---
title: "Corticosteroid-refractory checkpoint inhibitor pneumonitis in lung cancer: heterogeneity and mecha"
id: "frontiers-in-immunology-11-corticosteroid-refractory-checkpoint-inhibitor-pneumonitis-in-lung-cancer"
canonical_url: "https://medichelpline.com/clinical-feed/frontiers-in-immunology-11-corticosteroid-refractory-checkpoint-inhibitor-pneumonitis-in-lung-cancer"
content_type: "clinical_feed_article"
specialty: "Infectious Disease"
source_name: "Frontiers in Immunology"
source_url: "https://www.frontiersin.org/articles/10.3389/fimmu.2026.1983756"
published_at: "2026-10-28T00:00:00.000Z"
evidence_level: "Journal Feed"
license: "CC-BY-NC-4.0 / Informational Use"
---
# Corticosteroid-refractory checkpoint inhibitor pneumonitis in lung cancer: heterogeneity and mecha
## Provenance & Clinical Metadata
- **Canonical URL:** https://medichelpline.com/clinical-feed/frontiers-in-immunology-11-corticosteroid-refractory-checkpoint-inhibitor-pneumonitis-in-lung-cancer
- **Specialty:** [Infectious Disease](https://medichelpline.com/clinical-feed/infectious-disease.md)
- **Primary Source:** Frontiers in Immunology
- **Source URL:** [Original Journal Publication](https://www.frontiersin.org/articles/10.3389/fimmu.2026.1983756)
- **Published At:** 2026-10-28T00:00:00.000Z
- **Evidence Rating:** Journal Feed
## Executive GIST (TL;DR)
- Corticosteroid-refractory **checkpoint inhibitor pneumonitis** (CIP) denotes persistent or worsening respiratory disease despite approximately 48–72 hours of guideline-concordant systemic glucocorticoids after exclusion of infection and alternative causes. - Incidence of CIP varies; in NSCLC trial data reported ~3.0%–6.3% and real-world estimates approach higher ranges; roughly 10% or more of CIP cases respond poorly to glucocorticoids. - Refractory CIP carries higher rates of respiratory failure, intensive care need, and mortality than corticosteroid-sensitive cases; reported outcomes are heterogeneous across small retrospective cohorts. - Radiology is heterogeneous: organizing pneumonia (OP), NSIP, hypersensitivity pneumonitis, and diffuse alveolar damage (DAD) patterns are described; DAD associates with higher-grade disease and worse prognosis. - BALF frequently shows lymphocytosis but data are inconsistent; BALF cytokine findings (including **IL-6**, IL-17A, CXCL10) vary across small studies and may reflect timing and assay differences rather than definitive phenotypes. - Candidate mechanistic drivers include persistent tissue-resident T-cell activation (notably CD8+), inflammatory T-cell–myeloid crosstalk, redundant cytokine networks, innate immune dysregulation, and, in some patients, transition toward maladaptive repair and fibrotic remodeling. - Second-line treatments (mycophenolate mofetil, infliximab, IVIG, tocilizumab, cyclophosphamide, calcineurin inhibitors) are supported chiefly by case reports, series, and retrospective data; no RCT-backed optimal sequence exists. - The authors propose a hypothesis-generating phenotype-to-therapy framework distinguishing lymphocyte-dominant, cytokine-dominant, fulminant DAD, innate/myeloid-dominant, and organizing/fibrotic-remodeling phenotypes while acknowledging overlap and temporal evolution. - Translational and clinical priorities include harmonized definitions, longitudinal multimodal phenotyping (including BALF and imaging), development of treatment-predictive biomarkers, and prospective comparative studies to evaluate mechanism-informed interventions.
