基于机器学习构建肝切除术后肝衰竭预测列线图模型及其预测价值

doi:10.3877/cma.j.issn.2095-3232.2026.02.008

目的

基于机器学习（ML）构建肝切除术后肝衰竭（PHLF）预测列线图模型，探讨其在PHLF预测中的应用价值。

方法

回顾性分析2021年6月至2023年6月于首都医科大学附属北京天坛医院行肝切除术235例患者的临床资料。患者均签署患者知情同意书，符合医学伦理学规定。其中男118例，女117例；年龄27~83岁，中位年龄57岁。按8∶2的比例将患者随机分为训练集和测试集。观察指标包括ALT、TB、ALB、Plt、PT、INR、AST和Plt比率指数（APRI）及全肝体积（TLV）、残肝体积（RLV）、脾体积（SV）、标准残肝体积（SRLV）、残肝体积/体重比（RLV/Weight，RLV/W）、脾-肝脏体积比（SV/LV）等影像学相关指标。在训练集中，采用LASSO回归和Logistic回归分析进行特征选择，以筛选与PHLF相关的关键风险因素，分别构建ML-nomogram和Log-nomogram模型。两种模型预测价值采用ROC曲线分析，AUC比较采用Delong检验。

结果

本研究中PHLF发生率16.2%（38/235），其中A级30例，B级8例。在训练集中，使用LASSO回归筛选肝硬化、Plt、APRI、SV、RLV/TLV、SV/LV等6个变量作为PHLF预测的关键特征，构建的ML-nomogram列线图在训练集和测试集的AUC分别为0.937和0.839。logistic回归则共筛选出Plt、SV和RLV/TLV等3个危险因素，构建Log-nomogram列线图在训练集和测试集的AUC为0.934和0.813。ML-nomogram和Log-nomogram均对PHLF发生具有较高的预测能力（Z=1.21，1.19；P>0.05）。

结论

本研究基于ML成功构建ML-nomogram列线图PHLF预测模型，该模型和Log-nomogram均对PHLF发生具有较高的预测效能，预测准确率高，应用简便。

关键词: 肝切除术, 肝衰竭, 机器学习, 列线图, 风险因素

Objective

To construct a nomogram model for predicting post-hepatectomy liver failure (PHLF) based on machine learning (ML), and evaluate its application value in PHLF prediction.

Methods

Clinical data of 235 patients who underwent hepatectomy in Beijing Tiantan Hospital affiliated to Capital Medical University from June 2021 to June 2023 were retrospectively analyzed. The informed consents of all patients were obtained and the local ethical committee approval was received. Among them, 118 patients were male and 117 female, aged from 27 to 83 years, with a median age of 57 years. All patients were randomly divided into the training and test sets at a ratio of 8∶2. ALT, TB, ALB, Plt, PT, INR, AST to Plt ratio index (APRI), total liver volume (TLV), remnant liver volume (RLV), spleen volume (SV), standard remnant liver volume (SRLV), remnant liver volume/weight ratio (RLV/Weight, RLV/W) and spleen volume/liver volume ratio (SV/LV) and other imaging indexes were observed. In the training set, LASSO regression and Logistic regression analyses were used to select features and screen key risk factors related to PHLF. ML-nomogram and Log-nomogram models were constructed, respectively. The predictive value of these two models was analyzed by the ROC curve. AUC between two models was compared by Delong test.

Results

The incidence of PHLF was 16.2%(38/235), including 30 cases of grade A and 8 cases of grade B. In the training set, 6 variables including liver cirrhosis, Plt, APRI, SV, RLV/TLV and SV/LV were screened by LASSO regression as the key features of PHLF prediction. The AUC of the constructed ML-nomogram in the training and test sets was 0.937 and 0.839, respectively. Logistic regression analysis screened 3 risk factors including Plt, SV and RLV/TLV. The AUC of the constructed Log-nomogram in the training and test sets was 0.934 and 0.813, respectively. Both ML-nomogram and Log-nomogram had high predictive ability for the incidence of PHLF (Z=1.21, 1.19; both P>0.05).

Conclusions

In this study, ML-nomogram is successfully constructed for predicting PHLF. Both ML-nomogram and Log-nomogram are convenient and have high prediction efficiency and accuracy.

