探究CT增强图像灰度直方图纹理分析技术在术前评估胰腺癌恶性程度的可行性，评价灰度直方图参数和胰腺癌病理分级、分化程度的关系。

回顾性分析扬州大学附属苏北人民医院术后病理证实的胰腺癌患者49例，其中高分化组27例，中-低分化组22例。选取2组肿瘤CT增强扫描静脉期横断位图像最大层面，采用Mazda软件勾画感兴趣区，并进行灰度直方图纹理分析，对高分化、中-低分化两组直方图特征参数进行统计分析。对有统计学意义的参数建立受试者工作特征曲线(receiver operating characterist，ROC)，最终对各参数其诊断效能进行比较分析。此外，对差异有统计学意义的参数与高分化、中-低分化组别之间的相关性进行检测。

采用灰度直方图分析提取的9个参数特征中，其中均值(Mean)、第1百分位数(Perc.01%)、第10百分位数(Perc.10%)、第50百分位数(Perc.50%)在高分化、中-低分化2组的差异有统计学意义(P<0.05)，变异度(Variance)、偏度(Skewness)、峰度(Kurtosis)、第90百分位数(Perc.90%)、第99百分位数(Perc.99%)在2组中差异无统计学意义(P>0.05)。采用Spearman相关性分析可知均值、Perc.01%、Perc.10%、Perc.50%与分化程度之间呈正相关(r值分别为0.04、0.29、0.32、0.33，P均<0.05)。评估胰腺癌恶性程度时，均值的诊断效能最高(AUC＝0.695)；当均值取值147.27时，其对应的敏感度和特异度分别为86.4%、44.4%；当Perc.01%取值117时，其对应的敏感度和特异度分别为95.5%、40.7%；当Perc.10%取值120时，其对应的敏感度和特异度分别为72.7%、63%；当Perc.50%取值146时，其对应的敏感度和特异度分别为86.4%、44.4%。

CT增强图像灰度直方图分析的特征参数在高分化、中-低分化胰腺癌之间存在差异，给术前评估胰腺癌恶性程度提供了新的方法。

To explore the feasibility of grey histogram texture analysis based on CT contrast enhancement in preoperative evaluation of the degree of malignancy of pancreatic cancer, and to evaluate the relationship between grey histogram parameters and pathological grade and differentiated degree of pancreatic cancer.

A total of 49 patients with postoperative pathology proven pancreatic cancer including high differentiation(27 cases)and low-moderate differentiation(22 cases)were analyzed retrospectively.Region of interest(ROI)was chosen at axial CT images with maximum enhancement of lesion and grey histogram analysis was performed using Mazda software.The histogram parameters were analyzed statistically in these two groups.Receiver operating characteristic(ROC)curve was established for parameters to compare diagnostic performance.In addition, the correlation was tested between parameters and groups.

In the 9 parameters extracted from grey histogram analysis among the two groups, the mean value, Perc.01%, Perc.10%, Perc.50% had statistical differences(P<0.05), but variance, skewness, kurtosis, Perc.90%, Perc.99% had not statistical differences(P>0.05). Spearman correlation analysis showed that the mean value, Perc.01%, Perc.10%, Perc.50% were positively correlated with the differentiated degree(r values were 0.04, 0.29, 0.32 and 0.33 respectively, P<0.05). When evaluating malignant grade of pancreatic cancer, the diagnostic efficiency of the mean value was the highest(AUC=0.695). When the cutoff value was 147.27 for mean value, the sensitivity and specificity were 86.4% and 44.4%.When the cutoff value was 117 for Perc.01%, the sensitivity and specificity were 95.5% and 40.7%.When the cutoff value was 120 for Perc.10%, the sensitivity and specificity were 72.7% and 63%.When the cutoff value was 146 for Perc.50%, the sensitivity and specificity were 86.4% and 44.4%.

The characteristic parameters of gray histogram texture analysis based on CT enhanced images are different between high and low-moderate grade pancreatic cancer, providing a new method for preoperative evaluation of the degree of malignancy of pancreatic cancer.