Article

Diagnosis of the Intraductal Component of Invasive Breast Cancer

Assessment With Mammography and Sonography
Doo Kyoung Kang, MD, Gyeong Sik Jeon, MD, Hyunee Yim, MD, Yong Sik Jung, MD
Objective. The aim of this study was to investigate mammographic and sonographic features and their sensitivities for depiction of the intraductal component associated with invasive ductal carcinoma (IDC). Methods. During a 1-year period, 132 patients with IDC underwent surgical treatment. All patients underwent mammography and high-resolution sonography, and the findings were reported according to the American College of Radiologys Breast Imaging Reporting and Data System lexicon. Tumors were classified as pure IDC and IDC with an intraductal component by histopathologic evaluation. We compared mammographic and sonographic features between the above 2 groups and attempted to correlate them with histopathologic findings. We also investigated separate and combined sensitivities, specificities, and accuracies of both mammography and breast sonography for showing intraductal components. Finally, imaging measurements were compared with pathologic measurements. Results. One hundred four (79%) of the 132 IDCs contained an intraductal component. Patients with IDC with an intraductal component showed calcifications on mammography and showed an echogenic halo, duct dilatation, calcifications, and increased vascularity in surrounding tissue on sonography more frequently than patients with pure IDC. The sensitivities of mammography, sonography, and their combined assessment for detection of an intraductal component were 55%, 80%, and 86%, respectively. The combined assessment (r = 0.90) measured the extent of the tumor more accurately than mammography (r = 0.71) or sonography (r = 0.79) separately. Conclusions. Combined assessment with mammography and sonography offers more accurate information for the presence of an intraductal component and the extent of a tumor than each separate assessment. Key words: breast neoplasms; breast sonography; mammography; pathologic examination.

Abbreviations BI-RADS, Breast Imaging Reporting and Data System; DCIS, ductal carcinoma in situ; EIC, extensive intraductal component; IDC, invasive ductal carcinoma; MRI, magnetic resonance imaging

Received May 10, 2007, from the Departments of Diagnostic Radiology (D.K.K., G.S.J.), Pathology (H.Y.), and General Surgery (Y.S.J.), Ajou University, School of Medicine, Suwon, Korea. Revision requested June 12, 2007. Revised manuscript accepted for publication July 11, 2007. Address correspondence to Doo Kyoung Kang, MD, Department of Diagnostic Radiology, Ajou University, School of Medicine, San 5, Woncheondong, Yeongtong-gu, Suwon, Kyongi-do 442-749, Korea. E-mail: kdklsm@ajou.ac.kr

uctal carcinoma is the most common type of malignancy arising from the breast parenchyma. Examination of invasive ductal carcinoma (IDC) reveals an intraductal component in 50% to 80% of pathologic studies.1,2 The presence of an intraductal component is a very important risk factor for local recurrence in breast-conserving therapy because tumors with an intraductal component can leave residual islands of intraductal carcinoma.35 Actually, 26% of patients with an extensive intraductal component (EIC) had a local recurrence compared with 7% of patients without an EIC.6 However, when a microscopic clear margin was obtained after breast-conserving surgery, there was no difference in local recurrence between EIC-positive and -negative cancer.7,8 Therefore, our ultimate goal is to accurately detect intraductal components and to completely remove them for achieving clear margins.

2007 by the American Institute of Ultrasound in Medicine J Ultrasound Med 2007; 26:15871600 0278-4297/07/$3.50

Diagnosis of the Intraductal Component of Invasive Breast Cancer

The mammographic diagnosis of an intraductal component is based mainly on the presence of calcifications. Calcifications outside or adjacent to a mass or malignant calcifications without a tumor shadow are characteristic of intraductal components.911 Especially, when greater than 3 cm in extent, they are suggestive of a lesion with an EIC.9 Cancers showing calcifications on mammograms have been shown to be associated with an EIC in 22% to 63% of cases.912 However, mammography is of little value in cases without calcifications: mammographic detection of intraductal carcinomas without calcifications may be quite difficult, especially in dense breasts; therefore, approximately 23% of intraductal carcinomas are not visible on mammography.1315 On the other hand, it is easier to visualize noncalcified intraductal carcinomas than calcified intraductal carcinomas on sonography.16 Sonography can also show the intraductal component associated with invasive carcinoma. Sonographic features that characterize intraductal components of cancer are a dilated duct (or a tubular hypoechoic structure) radiating from the tumor and the presence of satellite lesions (or daughter tumors) around the main tumor.2,11,17 However, the sensitivity of sonography for detection of intraductal components varies from 21% to 89%.1012,17 The results depend mainly on the case selection method, the investigators skill, and the spatial resolution of the equipment.18 Furthermore, the effectiveness of sonography is partly lessened because of its limited capability to depict microcalcifications, although some investigators have reported that sonography depicts clustered microcalcifications in breast cancers.1921 Therefore, mammography and sonography appear to be complementary. The aims of this study, therefore, were to assess the mammographic and sonographic features of the intraductal component associated with IDC and to investigate separate and combined sensitivities of mammography and breast sonography for showing the intraductal component in patients with IDC. In addition, to our best knowledge, a study with application of the American College of Radiologys Breast Imaging Reporting and Data System (BI-RADS) ultrasound lexicon has not been reported previously.22
1588

