An ensemble deep learning model for tumor bud detection and risk stratification in colorectal carcinoma.

用于结直肠癌肿瘤芽检测和风险分层的集成深度学习模型。

• 批准号: 10564824
• 负责人: Wei Chen
• 金额: $ 54.37万
• 依托单位: OHIO STATE UNIVERSITY
• 依托单位国家: 美国
• 财政年份: 2023
• 资助国家: 美国
• 起止时间: 2023-07-01 至 2028-06-30
• 项目状态: 未结题
• 来源: https://reporter.nih.gov/project-details/10564824
• 关键词: Agreement; American; American Cancer Society; Biological; Cancer Etiology; Cell Adhesion; Cell Nucleus; Cell Polarity; Cells; Cessation of life; Clinical; Clinical Management; Color; Colorectal; Colorectal Cancer; Compensation; Computer Assisted; Computer software; Computer-Assisted Image Analysis; Consensus; Consumption; Cytokeratin; Decision Making; Detection; Diagnostic; Disease; Disease-Free Survival; Evaluation; Exclusion; Future; Guidelines; Head and Neck Squamous Cell Carcinoma; Health; Hematoxylin and Eosin Staining Method; Human; Image; Image Analysis; Incidence; International; Interobserver Variability; Large Intestine Carcinoma; Learning; Learning Module; Malignant Neoplasms; Malignant neoplasm of cervix uteri; Malignant neoplasm of esophagus; Manuals; Mathematics; Metastatic Neoplasm to Lymph Nodes; Methods; Microsatellite Instability; Mission; Modality; Modeling; Morphology; Neoplasm Metastasis; Outcome; Pathologic; Pathologist; Pathology; Pathology Report; Patient Care; Patient Selection; Patient-Focused Outcomes; Patients; Performance; Play; Prevalence; Prognostic Factor; Progression-Free Survivals; Protocols documentation; Public Health; Qualifying; Recommendation; Regional Anatomy; Reproducibility; Research; Role; Scientist; Shapes; Slide; Software Tools; Specimen; Squamous cell carcinoma; Stains; Standardization; Statistical Data Interpretation; System; Systems Analysis; Techniques; Time; Training; Tumor stage; United States; United States National Institutes of Health; Variant; Visual; Woman; Work; automated image analysis; cancer risk; cancer type; chemotherapy; cohort; college; colon cancer patients; colorectal cancer risk; colorectal cancer treatment; computer aided detection; deep learning; deep learning algorithm; deep learning model; differential geometry; disorder risk; epithelial to mesenchymal transition; experience; image registration; improved; improved outcome; innovation; lifetime risk; malignant breast neoplasm; malignant stomach neoplasm; men; mortality; neoplastic cell; novel; novel strategies; outcome prediction; patient stratification; personalized cancer therapy; quantitative imaging; risk stratification; symposium; theories; tool; tumor; whole slide imaging

PROJECT ABSTRACT Colorectal cancer (CRC) is the fourth most common cancer, and the second leading cause of cancer death in the United States, with an estimated incidence of 151,030 new cases in 2022. According to the American Cancer Society, the lifetime risk of developing colorectal cancer is 1 in 23 for men and 1 in 25 for women. Tumor budding is a prognostic factor in colorectal cancer with potential to risk stratify patients and possibly guide treatment decisions. It is defined as the presence of a single tumor cell or a cell cluster consisting of fewer than five tumor cells at the invasive tumor front. Unfortunately, tumor budding is not routinely disclosed in pathology reports due to lack of reproducible methods in identifying tumor buds from H&E slides. The prevalence, mortality, and risk of colorectal cancer as well as the potential of tumor budding as a prognostic factor necessitate an accurate, easy- to-use, reproducible system to identify tumor budding. We aim to develop a computer-aided image analysis system to standardize the quantitative criteria used to define tumor budding from H&E slides. In addition to identifying tumor buds, the system will correlate tumor buds with several outcomes (microsatellite instability status, overall survival, progression free survival, and disease free survival). As part of the proposed computer- aided image analysis system, we will first develop a sophisticated method for color deconvolution to compensate for color variations. This will be followed by deformable image registration and deep learning modules to differentiate tumor from non-tumor regions. The study will show that machines can be trained using deep learning to identify different anatomical regions within H&E slides of colorectal patients. From thereon, we will rely on scale-space theory and alpha-shapes to identify tumor buds and hotspots. We will use mathematical morphology and differential geometry to extract visually meaningful imaging features from tumor buds and hotspots. We will explore the potential of these imaging features along with features produced by our unsupervised multiple instance learning in predicting outcomes. The proposed research will help identify the association of tumor budding to colorectal cancer outcomes. The model will be subjected to rigorous statistical analysis for accuracy and reproducibility. The project will result in innovative software tools that facilitate the selection for personalized cancer therapies for colorectal patients.

项目摘要 结直肠癌 (CRC) 是第四大常见癌症，也是癌症死亡的第二大原因 美国，预计 2022 年新发病例数为 151,030 例。根据美国癌症协会的数据 社会上，男性终生患结直肠癌的风险为二十分之一，女性为二十五分之一。肿瘤出芽 是结直肠癌的一个预后因素，有可能对患者进行风险分层并可能指导治疗 决定。它被定义为存在单个肿瘤细胞或由少于五个肿瘤组成的细胞簇 位于侵袭性肿瘤前沿的细胞。不幸的是，由于以下原因，肿瘤出芽通常不会在病理报告中披露： 缺乏从 H&E 载玻片中识别肿瘤芽的可重复方法。患病率、死亡率和风险 结直肠癌以及肿瘤出芽作为预后因素的潜力需要一种准确、简单的方法 易于使用、可重复的系统来识别肿瘤出芽。我们的目标是开发计算机辅助图像分析 系统标准化用于定义 H&E 载玻片肿瘤出芽的定量标准。此外 识别肿瘤芽后，系统会将肿瘤芽与多种结果相关联（微卫星不稳定性） 状态、总生存期、无进展生存期和无病生存期）。作为拟议计算机的一部分 辅助图像分析系统，我们将首先开发一种复杂的颜色反卷积方法来补偿 用于颜色变化。接下来是可变形图像配准和深度学习模块 区分肿瘤与非肿瘤区域。该研究将表明可以使用深度学习来训练机器 识别结直肠患者 H&E 幻灯片中的不同解剖区域。从此，我们将依靠 尺度空间理论和阿尔法形状来识别肿瘤芽和热点。我们将使用数学形态学 和微分几何从肿瘤芽和热点中提取视觉上有意义的成像特征。我们将 探索这些成像特征以及我们无监督多重生成的特征的潜力 预测结果的实例学习。拟议的研究将有助于确定肿瘤的关联 结直肠癌的结果正在萌芽。该模型将经过严格的统计分析以确保准确性 和再现性。该项目将带来创新的软件工具，促进个性化选择 结直肠癌患者的癌症治疗。