Abstract: Value of enhanced pelvic computed tomography and magnetic resonance imaging in the diagnosis and multivisceral resection of locally recurrent rectal cancer        LI Zhen-yang*，ZHOU Lin-jiang，ZHOU Min-wei，et al. *Department of General Surgery，Huashan Hospital Affiliated to Fudan University，Shanghai 200040，China
Corresponding author：XIANG Jian-bin，E-mail：xjbzhw@163.com
LI Zhen-yang and ZHOU Lin-jiang are the first authors who contributed equally to the article.
Abstract    Objective    To analyze the clinical value of computed tomography (CT) and magnetic resonance imaging (MRI) in patients with locally recurrent rectal cancer (LRRC). Methods    The complete clinical and pathological data of 48 LRRC patients treated by surgery in Huashan Hospital Affiliated to Fudan University from January 2012 to December 2018 were analyzed retrospectively. According to the criteria of Memorial Sloan Kettering Cancer Center (MSKCC)，the patients were divided into central，anterior，posterior，and lateral subtypes. All patients received preoperative abdominal enhanced CT，pelvic plain and enhanced MRI. The preoperative diagnostic accuracy of LRRC and the predictive value of tumor invasion scope by enhanced pelvic CT and MRI were compared. Results    Combined with postoperative pathological data，41 patients (85.4%) were diagnosed with postoperative locally recurrence of rectal cancer in 48 cases. The accuracy of MRI was higher than CT significantly ［91.7% (44/48) vs. 77.1% (37/48)，χ2=3.872，P=0.049］. However，there was no significant difference in the sensitivity ［95.1% (39/41) vs. 87.8% (36/41)，χ2=1.406，P=0.236］ between the two methods. The accuracy of pelvic enhanced CT and MRI in evaluating tumor invasion scope was 70.8% (34/48) and 68.7% (33/48) respectively，with no significant difference (χ2=0.049，P=0.824). The diagnostic accuracy of enhanced CT for tumor invasion was 100% (7/7) in posterior subtypes，71.4% (5/7) in lateral subtypes，69.6% (16/23) in central subtypes and 54.6% (6/11) in anterior subtypes. The diagnostic accuracy of enhanced MRI was 78.3% (18/23) in central subtypes，63.6% (7/11) in anterior subtypes，57.1% (4/7) in posterior subtypes and 57.1%(4/7) in lateral subtypes. All the surgeries were successfully completed，with an R0 resection rate of 85.4% (41/48). Conclusion    Enhanced MRI is recommended as an auxiliary method for qualitative diagnosis of LRRC. Enhanced CT has a high detection rate for posterior subtypes of LRRC with sacrococcygeal invasion，while enhanced MRI has a high accuracy in the determination of central subtypes. Multi-modality image fusion could be helpful to determine the surgical strategy of LRRC.

Key words: locally recurrent rectal cancer, computed tomography（CT）, magnetic resonance imaging（MRI）, multi-modality image fusion

摘要： 目的    探讨盆腔增强CT和MRI在局部复发直肠癌（LRRC）多器官联合切除中的价值。方法    回顾性分析2012 年 1 月至 2018 年 12 月复旦大学附属华山医院外科收治的行手术治疗的48 例LRRC病人的临床资料，病理组织保存完整。LRRC根据纪念斯隆-凯特琳癌症中心（MSKCC）标准分为中央型、前向型、后向型和侧向型。术前均行腹部增强CT和盆腔平扫+增强MRI检查。比较CT和MRI对LRRC术前诊断准确率、对邻近器官侵犯的预测价值，并分析二者融合图像对手术的指导作用。结果    48例病人中41例（85.4%）经术后病理学检查确诊直肠癌术后复发。MRI和CT对LRRC检出率分别为95.1%（39/41）和87.8%（36/41），差异无统计学意义（χ2=1.406，P=0.236），MRI诊断LRRC的准确率高于CT［91.7%（44/48） vs. 77.1%（37/48），χ2=3.872，P=0.049］。盆腔增强CT及MRI鉴别LRRC累及邻近器官的准确率分别为70.8%（34/48）和 68.7%（33/48），差异无统计学意义（χ2=0.049，P=0.824）。各 LRRC 亚型分析提示，增强CT的鉴别准确率：后向型100%（7/7）、侧向型71.4%（5/7）、中央型69.6%（16/23）、前向型54.6%（6/11）；盆腔增强MRI的鉴别准确率：中央型78.3%（18/23）、前向型63.6%（7/11）、后向型57.1%（4/7）和侧向型57.1%（4/7）。所有病人均顺利完成手术，R0切除率为 85.4%（41/48）。结论    推荐增强MRI检查作为 LRRC诊断辅助手段；增强CT对后向型LRRC骶尾骨侵犯检出率高，而增强MRI判断中央型LRRC准确率较高；多模态融合图像有助于确定LRRC手术切除范围并制定策略。

关键词: 局部复发直肠癌, 计算机断层扫描, 磁共振成像, 多模态影像融合