Identifying risk factors and biomarkers for severe CIP in lung cancer patients– a retrospective case series study

ABSTRACT

Background

Immune checkpoint inhibitors (ICIs) have revolutionized cancer therapy, but they can induce immune-related adverse events, including immune checkpoint inhibitor-associated pneumonia (CIP), a severe lung complication. CIP, particularly Grades 3–4, is associated with poor prognosis, indicating a critical need for research on this issue. Our study aimed to investigate the risk factors and biomarkers associated with severe CIP in lung cancer patients treated with ICIs, where OS represents overall survival and PFS denotes progression-free survival.

Methods

We conducted a retrospective analysis of 106 lung cancer patients with CIP at the First Affiliated Hospital of Zhejiang University from 2019 to 2023, categorized into four severity grades.

Results

The median time to onset of CIP was 5.17 months. Patients with Grade 3–4 CIP had a median PFS of 6.5 months and OS of 11.2 months. Univariate analysis identified phosphocreatine kinase below 61.5 U/l, Forced Vital Capacity (FVC) below 1.96, and BMI below 21.26 as predictive factors for Grades 3–4 CIP. Multivariate analysis confirmed that a decreased FVC was a significant predictor.

Conclusions

A decreased FVC below 1.96 emerged as a predictive factor for Grades 3–4 CIP, highlighting the importance of monitoring FVC in patients receiving ICIs.

Plain Language Summary

Our study sheds light on the potential risks of a serious lung condition called checkpoint inhibitor pneumonitis (CIP) that can occur in patients treated with immune-boosting drugs known as immune checkpoint inhibitors (ICIs). We found that a simple breathing test, measuring something called forced vital capacity (FVC), can predict which patients might develop severe CIP. A low FVC reading suggests that these patients should be monitored closely or treated with extra care. By understanding this, doctors can better protect patients and help them make informed decisions about their treatment. Our findings also open the door for further research to improve the safety and effectiveness of these life-saving drugs.

1. Introduction

Immune checkpoint inhibitors (ICIs), including antibodies that target programmed cell death protein 1 (PD-1) or its ligand PD-L1, have the potential to reactivate the anticancer activity of dysregulated T cells, leading to the elimination of tumor cells. However, this reactivation of T cells can also trigger undesirable side effects, such as inflammation in multiple organs, which are collectively termed as immune-related adverse events (irAEs) [1]. Among these, immune checkpoint inhibitor-related pneumonitis (CIP) is a prevalent pulmonary complication encountered in patients receiving ICI treatment [2]. While clinical trials have indicated a relatively low incidence of CIP, approximately 3‒5%, in the context of non-small cell lung cancer treatment, real-world clinical studies have reported higher rates, ranging from 5 to 19% [2,3]. Consequently, it is crucial to identify risk factors associated to CIP. To date, research on ICI-related complications has predominantly relied on retrospective analyses with small sample sizes, as well as meta-analyses and systematic reviews [4,5]. The clinical manifestations of CIP are highly variable, encompassing asymptomatic cases to severe respiratory complications, including dyspnea, cough, respiratory failure, and, in some instances,death [6]. Several risk factors have been implicated in CIP, such as age, smoking history, preexisting lung disease, and prior chest radiation [2,7,8]. This study conducts a retrospective review of clinical data and treatments outcomes in lung cancer patients undergoing ICI therapy. Additionally, it investigates the influence of cytokines on the risk of developing Grade 3–4 CIP.

2. Methods

2.1. Patients

The study presents a retrospective analysis of the medical records of 2673 cancer patients were administered Immune checkpoint inhibitors (ICIs), specifically targeting the PD-1/PD-L1 pathways, at the First Affiliated Hospital of Zhejiang University from 2019 to 2023. Among these, 106 cases of Immune checkpoint inhibitor associated pneumonia(CIP) were identified in patients diagnosed with lung cancer (see Figure 1). Comprehensive Data, encompassing patient demographics, clinical features, and survival outcomes, were extracted from the electronic medical records. The research was conducted in accordance with the ethical standards of the Helsinki Declaration, as revised in 2013, and received approval from the Ethics Committee of the First Affiliated Hospital of Zhejiang University (Approval No: ZJU1AE2023–0520-Quick). Given the retrospective nature of the study, written informed consent from the the participants was not necessary. and this requirement was waived by the Ethics Committee of the First Affiliated Hospital of Zhejiang University.