Key words: Hepatectomy, Liver failure, Machine learning, Nomogram, Risk factor

表1 训练集与测试集肝切除患者基线数据比较

参数	训练集	测试集	统计值	P值
PHLF[例(%)]	30(16)	7(15)	χ²=0.038	0.845
男性[例(%)]	102(54)	18(39)	χ²=3.259	0.071
年龄(岁, $x ¯$ ±s)	57±28	56±9	t=0.171	0.865
肝炎[例(%)]	53(28)	7(16)	χ²=2.184	0.139
肝硬化[例(%)]	26(14)	4(9)	χ²=1.010	0.315
手术时间[h, M(Q₁, Q₃)]	5(4, 6)	5(4, 6)	Z=-0.647	0.517
出血量[ml, M(Q₁, Q₃)]	250(200, 400)	275(150, 350)	Z=-0.971	0.331
肝门阻断[例(%)]	36(19)	8(17)	χ²=0.031	0.859
ALT[U/L, M(Q₁, Q₃)]	21(15, 29)	24(15, 31)	Z=-0.646	0.518
TB[µmol/L, M(Q₁, Q₃)]	14(10, 17)	14(10, 18)	Z=-0.693	0.488
ALB(g/L, $x ¯$ ±s)	40±4	40±4	t=-0.830	0.407
Plt(×10⁹/L, $x ¯$ ±s)	199±65	197±51	t=0.156	0.876
PT[s, M(Q₁, Q₃)]	11(10, 11)	11(10, 11)	Z=-0.063	0.950
INR( $x ¯$ ±s)	1.13±0.92	1.02±0.08	t=0.823	0.411
ALBI( $x ¯$ ±s)	-2.63±0.03	-2.64±0.06	t=0.282	0.778
APRI[M(Q₁, Q₃)]	0.28(0.21, 0.43)	0.31(0.21, 0.44)	Z=-0.300	0.764
SV[ml, M(Q₁, Q₃)]	176(164, 199)	176(159, 198)	Z=-0.204	0.838
RLV/TLV( $x ¯$ ±s)	0.74±0.08	0.74±0.08	t=-0.153	0.878
SRLV(ml, $x ¯$ ±s)	661±148	648±135	t=0.529	0.597
RLV/W( $x ¯$ ±s)	18±14	17±4	t=0.595	0.552
SV/LV[M(Q₁, Q₃)]	0.12(0.11, 0.14)	0.13(0.11, 0.14)	Z=-0.943	0.346

表1 训练集与测试集肝切除患者基线数据比较

参数	训练集	测试集	统计值	P值
PHLF[例(%)]	30(16)	7(15)	χ²=0.038	0.845
男性[例(%)]	102(54)	18(39)	χ²=3.259	0.071
年龄(岁, $x ¯$ ±s)	57±28	56±9	t=0.171	0.865
肝炎[例(%)]	53(28)	7(16)	χ²=2.184	0.139
肝硬化[例(%)]	26(14)	4(9)	χ²=1.010	0.315
手术时间[h, M(Q₁, Q₃)]	5(4, 6)	5(4, 6)	Z=-0.647	0.517
出血量[ml, M(Q₁, Q₃)]	250(200, 400)	275(150, 350)	Z=-0.971	0.331
肝门阻断[例(%)]	36(19)	8(17)	χ²=0.031	0.859
ALT[U/L, M(Q₁, Q₃)]	21(15, 29)	24(15, 31)	Z=-0.646	0.518
TB[µmol/L, M(Q₁, Q₃)]	14(10, 17)	14(10, 18)	Z=-0.693	0.488
ALB(g/L, $x ¯$ ±s)	40±4	40±4	t=-0.830	0.407
Plt(×10⁹/L, $x ¯$ ±s)	199±65	197±51	t=0.156	0.876
PT[s, M(Q₁, Q₃)]	11(10, 11)	11(10, 11)	Z=-0.063	0.950
INR( $x ¯$ ±s)	1.13±0.92	1.02±0.08	t=0.823	0.411
ALBI( $x ¯$ ±s)	-2.63±0.03	-2.64±0.06	t=0.282	0.778
APRI[M(Q₁, Q₃)]	0.28(0.21, 0.43)	0.31(0.21, 0.44)	Z=-0.300	0.764
SV[ml, M(Q₁, Q₃)]	176(164, 199)	176(159, 198)	Z=-0.204	0.838
RLV/TLV( $x ¯$ ±s)	0.74±0.08	0.74±0.08	t=-0.153	0.878
SRLV(ml, $x ¯$ ±s)	661±148	648±135	t=0.529	0.597
RLV/W( $x ¯$ ±s)	18±14	17±4	t=0.595	0.552
SV/LV[M(Q₁, Q₃)]	0.12(0.11, 0.14)	0.13(0.11, 0.14)	Z=-0.943	0.346

