Development of Novel Computer-aided Diagnosis Systems

新型计算机辅助诊断系统的开发

• 批准号: RGPIN-2014-05215
• 负责人: Mandal, Mrinal
• 金额: $ 1.82万
• 依托单位: University of Alberta
• 依托单位国家: 加拿大
• 项目类别: Discovery Grants Program - Individual
• 财政年份: 2015
• 资助国家: 加拿大
• 起止时间: 2015-01-01 至 2016-12-31
• 项目状态: 已结题
• 来源: https://www.nserc-crsng.gc.ca/ase-oro/Details-Detailles_eng.asp?id=588277
• 关键词: Development; Novel; Computer; aided; Diagnosis

A phenomenal growth in the use of digital images has been observed in recent times in medical diagnosis by physicians. Picture archiving and communications system (PACS) are being widely used in health care as it offers several advantages such as low cost, improved quality, and flexibility in sharing patient data among the health care professionals. It also opens up the possibility of using computers to speed up diagnosis and reduce subjectiveness. The computer-aided diagnosis (CAD) system are already being used for breast cancer screening in many countries, and is expected to be extended for diagnosis of many other kinds of diseases. The image analysis techniques play a key role in developing a CAD system. However, the characteristics of images vary significantly across different modalities (e.g., X-ray, CT, MRI, histopathology), tissue/organ types, and diseases. It is a huge challenge to develop image analysis techniques for such a wide variety of images. In this proposal, we propose to carry on research in two areas: Computer aided diagnosis of Skin Cancer and Tuberculosis (TB). The skin cancer, especially Melanoma, and drug-resistant TB ared deadly diseases with high mortality rate. Early screening of these diseases through computer-aided diagnosis would be very useful in health care. In skin cancer, the proposed research will explore development of histopathological image analysis techniques to improve the diagnosis of Melanoma and Basal cell (the most frequent type of skin cancer) cancer. Techniques would be developed for analyzing both epidermis and dermis areas in a histopathological image. Techniques will also be developed for detecting clusters of nuclei, which are observed in the junction of epidermis and dermis area. Due to the noisy nature of the images, the performance of existing mitosis detection technique is low, and we would explore developing a robust mitosis detection technique based on the sparse representation and classification technique. Appropriate techniques for 3D whole slide histopathological image reconstruction and visualization will be developed to enable the pathologist to see a slide in 3D. In TB diagnosis, robust techniques would be developed for more accurate diagnosis. In addition to cavity, and acinar shadows, the "volume loss" is a hallmark of TB as a result of its destructive and fibrotic nature. The Hilar and mediastinal lymphadenopathy are good indicators of HIV-related TB, and primary TB. Techniques will be developed to detect these indicators, and reduce the false positive rate of typical primary and postprimary TB. Experiments will be conducted with larger databases and with a wider variety of images such as images with acinar shadows, atypical or normal radiographs without acinar shadows, and radiographs with nodules but without acinar shadows. The proposed research will establish a detailed understanding of several important issues related to computer-aided diagnosis of skin cancer, and TB. To our knowledge, we are the first group working on developing a computer-aided diagnosis system for melanoma. We are also one of very few groups who are working on development of efficient computer-aided diagnosis system for TB. A successful development is expected to reduce the diagnosis time and error, and will result in intellectual property.

近年来，在医生的医疗诊断中，数字图像的使用出现了惊人的增长。图像存档和通信系统(PACS)因其具有低成本、高质量和灵活地在医疗专业人员之间共享患者数据等优点而被广泛应用于医疗保健领域。它还为使用计算机加快诊断速度和减少主观性开辟了可能性。计算机辅助诊断(CAD)系统已经在许多国家用于乳腺癌筛查，并有望扩展到许多其他类型的疾病的诊断。图像分析技术在开发CAD系统中起着至关重要的作用。然而，不同的模式(例如，X射线、CT、MRI、组织病理学)、组织/器官类型和疾病，图像的特征显著不同。开发针对如此广泛种类的图像的图像分析技术是一个巨大的挑战。 在这个方案中，我们建议在两个领域进行研究：皮肤癌和结核病的计算机辅助诊断。皮肤癌，特别是黑色素瘤和耐药结核病是死亡率较高的致命性疾病。通过计算机辅助诊断对这些疾病进行早期筛查将在医疗保健中非常有用。在皮肤癌方面，拟议的研究将探索组织病理图像分析技术的发展，以提高对黑色素瘤和基底细胞癌(最常见的皮肤癌类型)的诊断。将开发出在组织病理学图像中分析表皮和真皮区域的技术。还将开发检测在表皮和真皮区域交界处观察到的核簇的技术。由于图像的噪声性质，现有的有丝分裂检测技术性能较低，我们将探索开发一种基于稀疏表示和分类技术的稳健的有丝分裂检测技术。将开发适用于3D整个切片的组织病理学图像重建和可视化的技术，以使病理学家能够在3D中看到切片。在结核病诊断方面，将开发稳健的技术，以实现更准确的诊断。除了空洞和腺泡阴影外，由于结核病具有破坏性和纤维化的性质，“体积损失”也是结核病的一个标志。肺门和纵隔淋巴结肿大是HIV相关结核病和原发结核病的良好指标。将开发技术来检测这些指标，并降低典型的初治和初治后结核病的假阳性率。将使用更大的数据库和更多种类的图像进行实验，例如带有腺泡阴影的图像、没有腺泡阴影的非典型或正常X射线照片以及带有结节但没有腺泡阴影的X射线照片。 这项拟议的研究将建立对皮肤癌和结核病的计算机辅助诊断相关的几个重要问题的详细理解。据我们所知，我们是第一个致力于开发黑色素瘤计算机辅助诊断系统的小组。我们也是为数不多的致力于开发有效的结核病计算机辅助诊断系统的小组之一。一项成功的开发有望减少诊断时间和错误，并将产生知识产权。