基于功能磁共振在体评价直肠癌胶原含量—与太赫兹离体谱分析及病理对照研究

Extracellular matrix especially collagen plays a significant role in rectal cancer invasion and metastasis. Analyzing the collagen in rectal cancer will contribute to figuring out the mechanism of tumor invasion thus, in turns assisting individual diagnosis, treatment and prognosis evaluation. However, there is no effective evaluation method of collagen not only in vivo, but also in quantitatively immunohistochemical staining. The new MR imaging technologies such as CEST and DKI et al which have been applied to protein imaging and tissue structure analysis, may have the potential to provide new solutions to evaluate tumor collagen in vivo. Terahertz spectroscopy can be applied to analyze biological materials, and by combing with immunohistochemistry in vitro, these two methods can deliver a quantitative analysis of collagen. This study proposed a collaborative method: to analyze the collagen content in vivo by functional MRI, in the mean time to study the imaging sensitivity and specificity according to the pathological result; and to quantitatively analyze the collagen content in vitro by Terahertz spectroscopy and immunohistochemistry, to improve the evaluation accuracy. This method will objectively provide an effective evaluation of collagen in rectal cancer in vivo, which will effectively guide its treatment.

肿瘤细胞外基质尤其是胶原蛋白在直肠癌浸润及转移过程中起着极其重要的作用。分析癌组织内的胶原蛋白有助于直肠癌浸润机制的研究，并实现个性化诊断、治疗和预后评估。目前不仅缺乏在体对直肠癌胶原的评价方法，离体状态免疫组化法进行胶原的定量分析也存在不足。CEST和DKI等MR成像技术在蛋白显像和组织结构分析中已经获得临床应用，有望成为在体分析肿瘤胶原蛋白的方法。太赫兹谱分析可用于生物材料分析，联合免疫组化可能实现对离体癌组织胶原的定量分析。本研究拟利用功能MR成像实现在体分析直肠癌胶原含量，结合病理结果判断成像的敏感性和特异性；联合太赫兹谱分析与免疫组化进行离体肿瘤胶原定量分析，以提高定量的准确性。从而为客观评价在体直肠癌组织内胶原蛋白提供有效的测定方法，对于指导直肠癌个性化治疗有重要意义。

肿瘤细胞外基质尤其是胶原组织在直肠癌浸润及转移过程中起着极其重要的作用。分析癌组织内的胶原含量有助于直肠癌浸润机制的研究，并实现个性化诊断、治疗和预后评估。目前在术前对直肠癌胶原含量评估受限。本研究对2016/01-2017/12于我院就诊的123名直肠癌患者行常规MRI平扫及增强扫描后进行DWI成像，并对其术后组织病理切片进行Van Gieson染色，从而分析其胶原组织含量，并对入组患者T2WI、DWI、ADC及增强图像上分别对病灶进行3D层面的勾画，提取组学特征5650个，将2016/01-2017/06入组的92名患者纳入原始集，将2017/07-2017/12入组的31名患者纳入验证集，通过对原始集患者进行组学特征和胶原含量的分析，并将预测模型代入验证集中进行验证。最后得出由组学特征构建的胶原含量评估模型，其在原始集中预测准确性为0.88，在验证集中准确性为0.71。本研究通过影像组学分析结合术后标本病理染色分析，建立了直肠癌胶原含量的预测模型，实现了在体直肠癌胶原含量评估，从而为客观评价在体直肠癌组织内胶原含量提供有效的测定方法，对于指导直肠癌个性化治疗有重要意义。