表2 基于ML筛选预测PHLF的关键特征变量的相关系数

变量	LASSO回归	Logistic回归
变量	回归系数	回归系数	标准误
肝硬化	0.028 659 5	0.273 5	0.804 9
Plt	-0.003 362 9	-0.018 2	0.007 2
APRI	0.561 609 0	0.565 4	1.119 6
SV	0.005 978 8	0.014 4	0.010 6
RLV/TLV	-4.603 047 0	-16.938 4	3.983 0
SV/LV	1.102 655 4	2.073 6	7.975 5

表2 基于ML筛选预测PHLF的关键特征变量的相关系数

变量	LASSO回归	Logistic回归
变量	回归系数	回归系数	标准误
肝硬化	0.028 659 5	0.273 5	0.804 9
Plt	-0.003 362 9	-0.018 2	0.007 2
APRI	0.561 609 0	0.565 4	1.119 6
SV	0.005 978 8	0.014 4	0.010 6
RLV/TLV	-4.603 047 0	-16.938 4	3.983 0
SV/LV	1.102 655 4	2.073 6	7.975 5

图1 PHLF预测ML-nomogram列线图注：ML-nomogram为机器学习列线图，PHLF为肝切除术后肝衰竭，SV为脾体积，APRI为AST和Plt比率指数，TLV为全肝体积，RLV为残肝体积，SV/LV为脾-肝脏体积比

表3 PHLF预测Logistic多因素回归分析

变量	b	SE(b)	P值	OR值	95% CI
Plt(×10⁹/L)	-0.022	0.008	0.006	0.979	0.979~0.988
SV(ml)	0.028	0.013	0.038	1.028	1.007~1.032
RLV/TLV	-16.815	4.776	<0.001	4.475×10^-8	8.889×10^-12~5.008×10^-5

表3 PHLF预测Logistic多因素回归分析

变量	b	SE(b)	P值	OR值	95% CI
Plt(×10⁹/L)	-0.022	0.008	0.006	0.979	0.979~0.988
SV(ml)	0.028	0.013	0.038	1.028	1.007~1.032
RLV/TLV	-16.815	4.776	<0.001	4.475×10^-8	8.889×10^-12~5.008×10^-5

图2 PHLF预测Log-nomogram列线图注：Log-nomogram 为Logistic回归列线图，PHLF为肝切除术后肝衰竭，APRI为AST和Plt比率指数，TLV为全肝体积，RLV为残肝体积，SV/LV为脾-肝脏体积比

表4 两种PHLF预测列线图ROC参数

列线图	AUC	95%CI	截断值	灵敏度	特异度
ML-nomogram
训练集	0.937	0.902~0.971	0.071	1.000	0.747
测试集	0.839	0.635~1.000	0.129	0.857	0.821
Log-nomogram
训练集	0.934	0.898~0.970	0.146	0.968	0.797
测试集	0.813	0.590~1.000	0.148	0.857	0.821

表4 两种PHLF预测列线图ROC参数

列线图	AUC	95%CI	截断值	灵敏度	特异度
ML-nomogram
训练集	0.937	0.902~0.971	0.071	1.000	0.747
测试集	0.839	0.635~1.000	0.129	0.857	0.821
Log-nomogram
训练集	0.934	0.898~0.970	0.146	0.968	0.797
测试集	0.813	0.590~1.000	0.148	0.857	0.821

