Zhao Meixin, Kluge Kilian, Papp Laszlo, Grahovac Marko, Yang Shaomin, Jiang Chunting, Krajnc Denis, Spielvogel Clemens P, Ecsedi Boglarka, Haug Alexander, Wang Shiwei, Hacker Marcus, Zhang Weifang, Li Xiang
Department of Nuclear Medicine, Peking University Third Hospital, 49 North Garden Road, Haidian District, Beijing, 100191, People's Republic of China.
Department of Biomedical Imaging and Image-guided Therapy, Division of Nuclear Medicine, Medical University of Vienna, Währinger Gürtel 18-20, Floor 3L, 1090, Vienna, Austria.
Eur Radiol. 2022 Oct;32(10):7056-7067. doi: 10.1007/s00330-022-08999-7. Epub 2022 Jul 28.
This study investigates the ability of machine learning (ML) models trained on clinical data and 2-deoxy-2-[18F]fluoro-D-glucose(FDG) positron emission tomography/computed tomography (PET/CT) radiomics to predict overall survival (OS), tumor grade (TG), and histologic growth pattern risk (GPR) in lung adenocarcinoma (LUAD) patients.
A total of 421 treatment-naive patients with histologically-proven LUAD and available FDG PET/CT imaging were retrospectively included. Four cohorts were assessed for predicting 4-year OS (n = 276), 3-year OS (n = 280), TG (n = 298), and GPR (n = 265). FDG-avid lesions were delineated, and 2082 radiomics features were extracted and combined with endpoint-specific clinical parameters. ML models were built for the prediction of 4-year OS (M4OS), 3-year OS (M3OS), tumor grading (MTG), and histologic growth pattern risk (MGPR). A 100-fold Monte Carlo cross-validation with 80:20 training to validation split was employed as a performance evaluation for all models. The association between the M4OS and M3OS predictions with OS was assessed by the Kaplan-Meier survival analysis.
The area under the receiver operator characteristics curve (AUC) was the highest for M4OS (AUC 0.88, 95% confidence interval (CI) 86.7-88.7), followed by M3OS (AUC 0.84, CI 82.9-84.9), while MTG and MGPR performed equally well (AUC 0.76, CI 74.4-77.9, CI 74.6-78, respectively). Predictions of M4OS (hazard ratio (HR) -2.4, CI -2.47 to -1.64, p < 0.05) and M3OS (HR -2.36, CI -2.79 to -1.93, p < 0.05) were independently associated with OS.
ML models are able to predict long-term survival outcomes in LUAD patients with high accuracy. Furthermore, histologic grade and predominant growth pattern risk can be predicted with satisfactory accuracy.
• Machine learning models trained on pre-therapeutic PET/CT radiomics enable highly accurate long-term survival prediction of patients with lung adenocarcinoma. • Highly accurate survival predictions are achieved in lung adenocarcinoma patients despite heterogenous histologies and treatment regimens. • Radiomic machine learning models are able to predict lung adenocarcinoma tumor grade and histologic growth pattern risk with satisfactory accuracy.
本研究调查了基于临床数据和2-脱氧-2-[¹⁸F]氟-D-葡萄糖（FDG）正电子发射断层扫描/计算机断层扫描（PET/CT）影像组学训练的机器学习（ML）模型预测肺腺癌（LUAD）患者总生存期（OS）、肿瘤分级（TG）和组织学生长模式风险（GPR）的能力。
回顾性纳入421例未经治疗且经组织学证实为LUAD并具备可用FDG PET/CT影像的患者。评估了四个队列以预测4年总生存期（n = 276）、3年总生存期（n = 280）、肿瘤分级（n = 298）和组织学生长模式风险（n = 265）。勾画出FDG摄取病灶，提取2082个影像组学特征并与特定终点的临床参数相结合。构建了用于预测4年总生存期（M4OS）、3年总生存期（M3OS）、肿瘤分级（MTG）和组织学生长模式风险（MGPR）的ML模型。采用80:20训练与验证分割的100倍蒙特卡洛交叉验证作为所有模型的性能评估。通过Kaplan-Meier生存分析评估M4OS和M3OS预测与总生存期之间的关联。
M4OS的受试者操作特征曲线下面积（AUC）最高（AUC 0.88，95%置信区间（CI）86.7 - 88.7），其次是M3OS（AUC 0.84，CI 82.9 - 84.9），而MTG和MGPR表现相当（AUC分别为0.76，CI 74.4 - 77.9，CI 74.6 - 78）。M4OS（风险比（HR）-2.4，CI -2.47至-1.64，p < 0.05）和M3OS（HR -2.36，CI -2.79至-1.93，p < 0.05）的预测与总生存期独立相关。
ML模型能够高精度预测LUAD患者的长期生存结局。此外，组织学分级和主要生长模式风险也能以令人满意的精度进行预测。
• 基于治疗前PET/CT影像组学训练的机器学习模型能够对肺腺癌患者进行高精度的长期生存预测。
• 尽管组织学类型和治疗方案存在异质性，但肺腺癌患者仍能实现高精度的生存预测。
• 影像组学机器学习模型能够以令人满意的精度预测肺腺癌肿瘤分级和组织学生长模式风险。
