Confocal video-mosaicking microscopy to guide surgery of superficially spreading skin cancers

1个来自PANEL项目R 01 CA 240771 2浅表扩散型皮肤癌，如恶性雀斑样痣黑色素瘤（Lentigo maligna melanomas，Lentigo）和非黑色素瘤 3皮肤癌（NMSCs）多发生于老年患者，以弥漫性亚临床亚表面扩散为主， 4个区域，边缘定义不清，难以检测。为了治疗这些癌症，皮肤科医生- 5定期进行大量的标测活检，以确定扩散和边缘，然后进行手术治疗。 6切除时留有较宽的“安全”边缘。毫不奇怪，这种“盲目”的方法会导致 7边缘，正常皮肤的过度采样，太多的假阳性和假阴性，以及太多的 8正常皮肤组织。可以帮助解决这个问题的是反射共焦显微镜（RCM）成像， 9例直接在患者身上非侵入性描绘边缘。RCM成像检测体内皮肤癌的灵敏度 10个为85-95%，特异性为80- 70%。2016年，医疗保险和医疗补助服务中心授予了reim- 皮肤RCM成像的11个bursement代码。RCM成像现在越来越多地用于非侵入性地 12引导诊断，使患者免于良性病变的活检。虽然20年的努力迄今为止， 由于成像引导诊断，新兴的应用是在成像中引导治疗。我们建议 14创建一种称为RCM视频马赛克的方法，以非侵入性地绘制大面积皮肤癌边缘 15例患者，增加了正常组织的采样、准确性和保留。父项目的具体目标 16是（1）开发一个实时和强大的RCM视频马赛克的方法，并纳入手持 17共聚焦显微镜用于床边，（2）测试该方法的图像质量和临床可接受性， 18和（3）对100例患者进行前瞻性测试，对LMM边缘和浅表NMSC进行视频镶嵌 19个边缘，然后根据术后病理学进行验证。 20.拟议的阿纳贝尔阿方索女士多样化项目 21.为阿纳贝尔阿方索女士提议的项目将以母项目为基础，并以完全不同的方式加以扩展。 22新的方向在母项目中，RCM视频马赛克将由Mohs外科医生在手术期间进行视觉读取。 23格里。阿纳贝尔项目的创新将是开发一种新的基于机器学习的算法 24，用于视频拼接中LMM边缘的自动检测和映射，以实现自动化，标准化和 提高莫氏外科医生在床边进行视频拼接的速度和效率。具体目标 阿纳贝尔的项目的26个目标是（1）开发一种图像分类和分割算法， 27分析了RCM视频马赛克，突出显示了LMM中具有诊断意义的区域（恶性与正常） 28个边缘，并为Mohs外科医生提供实时诊断反馈;（2）测试和验证算法， 29个RCM视频和视频马赛克从Leblon获得，涉及50名患者。获取视频和视频- 在父项目期间将出现30个马赛克。

1 FROM PARENT PROJECT R01CA240771 2 Superficially spreading types of skin cancers such as lentigo maligna melanomas (LMMs) and non-melanoma 3 skin cancers (NMSCs) occur mostly on older patients, with diffuse sub-clinical sub-surface spread over large 4 areas and with poorly defined margins that are difficult to detect. To treat these cancers, dermatologists rou- 5 tinely perform a large number of mapping biopsies to determine the spread and margins, followed by surgical 6 excision with wide "safety" margins. Not surprisingly, such a "blind" approach results in under-sampling of the 7 margins, over-sampling of normal skin, too many false positives and false negatives, and too much loss of 8 normal skin tissue. What may help address this problem is reflectance confocal microscopy (RCM) imaging to 9 noninvasively delineate margins, directly on patients. RCM imaging detects skin cancers in vivo with sensitivity 10 of 85-95% and specificity 80-70%. In 2016, the Centers for Medicare and Medicaid Services granted reim- 11 bursement codes for RCM imaging of skin. RCM imaging is now being increasingly used to noninvasively 12 guide diagnosis, sparing patients from biopsies of benign lesions. While the two-decade effort to date was fo- 13 cused on imaging-guided diagnosis, emerging applications are in imaging to guide therapy. We propose to 14 create an approach called RCM video-mosaicking, to noninvasively map skin cancer margins over large areas 15 on patients, with increased sampling, accuracy and sparing of normal tissue. The parent project specific aims 16 are (1) to develop a real-time and robust RCM video-mosaicking approach and incorporate into a handheld 17 confocal microscope for use at the bedside, (2) to test the approach for image quality and clinical acceptability, 18 and (3) to prospectively test on 100 patients, with video-mosaicking of LMM margins and superficial NMSC 19 margins, followed by validation against post-surgical pathology. 20 PROPOSED DIVERSITY SUPPLEMENT PROJECT FOR Ms. ANABEL ALFONSO 21 The proposed project for Ms. Anabel Alfonso will build upon the parent project and extend it in a completely 22 novel direction. In the parent project, RCM video-mosaics will be visually read by Mohs surgeons during sur- 23 gery. The innovation in Anabel’s project will be the development of a new machine learning-based algorithm 24 for automated detection and mapping of LMM margins in video-mosaics, toward automating, standardizing and 25 increasing the speed and efficiency of video-mosaicking for Mohs surgeons at the bedside. The specific aims 26 for Anabel’s project are to (1) develop an image classification and segmentation algorithm that automatically 27 analyzes RCM video-mosaics, highlights diagnostically significant areas (malignant versus normal) in LMM 28 margins and provides real-time diagnostic feedback to Mohs surgeons; (2) to test and validate the algorithm on 29 RCM videos and video-mosaics obtained from LMMs on 50 patients. Acquisition of the videos and video- 30 mosaics will occur during the parent project.