## Clinical Analysis & Structured Key Points
Frontiers | Corticosteroid-refractory checkpoint inhibitor pneumonitis in lung cancer: biological heterogeneity and mechanism-informed management REVIEW article Front. Immunol. , 28 October 2026 Sec. Cancer Immunity and Immunotherapy Volume 17 - 2026 | https://doi.org/10.3389/fimmu.2026.1983756 Published in Frontiers in Immunology Cancer Immunity and Immunotherapy 7 impact factor 11.3 citescore Editor & Reviewers Edited by Y S Yoo Sun Kim Reviewed by H C Huanwen Chen J D Jiaxi Deng 1 A 1 Anonymous reviewer Outline Figures and Tables Figure 1 View in article Table 1 Evidence summary of second-line and rescue therapies for corticosteroid-refractory checkpoint inhibitor pneumonitis. View in article Table 2 Proposed phenotype-to-therapy matching framework for corticosteroid-refractory checkpoint inhibitor pneumonitis. View in article REVIEW article Front. Immunol. , 28 October 2026 Sec. Cancer Immunity and Immunotherapy Volume 17 - 2026 | https://doi.org/10.3389/fimmu.2026.1983756 Corticosteroid-refractory checkpoint inhibitor pneumonitis in lung cancer: biological heterogeneity and mechanism-informed management R Z Rong Zhang 1,2 † * X Z Xiaoxue Zhang 1,2 X H Xiulan He 1,2 M X Min Xiang 1,2 1. Department of Respiratory and Critical Care Medicine, West China Tianfu Hospital, Sichuan University, Chengdu, China 2. Department of Respiratory and Critical Care Medicine, West China Hospital, Sichuan University, Chengdu, Sichuan, China Article metrics View details Abstract Immune checkpoint inhibitors (ICIs) have transformed the treatment of lung cancer, but checkpoint inhibitor pneumonitis (CIP) remains a potentially life-threatening immune-related adverse event. A subset of patients fails to improve despite adequate glucocorticoid therapy and develops corticosteroid-refractory CIP, which is associated with respiratory failure, substantial mortality, and considerable therapeutic uncertainty. This review critically examines the clinical definition, phenotypic characteristics, immunological mechanisms, and therapeutic evidence for corticosteroid-refractory CIP. Emerging data suggest that refractoriness should not be regarded as a uniform failure of glucocorticoid treatment, but rather as a biologically heterogeneous and dynamically evolving syndrome involving persistent tissue-resident T-cell activation, inflammatory T-cell–myeloid crosstalk, cytokine-network redundancy, innate immune dysregulation, and, in some patients, transition toward maladaptive tissue repair and structural lung remodeling. Current second-line treatments, including mycophenolate mofetil, infliximab, intravenous immunoglobulin, tocilizumab, cyclophosphamide, and calcineurin inhibitors, are supported predominantly by retrospective studies and case-report and case-series evidence, with no established optimal treatment sequence. We therefore propose a hypothesis-generating phenotype-to-therapy framework that distinguishes lymphocyte-dominant, cytokine-dominant, fulminant acute lung injury/diffuse alveolar damage, innate/myeloid-dominant, and organizing/fibrotic-remodeling phenotypes while recognizing their potential overlap and temporal evolution. Translating this framework into clinical practice will require harmonized definitions, longitudinal multimodal phenotyping, treatment-predictive biomarkers, and prospective comparative studies. Ultimately, management should move beyond empirical escalation of immunosuppression toward identifying the dominant biological process at a given disease stage and matching treatment intensity and mechanism accordingly. 1 Introduction Immune checkpoint inhibitors (ICIs) have transformed the treatment of multiple malignancies, particularly advanced lung cancer. By targeting the programmed cell death protein 1/programmed death-ligand 1 (PD-1/PD-L1) axis or cytotoxic T-lymphocyte-associated antigen 4 (CTLA-4), they relieve immunosuppression within the tumor microenvironment and reinvigorate antitumor T-cell responses ( 1 – 3 ). In non-small cell lung cancer (NSCLC), pembrolizumab, nivolumab, atezolizumab, and durvalumab are used in first-line, consolidation, and perioperative settings with significant survival gains ( 1 , 2 ), and pivotal trials including KEYNOTE-024, CheckMate-017/057, and IMpower133 confirmed prolonged survival in NSCLC and extensive-stage