Materials and Methods

Study Population From March 2006 to February 2007, 4191 consecutive patients underwent both mammography and sonography for clarification of suspicious breast findings. Three hundred ninety-four patients were recommended for tissue confirmation of suspicious lesions (BI-RADS category 4 or 5); 373 of them actually underwent mammographically guided needle localization with excision or sonographically guided core needle biopsy, and 187 patients were confirmed as having IDC or ductal carcinoma in situ (DCIS). Finally, 175 patients were surgically treated, except 12 patients who were scheduled for neoadjuvant chemotherapy or palliative treatment. A total of 132 patients were confirmed as having IDC and were enrolled in this study, excluding the remaining 43 patients with DCIS. Our Institutional Review Board did not require its approval or informed consent for the retrospective evaluation of patients records and data. The mean age of the study population was 46 years (range, 2279 years). Thirty-three patients (32%) underwent breast-conserving surgery, and 71 patients (68%) underwent mastectomy, including modified radical mastectomy. The diagnoses of the invasive components included ductal carcinoma, not otherwise specified (n = 126); mucinous carcinoma (n = 4); and medullary carcinoma (n = 2). Mammographic and Sonographic Examinations Mammograms from all 132 patients were obtained with a Senographe DMR mammography unit (GE Healthcare, Milwaukee, WI). Using the BI-RADS mammography lexicon, 1 experienced breast radiologist (D.K.K.) prospectively interpreted the mammograms. At that time, we regarded a mass or focal asymmetry as soft tissue density and then simply classified the mammographic findings as calcifications alone, calcifications with soft tissue density, and soft tissue density alone for statistical analysis. Whole-breast sonography was performed by a radiologist (D.K.K.) with 8 years of experience in breast sonography using an Acuson Sequoia ultrasound system (Siemens Medical Solutions,
J Ultrasound Med 2007; 26:15871600

Kang et al

Mountain View, CA) with an 8- to 13-MHz linear array transducer. Radial scanning techniques and careful analysis of the ductal structures were used to evaluate the extent of the tumors. For evaluation of the capability of US to depict microcalcifications, breast sonography was initially performed without information on mammographic findings. Thereafter, sonography was again performed immediately with careful correlation with mammographic findings, and the findings were finally reported according to the BI-RADS ultrasound lexicon.22 Histopathologic Evaluation In the Department of Pathology, surgical specimens were serially sliced at 5-mm intervals, prepared as paraffin-embedded sections, and stained with hematoxylin-eosin. The sections stained in each case were reviewed by 1 experienced pathologist (H.Y.) without knowledge of the results of the mammographic and sonographic examinations. The specimens were evaluated according to the following histopathologic features: histologic type, Black nuclear grade, modified Bloom-Richardson histologic grade, and the presence of hormone receptors (estrogen receptor, progesterone receptor, and c-ErbB2). Tumors were classified into 2 groups, based on the absence or presence of in situ component: pure IDC with no intraductal component and IDC with an intraductal component. In the IDC with an intraductal component group, EIC referred to invasive breast cancer wherein at least 25% of the tumor was composed of an intraductal component, and the intraductal component extended beyond the invasive tumor.23 In addition to size measurement of the invasive component (T status), the maximum extent including the intraductal component around the main tumor was also measured. The overall stage of cancer was classified according to American Joint Committee on Cancer TNM staging of breast cancer.24 Data Analysis We retrospectively analyzed the data of each examination in our study population. First, we compared the mammographic and sonographic features between the 2 groups and attempted to correlate them with the histopathologic findings.
J Ultrasound Med 2007; 26:15871600

Second, we investigated the separated and combined sensitivities, specificities, and accuracies of mammography and breast sonography for showing the intraductal component associated with IDC. Referring to earlier reports,9,10,12 calcifications around the mass or malignant calcifications without a mass on mammography and duct dilatation radiating from the tumor or satellite lesions around the main tumor on sonography were considered suspicious for an intraductal component of IDC (Figure 1). We also assessed the diagnostic performance, using additional sonographic features that were significantly different between pure IDC and IDC with an intraductal component. Third, we compared the imaging tumor size with the pathologic tumor size whenever it was possible to preoperatively measure them on both mammography and sonography. Statistical Analysis SPSS version 11.9 (SPSS Inc, Chicago, IL) and Microsoft Excel 2000 (Microsoft Corporation, Redmond, WA) software packages were used for the statistical analyses. Statistical comparisons of the characteristics between the subgroups of cases were performed with the Pearson 2 statistic or, when appropriate, the Fisher exact test. Separate and combined sensitivities, specificities, and accuracies for both mammography and breast sonography for showing intraductal components were calculated, and these diagnostic performances were compared among the above 3 methods by a McNemar test. We also used the Student t test and linear regression analysis for comparison of imaging and pathologic sizes. P < .05 was considered indicative of a significant difference. Correlation coefficients between 0 and 0.2 were regarded as no correlation, between 0.2 and 0.4 as a weak correlation, between 0.4 and 0.7 as a moderate correlation, and greater than 0.7 as a strong correlation.

