Prevalence of corticosteroid-refractory/resistant (CS-R/R) pneumonitis
During the study period, 1187 patients receiving ICI therapy were entered into the REISAMIC registry. Among them, 583 experienced irAEs, including 61 with ICI-pneumonitis. Of the 48 patients with ICI-pneumonitis who were given corticosteroid therapy, five (10%) had CS-R/R disease; these five patients accounted for 0.42% of the ICI-treated population (5/1187). Of these five patients, two received second-line immunosuppressive therapy and were included in our study; the other three died of severe ICI-pneumonitis after either an infectious complication or limitation of life-sustaining treatments within a month after the pneumonitis diagnosis.
Immune checkpoint inhibitor-induced (ICI-) pneumonitis
We identified nine patients, five men and four women, with CS-R/R ICI-pneumonitis diagnosed between July 2015 and June 2021, five from the REISAMIC registry and four via the immunotoxicity panel (Figure 1).
One patient underwent chest radiotherapy four years before the diagnosis of ICI-pneumonitis. None had COPD. One patient with Hodgkin’s lymphoma had taken bleomycin, which had not induced interstitial pneumonitis.
Median time from ICI initiation to ICI-pneumonitis was 92 days (13–567). All nine patients had dyspnoea, three had a fever, one had new crackles, and none had a cough. The CTCAE grade was ≥3 in eight patients (Table 1).
To exclude others diagnosis, different exams were performed depending on the patient condition. In the four patients who underwent bronchoscopy, the lavage fluid showed median counts of 19% for lymphocytes, 19% for neutrophils, and 57% for macrophages. Microbial investigations in blood and BAL were negative. Echocardiography, performed in three patients, showed no evidence of heart failure.
We examined the CT images at the diagnosis of ICI-pneumonitis, which were acquired within 12 hours of patient consultation for breathing symptoms. In eight of the nine patients, CT images obtained before the onset of pneumonitis showed a normal pulmonary parenchyma. Figure 2 shows an example of CT changes in corticosteroid-refractory pneumonitis. The CT abnormalities involved over 30% of the parenchyma in eight patients. Some patients had more than one elementary lesion and/or pattern. The most frequent elementary lesions were ground glass opacities (87.5%), reticulations (75%), septal thickening (75%), and condensations (62.5%). The main pattern was unclassified pneumonitis (75%) and hypersensitivity pneumonitis (25%). Pleuritis was noted in three patients.
Table 1. Characteristics of patients with ICI-pneumonitis treated with corticosteroids or IS treatment included in the study
|
Corticosteroid-resistant/refractory
pneumonitis
n=9
|
Corticosteroid-responsive pneumonitis
n=43
|
|
|
Immunosup-pressive therapy
n=6
|
No immuno-suppressive therapy
n=3
|
|
Patient characteristics
|
|
|
Median age, years (range)
|
69 (55–79)
|
65 (29–87)
|
Smoke exposure, n (%)
|
7 (78)
|
26 (60)
|
Prior pulmonary fibrosis, n (%)
|
1 (11)
|
0 (0)
|
Prior auto-immune disease, n (%)
|
1 (11)
|
4 (9)
|
Immune check point inhibitor
|
*
|
|
|
PD1
|
5 (83)
|
3 (100)
|
38 (88)
|
PDL1
|
1 (16)
|
|
5 (12)
|
Types of cancer
|
|
|
**
|
NSCLC
|
3 (50)
|
3 (100)
|
28 (65)
|
Hodgkin lymphoma
|
1 (17)
|
|
1 (2)
|
Melanoma
|
|
|
11 (26)
|
Breast cancer
|
1 (17)
|
|
-
|
Urothelial carcinoma
|
1 (17)
|
|
-
|
CTCAE grade at diagnosis, n (%)
|
|
|
|
1
|
|
|
1 (2)
|
