CT平扫和增强图像对480例患者肺磨玻璃结节的诊断价值

目的 探讨CT平扫及增强扫描对肺磨玻璃结节(GGN)良恶性及侵袭性的预测价值。 方法 回顾性分析2018年1月至2020年12月经手术病理证实、CT表现为GGN的480例患者,共540个GGN。根据病理分为良性组、非侵袭组与侵袭组,其中良性组57个、非侵袭组(不典型腺瘤样增生、原位腺癌、微浸润性腺癌)310个、侵袭组(浸润性腺癌)173个。记录一般资料(年龄、性别、吸烟史、肿瘤家族史、发病部位)、磨玻璃成分占比、形态、边界、分叶、毛刺、空泡征、支气管异常征、内部血管征、胸膜牵拉征、最长径、最短径、平扫CT值、增强动脉期CT值、静脉期CT值、强化程度ΔCT_A-N(CT值_动脉期-CT值_平扫)、ΔCT_V-N(CT值_静脉期-CT值_平扫)。采用ANOVA单因素方差分析、Kruskal-Wallis H检验、Pearson χ2检验评估计量资料及计数资料,组内相关系数(ICC)评估测量重复性。分别以结节良恶性及结节侵袭性为因变量,以有关因素为自变量进行二分类Logistic回归分析,并对有统计学意义的参数进行ROC曲线分析,计算曲线下面积(AUC)以及病理学诊断为金标准的诊断指标评估(包括临界值、灵敏度及特异度)。 结果 良性组、非侵袭组、侵袭组之间各结节定性及定量参数比较差异均有统计学意义(P<0.001)。Logistic回归分析结果显示,良性组与非侵袭组、侵袭组之间各结节磨玻璃成分占比、内部血管征、平扫CT值及ΔCT_V-N差异有统计学意义(P<0.001);ΔCT_V-N的ROC曲线下面积(AUC)为0.874,灵敏度为0.747,特异度为0.877;侵袭组与非侵袭组之间各结节最长径、分叶、边界、血管异常征差异有统计学意义,其中最长径AUC为0.851,灵敏度为0.746,特异度为0.813。 结论 CT平扫结合增强扫描对结节良恶性及侵袭性的预测具有重要意义,其中GGN强化程度ΔCT_V-N对良恶性的预测效能更高,最长径对侵袭性的预测效能更高。

Diagnostic value of unenhanced and enhanced CT images in ground glass pulmonary nodules

Objective To explore the predictive value of unenhanced CT scan and contrast-enhanced CT scan in benign, malignant and invasive ground glass pulmonary nodules(GGN). Methods The CT images of 480 patients with 540 GGNs who underwent lung curative resection during Jan. 2018 and Dec. 2020 were retrospectively analyzed, including 57 GGNs in the benign group, 310 in the non-invasive group(AAH+AIS+MIA), and 173 in the invasive group(MIA). The general data(age, gender, smoking history, family history of lung cancer, site), percentage of ground glass component, morphology, boundary, lobulation, burr, vacuolar sign, bronchial abnormality sign, internal vessel sign, pleural traction sign, longest diameter, shortest diameter, unenhanced CT value, CT values on enhancement in arterial phase, CT values on enhancement in venous phase, and degree of enhancement(ΔCT_A-N, ΔCT_V-N)were analyzed with one-way ANOVA, Kruskal-Wallis H test and Pearson χ2 test. Intra-group correlation coefficient(ICC)was used to evaluate the repeatability of measurement. Logistic regression analysis was performed by taking the nature of GGN(tumor or not)or invasiveness as the dependent variables, relative factors as independent variables. ROC curve analysis and the area under the curve(AUC)were calculated and the diagnostic criteria(including critical value, sensitivity and specificity)were evaluated based on the gold standard of pathological diagnosis. Results There were significant differences in the qualitative and quantitative parameters of nodules among the three groups(P<0.001). Logistic regression analysis showed statistically significant differences in the proportion of ground glass components, internal vessel sign, unenhanced CT values and degree of enhancement ΔCT_V-N(P<0.001). ROC curve showed AUC, sensitivity and specificity of ΔCT_V-N were 0.874, 0.747, and 0.877, respectively. There were significant differences in the longest diameter, lobulation, boundary and vascular abnormality of the nodules between the invasive and non-invasive groups. ROC curve showed that the longest diameter was an independent predictor(AUC=0.851, sensitivity=0.746, specificity=0.813). Conclusion Unenhanced CT scan combined with enhanced CT scan is of great significance in predicting benign, malignant and invasive GGNs. The enhancement degree of GGN(ΔCT_V-N)is very effective in predicting benign and malignant GGNs, and the longest diameter is very effective in predicting invasiveness.