基于影像学和血清学特征构建的列线图模型可较好预测原发性肝细胞癌的微血管浸润

张泳欣; 张水兴; 肖学红; 黄晓星; 杨昂; 唐秉航; 卢扬柏

doi:10.12122/j.issn.1674-4500.2022.04.10

基于影像学和血清学特征构建的列线图模型可较好预测原发性肝细胞癌的微血管浸润

doi: 10.12122/j.issn.1674-4500.2022.04.10

1.
中山市人民医院影像中心，广东中山 528403
2.
暨南大学附属第一医院影像中心，广东广州 510630
3.
中山市人民医院外科，广东中山 528403

基金项目:

中山市人民医院放射影像中心重点专科科研项目 T2020016

详细信息

作者简介:
张泳欣，博士，主治医师，E-mail: xinxin87824@126.com

通讯作者:
卢扬柏，博士，主治医师，Email: luyangbai2020@163.com

计量
- 文章访问数: 285
- HTML全文浏览量: 189
- PDF下载量: 20
- 被引次数: 0
出版历程
- 收稿日期: 2022-05-13
- 网络出版日期: 2022-07-25
- 刊出日期: 2022-07-20

Construction of a prediction model for predicting microvascular invasion in primary hepatocellular carcinoma

1.
Imaging Center, Zhongshan People's Hospital, Zhongshan 528403, China
2.
Imaging Center, the First Affiliated Hospital of Jinan University, Guangzhou 510630, China
3.
Department of Surgery, Zhongshan People's Hospital, Zhongshan 528403, China

摘要

摘要: 目的探讨术前基于影像学和血清学特征构建的列线图模型对肝癌微血管浸润(MVI)的预测价值。方法回顾性分析2015年1月~2020年12月于中山市人民医院接受切除或肝移植的548例肝细胞癌(HCC)患者的临床资料，最终纳入315例肝癌MVI患者，年龄53.2±11.5岁，肿瘤最大直径3.7~7.0 cm。收集患者临床及影像学资料并进行分析，采取单因素与多因素Logistic分析，筛查出能预测MVI的独立风险因素，构建预测HCC中MVI的列线图模型，利用ROC曲线、校准曲线和决策曲线对模型进行评估。结果 MVI (+)患者的中位生存时间为13月(95%CI：8.1~17.9)，1、3、5年无病生存率分别为50.6%、38.5%和30.9%(P < 0.05)；MVI (-)患者的中位生存时间为47月(95%CI：32.7~61.3)，1、3、5年无病生存率分别为77.9%、62.3%和38.8%(P < 0.05)。多因素Logistic回归分析显示，更大的肿瘤体积、突破肝外生长、缺乏或不完整假包膜、存在动脉期瘤周强化以及术前过高的球蛋白值是MVI (+)的独立危险因素(P < 0.05)。最终模型效能曲线下面积为0.895，95%CI为0.859-0.930，准确性为85.1%，敏感度为85.9%，特异性为84.1%。校准曲线显示预测概率与病理结果MVI (+)/MVI (-)概率有良好的一致性。决策曲线显示模型具有良好的临床应用价值。结论构建的列线图及预测模型能较好地术前预测MVI (+)的概率，可以根据MVI发生的风险调整HCC的治疗计划，以优化生存结果。
- 肝细胞癌 /
- 微血管浸润 /
- 危险因素 /
- 列线图
Abstract: Objective To investigate the value of a preoperative nomograph and prediction model based on imaging and serological characteristics for predicting microvascular invasion (MVI) in hepatocellular carcinoma (HCC). Methods Clinical data of 548 patients with HCC who underwent liver resection or liver transplantation from January 2015 to December 2020 in our Hospital were retrospectively included. A total of 315 patients with HCC (MVI+ or MVI-) with an average age of 53.2±11.5 years old and a maximum direct tumour of 3.7-7.0 cm were included. Clinical and imaging data were analyzed. Univariate and multivariate logistic analyses were used to screen out independent risk factors that could predict MVI, and a nomograph model was constructed to predict MVI in HCC, which was evaluated using a subject operating curve, calibration curve and decision curve. Results The median survival time of patients with MVI(+) was 13 months (95% CI: 8.1-17.9) while that of patients with MVI(-)was 47 months (95%CI: 32.7-61.3). The 1-year, 3-year and 5-year disease-free survival rates of patients with MVI(+) were 50.6%、38.5% and 30.9%, while that of patients with MVI(-) were 77.9%、62.3% and 38.8%, respectively. Multivariate logistic regression analysis showed that larger tumour size, extrahepatic growth, absence or incomplete pseudocapsule, presence of arterial peritumoral enhancement and high preoperative globulin value were independent risk factors for MVI(+). The final model efficacy were as follows: AUC=0.895, 95%CI: 0.859-0.930, accuracy: 85.1%, sensitivity: 85.9%, specificity: 84.1%. The calibration curve showed that the predicted probability was in good agreement with the MVI(+)/MVI(-) probability of the pathological results. Thus, the decision curve model displayed good clinical application value. Conclusion The constructed Nomograph and prediction model can better predict the probability of MVI(+) before surgery. It can aid in adjusting the treatment plan of HCC according to the risk of MVI to optimize the survival outcome.
- hepatocellular carcinoma /
- microvascular invasion /
- risk factors /
- nomograph

