第一部分  小细胞肺癌临床病理特征及预后分析

目的：本研究的目的是分析小细胞肺癌（SCLC, Small Cell Lung Cancer）患者的术后临床病理特征，探究影响无病生存期（DFS, Disease-Free Survival）和总生存期（OS, Overall Survival）的相关预后因素。

方法：调阅2005年7月至2016年4月期间SCLC手术患者资料共530例，经复阅临床病理及预后相关信息并排除不符合入组标准的病例，最终纳入247例单纯型SCLC手术病例进行临床病理预后分析。收集性别、年龄、吸烟史、原发肿瘤部位、原发肿瘤（T分期）、淋巴结转移（N分期）等临床相关因素，以及病理相关因素，包括胸膜侵犯、支气管侵犯、神经侵犯、气腔播散（STAS, Spread Through Air Spaces）、脉管瘤栓、癌细胞形态（圆形vs.梭形）、坏死、间质纤维化和肿瘤浸润淋巴细胞（TILs, Tumor Infiltrates Lymphocytes）；同时也回顾了神经内分泌标志物（CD56、Syn、ChrA）免疫组化染色结果。采用门诊病例记录及电话方式随诊患者的复发、远处转移和生存情况。生存分析采用Kaplan-Meier曲线和log-rank检验。

结果：247例患者中位DFS为98个月，1年、3年、5年DFS率分别为70.9%、54.4%、52.2%；中位OS未达到，1年、3年和5年生存率分别为94.2%、72.3%和65.4%。单因素分析显示影响患者DFS的临床病理特征包括性别、吸烟史、原发肿瘤T分期、区域淋巴结转移、细胞形态和TILs；影响OS的特征包括年龄、原发肿瘤T分期、区域淋巴结转移、远处转移、神经侵犯、细胞形态和TILs。多因素分析显示吸烟史、区域淋巴结转移、细胞形态是DFS的独立影响因素；年龄、原发肿瘤T分期、远处转移（脑、肝、骨）、神经侵犯、TILs是OS的独立影响因素。

结论：SCLC术后患者的预后相关因素包括分期、年龄、吸烟、神经侵犯、细胞形态以及TILs数量等，而神经内分泌标志物的表达与生存率无显著相关性。尤其值得注意的是，细胞形态为梭形和TILs>30%是独立预后不良因素，提示病理微环境特征影响预后，值得进一步探究相关分子机制。  

关键词：小细胞肺癌，局限期，预后，生存分析

第二部分  小细胞肺癌病理异质性分层评价体系探索

目的：SCLC临床进展缓慢，缺乏分层治疗策略。鉴于近年来基于细胞系和动物实验的基础研究提示SCLC在细胞形态、神经内分泌（NE，Neuroendocrine）分化及谱系转录因子表达等方面存在明显的异质性。本研究拟从肿瘤异质性角度探究适用于临床的病理分层评价方法，重点关注从mRNA水平异质性到蛋白质水平异质性探索，预期借助蛋白质免疫组化检测提示分子异质性特征，从而为临床研发异质性相关分层治疗策略提供部分依据。

方法：本研究纳入247例单纯型SCLC术后样本，基于病理学形态观察分为经典型（Classic）和变异型（Variant）两组，选取其中48例用Nanostring nCounter技术定量NE基因及谱系转录因子（ASCL1、NEUROD1、POU2F3、YAP1）的mRNA表达，同步采用蛋白质免疫组化检测上述相关因子，经Spearman相关分析进行mRNA水平及蛋白质水平一致性检验。进一步在247例样本中通过组织芯片（TMAs, Tissue Micriarray）联合免疫组化检测方法验证肿瘤组织形态、NE分化及谱系转录因子分型与临床病理特征及预后的关联。

