Idiopathic pulmonary fibrosis (IPF), which has the histological pattern of usual interstitial pneumonia (UIP), is an irreversible and progressive fatal interstitial lung disease with a poor prognosis.1 Unfortunately, there is no effective drug for the treatment of IPF, and the median survival is about 3–5 years following diagnosis. The developmental trajectories of IPF vary greatly and can be divided into three categories: stable or slow progress; Rapid deterioration; Steady but interspersed with periods of rapid progress.2 Currently, there is no widely accepted IPF disease monitoring system, so it is difficult for clinicians to predict the disease process of each patient. Therefore, the search for new biomarkers from the underlying mechanism may provide a new method for the prognosis evaluation of IPF.

Krebs von den Lungen-6 (KL-6) is a high-molecular-weight glycoprotein expressed on the extracellular surface of alveolar type II cells and bronchial epithelial cells.3 As a chemokine, it can promote migration, proliferation and survival of pulmonary fibroblasts.4 In interstitial lung diseases, KL-6 has been studied as a nonspecific marker of fibrosis. Previous study found that the serum KL-6 level of IPF patients was significantly increased compared with healthy volunteers or other patients with interstitial lung diseases.5 C-X-C motif chemokine 13 (CXCL13) is a key mediator that guides B cells to inflammatory lesions, which is significantly elevated in many autoimmune syndromes and is associated with clinical activity in these diseases.6,7 C-X-C chemokine receptor 5 (CXCR5) is the B cell ligand of CXCL13, which has the role of mediating chemotaxis of lymphocytes.8 Repeated antigen stimulation enhances B-cell differentiation, another typical characteristic of systemic lupus erythematosus, rheumatoid arthritis, and many other autoantibody syndromes, which are also clinically relevant to IPF.9 Abnormal B cells were recently found in the pathogenesis of IPF, suggesting that CXCL13 may be a biomarker related to IPF etiology and reflecting the progress of fibrosis. Previous studies have shown that serum KL-6 and CXCL13 can be used as biomarkers for the diagnosis of IPF and the judgment of prognosis. However, its sensitivity and specificity to the prognosis of IPF are still unclear, and most of them are retrospective small sample size studies, the clinical application value still needs to be further discussed.

In this study, we aimed to identify the relationship between serum KL-6 and CXCL13 levels, computer-aided diagnosis (CAD) system for automatic quantitative analysis and objective indicators of pulmonary function, and determined the role of complex indicators including KL-6 and CXCL13 levels, pulmonary function, and CAD system for automatic quantitative analysis as markers on reasonable prediction of disease progression and death risk of IPF.

Increased circulating KL-6 levels in interstitial pneumonia are thought to be due to increased permeability of KL-6 following alveolar wall alveolar II cell regeneration and/or disruption of the air blood flow barrier in the affected lungs, which is associated with disease progression corresponding to HRCT. The pro-fibrogenic and anti-apoptotic effects of KL-6 increased the maximum effective concentrations of platelet-derived growth factor (PDGF), basic fibroblast growth factor (BFGF) and transforming growth factor-beta (TGF-β). These findings suggest that elevated levels of KL-6 in epithelial fluid may stimulate the fibrosis process of interstitial lung disease.4 CXCL13 is a chemokines produced by follicular dendritic cells, which recruit B cells to become secondary and tertiary lymphoid structures by binding to the homologous receptor CXCR5.14 Studies have shown that the expression of CXCL13 and clinical relevance in IPF patients are far greater than in chronic obstructive pulmonary disease (COPD) patients, indicating that the generation of this chemokines is not purely caused by the lung injury itself, but specific to the disease process of IPF.15 Increasing researches have indicated that the serum KL-6 and CXCL13 levels in IPF patients are higher than other lung diseases and interstitial lung diseases patients.5,15–17 At present study, we found that serum KL-6 and CXCL13 were better biomarkers for predicting the outcome of IPF patients, and their baseline levels were consistent with pulmonary function and the degree of changes in lesions quantified by CAD system, which could improve the accuracy of survival assessment for IPF patients.

As a prognostic predictor, we attempted to predict pulmonary function deterioration by measuring baseline levels of serum KL-6 and CXCL13 in IPF patients. Some research suggests that a high level of KL-6 is often associated with a significant decline in FVC (ΔFVC and %ΔFVC, respectively).18,19 Vuga et al. mentioned that serum CXCL13 level was weakly correlated with DLCO and FVC levels in recent studies, but this study did not compare the correlation between serum CXCL13 level and the dynamic changes in pulmonary function.15 Our study showed that the baseline level of serum KL-6 was highly correlated with the annual changes of pulmonary function (FVC, DLCO and TLC). The baseline level of serum CXCL13 was moderately or highly correlated with the annual changes of pulmonary function (FVC, DLCO and TLC), suggesting that IPF patients with higher serum kl-6 and CXCL13 baseline levels were more likely to have progressive deterioration of pulmonary function. Studies have shown that it is more accurate to reflect the prognosis of IPF according to the degree of reduced pulmonary function than histologic differences.20 Flaherty et al. demonstrated that short-term changes of FVC have a strong prediction for long-term survival of patients diagnosed with UIP, thus, baseline levels of serum KL-6 and CXCL13 may be markers of IPF progression.21

As HRCT can accurately reflect the pathological changes of IPF lesions, it is commonly used as a non-invasive assessment method for patients with diffuse infiltrative pulmonary disease.22 Sakamoto et al. reported that the correlation between KL-6 level and HRCT abnormal fibrosis progression in fibrous NSIP patients was statistically significant.23 However, there are few studies on the correlation between serum CXCL13 level and imaging.

