肿瘤间质比及联合免疫评分对预测脊柱脊索瘤患者生存预后的意义

Significance of tumor-stroma ratio and tumor-stroma ratio combined with immunescore in predicting the survival and prognosis of spinal chordoma patients

中文关键词:  脊索瘤  肿瘤间质比  生存分析  预测因素  免疫评分

英文关键词:Chordoma  Tumor-stroma ratio  Survival analysis  Prognostic biomarker  Immunescore

中文摘要:

  【摘要】 目的：分析脊柱脊索瘤中肿瘤间质比（tumor-stroma ratio，TSR）、免疫评分（immunescore，IS）与患者预后之间的关系，探讨TSR及其联合IS在预测患者预后中的临床价值。方法：回顾性分析77例脊柱脊索瘤患者的临床资料，其中男性54例，女性23例。由2位病理科医师对病理切片中的TSR进行独立评估，利用X-tile软件获得局部无复发生存期（local relapse-free survival，LRFS）及总体生存时间（overall survival，OS）具有最小对数秩P值的点，利用该点将患者分为高TSR组和低TSR组；应用免疫组化法对77例脊索瘤标本进行CD3+和CD8+肿瘤浸润性淋巴细胞（tumor-infiltrating lymphocytes，TILs）子集的检测，再对其进行自动图像分析，得出IS，并将患者分为高IS组和低IS组。采用Kaplan-Meier方法对临床和病理参数（年龄、性别、肿瘤大小、位置、术前复发、周围组织浸润、级别、分期、切除方式、出血和坏死情况、TSR和IS）与患者结局（LRFS、OS）的关系进行单因素生存分析；纳入单因素分析有显著统计学意义的变量，使用Cox比例风险模型分析患者LRFS和OS的独立危险因素；Pearson′s相关性分析两个连续变量之间的关系。应用受试者工作特征（receiver operating characteristic, ROC）曲线比较TSR联合IS和TSR或IS单独使用时的预测能力。应用Bland-Altman一致性分析评估两位评估者之间TSR测量的一致性。结果：两位评估者在TSR评估方面存在很强的相关性（r=0.924，P<0.001）；Bland-Altman证实两位评估者之间TSR数据的平均差异较小，有良好的一致性（P=0.292）。单变量分析显示TSR、IS、年龄、周围肌肉浸润、手术切除方式与LRFS存在相关性（P<0.05）。TSR、IS、周围肌肉浸润、肿瘤分期、手术切除方式与OS存在相关性（P<0.05）。TSR与IS呈正相关（P<0.05），高IS预示着良好的临床预后，而低TSR和低IS患者存活率最低。LRFS的多变量Cox分析显示周围肌肉浸润、TSR和IS可独立预测预后（P<0.05），OS的多变量Cox分析显示TSR是OS的唯一预测因素（P＝0.011）。ROC分析显示TSR在预测LRFS和OS方面的能力与IS相当 [LRFS：AUC（TSR）=0.565，AUC（IS）=0.630；OS：AUC（TSR）=0.632，AUC（IS）=0.648]；将TSR纳入IS系统中可提高TSR对疾病复发和生存率预测的准确性 [LRFS：AUC（TSR+IS）=0.709；OS：AUC（TSR+IS）=0.727]。结论：TSR与患者的生存率相关，且是LRFS和OS的预测因素。在生存分析中纳入TSR可提高其预测预后的能力，将TSR纳入IS系统中可提高IS对疾病复发和生存率预测的准确性。

英文摘要:

  【Abstract】 Objectives: By analyzing the relationship between tumor-stroma ratio(TSR), immunescore(IS), and patient prognosis in spinal chordoma, we aimed to determine the clinical significance of TSR and further investigate the predictive ability of TSR combined with IS. Methods: The clinical data of 77 patients with spinal chordoma were retrospectively analyzed, including 54 men and 23 women. All the spinal cord tumor cases fell into the class of classic pathology. TSR was evaluated on pathology slides by 2 independent pathologists, the local relapse-free survival(LRFS) and overall survival(OS) point with the smallest log rank P-value was obtained using X-tile software, and then the patients were divided it into high TSR and low TSR groups. Immunohistochemistry was applied to 77 tumor specimens for CD3+ and CD8+ tumor-infiltrating lymphocytes subset(TILs), automated image analysis was performed to derive IS, and the patients were classified into two groups: high and low on the basis of IS. A univariate Kaplan-Meier curve by log-rank test was used to explore the relationship between clinicopathological factors and patient outcomes. A multivariate Cox proportional hazards model was used to assess independent prognostic factors of LRFS and OS after adjusting for other clinical predictors that were significant in our univariate survival analysis. Pearson′s correlation test was used to observe the relationship between two continuous variables. Receiver operating characteristic(ROC) curves were used to compare the predictive power of TSR in combination with IS or TSR or IS alone. And Bland-Altman consistency analysis was used to assess the consistency of TSR measures between two assessors. All tests were two-sided, and P<0.05 was considered to be statistically significant. Results: There was a strong correlation between the two assessors for TSR assessment(r=0.924, P<0.001); Bland-Altman confirmed a small mean difference in TSR data between the two assessors with good agreement(P=0.292). Univariate analysis showed that TSR, IS, age, surrounding muscle invasion, type of surgery were correlated with LRFS(P<0.05). TSR, IS, surrounding muscle invasion, tumor stage, and type of surgerywere related to OS(P<0.05). TSR was positively correlated with IS(P<0.05), high IS indicated a good clinical prognosis, patients with low TSR combined with low IS had the lowest survival rates. Multivariate Cox analysis of LRFS showed that surrounding muscle invasion, TSR, and IS could independently predict prognosis(P<0.05), and multivariate Cox analysis of OS shows that TSR was the only predictor of OS(P=0.011). ROC analysis showed that incorporating TSR into the IS system improved the accuracy of the IS in predicting disease recurrence and survival. Conclusions: TSR is associated with patient survival and is a predictor of LRFS and OS. Inclusion of TSR in survival analysis improves its ability to predict prognosis, and inclusion of TSR in the IS system improves the accuracy of IS in predicting disease recurrence and survival.