Stanford University School of Medicine, Stanford, CA, USA.
Quantitative Sciences Unit, Department of Medicine, Stanford University School of Medicine, Stanford, CA, USA.
Lung Cancer. 2021 May;155:61-67. doi: 10.1016/j.lungcan.2021.03.007. Epub 2021 Mar 11.
Lung cancer survivors have a high risk of developing a second primary lung cancer (SPLC). While national screening guidelines have been established for initial primary lung cancer (IPLC), no consensus guidelines exist for SPLC. Furthermore, the factors that contribute to SPLC risk have not been established. This study examines the potential for using serum metabolomics to identify metabolite biomarkers that differ between SPLC cases and IPLC controls.
In this pilot case-control study, we applied an untargeted metabolomics approach based on ultrahigh performance liquid chromatography-tandem mass spectroscopy (UPLC-MS/MS) to serum samples of 82 SPLC cases and 82 frequency matched IPLC controls enrolled in the Boston Lung Cancer Study. Random forest and unconditional logistic regression models identified metabolites associated with SPLC. Candidate metabolites were integrated into a SPLC risk prediction model and the model performance was evaluated through a risk stratification approach.
The untargeted analysis detected 1008 named and 316 unnamed metabolites among all study participants. Metabolites that were significantly associated with SPLC (False Discovery Rate q-value < 0.2) included 5-methylthioadenosine (odds ratio [OR] = 2.04, 95 % confidence interval [CI] 1.39-3.01; P = 2.8 × 10) and phenylacetylglutamine (OR = 2.65, 95 % CI 1.56-4.51; P = 3.2 × 10), each exhibiting approximately 1.5-fold increased levels among SPLC cases versus IPLC controls. In stratifying the study participants across quartiles of estimated SPLC risk, the risk prediction model identified a significantly higher proportion of SPLC cases in the fourth compared to the first quartile (68.3 % versus 39.0 %; P = 0.044).
SPLC cases may have distinct metabolomic profiles compared to those in IPLC patients without SPLC. A risk stratification approach integrating metabolomics may be useful for distinguishing patients based on SPLC risk. Prospective validation studies are needed to further evaluate the potential for leveraging metabolomics in SPLC surveillance and screening.
肺癌幸存者发生第二原发性肺癌（SPLC）的风险较高。虽然已经为初始原发性肺癌（IPLC）制定了国家筛查指南，但尚无 SPLC 的共识指南。此外，导致 SPLC 风险的因素尚未确定。本研究旨在探讨使用血清代谢组学来识别 SPLC 病例和 IPLC 对照之间存在差异的代谢物生物标志物的可能性。
在这项初步的病例对照研究中，我们应用了一种基于超高效液相色谱-串联质谱（UPLC-MS/MS）的非靶向代谢组学方法，对波士顿肺癌研究中纳入的 82 例 SPLC 病例和 82 例频率匹配的 IPLC 对照的血清样本进行了分析。随机森林和无条件逻辑回归模型确定了与 SPLC 相关的代谢物。候选代谢物被整合到 SPLC 风险预测模型中，并通过风险分层方法评估了模型性能。
非靶向分析在所有研究参与者中检测到 1008 种命名代谢物和 316 种未命名代谢物。与 SPLC 显著相关的代谢物（错误发现率 q 值 < 0.2）包括 5-甲基硫代腺苷（比值比 [OR] = 2.04，95%置信区间 [CI] 1.39-3.01；P = 2.8×10）和苯乙酰谷氨酰胺（OR = 2.65，95%CI 1.56-4.51；P = 3.2×10），SPLC 病例与 IPLC 对照相比，这两种代谢物的水平均升高了约 1.5 倍。在按估计的 SPLC 风险四分位数对研究参与者进行分层时，风险预测模型在第四四分位数中识别出明显更高比例的 SPLC 病例（68.3% 比 39.0%；P = 0.044）。
与无 SPLC 的 IPLC 患者相比，SPLC 病例可能具有不同的代谢组学特征。整合代谢组学的风险分层方法可能有助于根据 SPLC 风险区分患者。需要前瞻性验证研究来进一步评估代谢组学在 SPLC 监测和筛查中的潜在应用。
