Breast Cancer Res Treat (2015) 149:751–759 DOI 10.1007/s10549-015-3278-6
CLINICAL TRIAL
Effect of menstrual cycle and menopausal status on apparent diffusion coefficient values and detectability of invasive ductal carcinoma on diffusion-weighted MRI Suyoung Shin • Eun Sook Ko • Rock Bum Kim • Boo-Kyung Han • Seok Jin Nam • Jung Hee Shin Soo Yeon Hahn
•
Received: 1 September 2014 / Accepted: 18 January 2015 / Published online: 1 February 2015 Ó Springer Science+Business Media New York 2015
Abstract The purpose of this study was to determine whether the apparent diffusion coefficient (ADC) and tumor detectability based on diffusion-weighted imaging (DWI) are affected by the menstrual cycle or menopausal status in breast cancer patients. Institutional review board approval was obtained, and the requirement for informed consent was waived. A total of 124 women with invasive ductal carcinoma not otherwise specified (IDC NOS) who underwent breast MRI with DWI were included in this study. Two radiologists retrospectively measured the ADCs of tumor and contralateral normal glandular tissue and scored the tumor detectability. The ADCs and detectability were compared to menstrual cycle and menopausal status, based on patient questionnaires. ADCs of tumors and contralateral tissue were significantly lower in postmenopausal women than in premenopausal women (P = 0.006 and P \ 0.001, respectively). Tumor detectability did not differ significantly between the premenopausal and postmenopausal groups (P = 0.454). S. Shin E. S. Ko (&) B.-K. Han J. H. Shin S. Y. Hahn Department of Radiology, Samsung Medical Center, Sungkyunkwan University School of Medicine, 50 Irwon-dong, Gangnam-gu, Seoul 135-710, Republic of Korea e-mail: [email protected] S. Shin Department of Radiology, Sharing and Happiness Hospital, Busan 612-035, Republic of Korea R. B. Kim Department of Preventive Medicine, Dong-A University School of Medicine, Busan, Korea S. J. Nam Department of Surgery, Samsung Medical Center, Sungkyunkwan University School of Medicine, Seoul, Republic of Korea
Normalized ADCs were not significantly lower in postmenopausal women compared to premenopausal women (P = 0.880). There was no statistically significant difference in the absolute, contralateral, and normalized ADCs (P = 0.091, 0.809, and 0.299, respectively), and the tumor detectability (P = 0.680) according to the menstrual cycle. Although ADCs of the IDC and normal glandular tissue in postmenopausal women were significantly lower than those in premenopausal women, the menstrual cycle did not affect tumor detectability and ADCs of IDC. Keywords Breast Breast cancer MRI Diffusionweighted imaging Menstrual cycle
Introduction Diffusion-weighted imaging (DWI) is a magnetic resonance imaging (MRI) technique that provides physiologic information of lesions by quantitatively measuring the diffusivity of water, known as the apparent diffusion coefficient (ADC). ADC values are obtained by measuring the mean diffusivity along three orthogonal directions and are affected by tissue cellularity, fluid viscosity, membrane permeability, and blood flow [1, 2]. Many studies have shown that malignant breast lesions usually have lower ADCs than benign lesions due to their increased cellularity restricting water diffusion [3–5]. Recent meta-analysis studies showed that combining DWI with conventional dynamic contrast-enhanced (DCE)-MRI improves the specificity from 72 % using DCE-MRI alone [6] to 85 % using both DCE-MRI and DWI [7]. In DCE-MRI, variation in contrast enhancement of normal breast parenchyma caused by normal hormonal fluctuation throughout the menstrual cycle is a well-known
123
752
phenomenon and could disturb lesion detection. Therefore, it is recommended that DCE-MRI of the breast be conducted during week 2 of the menstrual cycle in an effort to minimize background parenchymal enhancement [8, 9]. We questioned whether the menstrual cycle affected ADC values on DWI. Robust use of ADC values in daily practice requires that variations in ADC values according to normal hormonal fluctuations be minimal. While most previous studies showed no significant differences in ADC values of normal breast parenchyma during the menstrual cycle, there is still significant controversy [10–13]. One study proposed normalizing the ADC to normal glandular tissue in order to decrease the overlap of ADC values between benign and malignant lesions [14]. A recent study by Nissan et al. [15] investigated the parameters obtained using diffusion tensor imaging (DTI) of the breast under various hormonal conditions. Their results showed no significant differences in DTI parameters between premenopausal and postmenopausal women. Those prior studies have assessed the effects of the menstrual cycle or menopausal status on normal breast parenchyma and in small cohorts of lesions. Therefore, the purpose of our study was to evaluate whether the absolute, contralateral, and normalized ADC values and tumor detectability on DWI were affected by the menstrual cycle or menopausal status in breast cancer patients.
Materials and methods Patients Approval of the Institutional Review Board was obtained, and the requirement for informed consent was waived. Between May 2012 and August 2012, 176 patients underwent 1.5-T breast MRI at our institution for preoperative staging of recently diagnosed breast cancer. The interval between biopsy and MRI was typically 2 weeks. We aimed to evaluate the effect of the menstrual cycle under homogeneous conditions and exclude other factors that might influence ADC values such as lesion size, MRI findings, and pathologic type. Therefore, lesions were included in this study after meeting the following criteria: (1) final pathologic diagnosis was invasive ductal carcinoma not otherwise specified (IDC NOS); (2) presented as masses on MRI; and (3) largest diameter C1 cm. Patients were required to fill out a questionnaire before the MRI, which included questions concerning the menopausal status; the regularity of menstrual cycles (if patient was premenopausal); the date of the last menstrual period (LMP); and the use of hormonal therapy or oral contraceptives. Patients who had irregular menstrual cycles, were
123
Breast Cancer Res Treat (2015) 149:751–759
undergoing hormonal therapy, or were taking oral contraceptives were excluded. Altogether, we enrolled 124 patients in this study (mean age: 50.4 years; range: 30–77 years). The menstrual cycle was divided into 4 weeks. Week 1 began on the first day of menstruation. Image acquisition MRI was performed using a 1.5-T Achieva scanner (Philips Medical Systems, Best, The Netherlands) in prone position with a dedicated bilateral phased array breast coil. Our MRI protocol comprised turbo spin-echo T1- and T2weighted sequences and a 3-dimensional dynamic contrastenhanced sequence; all were bilateral axial scans. The scanning parameters for DCE-MRI were as follows: TR/ TE, 6.5/2.5; slice thickness, 1.5 mm; flip angle, 10°; matrix size, 376 9 374; and field of view, 32 9 32 cm. Before contrast agent injection, a spin-echo single-shot echo planar imaging sequence with diffusion-sensitizing gradients applied along the x-, y-, and z-axes was used before and after 180° pulses. These images were used to synthesize isotropic transverse DWI. DWI was performed with two b values (0 and 750 s/mm2). We chose a b value of 750 s/ mm2 based on prior reports [16, 17]. The scanning parameters for DWI were as follows: TR/TE, 15,000/66.7; slice thickness, 3 mm; matrix size, 156 9 158; and field of view, 30 cm. For dynamic contrast enhancement, a 0.1 mmol/kg bolus of gadobutrol (Gadovist; Bayer Schering PharmaAG, Berlin, Germany) was administered, followed by a 10-mL saline flush. Sequential pre- and postcontrast images were obtained at 30, 90, 150, 210, 270, and 330 s after contrast injection. Image subtraction was performed after the dynamic series. Image analysis MR images were retrospectively reviewed in consensus by two radiologists (K.E.S. and S.S; 8 and 1 years of experience interpreting breast MR images, respectively) blinded to the clinicopathological findings including mammographic density, hormonal receptor status, menstrual cycle, and menopausal status. Tumor size, defined as the longest tumor diameter, was measured in the second phase of contrast-enhanced images. To measure the ADC value, the two radiologists determined a region of interest (ROI) in consensus. Using contrast-enhanced MRI and DWI as a reference, the ROI was placed on the ADC map to include most, though not all, of the tumor to avoid a partial volume effect. The enhanced solid portion in the first or second phase of the MRI was used to position the ROI for the ADC measurements. The ROI was drawn on the slice in which the cancer showed the greatest diameter. In the case of
Breast Cancer Res Treat (2015) 149:751–759
multifocal or multicentric disease, the largest lesion was analyzed. The ADC value was automatically measured when the ROI was positioned. Care was taken to avoid cystic or necrotic portions of the tumor, which showed high-signal intensity on T2-weighted images. If a lesion was not hyperintense on DWI, then an ROI of equal size was drawn at the corresponding location as reflected on DCE-MRI. The mean size of the ROIs was 75.2 mm2 (range, 12.5–223.4 mm2). The same process was performed for the normal glandular tissue in the contralateral breast. The mean size of the ROIs in the contralateral breast was 61.5 mm2 (range, 15.6–98.4 mm2). The ADC values were measured in areas of homogeneous breast parenchyma that was identified from the T1-weighted images at the level of the nipple in order to minimize partial volume effects, as suggested by previous reports [10, 13]. If there was inadequate measurable tissue at the nipple level, a different site was selected, usually the upper outer breast. As we adopted and modified the concept of normalized ADC proposed by El Khouli et al. [14] for each lesion, the mean ADC for the invasive carcinoma (ADCl) and the contralateral normal glandular tissue (ADCg) was calculated. The normalized ADC (ADCn) was then calculated as follows: ADCn = ADCl/ADCg. To assess the tumor detectability on DWI, the radiologists identified lesions on DWI with two b values, indicated by an area with higher signal intensity than that of the surrounding parenchyma. Detectability was scored in consensus on DWI with a b value of 750 s/mm2 using a 5-point scale (1: not detectable, 2: slight, 3: fair, 4: moderate, 5: excellent). Histopathological analysis Pathological reports of biopsies or surgery were reviewed to determine the tumor type, histologic grade, molecular marker (estrogen receptor [ER], progesterone receptor [PR], and human epidermal growth factor receptor 2 [HER2]). ER and PR status were determined using the Allred score [18]. A positive result was defined as a total score of 3 or more. HER2 scores of tumors were graded as 0, 1, 2, and 3 with score 3 considered positive. Tumors with score 2 were sent for silver enhanced in situ hybridization (SISH) testing. The results were evaluated using the new recommendations of the American Society of Clinical Oncology/College of American Pathologists; a positive result for HER2 amplification was defined as a HER2 to Chr17 ratio of[2.2; equivocal for HER2 amplification was defined as a HER2/Chr17 ratio of 1.8–2.2; and negative for HER2 amplification was defined as a HER2/Chr17 ratio of \1.8 [19].