In previous studies, we have found that the reticular lesions and HC lesions corresponding to fibrosis are the most important predictors for the prognosis of IPF patients, and the detection of clinical progression of overall ILD lesions is an important approach to reduce mortality. Therefore, our single-center prospective cohort study assessed the correlation between baseline levels of serum KL-6 and CXCL13 and HRCT disease dynamics.

Our study showed that the extent of volume change of the reticular lesions and overall ILD lesions was highly correlated with the baseline levels of serum KL-6 and CXCL13. And the volume change of HC lesions was highly correlated with the baseline level of serum KL-6, and moderately correlated with the baseline level of serum CXCL13. The high correlation between the serum KL-6 and CXCL13 baseline levels of IPF patients and the progression of HRCT lesions may reflect the severity of pulmonary fibrosis in advance and predict disease progression. Combined with the quantitative evaluation of HRCT lesions by CAD system, it is helpful to accurately select IPF patients with rapid progression clinically.

The heterogeneous characteristics of IPF make the survival time of patients have great individual differences.24 Recent studies have found that serum KL-6 and CXCL13 are associated with IPF patient's survival. In a prospective study involving 152 patients with idiopathic interstitial pneumonia (including IPF), patients with serum KL-6 level >1000U/ml had worse prognosis than patients with lower serum KL-6 level.25 In addition, Yokoyama et al. also reported that in a retrospective study of 27 IPF patients, the median survival time of patients with serum KL-6 level>1000U/ml was significantly reduced.19 The CXCL13 can be used as a biomarker for the prognosis of IPF patients.15,17 IPF patients with high levels of CXCL13 are often associated with pulmonary hypertension and respiratory failure.15 In our study, patients who died during five years of observation had significantly higher baseline levels of serum KL-6 and CXCL13 than those who survived. ROC curve analysis confirmed that the serum KL-6 baseline level was higher than the cut-off value 800.0U/ml and the serum CXCL13 baseline level was higher than the cut-off value 62.0pg/ml and had high sensitivity and specificity for the prediction of mortality risk. In addition, we found that the serum markers (including KL-6 and CXCL13), pulmonary function (including FVC, DLCO and TLC) and HRCT evaluation (including reticular lesions, HC lesions, and overall ILD lesions) have good sensitivity and specificity for the determination of survival rate (AUC>0.8). Among them, KL-6 has the maximum area under the curve (AUC 0.955, P<.001, 95% CI: 0.918, 0.941), so it has a strong predictive value for the poor prognosis of IPF patients.

Some scholars proposed that among the IPF patients with stable condition, serum KL-6>1000U/ml indicated poor prognosis.19,25 Hamai et al. reported that 1040U/ml was used as the cut-off value of serum KL-6 of 65 IPF patients, and the 5-year survival rate of the patients in the high level group was poor.16 Currently, no study has clearly shown the optimal cut-off value of CXCL13 for the survival of IPF patients, but the high level of CXCL13 is correlated with the low survival time.15,17 In our study, 800.0U/ml and 62.0pg/ml were used as the cut-off value of serum KL-6 and CXCL13, respectively. The difference in KL-6 cut-off value between our study and other studies may be related to the group selection bias, treatment differences and different sample sizes, prospective and large sample size of IPF study can further verify our experimental results. Following the kaplan–Meier method was used to estimate the survival rate, and it was found that the survival time between the survival group and non-survival group was statistically significant. The median survival time of patients with low level KL-6 was over 56 months, while patients with high level KL-6 were only 23 months. Median survival was more than 53 months for patients with low levels of CXCL13, while patients with high level CXCL13 were only 26 months. We research showed that the baseline levels of serum KL-6 and CXCL13 had a good correlation for lifetime prediction. Our results are different from other studies on cut-off value of KL-6, which may be related to the selection bias, treatment difference and sample size of the enrolled population. Prospective IPF studies with large sample size can further verify our experimental results.

In conclusion, measurement of baseline levels of specific serum markers KL-6 and CXCL13, radiologic quantification, and longitudinal follow-up and analysis of pulmonary function can all predict poor prognosis, such as disease deterioration and mortality risk in IPF patients, and also can assist in stratified management of IPF patients and selection of clinical strategies. As a non-invasive marker, its clinical value and its association with long-term survival rate of patients still need to be confirmed by further studies on multi-center and expanded number of cases.

The authors declare that they have no conflicts of interest.