HTML全文

图 1 MVI(+)和MVI(-)的生存曲线

Figure 1. Survival curves of MVI(+) and MVI(-).

下载: 全尺寸图片幻灯片

图 2 根据多因素回归模型绘制列线图预测患者MVI(+)概率

Figure 2. The probability of MVI(+) in patients was predicted by drawing a line graph based on the multi- factor regression model.

下载: 全尺寸图片幻灯片

图 3 融合模型及各单一变量对肝癌MVI(+)/ MVI(-)的预测效能的ROC曲线

融合模型AUC=0.895(0.859, 0.930)，曲线cut-off值上的敏感度为0.859，特异性为0.841，准确度为0.851.

Figure 3. ROC curves of the fusion model and the predictive efficacy of each single variable for liver cancer MVI(+)/MVI(-).

下载: 全尺寸图片幻灯片

图 4 线图的校准曲线

Figure 4. The calibration curve of the line diagram.

下载: 全尺寸图片幻灯片

图 5 显示列线图的决策曲线分析

Figure 5. The decision curve analysis of the column line graph.

下载: 全尺寸图片幻灯片

表 1 315例患者临床及影像基线特征

Table 1. Baseline clinical and imaging features of 315 patients[n(%)]

因素	总计(n=315)	MVI(-)(n=145)	MVI(+)(n=170)	P
性別				0.579
女=0	42(13.3)	21(14.5)	21(12.4)
男=1	273(86.7)	124(85.5)	149(87.6)
年龄(岁, Mean±SD)	53.2±11.5	54.1±10.4	52.4±12.3	0.040
肿瘤最大直径(cm)*	4.9(3.7, 7.0)	3.7(2.6, 6.1)	5.9(3.6, 9)	< 0.001
肿瘤边缘				0.665
光滑=0	64(20.3)	31(21.4)	33(19.4)
不光滑=1	251(79.7)	114(78.6)	137(80.6)
生长部位				0.001
肝内生长=0	202(64.1)	107(73.8)	95(55.9)
肝外生长=1	113(35.9)	38(26.2)	75(44.1)
累及肝段				0.001
一个肝段=0	127(40.3)	73(50.3)	54(31.8)
多个肝段=1	188(59.7)	72(49.7)	116(68.2)
肿瘤数量				0.968
单个=0	280(88.9)	129(89)	151(88.8)
多个=1	35(11.1)	16(11)	19(11.2)
假包膜				0.093
有=0	162(51.4)	82(56.6)	80(47.1)
无=1	153(48.6)	63(43.4)	90(52.9)
动脉期瘤周强化				< 0.001
无=0	128(40.6)	87(60)	41(24.1)
有=1	187(59.4)	58(40)	129(75.9)
Stage分期				0.660
1=0	282(89.5)	131(90.3)	151(88.8)
≧2=1	33(10.5)	14(9.7)	19(11.2)
肝硬化				0.306
无=0	87(27.6)	36(24.8)	51(30)
有=1	228(72.4)	109(75.2)	119(70)
乙肝				0.091
无=0	59(18.7)	33(22.8)	26(15.3)
有=1	256(81.3)	112(77.2)	144(84.7)
术前谷丙转氨酶(U/L)				0.808
< 40=0	178(56.5)	83(57.2)	95(55.9)
> 40=1	137(43.5)	62(42.8)	75(44.1)
术前谷草转氨酶(U/L)				0.046
