160 A. Shiraishi et al. ms, TE: 9 ms, excitations: 1, FA: 30, FOV: 25 cm, matrix: 256~256, slice thickness: 3 4 mm, no gap) with fat saturation. The c

Transcription

1 Magnetic Resonance in Medical Sciences, Vol.2, No.4, p , 2003 MAJOR PAPER Extension of Ductal Carcinoma in Situ: Histopathological Association with MR Imaging and Mammography Akihiko SHIRAISHI 1 *, Yoshihisa KUROSAKI 1, Tadayuki MAEHARA 1, Masaru SUZUKI 2, and Masafumi KUROSUMI 3 1 Department of Radiology, Juntendo University School of Medicine Hongo, Bunkyo ku, Tokyo , Japan Departments of 2 Radiology and 3 Clinical Pathology, Saitama Cancer Center (Received August 28, 2003; Accepted December 17, 2003) The purpose of this study is to evaluate the capability of breast MRI (magnetic resonance imaging) and mammography in determining tumor extent and the detectability of ductal carcinoma in situ (DCIS) in association with histopathological features. Thirty women with breast cancer underwent 3D dynamic MRI.Twelve women had pure DCIS and 18 women had DCIS with microinvasion.we analyzed the results of preoperative MRI and mammography with histopathologic results, retrospectively.the mean lesion size was 55.1 mm from the histopathologic results. Twenty six lesions were detected through the MRI (a sensitivity of 86.7z).MRI depicted eight lesions without mammographically detected microcalciˆcation.in seven cases, MRI showed tumor extent accurately compared with mammography, and the combined diagnosis improved the accuracy of evaluating tumor extent.mri can complement mammography in guiding surgical treatment of DCIS by providing better assessment of the extent of the lesion. Keywords: breast, magnetic resonance imaging, mammography, ductal carcinoma in situ, diagnosis Introduction Because of its easy availability, low cost, and ability to detect microcalciˆcations, mammography is currently the imaging modality of ˆrst choice for breast evaluation.ductal carcinoma in situ (DCIS) accounts for 15z 20z of all detected breast cancers and 25z 56z of clinically occult cancers detected with mammography; it is found in 16z of asymptomatic women at autopsy. 1,2 Clustered microcalciˆcation is the most common mammographic feature of DCIS and is found in 85z 90z of cases. 3 However, the several limitations of mammography are well documented, including di culties in the assessment of dense breasts and the evaluation of DCIS without microcalciˆcation. Recently, breast conservation surgery has become an increasingly common treatment for DCIS. Accurate evaluation of the tumor extent is important for breast conserving treatment of DCIS. 4 Advanced breast MRI techniques are gaining *Corresponding author, Phone: { , Fax: { , E mail: asiraisi med.juntendo.ac.jp ground as adjuncts to mammography, although questions remain about their usefulness in imaging of in situ carcinomas.we evaluate the capability of breast MRI and mammography regarding lesion detectability and the extent of DCIS association with histopathologic features. Materials and Methods MRI studies were reviewed in thirty histologically veriˆed lesions (12 pure DCIS and 18 DCIS with microinvasive foci).the patients ranged in age from 34 to 70 years (average, 49.8 years). Nineteen of 30 lesions were treated with mastectomy, and 11 lesions were treated with breast conserving surgery. Mammography and MRI were performed within four weeks of surgery. Imaging technique The MRI examinations were performed with two dišerent 1.5T systems (Gyroscan, ACS NT, Philips; VISART, Toshiba).On both systems, a breast surface coil was used with the patient in the prone position.the entire single breast was imaged with a 3D spoiled gradient echo sequence (TR:

