Worldwide the most common reason for a. Role of High Frequency Ultrasonography in the Evaluation of Palpable Breast Masses in Chinese Women:

Transcription

1 Role of High Frequency Ultrasonography in the Evaluation of Palpable Breast Masses in Chinese Women: Alternative to Mammography? Wei Tse Yang, MBBS, FRCR, Chun On Mok, MBBS, FRCS, Waiter King, MD, FRCS Can, FRCS (Edinb), FACS, Alice Tang, MBBS, FRCR, Constantine Metreweli, MA, MBBChir (Cantab), FRCP, MRCS, DMRD, FRCR We prospectively assessed the accuracy of high resolution breast ultrasonography in the diagnosis of palpable breast masses in comparison to clinical palpation and x-ray mammography. Four hundred and eight Chinese women with palpable breast lumps had clinical assessment followed by ultrasonography of the breast, mammography (for women over 35 years), and fine needle aspiration cytology. Excisional biopsy or surgery was performed for suggestive lesions. The clinical, mammographic and ultrasound diagnoses were compared with the final pathologic diagnosis. In the determination of whether a lesion was malignant, the sensitivity, specificity, and positive predictive values were 97%, 97%, and 85%, respectively, for ultrasonography; 92%, 94%, and 84%, respectively, for mammography; and 88%, 92%, and 67%, respectively, for clinical evaluation. The specificity for combined clinical palpation and ultrasonography was higher (99%) than that for combined clinical palpation and mammography (96%). Addition of ultrasonography to combined clinical palpation and mammography increased specificity. Mammography in addition to combined clinical palpa tion and ultrasonography did not significantly improve the sensitivity, specificity, or positive predictive value. This limited usefulness raises the question as to whether it should be eliminated in the workup of a palpable mass in the average Chinese patient. Its main advantage is the detection of extended foci of carcinoma in situ related to a palpable mass, which often is undetected by ultrasonography. KEY WORDS: Breast masses; Ultrasonography; Mammography. Worldwide the most common reason for a woman to seek medical consultation for carcinoma of the breast is that she finds a palpable abnormality in her breast. This is even Received November 30, 1995, from the Departments of Diagnostic Radiology and Organ lmaging (W.T.Y., A.T., C.M.) and Surgery <C.O.M... W.K.), The Chinese University of Hong Kong, Prince of Wales Hospital, Shatin, N.T., Hong Kong. Revised manuscript accepted for publication May 27, Address correspondence and reprint requests to Wei Tse Yang, Department of Diagnostic Radiology and Organ lmaging, Chinese University of Hong Kong, Prince of Wales Hospital, Shatin, N.T., Hong Kong. true in most developed countries despite the extensive screening programs available. All patients referred to the breast clinic of the Prince of Wales Hospital with a palpable breast lump undergo a triple assessment (clinical examination, imaging, and fine needle aspiration cytology) when there is any doubt as to tne nature of the lesion. Awareness of some of the difficulties in mammographic interpretation that occur in Chinese women, who tend to have small breasts,' has led to addition of breast ultra sonography to our irnaging protocol. This prospective study was undertaken to assess the accuracy of ultrasonography in the diagnosis of palpable breast lumps and to compare it with x-ray mammography and clinical palpation by the American Institute of Ultrasound in Medicine J Ultrasound Med 15: , / 96/ $3.50