< 35=0	146(46.3)	76(52.4)	70(41.2)
> 35=1	169(53.7)	69(47.6)	100(58.8)
AST/ALT	1(0.8, 1.3)	0.9(0.7, 1.2)	1(0.8, 1.4)	0.003
术前碱性磷酸酶(U/L)				0.215
< 150=0	284(90.2)	134(92.4)	150(88.2)
> 150=1	31(9.8)	11(7.6)	20(11.8)
术前白蛋白(g/L)				0.293
< 55=0	95(30.2)	48(33.1)	47(27.6)
> 55=1	220(69.8)	97(66.9)	123(72.4)
术前球蛋白(g/L)				< 0.001
< 35=0	184(58.4)	125(86.2)	59(34.7)
> 35=1	131(41.6)	20(13.8)	111(65.3)
白球比*	1.4(1.2, 1.6)	1.4(1.2, 1.6)	1.5(1.2, 1.7)	0.057
术前AFP(ng/mL)				0.150
< 20=0	121(38.4)	64(44.1)	57(33.5)
20~200=1	65(20.6)	28(19.3)	37(21.8)
> 200=2	129(41.0)	53(36.6)	76(44.7)
*以中位数(四分位数间距)表示.MVI: 微血管浸润.

下载: 导出CSV

表 2 临床影像实验室变量单因素和多因素Logistic回归分析

Table 2. Univariate and multivariate logistic regression analysis of clinical imaging laboratory variables

项目	单因素		多因素
项目	OR (95%CI)	P	OR (95%CI)	P
性別
女=0	Ref
男=1	1.202(0.627~2.302)	0.580
年龄	0.987(0.968~1.006)	0.183
肿瘤最大直径(cm)	1.231(1.138~1.332)	< 0.001	1.237(1.116~1.372)	< 0.001
肿瘤边缘
光滑=0	Ref
不光滑=1	1.129(0.652~1.956)	0.665
生长部位
肝内生长=0	Ref
肝外生长=1	2.223(1.378~3.586)	0.001	2.391(1.213~4.713)	0.012
累及肝段
一个肝段=0	Ref
多个肝段=1	2.178(1.377~3.445)	0.001
肿瘤数量
单个=0	Ref
多个=1	1.014(0.501~2.054)	0.968
假包膜
有=0	Ref
无1	1.464(0.938~2.286)	0.093	0.171(0.064~0.456)	< 0.001
动脉期瘤周强化
无=0	Ref
有=1	4.72(2.91~7.655)	< 0.001	20.618(7.504~56.65)	< 0.001
Stage分期
1=0	Ref
≧2=1	1.177(0.568~2.441)	0.661
肝硬化
无=0	Ref
有=1	0.771(0.468~1.27)	0.307
乙肝
无=0	Ref
有=1	1.632(0.923~2.886)	0.092
术前谷丙转氨酶(U/L)
< 40=0	Ref
> 40=1	1.057(0.676~1.653)	0.808
术前谷草转氨酶(U/L)
< 35=0	Ref
> 35=1	1.573(1.007~2.46)	0.047
AST/ALT	1.621(1.065~2.469)	0.024
术前碱性磷酸酶(U/L)
< 150=0	Ref
> 150=1	1.624(0.751~3.514)	0.218
术前白蛋白(g/L)
< 55=0	Ref
> 55=1	1.295(0.8~2.098)	0.293
术前球蛋白(g/L)
< 35=0	Ref
> 35=1	11.758(6.664~20.749)	< 0.001	17.077(8.531~34.185)	< 0.001
白球比	0.889(0.695~1.136)	0.347
术前AFP(ng/mL)
< 20=0	Ref
20~200=1	1.484(0.809~2.722)	0.203
> 200=2	1.61(0.976~2.656)	0.062

