Case reports of immune-associated pneumonia with radiological cavity formation

Abstract

Introduction

Checkpoint inhibitor-associated pneumonitis (CIP) is typically characterized by various radiographic patterns, including organizing pneumonia (OP), nonspecific interstitial pneumonia (NSIP), acute interstitial pneumonia (AIP), usual interstitial pneumonia (UIP), hypersensitivity pneumonitis (HP), and diffuse alveolar damage (DAD). However, cavitary lung lesions on chest CT images have rarely been described in the literature.

Case presentation

This article reports two lung cancer patients treated with immune checkpoint inhibitors (ICIs) who developed cavitary lung lesions on chest CT. Histopathology of OP with lymphocytic infiltration was supported by multiplex immunofluorescence. After excluding infection and tumour progression, CIP was diagnosed on the basis of the patient’s history of ICI use, clinical presentation, imaging characteristics, laboratory findings, and pathological results. Both patients responded well to glucocorticoids, showing complete or near-complete resolution of cavitary lesions. After steroid discontinuation, the first patient underwent right upper lobectomy followed by adjuvant chemotherapy. After lung cancer recurrence, the patient received combined chemotherapy, immunotherapy, and radiotherapy, with no observed recurrence of CIP during follow-up. The second patient has been receiving traditional chinese medicine treatment since 2025 at another hospital.

Conclusion

CIP rarely presents with cavitary lung lesions on CT, and careful differentiation of CIP from infectious aetiologies is essential. Empirical management may lead to treatment failure; therefore, imaging and biopsy are crucial for precise patient care.

Introduction

Checkpoint inhibitor-associated pneumonitis (CIP), triggered by immune checkpoint inhibitors (ICIs), is recognized as one of the major immune-related adverse events associated with antitumour therapies. Retrospective epidemiological studies have revealed that the incidence of CIP ranges from 3.5% to 19% [1–6]. Several factors contribute to the incidence of CIP, including smoking history [7], preexisting pulmonary conditions [8–9], and the specific type of ICI administered [10–11]. Compared with their nonsmoking counterparts, individuals with a history of smoking are more likely to develop CIP. Furthermore, patients with preexisting conditions such as chronic obstructive pulmonary disease (COPD) and interstitial lung disease (ILD) face a significantly increased risk of CIP. A pooled analysis revealed that lung cancer patients with concurrent COPD are at an increased risk of developing CIP during ICI treatment, with an odds ratio (OR) of 1.54 [12].

As an exclusionary diagnosis, the clinical diagnosis is established on the basis of the following aspects: (1) a history of ICI administration; (2) the presence or exacerbation of respiratory symptoms, including cough and dyspnoea, either with or without fever [13–14]; (3) the manifestation of radiographic abnormalities such as ground-glass opacities (GGOs), consolidation, or interstitial patterns on chest CT images [15–16]; (4) the lack of response of lung lesions to antibiotics but a positive response to glucocorticoids; and (5) the exclusion of other potential aetiologies such as lung infection, tumour progression, heart failure, and pulmonary embolism.

Alterations in chest CT findings play a crucial role in the diagnosis of CIP. CIP, categorized as interstitial pneumonia [17], can manifest as OP, NSIP, AIP or acute respiratory distress syndrome (ARDS), UIP, HP, and DAD. Among these, OP and NSIP are the most prevalent forms. Nevertheless, reports of cavitary lesions on chest CT images are scarce. Herein, we report two cases in which CIP was characterized by the presence of cavitary lesions on imaging and was confirmed via histopathological examination.

Case presentation

Case 1

A 61-year-old male with a history of pre-COPD and a 30-pack-year smoking history presented to our clinic for cough and slight haemoptysis. Chest CT revealed a soft tissue density shadow measuring 5.0 cm × 3.5 cm, with irregular margins in the hilum of the right upper lung. The right upper pulmonary artery was encircled by a mass accompanied by obstructive pneumonia and emphysema (Fig. 1A-B). Whole-body PET-CT revealed increased uptake of FDG in the mass in the right upper hilar region, while no such increase was observed in other areas.

Fig. 1.