753
Statistical analysis Before testing for differences in clinicopathologic factors according to the menopausal status and menstrual cycle, the Shapiro–Wilk test for normality and Levene’s F test for equal variance were performed for the continuous data. Data that did not show normality and equal variance were analyzed using the Kruskal–Wallis test or Mann–Whitney U test. For normal data with equal variance, we used the Student’s t test or the analysis of variance (ANOVA). For categorical data, the Fisher’s exact test was used. The Kruskal–Wallis test was used to detect any significant differences in the tumor size, age, and the absolute and normalized ADC values in breast cancer and the contralateral normal glandular tissue according to four different weeks of the menstrual cycle. The Fisher’s exact test was used to detect any significant difference in the mammographic density, histologic grade, hormone receptor status, and HER2 status according to the four different weeks of menstrual cycle. The Mann–Whitney U test was used to compare the premenopausal and postmenopausal groups for tumor size and patient age. For comparisons of the mammographic density, histologic grade, hormone receptor status, and HER2 status, the Fisher’s exact test was used. The Kruskal–Wallis test and Mann–Whitney U test were used to evaluate differences in the mean ADC values of the breast cancer according to the menstrual cycle or menopausal status. The Fisher’s exact test was used to evaluate whether the tumor detectability differed according to the menstrual cycle and menopausal status. The relationship between ADC values and clinicopathologic factors was evaluated using the Spearman’s correlation test, Kruskal–Wallis test, and Student’s t test after confirming normality and equal variance. Spearman’s correlation coefficients were used to examine the possible correlation between tumor size or age and ADC values. The Kruskal–Wallis test was performed to determine whether the mammographic density and histologic tumor grade were correlated to the ADC value. The Kruskal– Wallis test was performed to evaluate whether there were significant changes in the contralateral and normalized ADC values according to the mammographic density. The Student’s t test was performed to evaluate whether the hormone receptor status and HER2 status affected the ADC values. For the multivariate analysis, multiple linear regression analysis was used to identify independent variables associated with the ADC value. All analyses were performed using statistical software (SPSS version 21.0; IBM Corp., Armonk, NY, USA) with P \ 0.05 (two-sided test) defined as the threshold for a statistically significant difference.
123
123 28 (22.6)
36 (29.0)
Negative
77 (62.1)
Negative
17 (68.0)
8 (32.0)
5 (20.0)
20 (80.0)
5 (20.0)
20 (80.0)
14 (70.0)
6 (30.0)
4 (20.0)
16 (80.0)
3 (15.0)
17 (85.0)
5 (25.0)
8 (40.0) 7 (35.0)
10 (50.0)
9 (45.0)
1 (5.0)
0 (0.0)
23.0 (13.8–34.5)
45.5 (41.0–47.0)
Week 2 (n = 20)
8 (57.1)
6 (42.9)
3 (21.4)
11 (78.6)
3 (21.4)
11 (78.6)
3 (21.4)
4 (28.6) 7 (50.0)
8 (57.1)
5 (35.7)
1 (7.1)
0 (0.0)
21.5 (15.0–26.3)
44.0 (33.5–47.3)
Week 3 (n = 14)
10 (71.4)
4 (28.6)
2 (14.3)
12 (85.7)
2 (14.3)
12 (85.7)
1 (7.1)
7 (50.0) 6 (42.9)
3 (21.4)
8 (57.1)
2 (14.3)
1 (7.1)
16.5 (13.8–32.5)
48.0 (41.0–51.0)
Week 4 (o = 14)
IQR interquartile range, ER estrogen receptor, PR progesterone receptor, HER2 human epidermal growth factor receptor 2
47 (37.9)
Positive
HER2 (N, %)
88 (71.0)
Positive
PR (N, %)
96 (77.4)
Negative
10 (40.0)
35 (28.2)
Positive
ER (N, %)
3
1 2
7 (28.0)
17 (68.0)
2 (8.0) 13 (52.0)
36 (29.0)
4
1 (4.0)
0 (0.0)
26.0 (20.0–35.0)
43.0 (40.0–47.0)
Week 1 (n = 25)
Premenopausal (n = 73)
37 (29.8) 52 (41.9)
63 (50.8)
3
Histologic grade (N, %)
5 (4.0) 20 (16.1)
1
2
Mammographic density (N, %)
48.0 (43.0–55.0) 22.0 (15.0–31.5)
Age, years (median, IQR)
Tumor size, mm (median, IQR)
Total (n = 124)
Table 1 Clinicopathologic features of 124 patients according to menstrual cycle and menopausal status
0.872
1.000
0.928
0.056
0.193
0.211
0.094
P value*
49 (67.1)
24 (32.9)
14 (19.2)
59 (80.8)
13 (17.8)
60 (82.2)
19 (26.0)
21 (28.8) 33 (45.2)
28 (38.4)
39 (53.4)
5 (6.8)
1 (1.4)
23.0 (15.0–31.0)
45.0 (41.0–48.0)
Subtotal (n = 73)
P value
28 (54.9)
23 (45.1)
22 (43.1)
29 (56.9)
15 (29.4)
36 (70.6)
16 (31.4)
16 (31.4) 19 (37.3)
8 (15.7)
24 (47.1)
15 (29.4)
4 (7.8)
22.0 (12.0–33.0)
0.191
0.005
0.190
0.697
\0.001
0.460
56.0 (52.0–63.0) \0.001
Postmenopausal (n = 51)
754 Breast Cancer Res Treat (2015) 149:751–759
Breast Cancer Res Treat (2015) 149:751–759
755
was no significant difference in the absolute ADC values according to the mammographic density (P = 0.487). However, the contralateral ADC values differed significantly according to the mammographic density. The median ADC values of the contralateral breast were 1.352 9 10-3 mm2/s for grade 1, 1.568 9 10-3 mm2/s for grade 2, 1.598 9 10-3 mm2/s for grade 3, and 1.683 9 10-3 mm2/s for grade 4 (P = 0.026). The median normalized ADC values were 0.732 for grade 1, 0.624 for grade 2, 0.618 for grade 3, and 0.601 for grade 4 (P = 0.690). The histologic grade was independently correlated with the absolute ADC value, and a higher histologic tumor grade showed a lower mean ADC value according to the multiple linear regression analysis (Table 4). Menopausal status was independently correlated with the absolute ADC value (P = 0.003).