small cell lung cancer ( 3 – 6 ). With widespread adoption, however, immune-related adverse events (irAEs) have become a major factor limiting ICI use and may affect virtually every organ system ( 7 ); among these, checkpoint inhibitor pneumonitis (CIP) is one of the most severe, owing to its insidious onset, rapid progression, and potentially fatal course ( 8 , 9 ). Guidelines from the American Society of Clinical Oncology (ASCO), the European Society for Medical Oncology (ESMO), and the National Comprehensive Cancer Network (NCCN) provide a severity-graded framework for the management of CIP ( 10 – 12 ). Glucocorticoids are the first-line treatment for CIP ( 10 – 12 ), and most patients respond to adequate corticosteroid therapy, with reported response rates of approximately 70%-80% ( 13 ). A proportion respond poorly or relapse during tapering, manifesting as a corticosteroid-refractory or corticosteroid-dependent course ( 14 , 15 ); these patients typically have more severe disease, often requiring intensive care, and their mortality risk is markedly increased, making early recognition and effective management a pressing clinical problem ( 16 ). Corticosteroid-refractory CIP presents several challenges: no internationally standardized definition exists, and differing criteria across guidelines and studies affect trial eligibility and cross-study comparability; prognosis is poor, with a high risk of respiratory failure and death; evidence for second-line and salvage therapy derives largely from case reports and small retrospective series, with no strategy supported by randomized controlled trials (RCTs); and the mechanisms of disease onset and glucocorticoid resistance remain incompletely defined, involving persistent T-cell activation, cytokine imbalance, and innate immune dysregulation ( 17 , 18 ). This review integrates current clinical and translational evidence to characterize corticosteroid-refractory CIP, critically examines candidate mechanisms of glucocorticoid refractoriness and available rescue therapies, and proposes a phenotype-to-therapy framework for transitioning from empirical immunosuppressive escalation toward mechanism-informed precision management. 2 Definition and clinical features of corticosteroid-refractory CIP 2.1 Current definitions The ASCO, ESMO, and NCCN guidelines use broadly similar response-based approaches to identify corticosteroid-refractory CIP, although no universally standardized definition has been established. ASCO recommends treatment escalation when no clinical improvement is observed after approximately 48 hours of appropriate glucocorticoid therapy ( 11 ), whereas ESMO recommends escalation when no improvement is observed after 72 hours ( 10 ). NCCN similarly recommends reassessment within 48 hours for grade 3–4 CIP and consideration of additional immunosuppression if no improvement occurs ( 19 ). Adequate glucocorticoid therapy is defined as guideline-concordant systemic treatment matched to disease severity, generally prednisone or an equivalent at 1–2 mg/kg/day for grade 2 CIP and intravenous methylprednisolone at 1–2 mg/kg/day for grade 3–4 CIP, with response assessed over approximately 48–72 hours. Accordingly, we define corticosteroid-refractory CIP as persistent or worsening clinical or respiratory manifestations despite approximately 48–72 hours of guideline-concordant systemic glucocorticoid therapy, after appropriate evaluation for infection, tumor progression, and other alternative causes of deterioration. Corticosteroid-refractory CIP should be distinguished from corticosteroid-dependent, recurrent, and chronic CIP. In this review it refers to persistent deterioration or lack of meaningful improvement despite adequate glucocorticoid therapy within the guideline response-assessment window, typically 48-72h ( 10 – 12 , 15 ). Corticosteroid-dependent CIP initially improves but relapses during tapering or cannot be weaned, which may overlap therapeutically with refractory disease without indicating primary glucocorticoid failure ( 14 , 15 , 20 ). Recurrent CIP denotes a new episode after previous improvement or resolution, including recurrence after ICI re-exposure ( 21 , 22 ), whereas chronic CIP has been defined as pneumonitis persisting or worsening during tapering after ICI discontinuation and requiring immunosuppression for at least 12 weeks ( 23 ). Notably, a multicenter longitudinal analysis by Zhou et al. independently identified a chronic phase beginning at approximately 13 weeks after CIP onset, broadly consistent with this temporal threshold, although their cohort was not restricted to corticosteroid-refractory disease ( 24 ). Pooling refractory, dependent, recurrent, and chronic phenotypes can therefore obscure differences in disease biology, treatment timing, and outcomes. 