The diagnostic sensitivity of dynamic contrast-enhanced magnetic resonance imaging and breast-specific gamma imaging in women with calcified and non-calcified DCIS

Abstract

Background

Early detection of breast cancer reduces mortality. Therefore, diagnosis of ductal carcinoma in situ (DCIS) is important.

Purpose

To compare the sensitivities of dynamic contrast-enhanced magnetic resonance imaging (DCE-MRI) and breast-specific gamma imaging (BSGI) in pathologically proven calcified and non-calcified DCIS.

Material and Methods

Thirty-five patients with pathologically diagnosed DCIS from 1 June 2009 through 31 December 2011, underwent a protocol involving both breast MRI and BSGI. Each image was assessed by a separate dedicated breast radiologist. All lesions were divided into two groups; with or without microcalcifications on mammograms. In cases without microcalcifications, we recorded the mass, asymmetry, or negative findings on mammography. On MRI, the enhancement pattern was categorized as mass or non-mass-like enhancement. On BSGI, the uptake pattern was analyzed. The histopathological features of the lesions were obtained. Statistical analysis of the sensitivity of each modality was performed using McNemar’s test.

Results

Thirty-five women with a mean age of 48 years (range, 26–69 years) were enrolled in the study. The total sensitivities of MRI and BSGI in the 35 cases were 91.4% (32 of 35 DCIS) and 68.6% (24 of 35 DCIS), respectively. Eighteen cases with DCIS displayed microcalcifications on mammography, while 17 cases did not. Of these 17 cases without microcalcifications on mammography, 88.2% (15 of 17 DCIS) were detected by MRI and 52.9% (9 of 17 DCIS) by BSGI. Of 18 cases with microcalcifications on mammography, 94.4% (17 of 18 DCIS) were detected by MRI and 83.3% (15 of 19 DCIS) by BSGI.

Conclusion

MRI showed a higher sensitivity for the detection of calcified and non-calcified DCIS and is more helpful than BSGI in cases without microcalcifications on mammography.

Keywords

magnetic resonance imaging breast specific gamma imaging sensitivity microcalcification mammography

Introduction

Ductal carcinoma in situ (DCIS) is a breast malignancy, characterized pathologically by proliferation of malignant ductal epithelial cells in the lining of the terminal duct lobular unit without invasion through the basement membrane. Approximately 30–50% of DCIS cases progress to invasive breast cancer (1). Recently, the incidence of DCIS has been rising steadily due to more widespread use of screening mammography (2).

The vast majority of mammographically detected DCIS manifests with microcalcifications with coarse, heterogeneous, or fine pleomorphic morphology and segmental or linear branching distribution (3,4). However, not all DCIS is calcified. A low-grade DCIS lesion without necrosis is less likely to manifest with microcalcifications (5). Limitations in the sensitivity and specificity of screening mammography have led to the investigation of adjunct breast imaging modalities.

On breast magnetic resonance imaging (MRI), DCIS manifests with various findings, such as a mass with a wash-out or plateau pattern upon kinetic analysis or a non-mass-like enhancement (6,7). With the development of higher spatial resolution techniques, MRI has become the preferred tool for DCIS detection (8). Studies of MRI detection of breast cancer have reported high sensitivity for invasive cancer, but the sensitivity for DCIS has been investigated by few studies.

Regarding breast-specific gamma imaging (BSGI), malignant cells generally appear as “dark spots”. BSGI with a radiotracer uses high-resolution gamma cameras that are placed adjacent to the breast while in compression, similar to mammographic positioning. Recent studies have reported the potential of BSGI as a valuable adjunct imaging modality (9). However, limited data in the setting of DCIS exist. Current data show no statistically significant difference in the sensitivities of MRI and BSGI in DCIS (10).

The purpose of this study was to compare retrospectively the sensitivities of MRI and BSGI in pathologically diagnosed DCIS, in cases with or without microcalcification on mammography.