2 638 PALPABLE BREAST MASSES IN CHINESE WOMEN J Ultrasound Med 15: ,1996 MATERIALS AND METHODS Four hundred and eight women with palpable breast lumps seen at the breast clinic of the Prince of Wales Hospital had detailed clinical assessment, followed by breast ultrasonography and x-ray mammography. For women under the age of 35 years, x-ray mammography was performed only if the sonogram was suggestive of cancer to ensure screening of the asymptomatic breast. All palpable breast lumps (408) underwent fine needle aspiration. Excisional biopsy or surgery was performed for suggestive lesions (241 lesions). The clinical, mammographic, and ultrasonographic diagnoses were compared with the final pathologic diagnosis or with cytologic diagnosis for those patients who did not undergo excision biopsy or surgery (157 lesions). Of this latter group, 135 patients with negative cytologic results for malignancy were followed up either clinically or with imaging for a duration of 11 to 23 months (mean, 15 months). No further cancers were detected. Eighteen patients were lost to follow up. Whole breast ultrasonography was performed in all patients. An Aloka 650 (Aloka, Tokyo, Japan) with a high-resolution (10 MHz) mechanical sector probe and an ATL Ultramark 9 HDI (ATL, Bothell, WA) scanner with high-resolution dynamic linear (10 to 5 MHz) probe were used employing direct skin contact. The patients were placed in the supine oblique position with the side being examined raised slightly to distribute the breast tissue evenly over the pectoral muscle, and the ipsilateral arm raised to tense the breast as much as possible. Scans were obtained both transversely and longitudinally, taking special care to examine the periphery of the breast to avoid missing any occult lesions. Fine needle aspiration was performed under direct sonographic guidance with the needle tip seen within the lesion as a well-defined echogenic focus to confirm the correct position. X-ray mammography was performed using a GE Senographe DMR unit (General Electric Medical Systems, Milwaukee, WD and Kodak minr filmcassette-screen system (Eastman Kodak, Rochester, NY). Standard mediolateral oblique and craniocau dad views were obtained, with additional views as deemed necessary depending on the site of the lesion. Diagnostic Criteria The ultrasonograms and mammograms were obtained and interpreted by two radiologists who were unaware of the clinical impression. Differential and specific mammographic and sonographic diagnoses were formulated on the basis of the criteria recorded for breast masses.2-s The results of each diagnostic method were classified into four diagnostic categories: benign, indeterminate, suggestive, or malignant. To evaluate the diagnostic methods in this specific setting, sensitivity, specificity, accuracy, positive predictive value, and negative predictive value were calculated for individual tests and combination of tests. Malignant, suggestive, and indeterminate results were grouped together as they required the same clinical management. For determination of the indices in multimodality evaluation, the test was considered positive if at least one method indicated the lesion was indeterminate or malignant, and it was deemed negative if all methods indicated the lesion was benign. The parameters of individual tests were compared with the McNemar test for paired dichotomous data. RESULTS Series and Pathologic Findings The 408 patients with palpable breast lumps who had clinical evaluation, breast ultrasonography, aspiration cytology, or excision histology formed the study group. Patients over 35 years of age (226 women) also had x-ray mammography. The patients ranged in age from 13 to 85 years (mean, 37 years) (Table 1). This group was divided into six diagnostic categories: malignancy, fibroadenoma, benign breast change, cyst, normal breast, and miscellaneous (Table 2). The age distribution of the patients with breast cancer is listed in Table 3. The distribution of histologic findings in the 67 malignant breast tumors is given in Table 4. The truepositive and false-negative rates of each diagnostic method in relation to the size of the cancer are given in Table 5. The sensitivity, specificity, accuracy, positive predictive value, and negative predictive value of individual diagnostic methods and combinations of methods are given in Table 6. Table 1: Age Distribution of 408 Women with Palpable Breast Lumps Age (years) Under Over 50 Number(%) 190 (46.6) 73 (17.9) 89 (21.8) 56 (13.7)