Results
Pathologic Characteristics One hundred four (79%) of the 132 IDCs contained an intraductal component. Fifty-two tumors contained an intraductal component occupying 1% to 24% of the area of the tumor,
1589

Diagnosis of the Intraductal Component of Invasive Breast Cancer

and 52 tumors contained an EIC. The remaining 28 tumors, which were pure IDCs, did not contain any intraductal component in or around the main tumor. There were no statistical differences
Figure 1. Infiltrating ductal carcinoma with an EIC in a 39-year-old woman. A, Mediolateral oblique mammogram shows a 2.5-cm irregularly shaped hyperdense mass (short arrows) with a spiculated margin in the left breast. There are clustered pleomorphic microcalcifications (long arrows) within the mass. B, Radial sonogram shows a 2.8-cm microlobular hypoechoic mass in the right breast. The mass has an echogenic halo, associated duct dilatation (long arrow), and associated satellite nodules (short arrows) around the mass. These masses were confirmed as IDC with EICs. The maximum size of the malignant process was 3.5 cm on histopathologic analysis. The extent of the malignancy was consequently underestimated on both mammography and sonography.

in patient age (P = .386), symptoms (P = .059), tumor size (T status, P = .346), regional lymph node status (P = .515), cancer stage (P = .811), nuclear grade (P = .658), and histologic grade (P = .189) between patients with pure IDC and patients with IDC and an intraductal component (Table 1). Mammographic and Sonographic Features The most common mammographic feature of pure IDC was soft tissue density alone (86%; Figure 2), followed by calcifications with soft tissue density (7%) and negative findings (7%), whereas the most common feature of IDC with an intraductal component was calcifications with soft tissue density (48%; Figure 3), followed by soft tissue density alone (42%), calcifications alone (7%), and negative findings (3%; Table 2). Invasive ductal carcinoma with an intraductal component showed calcifications on mammography more frequently (55%) than pure IDC (7%). As for the mass, IDC with an intraductal component showed a more irregular shape (Figure 4) than pure IDC (P < .001), but margin and density were not significantly different between the 2 groups (Table 3). We could not obtain statistically significant results on calcifications because only a small number of cases with pure IDC revealed calcifications on mammography. All cases could be detected by sonography. On the other hand, sonography with knowledge of mammographic findings showed calcifications in 48 (81%) of 59 cases in which calcifications were revealed on mammography, whereas initial sonography without mammographic information showed calcifications in only 34 (58%) cases. Sonographic findings that showed significant differences between pure IDC and IDC with an intraductal component included the lesion boundary (P = .014), duct dilatation (P = .003), calcifications (P < .001), and vascularity (P = .002). Pure IDC showed an abrupt interface (Figure 2) and increased vascularity in or immediately adjacent to lesions more frequently, whereas IDC with an intraductal component showed an echogenic halo (Figure 4), duct dilatation (Figure 5), calcifications (Figure 6), and increased vascularity in surrounding tissue more frequently (Figure 7 and Table 4).
J Ultrasound Med 2007; 26:15871600

1590

Kang et al

Detection of Intraductal Components On mammography, we could detect intraductal component in 57 of 104 IDCs with an intraductal components. The sensitivity, specificity, and accuracy of mammography for detection of an intraductal component were 55%, 93%, and 63%, respectively (Figure 8). In cases in which duct dilatation and satellite lesions on sonography were considered suspicious for an intraductal component, we could detect intraductal components in 65 cases. On the other hand, when microcalcifications inside or outside mass and increased vascularity in surrounding tissue were included in the sonographic criteria for an intraductal component, we could detect intraductal components in 83 cases. Therefore, according to the latter sonographic criteria for detection of an intraductal component, the overall sensitivity, specificity, and accuracy were 80%, 79%, and 80%. In comTable 1. Characteristics of 132 Patients With IDC
Characteristic Pure IDC (n = 28)

bined assessment with both mammography and sonography, we were able to detect intraductal components in 89 cases, and the sensitivity, specificity, and accuracy were 86%, 79%, and 84% (Figure 8). According to the McNemar test, the sensitivity and accuracy of the combined assessment were significantly higher than those of mammography or sonography separately, whereas the specificity was not different significantly. The sensitivity and accuracy of sonography using traditional criteria of duct dilatation and satellite lesions were not significantly different compared with those of mammography (Figure 8). However, when the presence of calcifications and diffusely increased vascularity in the surrounding tissue were used as additional criteria for an intraductal component, the sensitivity and accuracy of sonography were significantly higher than those of mammography.

IDC With Intraductal Component (n = 104)

Mean age, y Symptom, n (%) Asymptomatic (n = 36) Palpable mass (n = 93) Nipple discharge (n = 3) Surgery, n (%) BCS (n = 45) Mastectomy (n = 87) Tumor size, n (%) T1 (n = 55) T2 (n = 69) T3 (n = 6) T4 (n = 2) Regional lymph node status, n (%) N0 (n = 71) N1 (n = 39) N2 (n = 15) N3 (n = 7) Stage, n (%) I (n = 38) II (n = 70) III (n = 24) Nuclear grade, n (%) 1 (n = 60) 2 (n = 57) 3 (n = 8) Histologic grade, n (%) 1 (n = 19) 2 (n = 48) 3 (n = 59) BCS indicates breast-conserving surgery.

47.5 3 (10.7) 24 (85.7) 1 (3.6) 12 (42.9) 16 (57.1) 13 12 2 1 16 9 1 2 (46.4) (42.9) (7.1) (3.6) (57.1) (32.1) (3.6) (7.1)

45.6 33 (31.7) 69 (66.3) 2 (1.9) 33 (31.7) 71 (68.3) 42 57 4 1 55 30 14 5 (40.4) (54.8) (3.9) (1.0) (52.9) (28.8) (13.5) (4.8)

.386 .059

.270

.346

.515

9 (32.1) 15 (53.6) 4 (14.3) 13 (46.4) 9 (32.1) 2 (7.1) 6 (21.4) 6 (21.4) 12 (42.9)

29 (27.9) 55 (52.9) 20 (19.2) 47 (45.2) 48 (46.2) 6 (5.8) 13 (12.5) 42 (40.4) 47 (45.2)