2 160 A. Shiraishi et al. ms, TE: 9 ms, excitations: 1, FA: 30, FOV: 25 cm, matrix: 256~256, slice thickness: 3 4 mm, no gap) with fat saturation. The coronal dynamic series comprised four scan times: before and 2, 6, 10 min after intravenous injection of 0.1 mmolwkg Gd DTPA. MPR (multiplanar reconstruction) images were obtained with parasagittal images after examination. In the mammographic examinations, magniˆcation and Wor spot compression views were obtained in addition to the standard oblique and craniocaudal projections (Mammorex, MGS 110B Toshiba). Pathological resection and method In breast conserving surgery, the received material is prepared as 5 mm thick slices throughout the nipple margin to the distal end to evaluate the tumor extent precisely. In cases where mastectomy was performed, a 5 mm sliced section was obtained in parallel to the line passing through the nipple and the tumor. The tumor diameter was estimated by pathologists through tumor extent mapping with the results of the microscopic examination. Imaging analysis MRI and mammography examinations were separately evaluated by two radiologists (AS, MS) with extensive experience in breast MR imaging and mammography. The reviewers were unaware of the histopathologic results. The consensus of the two radiologists as present was considered positive. In mammography, mass lesions, asymmetric opacities, architectural distortion, and malignant microcalciˆcations were considered positive ˆndings. MRI ˆndings were considered to be positive for tumorifafocalareaofenhancement,compared with normal breast tissue, was seen after contrast administration; the ˆndings were interpreted as negative if no enhancement was seen or if dišuse and uniform enhancement of the breast parenchyma was seen. We analyzed the preoperative MRI and mammography for tumor detectability and tumor extent, and the greater lesion size between mammographyandmriwasselectedinacombined diagnosis. The tumor size was measured as the maximal diameter in each modality, and we evaluated the accuracy of demonstrated tumor extent. For lesions detected in both modalities, we determined the dišerence in size between the imaging and histopathologic results as follows: within 10 mm, mm, and more than 20 mm. We also evaluated size discrepancy as a percentage: the dišerence in size divided by the actual size of the specimen. Results Of 30 lesions, the mean lesion size was 55.1 mm from the histopathologic results. MRI depicted 26 lesions (a sensitivity of 86.7z), mammography depicted 23 lesions (a sensitivity of 76.7z), and 22 lesions were detected with both modalities. Of the 26 lesions detected by MRI, four lesions without dense breast could not be depicted with mammography. In eight cases, mammography did not detect microcalciˆcations. All lesions in these eight cases were identiˆed with MRI. In one case, a lesion was identiˆed with mammography as a clustered microcalciˆcation, although no abnormality was detected with MRI. Three lesions (less than 10 mm in two lesions incidentally detected with microscopic examination and 20 mm in one lesion detected with ultrasound) were not detected with either modality. For lesions detected with both modalities, the mean histopathological tumor size was 40.9 mm (ranging from 10 to 90 mm). With MRI, the tumor size was underestimated or overestimated within 10mm in 13 cases. MRI overestimated tumor size by more than 20 mm in one case; in this case, proliferative ˆbrocystic change was found around the malignant lesion histologically. In seven cases, MRI underestimated tumor size by more than 20 mm. MRI could not show the microscopic extent of the tumor. Mammography evaluated tumor size with a distinction range within 10 mm in 10 cases. In seven cases, mammography underestimated tumor size by more than 20 mm. In seven cases, MRI was superior to mammography, and in three cases mammography was more accurate than MRI in evaluating the tumor extent. In the diagnoses combining MRI and mammography, the distinction range was more than 20 mm in four cases if the greater size was selected. (Table 1 shows the results demonstrating tumor extent.) MRI evaluated the percentage discrepancy within 20z in 12 cases. (Table 2 shows the percentage discrepancy in size.) Discussion Breast MRI has recently evolved, and many studies have documented the improvements in diagnostic performance, showing a higher sensitivity to invasive carcinoma than is provided with mammography. 5 7 The use of MRI as an adjunct to mammography resulted in an increase in diagnostic accuracy.anearlysteepincreaseinsignalintensity followed by a decrease or plateau in signal intensity has been reported as a sign of malignancy, and a continuous increase in signal intensity is regarded Magnetic Resonance in Medical Sciences

