Lifrange E, Kridelka F, Colin C
Liège University, Breast Department, C.H.R. Citadelle, Belgium.
Eur J Radiol. 1997 Jan;24(1):39-47. doi: 10.1016/s0720-048x(96)01114-x.
To determine the advantages and limitations of a combined stereotaxic fine-needle aspiration biopsy and needle-core biopsy in the diagnosis of 353 nonpalpable breast lesions with special attention given to the collection of follow-up data.
353 nonpalpable breast lesions underwent 'one pass' stereotaxic fine-needle aspiration (21 gauge needle) and needle-core biopsy (18 gauge needle) at our institution from January 1990 to October 1993. Stereotaxic biopsies were carried out by means of an 'add-on unit'. Surgical biopsy was usually recommended for highly suspicious radiologic patterns and/or needle biopsy reports classified as atypical or malignant. In all other cases mammographic follow-up was advised at 6 months and then annually for 3 years. The data were collected retrospectively during September 1995 (theoretical average follow-up of greater than 3 years).
Following the combined needle biopsy technique procedure, surgery was recommended for 83 lesions. Fifty-four cancers were associated to these suspicious lesions. Because of changing radiological or clinical pattern during follow-up (mean follow-up: 22 months), 11 cancers were detected among the 270 lesions initially considered not to need surgery. Forty-three percent of the 65 malignant lesions were initially read as having less than highly suspicious mammographic features. There was no significant difference between the sensitivity and the specificity of one pass fine-needle aspiration biopsy (57% and 96% respectively) and needle-core biopsy (60% and 97% respectively), but noncontributive samples were not included in the false negative diagnoses and atypical samples were included in the true positive diagnoses. Of the 11 missed cancers, nine were manifested initially by clusters of calcifications. Our diagnostic approach was significantly less sensitive (P = 0.006) and less specific (P = 0.032) in cases of clusters of calcifications (31% false negative diagnoses) than in cases of soft-tissue masses (5.5% false negative diagnoses). In this study, an average delay in diagnoses of 22 months was responsible for a significantly increased percentage of axillary node positive invasive cancer (P < 0.001) and six of the 11 missed cancers were palpable at the time of the delayed diagnosis. For the nine cancers initially manifested by calcifications, the 22 months delay in diagnosis was responsible for a nonsignificant increase of microinvasive type at the expense of carcinoma in situ.
Our enthusiasm with the sensitivity of this double stereotaxic needle sampling has been tempered by the results of this reanalysis in the light of a mean theoretical follow-up of three years. Our diagnostic approach was adequate in the presence of soft-tissue masses but not valid in the presence of clustered calcifications. When dealing with calcifications, multiple samplings must be done in order to improve the sensitivity of the diagnosis. Furthermore, this study does not favour the theory that the majority of mammographically detected cancers are indolent and highlights the poor sensitivity of the mammographic follow-up of nonpalpable lesions.
确定立体定向细针穿刺活检与针芯活检相结合在诊断353例不可触及乳腺病变中的优势与局限性，并特别关注随访数据的收集。
1990年1月至1993年10月，在本机构对353例不可触及乳腺病变进行了“一次穿刺”立体定向细针穿刺（21号针）和针芯活检（18号针）。立体定向活检通过“附加装置”进行。对于高度可疑的放射学表现和/或针吸活检报告分类为非典型或恶性的情况，通常建议进行手术活检。在所有其他情况下，建议在6个月时进行乳腺X线随访，然后每年随访3年。这些数据于1995年9月进行回顾性收集（理论平均随访时间超过3年）。
采用联合针吸活检技术后，建议对83个病变进行手术。54例癌症与这些可疑病变相关。由于随访期间放射学或临床特征发生变化（平均随访时间：22个月），在最初认为无需手术的270个病变中检测到11例癌症。65例恶性病变中有43%最初的乳腺X线特征被解读为低于高度可疑。一次穿刺细针穿刺活检（敏感性和特异性分别为57%和96%）与针芯活检（敏感性和特异性分别为60%和97%）之间的敏感性和特异性无显著差异，但非诊断性样本未纳入假阴性诊断，非典型样本纳入真阳性诊断。在11例漏诊的癌症中，9例最初表现为钙化灶。与软组织肿块病例（假阴性诊断率为5.5%）相比，我们的诊断方法在钙化灶病例中的敏感性显著降低（P = 0.006），特异性也显著降低（P = 0.032）（假阴性诊断率为31%）。在本研究中，平均诊断延迟22个月导致腋窝淋巴结阳性浸润性癌的比例显著增加（P < 0.001），11例漏诊癌症中有6例在延迟诊断时可触及。对于最初表现为钙化的9例癌症，22个月的诊断延迟导致微浸润型癌症略有增加，原位癌有所减少，但差异无统计学意义。
根据三年的平均理论随访结果进行的重新分析，使我们对这种双重立体定向针吸采样敏感性所抱的热情有所降温。我们的诊断方法在存在软组织肿块时是足够的，但在存在簇状钙化时无效。处理钙化时，必须进行多次采样以提高诊断的敏感性。此外，本研究不支持大多数乳腺X线检测到的癌症生长缓慢这一理论，并突出了对不可触及病变进行乳腺X线随访时敏感性较差的问题。