.811

.658

.189

J Ultrasound Med 2007; 26:15871600

1591

Diagnosis of the Intraductal Component of Invasive Breast Cancer

In 51 (49%) of 104 cases, both mammographic and sonographic findings were positive (Figure 3). In 6 (6%) of 104 cases, mammographic findings were positive, whereas sonographic findings were false-negative. In 32 (31%) of 104 cases, sonographic findings were positive, whereas mammographic findings were false-negative
Figure 2. Pure IDC in a 49-year-old woman. A, Mediolateral oblique mammogram shows a hyperdense mass (arrow) with an obscured margin in the right breast. Calcifications are not shown around the mass. The mass measured 3 cm in maximum dimension. B, Transverse sonogram shows an oval hypoechoic mass (arrow) with circumscribed margins measuring 2.6 cm. Duct dilatation or satellite lesions are not shown around the mass. The mass was confirmed as IDC without an intraductal component. Histopathologic examination of the mass indicated a high nuclear grade and a poorly differentiated histologic grade. The mass measured 2.5 cm on pathologic examination and was consequently overestimated on mammography.

(Figure 5). In 15 (14%) of 104 cases, the intraductal component was missed by both mammography and sonography (Figure 4). In the combined assessment of all 132 invasive carcinomas, we misdiagnosed 6 (21%) of 28 cases of pure IDC as IDC with an intraductal component. These falsepositive cases were revealed as clustered puncFigure 3. Infiltrating ductal carcinoma with an intraductal component in a 53-year-old woman. A, Mediolateral oblique mammogram shows a 3-cm oval hyperdense mass with microlobulated margins in the right breast. There are segmental pleomorphic microcalcifications (arrows) within and around the mass. B, Radial sonogram shows a 3.2-cm oval hypoechoic mass with microlobulated margins and calcifications (long arrow) within the mass. There is duct dilatation (short arrow) extending from the main tumor. The mass was confirmed as IDC containing an intraductal component. The malignant process measured 4 cm on histopathologic examination and was underestimated on both mammography and sonography.

1592

J Ultrasound Med 2007; 26:15871600

Kang et al

tate (n = 1) and amorphous (n = 1) calcifications around the mass on mammography and duct dilatation (n = 4) around the main tumor on sonography. However, they were all confirmed as fibrocystic changes, such as epithelial hyperplasia or adenosis. Measurement of Lesion Extent In the cases of pure IDC (n = 23), the mean size of the lesions SD was 26.4 12.2 mm (range, 1560 mm). Mammography underestimated the size in 8 cases (range, 125 mm) and overestimated it in 7 cases (range, 212 mm). Sonography underestimated the size in 10 cases (range, 110 mm) and overestimated it in 12 cases (range, 112 mm). In the cases of pure IDC, the size measured by mammography (r = 0.80; P < .01) and sonography (r = 0.91; P < .01) showed a strong correlation with pathologic size. In the cases of IDC with an intraductal component (n = 77), the mean size of the lesions was 26.6 12.3 mm (range, 570 mm). Mammography underestimated the size in 39 cases (range, 130 mm) and overestimated it in 29 cases (range, 125 mm). Sonography underestimated the size in 36 cases (range, 117 mm) and overestimated it in 39 cases (range, 119 mm; Table 5). The size measured by mammography (r = 0.69; P < .01) and sonography (r = 0.75; P < .01) showed a moderate correlation with the pathologic size. Cases containing an EIC showed the poorest correlation on both mammography (r = 0.47; P < .01) and sonography (r = 0.58; P < .01). The size measured by combined assessment of mammography and sonography showed the highest correlation in all pure IDC (r = 0.97; P < .01) and IDC with an intraductal component (r = 0.88; P < .01) compared with separate mammographic or sonographic assessment. Cases containing an EIC showed a strong correlation (r = 0.80; P < .01) between imaging and pathologic tumor sizes (Table 6).

dence of an EIC has previously been reported to be 13% to 39%. Recently, Elling et al25 retrospectively analyzed 250 resected surgical specimens of invasive breast cancer to determine the presence of intraductal components within the primary tumors and found that DCIS of varying extent was identified in addition to the invasive carcinoma in 127 (50.8%) of the 250 specimens. In our study, 104 (79%) of 132 IDCs contained an intraductal component, and 52 cases (39%) were EIC positive. The detection of an intraductal component is clinically important because an intraductal component associated with invasive cancer is one of the major reasons for a positive margin and recurrence after breast conservation surgery.4 Therefore, therapeutic management may differ depending on the presence and extent of an intraductal component. An earlier study actually suggested that computed tomographic imageguided lumpectomy diminished the positive surgical margin rate to 21% compared with 42% of cases without computed tomographic imageguided lumpectomy (P = .0055).26 The mammographic diagnosis of intraductal components is based mainly on the presence of calcifications.11 However, approximately 23% of intraductal carcinomas do not reveal calcifications on mammography.13,14 In cases without calcifications, mammography is of little value in detection of intraductal carcinomas. In our cases, 47 (45%) of 104 IDCs with an intraductal component, especially 16 (31%) of 52 cases containing an EIC, did not show calcifications on mammography. Among these 47 cases without

Table 2. Correlation of Mammographic and Histopathologic Findings in IDC

Finding Pure IDC (n = 28) IDC With Intraductal Component (n = 104)

Discussion
Invasive ductal carcinoma is the most common type of breast malignancy and often has an intraductal component that is considered a noninvasive component of the tumor. An intraductal component associated with IDC has been found in 50% to 80% of cases.1,2 Particularly, the inciJ Ultrasound Med 2007; 26:15871600

Calcifications only, n (%) Calcifications with soft tissue density, n (%) Calcifications + mass Calcifications + asymmetry Soft tissue density only, n (%) Mass Asymmetry Negative, n (%) Total, n (%)

0 2 (7) 2 0 24 (86) 21 3 2 (7) 28 (100)

7 (7) 50 (48) 39 11 44 (42) 38 6 3 (3) 104 (100)

1593

Diagnosis of the Intraductal Component of Invasive Breast Cancer

calcifications, 32 (76%) showed calcifications on sonography. In contrast, sonography is less sensitive than mammography in showing calcifications. The low capability to visualize microcalcifications remains a major limitation of sonography for diagnosing intraductal carcinoma. In one earlier study, sonography depicted
Figure 4. Infiltrating ductal carcinoma with and intraductal component in a 60-year-old woman. A, Craniocaudal mammogram shows a 1.5-cm irregular hyperdense mass (arrow) with spiculated margins. Calcifications are not shown within or around the mass. B, Transverse sonogram shows a 1.8-cm irregular hypoechoic mass (arrow) with spiculated margins. The mass is associated with an echogenic halo and posterior acoustic shadowing. Duct dilatation or satellite lesions are not shown around the mass. The malignant process was confirmed as IDC containing an intraductal component. The malignant process measured 2.2 cm on histopathologic examination and was underestimated on mammography and sonography.

associated breast abnormalities in only 45% of mammographically detected microcalcifications.27 On the other hand, the visibility on sonography becomes much higher for calcifications associated with malignant tumors or calcification clusters larger than 10 mm, although some isolated calcifications or small groups of calcifications may not be seen on sonography.16 Furthermore, sonography is highly dependent on the operators experience and technique.18 In this study, we could initially detect only 58% of mammographically detected calcifications on sonography without knowledge of the mammographic findings. However, when the scans were performed again with mammographic knowledge of the presence and areas of microcalcifications, we were able to detect calcifications in 81% of cases. Accordingly, with no knowledge of mammographic findings, some microcalcifications or small lesions could be missed by sonography. Dynamic contrast-enhanced breast magnetic resonance imaging (MRI) has a high degree of sensitivity for the diagnosis of invasive breast Table 3. Mammographic Findings in Pure IDC and IDC With an Intraductal Component
IDC With Intraductal Component (n = 104)

Mammographic Lexicon

Pure IDC (n = 28)

Mass Shape Round Oval Lobular Irregular Margin Circumscribed Microlobulated Obscured Indistinct Spiculated Density High Equal Calcifications Type Punctate Amorphous Fine, pleomorphic Fine, linear Distribution Regional Clustered Segmental

3 5 6 9 1 0 3 9 10 15 8

3 3 6 65 0 9 10 26 32 48 29

<.001

.228

.813

1 1 0 0 0 2 0

9 12 19 17 3 25 29

.148

.305

1594

J Ultrasound Med 2007; 26:15871600

Kang et al

cancer, with sensitivity in excess of 90%.2830 However, MRI often does not display intraductal carcinoma because of weak contrast enhancement and its small size; therefore, approximately 30% to 40% of intraductal carcinomas are not shown on MRI.3134 Moreover, normal glandular tissue or benign lesions are occasionalFigure 5. Infiltrating ductal carcinoma with an intraductal component in a 40-year-old woman. A, Mediolateral oblique mammogram shows a 2.4-cm irregular isodense mass (arrow) with spiculated margins and architectural distortion. Calcifications are not shown within or around the mass. B, Radial sonogram shows a 3.3-cm irregular hypoechoic mass with spiculated margins. There is duct dilatation extending from the main tumor (arrow). The malignant process was confirmed as IDC containing an intraductal component. The malignant process measured 3.5 cm on histopathologic examination and was underestimated on mammography.

Figure 6. Infiltrating ductal carcinoma with an EIC in a 38-yearold woman. A, Craniocaudal mammogram shows multifocal hyperdense masses (short arrows) in the left breast. There are regional pleomorphic calcifications (long arrows) within and around the masses. B, Radial sonogram shows an irregular hypoechoic mass with indistinct margins in the right breast. There are calcifications (arrows) within and around the mass. C, There are multiple satellite lesions (arrows) around the main tumor.

J Ultrasound Med 2007; 26:15871600

1595

Diagnosis of the Intraductal Component of Invasive Breast Cancer

Figure 7. Infiltrating ductal carcinoma with an EIC in a 42-yearold woman. A, Craniocaudal mammogram shows a 2.5-cm hyperdense mass (short arrows) in the left breast. There are regional punctate calcifications (long arrows) within and around the mass. B, Radial sonogram shows a 2.8-cm irregular isoechoic mass with indistinct margins. Calcifications (long arrow) are shown within the mass, and there is duct dilatation (short arrows) extending from the mass. C, Color Doppler sonogram shows increased vascularity in the surrounding tissues.

ly enhanced on MRI and can mimic a malignant tumor.35 Therefore, the sensitivity of MRI for detection of an intraductal component has been reported to range widely from 50% to 93%.11,12,17,27,35,36

Table 4. Sonographic Findings in Pure IDC and IDC With an Intraductal Component
IDC With Intraductal Spread (n = 104)

Sonographic Lexicon

Pure IDC (n = 28)