Material and Methods

Study participants

We reviewed retrospectively the records of patients diagnosed with pathologically proven DCIS of the breast. Our institutional review board approved the study protocol and waived the requirement for informed consent. Sixty-three patients with pathologically diagnosed pure DCIS based on final pathological reports from lumpectomy or mastectomy specimens, from 1 June 2009 through 31 December 2011 were included in the present study. We excluded 20 cases due to absence of BSGI, eight cases due to absence of MRI. Nine cases of DCIS with negative finding on mammography were detected by ultrasound. Finally, 35 women were included in the study.

Data collection

Mammography

All patients underwent mammography with standard imaging planes, including crandiocaudal (CC) and mediolateral oblique (MLO) views using the Senographe DS system (GE Healthcare, Milwaukee, WI, USA).

MRI

All patients underwent breast MRI on a 3.0-T MR system (Achieva; Philips Medical Systems, Best, The Netherlands) using a seven-channel dedicated breast coil in the prone position. MRI was performed using the following sequences. An axial, fat-suppressed, fast spin-echo T2-weighted image were obtained. Intravenous injection of gadolinium (0.1 mmol/kg body weight of Gd-DTPA; Magnevist, Bayer Schering Pharma AG, Berlin, Germany) was performed for dynamic contrast enhancement study. Pre- and postcontrast three-dimensional (3D) T1-weighted (T1W) gradient echo with fat saturation sequences were obtained in a dynamic fashion. Dynamic contrast-enhanced T1W images in the axial plane were obtained using a 3D radio-frequency spoiled gradient-echo sequence with a repetition time of 4.7 ms, an echo time of 2.3 ms (4.7/2.3), a flip angle of 10°, a field of view (FOV) of 340 mm, and an acquisition time of 1 min 17 s. The total acquisition time was <10 min. After imaging, postprocessing with subtraction was performed.

BSGI

BSGI (6800 Gamma Camera; Dilon Technologies, Newport, VA, USA) was performed after biopsy demonstrating DCIS to evaluate occult foci as well as to determine the extent of disease for surgical planning. BSGI was performed after injection of 925–1110 MBq 99mTc-sestamibi through the antecubital vein contralateral to the breast lesion. BSGI was performed 10 min after injection of the radioisotope. Patients were placed in a seated position, and CC and MLO images of the breasts were obtained. The acquisition time for each image was approximately 5 min, and >100,000 counts per image were defined as the minimal range.

Image interpretation

MR images, BSGI images, and mammographic images were reviewed retrospectively by two radiologists unaware of lesion pathology. Discrepancies were resolved by consensus.

The presence of existing microcalcifications on mammography was recorded and divided into two groups, with or without microcalcifications. Mammographic findings were scored as follows: negative (no microcalcification, mass, or asymmetry, score of 0), microcalcification (pleomorphic, linear branching or amorphous morphology, score of 1), mass (score of 2), and asymmetry (score of 3).

MRI findings were classified as follows: negative (score of 0), mass (score of 1), and non-mass like enhancement (score of 2). All studies were classified as positive or negative. A score of 0 was classified as negative and a score of 1 or 2 as positive.

BSGI images were classified as normal (score of 0) with no focal or diffuse uptake; benign (score of 1) with minimal patchy uptake; probably benign (score of 2), with scattered patchy uptake; probably abnormal (score of 3) with mild focal uptake; and abnormal (score of 4) with marked focal uptake. Patients with a score of 0, 1, or 2 were defined as negative, and those with a score of 3 or 4 were defined as positive.

Histopathology

Histopathological information regarding the lesions was obtained from the medical records. Histological sections were reviewed by an experienced breast pathologist on our institution. The pathologic diagnosis was used as the reference standard. Pathologic diagnoses of suspicious lesions were determined using breast tissue obtained from percutaneous biopsy retrieval by ultrasonographic guidance using a 14-gauge spring-loaded needle and subsequent surgical excision following cancer-positive biopsy. Breast-conserving surgery or mastectomy was performed in all patients.

