目的：呼吸分析技术具有无创、简便和快速的特点，为重大慢性疾病早期筛查与诊断提供了一种高应用前景的方法。当前，解决呼吸分析诊断重大慢性疾病走向临床应用的挑战在于，发展先进的呼吸气体分析技术和仪器设备，对特定疾病人群进行实时、在线、高灵敏度、高选择性的呼吸气体分析，建立呼吸生物标记物与人体异常（疾病、 健康、外部环境影响）诊断参数的定量相关性。

方法：使用基于腔衰荡光谱（cavity ring-down spectroscopy, CRDS）技术的呼吸丙酮分析仪测量了200名健康志愿者和17例1型糖尿病（type 1 diabetes, T1D）患者的呼吸丙酮，同时对实验室前期积累的65名健康志愿者、300例2型糖尿病（type 2 diabetes, T2D）患者和22例T1D患者的呼吸丙酮数据进行再分析，以上志愿者均控制运动、压力、饮食以及疾病等条件，评价空腹呼吸丙酮及联合其影响因素在糖尿病筛查中的作用。使用基于CRDS技术的呼出气一氧化氮（fractional exhaled nitric oxide, FeNO）检测分析系统，测量了456例原发性肺癌患者和284例健康志愿者FeNO含量。对FeNO浓度数据与临床指标进行相关统计分析，评估FeNO在肺癌筛查中的应用。

结果：T1D患者的空腹呼吸丙酮浓度（2.24±1.43 ppmv）明显高于健康人（1.43±0.55 ppmv）与T2D患者（1.41±0.73 ppmv）；男性患者空腹呼吸丙酮浓度均高于女性患者；健康人空腹呼吸丙酮浓度与年龄成正相关关系（R=0.312）；T1D患者空腹呼吸丙酮浓度与BMI成正相关（R=0.33）；T2D患者空腹呼吸丙酮浓度与身高成正相关（R=0.18）；空腹呼吸丙酮浓度对T1D的诊断有意义（AUC=0.853），联合BMI对T1D的诊断有意义（AUC=0.886）；空腹呼吸丙酮浓度联合性别以及身高对T2D的诊断有意义（AUC=0.672）；原发性肺癌患者的FeNO水平（34.13±17.21 ppbv）高于健康对照者(20.95+12.12 ppbv)。各组间在肺癌类型、分期、性别、吸烟情况、年龄、BMI、禁食等方面均无显著差异。FeNO水平有助于评估肺癌患者（AUC=0.728，敏感性=0.798，特异性=0.55，P<0.01）。

结论：呼吸丙酮联合糖尿病患者临床指标在T1D和T2D的诊断上有应用潜力。基于CRDS技术的FeNO分析有望为肺癌提供一定诊断依据。

Objective: the breath analysis technique has characteristics of noninvasive, simple and rapid, which provides a promising method for early screening and diagnosis of major chronic diseases. At present, to solve the challenge to the clinical application of breath analysis in the diagnosis of major chronic diseases is to develop advanced breathing gas analysis technology and equipment, a real-time, online, highly sensitive and selective breath gas analysis was conducted on a specific disease population, to establish a quantitative correlation between breath biomarkers and diagnostic parameters of human abnormalities (disease, health, external environmental impact).

Method: Used the breathing acetone analyzer based on the technology of cavity ring-down spectroscopy (CRDS), measured acetone in breath of 200 healthy volunteers and 17 cases of patients with type 1 diabetes (T1D), at the same time, the acetone in breath datas of 65 healthy volunteers, 300 type 2 diabetes (T2D) patients and 22 T1D patients accumulated in the previous laboratory work were re-analyzed. All the above volunteers were all sampled under conditions including exercise, stress, diet and disease, to evaluate the role of fasting acetone and associated with its influence factors in diabetes screening. Fractional exhaled nitric oxide (FeNO) levels in 456 patients with primary lung cancer and 284 healthy volunteers were measured by using a FeNO detection and analysis system based on CRDS technology. FeNO concentration data and clinical indicators were statistically analyzed to evaluate the application of FeNO in lung cancer screening.

Results: The concentration of acetone in breath during fasted of T1D patients (2.24±1.43 ppmv) was significantly higher than that of healthy people (1.43±0.55 ppmv) and T2D patients (1.41±0.73 ppmv). The concentration of acetone in male patients was higher than that in female patients. There was a positive correlation between the concentration of acetone in fasting and the age of healthy people (R=0.312). There was a positive correlation between fasting acetone concentration and BMI in T1D patients (R=0.33). The concentration of acetone in T2D patients was positively correlated with their height (R=0.18). The concentration of acetone in fasting respiration was significant for the diagnosis of T1D (AUC=0.853), and the combination of BMI was significant for the diagnosis of T1D (AUC=0.886). The combination of gender and height of acetone concentration in fasting respiration were significant for the diagnosis of T2D (AUC=0.672). FeNO levels in patients with primary lung cancer (34.13±17.21 ppbv) were higher than those in healthy controls (20.95+12.12 ppbv). There were no significant differences in lung cancer type, stage, gender, smoking status, age, BMI, fasting, etc. FeNO levels were helpful in assessing lung cancer patients (AUC=0.728, sensitivity =0.798, specificity =0.55, P<0.01).

Conclusion: clinical indicators of respiratory acetone combined with diabetes mellitus have potential application in the diagnosis of T1D and T2D. FeNO analysis based on CRDS technology is expected to provide some diagnostic basis for lung cancer.