Shape Oval Round Irregular Orientation Parallel Not parallel Margin Circumscribed Indistinct Angular Microlobulated Spiculated Lesion boundary Abrupt interface Echogenic halo Echo pattern Complex Hypoechoic Isoechoic Posterior acoustic features No posterior acoustic features Enhancement Shadowing Combined pattern Surrounding tissue Duct dilatation Cooper ligament changes Architectural distortion Calcifications No calcifications Microcalcifications inside or outside mass Vascularity Not present Present in or immediately adjacent to lesion Diffusely increased vascularity in surrounding tissue

7 2 19 22 6 5 2 2 14 5 12 16 3 22 3

18 3 83 85 19 9 18 11 46 20 21 83 2 82 20

.252

.705

.472

.014

.091

11 10 6 1 3 13 5

52 15 27 10 42 39 16

.102

.003 .391 .773

26 2

58 46

<.001

10 17

38 34

.002

32

1596

J Ultrasound Med 2007; 26:15871600

Kang et al

Several studies reported that the most characteristic finding of an intraductal component on sonography is a hypoechoic or anechoic tubular structure extending from the tumor.37,38 With this sonographic diagnostic feature, Akashi-Tanaka et al10 reported that sonography had poor sensitivity of 35% compared with mammography and contrast-enhanced computed tomography, and Hata et al12 also reported a low 20.6% sensitivity rate of sonography for detection of intraductal
A

spreading. Using high-resolution sonography, Sundararajan et al17 reported sensitivity of 57.1% and specificity of 84.2% in 46 patients. However, they analyzed significance referring to only widely spreading EICs, whereas we included all cases regardless of the amount of the intraductal component. Furthermore, to the best of our knowledge, none of the earlier studies used the American College of Radiologys BI-RADS ultrasound lexicon, and a study with application of

Figure 8. Sensitivity, specificity, and accuracy of detection of an intraductal component by mammography (MMG), sonography (US), and combined assessment (MMG+US). A and B, The sensitivity and accuracy of the combined assessment were significantly higher than those of mammography or sonography. However, the sensitivity and accuracy of US1 using traditional criteria of duct dilatation and satellite lesions were not significantly different compared with those of mammography. C, The specificity was not significantly different between the above 3 methods. US1 indicates that duct dilatation and satellite lesions were considered diagnostic criteria for an intraductal component; and US2, duct dilatation, satellite lesions, calcifications, and vascularity were considered as diagnostic criteria for an intraductal component.

J Ultrasound Med 2007; 26:15871600

1597

Diagnosis of the Intraductal Component of Invasive Breast Cancer

Table 5. Mean Overestimates and Underestimates of Tumor Size by Mammography and Sonography
Findings by Imaging Study Cases Underestimates Mean SD, mm Range, mm Cases Overestimates Mean SD, mm Range, mm

Mammography IDC (n = 23) IDC with intraductal component (n = 77) Sonography IDC (n = 23) IDC with intraductal component (n = 77) Combined IDC (n = 23) IDC with intraductal component (n = 77)

8 39 10 36 6 36

7.6 7.6 7.9 6.9 4.5 3.0 7.5 5.1 2.8 1.8 4.9 3.7

125 130 110 117 15 116

7 29 12 39 9 29

6.1 3.8 7.7 5.3 3.9 2.9 6.5 3.8 3.6 2.0 5.3 3.9

212 125 112 119 17 120

the BI-RADS lexicon for evaluation of intraductal components has not been reported previously. We compared the mammographic and sonographic features of pure IDC and IDC with an intraductal component using the BI-RADS lexicon. Several studies reported a spiculated margin, marked hypoechogenicity, a thick echogenic rim, and posterior acoustic shadowing as the sonographic findings of invasive cancer, whereas a microlobulated mass with mild hypoechogenicity, ductal extension, calcifications, and normal acoustic transmission is sonographically characteristic of intraductal carcinoma.16,39,40 However, in the above-mentioned studies, cases containing an EIC or adjacent intraductal carcinomas were classified as invasive carcinoma; that is a critical limitation because invasive carcinoma can have various features according to the proportion of the intraductal component. In this study, we showed that pure IDC frequently showed an abrupt interface, whereas IDC with an intraductal component frequently showed an Table 6. Correlation Between Pathologic and Imaging Measurements of Tumor Size
Findings by Imaging Study r P

Mammography Pure IDC (n = 23) Intraductal component (n = 77) EIC (n = 32) Sonography Pure IDC (n = 23) Intraductal component (n = 77) EIC (n = 32) Combined Pure IDC (n = 23) Intraductal component (n = 77) EIC (n = 32)

0.71 0.80 0.69 0.47 0.79 0.91 0.75 0.58 0.90 0.97 0.88 0.80

<.01 <.01 <.01 <.01 <.01 <.01 <.01 <.01 <.01 <.01 <.01 <.01

echogenic halo. This finding implies that an echogenic halo could be considered a finding indicative of an intraductal component. In addition to duct dilatation and satellite lesions, microcalcifications inside or outside a mass and increased vascularity in surrounding tissue could also be influential findings for an intraductal component, and we were able to additionally detect as many as 19 cases of intraductal components. However, the margin, echo pattern, and posterior acoustic features were not indicative of an intraductal component. In our study, the sensitivity and specificity of sonography for detection of an intraductal component were 80% and 79%, respectively. The most common cause of the 21 false-negative findings was the inability of breast sonography to show duct dilatation, satellite lesions, or calcifications around the main tumor, indicating that sonography, even highresolution sonography, still has limitations for the depiction of a subtle intraductal component. Therefore, breast sonographic findings should be correlated with mammography for detection of an accompanying intraductal component. In our study, combined assessment with mammography and sonography could show an intraductal component more accurately than either mammography or sonography alone. The intraductal component may extend farther than the invasive component. It can expand the tumor volume, extend diffusely over the entire breast, and sometimes develop to form satellite neoplasms. The distance of extension varies from a few millimeters to 4 cm or greater. Mammographic measurement has a poor correlation with pathologic measurement. The main difficulty in mammographic measurement is
J Ultrasound Med 2007; 26:15871600