3 J Ultrasound Med 15: ,1996 YANGETAL Table 2: Histologic or Cytologic Diagnosis in 408 Patients with Palpable Breast Lumps Diagnostic Category Number % Malignancy 67 (16.4) Fibroadenoma 169 (41.4) Cyst 50 (12.3) Benign breast change 78 (19.1) (fibnocysticchange,fibroadenos~. papilloma) Normal breast 26 (6.4) Miscellaneous 18 (4.4) (abscesses, galactocele, inflammation, fat necros~) Includes one patient with lobular carcinoma in situ. Breast Ultrasonography Carcinoma was diagnosed correctly by ultrasonography in 64 of 66 patients with a proved malig nancy (sensitivity 97%). Ultrasonography did not miss any lesion. However, in two patients, the lesion was seen but diagnosed as benign. Of the correctly diagnosed cancers (true-positive results), four measured less than 1 cm in diameter and 23 measured less than 2 cm in diameter. In each of three patients, an irregular hypoechoic tumor measuring under 1 cm with an increased depth to width ratio was diagnosed correctly as malignant. Ultrasonography, however, did not allow diagnosis of associated ductal carcinoma in situ components outside the mass in one of these patients (Fig. 1). In a fourth patient, the diagnosis of intraductal carcinoma was made sonographically in view of the finding of irregular ductal dilatation with solid tubular echoes within (Fig. 2). Ultrasonography also allowed diagnosis of a clinically palpable but mammographically occult 7 mm focus of recurrent ductal carcinoma in situ in the same patient 6 months later. A single patient with lobular carcinoma in situ was excluded from computation of results of breast cancer diagnosis (Tables 5 and 6), as this is regarded as a marker of increased Table 3: Age Distribution of 67 Women with Breast Cancer Age (years) Under Over SO Number(%) 2 (3.0) 10 (14.9) 14 (20.9) 41 (61.2) Includes one patient with lobular carcinoma in situ. risk for invasive carcinoma elsewhere in the breast rather than as a premalignant lesion. This lesion is often an incidental finding on histologic examination and has no typical clinical or radiologic features,6 Eleven false-positive cases were found, seven of which were also false-positive on x-ray mammography with breast ultrasonography. Histologic examination in these cases revealed fibroadenomas in five patients, florid fibrocystic change in three (including sclerosing adenosis in one), ductal papilloma in two, and abscess in one patient. X-ray Mammography X ray mammography correctly indicated the presence of malignancy in 54 of 59 patients with proved malignancy. Eight patients with histologically proved cancer had not undergone preoperative x-ray mammography either because the tumor was too large and ulcerative or because the patient was unable to sit up for the examination. Typical features of malignancy were noted in 40 patients and indeterminate features in 14 patients. One patient had asymmetrical breast density as the only indication of malignancy, in two patients a mass was interpreted as benign, and in two patients x-ray mammography was negative for malignancy owing to density of breast tissue. Ten additional patients had falsepositive results for malignancy; two of these patients had microcalcifications and eight had a poorly defined mass or trabecular distortion. Comparison Between X-ray Mammography and Ultrasonography Two patients with cancer were misdiagnosed by both imaging techniques. In one, a malignant mass Table 4: Histopathology of 67 Malignant Breast Tumors Histological Type Number (%) Carcinoma in situ Ductal 1 (1.5) Lobular 1 (1.5) Invasive Ductal 38 (56.7) Lobular 3 (4.5) Mucinous 6 (9.0) Invasive carcinoma + carcinoma in situ 14 (20.9) Carcinoma 4 (6.0) (FNAC, histology not available} FNAC, Fine needle aspiration cytology

4 MO PALPABLE BREAST MASSES IN CHINESE WOMEN J Ultrasound Med 15: , 1996 Table 5: Comparison of Modalities According to Cancer Size Ultrasonography Mnmmography Clinical Examinntion Carcinomas Number TP (%) FN (%) TP ('k) FN (%) TP (%) FN (%) < 1 cm 4 4/ 4 (100) 0 (0) 4/ 4 UOO) 0 (0) 3/ 4 (75) 1/ 4 (25) 1 - < 2cm 20 19/20 (95) 1/ 20 ( ) 16/ 19 (84) 3/ 19 (16) 19/ 20 (95) 1/ 20 (5) 2 - < 5cm 40 39/ 40 (98) 1/40 (2) 33/ 35 (94) 2/35 (6) 34/ 40 (85) 6/40 {15) Ol: 5cm (100) 1i1lnl 66. T ', True po10itivc; FN, false negative. Exd udl's one pntient with lobular carcinoma in situ. 0 (0) 1/1 (100) 0 (0) 2/ 2 (100) 0 (0) was diagnosed incorrectly as