Crippa F, Agresti R, Seregni E, Greco M, Pascali C, Bogni A, Chiesa C, De Sanctis V, Delledonne V, Salvadori B, Leutner M, Bombardieri E
Division of Nuclear Medicine, National Cancer Institute, Milan, Italy.
J Nucl Med. 1998 Jan;39(1):4-8.
The presurgical, noninvasive staging of axillary nodes for metastases was prospectively investigated in 68 patients who were diagnosed with primary breast cancer using PET with 18F-fluorodeoxyglucose (FDG). Four patients had bilateral nodules; therefore, the total number of evaluable cases was 72. Visual analyses of attenuation-corrected PET images and standardized uptake values (SUVs) of FDG uptake in carcinomas were compared with histopathological surgical findings. The SUV distribution differences between carcinomas with and without axillary metastases were evaluated by means of statistical and receiver operating characteristics analyses.
PET correctly classified 64 of the 72 cases; four false-positive and four false-negative PET results were found. The overall sensitivity, specificity and accuracy of PET for axillary metastases were 85%, 91% and 89%, respectively. With respect to the clinical axillary stage of the patients (TNM, or tumor-node-metastasis, classification), we obtained the following results: N0 patients, sensitivity = 70%, specificity = 92%, accuracy = 86%; N1a patients, sensitivity = 85.5%, specificity = 100%, accuracy = 95%; and N1b-2 patients, sensitivity = 100%, specificity = 67%, accuracy = 87%. The median SUV in carcinomas with axillary metastases (4.6) was significantly higher than that in carcinomas without metastases (2.9), but there was a great SUV overlap between the two groups (interquartile ranges = 2.7-7.2 and 1.9-4.5, respectively). Analysis of the receiver operating characteristics curve showed that a high sensitivity of SUV in predicting axillary metastases was associated with a very low specificity and vice versa. With the best SUV cutoff value of 2.9, the sensitivity and specificity were 74% and 56%, respectively.
PET showed good overall diagnostic accuracy in the detection of axillary metastases (86%). The very high accuracy (95%) in N1a patients is of particular importance. False-negative PET findings, however, can be encountered. SUVs of breast carcinoma cannot predict the spread of the disease to the axilla, even if higher values are often associated with axillary metastases. Any decision on the use of PET in the presurgical staging of breast cancer should be incorporated into a more general debate on axillary management. In selected patients with a very low probability of axillary metastases (T1a), in whom axillary surgery can already be avoided according to data from follow-up studies, 18F-FDG PET could be proposed as a noninvasive imaging modality to improve the diagnosis of axillary relapses.
对68例经18F-氟脱氧葡萄糖（FDG）PET诊断为原发性乳腺癌的患者进行前瞻性研究，以对腋窝淋巴结转移进行术前无创分期。4例患者有双侧结节；因此，可评估病例总数为72例。将衰减校正后的PET图像的视觉分析以及癌组织中FDG摄取的标准化摄取值（SUV）与组织病理学手术结果进行比较。通过统计分析和受试者工作特征分析评估有腋窝转移和无腋窝转移的癌组织之间的SUV分布差异。
PET正确分类了72例中的64例；发现4例假阳性和4例假阴性PET结果。PET对腋窝转移的总体敏感性、特异性和准确性分别为85%、91%和89%。关于患者的临床腋窝分期（TNM，即肿瘤-淋巴结-转移分类），我们得到了以下结果：N0期患者，敏感性 = 70%，特异性 = 92%，准确性 = 86%；N1a期患者，敏感性 = 85.5%，特异性 = 100%，准确性 = 95%；N1b - 2期患者，敏感性 = 100%，特异性 = 67%，准确性 = 87%。有腋窝转移的癌组织的SUV中位数（4.6）显著高于无转移的癌组织（2.9），但两组之间存在很大的SUV重叠（四分位间距分别为2.7 - 7.2和1.9 - 4.5）。受试者工作特征曲线分析表明，SUV在预测腋窝转移时高敏感性与非常低的特异性相关，反之亦然。最佳SUV临界值为2.9时，敏感性和特异性分别为74%和56%。
PET在检测腋窝转移方面显示出良好的总体诊断准确性（86%）。N1a期患者的极高准确性（95%）尤为重要。然而，可能会出现PET假阴性结果。乳腺癌的SUV不能预测疾病向腋窝的扩散，即使较高的值通常与腋窝转移相关。关于在乳腺癌术前分期中使用PET的任何决策都应纳入关于腋窝处理的更广泛讨论中。在腋窝转移可能性非常低（T1a）的特定患者中，根据随访研究数据已经可以避免腋窝手术，18F - FDG PET可作为一种无创成像方式来改善腋窝复发的诊断。
