Cutaneous neoplasias in epidermolysis bullosa: molecular characterization and application of augmented intelligence

大疱性表皮松解症（EB）是一种罕见而严重的皮肤病。皮肤很脆弱，轻微的伤害会导致水泡和伤口，患者经常遭受疼痛和感染。作为一个严重的并发症，EB患者会发展为皮肤癌，尤其是鳞状细胞癌（SCC）。EB中的SCCs发生在年轻时，具有侵袭性，因此早期发现是关键。然而，识别这种癌症是困难的。在这个项目中，我们将使用人工智能（AI）训练计算机识别EB皮肤上的SCC。这种方法最近在皮肤科被证明是成功的，经过数千张照片的训练，人工智能在识别“正常”皮肤上的某些皮肤癌方面与皮肤科医生不相上下。然而，EB患者队列和国家中心的图像收集很少。为了解决这个问题，该项目联合了几个国际参考中心，收集了大量EB SCCs和EB皮肤的照片，用于计算机培训。其目标是为医生和患者创建一个智能手机应用程序。将人工智能应用于罕见皮肤病是改善诊断、治疗和患者生活质量的重要一步。EB的另一个特征是EB痣，但人们对其皮肤发现知之甚少。这些大色斑出现在高达14%的EB患者中，生长迅速，并表现出黑色素瘤的某些特征，黑色素瘤是人类最严重的皮肤癌。虽然EB痣很常见，但患者和医生都很担心。关于EB nevi的知识很少，而且仅仅是来自小的案例研究。一些患者为何以及如何患上EB痣尚不清楚。到目前为止，还没有EB痣转化为黑色素瘤，这可能只是反映了患者寿命的缩短。EB痣是否是真正的痣也不清楚，它们携带BRAF或NRAS基因的某些遗传变化；我们认为它们是，这意味着它们可能会演变成癌症。在这个项目中，我们将加深我们对EB痣的临床表现、行为（即大小、位置、进展）和原因的了解。为此，我们将加入几个国际参考中心的患者队列，收集EB nevi的照片、临床数据和组织样本。从组织样本的切片中，色素细胞（黑素细胞）和在EB中发生故障的皮肤结构的特定标记物将被染色和评估。分离DNA，进行BRAF和NRAS基因的遗传检查。然后，我们将在实验室中通过培养黑色素细胞和EB角化细胞并评估它们的行为和相互作用来模拟EB痣。由此，我们将更好地了解EB nevi是如何发展的，他们走的是什么路线。总之，这两个项目部分将填补EB知识的空白，并可以作为罕见病大领域的范例。最重要的是，新发现将直接使EB患者受益。

Epidermolysis bullosa (EB) is a rare and serious skin condition. The skin is fragile, minor injuries lead to blisters and wounds, and patients suffer from frequent pain and infections. As a severe complication, patients with EB develop skin cancer, particularly squamous cell carcinoma (SCC). SCCs in EB occur at a young age and are aggressive, so early discovery is key. However, recognizing this cancer is difficult. In this project, we will train a computer using artificial intelligence (AI) to recognize SCC on EB skin. This method has recently proven successful in dermatology, and AI was equal to dermatologists in recognizing certain skin cancers on “normal” skin after being trained on thousands of photographs. However, EB patient cohorts and thus image collections in national centers are small. To solve this, this project joins several international reference centers to amass a large number of photographs of EB SCCs and EB skin for computer training. The goal is to create a smartphone application for both physicians and patients. Applying AI in rare skin disease is a new and important step to improve diagnostics, therapy, and patients’ quality of life.Another characteristic, but poorly understood skin finding in EB, is the EB nevus. These large pigmented spots occur in up to 14% of EB patients, grow rapidly, and show certain characteristics of melanoma, the most severe skin cancer in humans. Although EB nevi are common, they are worrisome to patients and physicians. Knowledge about EB nevi is sparse and results solely from small case studies. Why and how some patients develop EB nevi is unclear. To date, no EB nevus has transformed to melanoma, which may simply reflect the reduced lifespan of patients. It is also unknown whether EB nevi are in fact true nevi, which carry certain genetic changes in BRAF or NRAS genes; we think they are, which means they could turn into cancer.In this project, we will deepen our understanding of the clinical appearance, behavior (i.e., size, location, progression), and cause of EB nevi. For this, we will join patient cohorts of several international reference centers and collect photographs, clinical data, and tissue samples of EB nevi. From sections of tissue samples, specific markers of the pigmentary cells (melanocytes) and of the structures holding skin together that malfunction in EB will be stained and assessed. DNA will be isolated and genetic examination regarding BRAF and NRAS genes will be performed. We will then model the EB nevus in the laboratory by cultivating melanocytes and EB keratinocytes together and assessing their behavior and interaction. From this, we will better understand how EB nevi develop and what course they take.Together, both project parts will fill gaps in EB knowledge and can serve as examples for the large field of rare diseases. Most importantly, the new findings will directly benefit patients with EB.