Sneige N, Tulbah A
Department of Anatomic Pathology, The University of Texas M.D. Anderson Cancer Center, Houston 77030, USA.
Diagn Cytopathol. 2000 Jul;23(1):29-34. doi: 10.1002/1097-0339(200007)23:1<29::AID-DC7>3.0.CO;2-8.
We investigated the diagnostic utility and accuracy of touch imprints (TIs) prepared from core-needle biopsy (CNB) specimens of nonpalpable breast abnormalities. We reviewed air-dried, Diff-Quik-stained TIs prepared from 172 consecutive CNB specimens obtained with stereotactic or sonographic guidance. Using criteria established for fine-needle aspirates, TIs were categorized as benign, atypical, suspicious, malignant, or unsatisfactory (i.e., showing fewer than six benign epithelial cell clusters or cell distortion). Cytologic diagnoses of TIs were then correlated with the histologic diagnoses of corresponding CNB specimens. CNB specimens were histologically diagnosed as carcinoma (102 cases), benign (59 cases), low-grade phyllode tumor (six cases), and atypical ductal hyperplasia (five cases). TIs were cytologically diagnosed as malignant (63 cases), benign (35 cases), suspicious (19 cases), atypical (18 cases), and unsatisfactory (37 cases). Correlation of the cytologic and histologic diagnoses showed that five TIs diagnosed as benign were false-negative results for histologically diagnosed carcinomas (four cases) and phyllodes tumor (one case). False-negative results were attributed to poor representation of malignant cells. Two TIs diagnosed as suspicious were false results for two histologically diagnosed fibroadenomas. The false suspicious findings resulted from TIs with high cellularity, cytologic atypia, or no familiar (i.e., as seen on fine-needle aspirates) smear pattern. Unsatisfactory TIs were noted in both benign (44%) and malignant (11%) CNB specimens. When lesions categorized as suspicious were grouped with the malignant cases and those classified as atypical were grouped with the negative cases, TI sensitivity and specificity, were 83% and 95%, respectively. Fibroadenomas are difficult to identify on TIs and are likely to be misdiagnosed as suspicious. While high- and intermediate-grade carcinomas are easily categorized using TIs, low-grade carcinomas are best categorized as suspicious because of overlapping cytologic features with proliferative breast lesions. Increased experience with cytologic analysis of TIs improves the accuracy of cytologic diagnoses.
我们研究了从不可触及乳腺异常的粗针活检（CNB）标本制备的触摸印片（TI）的诊断效用和准确性。我们回顾了172例在立体定位或超声引导下获得的连续CNB标本制备的空气干燥、Diff-Quik染色的TI。根据细针穿刺抽吸物建立的标准，TI被分类为良性、非典型、可疑、恶性或不满意（即显示少于六个良性上皮细胞簇或细胞变形）。然后将TI的细胞学诊断与相应CNB标本的组织学诊断进行关联。CNB标本的组织学诊断为癌（102例）、良性（59例）、低级别叶状肿瘤（6例）和非典型导管增生（5例）。TI的细胞学诊断为恶性（63例）、良性（35例）、可疑（19例）非典型（18例）和不满意（37例）。细胞学和组织学诊断的关联显示，5例诊断为良性的TI是组织学诊断为癌（4例）和叶状肿瘤（1例）的假阴性结果。假阴性结果归因于恶性细胞的代表性差。2例诊断为可疑的TI是2例组织学诊断为纤维腺瘤的假结果。假可疑结果是由细胞丰富、细胞学非典型或无熟悉（即细针穿刺抽吸物所见）涂片模式的TI导致的。在良性（44%）和恶性（11%）CNB标本中均发现不满意的TI。当将分类为可疑的病变与恶性病例归为一组，将分类为非典型的病变与阴性病例归为一组时，TI的敏感性和特异性分别为83%和95%。纤维腺瘤在TI上难以识别，很可能被误诊为可疑。虽然高级别和中级别癌很容易通过TI进行分类，但低级别癌由于与增生性乳腺病变的细胞学特征重叠，最好分类为可疑。增加对TI细胞学分析的经验可提高细胞学诊断的准确性。