Figure 1.

Figure diagram of patient enrollment. CIP, immune checkpoint inhibitor-related pneumonitis.

3. Data collection

Clinical data for all enrolled patients were meticulously collected, encompassing a range of parameters: age at the time of treatment initiation, gender, smoking history, Eastern Cooperative Oncology Group Performance Status (ECOG PS), tumor-lymph node-metastasis (TNM) staging in accordance with the American Joint Committee on Cancer (8th edition), histological findings, PD-L1 expression levels, treatment protocols, cytokine levels both pre- and post- CIP, T, B, and natural killer cells (TBNK cells), imaging outcomes, pulmonary function tests results, clinical symptomatology, complete blood counts, biochemical markers, glycosylated hemoglobin levels, cortisol function and thyroid function, among other pertinent variables. Trained healthcare professionals meticulously recorded and verified the clinical data for each participant. The follow-up period was extended up to and including 1 June 2023. Progression-free survival (PFS) is defined as the interval from the commencement of treatment to the first occurrence of disease progression in any form or death from any cause. Overall survival (OS) denotes the time span from the start of treatment to the time of death due to any reason.

4. Diagnosis of CIP

Immune checkpoint inhibitor-associated pneumonia(CIP) was identified through chest imaging, which revealed new infiltrates post-ICI treatment. The diagnosis process involved ruling out alternative causes such as new lung infections and tumor progression and required the presence of respiratory distress and/or additional symptoms such as cough and exertional dyspnea [2]. Given that CIP is a diagnosis of exclusion, each case was meticulously reviewed to ensure diagnostic accuracy. Patients with definitive alternative diagnoses, including progressing cancer, lung infection, heart failure, and radiation pneumonitis were excluded from the study [9,10]. Researchers conducted a thorough review of chest CT scans for all participants, considering both disease progression and clinical presentation. Patients who met the aforementioned criteria were classified into the CIP group. For cases with a suspected diagnosis of CIP,a consensus was reached through discussion among two leading respiratory physicians and one radiologist. The National Comprehensive Cancer Network (NCCN) guidelines were followed to categorize CIP severity based on clinical and radiological features [9]. The grading is as follows: Grade 1 indicates asymptomatic cases with lesions limited to one lobe or less than 25% of the lung parenchyma. Grade 2 denotes new or severe respiratory symptoms, such as cough, shortness of breath, fever, chest pain, and an increased need for oxygen; Grade 3 is characterized by severe symptoms affecting all lung lobes or more than 50% of the lung parenchyma, significantly impeding daily activities; Grade 4 represents life-threatening respiratory impairment. These criteria were applied for the grading of CIP in subsequent analyses.For the sake of statistical analysis, we have grouped Grade 1 and Grade 2 CIP cases under the category of “mild CIP” Correspondingly, Grade 3 and Grade 4 CIP cases have been consolidated and are hereby referred to as “severe CIP.”

5. Statistical analysis

Descriptive statistics were employed to summarize the medical histories and clinical parameters of the study cohorts. For categorical variables, counts and percentages were generated, while the mean and standard deviation (SD) were calculated for continuous variables. Categorical data were analyzed using Fisher’s exact test or Pearson’s χ2 test, as appropriate.Continuous variables were assessed using the Mann-Whitney U test. One-way ANOVA was utilized for multiple comparisons, and non-parametric analyses, specifically rank-sum tests, were selected for multiple comparisons and non-parametric analyses, specifically rank-sum tests, were selected for analyzing data that were heteroscedastic or not normally distributed. The Kaplan-Meier analysis was applied to estimate the time from the initiation of immune checkpoint inhibitors to the occurrence of CIP, as well as the median survival and progression-free survival of CIP patients. Univariate logistic regression was conducted to evaluate the significance of individual factors in predicting CIP. Factors that were found to be statistically significant (p < 0.05) were subsequently included in a multivariate analysis to identify risk factors associated with CIP severity. The results are presented as Odds Ratios (OR) with their corresponding 95% Confidence Intervals (CI). Establishing causality using Directed Acyclic Graphs (DAGs). All statistical tests were two-tailed,and the significance level was set at 5%.

