Because there is no international standard for the diagnosis, nomenclature and classification of IPAF, studies on IPAF are limited. The diagnosis of IPAF is based on laboratory indicators, chest HRCT, pathological biopsy and other clinical manifestations, which is more complex and less practical. Therefore, the identification of biomarkers of IPAF can not only improve the level of diagnosis of IPAF, but also help to understand the pathophysiological mechanism of the occurrence and development of IPAF. So far, there are few studies on IPAF biomarkers.
KL‑6 is a MUC-1 mucin, commonly found in regenerative type Ⅱ alveolar epithelial cells [17, 18]. Interstitial pneumonia will proliferate type Ⅱ alveolar epithelial cells, resulting in an increase in KL-6 concentration and due to the damage leads to an increase in vascular permeability, allowing KL-6 to enter the bloodstream, so the concentration of KL-6 in the serum of patients with ILD increases [19, 20]. SP-A is a member of the water-soluble C-type lectin family and is an important part of the lung's innate immune system . The pathogenesis of IPF may be related to abnormal endoplasmic reticulum processing of lung surfactant proteins . The expression of SP-A1 gene is up-regulated by genetic analysis of lung biopsy in IPF patients, and SP-A2 gene defects are also associated with familial IPF pathogenesis [23, 24]. In Japan, serum SP-A and KL-6 levels are widely used as biomarkers for diagnosis, severity assessment and prognosis prediction of ILD patients . These findings collectively indicate that serum SP-A and KL-6 can act as a surrogate marker for the active process of disease progression  . However, it is not known whether the changes of SP-A and KL-6 levels, especially in the serum of patients with IPAF, can reflect the correlation between the changes and the progression of IPAF patients.
Our study found that compared with the healthy control group, the serum levels of SP-A and KL-6 in IPAF patients were significantly increased (P < 0.01). Our findings are consistent with the results from the report by Xue et al . When a cut-off value of 27.6 ng/mL was used, the sensitivity and specific of using serum SP-A levels as a diagnostic biomarker was 87.5% and 92%, respectively. When a cut-off value of 344 U/mL was used, the sensitivity and specific of using serum KL-6 levels as a diagnostic biomarker was 98.4% and 96%, respectively. Indicating that serum SP-A and KL-6 might be a promising biomarker for diagnosis of IPAF patients. In IPF patients, cut-off values were set as the levels that resulted in the optimal diagnostic accuracy for SP-A and KL-6: 476 U/mL for KL-6 and 44.0 ng/mL for SP-A . It is suggested that different levels of criticality may be required in ILD patients of different subtypes.
In IPAF, the serum SPA, KL-6 and %DLCO were significantly negatively correlated (P < 0.05) but not significantly correlated with FVC and FEV1(P > 0.05). Interestingly, an IPF cohort study by Sokai et al also reported a similar correlation between KL-6 and pulmonary function parameters such as FVC and DLCO , so it is possible that the severity of the disease before treatment may be more accurately reflected by DLCO than FVC in IPAF.
Among the follow-up patients, the levels of serum SP-A and KL-6 in patients with progressive disease were significantly higher after treatment, while the levels of serum SP-A and KL-6 were significantly lower in patients with improved condition, suggesting that the levels of serum SP-A and KL-6 may be effective serum biomarkers for monitoring IPAF activity. Arai et al  believe that the periodic measurement of KL-6 and SP-D levels will be useful in the evaluation of disease behavior and treatment response of idiopathic fibrotic nonspecific interstitial pneumonia.
Herein, we used Spearman's correlation test to study the correlation between Delta SP-A, Delta KL-6 and changes of lung function parameters (Delta DLCO, Delta FVC and Delta FEV1) to further explore the role of SP-A and KL-6 in the monitoring of prognoses in patients with IPAF. The results showed that Delta SP-A and Delta KL-6 were significantly negatively correlated with Delta DLCO, Delta FVC, and Delta FEV1 (P < 0.01). However, it is worth noting that in this study, serum SP-A and KL-6 levels before treatment were not related to FVC. Previous report also indicates that IPAF patient’s serum levels of both KL-6 and SP-D at baseline showed a negative correlation with %DLCO, but not with FVC . However, the change in FVC is still a reliable, effective, and responsive clinical status indicator that can be used as the primary endpoint of ILD crucial treatment studies. In the report of Lee et al. , serum SP-A and KL-6 levels in CTD-ILD were significantly negatively correlated with FVC and DLCO. This study has reported similar negative correlations between KL-6 and respiratory parameters . These studies also confirmed the relationship between SP-A and KL-6 with disease activity, suggesting that SP-A can also be used as an indicator to judge the prognosis of ILD. Therefore, we confirmed that the serum level of SPA and KL-6 reflected the severity of IPAF in terms of pulmonary function deterioration.
At present, the diagnosis of ILD mainly depends on HRCT or invasive transbronchial lung biopsy, which cloud be affected by various factors or cause great pain to patients . Although lung function test is known to be non-invasive, it is highly dependent on patient cooperation. Particularly, many elderly patients are usually unable to cooperate with the implementation of pulmonary function tests. Throughout the patient's disease, serum SP-A and KL-6 testing are easier to perform than frequent repeated lung function tests, x-rays, and invasive transbronchial lung biopsies . The correlation analysis of this study suggests that when the pulmonary function tests is difficult, we can make a preliminary evaluation and prediction of the patient's condition according to the expression level of the above two markers.
Our results also showed that although there was a significant correlation between Delta SP-A and Delta KL-6, the correlation coefficient was not high, suggesting that each marker may represent a different pathophysiological mechanism.
Our research had some limitations. First of all, due to the retrospective nature of the study, data related to symptom initiation and partial examination were missing. In addition, there was no analysis of chest HRCT manifestations, clinical symptoms and drug treatment. Second, the sample size of this study is small. Future studies with larger sample sizes are needed to verify this finding. Third, retrospective analysis of the course of disease does not allow changes in treatment and follow-up time, so it is difficult to avoid the impact of other confounding factors.
To sum up, this study showed that the levels of serum SP-A and KL-6 in patients with IPAF were significantly increased, which was negatively correlated with diffusion function. It was found that the levels of SP-A and KL-6 increased with the aggravation of the disease and decreased with the remission of the disease. As far as we know, we are the first to report the changes of serum SP-A and KL-6 levels with disease in patients with IPAF.