Chest CT scans in two patients. A-B CT images at the time of initial diagnosis in the patient of case 1. C-D CT images after receiving 3 cycles of chemotherapy combined with ICI treatment, showing the cavity lesions in the upper lobe of the right lung in the patient of case 1. E-F CT images showing cavity disappear after glucocorticoid treatment in the patient of case 1. G-H CT images at the time of initial diagnosis in the patient of case 2. I-J CT images after receiving 4 cycles of chemotherapy combined with ICI treatment, showing the cavity imaging in the posterior upper lobe and dorsal lower lobe of right lung in the patient of case 2. K-L CT images showing smaller cavity lesions after glucocorticoid treatment for three weeks in the patient of case 2

To establish a diagnosis, a routine bronchoscopy was arranged. It revealed a neoplasm at the orifice of the right superior lobar bronchus. Immunohistochemical staining revealed positive results for ck-pan, CK5/6, and p40 but negative results for TTF-1, Napsin-A, CD56, CgA, syn, LCA, p53, and vim, thereby confirming the presence of squamous cell carcinoma. Additionally, lung cancer driver gene tests for EGFR, ALK, and ROS were negative. On the basis of the patient’s medical history, clinical manifestations, and ancillary examinations, a diagnosis of right lung squamous cell carcinoma T4N0M0 Stage IIIA with pulmonary artery invasion was established. The patient was administered paclitaxel (albumin-bound) at a dose of 220 mg/m² and carboplatin (with an AUC of 5) in combination with 200 mg of pembrolizumab for three cycles. The patient subsequently complained of right-sided chest pain, exacerbated by deep breathing and coughing. Compared with the previous scan, chest CT revealed a reduction in the size of the lesion, measuring 2.5 cm × 2.2 cm, adjacent to the right upper lung hilum; however, a cavitary lesion was detected in the right upper lobe (Fig. 1C-D). To characterize the cavitary lesion, a percutaneous lung biopsy was carried out. Haematoxylin and eosin (HE) staining revealed degeneration and necrosis of alveolar epithelial cells and proliferation of myofibroblasts and fibroblasts within the alveolar cavity, accompanied by infiltration of lymphocytes, plasma cells, and histiocytic cells, as well as thickening of the alveolar septum with inflammatory cell infiltration (Fig. 2A-B). No tumour cells were observed by H&E staining, with no evidence of associated vascular invasion, perineural infiltration, or other invasive growth patterns (Fig. 2A-B). Immunohistochemical staining revealed positive results for desmin (Fig. 2C) and SMA (Fig. 2D), negative staining for ALK, and positive staining for CK (alveolar epithelium), CK7 (alveolar epithelium), TTF-1 (alveolar epithelium), and a ki-67 proliferation index of 1%, further supporting the diagnosis of OP. Microbial next-generation sequencing (mNGS) of lung tissues failed to detect any specific pathogens. Acid-fast bacilli staining and sputum culture were negative. In light of the patient’s medical history, clinical features, and immunohistochemical findings, a diagnosis of CIP was considered. The patient was prescribed prednisone 30 mg orally once daily, with a tapering regimen of one tablet reduction per week over a total course of 6 weeks. Given that the lung cancer remained relatively stable, no additional antitumor therapy was administered during the corticosteroid treatment period. Chest CT imaging subsequently demonstrated the disappearance of the cavity in the right upper lobe and stability in the size of the mass lesion, measuring 2.5 cm × 2.2 cm, at the right upper hilum (Fig. 1E-F). The patient underwent right upper lobectomy at another hospital. Adjuvant treatment with two cycles of paclitaxel (albumin-bound) plus carboplatin was administered postoperatively, followed by regular surveillance. On March 2022, PET-CT revealed enlarged mediastinal lymph nodes with increased FDG uptake, and the patient subsequently received five cycles of combined therapy with paclitaxel (albumin-bound), carboplatin, and durvalumab, followed by three cycles of durvalumab monotherapy. Follow-up PET‒CT in April 2023 revealed further slight enlargement of the mediastinal lymph nodes with elevated SUV values. Mediastinal radiotherapy was then administered. Owing to financial constraints, subsequent treatment consisted of tislelizumab with intermittent cycles of paclitaxel (albumin-bound) plus carboplatin therapy (Fig. 3A). No recurrence of CIP was observed.

Fig. 2.

A-B HE staining of pulmonary cavity lesion in the first case’s lung tissues (A: 10 ×; B: 20×). C Desmin staining in the first case. D SMA staining in the first case. E-F HE staining in the second case’s lung tissues (E and F: 10 ×)

Fig. 3.

Timeline of the two patients. A Timeline of the first patient. B Timeline of the second patient

Case 2

A 74-year-old male with comorbidities, namely, COPD, hypertension, and atrial fibrillation, was referred to our clinic because of cough and mild haemoptysis. He had a 30-pack-year smoking history. Laboratory investigations revealed an SCC concentration of 1.7 ng/mL, a Cyfra21-1 concentration of 2.77 ng/mL, a CA125 concentration of 38 U/mL, and a C-reactive protein concentration of 13.2 mg/L, with no other abnormal results. Chest CT revealed a soft tissue density shadow in the hilum of the left upper lung, accompanied by obstructive pneumonia, left-sided pleural effusion, and enlarged lymph nodes in the mediastinum (Fig. 1G-H). Abdominal CT, bone scintigraphy, and head MRI did not reveal any tumour lesions.