6. Results

6.1. Participant characteristics

This study encompassed a total of 2,673 cancer patients who had been treated with immune checkpoint inhibitors. Among them, 106 individuals with lung cancer developed checkpoint inhibitor pneumonitis (CIP) following treatment (as depicted in Figure 1). The patients were stratified into two distinct categories based on their clinical and imaging profiles: Grades 1 and 2, and Grades 3 and 4. Table 1 offers a detailed account of the demographic data, clinical features, and therapeutic approaches of these patients.

Table 1.

Analysis of clinical characteristics and treatment of grades 1-2 and grades 3-4 CIP.

Various	Grade1-2 N = 89%)	Grade3-4 N = 17%)	P-value
Gender(Male)	83(93.2)	16(94.1)	0.90
Age mean (SD),Y Min-Max	67.69(6.75) 52-86	68.47(5.91) 59-79	0.66
Smoking history(yes)	83(93.2)	16(94.1)	0.71
COPD	30(33.7)	6(35.2)	0.90
BMI mean(SD)	23.00(3.55)	20.33(3.14)	0.04
Stage
I-II stage	5(5.6)	1(5.8)	0.97
III stage	32(30.2)	4(23.5)	0.32
IV stage	52(58.4)	12(70.5)	0.35
Chemotherapy Line
First-line chemotherapy	75(84.2)	4(23.5)	0.67
Second-line chemotherapy	10(11.2)	13(76.4)	0.33
Second-line chemotherapy and above	4(4.4)	0	0.49
Combined medication
Single-agent	5(5.6)	3(17.6)	0.22
Combined chemotherapy	77(86.5)	12(70.5)	0.20
Combined with vascular-targeted therapy	2(2.2)	2(11.7)	0.44
Combined chemotherapy + vascular targeted therapy	5(5.6)	0	–
Pathology
Squamous cell carcinoma	58(56.1)	7(41.1)	0.03
Adenocarcinoma	16(17.9)	6(35.2)	0.35
Small cell lung cancer	11(12.3)	3(17.6)	–
Other neuroendocrine tumors	1(1.1)	1(5.8)	–
Adenosquamous carcinoma	3(3.3)	0	–
Metastatic sites
Brain	3(3.3)	1(5.8)	0.62
Lung	17(19.1)	3(17.6)	0.89
Liver	4(4.4)	2(11.7)	0.54
Bone	13(14.6)	5(29.4)	0.26
Combined　with radiotherapy	41(46)	4(23.5)	0.09

Chronic obstructive pulmonary disease (COPD) is a long-term lung condition characterized by difficulty in breathing and persistent cough. It is a major cause of morbidity and mortality worldwide; Checkpoint Inhibitor Pneumonitis (CIP): This term refers to a type of lung inflammation that can occur in patients treated with immune checkpoint inhibitors, a class of drugs used in cancer therapy; Body Mass Index (BMI): This is a numerical value derived from an individual’s weight in kilograms divided by the square of their height in meters. BMI serves as a widely recognized international standard for assessing body weight relative to height and is an indicator of overall health status. It is a crucial tool in identifying individuals who may be underweight, overweight, or have obesity, which can be associated with various health risks.

A higher incidence of CIP was observed in male patients, particularly in those with squamous cell carcinoma undergoing first-line therapy. Furthermore, the occurrence of CIP was frequently associated with concurrent chemotherapy. Notably, there were no significant disparities between the two severity groups (Grade 1–2 and Grade 3–4) regarding patient age, gender, smoking history, chronic obstructive pulmonary disease (COPD), sites of metastasis, and receipt of radiotherapy. However, significant differences were noted in the prevalence of squamous cell carcinoma (56.1% vs. 41.1%, p = 0.03) and body mass index (BMI) (23.00 ± 3.55 vs. 20.33 ± 3.14, p = 0.04).

7. Onset of CIP

The median duration from the commencement of immune checkpoint inhibitor (ICI) therapy to the development of checkpoint inhibitor pneumonitis (CIP) was determined to be 5.17 months (95% CI 4.61–5.72 months) (as illustrated in Figure 2(a)). Variations among subjects were significant. Patients with Grade 4 CIP exhibited the earliest median onset at 2.47 months (95% CI 0–5.12 months) (Figure 2(a)), correlating with an incidence rate of 4.7% (5/106) (Figure 2(c)). The majority of the patients were clinically categorized as Grade 2, constituting 58.5% of the cohort (62/106) (Figure 2(c)), with a median onset time of 4.64 months (95% CI 3.83–5.44 months) (Figure 2(a)). Grade 3 CIP had an incidence rate of 11.3% (12/106) (Figure 2(c)), with a median onset time of 5.17 months (95% CI 3.47–6.86 months) (Figure 2(a)). Upon conducting an analysis that grouped Grade 1 and Grade 2 with Grade 3 and Grade 4, no significant significant difference was observed in the timing of CIP development (5.37 months vs 3.83 months, p = 0.25)

Figure 2.