Results Absolute, normalized, and contralateral ADC values according to menstrual cycle and menopausal status Among the 124 patients, 51 patients were postmenopausal, and 73 patients were premenopausal. The mean patient age, tumor size, mammographic density, histologic grade, and molecular markers are presented in Table 1. The percentage of PR-positive patients was higher in the premenopausal group than in the postmenopausal group (P = 0.005). There was no statistically significant difference in ER and HER2 expression between the groups. The tumor size and histologic grade did not significantly differ between the groups. Table 2 shows that the ADC values of the tumors and contralateral normal breast were significantly lower in the postmenopausal group compared to the premenopausal group (P \ 0.001 and P = 0.006, respectively). In premenopausal women, there was no significant difference in ADC values of the tumor and contralateral normal glandular tissue between the four menstrual weeks (P = 0.091 and 0.809, respectively). Contrary to the absolute ADC and contralateral ADC, the normalized ADC showed no significant difference between the premenopausal and postmenopausal women (P = 0.880; Figs. 1, 2). However, tumor detectability did not significantly differ between the premenopausal and postmenopausal groups or between the four different menstrual weeks (P = 0.454 and 0.680, respectively).
Discussion The breast is an organ affected significantly by hormones according to the menstrual cycle. During the proliferative phase (days 1–14), as estrogen levels rise, the intralobular stroma become dense, and stromal edema and infiltrate weaken. During the secretory phase (days 15–28), as progesterone levels rise, lobules are stimulated, and stromal edema and inflammatory cells increase again [10, 20]. As demonstrated in several studies, the composition of stromal and epithelial elements of the breast changes depending on the menstrual cycle due to hormonal variation. Together with these breast changes, the water content and molecular environment of the breasts are also expected to change [13]. Therefore, DWI, which is based on different signal intensities according to the degree of water content, is likely to be affected by the menstrual cycle. The question of whether the menstrual cycle affects the ADC values of
Univariate and multivariate analysis of factors affecting ADC values Univariate analysis was performed to evaluate the factors affecting absolute ADC values. The tumor size and histologic grade were correlated with the ADC values (Table 3). There
Table 2 Mean ADC value and tumor detectability based on DWI according to menopausal status and menstrual cycle Premenopausal (n = 73)
Postmenopausal (n = 51)
P value
Week 1 (n = 25)
Week 2 (n = 20)
Week 3 (n = 14)
Week 4 (n = 14)
P value
Subtotal (n = 73)
Contralateral ADC value (910-3 mm2/s)
1.616
1.609
1.710
1.656
0.809
1.640
1.480
\0.001
Absolute ADC value (910-3 mm2/s)
0.973
1.080
0.998
1.078
0.091
1.027
0.948
0.006
Normalized ADC value
0.619
0.686
0.603
0.669
0.299
0.644
0.648
0.880
1
0 (0.0)
0 (0.0)
0 (0.0)
0 (0.0)
0.680
0 (0.0)
1 (2.0)
0.454
2
3 (12.0)
1 (5.0)
3 (21.4)
4 (28.6)
11 (15.1)
4 (7.8)
3
4 (16.0)
2 (10.0)
1 (7.1)
1 (7.1)
8 (11.0)
9 (17.6)
4
4 (16.0)
2 (10.0)
2 (14.3)
3 (21.4)
11 (15.1)
7 (13.7)
5
14 (56.0)
15 (75.0)
8 (57.1)
6 (42.9)
43 (58.9)
30 (58.8)
Tumor detectability (%)
ADC apparent diffusion coefficient
123
756
Fig. 1 54-year-old postmenopausal woman with invasive ductal carcinoma of the left breast. a Axial contrast-enhanced T1-weighted subtraction image reveals a 38-mm oval mass with circumscribed margins. b Axial diffusion-weighted MRI (b value, 750 s/mm2) shows a high intensity mass. The tumor detectability was scored at 5. c Axial ADC map shows the same lesion with low signal intensity. The absolute ADC value within the indicated tumor ROI was 0.635 9 10-3 mm2/s, the contralateral ADC value was 1.345 9 10-3 mm2/s, and the normalized ADC value was 0.472
123
Breast Cancer Res Treat (2015) 149:751–759
Fig. 2 41-year-old woman with invasive ductal carcinoma of the left breast during her second menstrual week. a Axial contrast-enhanced T1-weighted subtraction image reveals a 24-mm oval mass with circumscribed margins. b Axial diffusion-weighted MRI (b value, 750 s/mm2) shows a high intensity mass. The tumor detectability was scored at 5. c Axial ADC map shows the same lesion with low signal intensity. The absolute ADC value within the indicated tumor ROI was 0.971 9 10-3 mm2/s, the contralateral ADC value was 1.718 9 10-3 mm2/s, and the normalized ADC value was 0.565
normal glandular tissue remains controversial. Some studies have reported that the ADC values of normal glandular tissue decreased in week 2 and increased during week 4, although these differences were not statistically
Breast Cancer Res Treat (2015) 149:751–759
757
Table 3 Univariate analysis of clinicopathologic variables with ADC values Patient number
Correlation coefficient with ADC values (rho, q)
Age
124
-0.089
Tumor size
124
-0.219
Median ADC values (IQR) (910-3 mm2/s)
0.328 0.015
Mammographic density
0.487
1
5
0.989 (0.794–1.071)
2
20
0.979 (0.797–1.039)
3
63
0.987 (0.895–1.104)
36
1.011 (0.871–1.139)
1
37
1.058 (0.997–1.165)
2
52
0.968 (0.861–1.056)
3
35
0.945 (0.807–1.041)
Positive
96
0.988 ± 0.163
Negative
28
1.018 ± 0.160
Positive
88
0.988 (0.864–1.103)
Negative
36
0.997 (0.898–1.119)
Positive
47
1.005 (0.897–1.123)
Negative
77
0.983 (0.865–1.083)
4 Histologic grade
P value
\0.001
ER
0.387
PR 0.484
HER2 0.187
ADC apparent diffusion coefficient, IQR interquartile range, ER estrogen receptor, PR progesterone receptor, HER2 human epidermal growth factor receptor 2
Table 4 Multiple linear regression analysis of variables independently associated with ADC values of breast tumor
Tumor size
Regression coefficient (b)*
95 % Confidence interval for b
P-value
-0.001
-0.003 to 0.001
0.158
-0.135 to -0.029
0.003
Menopausal status Pre-menopause Post menopause
0 -0.082
Histologic grade 1
0
2
-0.099
-0.164 to -0.033
0.003
3
-0.122
-0.193 to -0.050
0.001
ADC apparent diffusion coefficient * b value means that after adjusting for the other variables, if the P value is \0.05, the particular variable is linearly related to the ADC value
significant [10, 13]. Contrary to these results, one study showed a significant difference [11]. In our study, there was no statistically significant difference in the contralateral ADC values between the four menstrual weeks (P = 0.809). Many studies have also investigated the difference in tumor ADC values according to the menstrual cycle or menopausal status. One previous study reported that the absolute ADC values and normalized ADC values of both
benign and malignant breast lesions showed no significant difference between menstrual cycle and between premenopausal and postmenopausal women [14]. Our results partially agreed with these reports. In both studies, the absolute ADC values of the tumor were not affected by the menstrual cycle. However, in our study, the menopausal status affected the absolute ADC values of the tumors. These discrepancies may reflect the smaller sample size in the previous study [14], which only evaluated 60 malignant
123
758
lesions in 24 premenopausal patients and 36 postmenopausal patients and is less than half the number of patients in our study. In addition, we surmise that partial volume effects caused by the presence of fat may also contribute to the reduced absolute ADC values in postmenopausal women compared to premenopausal women. This hypothesis seems more plausible in the contralateral normal breast of the postmenopausal women. In our study, the absolute ADC values did not differ significantly according to mammographic density (P = 0.487), while the contralateral ADC values were significantly lower in fatty breasts (P = 0.026). The derived normalized ADC values did not differ significantly according to mammographic density (P = 0.690). These results support the hypothesis that a partial volume effect due to fat may affect ADC values. In our results, we questioned whether the effect of decreased absolute ADC values in the tumors of postmenopausal women was negated if the absolute ADC values were normalized to the normal glandular tissue, which was also low. El Khouli et al. [20] reported that the diagnostic performance of DWI in differentiating benign and malignant lesions was greatly improved by ADC normalization using remote ipsilateral glandular tissue. They found that the use of normalized ADCs improved the specificity from 72 to 92 %. In addition, ADC normalization significantly decreased the degree of overlap between benign and malignant lesions. Our results suggest that normalizing the ADC values could eliminate variation caused by menopausal status. Pathologically, O’Flynn et al. [10] found that diminishing serum estrogen and progesterone concentrations caused a reduction in the water-containing glandular epithelium, collagenization of the interlobular stroma, and thickening of the lobular basement membrane. Increased collagen is a known diffusion barrier [21, 22], and decreased water content may significantly decrease the mean ADC values and contribute to the lower ADC values in postmenopausal women compared to premenopausal women. According to our study, which is based on direct measurements of ADC and the scored tumor detectability on DWI, the menstrual cycle does not affect tumor detectability and ADC values assessed by DWI. As described earlier, whether ADC values were affected by the menstrual cycle is still controversial; Woodhams et al. asserted that DWI of the breast may not be recommended in week 2 of the menstrual cycle because the DWI contrast and ADC values between normal breast and breast tumor may be lessened during that period [11, 13, 23]. In our study, there was no statistically significant difference in the contralateral ADC values between the four menstrual weeks (P = 0.809). The absolute and normalized ADC values similarly did not show a significant difference according to the menstrual cycle (P = 0.091 and 0.299, respectively).