For diagnostic confirmation, a cryobiopsy was arranged. Immunohistochemical staining revealed positive results for CK, p40, P63, and CK5/6, while the results for CK7, TTF-1, Syn, and ALK were negative, suggesting the presence of squamous cell carcinoma. Moreover, lung cancer driver gene tests for EGFR, ALK, and ROS were negative. Considering the patient’s medical history, clinical manifestations, and auxiliary examination findings, a diagnosis of left lung squamous cell carcinoma T2N2M1a Stage IV with mediastinal lymph node and left pleural metastasis was made. The patient was treated with paclitaxel (albumin-bound) at a dose of 220 mg/m² and carboplatin (with an AUC of 5) in combination with tistelizumab (200 mg) for four cycles. Chest CT revealed that the size of the lesions in the left lung were decreased compared with that in the previous scan, and cavitary lesions were observed in the posterior upper lobe and dorsal lower lobe of the right lung (Fig. 1I-J). To characterize the cavitary lesion, a transbronchial lung biopsy (TBLB) was carried out. Pathology of the biopsy samples revealed that most alveolar walls were collapsed, the interstitial fibrous tissue was hyperplastic, lymphocytes and plasma cells were significantly infiltrated, and focal fibrinoid exudation was noted within the alveolar cavity (Fig. 2E-F). Bronchoalveolar lavage fluid (BALF) mNGS failed to detect any specific pathogens. Acid-fast bacilli staining, fungal staining, mycobacterial detection and bacterial culture were negative in the BALF. TB‒PCR was negative for the biopsy sample. On the basis of the patient’s medical history, clinical features, and pathological results, a diagnosis of CIP was considered. The patient was prescribed prednisone 30 mg orally once daily, with a tapering regimen of one tablet reduction per week. Chest CT re-examination performed 3 weeks after prednisone administration revealed a reduction in the size of the cavity in the right upper lobe and a reduction in the size of the mass lesion at the left upper hilum (Fig. 1K-L), and the patient was continuously treated with corticoids for three weeks. The patient did not receive further tumor directed chemotherapy, immunotherapy, or radiotherapy, but instead pursued traditional Chinese medicine treatment at another hospital. The patient currently remains in generally good condition (Fig. 3B).

Discussion

As of February 1, 2026, a PubMed search using the terms “immune checkpoint inhibitor pneumonia” and “cavity” retrieved very few case reports of cavitary chest imaging findings in patients with CIP. The literature indicates that CIP imaging features are primarily associated with NSIP and OP patterns [1, 18], while cavitation is relatively uncommon. These two cases suggest that cavitation can be a radiological manifestation of CIP. In clinical practice, new cavitary lesions outside the primary disease site should be carefully differentiated from specific infections, such as fungal infections and tuberculosis [19]. Empirical management may lead to treatment failure; therefore, imaging and biopsy are crucial for precise patient care. Differential diagnosis should be based on a comprehensive assessment of medical history, clinical presentation, auxiliary examinations, pathological findings, and treatment response.

These two patients had a history of ICI use and presented with cough, but they did not exhibit symptoms suggestive of tuberculosis (such as fever or night sweats) or fungal infection (such as haemoptysis). The aetiological results were negative, and the pathological findings supported the diagnosis of OP. Considering the patients’ ICI use history, clinical presentation, imaging features, laboratory results, mNGS findings, pathological results, and the regression of cavitary lesions after corticosteroid therapy, the evidence supports the diagnosis of CIP. However, it is important to note the limitations of mNGS, such as sampling error, its inability to distinguish colonization from infection, and its variable sensitivity for certain pathogens, such as Mycobacterium tuberculosis and Aspergillus, because of factors such as cell wall disruption efficiency. Therefore, clinical decisions therefore require integrating the patient’s history, symptoms, imaging, and pathology with both microbiological test results for a complete diagnostic picture.

Although pathology does not currently play a dominant role in the diagnosis of CIP, it is advisable to strengthen pathological assessment in cases where CIP cannot be clearly diagnosed, particularly in the presence of atypical imaging features, to obtain more definitive conclusions. In essence, CIP is a form of non‑infectious pulmonary inflammation confined to interstitial and alveolar infiltration, resembling the interstitial pneumonias observed in collagen vascular diseases [20–21]. Currently, pathological studies on ICI-induced CIP are limited. Some case reports have described pathological findings such as OP, interstitial pneumonia, diffuse alveolar damage, and even cases with no abnormal findings [22–23]. We report two cases with pathological findings supportive of OP.

