1 Department of Radiology, Affiliated Cancer Hospital of Zhengzhou University, Henan Cancer Hospital, 127 Dongming Rd, Zhengzhou, China 450008.
2 Department of Thoracic Surgery, Affiliated Cancer Hospital of Zhengzhou University, Henan Cancer Hospital, Zhengzhou, China.
AJR Am J Roentgenol. 2019 Apr;212(4):788-795. doi: 10.2214/AJR.18.20204. Epub 2019 Jan 23.
The purpose of this study was to explore the value of 3-T MRI for evaluating the preoperative T staging of esophageal cancer (EC) treated with neoadjuvant chemotherapy (NAC), with histopathologic confirmation.
This prospective study enrolled patients for whom endoscopic biopsy showed EC and pretreatment CT showed stage cT1N+M0 or cT2-T4aN0-N3M0. All patients received two cycles of NAC (paclitaxel and nedaplatin protocol) followed by 3-T MRI and surgical resection. Readers assigned a T category on MRI, and postoperative pathologic confirmation was considered the reference standard. Interreader agreement, the diagnostic accuracy of T staging on T2-weighted turbo spin-echo (TSE) BLADE (Siemens Healthcare), contrast-enhanced StarVIBE (Siemens Healthcare), high-resolution delayed phase StarVIBE, and the combination of the three sequences were analyzed and compared with postoperative pathologic T staging.
The study included 79 patients. Mean time between NAC and MRI was 23 days. Interreader agreements of T category assignment were excellent for T2-weighted TSE BLADE (κ = 0.810, p < 0.0001), contrast-enhanced StarVIBE (κ = 0.845, p < 0.0001), high-resolution delayed phase StarVIBE (κ = 0.897, p < 0.0001), and the combination of the three sequences (κ = 0.880, p < 0.0001). The highest accuracy for T0, T1, T2, and T4a lesions was on high-resolution delayed phase StarVIBE (96.2%, 92.4%, 91.1%, and 91.1% for reader 1; 94.9%, 89.9%, 91.1%, and 94.9% for reader 2), and the highest accuracy for T3 lesions was on T2-weighted TSE BLADE (92.4% and 94.9% for reader 1 and reader 2, respectively). Diagnostic accuracy of the combination of the three sequences was not improved compared with individual sequences.
High-resolution delayed phase StarVIBE had the highest diagnostic accuracy in staging EC after NAC for all T categories except T3, for which T2-weighted TSE BLADE had the highest accuracy. Combining all three sequences did not improve diagnostic accuracy.
本研究旨在探讨术前使用新辅助化疗（NAC）治疗食管癌（EC）的 3T MRI 对 T 分期的评估价值，并通过组织病理学证实。
本前瞻性研究纳入经内镜活检诊断为 EC 且术前 CT 显示 cT1N+M0 或 cT2-T4aN0-N3M0 的患者。所有患者均接受两个周期的 NAC（紫杉醇和奈达铂方案），随后进行 3T MRI 和手术切除。读者在 MRI 上分配 T 分期，术后病理证实为参考标准。分析并比较 T2 加权涡轮自旋回波（TSE）BLADE（西门子医疗）、对比增强 StarVIBE（西门子医疗）、高分辨率延迟期 StarVIBE 和三种序列联合的 T 分期的诊断准确性。
该研究纳入 79 例患者。NAC 与 MRI 之间的平均时间为 23 天。T 分期分配的读者间一致性极好，T2 加权 TSE BLADE（κ=0.810，p＜0.0001）、对比增强 StarVIBE（κ=0.845，p＜0.0001）、高分辨率延迟期 StarVIBE（κ=0.897，p＜0.0001）和三种序列联合（κ=0.880，p＜0.0001）。T0、T1、T2 和 T4a 病变的最高准确性是高分辨率延迟期 StarVIBE（读者 1 为 96.2%、92.4%、91.1%和 91.1%；读者 2 为 94.9%、89.9%、91.1%和 94.9%），T3 病变的最高准确性是 T2 加权 TSE BLADE（读者 1 和读者 2 分别为 92.4%和 94.9%）。三种序列联合的诊断准确性并未优于单独序列。
高分辨率延迟期 StarVIBE 对所有 T 分期（除 T3 外）的 NAC 后 EC 分期具有最高的诊断准确性，而 T2 加权 TSE BLADE 对 T3 分期具有最高的准确性。联合使用所有三种序列并未提高诊断准确性。
