通过机器学习和放射组学分析在CT扫描上对肝细胞癌手术后肝切除术后肝衰竭进行术前预测。

Preoperative prediction of post hepatectomy liver failure after surgery for hepatocellular carcinoma on CT-scan by machine learning and radiomics analyses.

作者信息

Famularo Simone, Maino Cesare, Milana Flavio, Ardito Francesco, Rompianesi Gianluca, Ciulli Cristina, Conci Simone, Gallotti Anna, La Barba Giuliano, Romano Maurizio, De Angelis Michela, Patauner Stefan, Penzo Camilla, De Rose Agostino Maria, Marescaux Jacques, Diana Michele, Ippolito Davide, Frena Antonio, Boccia Luigi, Zanus Giacomo, Ercolani Giorgio, Maestri Marcello, Grazi Gian Luca, Ruzzenente Andrea, Romano Fabrizio, Troisi Roberto Ivan, Giuliante Felice, Donadon Matteo, Torzilli Guido

机构信息

Hepatobiliary Surgery Unit, Fondazione Policlinico Universitario A. Gemelli, IRCCS, Catholic University of the Sacred Heart, Rome, Italy; IRCAD, Research Institute Against Cancer of the Digestive System, 1 Place de l'Hôpital, Strasbourg, 67091, France.

Department of Radiology, San Gerardo Hospital, Monza, Italy.

出版信息

Eur J Surg Oncol. 2025 Jul;51(7):109462. doi: 10.1016/j.ejso.2024.109462. Epub 2024 Nov 15.

DOI:10.1016/j.ejso.2024.109462

PMID:39592285

Abstract

INTRODUCTION

No instruments are available to predict preoperatively the risk of posthepatectomy liver failure (PHLF) in HCC patients. The aim was to predict the occurrence of PHLF preoperatively by radiomics and clinical data through machine-learning algorithms.

MATERIALS AND METHODS

Clinical data and 3-phases CT scans were retrospectively collected among 13 Italian centres between 2008 and 2022. Radiomics features were extracted in the non-tumoral liver area. Data were split between training(70 %) and test(30 %) sets. An oversampling was run(ADASYN) in the training set. Random-Forest(RF), extreme gradient boosting (XGB) and support vector machine (SVM) models were fitted to predict PHLF. Final evaluation of the metrics was run in the test set. The best models were included in an averaging ensemble model (AEM).

RESULTS

Five-hundred consecutive preoperative CT scans were collected with the relative clinical data. Of them, 17 (3.4 %) experienced a PHLF. Two-hundred sixteen radiomics features per patient were extracted. PCA selected 19 dimensions explaining >75 % of the variance. Associated clinical variables were: size, macrovascular invasion, cirrhosis, major resection and MELD score. Data were split in training cohort (70 %, n = 351) and a test cohort (30 %, n = 149). The RF model obtained an AUC = 89.1 %(Spec. = 70.1 %, Sens. = 100 %, accuracy = 71.1 %, PPV = 10.4 %, NPV = 100 %). The XGB model showed an AUC = 89.4 %(Spec. = 100 %, Sens. = 20.0 %, Accuracy = 97.3 %, PPV = 20 %, NPV = 97.3 %). The AEM combined the XGB and RF model, obtaining an AUC = 90.1 %(Spec. = 89.5 %, Sens. = 80.0 %, accuracy = 89.2 %, PPV = 21.0 %, NPV = 99.2 %).

CONCLUSIONS

The AEM obtained the best results in terms of discrimination and true positive identification. This could lead to better define patients fit or unfit for liver resection.

摘要

引言

目前尚无仪器可在术前预测肝癌患者肝切除术后肝衰竭（PHLF）的风险。本研究旨在通过放射组学和临床数据，利用机器学习算法在术前预测PHLF的发生。

材料与方法

回顾性收集了2008年至2022年间意大利13个中心的临床数据和三期CT扫描图像。在非肿瘤肝区提取放射组学特征。数据被分为训练集（70%）和测试集（30%）。对训练集进行过采样（ADASYN）。采用随机森林（RF）、极端梯度提升（XGB）和支持向量机（SVM）模型预测PHLF。在测试集中对指标进行最终评估。将最佳模型纳入平均集成模型（AEM）。

结果

收集了500例连续的术前CT扫描图像及相关临床数据。其中，17例（3.4%）发生了PHLF。每位患者提取了216个放射组学特征。主成分分析选择了19个维度，解释了>75%的方差。相关临床变量包括：肿瘤大小、大血管侵犯、肝硬化、大范围切除和终末期肝病模型（MELD）评分。数据被分为训练队列（70%，n = 351）和测试队列（30%，n = 149）。RF模型的曲线下面积（AUC）= 89.1%（特异性= 70.1%，敏感性= 100%，准确率= 71.1%，阳性预测值= 10.4%，阴性预测值= 100%）。XGB模型的AUC = 89.4%（特异性= 100%，敏感性= 20.0%，准确率= 97.3%，阳性预测值= 20%，阴性预测值= 97.3%）。AEM结合了XGB和RF模型，AUC = 90.1%（特异性= 89.5%，敏感性= 80.0%，准确率= 89.2%，阳性预测值= 21.0%，阴性预测值= 99.2%）。