[1]	Michalopoulos GK, Bhushan B. Liver regeneration: biological and pathological mechanisms and implications[J]. Nat Rev Gastroenterol Hepatol, 2021, 18(1): 40-55.DOI: 10.1038/s41575-020-0342-4.
[2]	Baumgartner R, Engstrand J, Rajala P, et al. Comparing the accuracy of prediction models to detect clinically relevant post-hepatectomy liver failure early after major hepatectomy[J]. Br J Surg, 2024, 111(1): znad433.DOI: 10.1093/bjs/znad433.
[3]	Nishio T, Taura K, Koyama Y, et al. Current status of preoperative risk assessment for posthepatectomy liver failure in patients with hepatocellular carcinoma[J]. Ann Gastroenterol Surg, 2023, 7(6): 871-886.DOI: 10.1002/ags3.12692.
[4]	Wang J, Zheng T, Liao Y, et al. Machine learning prediction model for post-hepatectomy liver failure in hepatocellular carcinoma: a multicenter study[J]. Front Oncol, 2022, 12: 986867.DOI: 10.3389/fonc.2022.986867.
[5]	Rahbari NN, Garden OJ, Padbury R, et al. Posthepatectomy liver failure: a definition and grading by the International Study Group of Liver Surgery (ISGLS)[J]. Surgery, 2011, 149(5): 713-724.DOI: 10.1016/j.surg.2010.10.001.
[6]	Sultana A, Brooke-Smith M, Ullah S, et al. Prospective evaluation of the International Study Group for Liver Surgery definition of post hepatectomy liver failure after liver resection: an international multicentre study[J]. HPB, 2018, 20(5): 462-469.DOI: 10.1016/j.hpb.2017.11.007.
[7]	Sparrelid E, Olthof PB, Dasari BVM, et al. Current evidence on posthepatectomy liver failure: comprehensive review[J]. BJS Open, 2022, 6(6): zrac142.DOI: 10.1093/bjsopen/zrac142.
[8]	Gilg S, Sandström P, Rizell M, et al. The impact of post-hepatectomy liver failure on mortality: a population-based study[J]. Scand J Gastroenterol, 2018, 53(10-11): 1335-1339.DOI: 10.1080/00365521.2018.1501604.
[9]	Niederwieser T, Braunwarth E, Dasari BVM, et al. Early postoperative arterial lactate concentrations to stratify risk of post-hepatectomy liver failure[J]. Br J Surg, 2021, 108(11): 1360-1370.DOI: 10.1093/bjs/znab338.
[10]	Merath K, Tiwari A, Court C, et al. Postoperative liver failure: definitions, risk factors, prediction models and prevention strategies[J]. J Gastrointest Surg, 2023, 27(11): 2640-2649.DOI: 10.1007/s11605-023-05834-2.
[11]	Balzan S, Belghiti J, Farges O, et al. The “50-50 criteria” on postoperative day 5: an accurate predictor of liver failure and death after hepatectomy[J]. Ann Surg, 2005, 242(6): 824-828, discussion828-9.DOI: 10.1097/01.sla.0000189131.90876.9e.
[12]	Mullen JT, Ribero D, Reddy SK, et al. Hepatic insufficiency and mortality in 1, 059 noncirrhotic patients undergoing major hepatectomy[J]. J Am Coll Surg, 2007, 204(5): 854-862; discussion862-4.DOI: 10.1016/j.jamcollsurg.2006.12.032.
[13]	Hobeika C, Fuks D, Cauchy F, et al. Impact of cirrhosis in patients undergoing laparoscopic liver resection in a nationwide multicentre survey[J]. Br J Surg, 2020, 107(3): 268-277.DOI: 10.1002/bjs.11406.
[14]	Timchenko NA. Aging and liver regeneration[J]. Trends Endocrinol Metab, 2009, 20(4): 171-176.DOI: 10.1016/j.tem.2009.01.005.
[15]	Zou H, Yang X, Li QL, et al. A comparative study of albumin-bilirubin score with child-pugh score, model for end-stage liver disease score and indocyanine green R15 in predicting posthepatectomy liver failure for hepatocellular carcinoma patients[J]. Dig Dis, 2018, 36(3): 236-243.DOI: 10.1159/000486590.
[16]	Ross SW, Seshadri R, Walters AL, et al. Mortality in hepatectomy: Model for End-Stage Liver Disease as a predictor of death using the National Surgical Quality Improvement Program database[J]. Surgery, 2016, 159(3): 777-792.DOI: 10.1016/j.surg.2015.08.021.
