The value of preoperative prediction of histological grade of lung adenocarcinoma based on clinical and CT features
-
摘要:
目的 探索CT影像学特征和临床信息与肺浸润性非黏液腺癌病理组织学分级之间的关系，构建预测组织学分级的模型，并进行模型可视化。 方法 收集2021年7—12月浙江省肿瘤医院收治的肺浸润性非黏液腺癌患者313例，按照组织学分级分为中低级别组和高级别组。收集患者临床资料及CT影像学资料，进行统计学分析，分别建立影像学模型和联合模型，绘制列线图和ROC曲线计算AUC，并使用DeLong检验进行比较。 结果 多因素分析显示，CT影像学的长径、实性成分长径、实性比例、空气支气管征均为组织学分级的影响因素(P＜0.05)，建立影像学模型，AUC为0.879；纳入临床资料及CT影像学征象进行多因素分析，结果显示，吸烟、神经元特异性烯醇化酶(NSE)、长径、实性成分长径、实性比例、空气支气管征均为组织学分级的影响因素(P＜0.05)，建立联合模型，AUC为0.899。DeLong检验显示影像模型和联合模型AUC差异无统计学意义(P=0.070)。 结论 CT及血清肿瘤学指标对鉴别肺浸润性非黏液腺癌的中低级别和高级别具有一定的预测价值，有望在术前预测浸润性腺癌的组织学分级。 Abstract:Objective To explore the relationship between both CT imaging features as well as clinical information and the histological grading of lung invasive non-mucinous adenocarcinomas, to construct a model for predicting their histological grading, as well as to visualize the model. Methods Clinical and CT features of 313 patients with invasive non-mucinous lung adenocarcinoma treated at Zhejiang Cancer Hospital from July to December 2021 were collected. Patients were divided into moderate-to-low-grade and high-grade groups based on histological grading. Imaging and combined models were developed through statistical analyses. Nomograms were constructed, and receiver operating characteristic (ROC) curves were generated to calculate the area under the curve (AUC). The DeLong test was used to compare AUCs. Results Multivariate analysis showed that the long diameter of CT imaging, the long diameter of solid components, the proportion of solid, and air bronchogram were influencing factors of histological grade (P < 0.05). The imaging model achieved an AUC of 0.879. Multivariate analysis of clinical data and CT imaging signs showed that there were statistically significant differences in smoking, neuron-specific enolase (NSE), the long diameter of CT imaging, the long diameter of solid component, the proportion of solid, and air bronchogram (P < 0.05). The combined model yielded an AUC of 0.899. The DeLong test showed no statistically significant difference in AUC between the imaging and combined models (P=0.070). Conclusion CT imaging features, together with serum tumor markers, provide valuable predictive information for distinguishing moderate-to-low-grade from high-grade invasive non-mucinous lung adenocarcinoma, supporting their potential role in preoperative histological grading. -
Key words:
- Lung adenocarcinoma /
- Lung cancer /
- Computed tomography /
- Grading system /
- Nomogram
-
图 1 肺浸润性非黏液腺癌病理组织学分级预测的列线图
Figure 1. Nomogram for predicting the histopathological grade of invasive non-mucinous lung adenocarcinoma
图 2 肺浸润性非黏液腺癌的病理组织学分级预测模型的ROC曲线
Figure 2. The ROC curve of the histopathological grading prediction model for invasive non-mucinous lung adenocarcinoma
表 1 2组肺腺癌患者临床资料及CT影像学征象比较
Table 1. Comparison of clinical data and CT imaging data of lung adenocarcinoma between two groups