123
Breast Cancer Res Treat (2015) 149:751–759
Finally, contrary to the concern raised by Woodhams et al. [23], tumor detectability on DWI did not differ statistically according to the menstrual cycle (P = 0.680). Most likely, the change in water content associated with the menstrual cycle is subtle and insufficient to significantly affect tumor ADC values or detectability on DWI. In postmenopausal women, O’Flynn reported that the lower ADC values in postmenopausal women may reduce DWI contrast and disturb tumor detection [10]. Contrary to O’Flynn’s concern, tumor detectability in postmenopausal women did not significantly differ from that in premenopausal women in the present study. We suspect that the similar tumor detection on DWI between premenopausal and postmenopausal women is caused by the simultaneously decreased ADC values in the normal glandular tissue and tumor. Our results showed that the menopausal status was an independent factor affecting absolute ADC values of tumors and disappeared when the ADC values were normalized to normal glandular tissue. Our finding that higher histologic tumor grades were negatively associated with the mean ADC values is consistent with previous studies. According to Costantini and Razek et al., cases with a higher histological grade, larger tumor size, or abnormal axillary lymph nodes showed lower ADC values, which may indicate aggressive tumor behavior [24, 25]. In our study, tumor size was a factor associated with the ADC value according to the univariate analysis, but not the multivariate analysis. Our study has several limitations. First, each menstrual cycle group included a relatively small sample size. Larger sample sizes are needed to apply these results to the general patient population. Second, the images were interpreted by only two reviewers in consensus. Third, all of the included lesions were imaged after the biopsy, which may have influenced the ADC values. However, as described earlier, MRI was typically performed 2 weeks after the biopsy. This interval should be enough to resolve changes induced by the biopsy. Further, when we reviewed the MR images, none of the cases showed findings that could be attributed to the biopsy. Therefore, we suspect that imaging after the biopsy should not have affected our results. In conclusion, the menstrual cycle did not affect the ADC values of normal glandular tissue, the absolute and normalized ADC values of IDC, or the tumor detectability. However, postmenopausal women showed significantly lower ADC values compared to premenopausal women without any loss of tumor detectability, which was restored when lesions were normalized to normal glandular tissue.
Conflict of interest The authors declare that they have no conflicts of interest.
Breast Cancer Res Treat (2015) 149:751–759
References 1. Partridge SC, Mullins CD, Kurland BF et al (2010) Apparent diffusion coefficient values for discriminating benign and malignant breast MRI lesions: effects of lesion type and size. AJR Am J Roentgenol 194(6):1664–1673 2. Marini C, Iacconi C, Giannelli M, Cilotti A, Moretti M, Bartolozzi C (2007) Quantitative diffusion-weighted MR imaging in the differential diagnosis of breast lesion. Eur Radiol 17(10):2646– 2655 3. Woodhams R, Matsunaga K, Kan S et al (2005) ADC mapping of benign and malignant breast tumors. Magn Reson Med Sci 4(1): 35–42 4. Woodhams R, Matsunaga K, Iwabuchi K et al (2005) Diffusionweighted imaging of malignant breast tumors: the usefulness of apparent diffusion coefficient (ADC) value and ADC map for the detection of malignant breast tumors and evaluation of cancer extension. J Comput Assist Tomogr 29(5):644–649 5. Park MJ, Cha ES, Kang BJ, Ihn YK, Baik JH (2007) The role of diffusion-weighted imaging and the apparent diffusion coefficient (ADC) values for breast tumors. Korean J Radiol 8(5):390–396 6. Peters NH, Borel Rinkes IH, Zuithoff NP, Mali WP, Moons KG, Peeters PH (2008) Meta-analysis of MR imaging in the diagnosis of breast lesions. Radiology 246(1):116–124 7. Chen X, Li WL, Zhang YL, Wu Q, Guo YM, Bai ZL (2010) Meta-analysis of quantitative diffusion-weighted MR imaging in the differential diagnosis of breast lesions. BMC Cancer 10:693 8. Kuhl CK, Bieling HB, Gieseke J et al (1997) Healthy premenopausal breast parenchyma in dynamic contrast-enhanced MR imaging of the breast: normal contrast medium enhancement and cyclical-phase dependency. Radiology 203(1):137–144 9. Muller-Schimpfle M, Ohmenhauser K, Stoll P, Dietz K, Claussen CD (1997) Menstrual cycle and age: influence on parenchymal contrast medium enhancement in MR imaging of the breast. Radiology 203(1):145–149 10. O’Flynn EA, Morgan VA, Giles SL, deSouza NM (2012) Diffusion weighted imaging of the normal breast: reproducibility of apparent diffusion coefficient measurements and variation with menstrual cycle and menopausal status. Eur Radiol 22(7):1512– 1518 11. Clendenen TV, Kim S, Moy L et al (2013) Magnetic Resonance Imaging (MRI) of hormone-induced breast changes in young premenopausal women. Magn Reson Imaging 31(1):1–9 12. Abdel Razek AA, Gaballa G, Denewer A, Tawakol I (2010) Diffusion weighted MR imaging of the breast. Acad Radiol 17(3):382–386
759 13. Partridge SC, McKinnon GC, Henry RG, Hylton NM (2001) Menstrual cycle variation of apparent diffusion coefficients measured in the normal breast using MRI. J Magn Reson Imaging 14(4):433–438 14. Ei Khouli RH, Jacobs MA, Mezban SD et al (2010) Diffusionweighted imaging improves the diagnostic accuracy of conventional 3.0-T breast MR imaging. Radiology 256(1):64–73 15. Nissan N, Furman-Haran E, Shapiro-Feinberg M, Grobgeld D, Degani H (2014) Diffusion-tensor MR imaging of the breast: hormonal regulation. Radiology 271(3):672–680 16. Padhani AR, Liu G, Koh DM et al (2009) Diffusion-weighted magnetic resonance imaging as a cancer biomarker: consensus and recommendations. Neoplasia 11(2):102–125 17. Pereira FP, Martins G, Figueiredo E et al (2009) Assessment of breast lesions with diffusion-weighted MRI: comparing the use of different b values. AJR Am J Roentgenol 193(4):1030–1035 18. Allred DC, Harvey JM, Berardo M, Clark GM (1998) Prognostic and predictive factors in breast cancer by immunohistochemical analysis. Mod Pathol 11(2):155–168 19. Dietel M, Ellis IO, Hofler H et al (2007) Comparison of automated silver enhanced in situ hybridisation (SISH) and fluorescence ISH (FISH) for the validation of HER2 gene status in breast carcinoma according to the guidelines of the American Society of Clinical Oncology and the College of American Pathologists. Virchows Arch 451(1):19–25 20. Ramakrishnan R, Khan SA, Badve S (2002) Morphological changes in breast tissue with menstrual cycle. Mod Pathol 15(12):1348–1356 21. Erikson A, Andersen HN, Naess SN, Sikorski P, Davies CDL (2008) Physical and chemical modifications of collagen gels: impact on diffusion. Biopolymers 89(2):135–143 22. Ko ES, Han BK, Kim RB et al (2014) Apparent diffusion coefficient in estrogen receptor-positive invasive ductal breast carcinoma: correlations with tumor-stroma ratio. Radiology 271(1): 30–37 23. Woodhams R, Ramadan S, Stanwell P et al (2011) Diffusionweighted imaging of the breast: principles and clinical applications. Radiographics 31(4):1059–1084 24. Costantini M, Belli P, Rinaldi P et al (2010) Diffusion-weighted imaging in breast cancer: relationship between apparent diffusion coefficient and tumour aggressiveness. Clin Radiol 65(12): 1005–1012 25. Razek AA, Gaballa G, Denewer A, Nada N (2010) Invasive ductal carcinoma: correlation of apparent diffusion coefficient value with pathological prognostic factors. NMR Biomed 23(6):619– 623
123
CLINICAL TRIAL
Effect of menstrual cycle and menopausal status on apparent diffusion coefficient values and detectability of invasive ductal carcinoma on diffusion-weighted MRI Suyoung Shin • Eun Sook Ko • Rock Bum Kim • Boo-Kyung Han • Seok Jin Nam • Jung Hee Shin Soo Yeon Hahn
•
Received: 1 September 2014 / Accepted: 18 January 2015 / Published online: 1 February 2015 Ó Springer Science+Business Media New York 2015
Abstract The purpose of this study was to determine whether the apparent diffusion coefficient (ADC) and tumor detectability based on diffusion-weighted imaging (DWI) are affected by the menstrual cycle or menopausal status in breast cancer patients. Institutional review board approval was obtained, and the requirement for informed consent was waived. A total of 124 women with invasive ductal carcinoma not otherwise specified (IDC NOS) who underwent breast MRI with DWI were included in this study. Two radiologists retrospectively measured the ADCs of tumor and contralateral normal glandular tissue and scored the tumor detectability. The ADCs and detectability were compared to menstrual cycle and menopausal status, based on patient questionnaires. ADCs of tumors and contralateral tissue were significantly lower in postmenopausal women than in premenopausal women (P = 0.006 and P \ 0.001, respectively). Tumor detectability did not differ significantly between the premenopausal and postmenopausal groups (P = 0.454). S. Shin E. S. Ko (&) B.-K. Han J. H. Shin S. Y. Hahn Department of Radiology, Samsung Medical Center, Sungkyunkwan University School of Medicine, 50 Irwon-dong, Gangnam-gu, Seoul 135-710, Republic of Korea e-mail: [email protected] S. Shin Department of Radiology, Sharing and Happiness Hospital, Busan 612-035, Republic of Korea R. B. Kim Department of Preventive Medicine, Dong-A University School of Medicine, Busan, Korea S. J. Nam Department of Surgery, Samsung Medical Center, Sungkyunkwan University School of Medicine, Seoul, Republic of Korea
