Leveraging Label-Free Digital Imaging to Enable Rapid On-Site Evaluation (ROSE) During Thyroid Biopsy Procedures

项目总结/摘要 Pathware正在开发Bioptic™，这是一种成像平台，用于在护理点分析活检样本。的 系统将快速地对未染色活检样本进行染色，并利用计算机辅助检测/诊断， 病理学家快速有效的评估。床旁诊断的典型障碍通过以下方式克服： 将新型计算显微镜与AI（人工智能）算法相结合， 临床感兴趣的区域。这项拟议中的研究的广泛，长期的目标是使真实的- 在护理点对病理学样本进行时间诊断。 甲状腺结节的发病率很高，通过细针穿刺（FNA）活检进行评估通常是一种有效的方法。 诊断流程中不可或缺的一步。活检样本通常不足以诊断， 重复活检或不必要的手术切除。样本充足性的快速现场评估（ROSE） 非常适合患者护理，因为它通过验证诊断质量来降低样本不足的风险。 即时护理尽管活检不充分的发生率很高，但ROSE在不到25%的甲状腺中进行 由于财务和操作障碍，包括病理学家的可用性，活检程序。当前 ROSE的替代方案（远程病理学、机器人显微镜和实时流媒体显微镜）受到障碍的困扰 由于染色样品的要求、技术挑战和法规限制，人工智能- 辅助ROSE有可能克服这些障碍，并提供一个上级护理标准。 该项目的近期目标是：1）展示未染色甲状腺的非破坏性成像 细胞学涂片与一种新的显微镜模式和2）训练人工智能算法对所获得的图像，为未来 开发ROSE充分性评估的计算机辅助检测算法。

Project Summary/Abstract Pathware is developing Bioptic™, an imaging platform to analyze biopsy samples at the point-of-care. The system will rapidly digitize unstained biopsy samples and utilize computer-aided detection/diagnostics to enable quick and efficient evaluations by pathologists. Typical barriers to point-of-care diagnostics are overcome by combining novel computational microscopy with AI (Artificial Intelligence) algorithms to direct on-site or remote pathologists to regions of clinical interest. This proposed research's broad, long-term objective is to enable real- time diagnostics of pathology samples at the point-of-care. Thyroid nodules have a high incidence rate, and assessment via fine-needle aspiration (FNA) biopsy is often an integral step in the diagnostic pipeline. The biopsy samples are frequently inadequate for diagnosis resulting in repeat biopsies or unnecessary surgical resections. Rapid On-Site Evaluation (ROSE) for sample adequacy is ideal for patient care, because it mitigates the risk of inadequate samples by verifying diagnostic quality at the point-of-care. Despite the high incidence of inadequate biopsies, ROSE is performed in less than 25% of thyroid biopsy procedures due to financial and operational barriers, including pathologist availability. The current alternatives to ROSE (telepathology, robotic microscopy, and live-streaming microscopy) are plagued by barriers to adoption due to the requirement of staining samples, technical challenges, and regulatory limitations. AI- assisted ROSE has the potential to overcome these barriers and provide a superior standard-of-care. The immediate goals of the proposed project are to 1) demonstrate non-destructive imaging of unstained thyroid cytology smears with a novel microscopy modality and 2) train AI algorithms on the acquired images for future development of computer-aided detection algorithms for ROSE adequacy assessments.