变量 中低级别组(n=249) 高级别组(n=64) 统计量 P值 性别[例(%)] 5.364a 0.021 男性 100(73.5) 36(26.5) 女性 149(84.2) 28(15.8) 年龄(x±s，岁) 61.5±9.3 61.2±10.9 -0.064b 0.949 吸烟[例(%)] 18.606a ＜0.001 否 201(85.2) 35(14.8) 是 48(62.3) 29(37.7) 细胞角蛋白19[例(%)] 1.911a 0.167 正常 209(81.0) 49(19.0) 超出 40(72.7) 15(27.3) 血清铁蛋白[例(%)] 2.087a 0.149 正常 213(81.0) 50(19.0) 超出 36(72.0) 14(28.0) NSE[例(%)] 5.000a 0.025 正常 232(81.1) 54(18.9) 超出 17(63.0) 10(37.0) CA125[例(%)] 6.082a 0.014 正常 242(80.9) 57(19.1) 超出 7(50.0) 7(50.0) CEA[例(%)] 6.136a 0.013 正常 226(81.6) 51(18.4) 超出 23(63.9) 13(36.1) SCC[例(%)] 0.003a 0.954 正常 229(79.5) 59(20.5) 超出 20(80.0) 5(20.0) 性质[例(%)] 72.822a ＜0.001 磨玻璃 27(100.0) 0 混合磨玻璃 147(95.5) 7(4.5) 实性 75(56.8) 57(43.2) 长径[M(P25, P75)，mm] 20.00(14.00, 27.00) 27.00(17.25, 38.75) -3.931c ＜0.001 短径[M(P25, P75)，mm] 14.00(10.00, 18.00) 18.00(13.00, 29.00) -4.282c ＜0.001 实性成分长径[M(P25, P75)，mm] 12.00(5.00, 20.00) 23.50(16.25, 35.00) -7.374c ＜0.001 实性比例[M(P25, P75)，%] 60.00(30.39, 100.00) 100.00(100.00, 100.00) -8.002c ＜0.001 钙化[例(%)] 2.011a 0.156 否 239(80.5) 58(19.5) 是 10(62.5) 6(37.5) 分叶[例(%)] 5.042a 0.080 无 59(88.1) 8(11.9) 浅分叶 60(73.2) 22(26.8) 深分叶 130(79.3) 34(20.7) 毛刺[例(%)] 11.077a 0.004 无 108(88.5) 14(11.5) 短毛刺 82(76.6) 25(23.4) 长毛刺 59(70.2) 25(29.8) 位置[例(%)] 16.513a ＜0.001 中央 16(51.6) 15(48.4) 周围 233(82.6) 49(17.4) 肺叶[例(%)] 8.300a 0.081 左肺上叶 80(83.3) 16(16.7) 左肺下叶 33(71.7) 13(28.3) 右肺上叶 75(86.2) 12(13.8) 右肺中叶 16(80.0) 4(20.0) 右肺下叶 45(70.3) 19(29.7) 空泡征[例(%)] 4.874a 0.027 否 133(75.1) 44(24.9) 是 116(85.3) 20(14.7) 月牙铲征[例(%)] 14.572a ＜0.001 否 201(75.8) 64(24.2) 是 48(100.0) 0 胸膜牵拉[例(%)] 6.329a 0.012 否 64(90.1) 7(9.9) 是 185(76.4) 57(23.6) 血管集束[例(%)] 0.718a 0.397 否 104(81.9) 23(18.1) 是 145(78.0) 41(22.0) 空气支气管征[例(%)] 23.713a ＜0.001 否 135(70.7) 56(29.3) 是 114(93.4) 8(6.6) 注：a为χ2值, b为t值，c为Z值。 
下载: 导出CSV
表 2 变量赋值情况
Table 2. Variable assignment
变量 赋值方法 组织学分级分组 中低级别组=0，高级别组=1 性别 男性=0，女性=1 吸烟 否=0，是=1 NSE 正常=0，超出=1 CA125 正常=0，超出=1 CEA 正常=0，超出=1 性质 磨玻璃=(0, 0)；混合磨玻璃=(1, 0)；实性=(0, 1) 分叶 无=(0, 0)；浅分叶=(1, 0)；深分叶=(0, 1) 毛刺 无=(0, 0)；短毛刺=(1, 0)；长毛刺=(0, 1) 位置 中央=0，周围=1 肺叶 左肺上叶=(0, 0, 0, 0)；左肺下叶=(1, 0, 0, 0)；右肺上叶=(0, 1, 0, 0)；右肺中叶=(0, 0, 1, 0)；右肺下叶=(0, 0, 0, 1) 空泡征 否=0，是=1 月牙铲征 否=0，是=1 胸膜牵拉 否=0，是=1 空气支气管征 否=0，是=1 注：其余连续性变量均以实际值赋值。
下载: 导出CSV
表 3 肺腺癌患者组织学分级预测因素的多因素logistic回归分析
Table 3. Multivariate logistic regression analysis of predictive factors for histological grade in patients with lung adenocarcinoma
变量 B SE Waldχ2 P值 OR值 95% CI 包含CT影像学资料 长径 0.341 0.125 7.413 0.006 1.407 1.100~1.799 实性成分长径 -0.343 0.126 7.402 0.007 0.710 0.554~0.909 实性比例 0.141 0.053 7.217 0.007 1.152 1.039~1.277 空气支气管征 1.467 0.503 8.500 0.004 4.335 1.617~11.622 包含临床及CT影像学资料 吸烟 1.358 0.601 5.110 0.024 3.889 1.198~12.625 长径 0.370 0.133 7.733 0.005 1.448 1.116~1.881 实性成分长径 -0.398 0.136 8.543 0.003 0.671 0.514~0.877 实性比例 0.157 0.057 7.720 0.005 1.170 1.047~1.308 空气支气管征 -1.248 0.525 5.646 0.017 0.287 0.103~0.804 NSE 1.566 0.743 4.445 0.035 4.786 1.116~20.513 注：本表仅列出差异有统计学意义的变量。注：A为影像学模型；B为联合模型。
下载: 导出CSV
-
[1] MOREIRA A L, OCAMPO P S S, XIA Y, et al. A grading system for invasive pulmonary adenocarcinoma: a proposal from the international association for the study of lung cancer pathology committee[J]. J Thorac Oncol, 2020, 15(10): 1599-1610. [2] HEGEDVS F, ZOMBORI-TÓTH N, KISS S, et al. Prognostic impact of the IASLC grading system of lung adenocarcinoma: a systematic review and meta-analysis[J]. Histopathology, 2024, 85(1): 51-61. [3] 张炳太, 张雷, 刘军, 等. 胸腔镜下肺叶与肺段切除术治疗早期非小细胞肺癌的临床效果观察[J]. 