2.2 Incidence and mortality The reported incidence of CIP varies with treatment regimen, tumor type, and study population: among patients with NSCLC receiving ICIs, clinical trial data indicate approximately 3.0%-6.3% and real-world estimates approach 13.0%-19.0% ( 25 , 26 ), and dual checkpoint blockade may increase risk relative to monotherapy ( 8 , 27 ). Approximately 10% or more of patients respond poorly to glucocorticoids and some progress to refractory disease requiring additional immunosuppression ( 25 , 28 ). Prognosis is poor, with a high risk of respiratory failure and death. Reported mortality rates are heterogeneous, but outcomes are consistently worse than in corticosteroid-sensitive disease, and available evidence derives primarily from single-center retrospective studies and case series; Beattie et al. reported a 90-day all-cause mortality/hospice referral of 50% among patients treated with second-line immunosuppression ( 29 ). Beyond short-term mortality, available longitudinal evidence suggests that CIP may follow a temporally evolving clinical course. In a multicenter retrospective study of 56 patients with CIP, Zhou et al. proposed three phases according to time from CIP onset: an acute phase (≤4.9 weeks), a subacute phase (4.9-13.1 weeks), and a chronic phase (≥13.1 weeks) ( 24 ). Across these phases, symptoms, performance status, and CIP grade generally decreased, while radiological inflammatory lesions progressively resolved. However, structural abnormalities persisted in a subset of patients, including traction bronchiectasis in 3 of 11 patients reaching the chronic phase. Histopathological findings also suggested evolution from predominantly acute fibrinous pneumonitis in the acute phase toward organizing pneumonia in the subacute phase and fibrotic remodeling at later stages. Circulating IL-6, IL-10, and hsCRP likewise decreased over time. Importantly, this cohort included patients with CIP irrespective of glucocorticoid responsiveness and therefore does not establish the natural history or long-term prognosis of corticosteroid-refractory CIP specifically. Nevertheless, these findings provide a useful temporal framework indicating that persistent disease may evolve from an inflammation-dominant process toward organizing or fibrotic remodeling, underscoring the need for longitudinal reassessment in patients with an apparently refractory course. 2.3 Clinical presentation 2.3.1 Symptoms Patients with corticosteroid-refractory CIP typically present with dyspnea, cough, fever, and hypoxemia, and some deteriorate rapidly ( 8 , 9 ). Compared with patients who remit after glucocorticoid therapy, they more often develop persistent hypoxemia and acute respiratory failure and more frequently require intensive care and mechanical ventilation ( 28 , 29 ). Continued deterioration despite adequate therapy may indicate ongoing immune-mediated inflammatory activation and should prompt escalated immunosuppression ( 11 ). 2.3.2 Computed tomography features High-resolution computed tomography (HRCT) is essential for the diagnosis, radiological classification, and longitudinal assessment of CIP ( 30 , 31 ). CIP is radiologically heterogeneous; common patterns include organizing pneumonia (OP), non-specific interstitial pneumonia (NSIP), hypersensitivity pneumonitis (HP), and diffuse alveolar damage (DAD), with OP generally most frequent ( 18 , 30 , 31 ), and in a recent cohort of 102 patients an OP pattern accounted for approximately 50% of cases ( 32 ). CT patterns may correlate with severity: the DAD or acute interstitial pneumonia pattern typically shows extensive bilateral ground-glass opacities and consolidation with septal thickening and traction bronchiectasis, indicating widespread alveolar–capillary barrier disruption, and is more frequent in high-grade CIP and associated with rapidly progressive respiratory failure and poorer outcomes ( 18 , 31 ). 