benign on both x-ray mammography and ultrasonography. In the second patient, a mass obscured by dense breast on x-ray mnmmography was wrongly diagnosed as benign on ultrasonography (Fig. 3). Overall, in breast cancer, x-rny mammography and ultrasonography were in agreement in 54 patients. Comparison Between Clinical Palpation and Ultrasonography Clinicnl palpation misdiagnosed nine of 66 histologically proved cancers as benign. The false-negative cases ind uded six fibroadenomas, one scar or fibro sis, one cyst, and one subcutaneous nodule. Clinical palpation additionally diagnosed cancer in 28 patients. Histologic examination revealed cysts in four cases, abscesses in four, fibroadenomas in five, fibrocystic change in six, ductal papillomas in two, phyllodes tumor in one, inflammation in three, and benign findings in three. Accuracy of combined evaluation methods was evaluated ('Table 5). Ultrasonography was significantly more sensitive and specific than clinical palpation (P < 0.05). The combination of ultrasonography and clinical palpation was more specific than that of x-ray mammography and clinical palpation (P < 0.05). Performing x-ray mammography )n addition to combined ultrasonography and clinical palpation did not significantly improve the results, although ultrasonography in addition to combined x-ray mammography and clinical palpation wa ~ significantly more specific (P < 0.05). Benign Conditions Ultrasonography had a sensitivity of 90% (45 of 50 patients) in the diagnosis of cysts, and a sensitivity of 82% (139 of 169 patients) in the diagnosis of fibroadenomas. In 23 patients, ultrasonography added information to that obtained with x-ray mammography and clinical palpation by revealing additional cysts and fibroadenomas. In 34 patients, ultrasonography revealed a diagnosis (e.g., cyst, fibroadenoma) obscured by radiographic breast density. Ultrasonography excluded a mass lesion in 40 patients with a palpable lump in a radiodense breast. Table 6: Comparison of Sensitivity, Specificity, Positive Predictive Value, Negative Predictive Value, and Accuracy Between Sing1e- and Multimodality Evaluation Methods in the Diagnosis of Cancer St!nsitivity Spccificity PPV NPV Accuracy Ultr.1sonogrn phy Mammogrnphy CUnical examination Ultrasonography plus clinical examination Mammography plus clinlcal examination Ultrasonography plus mammography plus clinical examination PPV, Positive predictive value; NPV, Negative predictive viilue.

5 J Ultrasound Med 15: , 1996 A YANGETAL 641 B Figure 1 A, Mammogram shows a dense mass with poorly defined margins and associated clusters of pleomorphic microcalcifications within. Separate foci of pleomorphic microcalcifications are also noted posterior to the mass (arrowheads). 8, Corresponding ultrasonogram in the same patient shows an 8 x 9 mm focal hypoechoic mass with irregular margins and increased depth to width ratio, suggesting a malignant lesion. Pathologic examination showed invasive ductal carcinoma. Ultrasonography failed to demonstrate the separate focus of microcaldfications seen on mammography, which was found to be ductal carcinoma in situ at pathologic examination. DISCUSSION Breast ultrasonography has evolved tremendously and gained clinical acceptance over the decades. Contrary to the previously held belief that it is only capable of detecting palpable breast cancer, ultrasonography can now detect not only early tumors under 1 cm in diameter but also intraductal tumor components using high-resolution real-time systems.7.s This claim, however, cannot be generalized as the accuracy of breast ultrasonography depends on the operator's experience and equipment used (rates range from 58% to 95'7 ).2.~ Current real-time high-resolution annular array sector Figure 2 A, Mammogram of right breast shows abnormal patchy tubular opacities in the inner quadrant, without any foci of microcalcifications or a dominant mass lesion. B, Corresponding ultrasonogram of the breast shows patchy tubular hypoechoic areas (short white arrows) corresponding to dilated ducts with internal echoes. Pathologic study revealed multiple foci of ductal carcinoma in situ.

