Mahmoudi S, Bernatz S, Ackermann J, Koch V, Dos Santos D P, Grünewald L D, Yel I, Martin S S, Scholtz J-E, Stehle A, Walter D, Zeuzem S, Wild P J, Vogl T J, Kinzler M N
Department of Diagnostic and Interventional Radiology, University Hospital Frankfurt, Frankfurt am Main, Germany.
Department of Diagnostic and Interventional Radiology, University Hospital Frankfurt, Frankfurt am Main, Germany; Dr. Senckenberg Institute for Pathology, University Hospital Frankfurt, Goethe University, Frankfurt am Main, Germany; Frankfurt Cancer Institute (FCI), Goethe University, Frankfurt am Main, Germany; University Cancer Center Frankfurt (UCT), University Hospital, Goethe University, Frankfurt am Main, Germany.
Clin Oncol (R Coll Radiol). 2023 May;35(5):e312-e318. doi: 10.1016/j.clon.2023.01.018. Epub 2023 Feb 1.
Intrahepatic cholangiocarcinoma (iCCA) and hepatocellular carcinoma (HCC) differ in prognosis and treatment. We aimed to non-invasively differentiate iCCA and HCC by means of radiomics extracted from contrast-enhanced standard-of-care computed tomography (CT).
In total, 94 patients (male, n = 68, mean age 63.3 ± 12.4 years) with histologically confirmed iCCA (n = 47) or HCC (n = 47) who underwent contrast-enhanced abdominal CT between August 2014 and November 2021 were retrospectively included. The enhancing tumour border was manually segmented in a clinically feasible way by defining three three-dimensional volumes of interest per tumour. Radiomics features were extracted. Intraclass correlation analysis and Pearson metrics were used to stratify robust and non-redundant features with further feature reduction by LASSO (least absolute shrinkage and selection operator). Independent training and testing datasets were used to build four different machine learning models. Performance metrics and feature importance values were computed to increase the models' interpretability.
The patient population was split into 65 patients for training (iCCA, n = 32) and 29 patients for testing (iCCA, n = 15). A final combined feature set of three radiomics features and the clinical features age and sex revealed a top test model performance of receiver operating characteristic (ROC) area under the curve (AUC) = 0.82 (95% confidence interval =0.66-0.98; train ROC AUC = 0.82) using a logistic regression classifier. The model was well calibrated, and the Youden J Index suggested an optimal cut-off of 0.501 to discriminate between iCCA and HCC with a sensitivity of 0.733 and a specificity of 0.857.
Radiomics-based imaging biomarkers can potentially help to non-invasively discriminate between iCCA and HCC.
肝内胆管癌（iCCA）和肝细胞癌（HCC）在预后和治疗方面存在差异。我们旨在通过从标准护理对比增强计算机断层扫描（CT）中提取的放射组学方法对iCCA和HCC进行无创鉴别。
回顾性纳入2014年8月至2021年11月期间接受腹部对比增强CT检查、组织学确诊为iCCA（n = 47）或HCC（n = 47）的94例患者（男性，n = 68，平均年龄63.3 ± 12.4岁）。通过为每个肿瘤定义三个三维感兴趣体积，以临床可行的方式手动分割增强的肿瘤边界。提取放射组学特征。采用组内相关分析和Pearson指标对稳健且非冗余的特征进行分层，并通过LASSO（最小绝对收缩和选择算子）进一步进行特征约简。使用独立的训练和测试数据集构建四种不同的机器学习模型。计算性能指标和特征重要性值以提高模型的可解释性。
将患者群体分为65例用于训练（iCCA，n = 32）和29例用于测试（iCCA，n = 15）。最终由三个放射组学特征以及年龄和性别临床特征组成的联合特征集，使用逻辑回归分类器时，测试模型的曲线下受试者操作特征（ROC）面积表现最佳，曲线下面积（AUC） = 0.82（95%置信区间 = 0.66 - 0.98；训练ROC AUC = 0.82）。该模型校准良好，约登J指数表明最佳截断值为0.501，用于区分iCCA和HCC，灵敏度为0.733，特异性为0.857。
基于放射组学的影像生物标志物可能有助于对iCCA和HCC进行无创鉴别。