(a,b) Kaplan- Meier estimates for the initiation of immune checkpoint inhibitors to the occurrence of CIP. stratified by disease grade and showing medians and 95%CI. (c,d) Bar charts showing the clinical and imaging manifestations of CIP. (a) Relative frequencies of CIP grade. (b)Relative frequencies of clinical symptoms at the time of CIP diagnosis.

8. Clinical symptoms of CIP

As depicted in Figure 2(c), over two-thirds of the patients were diagnosed with Grade 1–2 CIP, with Grade 1 pneumonia occurring in 25.4% (27/106) and Grade 2 CIP in 58.4% (62/106). Only a small proportion of individuals developed Grade 4 CIP (4.7%, 5/106), while 12 patients had Grade 3 CIP (11.3%, 12/106). Approximately 28.3% (30/106) of CIP patients were diagnosed using radiological findings only, as these patients were asymptomatic at the onset of the disease. The common clinical symptoms of CIP included cough (42.5%, 45/106), shortness of breath (50.9%, 54/106), thoracodynia (5.7%, 6/106), fever (25.5%, 27/106), and fatigue (19.8%, 21/106, Figure 2(c)). The major clinical symptoms observed were cough and shortness of breath (Figure 2(c)).

9. The relationship between CIP severity and patient prognosis

The outcomes of the group analysis based on The outcomes of CIP severity are presented in Figure 3. The median progression-free survival (mPFS) for all CIP patients was 8.37 months (95% CI 7.29–9.44 months), as shown in Figure 3(a), The median overall survival (mOS) was 24.18 months (95% CI 21.17–27.18 months), The median overall survival (mOS). Among the 89 patients with Grade 1-2CIP, 75 experienced disease progression, while of the 17 patients with Grade 3–4 CIP, 16 showed progression. Although difference in the median PFS (mPFS) between Grade 1-2 and Grade 3-4, this did not reach statistical significance (mPFS 8.47 months vs. 6.50 months, HR = 3.36, log-rank p = 0.07), as illustrated in Figure 3(a). Among the 89 Grade 1–2 patients, 32 succumbed to the disease, and and of the 9 Grade 3–4 patients, 3 died. The median OS (mOS) was significantly shorter for Grade 3–4 patients (mOS 23.15 months vs. 11.20 months, HR = 8.97, log-rank p = 0.003), as shown in Figure 3(b).

Figure 3.

The Kaplan-Meier approach was used to estimate the progression-free survival (PFS) and overall survival (OS) in patients of different clinical grades. PFS: TFrom using immune checkpoint inhibitors to tumor progression; OS: the time from the use of immune checkpoint inhibitors to death from any cause.

In the subgroup of 27 Grade 1 CIP patients, 23 experienced disease progression, and 10 died. Their median progression-free survival (mPFS) was 10.7 months (95%CI 7.87–13.53 months) and and median overall survival (mOS) was 24.4 months (95%CI 23.15‒25.65 months), as indicated in Figure 3(c,d), respectively. Among the 62 Grade 2 patients, 52 had disease progression, and 22 died. Their mPFS was 7.94 months (95%CI 6.58–9.22 months) and the mOS was 31.17 months (95%CI 14.39–47.95 months) as seen in Figure 3(c,d). For the 12 Grade 3 patients, 11 showed disease progression, with 6 died. Their mPFS was 7.95 months (95%CI 5.20–9.73), and the mOS was 19.8 months (95%CI 6.20–33.40 months), as represented in Figure 3. All five Grade 4 patients developed disease progression, with 3 deaths. Their mPFS was 3.83 months (95% CI 7.29–9.44 months), and the mOS were 4.03 months (95% CI 3.15–4.92 months) as detailed in Figure 3(c,d).

Previous studies have indicated that factors such as radiation therapy, preexisting chronic obstructive pulmonary disease (COPD), a reduced forced expiratory volume in one second (FEV1) to forced vital capacity (FVC) ratio, and elevated ferritin levels are likely contributors to the severity of immune checkpoint inhibitors. Utilizing data from Tables 1 and 2, we designated FVC reduction as the exposure factor and the severity of CIP as the outcome measure. Our analysis encompassed variables such as the FEV1/FVC ratio, COPD, age, smoking history, low body mass index (BMI), elevated ferritin, PD-L1 expression, radiotherapy, and squamous cell carcinoma. We graphically represented these relationships in a directed acyclic graph, as depicted in Figure 4. It was identified that age, low BMI, radiotherapy, squamous cell carcinoma, and smoking history are confounding factors that require adjustment in our analysis.

Table 2.

Logistic regression analyses of potential risk factors for grade 3-4 CIP.

	Univariate analysis			Multivariate analysis
Various	ORs	95%CI	P-value	ORs	95%CI	P -value
Non-adenocarcinoma vs. Adenocarcinoma	2.01	0.66-6.14	0.66
Non-squamous cell carcinoma vs. Squamous cell carcinoma	0.32	0.11-0.93	0.32
Non-IV stage vs. IV stage	1.71	0.55-5.26	0.35
Non-1st line vs. 1st line	0.44	0.17-2.13	0.61
BMI	0.81	0.66-0.99	0.04
*BMI <21.26 VS BMI > 21.26	0.11	0.02-0.59	0.01
Non-monotherapy vs. Monotherapy	3.6	0.77-16.78	0.10
Non-combined chemotherapy vs. Chemotherapy	0.37	0.11-1.25	0.11
No radiotherapy vs. Radiotherapy	0.36	0.11-1.19	0.09
Whether coronary heart disease	3.60	0.92-14.05	0.07
Whether COPD	1.07	0.36-3.18	0.90
PDL1 low expression or 0 vs. high expression	0.50	0.31-11.52	0.50
Lymphocytes	0.74	2.26-2.17	0.59
CRP	0.99	0.97-1.02	0.48
Eosinophils	5.41	0.96-30.58	0.06
phosphocreatine kinase	0.98	0.96-1.00	0.04
*phosphocreatine kinase <61.5 U/l VS phosphocreatine kinase >61.5 U/l	0.27	0.08-0.90	0.03	0.20	0.01-2.88	0.24
Alanine transaminase	0.98	0.93-1.04	0.51
Total bilirubin	1.04	0.90-1.20	0.60
Ferritin	1.002	1.00-1.004	0.02
Albumin	0.98	0.88-1.09	0.98
IgE	1.002	0.99-1.01	0.65
FVC	0.18	0.03-0.96	0.04
*FVC <1.96 VS FVC > 1.96	0.03	0.003-0.365	0.006	0.03	0.03-0.39	0.007
FEV1%	0.99	0.93-1.04	0.58
FVC%	0.94	0.87-1.01	0.08
FEV1/FVC	1.08	0.96-1.20	0.19
DLCO%	0.95	0.89-1.01	0.11
Absolute values of T lymphocytes (CD3)	1.00	0.995-1.001	0.20
Absolute values of helper/inducer T lymphocytes (CD3+, CD4)	1.00	0.99-1.003	0.45
Absolute values of suppressor/cytotoxic T lymphocytes (CD3+, CD8+)	0.99	0.98-1.001	0.09

*Subsequent statistics when confirming that FVC reduction is the only independent risk factor for Grade 3–4 CIP, univariate analysis was performed on FVC < 1.96 VS FVC > 1.96, and multivariate logistic regression analysis was performed on FVC < 1.96 vs FVC > 1.96 and creatine kinase; Given that incorporating Body Mass Index (BMI) into the multivariate analysis would render the overall model invalid, we have elected to exclude BMI from the multivariate assessment. Forced vital capacity (FVC) refers to the maximum amount of air that can be exhaled as quickly as possible after inhaling as much as possible.

BMI (The Body Mass Index): The calculation formula is: BMI = weight ÷ height 2; FEV1: which stands for Forced Expiratory Volume in 1 second.; COPD: Chronic obstructive pulmonary disease; FEV1/FVC, also known as the one-second rate, denotes the ratio of the forced expiratory volume in the first second to the total forced expiratory volume; DLCO:Carbon monoxide diffusion capacity; IgE: Immunoglobulin E; CRP: C-reactionproteinPD-L1: Programmed cell death ligand 1.