6 642 PALPABLE BREAST MASSES IN CHINESE WOMEN systems produce optimal images in terms of resolution, contrast, and speckle pattern, that allow small lesions under 1 cm to be visualized in accurate detaij.s These technical features certainly contribute to the high sensitivity and accuracy of our results, and in addition, characteristic features of breast cancer have become better defined. Other contributing factors to the good results include the generally larger size of lesions evaluated, compared to the usually much smaller tumors detected by screening in the West, and the differences in size and glandular composition of typical Asian and Western breasts.! This study shows that despite reports to the contrary,m,n different types of breast lesions do have specific sonographic features tha t allow a correct diagnosis of cancer to be made with a sensitivity and specificity of 97% and 97%, respectively. This compares with reported accuracies of sonographic diagnosis of breas t cancer in palpable masses, which range from 72% to 92 %.1.12~1 5 Sickles and coworkers, however, reported an accuracy of only 58%.9 Multiple factors contribute to this large variation in rates, including ultrasound scanner technology, tumor size, criteria for characterization of suggestive lesions, and characteristics of the tumor and surrounding breast tissue. JUltrasound Med 1 5:637~, 1996 In the present study, two of 67 histologically proved cancers were misdiagnosed. Both lesions were diagnosed incorrectly as fibroadenomas, but sonographically guided fine needle aspiration cytology provided the true diagnosis. Both lesions were well-defined nodules with a homogeneous echotexture; hence, the impression was that they were most likely benign. In retrospect, however, failure to appreciate subtle margin irregularities and microlobulations in one patient (Fig. 3) probably accounted for the false-negative diagnosis. The characteristic ultrasonographic features of breast cancer have been well described, most commonly consisting of a hypoechoic tumor nidus, a surrounding hyperechoic halo, and distal acoustic shadowing.:!. 4 The larger tumors tend to show a heterogeneous internal echo pattern, with variable through transmission. There may be either shadowing, enhancement, or no effect. It is the carcinomas with no associated secondary changes that represent the greatest diagnostic challenge and controversy for ultrasonography. These tumors need to be differentiated from benign masses such as fibroadenomas and diffuse benign disease.:z~ Sonographically guided fine needle aspira.tion cytology has a definite role in such situations, as it can further improve the specificity.7 Figure 3 A, Mammography shows bilateral dense breasts with no visible dominant masses, architectural distortion, or dusters of microcalcifications suggestive of malignancy. Minimal asymmetry of breast density, increased on the right, is noted. 8, A focal, ovoid, well-defined hypoechoic right breast lesion was seen in the same patient on ultrasonography. This lesion was considered benign but was found to be infiltrative ductal carcinoma at pathologic examination. In retrospect, failure to appreciate the subtle anterior marginal irregularity and microlobulations (nrrowlrrnds) probably accounts for the false-negative diagnosis. A B

7 J Ultrasound Med 15: , 1996 YANGETAL 643 Our x-ray mammographic sensitivity of 92% is compatible with established sensitivities of 80 to 90% for detection of carcinoma of the breast.9,t6,t7 The false-negative rate is higher in women 50 years old and younger due to obscuration of breast lesions by the higher density breast parenchyma in these women.ts,t9 The ability of ultrasonography to image the radiodense breast has led to its use as the first investigation in patients under 35 years of age in several centers.2o.2t Breast ultrasonography is of particular value in investigating an area of radiologic density as well as in patients with clinically evident nodular breasts considered to represent fibroadenosis. Its main advantage is in the exclusion of a circumscribed mass lesion in these areas. Ultrasonography has the highest accuracy in the diagnosis of benign lesionsto.n.n and is particularly accurate in the diagnosis of simple cysts,to.n which can decrease the rate of unnecessary biopsies for benign lesions by up to 25%.24 Its further advantages are the absence of radiation hazard and the ability to precisely guide aspiration of any suggestive lesion under direct vision. In this study, it can be seen that ultrasonography is equivalent to x-ray mammography in diagnostic accuracy. When malignancy was present, no tumor mass was missed by ultrasonography. In its ability to detect small tumor masses (less than 1 cm), ultrasonography was equal to x-ray mammography. The smallest