were generated by positioning a region of interest (ROI) within the lesion on subjectively recognized areas of maximal contrast enhancement. All lesions were classified using the BIRADS lexicon [17] on the basis of kinetics curves and morphological findings. Subtracted images were subsequently superimposed on the DWI images (b = 1000 s/mm2 ) and on the ADC map for cancer lesion detection. A circular ROI measuring 3–6 mm in diameter was manually drawn on the ADC map, on the slice where the lesion reached the greatest diameter, and the ADC value was automatically calculated. ADC measurements were made only on the enhanced solid portion to avoid areas of T2 shine-through, i.e., the necrotic core of the tumor, which appears hyperintense on ADC maps and hypointense on subtracted images. Breast specimens obtained by core biopsy or after surgery were analyzed by a pathologist with more than 15 years’ experience. Histological diagnosis was performed according to the WHO classification. Tumor histopathological grade was assessed according to the NGS using a numerical scoring system for tubule formation, pleomorphism, and mitotic count. The total score ranges from 3 to 9. A total score of 3–5 was interpreted as grade 1 (G1), a total score of 6 or 7 was interpreted as grade 2 (G2) and a total score of 8 or 9 was interpreted as grade 3 (G3).

2146

V. Cipolla et al. / European Journal of Radiology 83 (2014) 2144–2150

Table 1 Histopathological and immunohistochemical results. Grading

Invasive ductal

Invasive lobular

ER+/Pg+

ER±/PgR± Ki67+

HER2+

TN

G1 G2 G3

29 16 28

3 8 12

25 1 4

6 13 (1 PgR−) 13 (1 PgR−)

1 9 (2 ER−/PgR−, 2 PgR−) 14 (7 ER−/PgR−)

0 1 9

ER, estrogen receptor; PgR, progesterone receptor; HER2, human epidermal growth factor receptor 2; TN, triple negative.

In addition, immunohistochemistry analysis (IHC) was performed to evaluate molecular receptors (ER, PR and HER2) and for Ki-67 index calculation. Assessment of ER and PR status was carried out by IHC using Dako monoclonal antibody, dilution 1:100. Only nuclear reactivity was taken into account for ER. HER2 status was reassessed according to recently published guidelines [18] using the Hercep Test (Dako, Glostrup, Denmark). Samples yielding an equivocal IHC result were subjected to fluorescence in situ hybridization (FISH) analysis. A ratio of HER2 gene signals to chromosome 17 signals of more than 2.2 was used as a cut-off value to define HER2 gene amplification. The Mib-1 monoclonal antibody (1:200 dilution; Dako, Glostrup, Denmark) was used to assess Ki-67, which was reported as the percentage of immunoreactive cells out of 2000 tumor cells in randomly selected, high-power fields surrounding the core of the tumor. ER and PR status was considered to be positive if the expression was ≥10% and negative if the expression was 0.01). Correlation between tumor grading (G1, G2 and G3) and other prognostic factors yielded the following results: in G1, 27 lesions measured 14% in 6 lesions (18.7%). In G2, 12 lesions measured