下载: 导出CSV

表 3 融合模型及单一变量对肝癌MVI的预测效能

Table 3. Prediction efficiency of fusion model and single variable for MVI of liver cancer

因素	AUC	敏感度(%)	特异性(%)	准确性(%)
融合模型	0.895 (0.859~0.930)	85.9	84.1	85.1
肿瘤最大直径	0.679 (0.621~0.738)	65.3	62.8	64.1
生长部位	0.590 (0.527~0.652)	44.1	73.8	57.8
假包膜	0.547 (0.484~0.611)	52.9	56.6	54.6
动脉期瘤周强化	0.679 (0.619~0.740)	75.9	60.0	68.6
术前球蛋白	0.758 (0.703~0.812)	65.3	86.2	74.9
注：构建融合模型的变量包括肿瘤最大直径、生长部位、假包膜、动脉期瘤周强化以及术前球蛋白.

下载: 导出CSV

参考文献(39)

[1]	Uemura T, Näppi JJ, Ryu Y, et al. A generative flow-based model for volumetric data augmentation in 3D deep learning for computed tomographic colonography[J]. Int J CARS, 2021, 16(1): 81-9. doi: 10.1007/s11548-020-02275-z
[2]	Torre LA, Bray F, Siegel RL, et al. Global cancer statistics, 2012 [J]. CAA Cancer J Clin, 2015, 65(2): 87-108. doi: 10.3322/caac.21262
[3]	Zeng HM, Chen WQ, Zheng RS, et al. Changing cancer survival in China during 2003-15: a pooled analysis of 17 population-based cancer registries[J]. Lancet Glob Health, 2018, 6(5): e555-67. doi: 10.1016/S2214-109X(18)30127-X
[4]	Vitale A, Peck-Radosavljevic M, Giannini EG, et al. Personalized treatment of patients with very early hepatocellular carcinoma[J]. J Hepatol, 2017, 66(2): 412-23. doi: 10.1016/j.jhep.2016.09.012
[5]	Zhang XP, Chen ZH, Zhou TF, et al. A nomogram to predict early postoperative recurrence of hepatocellular carcinoma with portal vein tumour thrombus after R0 liver resection: a large-scale, multicenter study[J]. Eur J Surg Oncol, 2019, 45(9): 1644-51. doi: 10.1016/j.ejso.2019.03.043
[6]	Pan YX, Chen JC, Fang AP, et al. A nomogram predicting the recurrence of hepatocellular carcinoma in patients after laparoscopic hepatectomy[J]. Cancer Commun (Lond), 2019, 39(1): 55. doi: 10.1186/s40880-019-0404-6
[7]	Lee S, Kang TW, Song KD, et al. Effect of microvascular invasion risk on early recurrence of hepatocellular carcinoma after surgery and radiofrequency ablation[J]. Ann Surg, 2021, 273(3): 564-71. doi: 10.1097/SLA.0000000000003268
[8]	Qi YP, Zhong JH, Liang ZY, et al. Adjuvant transarterial chemoembolization for patients with hepatocellular carcinoma involving microvascular invasion[J]. Am J Surg, 2019, 217(4): 739-44. doi: 10.1016/j.amjsurg.2018.07.054
[9]	Song T, Wang C, Guo C, et al. Pentraxin 3 overexpression accelerated tumor metastasis and indicated poor prognosis in hepatocellular carcinoma via driving epithelial-mesenchymal transition[J]. J Cancer, 2018, 9(15): 2650-8. doi: 10.7150/jca.25188
[10]	Lim KC, Chow PKH, Allen JC, et al. Microvascular invasion is a better predictor of tumor recurrence and overall survival following surgical resection for hepatocellular carcinoma compared to the Milan criteria[J]. Ann Surg, 2011, 254(1): 108-13. doi: 10.1097/SLA.0b013e31821ad884
[11]	Mazzaferro V, Llovet JM, Miceli R, et al. Predicting survival after liver transplantation in patients with hepatocellular carcinoma beyond the Milan criteria: a retrospective, exploratory analysis[J]. Lancet Oncol, 2009, 10(1): 35-43. doi: 10.1016/S1470-2045(08)70284-5
