Yeh Eren, Slanetz Priscilla, Kopans Daniel B, Rafferty Elizabeth, Georgian-Smith Dianne, Moy Linda, Halpern Elkan, Moore Richard, Kuter Irene, Taghian Alphonse
Department of Radiology, Massachusetts General Hospital, 15 Parkman St., ACC-219, Boston, MA 02114, USA.
AJR Am J Roentgenol. 2005 Mar;184(3):868-77. doi: 10.2214/ajr.184.3.01840868.
The objective of our study was to determine the relative accuracy of mammography, sonography, and MRI in predicting residual tumor after neoadjuvant chemotherapy for breast cancer as compared with the gold standards of physical examination and pathology.
Forty-one women with stage IIB-III palpable breast cancer were prospectively enrolled in a study investigating the effects of sequential single-agent chemotherapy (doxorubicin followed by paclitaxel or vice versa) on tumor imaging. The study cohort consisted of the first 31 patients (age range, 31-65 years; mean, 45 years) who completed the protocol. All underwent physical examination, mammography, sonography, and MRI before and after receiving each neoadjuvant chemotherapeutic drug. Imaging studies were reviewed by two radiologists using conventional lexicons for lesion analysis, and the findings were compared with clinical response and pathology results.
Complete, partial, and stable clinical response as defined by clinical examination was seen in 15, 14, and two of the 31 patients, respectively. Agreement rates about the degree of response were 32%, 48%, and 55%, respectively, for mammography, sonography, and MRI compared with clinical evaluation and did not differ statistically. Agreement about the rate of response as measured by clinical examination, mammography, sonography, and MRI compared with the gold standard (pathology) was 19%, 26%, 35%, and 71%, respectively. Of the four, MRI agreed with the gold standard significantly more often (p < 0.002 for all three paired comparisons with MRI). When there was disagreement with the gold standard, none of the four exhibited a significant tendency to either under- or overestimate.
MRI appears to provide the best correlation with pathology-better than physical examination, mammography, and sonography-in patients undergoing neoadjuvant chemotherapy. However, MRI may overestimate (6%) or underestimate (23%) residual disease in approximately 29% of the patients (95% confidence interval, 14-48%).
我们研究的目的是确定与体格检查和病理学的金标准相比，乳腺X线摄影、超声和磁共振成像（MRI）在预测乳腺癌新辅助化疗后残留肿瘤方面的相对准确性。
41例可触及的IIB - III期乳腺癌女性前瞻性纳入一项研究，该研究调查序贯单药化疗（多柔比星后接紫杉醇或反之）对肿瘤成像的影响。研究队列由完成方案的前31例患者组成（年龄范围31 - 65岁；平均45岁）。所有患者在接受每种新辅助化疗药物前后均接受了体格检查、乳腺X线摄影、超声和MRI检查。两名放射科医生使用传统术语对影像研究进行病变分析，并将结果与临床反应和病理结果进行比较。
31例患者中，根据临床检查定义的完全、部分和稳定临床反应分别见于15例、14例和2例。与临床评估相比，乳腺X线摄影、超声和MRI对反应程度的一致率分别为32%、48%和55%，且无统计学差异。与金标准（病理学）相比，通过临床检查、乳腺X线摄影、超声和MRI测量的反应率一致率分别为19%、26%、35%和71%。在这四种检查中，MRI与金标准的一致性明显更高（与MRI的所有三次配对比较p < 0.002）。当与金标准不一致时，这四种检查均未表现出明显的低估或高估倾向。
在接受新辅助化疗的患者中，MRI似乎与病理学的相关性最佳——优于体格检查、乳腺X线摄影和超声。然而，MRI可能在约29%的患者中高估（6%）或低估（23%）残留疾病（95%置信区间，14 - 48%）。