1598

Kang et al

poor demarcation of lesions in dense breasts.41 Discrepancies between mammographic and histopathologic size are also due to radiating spicules and indistinct margins of tumors.38 Moreover, the maximum extent of a tumor containing an intraductal component cannot be measured exactly on mammography. In this study, mammographic measurement of lesion size in IDC with an intraductal component showed a weaker correlation (r = 0.69) than pure IDC (r = 0.80). Especially, EIC-positive cases showed the poorest correlation (r = 0.47) between mammographic and histopathologic size. On the other hand, combined assessment showed a strong correlation in both pure IDC (r = 0.97) and IDC with an intraductal component (r = 0.88), even in EIC-positive cases (r = 0.80). These findings are critical for planning of treatment, especially when selecting either breastconserving surgery or mastectomy, to determine the range of excision for breast conservation and to justify the demand for sonography as well as mammography as an essential part of the diagnostic workup for breast cancer staging. Our study had some limitations. First, we excluded cases with DCIS, although it should be considered as an intraductal component. However, our study was focused on detection of an intraductal component in patients with IDC that was proven by percutaneous biopsy. Moreover, the therapeutic strategy of DCIS differs from that of IDC. Second, we included all staged cancers in the study population, although detection of an intraductal component is useful in patients who plan to undergo breast-conserving surgery. However, we focused on evaluation of mammographic and sonographic features to differentiate the absence or presence of an intraductal component rather than clinical usefulness for therapeutic strategy decisions. This clinical application for patients who plan to undergo breast-conserving surgery should be investigated in the future. Third, we did not minutely classify patients according to the extent of the intraductal component. Moreover, the definition of EIC by Schnitt et al23 is focused on the percentage of the intraductal component rather than the extent. Nevertheless, lesions with a prominent intratumoral intraductal component tend to have intraductal carcinoma outside the main tumor,42 and
J Ultrasound Med 2007; 26:15871600

we focused on the absence or presence of an intraductal component rather than the degree of the intraductal extent outside the main tumor. In summary, IDC with an intraductal component, compared with pure IDC, showed calcifications more frequently on mammography and showed an echogenic halo, duct dilatation, calcifications, and increased vascularity in surrounding tissue more frequently on sonography. Because mammography has a difficulty in showing intraductal carcinomas without calcifications, and even MRI has its limitations in assessing intraductal components, sonography is an excellent supplementary means for detecting intraductal components. Moreover, combined assessment with mammography and sonography detected intraductal components more sensitively and measured the extent of the tumor more accurately than separate mammographic and sonographic assessments.

References
1. Ohtake T, Abe R, Kimijima I, et al. Intraductal extension of primary invasive breast carcinoma treated by breastconservative surgery: computer graphic three-dimensional reconstruction of the mammary duct-lobular systems. Cancer 1995; 76:3245. Rizzatto GJ. Towards a more sophisticated use of breast ultrasound. Eur Radiol 2001; 11:24252435. Lindley R, Bulman A, Parsons P, et al. Histologic features predictive of an increased risk of early local recurrence after treatment of breast cancer by local tumor excision and radical radiotherapy. Surgery 1989; 105:1320. Holland R, Connolly JL, Gelman R, et al. The presence of an extensive intraductal component following a limited excision correlates with prominent residual disease in the remainder of the breast. J Clin Oncol 1990; 8:113118. Kollias J, Gill PG, Beamond B, Rossi H, Langlois S, VernonRoberts E. Clinical and radiological predictors of complete excision in breast-conserving surgery for primary breast cancer. Aust NZ J Surg 1998; 68:702706. Boyages J, Recht A, Connolly JL, et al. Early breast cancer: predictors of breast recurrence for patients treated with conservative surgery and radiation therapy. Radiother Oncol 1990; 19:2941. Gage I, Schnitt SJ, Nixon AJ, et al. Pathologic margin involvement and the risk of recurrence in patients treated with breast-conserving therapy. Cancer 1996; 78:19211928. Schnitt SJ, Abner A, Gelman R, et al. The relationship between microscopic margins of resection and the risk of local recurrence in patients with breast cancer treated with breast-conserving surgery and radiation therapy. Cancer 1994; 74:17461751.

2. 3.

4.

5.

6.

7.

8.

1599

Diagnosis of the Intraductal Component of Invasive Breast Cancer

9.

