Improving the Generalizability of Deep Neural Networks by Teaching them Lung Cancer Pathophysiology

通过教授肺癌病理生理学来提高深度神经网络的通用性

• 批准号: 10529498
• 负责人: Axel Herve Masquelin
• 金额: $ 3.32万
• 依托单位: UNIVERSITY OF VERMONT & ST AGRIC COLLEGE
• 依托单位国家: 美国
• 财政年份: 2022
• 资助国家: 美国
• 起止时间: 2022-08-01 至 2023-08-30
• 项目状态: 已结题
• 来源: https://reporter.nih.gov/project-details/10529498
• 关键词: Achievement; Address; Advisory Committees; Age-Years; Anxiety; Attention; Benign; Biological; Biological Markers; Caliber; Cancer Detection; Characteristics; Chronic; Chronic Obstructive Pulmonary Disease; Classification; Clinical; Data; Data Set; Detection; Diagnostic; Early Diagnosis; Educational process of instructing; Failure; Fibrosis; Foundations; Functional disorder; Future; Genetic; Genomics; Goals; Growth; Guidelines; Healthcare Systems; Histologic; Image; Individual; Inflammation; Investigation; Knowledge; Learning; Left; Lung; Lung nodule; Machine Learning; Malignant - descriptor; Malignant Neoplasms; Malignant neoplasm of lung; Measures; Medical; Methodology; Modality; Modeling; Morphology; Network-based; Nodule; Non-Small-Cell Lung Carcinoma; Outcome; Pathogenesis; Patients; Pattern; Performance; Phenotype; Physiological; Play; Precancerous Conditions; Predisposition; Preventive service; Procedures; Process; Pulmonary Emphysema; Pulmonary Inflammation; Quality of life; Recommendation; Research Personnel; Risk; Risk Factors; Role; Scanning; Sensitivity and Specificity; Smoker; Smoking History; Solid; Specific qualifier value; Specificity; Spirometry; Testing; Training; Training Programs; United States; Work; accurate diagnosis; adjudicate; airway obstruction; base; cancer diagnosis; career; clinically significant; computed tomography screening; convolutional neural network; cost; deep learning; deep neural network; diagnostic accuracy; diagnostic biomarker; directed attention; endophenotype; follow-up; genetic variant; high risk; histological image; idiopathic pulmonary fibrosis; imaging biomarker; improved; insight; low dose computed tomography; lung imaging; mortality; novel; novel strategies; radiologist; screening; trait; tumor; vector

PROJECT SUMMARY Diagnostic and treatment approaches for non-small cell lung cancer (NSCLC) have evolved over the last decade from primarily empirical methodologies to objective strategies that rely on clinical characteristics of the patient and morphological features of the nodule. Following recommendations by the United States Preventive Service Task Force (USPSTF), high-risk individuals are screened yearly with low-dose computed tomography (LDCT) as this provides high sensitivity with acceptable specificity for lung cancer. However, the introduction of LDCT as the primary screening modality for lung cancer has increased detection rates of indeterminate pulmonary nodules that then require invasive investigation. This decreases the quality of life for at-risk individuals through repeated follow-ups and procedures, and greatly increases anxiety over what usually turns out to a benign nodule. In this proposal, we aim to improve upon these outcomes by determining the features that convolutional neural networks (CNNs) utilize when classifying lung nodules as either or benign. We will also determine if providing CNNs with pre-specified histologic image features known to be associated with lung cancer improves their ability to generalize to novel images outside the image set used to train them. The central hypothesis of this proposal is that increasing the attention of a CNN on LDCT image features that are accepted as being pathophysiologically relevant will improve its generalizability to novel images and thus its ability to accurately distinguish between malignant versus benign nodules. In the F99 Aim of this proposal, we will address this hypothesis by utilizing LDCT images from the National Lung Screening Trial (NLST) together with concept activation vectors to determine which parenchymal and tumor-specific features are used by CNNs to classify lung nodules. In the K00 aim, we will determine if endophenotypes extracted from the COPDgene LDCT image set can be used to improve CNN generalizability. Completion of these aims will lead to an increased understanding of the morphologic biomarkers of lung cancer inherent in LDCT images of the lung that are most important for accurate diagnosis. This will have potential application to the improvement of CNN classification performance in other medical domains. In addition, by adhering to the training program outlined in this proposal I will gain high levels of expertise in image biomarkers, early cancer pathogenesis and detection, genetic networks, and genomics. These will collectively serve as a solid foundation for my future career as an independent biomedical investigator.

项目摘要 非小细胞肺癌（NSCLC）的诊断和治疗方法在过去十年中不断发展 从主要的经验方法到依赖于患者临床特征的客观策略 和结节的形态特征。根据美国预防服务局的建议， 工作组（USPSTF），每年对高危人群进行低剂量计算机断层扫描（LDCT）筛查 因为这为肺癌提供了高灵敏度和可接受的特异性。然而，LDCT的引入 作为肺癌的主要筛查方式， 然后需要侵入性检查的结节。这降低了高危人群的生活质量， 重复的后续行动和程序，并大大增加了焦虑，什么通常是良性的， 结节。在这个提议中，我们的目标是通过确定卷积的特征来改善这些结果。 神经网络（CNN）在将肺结节分类为良性或良性时使用。我们还将确定， 为CNN提供已知与肺癌相关的预先指定的组织学图像特征， 他们的能力，以推广到新的图像以外的图像集用来训练他们。的中心假设 该建议是，增加CNN对LDCT图像特征的关注， 病理生理相关性将提高其对新图像的概括性， 以准确区分恶性结节和良性结节。在本提案的F99目标中，我们 我将通过利用国家肺筛查试验（NLST）的LDCT图像来解决这一假设 使用概念激活向量来确定CNN使用哪些实质和肿瘤特异性特征 对肺结节进行分类在K00目标中，我们将确定从COPD基因中提取的内表型是否 LDCT图像集可用于提高CNN的泛化能力。这些目标的实现将导致 了解肺部LDCT图像中固有的肺癌形态学生物标志物， 这对准确诊断很重要。这将对CNN分类的改进具有潜在的应用价值 在其他医疗领域的表现。此外，通过遵守本建议书中概述的培训计划， 我将获得高水平的专业知识，在图像生物标志物，早期癌症发病机制和检测，遗传 网络和基因组学。这些将共同为我未来的职业生涯奠定坚实的基础， 独立的生物医学调查员