34 patients were diagnosed with CIP from July 2020 to December 2022 at Peking University First Hospital. Among them, 6 presented with grade 1 CIP and improved after discontinuation of immunotherapy. 6 patients were excluded due to the absence of bronchoscopy and 1 patient was excluded due to concurrent Covid-19 infection. Ultimately, 21 patients were included in this study. Figure 1 shows the flowchart of the overall screening process.
Baseline demographic characteristics are shown in table 1. The median age was 61 years old. The majority were male (85.7%). There were 11 cases of lung cancer, 3 Renal cell carcinoma, 2 bladder cancer, 1 esophageal cancer, 1 gastric cancer, 1 endometrial cancer, 1 head and neck squamous cell carcinoma, and 1 colon cancer. Among the 21 enrolled cases, PD-1 inhibitors were used in 16 (76.2%) cases. The median cycle of immunotherapy before the onset of CIP was 4 cycles and the median time from the onset of symptoms to diagnosis was 2 weeks. First-line immunotherapy was used in 17 (81.0%) patients. 9 (42.9%) patients received thoracic radiotherapy before the onset of CIP.
Cell differential count of BALF was performed in 19 patients, and lymphocyte subsets were tested in 15 cases. 13 patients underwent lung biopsy, 8 TBCB and 5 TBLB.
Factors associated with the severity of CIP
Of the 21 patients enrolled, there were 14 patients presenting with grade 2 CIP (defined as low-grade group), 7 presented with grade 3 or higher CIP (defined as high-grade group). Table 2 shows factors associated with the severity of CIP. The median age was 60 years old in the low-grade group and 65 years old in the high-grade group. Patients with previous ILD were more likely to develop high-grade CIP, compared with patients without previous ILD, 83.3%(5/6) versus 15.4%(2/15), respectively (P=0.006). No correlation was found between severity of CIP and smoking status, thoracic radiotherapy history, treatment line or tumor type. The median neutrophil lymphocyte ratio (NLR) was higher in the high-grade group than in the low-grade group, 6.33 versus 3.57, respectively (P=0.025). Cases in the high-grade group had lower lymphocytes and higher eosinophils, neutrophils, LDH and hsCRP, though no statistically significant difference was reached.
As for the cell differential of the BALF, the median lymphocyte percentage was higher in the high-grade group than low-grade group, 56.0% versus 21.5%, respectively (P=0.032). There was no significant correlation between BALF lymphocyte subsets and the severity of CIP. Table 4 shows binary logistic regression analysis of univariate factors related to the grade of CIP. Patients with previous ILD were more likely to develop high-grade CIP than patients without previous ILD (OR= 32.5, 95%CI 2.284-443.145, P=0.009). Higher BALF lymphocyte percentage was associated with higher possibility of high-grade CIP (OR =1.095, 95% CI 1.001, 1.197, P=0.047).
No relationship was found between pathological type and severity of CIP. In the high-grade group, only 3 cases underwent biopsy, with 2 cases of organizing pneumonia (OP) and 1 acute fibrinous and organizing pneumonia (AFOP). In the low-grade group, 10 cases underwent biopsy, with 3 cases of OP, 2 AFOP, 3 nonspecific interstitial pneumonia (NSIP), and 2 diffuse alveolar damage(DAD).
Risk factors for recurrence of CIP
After diagnosis, glucocorticoids were administered in 20 cases while 1 patient improved merely by ICI cessation. Among the 20 cases receiving steroid treatment, CIP showed rapid improvement in 18 patients, while 2 deteriorated despite timely and adequate treatment with steroids. Of the 2 steroid-resistant patients, one received infliximab but still progressed and eventually died, while the other received tocilizumab and clinically improved.
During steroid tapering, 5 (23.8%) patients experienced CIP recurrence, while 16 (76.2%) patients reported successful tapering without CIP recurrence. The median time to recurrence was 2 months after the administration of steroids. Median steroid dose at relapse was 10 mg. CIP recurrence was treated by increasing steroid dosage in combination with azathioprine in 2 patients, and increasing steroid dosage was given to 4 patients without adding any immunosuppressive agents.
Factors associated with CIP recurrence are shown in Table 3. The median age in the CIP recurrence (R-CIP) and non-recurrence groups was 54 and 61 years old, respectively (P=0.431). The LDH, hsCRP, lymphocyte, neutrophil, eosinophil count, and NLR had no significant difference between the two groups. The median BALF lymphocyte percentage was higher in the recurrence group than the non-recurrence group, 56.0% versus 20.0%, respectively (P=0.012). CD4/CD8 ratio ≧1 was found in 80%(4/5) of patients in the recurrence group but only 20%(2/10) in the non-recurrence group (P=0.047).
Table 5 shows binary logistic regression analysis of univariate factors related to recurrence of CIP. Higher BALF lymphocyte percentage was associated with higher possibility of CIP recurrence (OR =1.123, 95% CI 1.005, 1.255, P=0.040). Patients with BALF lymphocyte CD4/CD8 ratio ≧1 were more likely to develop recurrence than those with CD4/CD8 ratio <1 (OR=16.00, 95% CI 1.093,234.248, P=0.043). Due to limited sample size, multivariate factor regression analysis was not conducted.
As for CIP histologic features (Table 3), in the recurrence group, lung biopsy was performed in 3 patients, with 2 OP and 1 AFOP. Whereas in the non-recurrent group, of the 10 patients who underwent biopsy, 3 presented as OP, 2 AFOP, 3 NSIP and 2 DAD. Due to the limited sample size, no analysis of variance was performed.
Prognosis
The Kaplan-Meier curves of the low-grade CIP group and high-grade CIP group are shown in Figure 2. At the time of data collection, 6 patients had died, 1 of Grade 4 CIP, 1 of infection, and 4 of disease progression. By the end of the follow-up period, a trend toward a survival benefit in the low-grade group could be seen in the survival curves, however, the difference did not reach statistical significance (P=0.219). The median survival had not yet been reached. The Kaplan-Meier curves of the recurrence group and non-recurrence group are shown in Figure 3. The survival time had no significant difference between the recurrent group and non-recurrent group (P=0.596). And the median survival had not been achieved by the end of follow-up.
Safety of Lung Biopsy
There were 8 patients in this study who underwent TBCB. 3 developed pneumothorax after TBCB, and 2 required closed thoracic drainage. All the 3 patients recovered well without sequela. There were no procedure-related mortalities, prolonged air leak, severe bleeding, acute exacerbation, respiratory failure, or respiratory infection. Five patients in this study underwent TBLB, none had pneumothorax or hemorrhage that required intervention. Among the 13 patients who underwent lung biopsy, 2 patients died due to disease progression. The remaining 11 patients are still alive and in follow-up.