Castello Angelo, Manco Luigi, Cattaneo Margherita, Orlandi Riccardo, Rosso Lorenzo, Croci Giorgio Alberto, Florimonte Luigia, Scribano Giovanni, Turra Alessandro, Ferrero Stefano, Nosotti Mario, Carrafiello Gianpaolo, Castellani Massimo, Mendogni Paolo
Department of Nuclear Medicine, Fondazione IRCCS Ca' Granda, Ospedale Maggiore Policlinico, 20122 Milan, Italy.
Medical Physics Unit, University Hospital of Ferrara, 44124 Ferrara, Italy.
Bioengineering (Basel). 2025 Jun 30;12(7):721. doi: 10.3390/bioengineering12070721.
This study aimed to develop machine learning (ML) models to predict recurrence in thymoma patients using conventional and radiomic signatures extracted from preoperative [F]FDG PET/CT. A total of 50 patients (25 males, 25 females; mean age 63.3 ± 14.2 years) who underwent thymectomy and preoperative [F]FDG PET/CT between 2012 and 2022 were retrospectively analyzed. Radiomic analysis was performed using free-from-recurrence (FFR) status as a reference. Clinico-metabolic PET parameters were collected, and thymoma lesions were manually segmented on [F]FDG PET/CT. A total of 856 radiomic features (RFts) were extracted from PET and CT datasets following IBSI guidelines, and robust RFts were selected. The dataset was split into training (70%) and validation (30%) sets. Two ML models (PET- and CT-based, respectively), each with three classifiers-Random Forest (RF), Support-Vector-Machine, and Tree-were trained and internally validated using RFts and clinico-metabolic signatures. A total of 50 ROIs were selected and segmented. FFR was observed in 84% of our cohort. Forty-three robust RFts were selected from the CT dataset and 16 from the PET dataset, predominantly wavelet-based RFts. Additionally, three metabolic PET parameters were selected and included in the PET Model. Both the CT and PET models successfully discriminated against FFR after surgery, with the CT Model slightly outperforming the PET Model across different classifiers. The performance metrics of the RF classifier for the CT and PET models were AUC = 0.970/0.949, CA = 0.880/0.840, Precision = 0.884/0.842, Recall = 0.880/0.846, Specificity = 0.887/0.839, Sensitivity = 0.920/0.844, TP = 81.8%/83.3%, and TN = 92.9%/84.6%, respectively. ML-models trained on PET/CT radiomic features show promising results for predicting recurrence in patients with thymomas, which could be potentially applied in clinical practice for a better personalized treatment strategy.
本研究旨在开发机器学习（ML）模型，以利用从术前[F]FDG PET/CT中提取的传统特征和影像组学特征来预测胸腺瘤患者的复发情况。对2012年至2022年间接受胸腺切除术及术前[F]FDG PET/CT检查的50例患者（25例男性，25例女性；平均年龄63.3±14.2岁）进行回顾性分析。以无复发（FFR）状态为参考进行影像组学分析。收集临床代谢PET参数，并在[F]FDG PET/CT上手动分割胸腺瘤病变。按照国际生物医学影像标准倡议（IBSI）指南从PET和CT数据集中提取了总共856个影像组学特征（RFts），并选择了稳健的RFts。将数据集分为训练集（70%）和验证集（30%）。分别使用PET和CT数据集训练了两个ML模型（分别基于PET和CT），每个模型都有三个分类器——随机森林（RF）、支持向量机和决策树，并使用RFts和临床代谢特征进行内部验证。总共选择并分割了50个感兴趣区域（ROI）。在我们的队列中，84%的患者观察到FFR。从CT数据集中选择了43个稳健的RFts，从PET数据集中选择了16个，主要是基于小波的RFts。此外，选择了三个代谢PET参数并纳入PET模型。CT和PET模型均成功区分了术后的FFR情况，在不同分类器中，CT模型的表现略优于PET模型。CT和PET模型的RF分类器的性能指标分别为：曲线下面积（AUC）=0.970/0.949，准确率（CA）=0.880/0.840，精确率（Precision）=0.884/0.842，召回率（Recall）=0.880/0.846，特异性（Specificity）=0.887/0.839，敏感性（Sensitivity）=0.920/0.844，真阳性率（TP）=81.8%/83.3%，真阴性率（TN）=92.9%/84.6%。基于PET/CT影像组学特征训练的ML模型在预测胸腺瘤患者复发方面显示出有前景的结果，这可能潜在地应用于临床实践以制定更好的个性化治疗策略。
