基于多光子成像的胃浆膜面胶原变化与胃癌腹膜转移的相关性研究

About 44-54% advanced gastric cancer patients will develop to peritoneal metastasis after radical gastrectomy. However, it is difficult to predict which patient will have peritoneal metastasis after radical gastrectomy. Therefore, it is urgent in clinc to build up a prediction model to predict peritoneal metastasis after surgery. In our previous study, we found that serosal collagen played an importantl role in gastric cancer peritoneal metastasis by using multiphoton imaging and collagen quantification. Density of serosal collagen was significantly increased with uneven distribution, collagen area and collagen crosslink were also increased, and collagen arrangement was inclined to the orientation of tumor invasion in gastric cancer with peritoneal metastasis, compared with gastric cancer without peritoneal metastasis. We then hypothesize that alteration of collagen area, collagen density and collagen arrangement in gastric serosa is closely associated with gastric cancer peritoneal metastasis, and peritoneal matastasis can be precisely predicted by collagen quantification. We propose to further validate the correlation between serosal collagen alteration and peritoneal metastasis in different pathological type of gastric cancer. Then, based on multiphoton imaging and collagen quantification, we will construct a prediction model to predict peritoneal metastasis and find the machnism between serosal collagen alteration and perotoneal metastasis in animal experiments. Finally, a prospective clinical trial will be conducted to confirm the sensitivity, specificity, accuracy, positive predictive value and negative predictive value of the prediction model using multiphoton imaging and collagen quantification. Our research is expected to offer a novel approach to predict gastric cancer peritoneal metastasis after radical gastrectomy, which will be helpful for decision-making for gastric cancer patients after radical gastrectomy.

进展期胃癌患者在接受根治术后发生腹膜转移的概率约44-54%。如何早期判断胃癌根治术后是否会发生腹膜转移，是目前临床精准治疗急需解决的问题。我们前期通过多光子成像及胶原定量分析发现：有腹膜转移组胃癌较无腹膜转移组胃癌胃浆膜面胶原密度增加且分布不均，胶原面积增大、交联增多，胶原纤维排列更倾向于肿瘤侵袭方向。因此，我们假设：胃癌浆膜面胶原面积、胶原密度和胶原纤维排列方向的改变与腹膜转移密切相关，通过胶原定量可以预测胃癌的腹膜转移。我们拟在前期研究基础上进一步验证不同生物学特性的胃癌浆膜面胶原密度、胶原面积和胶原纤维排列方向与腹膜转移的关系，明确引起这些胶原变化的相关机制，构建基于多光子成像和胶原定量分析的胃癌术后腹膜转移预测模型，进而开展前瞻性临床试验验证胃癌术后腹膜转移预测模型的敏感性、特异性、准确性、阳性预测值、阴性预测值，从而为胃癌术后腹膜转移的预判及精准治疗提供帮助。

进展期胃癌根治术后发生腹膜转移的概率约44-54%，腹膜转移是进展期胃癌患者死亡的主要原因。目前临床尚缺乏一种可预判胃癌根治术后是否发生腹膜转移的方法。本项目应用多光子成像技术及胶原定量分析进行胃癌胃浆膜面胶原变化与腹膜转移的相关性研究，并构建胃癌术后腹膜转移的预测模型。我们从训练集中的198例T4期胃癌患者的胃浆膜面选取与肿瘤交界的区域，通过多光子成像，提取胶原的形态特征及纹理特征，运用LASSO回归筛选出4个关键胶原特征，构建胶原特征评分，通过竞争风险回归分析胶原特征评分与腹膜转移的关系，构建基于胶原特征评分的T4期胃癌腹膜转移的预测模型列线图，并在验证集的115例T4期胃癌患者中对预测模型进行验证。我们发现：胶原特征评分与T4期胃癌术后腹膜转移密切相关，胶原特征评分越高，术后发生腹膜转移的风险也越大，预后也越差。在训练集中，高胶原特征评分组的3年腹膜转移累积发生率（58.66%，95% CI：52.81%-64.50%）显著高于低胶原特征评分组（29.69%，95% CI：25.71%-33.66%），差异具有统计学意义（SHR：2.59; P <0.001）。在验证集中，高胶原特征评分组的3年腹膜转移累积发生率（52.63%；95% CI：41.03%-64.23%）也显著高于低胶原特征评分组（22.92%；95% CI：18.69%-27.14%），差异也具有统计学意义（SHR：3.01；P=0.004）。胶原特征评分是腹膜转移的独立预测因素，由胶原特征评分、肿瘤大小、分化程度及淋巴结N分期共同组成的预测模型，其总体敏感性、特异性、准确性、阳性预测值、阴性预测值分别为82.0%、72.4%、75.8%、62.1%、88.0%。该模型在训练集和验证集的C-index分别为0.794和0.716，可较好地预测T4期胃癌术后1年、2年和3年内的腹膜转移概率，为胃癌术后腹膜转移的预判及精准治疗提供帮助。.本项目共发表标注由国自然基金（81773117）资助的SCI论文15篇（代表作Nature Communications、JAMA Surgery、JAMA Network Open），获得国家授权发明专利3项，完成预期目标（SCI 3篇，专利2项）。.综上，本项目按计划如期完成任务，顺利结题。