Immunotherapy causing pneumonitis in a patient with non-small cell lung cancer (NSCLC)

Abstract

Our patient, who had been previously diagnosed with non-small cell lung cancer, presented with progressive dyspnoea after receiving second-line immunotherapy treatment with atezolizumab. Chest CT scan showed bilateral lung architectural distortion, bronchial dilatation, consolidative opacities, ground-glass opacities and linear opacities concerning for either infectious lung disease or treatment-related lung disease. A diagnostic bronchoscopy was performed and no evidence of malignancy or infection was detected. Discontinuing atezolizumab with the addition of oral corticosteroid improved the patient’s respiratory symptoms but the patient required continuous oxygen supplementation. Later, the patient was found to have radiologic findings suggestive of further progression of his pneumonitis after completion of a course of corticosteroid treatment and required another course of oral prednisone. Immune-mediated pneumonitis could present with mild to severe respiratory symptoms with a wide range of clinical and radiologic features and physicians should be aware of this diagnosis of exclusion. Although patients may experience progressive disease with or without immunotherapy rechallenge, most of these cases can be managed successfully with favourable outcomes.

Background

Treatment options for advanced non-small cell lung cancer (NSCLC) have developed beyond targeting of cell cycle and growth such as cisplatin and paclitaxel. Systemic therapy chosen depends on the presence of driver mutations identified in genes such as epidermal growth factor receptor (EGFR), anaplastic lymphoma kinase (ALK) fusion oncogene, reactive oxygen species (ROS1), and proto-oncogenes that are involved in cell signal pathways that control cell growth (BRAF and KRAS). In the absence of such driver mutations, patients may be eligible for immunotherapy that targets immune checkpoint inhibitors such as programmed death 1 (PD1) receptors on T-cells and PD-ligand 1 (PD-L1) on tumour cells which inhibit T-cell-induced antitumor activity.¹

Case presentation

A 72-year-old man who was an active smoker with a history of over 50 pack-years tobacco use and chronic obstructive pulmonary disease underwent a screening chest CT scan which demonstrated a left upper lobe non-calcified nodular opacity that measured around 25 mm in long axis and enlarged mediastinal lymph nodes at stations 5 and 12 (figure 1). Subsequently, a positron emission tomography (PET)–CT scan showed intense fluorodeoxyglucose (FDG) uptake in the left upper lobe nodular opacity and in the mediastinal lymph nodes stations 5 and 12 (figure 2). An intraoperative bronchoscopy and staging mediastinoscopy was performed and the pathology report showed scattered atypical clusters of cells present within the lymph nodes and were negative for cancer. The patient then underwent a diagnostic left upper lobe surgical wedge resection and removal of remaining station 5 mediastinal lymph nodes. Pathology showed left upper lobe poorly differentiated adenocarcinoma with detection of malignancy in station 5 lymph nodes: pathologic stage pT2aN2M0 with positive surgical margins/clinical stage III disease. Because a completion lobectomy was deemed to have been too technically difficult, the patient was referred to haematology/oncology for treatment of his NSCLC and received adjuvant chemotherapy paclitaxel and carboplatin in conjunction with radiation therapy. Two months after completing treatments, a follow-up PET–CT scan showed imaging findings suggestive of partial response to treatment with regards to the aforementioned lymph nodes. However, a follow-up PET–CT 3 months later showed imaging findings suggestive of progressive disease. Given the concern for progressive disease, the patient was recommended to start second-line treatment. The patient’s lung cancer did not have an EGFR or BRAF mutation or an ALK gene rearrangement. Since the patient had progressive disease after carboplatin and paclitaxel chemotherapy, he qualified for second line immunotherapy with either atezolizumab or nivolumab irrespective of being positive for PD-L1 expression. Therefore, a repeat biopsy was not performed, and the patient was started on atezolizumab. Two months later, he developed progressive shortness of breath.

Figure 1.

Screening chest CT scan shows a left upper lobe non-calcified nodular opacity that measures around 25 mm in long-axis concerning for lung cancer.

Figure 2.

(A,B) Positron emission tomography (PET)–CT scan shows a left upper lobe nodular opacity with intense fluorodeoxyglucose (FDG) uptake concerning for lung cancer. PET–CT scan also shows left upper lobe and aortopulmonary lymph nodes with intense FDG uptake concerning for regional lymphadenopathy.

Investigations

A chest CT scan demonstrated interval development of bilateral lung architectural distortion, bronchial dilatation, consolidative opacities, ground-glass opacities and linear opacities concerning for either infectious lung disease or treatment-related lung disease (figure 3). Atezolizumab therapy was halted but the patient continued to develop progressive shortness of breath resulting in hospitalisation and a diagnostic bronchoscopy was performed and no evidence of malignancy or infection was detected. The patient did have a positive nasopharyngeal swab for influenza A.

Figure 3.

Chest CT scan shows bilateral lung architectural distortion, bronchial dilatation, consolidative opacities, ground-glass opacities and linear opacities 2 months after starting immunotherapy.

Differential diagnosis

Differential diagnosis for this patient included infections (eg, viral infection, bacterial infection), immunotherapy-induced pneumonitis, chemoradiation-induced pneumonitis and progressive lung cancer.

