研究背景 

肺癌已成为发展中国家和发达国家的主要死亡原因，对个人和医疗保健系统造成了巨大的社会经济负担。由于CT筛查的应用，I期肺癌检出率显著提高，有效改善了病人的生存率。肺腺癌（Lung adenocarcinoma, LUAD）为目前非小细胞肺癌的主要病理亚型，手术切除为I期肿瘤标准治疗方式，但术后仍有部分患者（14-36%）复发，是影响预后的主要原因。因此，研究并发现影响复发相关的临床病理因素或者肿瘤标记物，可以更加精准的预测Ⅰ期肺癌复发风险，可用于指导制订术后随访方案和辅助治疗策略，是进一步提高Ⅰ期肺癌疗效的有效手段。

研究目的

本研究旨在分析并鉴定I期LUAD术后复发相关的临床病理特征和肿瘤标记物。

研究方法

以2014年1月至2018年10月在中国医学科学院肿瘤医院胸外科行肺癌手术切除，并经病理检查证实的I期原发性腺癌患者作为研究对象，最终入组患者为603例。所有患者临床病理资料通过病历系统获得，同时对病理特征（包括主要组织病理亚型、肿瘤分化和周围侵犯特征）和影像学特征（病灶影像学成分、毛刺征、胸膜牵拉征和空泡征）进行复阅，临床资料及随访资料完整。提取公共生物信息数据库RNA-seq数据，并采用免疫组化法检测I期LUAD肿瘤组织和配对的正常肺组织中GSTA1蛋白的表达情况。通过Kaplan-Meier法、单因素和多因素的Cox回归分析法、分层分析法，以及逆概率加权（Inverse probability of treatment weighting, IPTW）法探索影响I期LUAD术后无复发生存期（Recurrence free survival, RFS）的危险因素。

研究结果

影响I期LUAD患者术后RFS的临床病理因素

共纳入603例I期LUAD患者，包括334例（44.6%）女性和269例（55.4%）男性，中位年龄为60岁（范围53 - 65岁）。病理亚型分类为附壁型为主（13.3%）、腺泡型为主（62.5%）、乳头状为主（16.8%）、微乳头型（1.0%）为主和实体型（6.4%）为主。截止至随访日期，复发67例，所有患者的中位RFS为71.3个月（范围：51.0 - 79.1个月），复发患者的中位RFS为26.0个月（范围：12.8 - 38.0个月）。经过单因素分析筛选与复发相关因素，多因素Cox回归分析提示有合并症（HR：4.83，95%CI：2.91 - 8.01；P < 0.001）、影像学结节纯实性成分（HR：1.89，95%CI：1.06 - 3.37；P = 0.031）、肿瘤大小为2 - 3 cm（HR：1.84，95%CI：1.07 - 3.17；P = 0.027）、组织病理亚型为乳头型或微乳头型或实体型为主（HR：1.87，95%CI：1.09 - 3.23；P < 0.001）、有STAS（HR：3.14，95%CI：1.44 - 6.84；P = 0.004）和有神经侵犯（HR：2.70，95%CI：1.22 - 5.99；P = 0.015）是I期LUAD术后RFS的危险因素。

通过单因素和多因素分析、IPTW分析和建立列线图，我们发现STAS对I期腺癌RFS影响较为显著，根据预后将其分为STAS（+）组和STAS（-）组两组进一步分析。STAS（+）组的复发率高于 STAS（-）组（22.5%比3.2%，P < 0.001）。基于pT分层的Cox比例风险模型的结果表明，在pT1b和pT1c分层中，STAS（+）患者的复发风险分别比STAS（-）患者高4.56倍（95%CI：1.56 - 13.33；P=0.006）和3.16倍（95%CI：1.07 - 9.33；P = 0.038）。pT1b/c、pT1b和 pT1c的STAS（-）组患者分别与全部pT1a患者相比较，RFS差异均无统计学意义[(84.97 ± 0.72) 比 (84.05 ± 1.11)个月，(85.60 ± 0.74) 比 (84.05 ± 1.11)个月，(81.49 ± 1.63) 比 (84.05 ± 1.11)个月；均P > 0.05]。IPTW调整前后，pT1b/c中STAS（+）组与STAS（-）组患者的RFS差异均有统计学意义[(72.50 ± 2.23) 比 (85.12 ± 0.72)个月，(77.74 ± 1.12) 比 (84.59 ± 0.64)个月，均P < 0.001]。