2
|
1 (17)
|
|
19 (44)
|
3
|
3 (50)
|
3 (100)
|
14 (33)
|
4
|
2 (33)
|
|
9 (21)
|
Corticosteroid treatment
|
|
|
|
Intravenous methylprednisolone bolus, n (%)
|
5 (83)
|
2 (67)
|
11 (26)
|
Oral prednisone dose, mg, median (range)
|
80 (35–160)
|
80 (80–120)
|
60 (35–120)
|
Immunosuppressive therapy, n (%)
|
|
-
|
-
|
Cyclophosphamide
|
4 (67)
|
-
|
-
|
Infliximab
|
1 (17)
|
-
|
-
|
Intravenous immunoglobulins
|
1 (17)
|
-
|
-
|
Response to immunosuppressive therapy
|
|
-
|
-
|
Improved
|
2 (33)
|
-
|
-
|
Worsened
|
4 (67)
|
-
|
-
|
CTCAE grade with treatment, n (%)
|
with immunosuppres-sant
|
with corticosteroid
|
with corticosteroid
|
1
|
2 (33)
|
|
32 (74)
|
3
|
|
|
11 (26)
|
4
|
1 (17)
|
|
|
5
|
3 (50)
|
3 (100)
|
|
Cause of death, n (%)
|
***
|
***
|
|
Pneumonitis
|
3 (50)
|
3 (100)
|
|
Cancer
|
1 (17)
|
|
21 (49)
|
Other irAE
|
1 (17)
|
|
|
|
CTCAE: Common Terminology Criteria for Adverse Events; irAE: immune-related adverse event; NSCLC: non-small-cell lung cancer
* Of which 3 combinations: Pembrolizumab and Permetrexed, Pembrolizumab and Trastuzumab, Atezolizumab and Carbozantinib
** Including one renal cancer, one colon and one thyroid
*** Median follow-up time 27 days (8 – 916)
Corticosteroid therapy
Table 1 shows the main features of corticosteroid therapy and Figure 3 the time intervals separating events. A single patient (#1) had steroid-resistant pneumonitis with a relapse at 15 mg/day of prednisone. Compared to the other patients, his dyspnoea was less severe (CTCAE grade 2) and the time from ICI initiation to ICI-pneumonitis was longer (169 days vs. a median of 83 days).
Of the 48 patients with ICI-pneumonitis in the REISAMIC registry, 43 responded to corticosteroid therapy, of which 32 experienced complete resolution of the lung disease (Table 1).
Patients given immunosuppressive treatment
Six patients, described in Table 1, received immunosuppressive or immunomodulatory treatment. All six patients were treated with a PD-1 or PD-L1 inhibitor and had received at least one line of systemic chemotherapy. ICI was third-line treatment for patient #1 and fourth-line treatment for patient #4. ICI was combined with chemotherapy in first-line treatment for patient #2, #3, #5, #6.
Two patients (#2 and #4, Figure 3) required mechanical ventilation and two others (#5 and #6) standard oxygen supplementation before immunosuppressive therapy initiation. Four patients received 600 mg/m2 of intravenous cyclophosphamide (patients #1, #2, #3, and #6). The median number of cyclophosphamide doses was 3.5 (1–6). Patients #1 and #3 improved after 28 and 24 days, respectively. Neither had required supplemental oxygen before immunosuppressive therapy. One died because of cancer progression five months after cyclophosphamide therapy and the other was alive on day 598. Neither had relapsing pneumonitis after cyclophosphamide discontinuation.
Patients #2 and #6 failed to improve and died a few days after cyclophosphamide initiation due to the pneumonitis, despite administration of ruxolitinib, tocilizumab, and mechanical ventilation in patient #6.
Of the remaining two patients, one (#4) was given infliximab and one (#5) intravenous immunoglobulin. Neither improved. Patient #4 died due to ICI-induced Lambert-Eaton myasthenic syndrome with diaphragm paralysis and patient #5 died of ICI-pneumonitis. The two months mortality rate was 67% (6/9).
No infectious complications of immunosuppressive therapy were recorded.