Statistical analysis

To evaluate the diagnostic sensitivity of BSGI and MRI, statistical analysis using McNemar’s test was performed using the SPSS version 10.0 for Windows statistical software (SPSS Inc., Chicago, IL, USA).

Results

Thirty-five women (mean age, 48 years; range, 26–69 years) with biopsy-proven DCIS lesions were included in the study. Of these 35 women, 18 had pleomorphic, linear branching, or amorphous microcalcifications on mammography. Of 17 cases without microcalcifications, four displayed a mass, four had asymmetry, and nine had negative finding on mammography. On MRI, 15 cases displayed a mass, 17 cases had non-mass like enhancement, and three cases showed negative findings. On BSGI, seven cases had mild focal uptake, 17 cases had marked focal uptake, and 11 cases revealed negative finding. Seven DCIS lesions missed by mammography were detected by MRI and four DCIS lesions missed by mammography were detected by BSGI (Table 1).

Table 1.

Mammographic, MRI, and BSGI features of 35 cases diagnosed with DCIS.

			MRI			BSGI
			Negative	Mass	Non- mass-like enhancement	Negative	Minimal uptake	Scattered uptake	Mild focal uptake	Marked focal uptake
			3 (8.6)	15 (42.8)	17 (48.6)	5 (14.3)	3 (8.6)	3 (8.6)	7 (20)	17 (48.5)
Calcification (+)		18 (51.4)	1	7	10	2	1	4	11
Calcification (–)	Mass	4 (11.4)	3	1	1	2	1
	Asymmetry	4 (11.4)	1	3	1	3
	Negative	9 (25.7)	2	4	3	2	2	1	2	2

Data are numbers of patients with the percentage in parentheses.

BSGI, breast-specific gamma imaging; DCIS, ductal carcinoma in situ; MRI, magnetic resonance imaging.

Of 35 cases, MRI identified 32 cases with 91.4% sensitivity and BSGI identified 24 cases with 68.6% sensitivity. DCIS was detected only by MRI in 10 cases. In two cases, DCIS was detected only by BSGI. In 22 cases, both MRI and BSGI detected DCIS. In one case, DCIS was not detected by either modality. The total sensitivity of MRI of 91.4% in 35 patients was significantly higher than that of BSGI (P < 0.05; Table 2).

Table 2.

Sensitivities of MRI and BSGI for 35 cases.

		MR lesions		Total
		Positive	Negative	Total
Scan lesion	Positive	22 (62.9)	2 (5.7)	24 (68.6)
Scan lesion	Negative	10 (28.6)	1 (2.9)	11 (31.4)
Total		32 (91.4)	3 (8.6)	35 (100)

Data are numbers of patients with the percentages in parentheses.

P < 0.05.

Of 18 cases with microcalcifications on mammography, MRI identified 17 cases (94.4%) and BSGI (83.3%) identified 15 cases. In three cases, DCIS was detected only by MRI. In one case, DCIS was detected only by BSGI. In 14 cases, DCIS was detected by both modalities (Fig. 1). All 18 cases with microcalcifications were detected by MRI or BSGI. The sensitivity of MRI (94.4%) was not significantly higher than that of BSGI (83.3%; P = 0.625; Table 3).

Fig. 1.

A 61-year-old woman with DCIS in the left breast. (a) Magnified view shows pleomorphic microcalcifications at the mid portion of the left breast. (b) Breast-specific gamma imaging (BSGI) with crandiocaudal (CC) and mediolateral oblique (MLO) views show focal increased radiotracer uptake in the mid-outer portion of the left breast. (c) Sagittal contrast-enhanced MR images show a region of non-mass-like enhancement with clumped internal enhancement, measuring ∼5 × 3 cm. Enhancement extends to the nipple. Histopathology demonstrated DCIS with 4 × 2.5 cm extent. The extent of pathology was similar to that of MRI. Additionally, the tumor invaded to the nipple.

Table 3.

Sensitivities of MRI and BSGI for 18 cases with microcalcifications on mammography.

		MR lesions		Total
		Positive	Negative
Scan lesion	Positive	14 (77.8)	1 (5.6)	15 (83.3)
Scan lesion	Negative	3 (16.7)	0 (0)	3 (16.7)
Total		17 (94.4)	1 (5.6)	18 (100)

Data are numbers of patients with the percentages in parentheses.

P = 0.625.

Of 17 cases without microcalcifications on mammography, MRI identified 15 cases (88.2%) and BSGI identified nine (52.9%). In seven cases, DCIS was detected only by MRI (Fig. 2). In one case, DCIS was detected only by BSGI (Fig. 3). In eight cases, DCIS was detected by both modalities. In one case, DCIS was not detected by either modality. The sensitivity of MRI (88.2%) was of borderline statistical significance compared with sensitivity of BSGI (52.9%; P = 0.070; Table 4).

Fig. 2.

A 42-year-old woman with DCIS in the left breast. (a) Mammogram revealed negative finding. (b) BSGI images show heterogeneous radiotracer uptake in both breasts. No definite focal area of increased radiotracer uptake is evident. (c) MRI demonstrates non-mass-like enhancement in the upper inner portion of the left breast.

Fig. 3.

A 40-year-old woman with DCIS in the right breast. (a) Mammogram shows negative finding. (b) BSGI images demonstrates focal tracer uptake in the inner portion of the right breast. (c) Axial contrast-enhanced MR image shows no definite abnormal enhancing lesion.

Table 4.

Sensitivities of MRI and BSGI for 17 cases without microcalcifications on mammography.

		MR lesions		Total
		Positive	Negative	Total
Scan lesion	Positive	8 (47.1)	1 (5.9)	9 (52.9)
Scan lesion	Negative	7 (41.2)	1 (5.9)	8 (47.1)
Total		15 (88.2)	2 (11.8)	17 (100)

Data are numbers of patients with the percentages in parentheses.

P = 0.070.

Of 18 cases with microcalcifications on mammography, 10 cases revealed high grade, five were intermediate grade, and three were low grade. Of 17 cases without microcalcifications on mammography, only one was high grade, six were intermediate grade, and 10 were low grade.

The rate of non-mass-like enhancement was higher in high grade cases (73%) than low grade cases (46%) on MRI. In low grade lesions, six had non-mass-like enhancement, and seven had mass. In intermediate lesions, three had negative findings, three had non-mass-like enhancement, and five had mass. In high grade lesions, eight had non-mass-like enhancement, and three had mass (Table 5).

Table 5.

Correlation of lesion type with contrast-enhanced MRI findings (mass or non-mass-like enhancement [NMLE]) and histopathological grade.

Low	Intermediate	High
Negative	0	3 (27)	0
NMLE	6 (46)	3 (27)	8 (73)
Mass	7 (54)	5 (46)	3 (27)

Data are numbers of patients with the percentages in parentheses.

Discussion

The incidence of DCIS has been rising steadily. Currently, screening mammography is the most useful imaging modality for detection of DCIS. However, the reported sensitivity of mammography varies, from 22% to 86% (11,12). For this reason, investigators have sought adjunct breast imaging modalities. The sensitivity of mammography for detection of non-calcified DCIS has been reported as 50% (36/71), whereas that for detection of calcified DCIS was 100% (146/146) (13). Our results are in agreement with that report. The sensitivity of mammography for detection of non-calcified DCIS was 47% (8/17), whereas that for detection of calcified DCIS was 100% (18/18).

We compared the sensitivities of MRI and BSGI for detection of DCIS. The sensitivity of MRI was significantly higher than that of BSGI (91.4% vs. 68.6%, respectively; P < 0.05). The MRI sensitivity is similar to that reported previously (7,14,15). The BSGI sensitivity for the detection of DCIS (68.6%) was lower than those reported elsewhere (89% and 75%, respectively (14,16)). Notably, previous studies did not show statistical significance (10,14). However, our data reached statistical significance, indicating that all cases were pure DCIS.

