第一部分

II-IV期结直肠神经内分泌肿瘤的流行病学趋势及新型预后模型评估方法的探索

【目的】 本研究旨在阐明结直肠神经内分泌肿瘤（Neuroendocrine Neoplasms, NENs）全阶段的发病率趋势，分析II-IV期结直肠NENs患者的总生存期（Overall Survival, OS）和疾病特异性生存期（Disease-Specific Survival, DSS），并建立相关的列线图以进行风险分层。

【方法】从1975年至2019年，基于美国监测、流行病学和最终结果（The Surveillance, Epidemiology, and End Results, SEER）数据库中诊断为结直肠NENs的患者，评估其发病率的时间趋势。提取2010年至2016年间668例II-IV期结直肠NENs患者的临床数据进行生存分析。患者按7:3的比例随机分为训练队列和验证队列。通过单因素和多因素Cox回归分析确定影响OS的独立预后因素，并应用竞争风险分析探讨与DSS相关的风险因素。针对II-IV期结直肠NENs患者，分别构建了OS和DSS的列线图，并通过校准曲线、受试者工作特征（Receiver Operating Characteristic, ROC）曲线、时间依赖性曲线下面积（Area Under Curve , AUC）和决策曲线分析（Decision Curve Analysis, DCA）评估其预测能力。此外，使用我院62例结直肠NENs患者的独立队列作为外部验证队列。
    【结果】1975年至2019年间，结直肠NENs的发病率呈稳步上升趋势，年百分比变化（annual percentage change, APC）为4.50（95%置信区间[Confidence Interval, CI]：3.90–5.11，p<0.05）。2010年至2016年间，共纳入668例II-IV期结直肠NENs患者进行生存分析。治疗前的独立不良预后因素包括：低分化/未分化（OS的[Hazard Ratio,  HR]：4.66，95%CI：2.92-7.42；DSS的HR：4.79，95%CI：4.27-5.31）、更高的TNM分期（[III期 vs II期] OS的HR：2.22，95%CI：1.25-3.94；DSS的HR：2.69，95%CI：1.96-3.42。[IV期 vs II期] OS的HR：3.99，95%CI：2.03-7.83；DSS的HR：4.96，95%CI：4.14-5.78）、肝转移（OS的HR：1.61，95%CI：1.03-2.51；DSS的HR：1.86，95%CI：1.39-2.32）和脑转移（OS的HR：4.57，95%CI：1.66-12.58；DSS的HR：5.01，95%CI：4.15-5.87）。高龄也被确定为OS的风险因素（HR：2.03，95%CI：1.5-2.76），但对DSS无显著影响。在治疗方面，手术可显著延长OS（HR：0.62，95%CI：0.44-0.86）和DSS（HR：0.67，95%CI：0.29-1.05），但化疗和放疗未显示出显著效果。针对II-IV期结直肠NENs患者的OS和DSS列线图在训练队列、内部验证队列和外部验证队列中均表现出高准确性和强大的预测能力，能够预测1年、3年和5年的OS和DSS结果。此外，还建立了两个在线工具，用于OS和DSS预测，便于列线图评分计算、风险组确定以及个体患者的生存预测。

【结论】过去40年间，结直肠NENs的发病率呈稳步上升趋势，同时生存结局有所改善。更差的分化、更高的TNM分期、肝转移、脑转移以及未接受手术与较差的OS和DSS相关。高龄是OS的风险因素，但对DSS无显著影响。针对II-IV期结直肠NENs患者的列线图能够高精度预测1年、3年和5年的OS和DSS率，并实现风险分层。

第二部分

II-IV期结直肠神经内分泌肿瘤的基因特征及KMT2D突变的预后与治疗意义

【目的】结直肠NENs的遗传特征与转移机制尚未完全阐明，目前针对局部晚期及晚期患者的治疗选择有限。本研究旨在揭示II-IV期结直肠NENs的基因组异质性、潜在靶向突变及异常信号通路，并探讨其临床意义。

【方法】纳入46例II-IV期结直肠NENs患者，通过靶向二代测序分析体细胞突变谱，结合京都基因与基因组百科全书（Kyoto Encyclopedia of Genes and Genomes, KEGG）数据库通路富集、生存分析及精准医疗肿瘤数据库（Oncology Knowledge Base , OncoKB）注释工具，解析分期特异性分子特征。再利用基因表达综合数据库（Gene Expression Omnibus, GEO）数据库验证KMT2D基因的预后价值。

【结果】II期患者以NCOR2（17.4%）、PTPRT（17.4%）和ARID1A（13.1%）突变为特征，III期以EIF1AX（18.8%）和KMT2D（12.5%）为主，IV期则以CDKN1B（42.9%）、BRD4（28.6%）和NF1（28.6%）高频突变为典型。突变特征分析显示，C>T碱基转换（COSMIC Signature 1）及氧化应激相关特征（Signature17b）显著富集。KEGG通路分析揭示分期依赖性异常通路：II期以PI3K-AKT/MAPK通路为主，III期与细胞衰老/EGFR耐药相关，IV期涉及Polycomb抑制复合体及Fanconi贫血通路。50%患者携带潜在可靶向突变（如ARID1A、NF1、FGFR3），III/IV期比例更高。生存分析表明，KMT2D突变是独立预后危险因素（HR=12.87，p=0.012），其低表达与总生存期缩短显著相关（p <0.05）。

【结论】本研究首次系统阐明结直肠NENs的分期特异性基因组特征，为精准治疗提供潜在靶点（如TRK/mTOR抑制剂）。KMT2D突变可作为预后标志物，其功能缺失可能通过激活mTOR通路及免疫微环境重塑驱动肿瘤进展，为个体化治疗策略奠定理论基础。

Part 1. Epidemiological trends and novel prognostic evaluation approaches of patients with stage II-IV colorectal neuroendocrine neoplasms: a population-based study with external validation

Objective: This study aimed to clarify the incidence trend of all-stage colorectal neuroendocrine neoplasms (CRNENs), overall survival (OS), and disease-specific survival (DSS) of patients with stage II-IV CRNENs, and to establish relevant nomograms for risk stratification.