cyst identified was 3 mm, whereas the smallest malignancy detected was 7 mm in diameter. The major limitation of ultrasonography was in the diagnosis of associated foci of ductal carcinoma in situ without a palpable mass, owing to its inability to visualize microcalcifications outside a mass.to For this reason, ultrasonography has a limited role in breast screening. In conclusion we have demonstrated a high diagnostic accuracy of breast ultrasonography in the diagnosis of palpable breast cancer (sensitivity 97% and specificity 97%) as well as in the evaluation of benign breast conditions (overall accuracy 84%). Its role in the evaluation of a palpable breast mass is not only to define the size of a lesion and extent of disease with accuracy but also allow accurate guidance for biopsy. These results further suggest that the use of fine needle aspiration cytology may be limited to only the indeterminate, somewhat suggestive, or highly suggestive masses without significant detriment to the patient. The addition of x-ray mammography to clinical palpation and ultrasonography did not improve the sensitivity, specificity, or positive predictive value, underlining the limited usefulness of mammography in Chinese patients. Whether it should be eliminated in the workup of a palpable breast mass needs to be evaluated further by larger scale studies. REFERENCES 1. Alagaratnam TT, Wong J: Limitations of mammography in Chinese females. Clin Radiol36:175, Harper AP, Kelly Fry E, Noe JS, et al: Ultrasound in the evaluation of solid breast masses. Radiology 146:731, Harper P, Jackson VP, Bies J, et al: A preliminary analysis of the ultrasound imaging characteristics of malignant breast masses compared with x-ray mammographic appearances and the gross and microscopic pathology. Ultrasound Med Biol8:365, Cole-Beuglet C: Sonographic manifestation of malignant breast disease. Semin Ultrasound 3:51, Cole Beuglet C, Soriano RZ, Kurtz AB, et al: Fibroadenoma of breast-sonomammography correlated with pathology in 122 patients. AJR 140:369, Wheeler JE, Enterline HT: Lobular carcinoma of the breast in situ and infiltrating. Pathol Annu ll(pt 2):161, Fomage BD, Sneige N, Faroux MJ, et al: Sonographic appearance and ultrasound guided fine-needle aspira tion biopsy of breast carcinomas smaller than 1 cm. J Ultrasound Med 9:559, Mad jar H, Ladner HA, Sauerbrei W, et al: A preoperative staging of breast cancer by palpation, mammography and high-resolution ultrasound. Ultrasound Obstet Gynecol3:185, Sickles EA, Filly RA, Callen PW: Breast cancer detection with sonography and mammography: Comparison using state-of-the-art equipment. AJR 140:843, Rosner D, Blaird D: What ultrasonography can tell in breast mass that mammography and physical examination cannot. Surg Oncol 23: Donegan WL: Evaluation of a palpable breast mass. N Engl J Med 327:937, Guyer PB, Dewbury KC, Wanvick D, et al: Direct contact B scan ultrasound in the diagnosis of solid breast masses. Clin Radiol37:451, Kobayashi T: Gray-scale echography for breast cancer. Radiology 122:207, Maturo VG, Zusmer NR, Gilson AJ, et al: Ultrasonic appearances of mammary carcinoma with a dedicated whole-breast scanner. Radiology 142:713, Jellins J, Kossoff G, Reeve TS: Detection and classification of liquid-filled masses in the breast by gray scale echography. Radiology 125:205, Dodd GD: Present status of thermography, ultrasound and mammography in breast cancer detection. Cancer 39:2796, Guyer PB, Dewbury KC: Ultrasound of the breast in the symptomatic and x-ray dense breast. Clin Radiol 36:69, 1985

8 644 PALPABLE BREAST MASSES IN CHINESE WOMEN J Ultrasound Med 15: , Feig SA, Shaber GS, Patchefsky A, et al: Analysis of dinically occult and mammographically occult breast tumors. Aj R 1 28~403, Edeiken S. Mammography and palpable cancer of the breast. Cancer 61:261, Harper P, Kelly-Fry E: Ultrasound visualisation of the breast in symptomatic patients. Radiology 137:465, Kopans DB ~ Early breast cancer detection using techniques other than mammography. AJR 143:465, Van Dam PA, Van Goethem MLA, Kersschot E, et al: Palpable solid breast masses: Retrospective single- and multimodality evaluation of 201 lesions. Radiology 166:435, Cole-Beuglet C, Goldberg BB, Kurtz AB, et a!~ Ultrasound mammography: A comparison with radiographic mammography. Radiology 139:693, Hilton SVW, Leopold GR, Olson LK, et a!: Real-time breast sonography: Application in 300 consecutive patients. AJR 147:479, 1986 Erratum In Yang et al: "High Resolution Sonographic Detection of Axillary Lymph Node Metastases in Breast Cancer" J Ultrasound Med 15:241, 1996, sentence three in Patients and Methods should read as follows: The minimum number of lymph nodes identified at dissection was six; 70% of patients had more than 10 lymph nodes sampled at dissection.