Stomper PC, Connolly JL. Mammographic features predicting an extensive intraductal component in early-stage infiltrating ductal carcinoma. AJR Am J Roentgenol 1992; 158:269272. Akashi-Tanaka S, Fukutomi T, Miyakawa K, et al. Diagnostic value of contrast-enhanced computed tomography for diagnosing the intraductal component of breast cancer. Breast Cancer Res Treat 1998; 49:7986. Satake H, Shimamoto K, Sawaki A, et al. Role of ultrasonography in the detection of intraductal spread of breast cancer: correlation with pathologic findings, mammography and MR imaging. Eur Radiol 2000; 10:17261732. Hata T, Takahashi H, Watanabe K, et al. Magnetic resonance imaging for preoperative evaluation of breast cancer: a comparative study with mammography and ultrasonography. J Am Coll Surg 2004; 198:190197. Stomper PC, Connolly JL, Meyer JE, Harris JR. Clinically occult ductal carcinoma in situ detected with mammography: analysis of 100 cases with radiologic-pathologic correlation. Radiology 1989; 172:235241. Ikeda DM, Andersson I. Ductal carcinoma in situ: atypical mammographic appearances. Radiology 1989; 172:661666. Holland R, Peterse JL, Millis RR, et al. Ductal carcinoma in situ: a proposal for a new classification. Semin Diagn Pathol 1994; 11:167180. Moon WK, Myung JS, Lee YJ, Park IA, Noh DY, Im JG. US of ductal carcinoma in situ. Radiographics 2002; 22:269280. Sundararajan S, Tohno E, Kamma H, Ueno E, Minami M. Detection of intraductal component around invasive breast cancer using ultrasound: correlation with MRI and histopathological findings. Radiat Med 2006; 24:108114. Rizzatto G, Chersevani R, Abbona M, Lombardo VL, Macorig D. High-resolution sonography of breast carcinoma. Eur J Radiol 1997; 24:1119. Yang WT, Suen M, Ahuja A, Metreweli C. In vivo demonstration of microcalcification in breast cancer using high resolution ultrasound. Br J Radiol 1997; 70:685690. Ranieri E, DAndrea MR, DAlessio A, et al. Ultrasound in the detection of breast cancer associated with isolated clustered microcalcifications, mammographically identified. Anticancer Res 1997; 17:28312835. Rickard MT. Ultrasound of malignant breast microcalcifications: role in evaluation and guided procedures. Australas Radiol 1996; 40:2631. American College of Radiology. Breast Imaging Reporting and Data System (BI-RADS) Atlas. 4th ed. Reston, VA: American College of Radiology; 2003. Schnitt SJ, Connolly JL, Harris JR, et al. Pathologic predictors of early local recurrence in stage I and II breast cancer treated by primary radiation therapy. Cancer 1984; 53:10491057. Greene F, Page D, Fleming I, et al (eds). AJCC Cancer Staging Manual. 6th ed. New York, NY: Springer; 2002. Elling D, Vesper AS, Fiedler B, et al. Intraductal component in invasive breast cancer: analysis of 250 resected surgical specimens. Breast 2001; 10:405410.

26.

Uematsu T, Sano M, Homma K, et al. Value of three-dimensional helical CT image-guided planning for made-to-order lumpectomy in breast cancer patients. Breast J 2004; 10: 3337. Moon WK, Im JG, Koh YH, Noh DY, Park IA. US of mammographically detected clustered microcalcifications. Radiology 2000; 217:849854. Heywang SH, Wolf A, Pruss E, Hilbertz T, Eiermann W, Permanetter W. MR imaging of the breast with Gd-DTPA: use and limitations. Radiology 1989; 171:95103. Mussurakis S, Buckley DL, Drew PJ, et al. Dynamic MR imaging of the breast combined with analysis of contrast agent kinetics in the differentiation of primary breast tumours. Clin Radiol 1997; 52:516526. Fischer U, Kopka L, Grabbe E. Breast carcinoma: effect of preoperative contrast-enhanced MR imaging on the therapeutic approach. Radiology 1999; 213:881888. Ikeda DM, Birdwell RL, Daniel BL. Potential role of magnetic resonance imaging and other modalities in ductal carcinoma in situ detection. Magn Reson Imaging Clin N Am 2001; 9:345356. Kinkel K, Hylton NM. Challenges to interpretation of breast MRI. J Magn Reson Imaging 2001; 13:821829. Viehweg P, Lampe D, Buchmann J, et al. In situ and minimally invasive breast cancer: morphologic and kinetic features on contrast-enhanced MR imaging. MAGMA 2000; 11:129137. Orel SG, Mendonca MH, Reynolds C, Schnall MD, Solin LJ, Sullivan DC. MR imaging of ductal carcinoma in situ. Radiology 1997; 202:413420. Ando Y, Fukatsu H, Ishigaki T, Endo T, Miyazaki M. Intramammary extension of breast cancer: diagnosis with MR mammography. Breast Cancer 1998; 5:291300. Ikeda O, Nishimura R, Miyayama H, et al. Magnetic resonance evaluation of the presence of an extensive intraductal component in breast cancer. Acta Radiol 2004; 45:721725. Rizzatto G, Chersevani R. Breast ultrasound and new technologies. Eur J Radiol 1998; 27(suppl 2):S242S249. Tresserra F, Feu J, Grases PJ, et al. Assessment of breast cancer size: sonographic and pathologic correlation. J Clin Ultrasound 1999; 27:485491. Skaane P, Engedal K. Analysis of sonographic features in the differentiation of fibroadenoma and invasive ductal carcinoma. AJR Am J Roentgenol 1998; 170:109114. Chen SC, Cheung YC, Lo YF, et al. Sonographic differentiation of invasive and intraductal carcinomas of the breast. Br J Radiol 2003; 76:600604. Fornage BD, Toubas O, Morel M. Clinical, mammographic, and sonographic determination of preoperative breast cancer size. Cancer 1987; 60:765771. Fisher ER, Gregorio RM, Fisher B, et al. The pathology of invasive breast cancer: a syllabus derived from findings of the National Surgical Adjuvant Breast Project. Cancer 1975; 36:185.

10.

27.

28.

11.

29.

12.

30.

13.

31.

14. 15.

32. 33.

16. 17.

34.

35.

18.

36.

19.

37. 38.

20.

21.

39.

22.

40.

23.

41.

24. 25.

42.

1600

J Ultrasound Med 2007; 26:15871600