In cases with microcalcifications on mammography, the sensitivity of MRI (94.4%) was not significantly higher than that of BSGI (94.4%; P = 0.625).

In cases without microcalcifications on mammography, the sensitivity of MRI of 88.2% displayed borderline statistical significance (88.2%; P = 0.070) compared to that of BSGI. To our knowledge, no previous study has evaluated sensitivity divided into two groups according to existing microcalcifications on mammography. Additionally, the first comparison of two imaging modalities in cases without microcalcifications on mammography.

In cases detected by both modalities, MRI was superior to BSGI in terms of determining the malignant lesions. Additionally, MRI identified tumor extension to the nipple. Thus, this modality may assist determination of the appropriate surgical treatment (breast-conserving surgery or mastectomy) (Fig. 1).

Microcalcifications are the most common finding, but not all DCIS is calcified. The relationship between pathologic grade and mammographically detected microcalcifications remains controversial. Some studies reported a significant correlation (4,17,18). High-grade lesions had microcalcifications on mammography, although most cases with microcalcifications were not high-grade lesions. Thus, the presence of microcalcifications on mammography may correlate with the pathologic grade.

Several authors have analyzed difference of MRI features between high-grade and non-high-grade DCIS. Some studies reported that the rate of non-mass-like enhancement increased according to the nuclear grade (19,20). In our study, most of high grade DCIS presented as non-mass-like enhancement (73%) whereas low grade DCIS presented as a mass (54%) (Table 5). Kim et al. (21) showed that the dominant MRI findings of pure DCIS were non-mass-like enhancement and plateau curve pattern. However, several studies have reported that no correlation existed between MRI enhancement kinetics and nuclear grade of DCIS (6,22). Previous studies showed that measurement of the size of DCIS by MRI and pathology had a moderate correlation (19,23,24).

BSGI exhibits greater specificity than MRI (25). We did not evaluate the specificity, negative predictive value, and positive predictive value because all cases were confirmed as DCIS: no negative case was included in this study.

The present study possessed several limitations. First, this study was of a retrospective design, which could lead to selection bias. Second, we reviewed a small number of DCIS lesions; this may result in a low statistical power. Larger studies should be performed to determine more accurately the sensitivities of both modalities. Third, BSGI data in a previous report were reviewed by a nuclear radiologist. Fourth, interpretation of MRI findings in cases with small DCIS presenting as non-mass-like enhancement have a considerable inter-observer variability. We resolved the discrepancy by consensus but other radiologic studies may demonstrate such discrepancy in our opinion.

In conclusion, MRI showed a higher sensitivity for the detection of calcified and non-calcified DCIS and is more helpful than BSGI in cases without microcalcifications on mammography.

Footnotes

Funding

This study was supported by a grant of Ewha Womans University [5051085].

References

Recht

Rutgers

Fentiman

. The Fourth EORTC DCIS Consensus Meeting (Chateau Marquette, Heemskerk, The Netherlands, 23-24 January 1998)—conference report. Eur J Center 1998; 34: 1664–1669.

Sumner

3rd Koniaris

Snell

. Results of 23,810 cases of ductal carcinoma in situ. Ann Surg Oncol 2007; 14: 1638–1643.

Stomper

Connolly

Meyer

. Clinically occult ductal carcinoma in situ detected with mammography: analysis of 100 cases with radiologic-pathologic correlation. Radiology 1989; 172: 235–241.

Evans

Pinder

Wilson

. Ductal carcinoma in situ of the breast: correlation between mammographic and pathologic findings. Am J Roentgenol 1994; 162: 1307–1311.

Yang

Tse

. Sonographic, mammographic, and histopathologic correlation of symptomatic ductal carcinoma in situ. Am J Roentgenol 2004; 182: 101–110.

Jansen

Newstead

Abe

. Pure ductal carcinoma in situ: kinetic and morphologic MR characteristics compared with mammographic appearance and nuclear grade. Radiology 2007; 245: 684–691.

Rosen

Smith-Foley

de Martini

. BI-RADS MRI enhancement characteristics of ductal carcinoma in situ. Breast J 2007; 13: 545–550.