Figure 4.

Directed acyclic graph (DAG) of checkpoint inhibitor pneumonitis (CIP) severity that illustrates the complex relationships between various factors and their potential influence on the severity of checkpoint inhibitor pneumonitis (CIP).

The analysis of risk factor associated with the severity of checkpoint inhibitor pneumonitis (CIP) is presented in Table 2. A single-factor logistic regression analysis identified phosphocreatine kinase (OR = 0.98, 95% CI 0.96–1.00, p = 0.04), forced vital capacity(FVC) (OR = 0.18,95%CI 0.03–0.96, p = 0.04), and body mass index (BMI) (OR = 0.81,95%CI 0.66–0.99, p = 0.04) as significant predictors of CIP severity (Table 2). The optimal cutoff values for these variables were determined using the receiver operating characteristic (ROC) curve (Figure 5). For FVC, the cutoff value was 1.96, with an area under the curve (AUC) of 0.8, 95% CI 0.55–1.0, p = 0.03, sensitivity of 80%, and specificity of 89%. For BMI, the cutoff value was 21.26, with an AUC of 0.71, 95% CI 0.56–0.80, p = 0.04, sensitivity of 72%, and specificity of 78%. The cutoff value for creatine phosphokinase was 61.5, with an AUC of 0.67, 95% CI 0.54–0.80, p = 0.03, sensitivity of 55%, and specificity of 75%. In the univariate logistic regression analysis, FVC less than 1.96 versus FVC greater than 1.96 was found to be statistically significant (OR = 0.03, 95% CI 0.003–0.365, p = 0.006). Similarly, for BMI less than 21.26 versus BMI greater than 21.26, and creatine phosphokinase less than 61.5 U/L versus creatine phosphokinase greater than 61.5 U/L, there were statistically significant differences, with P-values of 0.01 and 0.03, respectively (see Table 2, marked with *). Since low BMI is a confounding factor that needs adjustment for the severity of CIP (as indicated in Figure 4), it was excluded from the multivariate logistic analysis. A multivariate logistic regression was constructed to compare phosphocreatine kinase levels less than 61.5 U/L versus greater than 61.5 U/L, and FVC less than 1.96 versus greater than 1.96. The multifactorial logistic regression analysis revealed that an FVC less than 1.96 (OR = 0.03, 95% CI 0.03–0.39, p = 0.007) was a significant independent risk factor for CIP severity (Table 2).

Figure 5.

ROC of FVC & phosphocreatine kinase & bmi for Grade 3-4 CIP.

10. Follow-up treatment status of patients with CIP

Of the 106 patients diagnosed with CIP, 95 (95/106,89.6%) discontinued ICI therapy. Among the cohort of patients who discontinued therapy, a subset of 29 individuals (out of 95, representing 30.5%) continued to utilize immune checkpoint inhibitors (ICIS) subsequent to their treatment phase. In contrast, the majority, comprising 64 patients (out of 95, or 85.3%), opted not to pursue further ICIS therapy. Within this group of patients who chose to continue with ICIS, it was noted that a single case involved a patient who had been classified with Grade 3–4 checkpoint inhibitor-related pneumonitis (CIP), encountering a relapse of CIP upon the resumption of treatment. The remaining 28 patients were classified as Grade 1–2. An additional 3 patients could not be evaluated for continued medication due to loss of follow-up or death. Half (15/29, 51.7%) of the 29 patients who continued treatment for CIP developed a recurrence. Glucocorticoids were administered to 89.6% (95/106) of the total number of patients who developed CIP; 10.3% (11/106) did not undergo steroid therapy. Chest CT scans were performed approximately one month after the initiation of steroid treatment. The results indicated that 44.2% (42/95) of the patients had achieved radiographic resolution, 16.8% (16/95) had progressed, 25.2% (24/95) maintained radiographic stability, 1.0% (1/95) achieved complete radiographic resolution, and 12.6% (12/95) were unable to undergo radiographic evaluation due to loss to follow-up or death.

11. Discussion

Immune checkpoint inhibitors (ICIs), targeting PD-1, PD-L1, and CTLA-4, have significantly advanced cancer treatment but also introduced immune-related adverse effects (irAEs) [11]. By activating T cells against tumor antigens [12,13]. ICIs can disrupt immune tolerance, leading to various irAEs, including pneumonia, endocrine disorders, and colitis [11,12,14–17]. CIP is particularly concerning, being the most common fatal irAE associated with anti-PD-1/PD-L1 therapies, causing about 35% of treatment-related deaths [18]. Our study investigated risk factors associated with the progression of checkpoint inhibitor pneumonitis (CIP) to severe disease in 106 patients.