[12]	Shi M, Guo RP, Lin XJ, et al. Partial hepatectomy with wide versus narrow resection margin for solitary hepatocellular carcinoma: a prospective randomized trial[J]. Ann Surg, 2007, 245(1): 36-43. doi: 10.1097/01.sla.0000231758.07868.71
[13]	PhD MRP, Md TVL, Md LA, et al. A systematic review of microvascular invasion in hepatocellular carcinoma: diagnostic and prognostic variability[J]. Ann Surg Oncol, 2013, 20(1): 325-39. doi: 10.1245/s10434-012-2513-1
[14]	Amin MB, Greene FL, Edge SB, et al. The Eighth Edition AJCC Cancer Staging Manual: continuing to build a bridge from a population-based to a more personalized approach to cancer staging [J]. CA Cancer J Clin, 2017, 67(2): 93-9. doi: 10.3322/caac.21388
[15]	Roayaie S, Blume IN, Thung SN, et al. A system of classifying microvascular invasion to predict outcome after resection in patients with hepatocellular carcinoma[J]. Gastroenterology, 2009, 137(3): 850-5. doi: 10.1053/j.gastro.2009.06.003
[16]	Zhao WC, Fan LF, Yang N, et al. Preoperative predictors of microvascular invasion in multinodular hepatocellular carcinoma [J]. Eur J Surg Oncol EJSO, 2013, 39(8): 858-64. doi: 10.1016/j.ejso.2013.04.003
[17]	Poté N, Cauchy F, Albuquerque M, et al. Performance of PIVKA-Ⅱ for early hepatocellular carcinoma diagnosis and prediction of microvascular invasion[J]. J Hepatol, 2015, 62(4): 848-54. doi: 10.1016/j.jhep.2014.11.005
[18]	Lambin P, Leijenaar RTH, Deist TM, et al. Radiomics: the bridge between medical imaging and personalized medicine[J]. Nat Rev Clin Oncol, 2017, 14(12): 749-62. doi: 10.1038/nrclinonc.2017.141
[19]	Xu X, Zhang HL, Liu QP, et al. Radiomic analysis of contrastenhanced CT predicts microvascular invasion and outcome in hepatocellular carcinoma[J]. J Hepatol, 2019, 70(6): 1133-44. doi: 10.1016/j.jhep.2019.02.023
[20]	Yang L, Gu DS, Wei JW, et al. A radiomics nomogram for preoperative prediction of microvascular invasion in hepatocellular carcinoma [J]. Liver Cancer, 2019, 8(5): 373-86. doi: 10.1159/000494099
[21]	Lee S, Kim SH, Lee JE, et al. Preoperative gadoxetic acid-enhanced MRI for predicting microvascular invasion in patients with single hepatocellular carcinoma[J]. J Hepatol, 2017, 67(3): 526-34. doi: 10.1016/j.jhep.2017.04.024
[22]	Zhang YF, Lei XZ, Xu LL, et al. Preoperative and postoperative nomograms for predicting early recurrence of hepatocellular carcinoma without macrovascular invasion after curative resection [J]. BMC Surg, 2022, 22(1): 233. doi: 10.1186/s12893-022-01682-0
[23]	Njei B, Rotman Y, Ditah I, et al. Emerging trends in hepatocellular carcinoma incidence and mortality[J]. Hepatology, 2015, 61(1): 191-9. doi: 10.1002/hep.27388
[24]	Llovet JM, Schwartz M, Mazzaferro V. Resection and liver transplantation for hepatocellular carcinoma[J]. Semin Liver Dis, 2005, 25(2): 181-200. doi: 10.1055/s-2005-871198
[25]	Tabrizian P, Jibara G, Shrager B, et al. Recurrence of hepatocellular cancer after resection: patterns, treatments, and prognosis[J]. Ann Surg, 2015, 261(5): 947-55. doi: 10.1097/SLA.0000000000000710
