Zhou Yi, Zhou Xue-Yan, Xu Yu-Chao, Ma Xue-Lei, Tian Rong
Department of Nuclear Medicine, West China Hospital, Sichuan University, Chengdu, China.
Key Laboratory of Bio-Resource and Eco-Environment of Ministry of Education, College of Life Sciences, Sichuan University, Chengdu, China.
Quant Imaging Med Surg. 2025 Jan 2;15(1):103-120. doi: 10.21037/qims-24-585. Epub 2024 Dec 30.
Radiomics features extracted from baseline F-fluorodeoxyglucose positron emission tomography (F-FDG PET) scans have shown promising results in predicting the treatment response and outcome of diffuse large B-cell lymphoma (DLBCL) patients. This study aimed to assess the influence of lesion selection approaches and segmentation methods on the radiomics of DLBCL in terms of treatment response and prognosis prediction.
A total of 522 and 382 patients pathologically diagnosed with DLBCL were enrolled for complete regression and 2-year event-free survival prediction, respectively. Three lesion selection methods (largest or hottest lesion, patient level) and five segmentation methods (manual and four semiautomatic segmentations) were applied. A total of 112 radiomics features were extracted from the lesions and at the patient level. The feature selection was performed via random forest, and models were constructed via eXtreme Gradient Boosting. The performance of all the models was evaluated via the area under the curve (AUC), which was compared via the Delong test.
The AUC values varied from 0.583 to 0.768 for the treatment response and prognosis prediction models on the basis of different lesion selection and segmentation methods. However, the prediction performance gap was not significant for each model (all P>0.05). The combined models (AUC =0.908 and 0.837 for treatment response and prognosis prediction, respectively) that incorporated radiomics and clinical features exhibited significant predictive superiority over the clinical models (AUC =0.622 and 0.636, respectively) and the international prognostic index model (AUC =0.623 for prognosis prediction) (all P<0.05).
Although there are differences in the selected radiomics features among lesion selection and segmentation methods, there is no substantial difference in the predictive power of each radiomics model. In addition, radiomics features have potential added value to clinical features.
从基线氟脱氧葡萄糖正电子发射断层扫描（F-FDG PET）图像中提取的放射组学特征在预测弥漫性大B细胞淋巴瘤（DLBCL）患者的治疗反应和预后方面已显示出有前景的结果。本研究旨在评估病变选择方法和分割方法对DLBCL放射组学在治疗反应和预后预测方面的影响。
分别纳入522例和382例经病理诊断为DLBCL的患者，用于完全缓解和2年无事件生存预测。应用了三种病变选择方法（最大或最热点病变、患者层面）和五种分割方法（手动分割和四种半自动分割）。从病变和患者层面共提取了112个放射组学特征。通过随机森林进行特征选择，并通过极端梯度提升构建模型。所有模型的性能通过曲线下面积（AUC）进行评估，并通过德龙检验进行比较。
基于不同的病变选择和分割方法，治疗反应和预后预测模型的AUC值在0.583至0.768之间变化。然而，每个模型的预测性能差距并不显著（所有P>0.05）。结合放射组学和临床特征的联合模型（治疗反应和预后预测的AUC分别为0.908和0.837）相对于临床模型（AUC分别为0.622和0.636）和国际预后指数模型（预后预测的AUC为0.623）表现出显著的预测优势（所有P<0.05）。
尽管病变选择和分割方法之间所选的放射组学特征存在差异，但每个放射组学模型的预测能力没有实质性差异。此外,放射组学特征对临床特征具有潜在的附加价值。
