The microvascular invasion pathologically diagnosed in hepatocellular carcinoma (HCC) has not been valued enough in previous studies. And there are no unified pathological diagnostic criteria for microvascular invasion in HCC. Besides that, microvascular invasion is not included as a routine pathological indicator for HCC. C-reactive protein to albumin ratio (CAR) is a new systemic inflammatory factor that is closely related to the proliferation, invasion, metastasis and other malignant biological behaviors of HCC.

To investigate the value of CAR in predicting microvascular invasion in single small HCC.

Participants (n=346) who were pathologically diagnosed with HCC following the resection of single liver tumor (diameter≤5 cm) were selected from the First Affiliated Hospital of Xinjiang Medical University from June 2017 to June 2021. Demographic data and calculated CAR were collected. ROC analysis was performed to estimate the predictive value and optimal cut-off value of CAR for microvascular invasion in single small HCC. All patients were divided into two groups according to the optimal cut-off value of CAR (≥0.03 or <0.03) , and those with similar propensity scores estimated by Logistic model were matched by 1∶1 nearest neighbor matching, yielding two comparison groups with highly similar clinical characteristics, and the microvascular invasion rate between them was compared. Logistic regression was used in the sensitivity analysis of the predictive value of CAR for microvascular invasion in single small HCC before and after the aforementioned matching.

One hundred and thirty-one (37.9%) cases were found with microvascular invasion and other 125 without (62.1%) . ROC analysis revealed that the AUC of CAR in predicting microvascular invasion in single small HCC was 0.787〔95%CI (0.697, 0.877) 〕, with 82.9% sensitivity and specificity 76.4% when the optimal cut-off point was determined as 0.03. Ninety-two of 145 cases with CAR<0.03 and 92 of 201 cases with CAR≥0.03 were successful matched, and the prevalence of microvascular invasion was found to be lower in the former group〔13.0% (12/92) vs 43.5% (40/92) 〕 (χ²=6.314, P=0.013) . Three Logistic models showed that CAR was an independent predictor of microvascular invasion in single small HCC, regardless of whether the participants were matched or not (P<0.05) .

CAR could be used as a predictor for microvascular invasion in single small HCC. CAR≥0.03 may indicate a very high probability of microvascular invasion.

既往对肝癌微血管侵犯病理诊断的重要性重视不够，目前国内外缺乏对微血管侵犯统一的病理诊断标准，也未将微血管侵犯列为病理常规诊断指标。C反应蛋白/白蛋白比值（CAR）作为新型系统性炎性因子，与肝癌的增殖、侵袭转移等恶性生物学行为密切相关。

探讨CAR预测单发小肝癌患者微血管侵犯的价值。

选择2017年6月至2021年6月在新疆医科大学第一附属医院行肝切除术的单个、肿瘤直径≤5 cm的术后病理检查证实为肝癌的患者346例为研究对象。收集患者一般资料，并计算CAR。绘制CAR预测单发小肝癌患者微血管侵犯的受试者工作特征（ROC）曲线，并计算CAR的最佳诊断截点，根据CAR最佳诊断截点将患者进行分组，采用1∶1最近邻居倾向性评分匹配（PSM）法将Logistic模型估计的倾向性评分相近患者进行配对，得到两组间各临床特征比较均衡性较高的样本。比较匹配后两组患者微血管侵犯率，采用Logistic回归分析评估匹配前、后CAR对单发小肝癌患者微血管侵犯的预测价值。

346例患者中微血管侵犯阳性131例（37.9%），微血管侵犯阴性215例（62.1%）。ROC曲线分析结果显示，CAR预测单发小肝癌患者微血管侵犯的灵敏度为82.9%，特异度为76.4%，ROC曲线下面积为0.787〔95%CI（0.697，0.877）〕，最佳诊断截点为0.03。根据CAR最佳诊断截点，将患者分为CAR<0.03组（A组，n=145）和CAR≥0.03组（B组，n=201）。采用1∶1最近邻居PSM法，共92对匹配成功，匹配后两组临床资料均衡。匹配后，B组患者微血管侵犯发生率〔43.5%（40/92）〕高于A组〔13.0%（12/92）〕（χ²=6.314，P=0.013）。采用3种Logistic回归分析结果显示，匹配前、后CAR均为单发小肝癌患者微血管侵犯的独立影响因素（P<0.05）。

CAR作为新型系统性炎症指标，可用于预测单发小肝癌微血管侵犯，当CAR≥0.03时提示单发小肝癌微血管侵犯发生率较高。