Normalized ADCs were not significantly lower in postmenopausal women compared to premenopausal women (P = 0.880). There was no statistically significant difference in the absolute, contralateral, and normalized ADCs (P = 0.091, 0.809, and 0.299, respectively), and the tumor detectability (P = 0.680) according to the menstrual cycle. Although ADCs of the IDC and normal glandular tissue in postmenopausal women were significantly lower than those in premenopausal women, the menstrual cycle did not affect tumor detectability and ADCs of IDC. Keywords Breast Breast cancer MRI Diffusionweighted imaging Menstrual cycle
Introduction Diffusion-weighted imaging (DWI) is a magnetic resonance imaging (MRI) technique that provides physiologic information of lesions by quantitatively measuring the diffusivity of water, known as the apparent diffusion coefficient (ADC). ADC values are obtained by measuring the mean diffusivity along three orthogonal directions and are affected by tissue cellularity, fluid viscosity, membrane permeability, and blood flow [1, 2]. Many studies have shown that malignant breast lesions usually have lower ADCs than benign lesions due to their increased cellularity restricting water diffusion [3–5]. Recent meta-analysis studies showed that combining DWI with conventional dynamic contrast-enhanced (DCE)-MRI improves the specificity from 72 % using DCE-MRI alone [6] to 85 % using both DCE-MRI and DWI [7]. In DCE-MRI, variation in contrast enhancement of normal breast parenchyma caused by normal hormonal fluctuation throughout the menstrual cycle is a well-known
123
752
phenomenon and could disturb lesion detection. Therefore, it is recommended that DCE-MRI of the breast be conducted during week 2 of the menstrual cycle in an effort to minimize background parenchymal enhancement [8, 9]. We questioned whether the menstrual cycle affected ADC values on DWI. Robust use of ADC values in daily practice requires that variations in ADC values according to normal hormonal fluctuations be minimal. While most previous studies showed no significant differences in ADC values of normal breast parenchyma during the menstrual cycle, there is still significant controversy [10–13]. One study proposed normalizing the ADC to normal glandular tissue in order to decrease the overlap of ADC values between benign and malignant lesions [14]. A recent study by Nissan et al. [15] investigated the parameters obtained using diffusion tensor imaging (DTI) of the breast under various hormonal conditions. Their results showed no significant differences in DTI parameters between premenopausal and postmenopausal women. Those prior studies have assessed the effects of the menstrual cycle or menopausal status on normal breast parenchyma and in small cohorts of lesions. Therefore, the purpose of our study was to evaluate whether the absolute, contralateral, and normalized ADC values and tumor detectability on DWI were affected by the menstrual cycle or menopausal status in breast cancer patients.
Materials and methods Patients Approval of the Institutional Review Board was obtained, and the requirement for informed consent was waived. Between May 2012 and August 2012, 176 patients underwent 1.5-T breast MRI at our institution for preoperative staging of recently diagnosed breast cancer. The interval between biopsy and MRI was typically 2 weeks. We aimed to evaluate the effect of the menstrual cycle under homogeneous conditions and exclude other factors that might influence ADC values such as lesion size, MRI findings, and pathologic type. Therefore, lesions were included in this study after meeting the following criteria: (1) final pathologic diagnosis was invasive ductal carcinoma not otherwise specified (IDC NOS); (2) presented as masses on MRI; and (3) largest diameter C1 cm. Patients were required to fill out a questionnaire before the MRI, which included questions concerning the menopausal status; the regularity of menstrual cycles (if patient was premenopausal); the date of the last menstrual period (LMP); and the use of hormonal therapy or oral contraceptives. Patients who had irregular menstrual cycles, were
123
Breast Cancer Res Treat (2015) 149:751–759
undergoing hormonal therapy, or were taking oral contraceptives were excluded. Altogether, we enrolled 124 patients in this study (mean age: 50.4 years; range: 30–77 years). The menstrual cycle was divided into 4 weeks. Week 1 began on the first day of menstruation. Image acquisition MRI was performed using a 1.5-T Achieva scanner (Philips Medical Systems, Best, The Netherlands) in prone position with a dedicated bilateral phased array breast coil. Our MRI protocol comprised turbo spin-echo T1- and T2weighted sequences and a 3-dimensional dynamic contrastenhanced sequence; all were bilateral axial scans. The scanning parameters for DCE-MRI were as follows: TR/ TE, 6.5/2.5; slice thickness, 1.5 mm; flip angle, 10°; matrix size, 376 9 374; and field of view, 32 9 32 cm. Before contrast agent injection, a spin-echo single-shot echo planar imaging sequence with diffusion-sensitizing gradients applied along the x-, y-, and z-axes was used before and after 180° pulses. These images were used to synthesize isotropic transverse DWI. DWI was performed with two b values (0 and 750 s/mm2). We chose a b value of 750 s/ mm2 based on prior reports [16, 17]. The scanning parameters for DWI were as follows: TR/TE, 15,000/66.7; slice thickness, 3 mm; matrix size, 156 9 158; and field of view, 30 cm. For dynamic contrast enhancement, a 0.1 mmol/kg bolus of gadobutrol (Gadovist; Bayer Schering PharmaAG, Berlin, Germany) was administered, followed by a 10-mL saline flush. Sequential pre- and postcontrast images were obtained at 30, 90, 150, 210, 270, and 330 s after contrast injection. Image subtraction was performed after the dynamic series. Image analysis MR images were retrospectively reviewed in consensus by two radiologists (K.E.S. and S.S; 8 and 1 years of experience interpreting breast MR images, respectively) blinded to the clinicopathological findings including mammographic density, hormonal receptor status, menstrual cycle, and menopausal status. Tumor size, defined as the longest tumor diameter, was measured in the second phase of contrast-enhanced images. To measure the ADC value, the two radiologists determined a region of interest (ROI) in consensus. Using contrast-enhanced MRI and DWI as a reference, the ROI was placed on the ADC map to include most, though not all, of the tumor to avoid a partial volume effect. The enhanced solid portion in the first or second phase of the MRI was used to position the ROI for the ADC measurements. The ROI was drawn on the slice in which the cancer showed the greatest diameter. In the case of
Breast Cancer Res Treat (2015) 149:751–759
multifocal or multicentric disease, the largest lesion was analyzed. The ADC value was automatically measured when the ROI was positioned. Care was taken to avoid cystic or necrotic portions of the tumor, which showed high-signal intensity on T2-weighted images. If a lesion was not hyperintense on DWI, then an ROI of equal size was drawn at the corresponding location as reflected on DCE-MRI. The mean size of the ROIs was 75.2 mm2 (range, 12.5–223.4 mm2). The same process was performed for the normal glandular tissue in the contralateral breast. The mean size of the ROIs in the contralateral breast was 61.5 mm2 (range, 15.6–98.4 mm2). The ADC values were measured in areas of homogeneous breast parenchyma that was identified from the T1-weighted images at the level of the nipple in order to minimize partial volume effects, as suggested by previous reports [10, 13]. If there was inadequate measurable tissue at the nipple level, a different site was selected, usually the upper outer breast. As we adopted and modified the concept of normalized ADC proposed by El Khouli et al. [14] for each lesion, the mean ADC for the invasive carcinoma (ADCl) and the contralateral normal glandular tissue (ADCg) was calculated. The normalized ADC (ADCn) was then calculated as follows: ADCn = ADCl/ADCg. To assess the tumor detectability on DWI, the radiologists identified lesions on DWI with two b values, indicated by an area with higher signal intensity than that of the surrounding parenchyma. Detectability was scored in consensus on DWI with a b value of 750 s/mm2 using a 5-point scale (1: not detectable, 2: slight, 3: fair, 4: moderate, 5: excellent). Histopathological analysis Pathological reports of biopsies or surgery were reviewed to determine the tumor type, histologic grade, molecular marker (estrogen receptor [ER], progesterone receptor [PR], and human epidermal growth factor receptor 2 [HER2]). ER and PR status were determined using the Allred score [18]. A positive result was defined as a total score of 3 or more. HER2 scores of tumors were graded as 0, 1, 2, and 3 with score 3 considered positive. Tumors with score 2 were sent for silver enhanced in situ hybridization (SISH) testing. The results were evaluated using the new recommendations of the American Society of Clinical Oncology/College of American Pathologists; a positive result for HER2 amplification was defined as a HER2 to Chr17 ratio of[2.2; equivocal for HER2 amplification was defined as a HER2/Chr17 ratio of 1.8–2.2; and negative for HER2 amplification was defined as a HER2/Chr17 ratio of \1.8 [19].