[17]	Wang YY, Zhong JH, Su ZY, et al. Albumin-bilirubin versus Child-Pugh score as a predictor of outcome after liver resection for hepatocellular carcinoma[J]. Br J Surg, 2016, 103(6): 725-734.DOI: 10.1002/bjs.10095.
[18]	Fagenson AM, Gleeson EM, Pitt HA, et al. Albumin-bilirubin score vs model for end-stage liver disease in predicting post-hepatectomy outcomes[J]. J Am Coll Surg, 2020, 230(4): 637-645.DOI: 10.1016/j.jamcollsurg.2019.12.007.
[19]	Santol J, Kim S, Gregory LA, et al. An APRI+ALBI-based multivariable model as a preoperative predictor for posthepatectomy liver failure[J]. Ann Surg, 2025, 281(5): 861-871.DOI: 10.1097/sla.0000000000006127.
[20]	Gao XX, Li JF. Current strategies for predicting post-hepatectomy liver failure and a new ultrasound-based nomogram[J]. World J Gastroenterol, 2024, 30(39): 4254-4259.DOI: 10.3748/wjg.v30.i39.4254.
[21]	Zhong X, Salahuddin Z, Chen Y, et al. An interpretable radiomics model based on two-dimensional shear wave elastography for predicting symptomatic post-hepatectomy liver failure in patients with hepatocellular carcinoma[J]. Cancers, 2023, 15(21): 5303.DOI: 10.3390/cancers15215303.
[22]	Kawaguchi T, Tokushige K, Hyogo H, et al. A data mining-based prognostic algorithm for NAFLD-related hepatoma patients: a nationwide study by the Japan study group of NAFLD[J]. Sci Rep, 2018, 8(1): 10434.DOI: 10.1038/s41598-018-28650-0.
[23]	Kawaguchi T, Yoshio S, Sakamoto Y, et al. Impact of decorin on the physical function and prognosis of patients with hepatocellular carcinoma[J]. J Clin Med, 2020, 9(4): 936.DOI: 10.3390/jcm9040936.
[24]	Xu X, Xing Z, Xu Z, et al. A deep learning model for prediction of post hepatectomy liver failure after hemihepatectomy using preoperative contrast-enhanced computed tomography: a retrospective study[J]. Front Med, 2023, 10: 1154314.DOI: 10.3389/fmed.2023.1154314.
[25]	Kang CM, Ku HJ, Moon HH, et al. Predicting safe liver resection volume for major hepatectomy using artificial intelligence[J]. J Clin Med, 2024, 13(2): 381.DOI: 10.3390/jcm13020381.
[26]	Yamaguchi Y, Yoshida S, Misumi T, et al. Multiple imputation for longitudinal data using Bayesian lasso imputation model[J]. Stat Med, 2022, 41(6): 1042-1058.DOI: 10.1002/sim.9315.
[27]	Li E, Ai F, Liang C. A machine learning model to predict the risk of depression in US adults with obstructive sleep apnea hypopnea syndrome: a cross-sectional study[J]. Front Public Health, 2024, 11: 1348803.DOI: 10.3389/fpubh.2023.1348803.
[28]	Bose G, Healy BC, Lokhande HA, et al. Early predictors of clinical and MRI outcomes using least absolute shrinkage and selection operator (LASSO) in multiple sclerosis[J]. Ann Neurol, 2022, 92(1): 87-96.DOI: 10.1002/ana.26370.
[29]	Ocak İ, Topaloğlu S, Acarli K. Posthepatectomy liver failure[J]. Turk J Med Sci, 2020, 50(6): 1491-1503.DOI: 10.3906/sag-2006-31.
[30]	Søreide JA, Deshpande R. Post hepatectomy liver failure (PHLF)–Recent advances in prevention and clinical management[J]. Eur J Surg Oncol, 2021, 47(2): 216-224.DOI: 10.1016/j.ejso.2020.09.001.
[31]	Santol J, Ammann M, Reese T, et al. Comparison of the LiMAx test vs. the APRI+ALBI score for clinical utility in preoperative risk assessment in patients undergoing liver surgery: a European multicenter study[J]. Eur J Surg Oncol, 2024, 50(4): 108048.DOI: 10.1016/j.ejso.2024.108048.
[32]	Yugawa K, Maeda T, Nagata S, et al. Impact of aspartate aminotransferase-to-platelet ratio index based score to assess posthepatectomy liver failure in patients with hepatocellular carcninoma[J]. World J Surg Oncol, 2022, 20(1): 248.DOI: 10.1186/s12957-022-02714-y.

