CT增强扫描灰度直方图纹理分析技术对结直肠癌术前恶性程度评估及预后的预测价值  

The value of gray histogram texture analysis technique of enhanced CT scanning in preoperative evaluation of malignant degree and prognosis of colorectal cancer

作　　者：徐园园 郑吟诗[1] 李依明[1] 刘阳[2] 李玉舟[1] XU Yuanyuan;ZHENG Yinshi;LI Yiming;LIU Yang;LI Yuzhou(CT Room,Shangqiu First People´s Hospital,Shangqiu,He'nan 476100,China;Department of Nuclear Medicine,Shangqiu First People´s Hospital,Shangqiu,He'nan 476100,China)

机构地区：[1]商丘市第一人民医院CT室,河南商丘476100 [2]商丘市第一人民医院核医学科,河南商丘476100

出　　处：《安徽医药》2024年第10期1961-1967,I0005,共8页Anhui Medical and Pharmaceutical Journal

摘　　要：目的探究CT增强扫描灰度直方图纹理分析技术对结直肠癌(CRC)术前恶性程度评估及对预后的预测价值。方法选取2018年2月至2021年2月于商丘市第一人民医院就诊经术后病理证实的94例CRC病人为研究对象,其中高分化28例、中分化组32例和低分化组34例。统计分析所有病人腹部动脉期CT增强扫描灰度直方图纹理分析特征参数,并采用受试者操作特征曲线(ROC曲线)分析特征参数对CRC病人术前恶性程度的诊断效能及预后的预测价值。采用多因素Cox回归分析筛选影响CRC病人预后不良的因素,构建列线图预测模型并进行评价。结果CT增强扫描灰度直方图均值、峰度、10%分位、50%分位在高、中、低分化三组中差异有统计学意义(142.39±21.78比130.62±18.79比122.34±14.57,0.51±0.14比0.89±0.24比1.11±0.31,129.63±27.42比113.75±22.13比102.64±17.85,145.51±20.39比134.52±17.34比126.09±14.53)(P<0.05)。评估CRC恶性程度时,上述指标均具有诊断效能,且联合诊断效能最佳。低分化、淋巴结转移、中性粒细胞与淋巴细胞比值(NLR)、血小板与淋巴细胞比值(PLR)是CRC病人预后不良的危险因素,而均值、10%分位、50%分位是其保护因素(P<0.05)。均值、10%分位、50%分位联合预测病人预后的曲线下面积(AUC)为0.83[95%CI:(0.76,0.91)],远大于各指标单独预测的AUC值。利用上述指标构建列线图模型,经内部验证(Bootstrap重采样1000次)后可知,该模型预测病人CRC病人预后不良AUC为0.84[95%CI:(0.71,0.92)],灵敏度为82.7%,特异度为75.6%。结论腹部动脉期CT增强扫描灰度直方图纹理分析特征参数在不同恶性程度CRC之间存在明显差异,可用于评估CRC病人术前恶性程度及预测预后,为临床CRC病人术前评估及预后提供了新的可能。Objective To explore the value of enhanced CT histogram texture analysis in evaluating preoperative malignancy degree and predicting prognosis of colorectal cancer(CRC).Methods A total of 94 patients with CRC who were treated in the First People's Hospital of Shangqiu City from February 2018 to February 2021 and confirmed by postoperative pathology were selected as the study objects,including 28 patients in the highly differentiated group,32 patients in the moderately differentiated group and 34 patients in the poorly differentiated group.The feature parameters of the gray histogram of enhanced CT scanning in abdominal arterial phase were analyzed statistically,and the predictive value of characteristic parameters on preoperative malignant degree diagnosis and prognosis of CRC patients was analyzed by receiver operating characteristic curve(ROC curve).Multivariate Cox regression analysis was used to screen the factors affecting the poor prognosis of CRC patients,and a nomogram prediction model was constructed and evaluated.Results There were significant differences in mean value,kurtosis,10%quantile and 50%quantile among high,medium and low differentiation groups(142.39±21.78 vs.130.62±18.79 vs.122.34±14.57,0.51±0.14 vs.0.89±0.24 vs.1.11±0.31,129.63±27.42 vs.113.75±22.13 vs.102.64±17.85,145.51±20.39 vs.134.52±17.34 vs.126.09±14.53,P<0.05).When evaluating the degree of CRC malignancy,the four indicators had certain diagnostic efficacy,and the combined diagnostic efficacy was the best.Low differentiation,lymph node metastasis,neutrophil/lymphocyte ratio(NLR)and platelet-to-lymphoccyte ratio(PLR)were independent risk factors for poor prognosis in CRC patients,while mean,10%and 50%were protective factors(P<0.05).The area under the curve(AUC)of the mean,10%quantile and 50%quantile combined to predict the prognosis of patients was 0.83[95%CI:(0.76,0.91)],which was much higher than the AUC value predicted by each index alone.The above indexes were used to construct a line chart model.After internal verificat

关 键 词：结直肠肿瘤 CT增强扫描 灰度直方图 纹理分析 评估 

分 类 号：R735.34[医药卫生—肿瘤]