3 Extension of Ductal Carcinoma in Situ 161 Table 1. DiŠerence in tumor extent between imaging and specimen 0 Å10 mm mm»21 mm Å10 mm mm»21 mm underestimation overestimation 8 (18 MRI 2 (9.1z) 8 (36.4z) 3 (13.6z) 6 (27.3z) 3 (13.6z) 0 0 mammography 1 (4.5z) 7 (31.8z) 3 (13.6z) 7 (31.8z) 3 (13.6z) 0 1 (4.5z) MRI{mammography 2 (9.1z) (36.4z) 4.2z) 3 (13.6z) 4 (18.2z) 0 1 (4.5z) MRI{mammography: combined evaluation with MRI and mammography Table 2. DiŠerence in size divided by actural size of specimen 0 20z 21 40z 41 60z 61 80z z 100zº MRI mammography as a sign of benignity. 8 Various authors reported MR ˆndings of DCIS; however, the evaluation of DCIS with MRI is reported to be limited. The reported sensitivities of MRI for the detection of DCIS have varied widely, ranging from 40z to 100z 9 13 ; moreover, there have been reports of false negative cases with variable size tumors. 5,14 16 DCIS demonstrates a variable enhancement pattern with MRI. Some investigators have identiˆed relatively slowly enhancing cancers and rapidly enhancing benign lesions. 17 DCIS shows clumped or linear enhancement along the ductal pattern. The ductal pattern is important in the diagnosis of DCIS. However, ˆbrocystic changes show clumped enhancement as well as DCIS, and the enhancement is sometimes delayed, suggesting benignity. 11 Therefore, if the ductal pattern is not identiˆed, DCIS is often indistinguishable from a benign lesion with MRI. The cure rate of DCIS with mastectomy is approximately 100z. Currently, demand for breast conservation treatment is increasing. Harris et al. describe lumpectomy and radiation therapy as reasonable for a DCIS within 3 cm in size. 18 Rosner et al. demonstrated that patients who underwent wedge resection had equivalent outcomes to those who underwent mastectomy for DCIS. 19 Therefore, accurate evaluation of the tumor extent is important when breast conservation treatment is considered. Mammography can demonstrate most DCIS lesions, but a more accurate imaging assessment of lesion extent is necessary for successful breast conservation. The extent of DCIS cannot always be accurately predicted with mammography. 20,21 The tumor extent is often underestimated with mammography. 22 As for size determination of breast cancer with MRI, several previous reports are available. 23,24 They concluded that MRI provided greater accuracy than did mammography in determining tumor size. Boetes et al. demonstrated with eight DCIS lesions that, while the size of the invasive part was correctly estimated, the DCIS component was underestimated by more than 1 cm in all patients. 12 However, little data is available for DCIS. In our study group, the detectability of DCIS was higher with MRI than with mammography. Microcalciˆcations and the microscopic extent of the lesions are not depicted with MRI; therefore, MRI alone may be insu cient. One false negative case with MRI was correctly diagnosed as a malignancy with mammography. Because MRI was able to depict DCIS without microcalciˆcations (Fig. 1), a combined evaluation would increase the sensitivity compared with a single modality. One 15 mm lesion was not identiˆed with MRI, suggesting that tumor size alone is not responsible for false negative results. We considered that mammography is able to depict lesions if microcalciˆcation exists, and that demonstration with MRI is related not only to lesion size, but also to the amount of tumor cells and the signal behavior of contrast enhancement. We analyzed the histologic results and the preoperative MRI results regarding tumor extent for DCIS. MRI in addition to mammography improved the diagnostic accuracy of evaluating tumor extent and reduced the number of cases in which the tumor margin was underestimated by more than 20 mm. In our hospital, surgeons have a policy of maintaining a 20 mm surgical margin for breast conservative surgery. Our results suggest that MRI is helpful for preoperatively evaluating tumor extent in DCIS. However, the presence of Vol. 2 No. 4, 2003