2.3.3 Laboratory findings No laboratory parameter specifically predicts corticosteroid-refractory CIP. Some patients show elevated C-reactive protein (CRP), lactate dehydrogenase, and neutrophil counts with lymphopenia. Lung injury-associated markers such as Krebs von den Lungen-6 (KL-6) and surfactant protein D may reflect parenchymal damage and have ancillary diagnostic value in ICI-related pneumonitis, although their capacity to predict glucocorticoid responsiveness requires validation ( 33 ), and serial KL-6 monitoring awaits prospective confirmation ( 34 ). Pro-inflammatory cytokines such as interleukin-6 (IL-6) are elevated in some patients, yet no single mediator predicts treatment failure, and although tocilizumab has shown potential value in selected irAEs and in a small number of corticosteroid-refractory CIP cases ( 35 ), treatment selection and response prediction remain highly uncertain ( 16 , 29 ). 2.3.4 Bronchoalveolar lavage fluid features Bronchoalveolar lavage (BAL) is an important ancillary investigation, principally to obtain lower respiratory tract specimens and exclude pulmonary infection, alveolar hemorrhage, and tumor progression; it may be considered in grade 2 or higher disease, particularly when infection cannot be excluded or imaging is atypical ( 10 , 11 ). No single bronchoalveolar lavage fluid (BALF) parameter can confirm or exclude CIP. BALF frequently demonstrates lymphocytosis, although findings are heterogeneous: Chen et al. reported a mean lymphocyte proportion of 36.7% ± 22.5% in 42 patients, with 81.0% meeting a lymphocytic pattern (>15%) and no difference by disease grade ( 36 ), whereas Guo et al. found higher proportions associated with high-grade disease and relapse during tapering ( 21 ). Because these studies are single-center, small, and retrospective, BALF lymphocyte proportions cannot reliably assess severity or prognosis. Lymphocyte subsets are similarly heterogeneous: Chen et al. found relative CD8 + predominance, with 73.8% of patients having a CD4 + /CD8 + ratio below 1 ( 36 ), whereas Guo et al. found a ratio above 1 associated with recurrence during corticosteroid tapering ( 21 ). Comparable changes occur in drug-induced lung injury, hypersensitivity pneumonitis, and other interstitial lung diseases and therefore lack specificity ( 37 , 38 ), and BALF neutrophilia cannot distinguish infection from immune-mediated injury ( 10 , 11 ). BALF cytokines have been explored as candidate biomarkers with inconsistent results: Kowalski et al. found significantly elevated BALF IL-6 and group-dependent elevations in BALF IL-17A, with BALF IL-6 showing potential discriminatory value for CIP ( 39 ); Si et al. reported increased IP-10/CXCL10 and CD3 + CD8 + T-cell proportions without differences in IL-6 ( 40 ); and Ploch et al. observed elevated BALF IL-6 in CIP and other immune-mediated conditions, although only three patients with CIP were included ( 41 ). These inconsistencies may reflect differences in case definitions, disease stage, sampling, and assay platforms. Single-cell sequencing has further shown BALF T-cell enrichment, cytotoxic T-cell activation, and expansion of pro-inflammatory monocytes/macrophages in CIP ( 42 ). Overall, BALF data support the concept that corticosteroid-refractory CIP is unlikely to represent a single inflammatory phenotype. Available cohorts suggest potentially distinct lymphocyte-dominant, T-cell subset-skewed, and myeloid/neutrophil-enriched patterns, accompanied by variable cytokine signatures involving IL-6, IL-17A, and CXCL10/IP-10. However, these putative phenotypes remain poorly defined because most published BALF studies enrolled heterogeneous CIP populations and were not designed to distinguish corticosteroid-responsive from corticosteroid-refractory disease. Small sample sizes, retrospective designs, variable timing of bronchoscopy relative to corticosteroid exposure, inconsistent exclusion of infection, and differences in cytokine assays further limit comparability. Thus, BALF cellular and cytokine profiling is currently best viewed as a research tool for mechanistic stratification rather than a validated method for defining corticosteroid-refractory CIP subgroups or guiding rescue therapy. 