Baseline characteristics of patients and comparison of study groups
One hundred eighty-one IPF patients from the ILD clinic registry were screened for eligibility (figure 1). Overall, twenty-nine patients were excluded for unavailability of CT scan within 12 months from diagnosis (n=21), history of thoracic malignancy (n=3), respiratory infection at the time of diagnosis (n=1) or for missing reports of baseline pulmonary function tests (n=4). One hundred fifty-two IPF patients in the ILD registry were therefore included in the analysis. Average clinical follow up time of the study population was 19.08 months (SD 12.8). The majority of IPF patients (n=94, 62%) had at least one LNE on CT performed within a year from diagnosis, while 58 patients (38%) had no LNE present. Within patient with mediastinal LNE, 62 (66%) had two or more mediastinal lymph node with short-axis diameter ≥ 10 mm, while 31 (33%) had 3 or more LNE. In this group, the mean short-axis diameter of largest lymph node was 12.4 mm (SD 2.4). LNE was predominantly distributed in paratracheal stations (n=93, 99% of LNE patients), while only 3 patients (3%) had LNE in the lower zones. The large majority of patients included in this study (n=135, 89%) received antifibrotic treatment for IPF during the follow up.
Baseline characteristics of the study groups by presence of LNE on baseline CT scan are reported in table 1. No significant differences were found at baseline between the two study groups as to demographics. Patients without LNE had lower GAP stage (p=0.038) and higher DLco % predicted at baseline (p=0.023) as compared to LNE+ patients. Diabetes was a more frequent comorbidity among patients with LNE (25.8% vs 6.9%, p=0.004).
On univariate Kaplan-Meyer analysis, IPF patients with evidence of mediastinal LNE on baseline CT scan of the chest had lower survival rates as compared to patients without LNE (median survival 37.8 months vs 44.5 months, log-rank p=0.025) (figure 2). This corresponded to a 2.65-fold increased risk of mortality for LNE+ patients on Cox proportional hazard analysis (95% CI 1.09-6.46, p=0.032) (table 2). After stratification of the study population by number of enlarged mediastinal lymph nodes, the involvement of three or more lymph nodes was found significantly associated with worse survival as compared to each of the other groups (Figure 2), with a 5.72-fold increased risk of mortality (HR 5.72, 95%CI 2.18-14.98, p<0.001). Within patients with mediastinal LNE at baseline, increasing dimensions of the largest lymph node were also associated with higher mortality (HR 1.17, 95% CI 1.04-1.32, p=0.01).
When hazard ratios were adjusted for GAP stage, only the presence of 3 or more mediastinal enlarged lymph nodes maintained a strong association with poorer survival (HR 5.03, 95% CI 1.86-13.62, p≤0.001)(table 2).
Time to significant functional decline defined as 10% absolute decline in FVC was not significantly different between patients with or without LNE (figure 3, table 2). On the other hand, patients with evidence of 3 or more mediastinal enlarged lymph nodes showed increased disease progression rates both on univariate analysis (HR 2.67, 95% CI 1.19-5.97, p=0.017) and after adjusting for GAP stage (HR 2.99, 95% CI 1.22-7.33, p=0.17)(table 2).
In order to further explore the association between mediastinal LNE and change in pulmonary function parameters, the annualized rates of change in absolute FVC and % predicted DLco were compared between patients with different degrees of lymph node involvement at baseline (figure 4, table 3). The rate of decline in FVC increased with the number of mediastinal lymph nodes involved and was largest in patients with 3 or more enlarged lymph nodes (-178 mL, SE 0.09), especially if compared to patients without LNE, who showed relative stability over 12 months (FVC change -4 mL, SE 0.07). However, the differences in the rates of change in FVC were not statistically significant as shown by the group-by-time interaction term in the model (p=0.332). The largest decline in DLco occurred in the group with 3 or more enlarged lymph nodes (-10.5%, SE 2.55) as compared with the other groups, although such difference did not meet statistical significance (group-by-time interaction p=0.349) (table3).
Longitudinal assessment of mediastinal LNE
117 patient cases who had available longitudinal imaging data were assessed for the presence of mediastinal LNE at the follow up CT scan. Average time difference between consecutive CT scans was 17.7 months (SD 10.4). LNE status did not change over time in 101 (86%) of these patients. Among the 48 patients without evidence of mediastinal LNE at baseline, 38 (79%) patients remained without mediastinal LNE at follow up, while among the 69 patients with baseline mediastinal LNE 63 (91%) maintained evidence mediastinal LNE over time. Among the 62 patients with baseline LNE who received antifibrotic treatment, 57 (92%) had mediastinal LNE at follow up. Mediastinal LNE disappeared in only 6 (9%) patients who had LNE at baseline, while 10 (11%) patients who did not have LNE at baseline developed LNE at follow up.