It is imperative to maintain vigilance for the clinical manifestations of checkpoint inhibitor pneumonitis (CIP), including nonspecific yet critical symptoms such as cough and dyspnea, which can be easily overlooked and may lead to delayed diagnosis and treatment [1,2,5]. Notably, our study revealed that 28.3% (30/106) of CIP patients were asymptomatic, highlighting the importance of imaging for CIP detection. However, radiological manifestations of CIP can mimic those of other pulmonary conditions, including infections, radiation pneumonia, heart failure, and tumor progression. necessitating further diagnostic procedures like pathological testing and bronchoscopy with bronchoalveolar lavage [2,5]. The results of our study highlight the critical relationship between the severity of CIP and patient outcomes.Specifically, patients with Grade 3–4 CIP who constitute a small fraction of the total patient population [5] (17/106, 16.04%, illustrated in Figure 3(b,d)), exhibit markedly poorer overall survival (OS). The overall survival rate for these patients was alarmingly reduced, with a median survival of merely 11.2 months, indicating a grave impact on their prognosis. This finding underscores the critical need for the early identification of patients at risk of developing severe CIP, allowing for timely interventions that may improve patient outcomes. Current discussions on the imaging manifestations of checkpoint inhibitor pneumonitis (CIP) and the risk factors associated with its severity have reached a level of maturity. However, due to the low incidence of this condition, only a few studies with small sample sizes have endeavored to uncover the factors influencing the severity of immunosuppressant-associated pneumonia.

Our investigation encompassed a cohort of over 2,000 cancer patients receiving immune checkpoint inhibitors (ICIs), culminating in the identification of 106 individuals with definitive checkpoint inhibitor pneumonitis (CIP). We meticulously analyzed the severity of the disease among these 106 CIP patients. Previous research has established that chest radiation, the concurrent use of epidermal growth factor receptor (EGFR) tyrosine kinase inhibitors with immune checkpoint inhibitors (ICIs), and preexisting pulmonary conditions, such as interstitial lung disease and chronic obstructive pulmonary disease (COPD), are associated with the development of checkpoint inhibitor pneumonitis (CIP) [19]. Previous chest radiation, lung disease, and combination therapy were identified as significant risk factors for the development of checkpoint inhibitor pneumonitis (CIP) following immune checkpoint inhibitor (ICI) therapy, with odds ratios of 3.33, 2.82, and 3.42, respectively [20]. And some previous studies suggesting a link between T lymphocyte ratios and CIP onset [21]. A recent study has identified older age, higher ECOG scores, and elevated D-dimer levels as common in patients with severe checkpoint inhibitor pneumonitis (CIP), along with a higher incidence of subsequent pulmonary infections. Despite these insights, the study’s small sample size precluded an analysis of risk factors for high-grade CIP [22]. This highlights the need for further research to clarify these risks.In our investigation, we incorporated these variables into our analysis of factors influencing the severity of checkpoint inhibitor pneumonitis (CIP). Nonetheless, we did not find a significant correlation among themIn our study, univariate logistic regression analysis identified phosphocreatine kinase, Forced Vital Capacity (FVC), and Body Mass Index (BMI) as potential predictors for severe Grade 3–4 checkpoint inhibitor pneumonitis (CIP). However, after employing directed acyclic graphs (DAGs) to clarify the relationships, we found that low weight was unlikely to be a direct causal factor and therefore was excluded from the multivariate regression analysis. Subsequent multivariate logistic regression analysis confirmed that only a reduced FVC, specifically below 1.96, was a significant risk factor for developing severe Grade 3–4 CIP. Despite this, our findings suggest that patients with a BMI below 21.26, indicative of low weight, should be closely monitored for the risk of severe CIP. Our findings carry substantial practical significance for clinical decision-making and corroborate existing medical knowledge. Firstly, assessing Forced Vital Capacity (FVC) is a straightforward procedure, requiring only a pre-treatment pulmonary function test for patients. For individuals with an FVC below 1.96, it is imperative to carefully balance the risks and benefits of treatment, ensuring they are fully apprised of the potential for developing severe checkpoint inhibitor pneumonitis (CIP). Furthermore, patients with low body weight must also be cautiously monitored for the risk of severe CIP, with medication dosages adjusted accordingly and close surveillance for any clinical signs related to CIP.While conditions like chronic obstructive pulmonary disease (COPD), prior chest radiation, and combination therapies might elevate the risk of developing checkpoint inhibitor pneumonitis (CIP), they do not seem to raise the risk of severe outcomes. Therefore, these factors should not preclude the use of immune checkpoint inhibitors (ICIs) but rather prompt close monitoring to facilitate prompt management should CIP occur.

