论著摘要 |【多模态】用于预测肺癌立体定向放射治疗后远处转移的多目标影像组学模型

2017-08-28 11:50:28 admin 23

Multi-objective radiomics model for predicting distant failure in lung SBRT.

发表日期:2017.6.7    来源:Phys Med Biol. 

作者:Zhou Z1Folkert MIyengar PWestover KZhang YChoy HTimmerman RJiang SWang J.

作者介绍

    1.Department of Radiation Oncology, UT Southwestern Medical Center, Dallas, TX, United States of America.

摘要

立体定向放射治疗(SBRT)在非理想手术患者的早期非小细胞肺癌患者中表现出较高的局部控制率。然而,SBRT后的远处转移仍然很常见。对于在SBRT治疗后发生早期远处转移的高风险患者,额外的全身治疗可能降低远处复发风险并提高总体存活率。因此需要一种可以正确分层高转移风险患者的策略。影像组学领域通过使用高通量提取定量成像特征,在预测治疗结果方面具有很大的潜力。影像组学预测模型的构建通常基于单一目标,如总精度或曲线下面积(AUC)。然而由于训练数据集中的正负事件不平衡,单一的目标可能并不是指导模式建立的理想选择。为了突破这些限制,我们提出一个多目标影像组学模型,将灵敏度和特异性同时视为目标函数。为了设计更加准确可靠的模型,提出一种迭代多目标免疫算法(IMIA)来优化这些目标函数。多目标影像组学模型与单目标模型相比更为敏感,同时保持同等的特异性和AUC水平。 IMIA表现优于传统的免疫启发多目标算法。

Stereotactic body radiation therapy (SBRT) has demonstrated high local control rates in early stage non-small cell lung cancer patients who are not ideal surgical candidates. However, distant failure after SBRT is still common. For patients at high risk of early distant failure after SBRT treatment, additional systemic therapy may reduce the risk of distant relapse and improve overall survival. Therefore, a strategy that can correctly stratify patients at high risk of failure is needed. The field of radiomics holds great potential in predicting treatment outcomes by using high-throughput extraction of quantitative imaging features. The construction of predictive models in radiomics is typically based on a single objective such as overall accuracy or the area under the curve (AUC). However, because of imbalanced positive and negative events in the training datasets, a single objective may not be ideal to guide model construction. To overcome these limitations, we propose a multi-objective radiomics model that simultaneously considers sensitivity and specificity as objective functions. To design a more accurate and reliable model, an iterative multi-objective immune algorithm (IMIA) was proposed to optimize these objective functions. The multi-objectiveradiomics model is more sensitive than the single-objective model, while maintaining the same levels of specificity and AUC. The IMIA performs better than the traditional immune-inspired multi-objective algorithm.


慧影医疗科技(北京)有限公司

地点:北京市海淀区中关村东升科技园B2-C103

电话:400-890-9020

邮箱:radcloud@huiyihuiying.com

关闭
图片
图片
  • 人工智能诊断云平台