结果：（1）形态学方面，Classic组和Variant组分别占81.8%（202/247）vs. 18.2%（45/247），Classic组肿瘤纤维化程度较低（p=0.021），INSM1表达较高（p=0.03）。Classic组有预后较好趋势，尽管差异尚未达显著统计学意义（p＞0.05）。（2）在mRNA水平，NE亚型分布如下：NE-high（83.3%，40/48），NE-low（16.7%，8/48），NE-high亚型与AJCC进展期（p=0.014）、Classic细胞形态（p=0.047）及INSM1高表达（p=0.017）显著相关；预后无显著统计学差异（p＞0.05），但NE-low亚型有预后较好趋势；谱系转录因子亚型方面，SCLC-A占比45.8%（22/48），SCLC-N为22.9%（11/48），SCLC-P为8.3%（4/48），SCLC-Y为22.9%（11/48），SCLC-A亚型患者显著集中于≤65岁年龄组（p=0.016）及Syn阳性表达组（p=0.08），各亚型预后无显著统计学差异（p＞0.05）。（3）在蛋白质水平，NE(INSM1)-high占比73.7%（182/247），NE(INSM1)-low占比26.3%（65/247），NE(INSM1)-high在形态上倾向于Classic，尚未达到显著统计学意义（p=0.062），NE(INSM1)-high患者支气管侵犯（p=0.043）、脉管瘤栓（p=0.002）、传统NE标记物（CD56、ChrA、Syn）阳性表达显著高于NE(INSM1)-low组（p＜0.05），NE(INSM1)-low亚型有预后稍好趋势，但两组间差异不显著（p＞0.05）。在蛋白质水平，谱系转录因子亚型分布如下：SCLC-A（38.9%, 96/247），SCLC-N（46.6%, 115/247），SCLC-P（9.3%, 23/247），SCLC-Y（5.3%, 13/247）；年龄≤65岁（p=0.019）、Classic形态（p=0.009）、NE标记物（INSM1、CD56、ChrA、Syn）阳性表达（p＜0.05）显著集中于SCLC-A及SCLC-N亚型，四组间预后无显著统计学差异（p＞0.05）。谱系转录因子蛋白质层面分子分型存在92.7%（229/247）不同程度共表达情况，其中双表达64例（25.9%），三表达129例（52.2%），四表达36例（14.6%），共表达H-score差值≤30的病例占比24.7%（61/247）。另外，个别病例还表现出异质性表达特征。

结论：通过48例手术样本mRNA和蛋白质定量检测验证了转录组测序为主的分子异质性特征可在蛋白质水平以免疫组化检测表征：INSM1是较好的NE分化标志物，而ASCL1、NEUROD1、POU2F3、YAP1在蛋白质水平可以区分分子分型，但它们存在共表达及异质性表达情况，提示单纯依靠谱系转录因子分型存在一定困难。细胞形态为Classic者倾向于NE-high及SCLC-A/-N亚型，而细胞形态为Variant者倾向于NE-low及SCLC-P/-Y亚型，根据“细胞形态-NE分化-谱系转录因子表达”的关联现象，构建以这几个层次为基础的病理异质性分层评价体系，或将有助于为临床研发分层治疗策略提供科学依据。

关键词：小细胞肺癌，形态学，分子分型，病理异质性，分层评价体系

Part Ⅰ   Clinicopathological features and prognosis of small cell lung cancer

Objective: The objective of this study was to analyze the clinical and pathological characteristics of patients with small cell lung cancer (SCLC) after curative surgery and to explore prognostic factors for disease-free survival (DFS) and overall survival (OS).

Methods: 530 patients undergoing surgery for SCLC from July 2005 to April 2016 were reviewed. After reviewing clinicopathological and prognostic information and excluding cases that did not meet the inclusion criteria, 247 patients with pure SCLC were finally included for clinicopathological prognostic analysis. Clinicopathological features were retrieved, including gender, age, smoking history, tumor location, and distant metastasis. Histopathological features were also reviewed by three pathologists, including primary tumor (T), lymph node metastasis (N), pleural invasion, bronchial invasion, nerve invasion, spread through air spaces (STAS), tumor thrombosis, major cell shape (round vs. spindle), tumor necrosis, stromal fibrosis, and tumor-infiltrating lymphocytes (TILs). Immunohistochemical staining of neuroendocrine markers (CD56, Syn, ChrA) was also reviewed. Recurrence, distant metastasis and survival were followed up by outpatient case records and telephone. Kaplan-Meier curves and log-rank tests were applied for survival analysis.

Results: The median DFS was 98 months, and the 1-year, 3-year, and 5-year DFS rates were 70.9%, 54.4%, and 52.2%, respectively. The median OS was not reached, and the 1-year, 3-year, and 5-year survival rates were 94.2%, 72.3%, and 65.4%, respectively. Univariate analysis revealed clinicopathological features with DFS (gender, smoking history, primary tumor, regional lymph node metastasis, major cell shape, and TILs) and OS (age, primary tumor, regional lymph node metastasis, distant metastasis, nerve invasion, major cell shape, and TILs). Multivariate analysis revealed DFS-related factors (smoking history, regional lymph node metastasis and major cell shape) and OS-related factors (age, primary tumor, distant metastasis in the brain, liver, bone, nerve invasion, and TILs).  

Conclusion: Postoperative prognostic factors of SCLC patients include stage, age, smoking, nerve invasion, cell morphology and TILs number. Neuroendocrine immunostaining markers showed no correlation with survival. Of interest, spindle cell type and TILs >30% are revealed as independent negative prognostic factors, and further molecular mechanisms need to be explored.