753
Statistical analysis Before testing for differences in clinicopathologic factors according to the menopausal status and menstrual cycle, the Shapiro–Wilk test for normality and Levene’s F test for equal variance were performed for the continuous data. Data that did not show normality and equal variance were analyzed using the Kruskal–Wallis test or Mann–Whitney U test. For normal data with equal variance, we used the Student’s t test or the analysis of variance (ANOVA). For categorical data, the Fisher’s exact test was used. The Kruskal–Wallis test was used to detect any significant differences in the tumor size, age, and the absolute and normalized ADC values in breast cancer and the contralateral normal glandular tissue according to four different weeks of the menstrual cycle. The Fisher’s exact test was used to detect any significant difference in the mammographic density, histologic grade, hormone receptor status, and HER2 status according to the four different weeks of menstrual cycle. The Mann–Whitney U test was used to compare the premenopausal and postmenopausal groups for tumor size and patient age. For comparisons of the mammographic density, histologic grade, hormone receptor status, and HER2 status, the Fisher’s exact test was used. The Kruskal–Wallis test and Mann–Whitney U test were used to evaluate differences in the mean ADC values of the breast cancer according to the menstrual cycle or menopausal status. The Fisher’s exact test was used to evaluate whether the tumor detectability differed according to the menstrual cycle and menopausal status. The relationship between ADC values and clinicopathologic factors was evaluated using the Spearman’s correlation test, Kruskal–Wallis test, and Student’s t test after confirming normality and equal variance. Spearman’s correlation coefficients were used to examine the possible correlation between tumor size or age and ADC values. The Kruskal–Wallis test was performed to determine whether the mammographic density and histologic tumor grade were correlated to the ADC value. The Kruskal– Wallis test was performed to evaluate whether there were significant changes in the contralateral and normalized ADC values according to the mammographic density. The Student’s t test was performed to evaluate whether the hormone receptor status and HER2 status affected the ADC values. For the multivariate analysis, multiple linear regression analysis was used to identify independent variables associated with the ADC value. All analyses were performed using statistical software (SPSS version 21.0; IBM Corp., Armonk, NY, USA) with P \ 0.05 (two-sided test) defined as the threshold for a statistically significant difference.
123
123 28 (22.6)
36 (29.0)
Negative
77 (62.1)
Negative
17 (68.0)
8 (32.0)
5 (20.0)
20 (80.0)
5 (20.0)
20 (80.0)
14 (70.0)
6 (30.0)
4 (20.0)
16 (80.0)
3 (15.0)
17 (85.0)
5 (25.0)
8 (40.0) 7 (35.0)
10 (50.0)
9 (45.0)
1 (5.0)
0 (0.0)
23.0 (13.8–34.5)
45.5 (41.0–47.0)
Week 2 (n = 20)
8 (57.1)
6 (42.9)
3 (21.4)
11 (78.6)
3 (21.4)
11 (78.6)
3 (21.4)
4 (28.6) 7 (50.0)
8 (57.1)
5 (35.7)
1 (7.1)
0 (0.0)
21.5 (15.0–26.3)
44.0 (33.5–47.3)
Week 3 (n = 14)
10 (71.4)
4 (28.6)
2 (14.3)
12 (85.7)
2 (14.3)
12 (85.7)
1 (7.1)
7 (50.0) 6 (42.9)
3 (21.4)
8 (57.1)
2 (14.3)
1 (7.1)
16.5 (13.8–32.5)
48.0 (41.0–51.0)
Week 4 (o = 14)
IQR interquartile range, ER estrogen receptor, PR progesterone receptor, HER2 human epidermal growth factor receptor 2
47 (37.9)
Positive
HER2 (N, %)
88 (71.0)
Positive
PR (N, %)
96 (77.4)
Negative
10 (40.0)
35 (28.2)
Positive
ER (N, %)
3
1 2
7 (28.0)
17 (68.0)
2 (8.0) 13 (52.0)
36 (29.0)
4
1 (4.0)
0 (0.0)
26.0 (20.0–35.0)
43.0 (40.0–47.0)
Week 1 (n = 25)
Premenopausal (n = 73)
37 (29.8) 52 (41.9)
63 (50.8)
3
Histologic grade (N, %)
5 (4.0) 20 (16.1)
1
2
Mammographic density (N, %)
48.0 (43.0–55.0) 22.0 (15.0–31.5)
Age, years (median, IQR)
Tumor size, mm (median, IQR)
Total (n = 124)
Table 1 Clinicopathologic features of 124 patients according to menstrual cycle and menopausal status
0.872
1.000
0.928
0.056
0.193
0.211
0.094
P value*
49 (67.1)
24 (32.9)
14 (19.2)
59 (80.8)
13 (17.8)
60 (82.2)
19 (26.0)
21 (28.8) 33 (45.2)
28 (38.4)
39 (53.4)
5 (6.8)
1 (1.4)
23.0 (15.0–31.0)
45.0 (41.0–48.0)
Subtotal (n = 73)
P value
28 (54.9)
23 (45.1)
22 (43.1)
29 (56.9)
15 (29.4)
36 (70.6)
16 (31.4)
16 (31.4) 19 (37.3)
8 (15.7)
24 (47.1)
15 (29.4)
4 (7.8)
22.0 (12.0–33.0)
0.191
0.005
0.190
0.697
\0.001
0.460
56.0 (52.0–63.0) \0.001
Postmenopausal (n = 51)
754 Breast Cancer Res Treat (2015) 149:751–759
Breast Cancer Res Treat (2015) 149:751–759
755
was no significant difference in the absolute ADC values according to the mammographic density (P = 0.487). However, the contralateral ADC values differed significantly according to the mammographic density. The median ADC values of the contralateral breast were 1.352 9 10-3 mm2/s for grade 1, 1.568 9 10-3 mm2/s for grade 2, 1.598 9 10-3 mm2/s for grade 3, and 1.683 9 10-3 mm2/s for grade 4 (P = 0.026). The median normalized ADC values were 0.732 for grade 1, 0.624 for grade 2, 0.618 for grade 3, and 0.601 for grade 4 (P = 0.690). The histologic grade was independently correlated with the absolute ADC value, and a higher histologic tumor grade showed a lower mean ADC value according to the multiple linear regression analysis (Table 4). Menopausal status was independently correlated with the absolute ADC value (P = 0.003).