4 162 A. Shiraishi et al. Fig year old woman with bloody nipple discharge Mediolateral oblique mammogram of the right (a) and left (b) breasts shows no deˆnite mass and malignant calciˆcations. Dynamic studies of coronal images (c) before and at (d) 2,(e) 6, and(f) 10 min after injection of Gd DTPA show clumped enhancement in the upper inner quadrant of the left breast. (g) TheMPRimage demonstrates linear enhancements toward the nipple in a ductal pattern,corresponding to a DCIS. ˆbrocystic changes reduces the diagnostic value. If there is only dišuse or patchy enhancement without linear fashion with MRI,histological examination is necessary in the cases where mammography depicted no malignant calciˆcations. In conclusion,mri improves lesion detectability and accuracy when evaluating tumor extent and it is able to depict DCIS without microcalciˆcations. Acknowledgments We wish to thank T. Ohtsuka,K. Ishiguri,and the technicians of the Saitama Cancer Center for their invaluable assistance. References 1. Stomper PC,Margolin FR. Ductal carcinoma in situ: the mammographer's perspective. AJR 1994; 162: Ores SG,Schnall MD,Newman RW,et al. MR imaging guided localization and biopsy of breast lesions: initial experience. Radiology 1994; 193: Evans AJ,Wilson ARM,Burrell HC,et al. Mammographic features of ductal carcinoma in situ (DCIS) present on previous mammography. Clinic Radiol 1999; 54: Elliott RL,Haynes AE,Bolin JA,et al. Stereotaxic needle localization and biopsy of occult breast lesions: ˆrst year's experience. Am Surg 1992; 58: Magnetic Resonance in Medical Sciences

5 Extension of Ductal Carcinoma in Situ Soderstrom CE, Harms SE, Copit DS, et al. Threedimensional RODEO breast MR imaging of lesions containing ductal carcinoma in situ. Radiology 1996; 201: Elvecrog EL, Lechner MC, Nelson MT. Nonpalpable breast lesions: correlation of stereotaxic largecore needle biopsy and surgical biopsy results. Radiology 1993; 188: Bone B, Aspelin P, Bronge L, et al. Sensitivity and speciˆcity of MR mammography with histopathological correlation in 250 breasts. Acta Radiologica 1996; 37: Baudu LD, Murakami J, Murayama S, et al. Breast lesions: correlation of contrast medium enhancement patterns on MR images with histopathologic ˆndings and tumor angiogenesis. Radiology 1996; 200: Fryberg ER, Bland KI. Overview of the biology and management of ductal carcinoma in situ of the breast. Cancer 1994; 74: Heywang SH. Contrast enhanced magnetic resonance imaging of the breast. Invest Radiol 1994; 29: Orel SG, Schnall MD, LiVolsi VA, et al. Suspicious breast lesions: MR imaging with radiologic pathologic correlation. Radiology 1994; 190: BoetesC,BarentzJO,MusRD,etal.MRcharacterization of suspicious breast lesions with a gadolinium enhanced turbo FLASHsubtraction technique. Radiology 1994; 193: Gilles R, Meunier M, Lucidarme O, et al. Clustered breast microcalciˆcations: evaluation by dynamic contrast enhanced subtraction MRI. J Comput Assist Tomogr 1996; 20(1): Fobben ES, Rubin CZ, Kalisher L, et al. Breast MR imaging with commercially available technique: radiologic pathologic correlation. Radiology 1995; 196: Gilles R, Guinebretiere J, Lucidarme O, et al. Non palpable breast tumors: diagnosis with contrastenhanced subtraction dynamic MR imaging. Radiology 1994; 191: Stomper PC, Connolly JL. Ductal carcinoma in situ of the breast: correlation between mammographic calciˆcation and tumor subtype. AJR 1992; 159: Holland R, Hendriks JHCL, Verbeek ALM, et al. Extent, distribution, and mammographic Whistological distribution of breast ductal carcinoma in situ. Lancet 1990; 335: Harris JR, Lippman ME, Veronesi U, et al. Breast cancer. N Engl J Med 1992; 327: RosnerD,BedwaniRN,VenaJ,etal.Noninvasive breast carcinoma: results of a national survey by the American College of Surgeons. Ann Surg 1980; 192: Gilles R, Zafrani B, Guinebretiere JM, et al. Ductal carcinoma in situ: MR imaging histopathologic correlation. Radiology 1995; 196: StomperPC,HermanS,KlippensteinDL,etal. Suspect breast lesions: ˆndings at dynamic gadolinium enhanced MR imaging correlated with mammographic and pathologic features. Radiology 1995; 197: Boetes C, Mus RDM, Holland R, et al. Breast tumors: comparative accuracy of MR imaging relative to mammography and US for demonstrating extent. Radiology 1995; 197: Boetes C, Strijik SP, Holland R, et al. Falsenegative MR imaging of malignant breast tumors. Eur Radiol 1997; 7: Westerhof JP, Fischer U, Moritz JD, et al. MR imaging of mammographically detected clustered microcalciˆcations: is there any value? Radiology 1998; 207: Vol. 2 No. 4, 2003