Innovative analytic methods in the imaging of upper tract urothelial carcinoma: Radiomics, Radiogenomics and creation of a new prognostic tool for systemic therapies of UTUC

上尿路上皮癌(UTUC)是一种罕见的恶性肿瘤，起源于输尿管和肾盂的尿路。微小和浅表的病变可以通过微创的保留肾脏手术来治疗。在诊断时，大约60%的UTUC是侵袭性的，需要根治性手术。对于局部晚期和转移性疾病，建议采用系统治疗。目前使用CT扫描、输尿管镜结合细胞学和活检进行诊断的研究报告准确率高达74%。尽管诊断工具有了很大的进步，但UTUC的分期在分级和侵袭性方面仍然需要改进，这对启动适当的治疗至关重要。放射组学是一种很有前途的基于成像的技术，有可能填补术前检查的空白。放射组学从标准成像中提取数据，并将其转换为临床有用的信息。与传统的成像放射组学分析纹理场不同，基于单个数据点的扩展形状建模称为体素。这些信息是基于直方图的统计数据，并遵循反映肿瘤特征的假设。放射组学在预测肿瘤的侵袭性、分级和突变负担等方面具有巨大的潜力。因此，放射组学可能会提供最准确的分期，预测系统治疗的反应，并评估进展的风险。我的研究员的目标是将标准的放射信息从分期图像转换为更广泛有用的临床参数。纪念斯隆-凯特琳癌症中心(MSKCC)的泌尿科拥有一个庞大的UTUC患者数据库，包括临床和基因组信息以及相应的成像。MSKCC的泌尿学和放射学合作专门研究了泌尿生殖系统恶性肿瘤的放射组学。在第一步中，我试图开发一种利用放射组学对UTUC进行风险分层的方法。因此，将提取图像的特征来识别侵袭性和浸润性UTUC。了解肿瘤的分级和浸润性对于提供保肾或根治性手术是至关重要的。其次，成像特征将与基因组改变相关联，以确定UTUC的基因组足迹。各种基因改变已被发现与肿瘤分级、侵袭性和进展风险相关。因此，我计划开发一个诺模图，利用放射基因组学来识别低风险和高风险患者。在第三步中，我将专注于基于成像的对系统治疗反应率的预测。因此，成像特征将被用来制定风险评分，以预测治疗反应，包括相关的临床信息和评估的基因突变。根据风险评分，治疗反应的概率可能是可预测的。根据目前的文献和初步的初步结果，放射组学在UTUC的临床实践中引入影像适应医学具有很高的潜力。

Upper tract urothelial carcinoma (UTUC) are rare malignant tumors arising from the urothel of the ureter and renal pelvis. Small and superficial lesions can be treated by minimal-invasive kidney sparing surgery. At the time of diagnosis about 60% of UTUC are invasive and require radical surgery. In locally advanced and metastatic disease systemic therapy is recommended. Current diagnostic work-up using CT scan, ureteroscopy with cytology and biopsy is reported with an accuracy of up to 74%. Despite significant advances in diagnostic tools, staging of UTUC is still in need of improvement in regards of grading and invasiveness which is crucial to initiate adequate treatment. Radiomics is a promising imaging-based technique with potential to fill the gap in preoperative work-up. Radiomics extracts data from standard imaging and converts them to clinical useful information. In contrast to conventional analysis of imaging radiomics analyses texture fields, extended shape modeling based on single data points, called voxels. These information are histogram-based statistics and follow the hypothesis to mirror tumor characteristics. Radiomics holds immense potential to predict tumor characteristics such as invasiveness, grading and mutational burden. Consequently, radiomics might allow most accurate staging, prognosticate response to systemic therapy and evaluate risk of progression.The objective of my fellowship is to convert standard radiological information from staging images into more widely useful clinical parameters. The Department of Urology at Memorial Sloan-Kettering Cancer Center (MSKCC) holds a vast database of UTUC patients including clinical and genomic information and corresponding imaging. The collaboration of Urology and Radiology at MSKCC has specialized on radiomics in genitourinary malignanciCes. In the first step I attempt to develop a method of risk stratifying UTUC by using radiomics. Therefore features of the images will be extracted to identify aggressive and infiltrative UTUC. The knowledge of tumor grading and infiltration are crucial to offer either kidney sparing or radical surgery. Secondly, imaging features will be correlated with genomic alterations to identify a genomic footprint of UTUC. Various genetic alterations have been identified to correlate with tumor grading, invasiveness and risk of progression. Therefore, I plan to develop a nomogram to identify low-risk and high-risk patients using radiogenomics. In the third step I will be focusing on imaging-based prediction of response rates to systemic therapy. Therefore, imaging features will be used to develop a risk score to predict treatment response, including relevant clinical information and genetic mutations assessed. According to the risk score probability of treatment response might be predictable. In line with the current literature and first preliminary results radiomics holds high potential to introduce imaging-adapted medicine in clinical practice of UTUC.