2.4 Risk factors Identifying patients at risk of progressing to corticosteroid-refractory CIP is important for early risk stratification, but existing evidence largely concerns the risk of CIP or of severe CIP rather than predictors of glucocorticoid treatment failure ( 10 , 11 ). Pre-existing interstitial lung disease or interstitial lung abnormalities, pulmonary fibrosis, and combination immunotherapy may increase risk ( 43 ), as may previous thoracic radiotherapy ( 44 , 45 ), whereas advanced age, smoking history, and chronic obstructive pulmonary disease have been implicated inconsistently ( 43 , 46 ). Although the reported incidence of CIP is relatively high in NSCLC ( 47 ), coexisting smoking-related lung injury, interstitial abnormalities, and prior radiotherapy make it uncertain whether NSCLC histology is an independent risk factor. These factors principally reflect susceptibility to CIP rather than validated predictors of glucocorticoid failure. The single-center retrospective study by Li et al. is among the few to address corticosteroid-refractory CIP directly, comparing 28 refractory with 38 corticosteroid-sensitive patients ( 16 ). Refractory patients more frequently had pre-existing interstitial lung disease, higher severity grades, higher neutrophil counts and procalcitonin, and lower albumin; on multivariate analysis, grade 3–4 disease at diagnosis and a higher absolute neutrophil count were independently associated with refractoriness ( 16 ). These factors nevertheless require prospective multicenter validation. 3 Mechanisms of corticosteroid-refractory checkpoint inhibitor pneumonitis Corticosteroid-refractory CIP is likely a biologically heterogeneous syndrome rather than a single mechanistic entity, involving persistent tissue-resident T-cell activation, inflammatory T-cell–myeloid crosstalk, cytokine and chemokine amplification ( 48 – 50 ). In some patients, failure of epithelial repair and structural remodeling may also contribute to persistent lung injury ( 51 , 52 ). Because most mechanistic studies have compared patients with and without CIP rather than steroid-sensitive and steroid-refractory disease, these pathways should be regarded as candidate mechanisms of refractoriness rather than established determinants of glucocorticoid resistance ( 16 , 53 ). Retrospective cohorts indicate that refractory disease is associated with greater initial severity, elevated peripheral neutrophil counts, and poor long-term outcomes, and pre-existing pulmonary abnormalities have been associated with severe CIP ( 16 , 17 , 53 ). 3.1 Sustained adaptive immune activation 3.1.1 Persistent CD8 + T
## Related Clinical Research

- [Sintilimab plus R-CHOP vs DA-EPOCH-R in PMBCL — article content unavailable on source page](https://medichelpline.com/clinical-feed/frontiers-in-immunology-15-sintilimab-plus-r-chop-versus-da-epoch-r-as-first-line-therapy-for-primary.md)
- [Editorial placeholder: Innovative therapeutic approaches for complex cancers — source content miss](https://medichelpline.com/clinical-feed/frontiers-in-immunology-12-editorial-innovative-therapeutic-approaches-for-complex-cancers-exploring-new.md)
- [Perioperative ivonescimab ± chemotherapy Yields High MPR in Resectable NSCLC (Phase 2)](https://medichelpline.com/clinical-feed/nature-1-perioperative-ivonescimab-with-or-without-chemotherapy-in-resectable-non-small.md)
- [Neoadjuvant durvalumab plus chemotherapy in resectable NSCLC: results from the phase 2 MDT-BRIDGE](https://medichelpline.com/clinical-feed/nature-0-neoadjuvant-durvalumab-plus-chemotherapy-in-resectable-or-borderline-resectable.md)
- [VISTA promotes tumour-intrinsic proliferation in mesothelioma cells](https://medichelpline.com/clinical-feed/british-journal-of-cancer-0-vista-drives-tumour-intrinsic-proliferation-in-mesothelioma-cells.md)

## Navigation
- [← Back to Infectious Disease Feed](https://medichelpline.com/clinical-feed/infectious-disease.md)
- [← All Clinical Specialties](https://medichelpline.com/clinical-feed.md)
## Medical & Regulatory Disclaimer

> [!CAUTION]
> MedicHelpline content is structured for research, educational, and professional discovery purposes. It does not constitute individual medical advice, clinical diagnosis, or treatment recommendations.
> Always verify dosing, contraindications, and regulatory alerts against official product labeling and primary regulatory sources before clinical decision-making.