Complex cystic breast lesions, which are defined as lesions

Transcription

1 ORIGINAL RESEARCH Value of Ultrasonographic Features for Assessing Malignant Potential of Complex Cystic Breast Lesions Jin-Peng Yao, MD, Yu-Zhi Hao, MD, Qing Chang, MD, Cheng-Yun Geng, MD, Yu Chen, MD, Wen-Peng Zhao, MD, Yan Song, MD, Xiang Zhou, MD Objectives To assess the value of ultrasonography (US) features for determining the malignant potential of complex cystic lesions. Methods Seventy-nine complex cystic lesions were reviewed retrospectively. They were classified into four types according to US features in type I, the masses have a thick outer wall, thick internal septa, or both; in type II, the masses are an intracystic type with one or more discrete solid mural lesions within a cyst; in type III, the masses contain mixed cystic and solid components and are at least 50% cystic portion in a mass; in type IV, there are predominantly (at least 50%) solid masses with eccentric or central cystic foci. Positive predictive values were calculated for all types. Results The frequency of malignancy was higher among type III and IV lesions than among the other two types. Lesions with a diameter greater than or equal to 20 mm, margins not circumscribed, resistance index greater than or equal to 0.7, and axillary abnormal nodes had a high probability of malignancy. Conclusions US is an important adjunct to evaluate the malignant potential of complexcysticlesions. Key Words breast; complex cystic lesion; ultrasonography Received May 9, 2016, from the Departments of Pathology (J.P.Y., Y.Z.H., Q.C., C.Y.G., Y.C., W.P.Z., X.Z.) and Ultrasonography (Y.S.), Cancer Hospital, Chinese Academy of Medical Sciences, Beijing, China. Revised manuscript accepted for publication June 27, Address correspondence to Xiang Zhou, MD, Department of Ultrasound, Cancer Hospital, Chinese Academy of Medical Sciences, No. 17, Pan Jia Yuan Nan-li, Beijing , China. zhou.xiang@yeah.net Abbreviations ACR, American College of Radiology; ADH, atypical ductal hyperplasia; BI-RADS, Breast Imaging Reporting and Data System; DCIS, ductal carcinoma in situ; PPV, positive predictive value; RI, resistance index; US, ultrasonography doi: /ultra Complex cystic breast lesions, which are defined as lesions consisting of cystic and solid components with a thick wall or internal septum/septa, are encountered with increasing frequency on ultrasonography (US). 1 Such lesions are associated with a variety of benign, atypical, and malignant pathologic diagnoses, including fibrocystic changes, intraductal papilloma, atypical ductal hyperplasia (ADH), ductal carcinoma in situ (DCIS), and infiltrating ductal carcinoma. 1,2 In the American College of Radiology Breast Imaging Reporting and Data System (ACR BI-RADS) US lexicon, complex cystic lesions containing a solid component, even if discreet, are classified as ACR BI-RADS category 4 (ACR4) and considered suspicious for malignancy, requiring histological verification following percutaneous biopsy and/or surgical excision for diagnostic purposes. 3 5 Radiologists face the challenge of differentiating complicated cysts (ACR3/probably benign lesions, >98% likelihood; short-term follow-up recommended) from complex cystic lesions (ACR4/suspicious abnormality; biopsy must be considered). However, the US features of complex cystic breast lesions remain undefined, 1,2,6 and their use for assessing the malignant potential of lesions has not been established. VC 2017 by the American Institute of Ultrasound in Medicine J Ultrasound Med 2017; 36:

2 Accurate descriptions and classifications of these lesions as detected with US will enhance the clarity of the US reporting system and permit more appropriate and specific subsequent recommendations. The purpose of this study was to analyze the US features of complex cystic breast masses for which subsequent fine-needle aspiration biopsy, core-needle biopsy, or surgical excision was performed to determine the value of US features for predicting the malignant potential of these cysts. Materials and Methods Institutional review board approval was obtained for this study, and the need for informed patient consent was waived as a result of the retrospective nature of the study. Cases observed between July 2010 and October 2015 were retrospectively identified by searching the institutional database for all consecutive US reports for patients with prospective breast lesions with a cystic component. Records were reviewed for 76 women with 79 masses for which subsequent fine-needle aspiration biopsy, core-needle biopsy, or surgical excision was performed. Data on patient demographics, clinical presentation, and histological examinations were extracted from the medical records. Breast US was performed for all 79 lesions with a 7- to 12-MHz linear array transducer (GE Logiq E9, GE Logiq9,GELogiqS6,GELogiq700[GEHealthcare, Milwaukee, WI], Philips iu 22 [Philips Medical Systems, Bothell, WA], or Siemens S2000 [Siemens Medical Solutions, Mountain View, CA]) by four radiologists, each with 4 to 20 years of experience in breast imaging. The US signs and classification of complex cystic masses were analyzed by two radiologists who each had at least 5 years of experience in breast imaging. The US signs include lesion size, margins, microcalcification, Doppler mode (resistance index [RI] > 0.7), and presence of abnormal axillary nodes. Masses were categorized as one of four types according to US findings 6 : type I, masses with a thick wall (>0.5 mm) or thick septa (>0.5 mm; Figure 1); type II, masses of an intracystic type with one or more discrete solid mural lesions within a cyst (Figure 2); type III, masses containing mixed cystic and solid components with the cystic portion occupying at least 50% of the mass (Figure 3); and type IV, masses that were predominantly (at least 50%) solid with eccentric or central cystic foci (Figure 4). All retrospective reviews were performed with knowledge of the study design but without knowledge of histological results. Discrepancies were resolved though discussion. All fine-needle aspiration and core-needle biopsies were performed under US guidance. Diagnosis was established with aspiration using a 22-gauge needle in 32 lesions from which fluid was sent for cytological examination. A median number of three tissue samples were Figure 1. Type I. Right breast mass in a 78-year-old woman noticed by the patient more than 1 year previously that caused no symptoms and gradually increased. A, Ultrasonography image showing a circumscribed, thick-walled complex cystic mass with thick septa (arrow). B, Photomicrograph of the subsequently excised specimen showing infiltrating ductal carcinoma with differentiation of squamous cell carcinoma in the fibrous capsule wall (arrow) (magnification, 340). 700 J Ultrasound Med 2017; 36:

3 Yao et al US Assessment of Malignant Potential of Complex Cystic Lesions obtained using an automated biopsy device with an 18gauge needle (Promex Technologies, Franklin, IN). The biopsy procedures were performed by radiologists with more than 5 years of experience in breast imaging. Statistical computations were performed using commercially available software (SPSS version 16.0, SPSS Inc, Chicago, IL). For all lesions, we calculated the positive predictive value (PPV) as the number of cancerous lesions divided by the total number of lesions per type for the four US types, and the chi-square test was applied for identifying differences among the four types of lesions. Multiple linear regression was used to analyze the correlation of malignancy (dependent variable) with lesion size, margins, microcalcification, blood-flow RI, and axillary abnormal nodes (independent variables), screening variables with the enter method. Differences with a P value of less than.05 were considered significant. Figure 2. Type II. Left breast mass in a 49-year-old woman noticed by the patient 8 years previously that gradually increased and caused nipple discharge for 7 months before treatment. A, Ultrasonography image showing a cystic mass with internal mural nodules (arrow) and anechoic portion. B, Photomicrograph of the subsequently excised specimen showing an intraductal papilloma with a papillary structure and no atypical cells (arrow) (magnification, 340). Figure 3. Type III. Left breast mass in a 76-year-old woman noticed by the patient 1 year previously that caused no symptoms and gradually increased. A, Ultrasonography image showing a complex cystic and solid mass (arrow) with a cystic portion occupying at least 50% of the mass. B, Photomicrograph of the subsequently excised specimen showing mucinous adenocarcinoma with intraductal papilloma (arrow) (magnification, 3100). J Ultrasound Med 2017; 36:

4 Results Themeanageofthe76femalepatientswiththe79 complex cysts included in this study was 49.5 years (range, years). Pathologic confirmation was performed by fine-needle aspiration (n 5 32) or coreneedle biopsy (n 5 47), and 49 of the 79 lesions were surgically resected. Histological analysis showed that 28 (35.4%) cysts were malignant, 3 (3.8%) were high-risk lesions, and 48 (60.8%) were benign lesions (Table 1). The malignancies included invasive ductal carcinoma (n 5 10), intraductal carcinoma (n 5 5), intraductal papillary carcinoma (n 5 4), non-hodgkin lymphoma (n 5 3), mucinous carcinoma (n 5 2), DCIS (n 5 2), metaplastic carcinoma (n 5 1), and phylloides sarcoma (n 5 1). After surgical excision, the three high-risk lesions were found to be intraductal papillomas with adjacent ADH. The histological diagnoses of the remaining 48 benign lesions were cysts (n 5 15), fibrocystic changes (n 5 9) including sclerosing adenosis and apocrine metaplasis, intraductal papilloma (n 5 6), intraductal papilloma with fibrocystic changes (n 5 5), galactoceles (n 5 5), abscesses (n 5 4), fat necrosis (n 5 2), fibroadenoma (n 5 1), and benign phyllodes tumor (n 5 1). A summary of the histological outcomes and US classifications for all lesions is provided in Table 1. Of the 79 complex cystic lesions on US, 14 (17.7%) were classified as type I, 18 (22.8%) as type II, 18 (22.8%) as type III, and 29 (36.7%) as type IV. The PPV for malignancy in each type of lesion were 7.1% for type I, 16.7% for type II, 61.1% for type III, and 44.8% for type IV. Among these four types, types III and IV were associated with the highest PPV. Moreover, the frequency of malignancy was significantly higher among type III and IV lesions compared with type I and II lesions (P 5.004). Table 2 lists the association of malignancy and US characteristics. Lesion size greater than or equal to 2 cm, absence of circumscribed margins, RI greater than or equal to 0.7, and axillary abnormal nodes were found to be independently related to the risk of malignancy (all P <.05). No association was observed between microcalcification and malignancy (P 5.071), even though 5 of the 6 cysts showing microcalcification were malignant (invasive ductal carcinoma [n 5 1], DCIS [n 5 1], intraductal papillary carcinoma [n 5 2], and non-hodgkin lymphoma [n 5 1]). The other cyst showing microcalcification was breast adenosis, with ductal dilatation and focal epithelial hyperplasia. Discussion Consistent with the general consensus regarding complex cystic masses, the lesions in our study corresponded to a wide range of diagnoses, with more than half being benign (48/79, 60.8%) and distributed across 10 histologic types. Moreover, the rate of malignancy (28/79, 35.4%), including eight histologic Figure 4. Type IV. Right breast mass in a 72-year-old woman noticed by the patient 1 year previously that caused no symptoms and gradually increased. A, Ultrasonography image showing a complex cystic and solid mass (arrow) with a cystic portion occupying at least 50% of the mass. B, Photomicrograph of the subsequently excised specimen showing intraductal papillary carcinoma, obvious atypical cells, and epithelial cells around, without myoepithelium (arrow) (magnification, 3100). 702 J Ultrasound Med 2017; 36:

5 types, was similar to a rate previously reported in the literature. 7 Thus, our findings confirm that any complex lesions must be characterized and carefully sampled, and a clear understanding of the US characteristics of complex cystic masses will better allow radiologists to guide clinicians in treatment planning. In this study, we analyzed the US imaging features with respect to the association between malignancy and complex cystic lesions, and attempted to identify lesion characteristics associated with a significantly higher rate of malignancy. As given in Table 2, lesion size greater than or equal to 2 cm, the absence of circumscribed Table 1. Correlation of Histopathologic Findings and US Types for 79 Complex Cystic Breast Lesions Findings Type I Type II Type III Type IV Total Benign Cyst IDP Fat necrosis Galactocele Abscess Fibroadenoma FCC IDP1FCC IDP1ADH Phyllodes tumor Malignant DCIS Intraductal CA Papillary CA IDC Mucinous CA Lymphoma Metaplastic CA Phylloides sarcoma Total Note: Data are the numbers of lesions. IDP, intraductal papilloma; FCC, fibrocystic change, including sclerosing adenosis and apocrine metaplasia; IDP1ADH, intraductal papilloma and adjacent atypical ductal hyperplasia; IDC, invasive ductal carcinoma; DCIS, ductal carcinoma in situ; CA, carcinoma. margins, RI greater than or equal to 0.7, and axillary abnormal nodes were found to be significant predictors of malignancy. Although microcalcification is an US imaging characteristic that is important for early diagnosis of breast cancer, 8 10 we found no association between microcalcification and malignancy in our series of complex cysts. Possible explanations for this result are as follows. First, microcalcification was seen on US in only 5 (16.1%)ofthe31malignantlesions(andconfirmedby mammography), and this rate is lower than that in the series by Morgan et al, 11 in which approximately half of breast cancers exhibited microcalcification. Second, microcalcification in lesions proven to be malignant is nearly always associated with DCIS or DCIS with a small area of invasion. 12,13 In the present study, four malignant pathological types were observed among the five of six cysts showing microcalcification. Therefore, the variety of pathological types may be another reason for the low incidence of microcalcification in complex cysts. Breast US has been considered most useful for distinguishing fluid-filled lesions from solid masses. Ultrasonography is a highly accurate modality for confirming the presence and location of a solid component within complex cystic lesions. 1,2,6,14,15 To assess the risk of malignancy among the cases included in the present study, we classified the masses into four types according to the proportion and distribution of internal cystic and solid components. Notably, in all categories, the PPV was greater than 2%, with even type I lesions having a PPV of 7.1% and the other categories having higher PPV values. Therefore, these lesions fit in category ACR4 (PPV between 2 and 95%) and must be biopsied. In addition, the greater PPVs observed for type III and IVlesionsversustypeIandIIlesionswereconsistent with the findings of Berg et al. 16 Thus, these classifications may allow the differentiation of masses more likely to be malignant from those less likely to be malignant. Table 2. Risk Factors Associated With Malignancy: Results of Univariate and Multivariate Analyses Standardized Coefficients t P Value 95% CI for B Variables Beta [lower bound ~ upper bound] Constant Size Margin Microcalcification RI Axillary Nodes J Ultrasound Med 2017; 36:

6 There were some limitations to our study. First, our study was retrospective in design, and selection bias was introduced by limiting entry to patients who had undergone US examination and biopsy, which does not represent the general population. Second, the results were derived from a relatively small sample of 79 cases, which is insufficient to adequately describe lesions with a low prevalence rate such as those with microcalcification. Further analysis in a larger series with more cases is needed to more clearly address this concern. Another limitation is the operator-dependent nature of US; interpretation of features may vary on the basis of images. However, even though biased interpretation of images was possible, the approach used in our study is representative of how complex cystic lesions in the breast are diagnosed. In summary, complex cystic masses can be associated with a wide range of diagnoses, from benign lesions to high-risk or atypical lesions or even malignant lesions. Ultrasonography imaging can permit appropriate categorization and thus be valuable in risk stratification of lesions according to the risk of malignancy. A more completeknowledgeoftheusappearancesofcomplexcysts may help radiologists to achieve greater specificity in the diagnosis of such cases and, in turn, improve recommendations for appropriate clinical management of these cysts. References 1. Doshi DJ, March DE, Crisi GM, Coughlin BF. Complex cystic breast masses: diagnostic approach and imaging-pathologic correlation. Radiographics 2007; 27:S53 S Chang YW, Kwon KH, Goo DE, et al. Sonographic differentiation of benign and malignant cystic lesions of the breast. J Ultrasound Med 2007; 26: Athanasiou A, Aubert E, Vincent Salomon A, Tardivon A. Complex cystic breast masses in ultrasound examination. Diagn Interv Imaging 2014; 95: Houssami N, Irwig L, Ung O. Review of complex breast cysts: implications for cancer detection and clinical practice. ANZ J Surg 2005; 75: Mendelson EB, Baum JK, Berg WA, Merritt CR, Rubin E. Ultrasonography. In: Breast Imaging Reporting and Data System (BI-RADS). 4th Edition. Reston, VA: American College of Radiology; Berg WA, Campassi CI, Ioffe OB. Cystic lesions of the breast: sonographic-pathologic correlation. Radiology 2003; 227: Berg WA, Sechtin AG, Marques H, Zhang Z. Cystic breast masses and the ACRIN 6666 experience. Radiol Clin North Am 2010; 48: Bae S, Yoon JH, Moon HJ, Kim MJ, Kim EK. Breast microcalcifications: diagnostic outcomes according to image-guided biopsy method. Korean J Radiol 2015; 16: Han XJ, Ren JH, Ma N, Tan QT, Wang SY. [The study in detection of microcalcification in early breast cancer by ultrasound and its correlation with pathohistology]. Zhonghua Yi Xue Za Zhi 2012; 92: Tse GM, Tan PH, Pang AL, Tang AP, Cheung HS. Calcification in breast lesions: pathologists perspective. J Clin Pathol 2008; 61: Morgan MP, Cooke MM, McCarthy GM. Microcalcifications associated with breast cancer: an epiphenomenon or biologically significant feature of selected tumors? J Mammary Gland Biol Neoplasia 2005; 10: BagnallMJ,EvansAJ,WilsonAR,etal.Predictinginvasioninmammographically detected microcalcification. Clin Radiol 2001; 56: Catteau X, Simon P, Noel JC. Predictors of invasive breast cancer in mammographically detected microcalcification in patients with a core biopsy diagnosis of flat epithelial atypia, atypical ductal hyperplasia or ductal carcinoma in situ and recommendations for a selective approach to sentinel lymph node biopsy. Pathol Res Pract 2012; 208: Kim SM, Kim HH, Kang DK, et al. Mucocele-like tumors of the breast as cystic lesions: sonographic-pathologic correlation. Am J Roentgenol 2011; 196: Vargas HI, Vargas MP, Gonzalez KD, Eldrageely K, Khalkhali I. Outcomes of sonography-based management of breast cysts. Am J Surg 2004; 188: Hsu HH, Yu JC, Lee HS, et al. Complex cystic lesions of the breast on ultrasonography: feature analysis and BI-RADS assessment. Eur J Radiol 2011; 79: J Ultrasound Med 2017; 36: