第一部分:目的：描绘成人组织细胞病患者的基因突变谱，探索基因突变与临床特征及预后的关系。方法：检索既往文献中与组织细胞病、血液系统肿瘤相关的基因，拟定出含183个基因的目标基因谱，用ER-seq建库测序技术及Illumina平台，对初治成人组织细胞病患者的肿瘤组织进行靶向测序。结果：共纳入144例成人组织细胞病患者，包括80例朗格汉斯细胞组织细胞增生症（Langerhans cell histiocytosis, LCH）、44例Erdheim-Chester病（Erdheim-Chester disease, ECD）、15例Rosai-Dorfman病、3例未分类树突细胞肿瘤、1例指状树突细胞肉瘤、1例进行性结节性组织细胞增生症。119（82.64%）例患者测出至少有1个致病突变，中位致病突变数为1（0-6）个。致病突变基因共39个，涉及MAPK通路（22个）、PI3K/AKT/mTOR通路（3个）、NFκB通路（4个）、转录因子及表观遗传学调控（3个）、细胞周期调控（4个）、其它（3个）。成人LCH中BRAFV600E突变27人（33.75％），与骨骼受累显著相关（P=0.005）；BRAFN486_P490del突变23人（占28.75％），与多系统受累（P=0.011）、甲状腺(P=0.002)、肝脏(P=0.001)、垂体（P=0.001）、肺(P=0.027)受累显著相关。ECD中，BRAFV600E突变27人（61.4%），与肺受累显著相关（P=0.031）；MAP2K1突变3人（6.8%）。Rosai-Dorfman病中，MAP2K1突变4例（26.7%），KRAS突变2例（13.3%）。未分类树突细胞肿瘤中，MAP2K1突变1例（33.3%），TP53突变1例（33.3%）。指状树突细胞肉瘤检出BRAFV600E突变；进行性结节性组织细胞增生症检出MAPK1突变。LCH中位随访13.9月，死亡1人，进展9人。ECD中位随访23.3月，死亡2人，进展12人，BRAF突变状态与LCH及ECD的无进展生存无显著关系。结论：大多数组织细胞病患者都可检出至少1个致病突变，突变主要涉及Ras/MAPK通路和PI3K/AKT/mTOR通路、NFκB通路、转录因子及表观遗传学调控、细胞周期调控。第二部分: 目的：研究LCH及ECD中用血浆cfDNA代替组织检出驱动致病突变的可行性。方法: 收集初治LCH和ECD患者的肿瘤组织、血浆cfDNA，用ER-seq建库测序技术及Illumina平台，对拟定的183个目标基因开展靶向测序。结果：研究纳入LCH 22例和ECD 15例。LCH中，血浆cfDNA相对于肿瘤组织检出BRAF/MAP2K1突变的灵敏度为66.7%，特异度为100%，阳性预测值为100%，阴性预测值为40%，两种方法的一致性为72.7%。ECD中，血浆cfDNA相对于肿瘤组织检出两种突变的灵敏度为92.3%，特异度为100%，阳性预测值为100%，阴性预测值为66.7%，两种方法的一致性为93.3%。结论：LCH患者中，血浆cfDNA相对于肿瘤组织检出BRAF/MAP2K1突变的特异性高，但灵敏度低。而ECD患者中的灵敏度和特异度都高。第三部分: 目的：研究成人LCH/ECD的细胞起源。方法：收集成人LCH和ECD患者的骨髓，磁珠分选骨髓CD34+细胞，用ER-seq建库测序技术及Illumina平台，对拟定的183个目标基因开展靶向测序。结果：LCH共6例，均为多系统受累，其中4例在肿瘤组织或血浆cfDNA中有BRAFN486_P490del突变，2例在肿瘤组织或血浆cfDNA中有MAP2K1E102_I103del突变，但6例均未在骨髓CD34+细胞中检出相应突变。ECD 1例，在其肿瘤组织和血浆cfDNA中均有BRAFV600E突变，组织突变频率2.5%，血浆cfDNA突变频率2.4%，骨髓CD34+细胞中也检出相应突变，突变频率0.33%。结论：ECD的肿瘤细胞可能起源于骨髓CD34+细胞。

Part I: Objective: To explore genetic landscape and its relation to clinical characteristics and prognosis of adult histiocytic neoplasm. Methods: A total of 183 genes related to histiocytosis and hematologic neoplasm were included in the panel. Targeted sequencing was conducted based on ER-seq and Illumina platform. ELISA and immunohistochemistry were performed to verify the function of candidate genes. Results: A total of 144 adult histiocytic neoplasm patients were enrolled, including 80 cases of Langerhans cell histiocytosis (LCH), 44 cases of Erdheim-Chester disease (ECD), 15 cases of Rosai-Dorfman disease, 3 cases of indeterminate dendritic cell tumor, one case of interdigitating dendritic cell sarcoma and one case of progressive nodular histiocytosis. Pathogenic mutations were discovered in 119 (82.64%) patients. Median number of mutations was 1 (0-6). Pathogenic mutations were discovered in 39 genes, involving MAPK pathway (22), PI3K/AKT/mTOR pathway (3), NFκB pathway (3), transcription factors and epigenetic modulation (3), cell cycle regulation (4) and others (3). In adult LCH, BRAFV600E mutation was identified in 27 (33.75%) patients, significantly related to bone involvement (P=0.005). BRAFN486_P490del mutation was discovered in 23 (28.75%) patients, which was significantly related to multi-system involvement (P=0.011), thyroid involvement (P=0.002), liver involvement (P=0.001), pituitary involvement (P=0.001), and lung involvement (P=0.027). In ECD, BRAFV600E mutation was identified in 27 (61.4%) patients, significantly related to lung involvement (P=0.031). MAP2K1 mutation was identified in 3 (6.8%) ECD patients. In Rosai-Dorfman disease, MAP2K1 mutation was discovered in 4 (26.7%) patients, while KRAS mutation was discovered in 2 (13.3%) patients. In indeterminate dendritic cell tumor, MAP2K1 mutation (33.3%) and TP53 mutation (33.3%) were discovered once. The interdigitating dendritic cell sarcoma patient carried a BRAFV600E mutation, while the progressive nodular histiocytosis carried a MAPK1 mutation. LCH patients were followed for 13.9 months. One patient died and progression occurred in 9 patients. There was no significant relation between BRAF mutation and progression-free survival. ECD patients were followed for 23.3 months. Two patient died and progression occurred in 12 patients. There was no significant relation between BRAF mutation and progression-free survival neither. Conclusion: Most histiocytic neoplasm patients carried at least one pathogenic mutation, which mainly involved Ras/MAPK pathway. PI3K/AKT/mTOR pathway, NFκB pathway, transcription factor and epigenetic modulation, and cell cycle regulation. Genetic landscape of adult LCH patients was different from that of pediatric LCH patients. Part II: Objective: To investigate the feasibility of utilizing plasma cfDNA to detect driver mutation of histiocytosis. Methods: Both Tumor tissue and baseline plasma cfDNA of LCH and ECD patients were collected. Targeted sequencing of the 183 genes was conducted based on ER-seq and Illumina platform. Results: Twenty-two LCH patients were recruited, including 13 patients with multi-system involved and 9 patients with single system involved. Fifteen ECD patients were also included. Compared with tumor tissue, sensitivity of detecting BRAF or MAP2K1 mutation in plasma cfDNA of LCH was 66.7%. Specificity was 100%. Positive predictive value was 100%. Negative predictive value was 40%. Agreement between tumor tissue and plasma cfDNA was 72.7%. In ECD, the sensitivity of plasma cfDNA detecting BRAF or MAP2K1 mutation was 92.3%. Specificity was 100%. Positive predictive value was 100%. Negative predictive value was 66.7%. Agreement between tumor tissue and plasma cfDNA was 93.3%. Conclusion: In LCH, specificity of plasma cfDNA was high, while sensitivity was limited. In ECD, both sensi. Part III: Objective: To explore the cell of origin in adult LCH and ECD. Methods: Collect bone marrow of adult LCH and ECD patients. CD34 magnetic beads were used to obtaine CD34+ cells. Targeted sequencing of the 183 genes was conducted based on ER-seq and Illumina platform. Results: Six multi-system LCH patients and 1 ECD patient were included. In LCH, 4 patients carried a BRAFN486_P490del mutation and 2 patients carried a MAP2K1E102_I103del mutation in tumor tissue of plasma cfDNA. But none of them had the corresponding mutation in bone marrow CD34+ cells. The ECD patient carried BRAFV600E mutation in both tumor tissue and plasma cfDNA. Variant allele frequency in tumor tissue was 2.5%, while that in plasma cfDNA was 2.4%. The mutation was also identified in his bone marrow CD34+ cells. Variant allele frequency was 0.33%. Conclusion: Tumor cells of ECD patients might arise from bone marrow CD34+ cells.