[26]	Lee KW, Park JW, Park JB, et al. Liver transplantation for hepatocellular carcinoma with bile duct thrombi[J]. Transplant Proc, 2006, 38(7): 2093-4. doi: 10.1016/j.transproceed.2006.06.034
[27]	Han J, Li ZL, Xing H, et al. The impact of resection margin and microvascular invasion on long-term prognosis after curative resection of hepatocellular carcinoma: a multi-institutional study[J]. HPB, 2019, 21(8): 962-71. doi: 10.1016/j.hpb.2018.11.005
[28]	Roayaie S, Obeidat K, Sposito C, et al. Resection of hepatocellular cancer ≤2 cm: results from two Western centers[J]. Hepatology, 2013, 57(4): 1426-35. doi: 10.1002/hep.25832
[29]	Yo-ichi Yamashita MD P, Eiji Tsuijita MD P, Md KT, et al. Predictors for microinvasion of small hepatocellular carcinoma ≤2 cm[J]. Ann Surg Oncol, 2012, 19(6): 2027-34. doi: 10.1245/s10434-011-2195-0
[30]	Cuccurullo V, di Stasio GD, Mazzarella G, et al. Microvascular invasion in HCC: the molecular imaging perspective[J]. Contrast Media Mol Imaging, 2018, 2018: 9487938.
[31]	Sun Z, Li Z, Shi XL, et al. Anatomic versus non-anatomic resection of hepatocellular carcinoma with microvascular invasion: a systematic review and meta-analysis[J]. Asian J Surg, 2021, 44(9): 1143-50. doi: 10.1016/j.asjsur.2021.02.023
[32]	Pawlik TM, Delman KA, Vauthey JN, et al. Tumor size predicts vascular invasion and histologic grade: implications for selection of surgical treatment for hepatocellular carcinoma[J]. Liver Transplant, 2005, 11(9): 1086-92. doi: 10.1002/lt.20472
[33]	Md SS, Md ON, Okuda K, et al. The significance of classifying microvascular invasion in patients with hepatocellular carcinoma [J]. Ann Surg Oncol, 2014, 21(3): 1002-9. doi: 10.1245/s10434-013-3376-9
[34]	Iguchi T, Shirabe K, Aishima S, et al. New pathologic stratification of microvascular invasion in hepatocellular carcinoma[J]. Transplantation, 2015, 99(6): 1236-42. doi: 10.1097/TP.0000000000000489
[35]	Jeong SH, Kim RB, Park SY, et al. Nomogram for predicting gastric cancer recurrence using biomarker gene expression[J]. Eur J Surg Oncol, 2020, 46(1): 195-201. doi: 10.1016/j.ejso.2019.09.143
[36]	Yin XY, Pang T, Liu Y, et al. Development and validation of a nomogram for preoperative prediction of lymph node metastasis in early gastric cancer[J]. World J Surg Oncol, 2020, 18(1): 2. doi: 10.1186/s12957-019-1778-2
[37]	Wang HH, Lu Y, Liu RK, et al. A non-invasive nomogram for preoperative prediction of microvascular invasion risk in hepatocellular carcinoma[J]. Front Oncol, 2021, 11((5)): 745085.
[38]	Li HG, Li T, Hu JH, et al. A nomogram to predict microvascular invasion in early hepatocellular carcinoma[J]. J Cancer Res Ther, 2021, 17(3): 652-7. doi: 10.4103/jcrt.JCRT_1714_20
[39]	Gu YF, Zheng FY, Zhang YX, et al. Novel nomogram based on inflammatory markers for the preoperative prediction of microvascular invasion in solitary primary hepatocellular carcinoma [J]. Cancer Manag Res, 2022, 14: 895-907. doi: 10.2147/CMAR.S346976