[1]	吕思怡, 王琰琪, 仇珺, 陈宇江, 高洁. 人工智能图像处理技术在口腔医学中的应用[J/OL]. 中华口腔医学研究杂志(电子版), 2026, 20(01): 40-46.
[2]	李瑞, 陈文浩, 徐镇, 蒋维荣, 李志文, 侯以森, 孟勇. 3D打印解剖模型联合腹腔镜手术视频在腹腔镜肝切除术临床教学的应用研究[J/OL]. 中华普外科手术学杂志(电子版), 2026, 20(01): 63-66.
[3]	严一杰, 张军, 孟繁杰, 关志宇. 肺结节-胸膜关系预测CT引导下肺穿刺活检后气胸风险的作用研究[J/OL]. 中华肺部疾病杂志(电子版), 2026, 19(01): 36-41.
[4]	杜东莲, 史志伟, 姚洁, 张树敏, 代卫斌. 基于CT影像组学和临床特征预测单发肺结节生长的临床意义[J/OL]. 中华肺部疾病杂志(电子版), 2026, 19(01): 56-61.
[5]	杨宇尘, 陈拥军. 三维可视化技术在肝癌转化治疗辅助决策中的应用[J/OL]. 中华肝脏外科手术学电子杂志, 2026, 15(01): 10-15.
[6]	戴宗伯, 张城硕, 郭庭维, 何知远, 赵昊宇, 张宇慈, 张佳林. 基于MRI影像组学机器学习构建肝细胞癌微血管侵犯预测模型[J/OL]. 中华肝脏外科手术学电子杂志, 2026, 15(01): 36-44.
[7]	钟文卿, 韩冰. 基于术前CT影像数据构建肝癌患者生存期Nomogram预测模型[J/OL]. 中华肝脏外科手术学电子杂志, 2026, 15(01): 53-58.
[8]	阙园, 赵增江, 郑潇, 黄从云, 欧阳再兴, 刘文瑛. 头戴式裸眼3D腔镜显示系统在腹腔镜肝切除术中的应用[J/OL]. 中华肝脏外科手术学电子杂志, 2026, 15(01): 59-65.
[9]	孙娟华, 白引苗, 孔胜男, 王梦雪, 王文慧, 张红梅. 胰腺癌患者化疗相关性恶心呕吐风险列线图构建及验证[J/OL]. 中华消化病与影像杂志(电子版), 2026, 16(02): 114-119.
[10]	张维娜, 潘亚娟, 徐敏. 晚期消化系统癌症手术患者器官/腔隙感染的风险预测模型的建立[J/OL]. 中华消化病与影像杂志(电子版), 2026, 16(02): 120-124.
[11]	慈娟娟, 吴俊成, 朱琴琴. 肝硬化合并食管胃静脉曲张破裂出血患者内镜治疗后再出血风险列线图的构建与验证[J/OL]. 中华消化病与影像杂志(电子版), 2026, 16(01): 47-52.
[12]	赵莉, 张敏伟, 郭浩东, 于海侠, 王光明, 李德春. 乳腺X线模型构建列线图预测乳腺癌HR及HER2的表达[J/OL]. 中华临床医师杂志(电子版), 2025, 19(10): 747-757.
[13]	王超, 张晓会, 李晓帆, 赵海丹. 维持性血液透析患者感染与心脑血管死亡风险的比较及联合预测模型构建[J/OL]. 中华临床医师杂志(电子版), 2025, 19(09): 675-681.
[14]	于少华, 苏飞, 芦永斌, 袁芳芸, 阚晓燕, 张涛, 侯小明. 基于血清电解质水平构建列线图模型对广泛期小细胞肺癌患者的预测价值[J/OL]. 中华临床医师杂志(电子版), 2025, 19(08): 574-581.
[15]	皇立媛, 浦洁, 王苏贵, 陈婷婷, 朱德慧, 胡雪. 中青年脑卒中患者应激障碍风险预测模型的构建与验证[J/OL]. 中华临床医师杂志(电子版), 2025, 19(07): 504-512.

阅读次数

全文

摘要

选择文件类型/文献管理软件名称

选择包含的内容

Construction of nomogram model for predicting post-hepatectomy liver failure based on machine learning and its predictive value

模态框（Modal）标题

选择文件类型/文献管理软件名称

选择包含的内容

Construction of nomogram model for predicting post-hepatectomy liver failure based on machine learning and its predictive value