The role of circulating tumor DNA from cerebrospinal fluid as a minimal-invasive biomarker for comprehensive genetic profiling and improved outcome prediction in patients with CNS lymphoma

原发性中枢神经系统淋巴瘤（PCNSL）是一种异质性脑癌类型。PCNSL临床结局的遗传因素知之甚少。由于样本量小或组织靶向不准确，从侵入性立体定向活检或脑手术获得的肿瘤材料的基因分型通常是不够的。此外，这些手术在一部分患者中引起术中和术后并发症，包括颅内出血和严重感染。循环肿瘤DNA（ctDNA）是包括淋巴瘤在内的肿瘤学中新兴的生物标志物。我们已经成功地将靶向捕获下一代测序（NGS）方法（CAPP-Seq，通过深度测序进行癌症个性化分析）转移到我们的机构，该方法可以对淋巴瘤患者的ctDNA进行超灵敏和全面的分析。我们优化了CAPP-Seq用于PCNSL，并证明ctDNA在血浆和脑脊液（CSF）中易于检测。此外，来自CSF的ctDNA似乎准确地反映了PCNSL肿瘤的突变景观和遗传组成（n=4）。在拟定的研究中，我们的目标是利用来自前瞻性多中心试验（n=84，DRKS 00005503）的可用诊断CSF，将CSF ctDNA确立为用于全面肿瘤基因分型、评估肿瘤负荷、表征PCNSL突变景观和预测PCNSL结局的无活检生物标志物。我们将进一步利用来自CSF ctDNA基因分型的信息以及传统的临床和放射学风险因素，开发一种新的综合风险模型，该模型可以准确和改善单因素传统模型的结果预测。如果成功的话，我们设想CSF测序作为一种微创方法来全面评估PCNSL基因型，而不需要侵入性手术或足够量的肿瘤DNA。此外，一个改进的和个性化的综合算法的结果预测可能会显着提高PCNSL患者的临床管理，并帮助指导未来的治疗。

Primary central nervous system lymphoma (PCNSL) represents a heterogeneous brain cancer type. Genetic factors underlying PCNSL clinical outcomes are poorly understood. Genotyping from tumor material achieved from invasive stereotactic biopsies or brain surgery is often insufficient due to small sample size or inaccurate tissue targeting. Moreover, these procedures cause intra- and postsurgical complications in a subset of patients, including intracranial hemorrhage and severe infections. Circulating tumor DNA (ctDNA) is an emerging biomarker across oncology, including lymphomas. We have successfully transferred a targeted capture next-generation sequencing (NGS) approach (CAPP-Seq, Cancer Personalized Profiling by Deep Sequencing) to our institution that allows ultrasensitive and comprehensive profiling of ctDNA in lymphoma patients. We optimized CAPP-Seq for its use in PCNSL and demonstrated that ctDNA is readily detectable in blood plasma and cerebrospinal fluid (CSF). Moreover, ctDNA from CSF accurately seems to mirror the mutational landscape and genetic composition of PCNSL tumors (n=4). In the proposed study, we aim to utilize available diagnostic CSF from a prospective multi-center trial (n=84, DRKS00005503) to establish CSF ctDNA as a biopsy-free biomarker for comprehensive tumor genotyping, assessment of tumor burden, characterization of PCNSL mutation landscapes, and prediction of PCNSL outcomes. We will further utilize information from CSF ctDNA genotyping together with conventional clinical and radiographic risk factors to develop a novel integrative risk models that allows accurate and improved outcome prediction over single-factor traditional models. If successfull, we envision a role of CSF sequencing as a minimal-invasive way to comprehensively assess PCNSL genotypes without the need for invasive surgical procedures or sufficient amounts of tumor DNA. Moreover, an improved and personalized integrative algorithm for outcome prediction might significantly enhance clinical management of patients with PCNSL with and help guide therapies in the future.