谷胱甘肽S转移酶1 （glutathione S-transferase alpha 1, GSTA1）表达与I期LUAD患者临床病理特征和RFS相关

经过生物信息学研究，发现GSTA1在I期LUAD细胞中的mRNA表达水平显著高于正常人肺细胞表达水平（P < 0.001），并且肿瘤越大其表达量越低（P = 0.023），差异均具有统计学意义。通过对本院样本进行检验，免疫组化结果显示I期LUAD组织中GSTA1蛋白表达水平显著高于配对的正常人肺组织（P < 0.001）。基于GSTA1免疫组化评分，将患者分为高表达和低表达两组，分析了GSTA1表达与临床病理特征的关系，结果显示两组间月经初潮年龄（P = 0.035）、STAS （P = 0.020）和复发情况（P = 0.046）有统计学差异。Pearson相关分析显示GSTA1表达量与微乳头型（P = 0.017）和实体型（P = 0.022）肿瘤细胞占比呈负相关，但与附壁型（P = 0.037）和腺泡型（P = 0.042）肿瘤细胞占比呈正相关。对数据库和本院数据进行单因素和多因素Cox回归分析和Kaplan-Meier分析，结果显示GSTA1低表达与RFS缩短有关，提示GSTA1低表达者预后更差，GSTA1是I期LUAD潜在的预后生物标志物。

结论

在603例I期LUAD病例中，合并症、影像学纯实性成分、肿瘤大小为2 - 3cm、组织学病理亚型以乳头型或微乳头型或实体型为主、有STAS和有神经侵犯是影响术后RFS的危险因素。

STAS增加了术后pT1b/cN0M0腺癌患者的复发风险，但不是pT1aN0M0腺癌患者术后复发风险。

在I期LUAD中，GSTA1在肿瘤组织中比在正常肺组织中表达更高，但相对低表达的腺癌RFS更短，可能与调控通路和肿瘤干性相关。

Background

Both developed and developing countries face a great deal of socio-economic burden from lung cancer, which is the leading cause of death. The widespread use of CT screening has led to an increase in detection rates of stage I lung cancer, increasing patient survival rates. In the current medical climate, lung adenocarcinoma (LUAD) represents the main subtype of non-small cell lung cancer. Surgical resection is the standard of care for stage I tumours, but recurrence remains a major cause of prognosis in a proportion of patients (14-36%) after surgery. Therefore, the study and identification of clinicopathological factors or tumour markers associated with recurrence can more accurately predict the risk of recurrence in stage I lung cancer, which can be used to guide the development of post-operative follow-up protocols and adjuvant treatment strategies, and is an effective means to further improve the outcome of stage I lung cancer.

Aims

The aim of this study was to analyse and identify the clinicopathological features and tumour markers associated with recurrence after stage I LUAD.

Methods

Patients with surgically resected lung cancer and pathologically confirmed stage I primary adenocarcinoma in the Department of Thoracic Surgery, Cancer Hospital, Chinese Academy of Medical Sciences were selected for the study from January 2014 to October 2018, and 603 patients were finally enrolled. All patients' clinicopathological data were obtained through the medical record system, while pathological features (including major histopathological subtypes, tumor differentiation and peripheral invasion features) and imaging features (including lesion imaging components, burr sign, pleural traction sign and vacuolation sign) were reviewed, and clinical data and follow-up data were complete. Public bioinformatics database RNA-seq data were extracted, while GSTA1 protein expression in stage I LUAD tumour tissue and paired normal lung tissue was detected using immunohistochemistry. The Kaplan-Meier method, univariate and multifactorial Cox regression analysis, stratified analysis and Inverse probability of treatment weighting (IPTW) were used to explore the risk factors affecting recurrence-free survival (RFS) after surgery for stage I LUAD.

