Virtual Fly-Over Colonoscopy System for Clinical Test-SBIR Phase 1

描述(申请人提供)：结直肠癌是美国第三种最常见的癌症，也是与癌症相关的死亡的第二大原因。然而，如果及早发现并切除息肉，结直肠癌基本上是可以预防的。虽然光学结肠镜(OC)是目前的黄金标准，但它可以检测到90%以上的结直肠肿瘤；它是侵入性的，可能会让患者感到不适、不便，并被患者视为不受欢迎。此外，即使是经验丰富的内窥镜医生也可能难以到达盲肠，导致对结肠的显示不完整。因此，虚拟结肠镜(VC)作为OC的替代品应运而生。在VC期间，使用一个虚拟相机来查看虚拟结肠的内壁，该内壁是通过腹部的CT扫描重建的。然而，目前的VC系统的临床吸引力有限，因为它们仅限于特定类型的息肉，可能会产生大量的假阳性，或者对大小为5-9 mm的显著息肉的检测率很低。我们建议通过这项SBIR工作实现商业化的新技术是一种改变游戏规则的、获得专利的VC可视化技术，称为“虚拟天桥”技术。这项技术对检测结肠息肉是敏感、有效和有效的。这项提案的总体目标是完成一种新的虚拟结肠镜可视化技术的开发和验证，该技术已获得路易斯维尔大学的专利。假设新的可视化技术将使人们能够更好地观察复杂的结肠拓扑结构，从而更好地检测息肉，特别是那些可能隐藏在隆起褶皱后面的息肉。目前的原型已经被用来评估20个临床数据集，取得了很好的结果。然而，在第二阶段之前，必须结合人工产物去除和用户友好的特征，在第二阶段中，该技术将被用于大规模临床验证试验，从而产生商业产品。此外，我们建议(1)在160个数据集的初步研究中产生更令人信服的初步数据，以及(2)在临床数据集中引入几个大小为5-9 mm的幻影息肉，以提供该技术有效性的统计意义。幻影息肉将被放置在传统上难以分析的位置，这对两个OC都构成了严重的检测问题 和当前的VC方法。根据文献，目前的OC方法导致61-91%(平均80%)的结肠观看。路易斯维尔大学在模仿经典OC的“飞越VC方法”中的研究结果是93.4%的收视率，新的“飞越”方法的收视率为97.5%。更重要的是改进了视角(“天空之眼”)，没有光学失真，增强了CAD功能，这将显著提高息肉识别能力，特别是在检测小结肠息肉、隐藏在半侧皱折后面的息肉，以及在解剖拐点处的折叠结肠段息肉时。我们预计，与今天的方法相比，难处理的息肉可视化能力的整体改进将超过30%，我们对与路易斯维尔大学将这项技术商业化感到兴奋。

DESCRIPTION (provided by applicant): Colorectal cancer is the third most common form of cancer and the second leading cause of cancer-related death in the US. Yet, if polyps are detected and removed early, colorectal cancer is largely preventable. Although optical colonoscopy (OC), the current gold standard, detects more than 90% of colorectal neoplasms; it is invasive and can be uncomfortable, inconvenient, and perceived as undesirable by patients. Furthermore, even experienced endoscopists may have difficulty reaching the cecum, resulting in incomplete visualizations of the colon. As a consequence, virtual colonoscopy (VC) has emerged as an alternative to OC. During VC, a virtual camera is used to view the internal walls of a virtual colon, reconstructed from CT scans of the abdominal cavity. However, current VC systems have had limited clinical appeal, as they are limited to specific types of polyps, may generate a large number of false positives, or have poor detection rates for significant polyps in the size range of 5-9 mm. The new technology we propose to commercialize through this SBIR work is a game changing, patented, visualization technique for VC, called the "virtual fly-over" technique. The technique is sensitive, effective, and efficient for detecting colon polyps. The overall objective of this proposal is to complete the development and validation of a novel visualization technique for virtual colonoscopy, which was patented by the University of Louisville. The hypothesis is that the new visualization technique will enable better viewing of the complex colonic topology, and hence a better capability to detect polyps, especially those that may be hidden behind haustral folds. The current prototype has been utilized to evaluate twenty clinical datasets, with excellent results. However, artifact removal and user friendly features must be incorporated prior to Phase II, in which the technology will be utilized in a larg scale clinical validation trial leading to a commercial product. Also, we propose to (1) generate more convincing preliminary data in a pilot study of 160 datasets, and (2) introduce several phantom polyps, in the size range of 5-9 mm, into the clinical datasets, in order to provide statistical significance of the technology's effectiveness. The phantom polyps will be placed in traditionally difficult-to-analyze positions, which pose significant detection problems for both OC and current VC methods. According to the literature, current OC methods result in a 61-91% (average 80%) viewing of the Colon. The University of Louisville's work in the "fly-through VC method", which mimics classic OC, results in 93.4 percent viewing, and the new "fly-over" method results in 97.5% percent viewing. Even more important is the improved point of view ("eye-in-the-sky"), the lack of optical distortion, and enhanced CAD functionality that will increase polyp recognition dramatically, especially when detecting small colon polyps, polyps hidden behind haustral folds, and polyps in folded colonic segments at anatomical inflection points. We anticipate the overall improvement in the ability to visualize difficult polyps to be upwards of 30% compared to today's methods, and we are excited about commercializing this technology with the University of Louisville.