Objective  To investigate the performance of using lung dynamic compliance (Cdyn) and airway resistance (RAW) levels to predict lung infection in elderly esophageal cancer patients who have undergone radiotherapy.

  Methods  A total of 298 elderly esophageal cancer patients who received radiotherapy at Shanxi Fenyang Hospital between October 2017 and July 2022 were retrospectively enrolled and their clinical data were collected. The patients were divided into an infection group (124 cases) and a non-infection group (174 cases) according to their status of lung infection. Then, in the infection group, CURB-65 score was used to assess the severity of the patients' lung infection and the patients were further divided into subgroups accordingly, with 36 cases in the mild infection subgroup, 58 cases in the moderate infection subgroup, and 30 cases in the severe infection subgroup. The levels of Cdyn, RAW, and infection indicators, including serum procalcitonin (PCT), interleukin-6 (IL-6), and angiotensin Ⅱ (Ang Ⅱ), were measured in both groups of patients and the differences in the findings were compared between the infection and the non-infection groups and among patients with infection of varying degrees of severity. The correlation between Cdyn and RAW and the levels of PCT, IL-6, and Ang Ⅱ was analyzed. Receiver operating characteristic (ROC) curve was used to evaluate the performance of predicting infection with Cdyn and RAW.

  Results  The Cdyn level of patients in the infection group was lower than that of patients in the non-infection group, while the RAW level of the infection group was higher than that of the non-infection group (P<0.05). Among the infection subgroup, the level of Cdyn of the mild infection subgroup was higher than those of the moderate and severe infection subgroups, while the levels of RAW, PCT, IL-6, and Ang Ⅱ of the mild infection subgroup were lower than those of the moderate severe subgroups. The level of Cdyn of the moderate infection subgroup was higher than that of the severe infection subgroup, while the RAW, PCT, IL-6, and Ang Ⅱ levels of the moderate infection subgroup were lower than those of the severe infection subgroup, with all difference being statistically significant (P<0.05). The Cdyn level of patients with lung infection was negatively correlated with PCT, IL-6, and Ang Ⅱ levels and the severity of infection (r=−0.501, −0.430, −0.367, and −0.484, respectively, P<0.05), while RAW was positively correlated with PCT, IL-6, and Ang Ⅱ levels and the severity of infection (r=0.483, 0.395, 0.374, and 0.423, respectively, P<0.05). The area under the curve (AUC) of Cdyn and RAW for predicting lung infection in elderly patients with esophageal cancer after radiotherapy were 0.898 (95% confidence interval CI: 0.857-0.930) and 0.823 (95% CI: 0.775-0.865), respectively, and the AUC of combined evaluation of Cdyn and RAW was 0.959 (95% CI: 0.930-0.979), which suggested that the predictive performance of combined evaluation was better than evaluation with Cdyn or RAW alone.

  Conclusion  When elderly esophageal cancer patients develop lung infection after radiotherapy, their Cdyn level is decreased, while the levels of RAW, PCT, IL-6, and Ang Ⅱ are increased. In addition, the levels of Cdyn and RAW are correlated with the PCT, IL-6, and Ang Ⅱ levels. The combined use of Cdyn and RAW shows good performance for predicting lung infection in patients.