Abstract：
BACKGROUND: Microbiota may be associated with esophageal squamous cell carcinoma (ESCC) development. However, it is not known the predictive value of microbial biomarkers combining epidemiological factors for the early detection of ESCC and precancerous lesions.
METHODS: A total of 449 specimens (esophageal swabs and saliva) were collected from 349 participants with different esophageal statuses in China to explore and validate ESCC-associated microbial biomarkers from genes level to species level by 16S rRNA sequencing, metagenomic sequencing and real-time quantitative polymerase chain reaction.
RESULTS: A bacterial biomarker panel including Actinomyces graevenitzii (A.g_1, A.g_2, A.g_3, A.g_4), Fusobacteria nucleatum (F.n_1, F.n_2, F.n_3), Haemophilus haemolyticus (H.h_1), Porphyromonas gingivalis (P.g_1, P.g_2, P.g_3) and Streptococcus australis (S.a_1) was explored by metagenomic sequencing to early detect the participants in Need group (low-grade intraepithelial neoplasia, high-grade intraepithelial neoplasia and ESCC) vs participants without these lesions as the Noneed group. Significant quantitative differences existed for each microbial target in which the detection efficiency rate was higher in saliva than esophageal swab. In saliva, the area under the curve (AUC) based on the microbial biomarkers (A.g_4 ∩ P.g_3 ∩ H.h_1 ∩ S.a_1 ∩ F.n_2) was 0.722 (95% CI 0.621-0.823) in the exploration cohort. Combining epidemiological factors (age, smoking, drinking, intake of high-temperature food and toothache), the AUC improved to 0.869 (95% CI 0.802-0.937) in the exploration cohort, which was validated with AUC of 0.757 (95% CI 0.663-0.852) in the validation cohort.
CONCLUSIONS: It is feasible to combine microbial biomarkers in saliva and epidemiological factors to early detect ESCC and precancerous lesions in China.
METHODS: A total of 449 specimens (esophageal swabs and saliva) were collected from 349 participants with different esophageal statuses in China to explore and validate ESCC-associated microbial biomarkers from genes level to species level by 16S rRNA sequencing, metagenomic sequencing and real-time quantitative polymerase chain reaction.
RESULTS: A bacterial biomarker panel including Actinomyces graevenitzii (A.g_1, A.g_2, A.g_3, A.g_4), Fusobacteria nucleatum (F.n_1, F.n_2, F.n_3), Haemophilus haemolyticus (H.h_1), Porphyromonas gingivalis (P.g_1, P.g_2, P.g_3) and Streptococcus australis (S.a_1) was explored by metagenomic sequencing to early detect the participants in Need group (low-grade intraepithelial neoplasia, high-grade intraepithelial neoplasia and ESCC) vs participants without these lesions as the Noneed group. Significant quantitative differences existed for each microbial target in which the detection efficiency rate was higher in saliva than esophageal swab. In saliva, the area under the curve (AUC) based on the microbial biomarkers (A.g_4 ∩ P.g_3 ∩ H.h_1 ∩ S.a_1 ∩ F.n_2) was 0.722 (95% CI 0.621-0.823) in the exploration cohort. Combining epidemiological factors (age, smoking, drinking, intake of high-temperature food and toothache), the AUC improved to 0.869 (95% CI 0.802-0.937) in the exploration cohort, which was validated with AUC of 0.757 (95% CI 0.663-0.852) in the validation cohort.
CONCLUSIONS: It is feasible to combine microbial biomarkers in saliva and epidemiological factors to early detect ESCC and precancerous lesions in China.
摘要：
背景：微生物群可能与食管鳞状细胞癌（ESCC）的发展有关。然而,目前尚不清楚微生物生物标志物结合流行病学因素对早期发现ESCC和癌前病变的预测价值.
方法：从中国不同食管状态的349名参与者中收集了449份标本（食管拭子和唾液），通过16SrRNA测序从基因水平到物种水平探索和验证ESCC相关的微生物生物标志物。宏基因组测序和实时定量聚合酶链反应。
结果：包括放线菌（A.g_1,A.g_2,A.g_3,A.g_4),具核梭杆菌(F.n_1,F.n_2,F.n_3),溶血嗜血杆菌(H.h_1),牙龈卟啉单胞菌(P.g_1、P.g_2、P.g_3)和南方链球菌(S.a_1)通过宏基因组测序进行探索，以早期检测需要组(低级别上皮内瘤变，高级别上皮内瘤变和ESCC）与无这些病变的参与者作为Noneed组。每个微生物靶标都存在显着的定量差异，其中唾液中的检测效率高于食管拭子。在唾液中，基于微生物生物标志物的曲线下面积(AUC)(A.g_4281P.g_3281H.h_1281S.a_1281F.在探索队列中，n_2）为0.722（95%CI0.621-0.823）。结合流行病学因素(年龄，吸烟,饮酒,摄入高温食物和牙痛)，在探索队列中，AUC提高到0.869(95%CI0.802-0.937)，在验证队列中，AUC为0.757（95%CI0.663-0.852）。
结论：在中国，将唾液中的微生物生物标志物与流行病学因素相结合，以早期发现ESCC和癌前病变是可行的。
方法：从中国不同食管状态的349名参与者中收集了449份标本（食管拭子和唾液），通过16SrRNA测序从基因水平到物种水平探索和验证ESCC相关的微生物生物标志物。宏基因组测序和实时定量聚合酶链反应。
结果：包括放线菌（A.g_1,A.g_2,A.g_3,A.g_4),具核梭杆菌(F.n_1,F.n_2,F.n_3),溶血嗜血杆菌(H.h_1),牙龈卟啉单胞菌(P.g_1、P.g_2、P.g_3)和南方链球菌(S.a_1)通过宏基因组测序进行探索，以早期检测需要组(低级别上皮内瘤变，高级别上皮内瘤变和ESCC）与无这些病变的参与者作为Noneed组。每个微生物靶标都存在显着的定量差异，其中唾液中的检测效率高于食管拭子。在唾液中，基于微生物生物标志物的曲线下面积(AUC)(A.g_4281P.g_3281H.h_1281S.a_1281F.在探索队列中，n_2）为0.722（95%CI0.621-0.823）。结合流行病学因素(年龄，吸烟,饮酒,摄入高温食物和牙痛)，在探索队列中，AUC提高到0.869(95%CI0.802-0.937)，在验证队列中，AUC为0.757（95%CI0.663-0.852）。
结论：在中国，将唾液中的微生物生物标志物与流行病学因素相结合，以早期发现ESCC和癌前病变是可行的。