Results Absolute, normalized, and contralateral ADC values according to menstrual cycle and menopausal status Among the 124 patients, 51 patients were postmenopausal, and 73 patients were premenopausal. The mean patient age, tumor size, mammographic density, histologic grade, and molecular markers are presented in Table 1. The percentage of PR-positive patients was higher in the premenopausal group than in the postmenopausal group (P = 0.005). There was no statistically significant difference in ER and HER2 expression between the groups. The tumor size and histologic grade did not significantly differ between the groups. Table 2 shows that the ADC values of the tumors and contralateral normal breast were significantly lower in the postmenopausal group compared to the premenopausal group (P \ 0.001 and P = 0.006, respectively). In premenopausal women, there was no significant difference in ADC values of the tumor and contralateral normal glandular tissue between the four menstrual weeks (P = 0.091 and 0.809, respectively). Contrary to the absolute ADC and contralateral ADC, the normalized ADC showed no significant difference between the premenopausal and postmenopausal women (P = 0.880; Figs. 1, 2). However, tumor detectability did not significantly differ between the premenopausal and postmenopausal groups or between the four different menstrual weeks (P = 0.454 and 0.680, respectively).
Discussion The breast is an organ affected significantly by hormones according to the menstrual cycle. During the proliferative phase (days 1–14), as estrogen levels rise, the intralobular stroma become dense, and stromal edema and infiltrate weaken. During the secretory phase (days 15–28), as progesterone levels rise, lobules are stimulated, and stromal edema and inflammatory cells increase again [10, 20]. As demonstrated in several studies, the composition of stromal and epithelial elements of the breast changes depending on the menstrual cycle due to hormonal variation. Together with these breast changes, the water content and molecular environment of the breasts are also expected to change [13]. Therefore, DWI, which is based on different signal intensities according to the degree of water content, is likely to be affected by the menstrual cycle. The question of whether the menstrual cycle affects the ADC values of
Univariate and multivariate analysis of factors affecting ADC values Univariate analysis was performed to evaluate the factors affecting absolute ADC values. The tumor size and histologic grade were correlated with the ADC values (Table 3). There
Table 2 Mean ADC value and tumor detectability based on DWI according to menopausal status and menstrual cycle Premenopausal (n = 73)
Postmenopausal (n = 51)
P value
Week 1 (n = 25)
Week 2 (n = 20)
Week 3 (n = 14)
Week 4 (n = 14)
P value
Subtotal (n = 73)
Contralateral ADC value (910-3 mm2/s)
1.616
1.609
1.710
1.656
0.809
1.640
1.480
\0.001
Absolute ADC value (910-3 mm2/s)
0.973
1.080
0.998
1.078
0.091
1.027
0.948
0.006
Normalized ADC value
0.619
0.686
0.603
0.669
0.299
0.644
0.648
0.880
1
0 (0.0)
0 (0.0)
0 (0.0)
0 (0.0)
0.680
0 (0.0)
1 (2.0)
0.454
2
3 (12.0)
1 (5.0)
3 (21.4)
4 (28.6)
11 (15.1)
4 (7.8)
3
4 (16.0)
2 (10.0)
1 (7.1)
1 (7.1)
8 (11.0)
9 (17.6)
4
4 (16.0)
2 (10.0)
2 (14.3)
3 (21.4)
11 (15.1)
7 (13.7)
5
14 (56.0)
15 (75.0)
8 (57.1)
6 (42.9)
43 (58.9)
30 (58.8)
Tumor detectability (%)
ADC apparent diffusion coefficient
123
756
Fig. 1 54-year-old postmenopausal woman with invasive ductal carcinoma of the left breast. a Axial contrast-enhanced T1-weighted subtraction image reveals a 38-mm oval mass with circumscribed margins. b Axial diffusion-weighted MRI (b value, 750 s/mm2) shows a high intensity mass. The tumor detectability was scored at 5. c Axial ADC map shows the same lesion with low signal intensity. The absolute ADC value within the indicated tumor ROI was 0.635 9 10-3 mm2/s, the contralateral ADC value was 1.345 9 10-3 mm2/s, and the normalized ADC value was 0.472
123
Breast Cancer Res Treat (2015) 149:751–759
Fig. 2 41-year-old woman with invasive ductal carcinoma of the left breast during her second menstrual week. a Axial contrast-enhanced T1-weighted subtraction image reveals a 24-mm oval mass with circumscribed margins. b Axial diffusion-weighted MRI (b value, 750 s/mm2) shows a high intensity mass. The tumor detectability was scored at 5. c Axial ADC map shows the same lesion with low signal intensity. The absolute ADC value within the indicated tumor ROI was 0.971 9 10-3 mm2/s, the contralateral ADC value was 1.718 9 10-3 mm2/s, and the normalized ADC value was 0.565
normal glandular tissue remains controversial. Some studies have reported that the ADC values of normal glandular tissue decreased in week 2 and increased during week 4, although these differences were not statistically
Breast Cancer Res Treat (2015) 149:751–759
757
Table 3 Univariate analysis of clinicopathologic variables with ADC values Patient number
Correlation coefficient with ADC values (rho, q)
Age
124
-0.089
Tumor size
124
-0.219
Median ADC values (IQR) (910-3 mm2/s)
0.328 0.015
Mammographic density
0.487
1
5
0.989 (0.794–1.071)
2
20
0.979 (0.797–1.039)
3
63
0.987 (0.895–1.104)
36
1.011 (0.871–1.139)
1
37
1.058 (0.997–1.165)
2
52
0.968 (0.861–1.056)
3
35
0.945 (0.807–1.041)
Positive
96
0.988 ± 0.163
Negative
28
1.018 ± 0.160
Positive
88
0.988 (0.864–1.103)
Negative
36
0.997 (0.898–1.119)
Positive
47
1.005 (0.897–1.123)
Negative
77
0.983 (0.865–1.083)
4 Histologic grade
P value
\0.001
ER
0.387
PR 0.484
HER2 0.187
ADC apparent diffusion coefficient, IQR interquartile range, ER estrogen receptor, PR progesterone receptor, HER2 human epidermal growth factor receptor 2
Table 4 Multiple linear regression analysis of variables independently associated with ADC values of breast tumor
Tumor size
Regression coefficient (b)*
95 % Confidence interval for b
P-value
-0.001
-0.003 to 0.001
0.158
-0.135 to -0.029
0.003
Menopausal status Pre-menopause Post menopause
0 -0.082
Histologic grade 1
0
2
-0.099
-0.164 to -0.033
0.003
3
-0.122
-0.193 to -0.050
0.001
ADC apparent diffusion coefficient * b value means that after adjusting for the other variables, if the P value is \0.05, the particular variable is linearly related to the ADC value
significant [10, 13]. Contrary to these results, one study showed a significant difference [11]. In our study, there was no statistically significant difference in the contralateral ADC values between the four menstrual weeks (P = 0.809). Many studies have also investigated the difference in tumor ADC values according to the menstrual cycle or menopausal status. One previous study reported that the absolute ADC values and normalized ADC values of both
benign and malignant breast lesions showed no significant difference between menstrual cycle and between premenopausal and postmenopausal women [14]. Our results partially agreed with these reports. In both studies, the absolute ADC values of the tumor were not affected by the menstrual cycle. However, in our study, the menopausal status affected the absolute ADC values of the tumors. These discrepancies may reflect the smaller sample size in the previous study [14], which only evaluated 60 malignant
123
758
lesions in 24 premenopausal patients and 36 postmenopausal patients and is less than half the number of patients in our study. In addition, we surmise that partial volume effects caused by the presence of fat may also contribute to the reduced absolute ADC values in postmenopausal women compared to premenopausal women. This hypothesis seems more plausible in the contralateral normal breast of the postmenopausal women. In our study, the absolute ADC values did not differ significantly according to mammographic density (P = 0.487), while the contralateral ADC values were significantly lower in fatty breasts (P = 0.026). The derived normalized ADC values did not differ significantly according to mammographic density (P = 0.690). These results support the hypothesis that a partial volume effect due to fat may affect ADC values. In our results, we questioned whether the effect of decreased absolute ADC values in the tumors of postmenopausal women was negated if the absolute ADC values were normalized to the normal glandular tissue, which was also low. El Khouli et al. [20] reported that the diagnostic performance of DWI in differentiating benign and malignant lesions was greatly improved by ADC normalization using remote ipsilateral glandular tissue. They found that the use of normalized ADCs improved the specificity from 72 to 92 %. In addition, ADC normalization significantly decreased the degree of overlap between benign and malignant lesions. Our results suggest that normalizing the ADC values could eliminate variation caused by menopausal status. Pathologically, O’Flynn et al. [10] found that diminishing serum estrogen and progesterone concentrations caused a reduction in the water-containing glandular epithelium, collagenization of the interlobular stroma, and thickening of the lobular basement membrane. Increased collagen is a known diffusion barrier [21, 22], and decreased water content may significantly decrease the mean ADC values and contribute to the lower ADC values in postmenopausal women compared to premenopausal women. According to our study, which is based on direct measurements of ADC and the scored tumor detectability on DWI, the menstrual cycle does not affect tumor detectability and ADC values assessed by DWI. As described earlier, whether ADC values were affected by the menstrual cycle is still controversial; Woodhams et al. asserted that DWI of the breast may not be recommended in week 2 of the menstrual cycle because the DWI contrast and ADC values between normal breast and breast tumor may be lessened during that period [11, 13, 23]. In our study, there was no statistically significant difference in the contralateral ADC values between the four menstrual weeks (P = 0.809). The absolute and normalized ADC values similarly did not show a significant difference according to the menstrual cycle (P = 0.091 and 0.299, respectively).