Results

Clinicopathological factors affecting postoperative RFS in patients with stage I LUAD

We included 603 subjects with stage I LUAD, including 334 women (44.6%) and 269 men (55.4%), whose median age was 60 (range 53-65 years). Pathological subtypes were classified as predominantly accessory (13.3%), predominantly alveolar (62.5%), predominantly papillary (16.8%), predominantly micropapillary (1.0%) and predominantly solid (6.4%). As of the follow-up date, there were 67 recurrences, with a median RFS of 71.3 months (range: 51.0 - 79.1 months) for all patients and 26.0 months (range: 12.8 - 38.0 months) for those with recurrences. After univariate analysis to screen for prognostic-related influences, multifactorial Cox regression analysis revealed the presence of comorbidities (HR: 4.83, 95% CI: 2.91 - 8.01; P < 0.001), a purely solid component of imaging nodules (HR: 1.89, 95% CI: 1.06 - 3.37; P = 0.031), a tumour size of 2 - 3 cm (HR. 1.84, 95% CI: 1.07 - 3.17; P = 0.027), a predominantly papillary or micropapillary or solid histopathological subtype (HR: 1.87, 95% CI: 1.09 - 3.23; P < 0.001), the presence of STAS (HR: 3.14, 95% CI: 1.44 - 6.84; P = 0.004) and the presence of nerve invasion ( HR: 2.70, 95% CI: 1.22 - 5.99; P = 0.015) were risk factors for postoperative RFS after stage I LUAD.

Using univariate and multifactorial analyses, IPTW analysis and the creation of columnar plots, we found a more significant effect of STAS on RFS in stage I adenocarcinoma, which was further analysed by dividing it into two groups based on prognosis: the STAS (+) group and the STAS (-) group. the recurrence rate was higher in the STAS (+) group than in the STAS (-) group (22.5% versus 3.2%, P < 0.001). Multi-factor Cox regression analysis based on pT stratification showed that the risk of recurrence was 4.56-fold (95% CI: 1.56 ~ 13.33; P = 0.006) and 3.16-fold (95% CI: 1.07 ~ 9.33; P = 0.038) higher in the pT1b and pT1c strata, respectively, for STAS (+) patients than for STAS (-) patients. pT1b /c, pT1b and pT1c patients in the STAS(-) group were not statistically different from all pT1a patients, respectively [(84.97 0.72) versus (84.05 1.11) months, (85.60 0.74) versus (84.05 1.11) months and (81.49 1.63) versus (84.05 1.11) months; all P > 0.05]. The difference in RFS between patients in the STAS (+) and STAS (-) groups in pT1b/c before and after IPTW adjustment was statistically significant [(72.50 ± 2.23) versus (85.12 ± 0.72) months, (77.74 1.12) versus (84.59 0.64) months; all P < 0.001].

GSTA1 expression correlates with clinicopathological features and RFS in patients with stage I LUAD

Based on bioinformatics analysis, it was determined that GSTA1 mRNA expression level in stage I LUAD tissues was significantly higher than that in normal lung tissue (P < 0.001), and the larger the tumor, the lower its expression (P = 0.023). Through the verification of the samples collected in our hospital, the immunohistochemical results showed that a significant increase in GSTA1 expression was observed in stage I LUAD tissues compared with normal lung tissue (P < 0.001). Based on GSTA1 immunohistochemical scores, the patients were divided into groups with high and low expression. A correlation was found between GSTA1 expression and clinicopathological features. According to the results, there were significant differences in menarche age (P = 0.035), STAS (P = 0.020) and recurrence (P = 0.046) between the two groups. Pearson correlation analysis showed that the expression of GSTA1 was negatively correlated with the proportion of tumor cells in microemulsion head type (P = 0.017) and solid type (P = 0.022), but positively correlated with the proportion of tumor cells in mural type (P = 0.037) and acinar type (P = 0.042). Univariate and multivariate Cox regression analysis and Kaplan Meier analysis were performed on the database and hospital data. The results showed that the low expression of GSTA1 was related to the shortening of RFS, suggesting that the prognosis of patients with low expression of GSTA1 was worse, and GSTA1 was a potential prognostic biomarker of stage I LUAD.

Conclusion

In 603 patients with stage I LUAD, comorbidity, purely solid imaging component, tumour size of 2 - 3 cm, histopathological subtype of predominantly papillary or micropapillary or solid type, presence of STAS and presence of nerve invasion were risk factors for postoperative RFS.

STAS increased the risk of postoperative recurrence in patients with pT1b/cN0M0 stage adenocarcinoma, but not in patients with pT1aN0N0 stage adenocarcinoma.

In stage I LUAD, GSTA1 was more highly expressed in tumour tissue than in normal lung tissue, but RFS was shorter in relatively low-expressing adenocarcinomas, possibly related to regulatory pathways and tumour stemness.