Simultaneous coaxial widefield imaging and reflectance confocal microscopy for improved diagnosis of skin cancers in vivo

皮肤科医生依靠视觉(临床广泛领域)和皮肤镜检查皮损来指导需求 活检2例。使用这种方法，敏感性很高，但特异性往往非常可变和较低，结果 每年数以百万计的良性病变活检中有3例。为了提高特异性，有几种光学技术 4正在被开发用于非侵入性检测皮肤癌。其中，反射共焦显微镜(RCM)是 5临床应用最先进，对皮肤癌的诊断具有高度的敏感性和特异性。 6在皮肤镜引导下的RCM成像，对皮肤癌的检测具有2倍的特异度，并降低了皮肤癌的发病率 7活检良恶性比单纯皮肤镜检查高2倍。2016年，该中心 8 Medicare和Medicaid服务获得RCM的当前程序术语(CPT)报销代码 9皮肤成像。RCM成像结合皮肤镜检查现已进入临床应用，节省了患者的时间。 10例来自不必要的良性病变活检。然而，为了得到广泛的接受和采用， 11关键挑战是临床大范围检查、皮肤镜检查和RCM成像目前的执行方式 12使用不同设备的三种不同程序。临床医生不能准确地知道RCM影像的位置。 13ES与临床广泛视野检查所见的周围周围病变的形态相关 14皮肤镜检查，导致诊断准确率较低且变化较大(尤其是敏感性、阳性和 15个阴性预测值)。我们提出了一种新的解决方案：(I)一种新的物镜，它具有集成的微透镜和 16个摄像头，以提供RCM成像位置周围皮肤表面的并发广域图像； 17(2)一种新的软件算法，用于在皮肤镜内基于宽视场图像跟踪RCM图像的位置 18视野；(3)一种新的诊断方法，将主动使用广域成像来定位RCM图像 19张皮肤镜图像。我们打算提供这一集成的广泛领域的临床、皮肤镜和RCM成像AP- 20进入临床，迈向更准确、一致和更快的RCM成像的新标准，以指导 21例病人护理。台式“模拟”物镜和微型相机的初步研究 22展示了用于RCM成像的出色光学切片(~2微米)和分辨率(~1微米)，以及准确和 23宽视场图像内RCM视场的可重复位置。RCM图像显示极好的细胞性 24和活体内的形态细节。我们的具体目标是：(1)研制一种手持反射式共焦微控制器。 25范围，集成广域摄像头；(2)开发实时广域成像的图像处理算法- 26老化引导跟踪皮肤镜视野内的RCM图像位置；(3)测试和验证性能 100名患者中有27名。尽管我们的主张是针对皮肤病的，但这项研究肯定会对 28在其他环境中成像，特别是使用微型共焦显微镜和内窥镜，它们具有 29个非常小的视场。我们是来自蒙大拿州立大学斯隆纪念分校的高度协作的团队 30东北大学凯特琳癌症中心和口径成像与诊断公司(以前为Lucid Inc.)。

1 Dermatologists rely on visual (clinical widefield) and dermoscopic examination of skin lesions to guide the need 2 for biopsy. With this approach, sensitivity is high, but specificity tends to be quite variable and lower, resulting 3 in millions of biopsies of benign lesions every year. To improve specificity, several optical technologies are 4 being developed to noninvasively detect skin cancer. Of these, reflectance confocal microscopy (RCM) is the 5 furthest advanced in clinical utility, proven for diagnosing skin cancers with high sensitivity and specificity. 6 RCM imaging, guided by dermoscopy, detects skin cancers with 2 times superior specificity, and reduces the 7 benign-to-malignant biopsy rate by 2 times, compared to that with dermoscopy alone. In 2016, the Centers for 8 Medicare and Medicaid Services granted current procedural terminology (CPT) reimbursement codes for RCM 9 imaging of skin. RCM imaging combined with dermoscopy is now advancing into clinical practice, sparing pa- 10 tients from unnecessary biopsies of benign lesions. However, toward widespread acceptance and adoption, a 11 key challenge is that clinical widefield examination, dermoscopy and RCM imaging are currently performed as 12 three separate procedures with separate devices. Clinicians do not precisely know the location of RCM imag- 13 es relative to the surrounding contextual lesion morphology that is seen with clinical widefield examination and 14 dermoscopy, resulting in lower and more variable diagnostic accuracy (particularly, sensitivity, positive and 15 negative predictive values). We propose a novel solution: (i) a new objective lens with an integrated micro- 16 camera, to deliver a concurrent widefield image of the skin surface surrounding the location of RCM imaging; 17 (ii) a new software algorithm for widefield image-based tracking of the location of RCM images within a dermoscopic 18 field of view; (iii) a new diagnostic approach that will proactively use widefield imaging to locate RCM images in 19 dermoscopic images. We intend to deliver this integrated widefield clinical, dermoscopic and RCM imaging ap- 20 proach into the clinic, toward a new standard for more accurate, consistent and faster RCM imaging to guide 21 patient care. Preliminary studies with a “mock” objective lens and micro-camera on a bench-top set-up 22 demonstrated excellent optical sectioning (~2 µm) and resolution (~1 µm) for RCM imaging, and accurate and 23 repeatable location of RCM fields-of-view within the widefield image. RCM images showed excellent cellular 24 and morphologic detail in vivo. Our specific aims are (1) to develop a handheld reflectance confocal micro- 25 scope with integrated widefield camera; (2) to develop image processing algorithms for real-time widefield im- 26 aging-guided tracking of RCM image locations within dermoscopic fields; (3) to test and validate performance 27 on 100 patients. Although our proposition is for skin lesions, the research will surely have wider impact for 28 imaging in other settings, particularly, with miniaturized confocal microscopes and endoscopes, which have 29 very small fields-of-view. We are a highly synergistic team from Montana State University, Memorial Sloan 30 Kettering Cancer Center, Northeastern University and Caliber Imaging and Diagnostics (formerly, Lucid Inc.).