Key words: Small cell lung cancer (SCLC), Limited-stage, Prognosis, Survival analysis

Part Ⅱ   Exploration of stratified evaluation system of small cell lung cancer

Objective: SCLC is characterized by slow clinical progression and lack of stratified treatment strategies. In view of the recent basic studies from cell lines and animal experiments, SCLC has obvious heterogeneity in cell morphology, NE differentiation and expression of lineage transcription factors. This study intends to explore the clinical pathological stratification evaluation method from the perspective of tumor heterogeneity, focusing on the exploration of the heterogeneity from mRNA level to protein level. It is expected to use immunohistochemical detection of protein to reveal the characteristics of molecular heterogeneity, so as to provide partial basis for clinical research and development of heterogenity-related stratification treatment strategies.  

Methods: In this study, 247 postoperative samples of pure SCLC were included and divided into two groups, Classic and Variant, based on pathological morphology observation. Nanostring nCounter technique was used to quantify the mRNA expression of NE genes and lineage transcription factors (ASCL1, NEUROD1, POU2F3, YAP1) in 48 cases, and the related factors were detected by immunohistochemistry simultaneously. Spearman correlation analysis was used to test the consistency of mRNA level and protein level. In addition, the association of morphology, NE differentiation and lineage transcription factors subtypes with clinicopathological features and prognosis was further verified by tissue microarray combined with immunohistochemistry in 247 samples.  

Results: (1) In terms of morphology, 81.8% (202/247) in the Classic group and 18.2% (45/247) in the Variant group. The degree of fibrosis in the Classic group was lower (p=0.021), and the expression of INSM1 was higher (p=0.03). The Classic group had a better prognosis, although the difference was not statistically significant (p>0.05). (2) At mRNA level, NE subtypes were distributed as follows: NE-high (83.3%, 40/48), NE-low (16.7%, 8/48), NE-high subtype was significantly correlated with AJCC progression stage (p=0.014), Classic (p=0.047) and INSM1 expression (p=0.017). There was no significant difference in prognosis (p>0.05), but NE-low subtype tended to have a better prognosis. In terms of lineage transcription factors subtypes, SCLC-A accounted for 45.8% (22/48), SCLC-N 22.9% (11/48), SCLC-P 8.3% (4/48), and SCLC-Y 22.9% (11/48). SCLC-A patients were significantly concentrated in ≤65 years old age group (p=0.016) and Syn positive expression group (p=0.08), and there was no significant difference in prognosis among the four subtypes (p>0.05). (3) At the protein level, NE(INSM1)-high accounted for 73.7% (182/247), and NE(INSM1)-low accounted for 26.3% (65/247). NE(INSM1)-high tended to be Classic in morphology, but did not reach statistical significance (p=0.062). The bronchial invasion (p=0.043), tumor thrombosis (p=0.002) and positive expression of traditional NE markers (CD56, ChrA, Syn) in NE(INSM1)-high patients were significantly higher than those in NE(INSM1)-low (p<0.05). NE(INSM1)-low subtype tended to have a better prognosis, but there was no significant difference between the two groups (p>0.05). The subtypes of lineage transcription factors at the protein level were as follows: SCLC-A (38.9%, 96/247), SCLC-N (46.6%, 115/247), SCLC-P (9.3%, 23/247), SCLC-Y (5.3%, 13/247); Age ≤65 years (p=0.019), Classic (p=0.009), positive expression of NE markers (INSM1, CD56, ChrA, Syn) were significantly concentrated in SCLC-A and SCLC-N subtypes (p<0.05), and there was no statistically significant difference in prognosis among the four subtypes (p>0.05). At the protein level, 92.7% (229/247) of the lineage transcription factors were co-expressed, including 64 (25.9%) with double-positive, 129 (52.2%) with triple-positive, and 36 (14.6%) with all-positive. The proportion of patients with ΔH ≤30 was 24.7% (61/247). In addition, individual cases also showed heterogeneous expression characteristics.  

Conclusion: Quantitative detection of mRNA and protein in 48 surgical samples verified that the characteristics of molecular heterogeneity dominated by transcriptome sequencing could be characterized by immunohistochemical detection at protein level: INSM1 is a reliable marker of NE differentiation, while ASCL1, NEUROD1, POU2F3 and YAP1 can distinguish molecular subtypes at the protein level, but their co-expression and heterogeneous expression suggesting that it is difficult to type by simply using lineage transcription factors. Those with Classic cell morphology tend to be NE-high and SCLC-A/-N subtypes, while those with Variant cell morphology tend to be NE-low and SCLC-P/-Y subtypes. According to the correlation of "Morphology-NE differentiation - lineage transcription factors expression", the establishment of stratified evaluation system of pathological heterogeneity based on these levels may provide scientific basis for clinical research and development of stratified treatment strategies.

Key words: Small cell lung cancer (SCLC), Morphology, Molecular subtypes, Pathological heterogeneity, Hierarchical evaluation system