123
Breast Cancer Res Treat (2015) 149:751–759
Finally, contrary to the concern raised by Woodhams et al. [23], tumor detectability on DWI did not differ statistically according to the menstrual cycle (P = 0.680). Most likely, the change in water content associated with the menstrual cycle is subtle and insufficient to significantly affect tumor ADC values or detectability on DWI. In postmenopausal women, O’Flynn reported that the lower ADC values in postmenopausal women may reduce DWI contrast and disturb tumor detection [10]. Contrary to O’Flynn’s concern, tumor detectability in postmenopausal women did not significantly differ from that in premenopausal women in the present study. We suspect that the similar tumor detection on DWI between premenopausal and postmenopausal women is caused by the simultaneously decreased ADC values in the normal glandular tissue and tumor. Our results showed that the menopausal status was an independent factor affecting absolute ADC values of tumors and disappeared when the ADC values were normalized to normal glandular tissue. Our finding that higher histologic tumor grades were negatively associated with the mean ADC values is consistent with previous studies. According to Costantini and Razek et al., cases with a higher histological grade, larger tumor size, or abnormal axillary lymph nodes showed lower ADC values, which may indicate aggressive tumor behavior [24, 25]. In our study, tumor size was a factor associated with the ADC value according to the univariate analysis, but not the multivariate analysis. Our study has several limitations. First, each menstrual cycle group included a relatively small sample size. Larger sample sizes are needed to apply these results to the general patient population. Second, the images were interpreted by only two reviewers in consensus. Third, all of the included lesions were imaged after the biopsy, which may have influenced the ADC values. However, as described earlier, MRI was typically performed 2 weeks after the biopsy. This interval should be enough to resolve changes induced by the biopsy. Further, when we reviewed the MR images, none of the cases showed findings that could be attributed to the biopsy. Therefore, we suspect that imaging after the biopsy should not have affected our results. In conclusion, the menstrual cycle did not affect the ADC values of normal glandular tissue, the absolute and normalized ADC values of IDC, or the tumor detectability. However, postmenopausal women showed significantly lower ADC values compared to premenopausal women without any loss of tumor detectability, which was restored when lesions were normalized to normal glandular tissue.
Conflict of interest The authors declare that they have no conflicts of interest.
Breast Cancer Res Treat (2015) 149:751–759
References 1. Partridge SC, Mullins CD, Kurland BF et al (2010) Apparent diffusion coefficient values for discriminating benign and malignant breast MRI lesions: effects of lesion type and size. AJR Am J Roentgenol 194(6):1664–1673 2. Marini C, Iacconi C, Giannelli M, Cilotti A, Moretti M, Bartolozzi C (2007) Quantitative diffusion-weighted MR imaging in the differential diagnosis of breast lesion. Eur Radiol 17(10):2646– 2655 3. Woodhams R, Matsunaga K, Kan S et al (2005) ADC mapping of benign and malignant breast tumors. Magn Reson Med Sci 4(1): 35–42 4. Woodhams R, Matsunaga K, Iwabuchi K et al (2005) Diffusionweighted imaging of malignant breast tumors: the usefulness of apparent diffusion coefficient (ADC) value and ADC map for the detection of malignant breast tumors and evaluation of cancer extension. J Comput Assist Tomogr 29(5):644–649 5. Park MJ, Cha ES, Kang BJ, Ihn YK, Baik JH (2007) The role of diffusion-weighted imaging and the apparent diffusion coefficient (ADC) values for breast tumors. Korean J Radiol 8(5):390–396 6. Peters NH, Borel Rinkes IH, Zuithoff NP, Mali WP, Moons KG, Peeters PH (2008) Meta-analysis of MR imaging in the diagnosis of breast lesions. Radiology 246(1):116–124 7. Chen X, Li WL, Zhang YL, Wu Q, Guo YM, Bai ZL (2010) Meta-analysis of quantitative diffusion-weighted MR imaging in the differential diagnosis of breast lesions. BMC Cancer 10:693 8. Kuhl CK, Bieling HB, Gieseke J et al (1997) Healthy premenopausal breast parenchyma in dynamic contrast-enhanced MR imaging of the breast: normal contrast medium enhancement and cyclical-phase dependency. Radiology 203(1):137–144 9. Muller-Schimpfle M, Ohmenhauser K, Stoll P, Dietz K, Claussen CD (1997) Menstrual cycle and age: influence on parenchymal contrast medium enhancement in MR imaging of the breast. Radiology 203(1):145–149 10. O’Flynn EA, Morgan VA, Giles SL, deSouza NM (2012) Diffusion weighted imaging of the normal breast: reproducibility of apparent diffusion coefficient measurements and variation with menstrual cycle and menopausal status. Eur Radiol 22(7):1512– 1518 11. Clendenen TV, Kim S, Moy L et al (2013) Magnetic Resonance Imaging (MRI) of hormone-induced breast changes in young premenopausal women. Magn Reson Imaging 31(1):1–9 12. Abdel Razek AA, Gaballa G, Denewer A, Tawakol I (2010) Diffusion weighted MR imaging of the breast. Acad Radiol 17(3):382–386
759 13. Partridge SC, McKinnon GC, Henry RG, Hylton NM (2001) Menstrual cycle variation of apparent diffusion coefficients measured in the normal breast using MRI. J Magn Reson Imaging 14(4):433–438 14. Ei Khouli RH, Jacobs MA, Mezban SD et al (2010) Diffusionweighted imaging improves the diagnostic accuracy of conventional 3.0-T breast MR imaging. Radiology 256(1):64–73 15. Nissan N, Furman-Haran E, Shapiro-Feinberg M, Grobgeld D, Degani H (2014) Diffusion-tensor MR imaging of the breast: hormonal regulation. Radiology 271(3):672–680 16. Padhani AR, Liu G, Koh DM et al (2009) Diffusion-weighted magnetic resonance imaging as a cancer biomarker: consensus and recommendations. Neoplasia 11(2):102–125 17. Pereira FP, Martins G, Figueiredo E et al (2009) Assessment of breast lesions with diffusion-weighted MRI: comparing the use of different b values. AJR Am J Roentgenol 193(4):1030–1035 18. Allred DC, Harvey JM, Berardo M, Clark GM (1998) Prognostic and predictive factors in breast cancer by immunohistochemical analysis. Mod Pathol 11(2):155–168 19. Dietel M, Ellis IO, Hofler H et al (2007) Comparison of automated silver enhanced in situ hybridisation (SISH) and fluorescence ISH (FISH) for the validation of HER2 gene status in breast carcinoma according to the guidelines of the American Society of Clinical Oncology and the College of American Pathologists. Virchows Arch 451(1):19–25 20. Ramakrishnan R, Khan SA, Badve S (2002) Morphological changes in breast tissue with menstrual cycle. Mod Pathol 15(12):1348–1356 21. Erikson A, Andersen HN, Naess SN, Sikorski P, Davies CDL (2008) Physical and chemical modifications of collagen gels: impact on diffusion. Biopolymers 89(2):135–143 22. Ko ES, Han BK, Kim RB et al (2014) Apparent diffusion coefficient in estrogen receptor-positive invasive ductal breast carcinoma: correlations with tumor-stroma ratio. Radiology 271(1): 30–37 23. Woodhams R, Ramadan S, Stanwell P et al (2011) Diffusionweighted imaging of the breast: principles and clinical applications. Radiographics 31(4):1059–1084 24. Costantini M, Belli P, Rinaldi P et al (2010) Diffusion-weighted imaging in breast cancer: relationship between apparent diffusion coefficient and tumour aggressiveness. Clin Radiol 65(12): 1005–1012 25. Razek AA, Gaballa G, Denewer A, Nada N (2010) Invasive ductal carcinoma: correlation of apparent diffusion coefficient value with pathological prognostic factors. NMR Biomed 23(6):619– 623
123