Methods: Among all patients diagnosed with CRNENs in the Surveillance, Epidemiology, and End Results (SEER) database from 1975 to 2019, temporal trends in incidence were assessed. Clinical data of 668 patients with stage II-IV CRNENs from 2010 to 2016 were extracted for survival analysis. Patients were randomly divided into a training cohort and a validation cohort at a ratio of 7:3. Univariate and multivariate cox regression analyses were utilized to identify independent prognostic factors affecting OS outcomes. Competing risk analysis was applied to investigate risk factors related to the DSS of CRNENs. Two nomograms specifically for OS and DSS were developed for patients with stage II-IV CRNENs, their prognostic capabilities were evaluated using calibration curves, receiver operating characteristic (ROC) curves, the time-dependent area under the curve (AUC), and decision-curve analysis (DCA). Our hospital's independent cohort of 62 patients with CRNENs was used as the external validation cohort.

Results:  In the period of 1975-2019, the incidence of CRNENs increased steadily with an annual percentage change (APC) of 4.50 (95% confidence interval [CI]: 3.90–5.11, P < 0.05). In total, 668 patients with stage II-IV CRNENs were included in the survival analysis from 2010 and 2016. Independent adverse prognostic factors for both OS and DSS of CRNENs prior treatment included poorly differentiated/undifferentiated (HR for OS: 4.66, 95%CI: 2.92-7.42; HR for DSS: 4.79, 95%CI: 4.27-5.31), higher TNM stage ([stage III vs stage II] HR for OS: 2.22, 95%CI: 1.25-3.94; HR for DSS: 2.69, 95%CI: 1.96-3.42. [stage IV vs stage II] HR for OS: 3.99, 95%CI: 2.03-7.83; HR for DSS: 4.96, 95%CI: 4.14-5.78), liver metastasis (HR for OS: 1.61, 95%CI: 1.03-2.51; HR for DSS: 1.86, 95%CI: 1.39-2.32), and brain metastasis (HR for OS: 4.57, 95%CI: 1.66-12.58; HR for DSS: 5.01, 95%CI: 4.15-5.87). Advanced age was also identified as a risk factor for OS (HR: 2.03, 95%CI: 1.5-2.76) but not DSS. In terms of treatment, surgery can significantly prolong OS (HR: 0.62, 95%CI: 0.44-0.86) and DSS (HR: 0.67, 95%CI: 0.29-1.05), but chemotherapy and radiation failed to show significance. The respective nomograms for OS and DSS for stage II-IV CRNENs demonstrated high accuracy and robust prediction value in predicting 1-year, 3-year, and 5-year OS and DSS outcomes in training, internal validation, and external validation cohorts. Besides, two online tools regarding OS and DSS prediction were established, facilitating nomogram score calculation, risk group determination, as well as survival prediction for each individual patient.

Conclusion: Over the past 40 years, the incidence of CRNENs presented increased steadily, along with improved survival outcomes. poorly differentiated/undifferentiated, higher TNM stage, liver metastasis, brain metastasis, and without receiving surgery were found to be associated with worse OS and DSS. Advanced age was a risk factor for OS but not DSS. Nomograms for patients with stage II-IV stage CRNENs are capable of predicting the 1-, 3-, and 5-year OS and DSS rates with high accuracy, and realize risk stratification.

Part 2. Stage-Specific Genomic Landscape and Prognostic/Therapeutic Implications of KMT2D Mutations in Colorectal Neuroendocrine Neoplasms (Stages II-IV)

Objective: The genomic characteristics and metastatic mechanisms of colorectal neuroendocrine neoplasms (NENs) remain poorly understood, limiting therapeutic options for advanced-stage patients. This study aimed to elucidate the stage-specific genomic heterogeneity, actionable mutations, and dysregulated pathways in stages II-IV colorectal NENs.

Methods: We enrolled 46 patients with stage II-IV colorectal NENs. Targeted next-generation sequencing (808-gene panel) was performed to analyze somatic mutations. KEGG pathway enrichment, survival analysis, and OncoKB annotation were applied to decipher molecular features. Prognostic validation of KMT2D was conducted using GEO datasets.

Results: Stage-specific mutation profiles were identified: NCOR2 (17.4%), PTPRT (17.4%), and ARID1A (13.1%) dominated in stage II; EIF1AX (18.8%) and KMT2D (12.5%) in stage III; CDKN1B (42.9%), BRD4 (28.6%), and NF1(28.6%) in stage IV. Mutational signature analysis revealed enrichment of C>T transitions in stage III, and Polycomb/Fanconi anemia pathways in stage IV. Actionable mutations (e.g., ARID1A, NF1, FGFR3) were detected in 50% of patients, with higher frequencies in stages III/IV. Survival analysis demonstrated that KMT2D mutation was an independent prognostic risk factor (HR=12.87, p=0.012), and its low expression correlated with shorter overall survival (p<0.05).

Conclusion: This study systematically delineates the stage-specific genomic landscape of colorectal NENs, identifying potential therapeutic targets (e.g., TRK/mTOR inhibitors). KMT2D mutation serves as a prognostic marker, and its loss-of-function may drive progression via mTOR activation and immune microenvironment remodeling, providing a foundation for personalized treatment strategies.