Anticancer therapy in patients after CIP treatment is often difficult because the patients may have some risk factors for recurrent ILD. Tumour burden, the severity of CIP, and whether radiographic evidence of CIP has resolved after treatment should be considered when deciding whether to proceed with ICI rechallenge. A retrospective study from the First Affiliated Hospital of Zhejiang University School of Medicine revealed that during continued ICI therapy, the likelihood of CIP recurrence or progression was 50%. Patients who resumed ICI treatment without discontinuation achieved significantly better outcomes than those who permanently discontinued therapy did, with the median progression-free survival (mPFS) duration extended by 6 months and the median overall survival (mOS) duration doubled. The decision to resume ICI therapy should be based on tumour control status, symptomatic improvement, and radiological findings [24]. In the first case, the patient experienced grade 2 CIP. Following glucocorticoid treatment, the pulmonary cavitary lesions significantly resolved. After discontinuation of glucocorticoids, the patient underwent right upper lobectomy for lung cancer and subsequently received chemotherapy and durvalumab. To date, no recurrence of CIP has been observed.

CIP is a form of interstitial pneumonia characterized by damage to the lung interstitium and alveoli. A more comprehensive analysis of the imaging features of the two patients suggested that the exact mechanism underlying cavity formation remains unclear. Potential mechanisms include the following aspects: (1) Extension of preexisting emphysema or bullae. ICIs may aggravate underlying emphysema or enlarge bullae through immune-mediated parenchymal damage. Local inflammatory responses can weaken alveolar walls, promoting cavity formation in structurally vulnerable areas. (2) Immune-mediated alveolar damage. Activated T-cells and macrophages triggered by checkpoint inhibitors may directly attack alveolar epithelium and endothelium, causing tissue injury resembling barotrauma. The necrosis and clearance of injured alveoli can lead to cystic or cavitary lesions visible on imaging. (3) Checkpoint inhibitor related vasculitis or ischemic injury. Immune-related vasculitis or microvascular injury could induce localized ischemic necrosis in lung tissue. As necrotic material is absorbed or expectorated, residual cavities may form a mechanism similar to cavitation in conditions like granulomatosis with polyangiitis. (4) Rapid clearance of inflammatory masses. The swift resolution of inflammatory consolidations may leave behind persistent air-filled cavities. The precise mechanisms warrant further investigation.

Although ICIs increase the antitumour efficacy of T cells, they may aberrantly amplify normal immune responses, leading to disrupted immune tolerance and triggering autoimmune-like inflammation in healthy tissues. These reactions are mediated by CD4 + T cells, CD8 + T cells, B lymphocytes, granulocytes, and cytokines [25–26]. Enhanced T-cell activity may target cross-reactive antigens shared by tumour and normal lung tissue, resulting in off-target toxicity. To further characterize lymphocytic infiltration, we performed multiplex immunofluorescence staining for T cells in lung cavity lesion sections from patients (Fig. 4). Increased infiltration of T lymphocytes, including CD3 + T lymphocytes, CD3 + CD4+ T lymphocytes, and CD3 + CD8+ T lymphocytes, was observed. These findings align with those of previous studies reporting significant increases in both CD4 + and CD8 + T lymphocytes in the lung tissue and BALF of CIP patients, indicating an excessive lymphocyte-mediated immune response [27–29]. However, larger-scale studies are needed to further validate the immune microenvironment characteristics in CIP.

Fig. 4.

Multiplex immunofluorescence staining on T cells in the patients’ pulmonary cavity lesion slides. A-E Multiplex immunofluorescence staining on T cells in the first patient. F-J Multiplex immunofluorescence staining on T cells in the second patient

This study has several limitations. First, it is a retrospective observational study. Second, the sample size is small. Third, the proposed mechanism for cavity formation remains unclear. Additionally, comprehensive immunophenotyping of the bronchoalveolar lavage fluid was not performed.

Authors’ contributions

Liang Guo, Huoyan Tang, Yu Yang contributed equally to this work. Liang Guo, Zhi Xu, Li Bai conceptualize the study design. Huoyan Tang, Yu Yang and Jia Chen were involved in data curation. Zhoukui Bi and Yangfan Lv were involved in methodology. All authors reviewed the manuscript

Funding

This work was supported by Grants from the Xinqiao Hospital of Army medical university (2024F042 and 2024YQB052) and the Chongqing Municipal Health Appropriate Technology Promotion Project (2025 JSTG016).

Data availability

All data generated or analyzed in this study are contained within this article. For any further inquiries, please contact the corresponding author.

Declarations

Ethics approval and consent to parti-cipate

This case was carried out ethically in compliance with the World Medical Association's Declaration of Helsinki. The data collection received approval from the Ethics Committees of the Xinqiao Hospital (Number 2025-126-01). No concerns regarding conflicts of interest or publication ethics.

Written informed consent was obtained from the two patients for participation in this study.

Consent for publication

Written informed consent for publication was obtained from the two patients.

Competing interests

The authors declare no competing interests.