中华全科医学, 2022, 20(3): 399-402. doi: 10.16766/j.cnki.issn.1674-4152.002362ZHANG B T, ZHANG L, LIU J, et al. Clinical effect observation of thoracoscopic lobectomy and segmentectomy in the treatment of early non-small cell lung cancer[J]. Chinese Journal of General Practice, 2022, 20(3): 399-402. doi: 10.16766/j.cnki.issn.1674-4152.002362 [4] TASOUDIS P, LOUFOPOULOS G, MANAKI V, et al. Long term outcomes after lobar versus sublobar resection for patients with non-small cell lung cancer: systematic review and individual patient data meta-analysis[J]. Lung Cancer, 2024, 195: 107929. DOI: 10.1016/j.lungcan.2024.107929. [5] SU H, XIE H, DAI C, et al. Procedure-specific prognostic impact of micropapillary subtype may guide resection strategy in small-sized lung adenocarcinomas: a multicenter study[J]. Ther Adv Med Oncol, 2020, 12: 1758835920937893. DOI: 10.1177/1758835920937893. [6] 陈昶, 朱余明, 姜格宁, 等. 肺段切除术和肺叶切除术治疗直径≤2 cm且术中冰冻证实微乳头和实性亚型阴性肺腺癌的多中心随机对照研究[J]. 中国胸心血管外科临床杂志, 2021, 28(11): 1292-1298.CHEN C H, ZHU Y M, JIANG G N, et al. Comparison of segmentectomy versus lobectomy for ≤2 cm lung adenocarcinoma with micropapillary and solid subtype negative by intraoperative frozen sections: a multi-center randomized controlled trial[J]. Chinese Journal of Clinical Thoracic and Cardiovascular Surgery, 2021, 28(11): 1292-1298. [7] 薛阳, 杜学松, 杨博, 等. 基于最新浸润性肺腺癌分级系统的手术前CT影像特征分析及预后评估[J]. 放射学实践, 2023, 38(9): 1121-1128.XUE Y, DU X S, YANG B, et al. Analysis of preoperative CT image features and prognostic assessment based on the latest classification system for invasive lung adenocarcinoma[J]. Radiol Practice, 2023, 38(9): 1121-1128. [8] PHILLIPS W W, GILL R R, MAZZOLA E, et al. Impact of nodule density in women with sublobar resection for stage ⅠA adenocarcinoma[J]. Ann Thorac Surg, 2021, 112(4): 1067-1075. [9] JING W, LIU M, LI W, et al. Prognostic implication of consolidation-to-tumor ratio in early lung adenocarcinoma: a retrospective cross-sectional study[J]. Quant Imaging Med Surg, 2024, 14(5): 3366-3380. [10] LU L, ZHA Z, ZHANG P, et al. Neuron-specific enolase promotes stem cell-like characteristics of small-cell lung cancer by downregulating NBL1 and activating the BMP2/Smad/ID1 pathway[J]. Oncogenesis, 2022, 11(1): 21. DOI: 10.1038/s41389-022-00396-5. [11] YAN P, HAN Y, TONG A, et al. Prognostic value of neuron-specific enolase in patients with advanced and metastatic non-neuroendocrine non-small cell lung cancer[J]. Biosci Rep, 2021, 41(8): BSR20210866. DOI: 10.1042/BSR20210866. [12] XU C M, LUO Y L, LI S, et al. Multifunctional neuron-specific enolase: its role in lung diseases[J]. Biosci Rep, 2019, 39(11): BSR20192732. DOI: 10.1042/BSR20192732. [13] PARK S, LEE S M, KIM S, et al. Volume doubling times of lung adenocarcinomas: correlation with predominant histologic subtypes and prognosis[J]. Radiology, 2020, 295(3): 703-712. [14] LI Z, WU W, PAN X, et al. Serum tumor markers level and their predictive values for solid and micropapillary components in lung adenocarcinoma[J]. Cancer Med, 2022, 11(14): 2855-2864. [15] 董浩, 邱勇刚, 汪鑫斌, 等. 基于高分辨率CT征象建立logistic回归模型对IA期肺腺癌高级别模式的预测价值[J]. 中国癌症杂志, 2023, 33(8): 768-775.DONG H, QIU Y G, WANG X B, et al. Predictive value of logistic regression model based on high-resolution CT signs for high-grade pattern in stage IA lung adenocarcinoma[J]. China Oncology, 2023, 33(8): 768-775. -
点击查看大图
图(2) / 表(3)
计量
- 文章访问数: 3
- HTML全文浏览量: 0
- PDF下载量: 0
- 被引次数: 0