While our retrospective study provides valuable insights, it is not without limitations, including the presence of multiple confounders and incomplete clinical data. Despite having a larger sample size than many contemporary studies, the findings should be interpreted with caution and require confirmation through further research, including prospective studies with larger cohorts. This is a direction we intend to pursue in our future work.

12. Conclusions

In conclusion, our study identifies forced vital capacity (FVC) below 1.96 as a predictor for severe checkpoint inhibitor pneumonitis (CIP), advocating for cautious ICI treatment in patients with low baseline FVC. Further research is needed to validate this biomarker and explore its mechanistic links to CIP severity.

Funding Statement

This work was supported by Zhejiang Provincial Clinical Research Center for Respiratory Disease [2022E50005 to Jianya Zhou]; Key R & D Program of Zhejiang Respiratory Disease [No. 2023C03069].

Article highlights

• Checkpoint inhibitor pneumonitis (CIP) is a rare but severe immune-related adverse event associated with immune checkpoint inhibitors (ICIs) therapy.

• Our study identified a low forced vital capacity (FVC) as a predictor for severe CIP, which can help in the early identification of patients at risk.

• Clinicians should consider FVC measurements before administering ICIs to identify patients who may require closer monitoring or alternative treatment strategies.

• Further validation of FVC as a predictive biomarker for CIP severity is needed, along with exploration of its biological mechanisms.

• The inclusion of a Plain Language Summary can improve article accessibility for patients and nonspecialist readers.

Disclosure statement

The authors have no relevant affiliations or financial involvement with any organization or entity with a financial interest in or financial conflict with the subject matter or materials discussed in the manuscript. This includes employment, consultancies, honoraria, stock ownership or options, expert testimony, grants or patents received or pending, or royalties.

Ethics approval and consent to participate

Ethics approval was granted by Committee of from the Ethics Committee of the First Affiliated Hospital of Zhejiang University (Approval No: ZJU1AE2023-0520-Quick). As the study was retrospective, any personal data (including any personal details, images, or videos) were not covered, and written informed consent from the subjects was not required.

Authors’ contributions

Conceptualization: [Jianying Zhou, Jianya Zhou] - Involved in the original idea and the development of the research project

Data Curation: [Guixian Wu, Binggen Wu, Ting Wang, Yuncui Gan, Nan Jiang, Yuekang Li] - Responsible for data collection, management, and quality control.

Formal Analysis: [Jingjing Qu, Jing Zheng, Guixian Wu] - Conducted the statistical analysis and computational modeling.

Funding Acquisition: [Jianya Zhou] - Secured the funding to support the research work.

Investigation: [Jianya Zhou] - Conducted the experiments and collected the data.

Methodology: [Jingjing Qu, Jing Zheng] - Developed the experimental design and methodology.

Project Administration: [Jianya Zhou] - Managed the project administration and supervision.

Resources: [Jinpeng Liu] - Provided essential resources and materials for the research.

Software: [Guixian Wu, Jingjing Qu, Jing Zheng] - Developed and maintained software used in the research.

Supervision: [Jianying Zhou, Jianya Zhou] - Oversaw and directed the research project.

Validation: [Jianya Zhou] - Ensured the validity of the experimental results.

Visualization: [Guixian Wu, Jingjing Qu, Jing Zheng] - Created visualizations and graphical representations of data.

Writing – Original Draft: [Guixian Wu, Binggen Wu] - Drafted the initial version of the manuscript.

Writing – Review & Editing: [Jingjing Qu, Jing Zheng] - Reviewed and edited the manuscript

Availability of data and materials

All data generated or analyzed are included in this published article. The datasets used and/or analyzed during the current study are available from the corresponding author upon reasonable request.