SBIR Phase II: Development of AI Software to Capture and Identify Circulating Rare Cells in Lung Patients

这项小企业创新研究（SBIR）二期项目的更广泛影响/商业潜力是一种新的癌症治疗液体活检产品，该产品使用人工智能（AI），可以从抽血中检测和分类癌症来源的罕见细胞（CRC）。有100多种不同类型的癌症，预计到2022年，美国将诊断出190多万例新的癌症病例，导致60多万人死亡（每天1670人死亡）。癌症是美国第二大常见死因，仅次于心脏病。目前正在为相当一部分癌症开发新的治疗方法，世界各地正在进行许多新疗法的临床试验。这种微创、高灵敏度的血液测试将以低成本监测治疗反应和进展，支持这些新的癌症治疗方法的发展。具体来说，通过一种侵入性更小、更全面的诊断工具，测试结果将使临床研究人员实时了解癌症肿瘤生物学，更好地了解癌症的异质性。这个小企业创新研究（SBIR）二期项目将人工智能（AI），特别是用于计算机视觉的深度学习神经网络，与CRC免疫荧光试剂集成到免疫荧光显微镜中。这项工作的主要目标是以高精度识别和分类crc。越来越多的证据表明，crc与癌症类型、分期、治疗反应、最小残留疾病和整体疾病进展相关。然而，在一个典型的血液样本中，有超过700万个血液伪影，只有很少的红细胞存在。目前分析crc的技术是昂贵的，冗长的，并且在自动化方面受到限制。为了满足项目的敏感性、特异性和运行时间要求，AI图像分析将进一步优化为：1)发现crc， 2)区分假阳性，3)将crc划分为临床相关类型。开发的人工智能架构将通过使用数千个临床样本的广泛培训来选择，并与专业表征的癌症血液病理图像进行比较。在证明了高灵敏度和特异性之后，开发工作将继续使人工智能革命性的结直肠癌液体活检测试成熟，以满足临床研究用途（RUO）的要求。对于癌症研究界，该产品将用于进行非临床实验室研究。该奖项反映了美国国家科学基金会的法定使命，并通过使用基金会的知识价值和更广泛的影响审查标准进行评估，被认为值得支持。

The broader impact/commercial potential of this Small Business Innovation Research (SBIR) Phase II project is a new cancer treatment liquid biopsy product using Artificial Intelligence (AI) that can detect and classify cancer derived rare cell (CRC) from a blood draw. There are over 100 different types of cancers and over 1.9 million new cancer cases are expected to be diagnosed in the US in 2022 resulting in over 600,000 deaths (1,670 deaths per day). Cancer is the second most common cause of death in the US, exceeded only by heart disease. New treatment therapies are being developed for a substantial proportion of cancers with many clinical trials for new therapies on-going world-wide. The minimally invasive, high sensitivity blood test will monitor therapeutic response and progression at low-cost, supporting development of these new cancer treatments. Specifically, with a less invasive and more comprehensive diagnostic tool, the test results will give clinical researchers real-time insights into cancer tumor biology, providing better understanding of cancer heterogeneity. This Small Business Innovation Research (SBIR) Phase II project combines Artificial Intelligence (AI), specifically deep learning neural networks used for computer vision, with CRC immunofluorescent reagents integrated into an immunofluorescent microscope. The main objective of this effort is to identify and classify CRCs with high accuracy. There is increasing evidence that CRCs are correlated with cancer type, staging, treatment response, minimal residual disease, and overall disease progression. However, in a typical blood sample, there are over 7 million blood artifacts with very few CRCs present. Current techniques to analyze CRCs are expensive, lengthy, and are limited in automation. To meet project sensitivity, specificity, and runtime requirements, the AI image analysis will be further optimized to: 1) find CRCs, 2) discriminate against false positives, and 3) classify CRCs into clinically relevant types. The developed AI architectures will be selected through extensive training using thousands of clinical samples compared to expertly characterized cancer blood pathology images. After high sensitivity and specificity are demonstrated, development work will continue to mature the AI-revolutionized CRC liquid biopsy test to meet clinical research use only (RUO) requirements. For the cancer research community, the product offering will be used in the conduct of non-clinical laboratory research.This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.