Lehman

. Magnetic resonance imaging in the evaluation of ductal carcinoma in situ. J Natl Cancer Inst Monogr 2010; 2010: 150–151.

Brem

Floerke

Rapelyea

. Breast-specific gamma imaging as an adjunct imaging modality for the diagnosis of breast cancer. Radiology 2008; 247: 651–657.

10.

Brem

Fishman

Rapelyea

. Detection of ductal carcinoma in situ with mammography, breast specific gamma imaging, and magnetic resonance imaging: a comparative study. Acad Radiol 2007; 14: 945–950.

11.

Menell

Morris

Dershaw

. Determination of the presence and extent of pure ductal carcinoma in situ by mammography and magnetic resonance imaging. Breast J 2005; 11: 382–390.

12.

Ernster

Ballard-Barbash

Barlow

. Detection of ductal carcinoma in situ in females undergoing screening mammography. J Natl Cancer Inst Monogr 2002; 94: 1546–1554.

13.

Mun

Shin

Kim

. Screening-detected calcified and non-calcified ductal carcinoma in situ: Differences in the imaging and histopathological features. Clin Radiol 2013; 68: e27–e35.

14.

Keto

Kirstein

Sanchez

. MRI versus breast-specific gamma imaging (BSGI) in newly diagnosed ductal cell carcinoma in situ: a prospective head-to-head trial. Ann Surg Oncol 2012; 19: 249–252.

15.

Berg

Gutierrez

Ness-Aiver

. Diagnostic accuracy of mammography, clinical examination, US, and MR imaging in preoperative assessment of breast cancer. Radiology 2004; 233: 830–849.

16.

Papantoniou

Valsamaki

Sotiropoulou

. Increased breast density correlates with the proliferation-seeking radiotracer (99m)Tc(V)-DMSA uptake in florid epithelial hyperplasia and in mixed ductal carcinoma in situ with invasive ductal carcinoma but not in pure invasive ductal carcinoma or in mild epithelial hyperplasia. Mol Imaging 2011; 10: 370–376.

17.

Barreau

de Mascarel

Feuga

. Mammography of ductal carcinoma in situ of the breast: review of 909 cases with radiographic-pathologic correlations. Eur J Radiol 2005; 54: 55–61.

18.

Kim

Kim do

. Noncalcified ductal carcinoma in situ: imaging and histologic findings in 36 tumors. J Ultrasound Med 2009; 28: 903–910.

19.

Baur A, Bahrs SD, Speck S, et al. Breast MRI of pure ductal carcinoma in situ: Sensitivity of diagnosis and influence of lesion characteristics. Eur J Radiol 2013;82:1731–1737.

20.

Liu

Peng

. MRI morphological classification of ductal carcinoma in situ (DCIS) correlating with different biological behavior. Eur J Radiol 2012; 81: 214–217.

21.

Kim

Son

Youk

. MRI findings of pure ductal carcinoma in situ: Kinetic characteristics compared according to lesion type and histopathologic factors. Am J Roentgenol 2011; 196: 1450–1456.

22.

Chan S, Chen JH, Agrawal G, et al. Characterization of pure ductal carcinoma in situ on dynamic contrast-enhanced MR imaging: Do nonhigh grade and high grade show different imaging features? J Oncol 2010; Epub 2010 Sep 21.

23.

Huang

Cheung

. MRI findings of cancers preoperatively diagnosed as pure dcis at core needle biopsy. Acta Radiol 2011; 52: 1064–1068.

24.

Marcotte-Bloch

Balu-Maestro

Chamorey

. MRI for the size assessment of pure ductal carcinoma in situ (DCIS): A prospective study of 33 patients. Eur J Radiol 2011; 77: 462–467.

25.

Brem

Petrovitch

Rapelyea

. Breast-specific gamma imaging with 99mTc-Sestamibi and magnetic resonance imaging in the diagnosis of breast cancer—a comparative study. Breast J 2007; 13: 465–469.