Treatment

The patient was started on oseltamivir and intravenous methylprednisolone of 60 mg daily and then changed to oral prednisone of 40 mg daily which was tapered after 4 weeks. The patient’s respiratory symptoms improved back to baseline before completing the course of prednisone but worsened approximately 2 weeks after the last dose of prednisone. A repeat chest CT scan showed progression of bilateral lung architectural distortion, bronchial dilatation, consolidative opacities, ground-glass opacities and linear opacities concerning for progression of treatment-related lung disease (figure 4). The patient was started on moderate-dose prednisone again and expected to be on a more prolonged course than previously.

Figure 4.

(A–C) Chest CT scan shows interval increase in bilateral lung architectural distortion, bronchial dilatation, consolidative opacities, ground-glass opacities and linear opacities while on second course of oral prednisone for immunotherapy-induced pneumonitis.

Outcome and follow-up

The patient’s respiratory symptoms have improved modestly and is able to tolerate more exertional activities but still requires continuous supplemental oxygen support. Serial imaging by PET–CT revealed the patient had a right adrenal gland mass that was highly suspicious for malignancy. He was started on pemetrexed chemotherapy.

Discussion

In this case, the patient presented with progressive dyspnoea on atezolizumab treatment. The diagnosis of immunotherapy-induced pneumonitis was made after careful exclusion of other pulmonary conditions such as infection and malignancy. The symptoms improved on discontinuation of atezolizumab and a course of prednisone. However, the patient relapsed after the last dose of prednisone, possibly due to the initial severe grade of pneumonitis and the relative rapid tapering of prednisone. Modest improvement was achieved with a more prolonged steroid course. There are no clinical studies to guide how long to treat patients with steroids or what dose is appropriate.

Cancer immunotherapy aims to release the inhibition of immune surveillance so that tumour recognition can be activated and result in anti-tumour cytotoxic activity.² Immune checkpoint inhibitors are developed antibodies against PD-1/PD-L1 or cytotoxic T-lymphocyte-associated protein 4 (CTLA4), which are important regulators of T-cell immune response.³ So far, these inhibitors have been applied to a variety of anti-cancer regimens, including melanoma, NSCLC, renal cell carcinoma, bladder cancer and Hodgkin lymphoma. The PD-L1 inhibitor atezolizumab is approved by the US Food and Drug administration (FDA) as a second-line treatment for metastatic NSCLC progressing during or following platinum-containing chemotherapy. The phase three trial, OAK, showed that NSCLC patients who progressed on platinum-based chemotherapy received atezolizumab every 3 weeks and had significant improved overall survival compared with patients who received docetaxel.⁴ Notably, patients benefited from atezolizumab regardless of their PD-L1 expression. Therefore, although positive PD-L1 staining enriches anti-tumour responsiveness and VENTANA SP142 assay (detection of PD-L1 expression) is recommended as a complimentary diagnostic test, it is not a requirement for clinical utility of atezolizumab in progressing NSCLC post-chemotherapy.

Despite important clinical benefits, the checkpoint inhibitors are also associated with a broad spectrum of immune-related adverse events including systemic symptoms, dermatologic or mucosal toxicity, diarrhoea, hepatotoxicity, endocrinopathies and pneumonitis.⁵ The checkpoint inhibitor-induced pneumonitis occurs approximately 3%–5% of patients overall, mostly from anti-PD-1/PD-L1 antibodies, either alone or in combination with anti-CTLA-4 inhibitors.^6–9 In addition, the incidence of all-grade and severe pneumonitis (grade 3 or 4) in NSCLC patients seems to be higher compared with that in melanoma patients treated with immune checkpoint inhibitors.¹⁰ The incidence of pneumonitis between PD-1/PD-L1 treatment naïve and previously treated patients were significantly higher in the former group (4.3% vs 2.8% P=0.03).¹¹ The treatment duration prior to the diagnosis varies, typically from 9 days to 19 months, with a median of 2.8 months.¹⁰ In patients who received atezolizumab, the incidence was 3.7% (38/1027), the median time to onset was 3.3 months (from 3 days to 18.7 months) and the median duration was 1.4 months (0 days to 12.6+ months).¹²

The most common presenting symptoms of pneumonitis are cough and dyspnoea. Fever and chest pain were less common and nearly one-third of the patient did not have any symptoms.¹⁰ Among patients who developed pneumonitis, more than 50% also experienced other checkpoint inhibitor-related toxicity.¹⁰ Less than 1% of treated patients died with concurrent infection and tumour progression. In addition, current and former smokers and those with underlying pulmonary conditions tend to have poorer prognosis.¹¹ Nevertheless, there is no reliable biomarker to predict its occurrence or severity.¹³ Notably, pneumonitis could occur in patients with prior exposure to radiation therapy to the lung.¹⁴

Drug-induced pneumonitis is a diagnosis of exclusion and infections and malignancy must be ruled out. There are no specific features or biomarkers. A diagnosis must combine clinical findings, imaging as well as exclusion of other pathogenic causes of pneumonitis. Chest imaging such as X-ray and CT scan and pulse oximetry should be obtained. The radiographic pattern varies, including cryptogenic organising pneumonia (the most common type), hypersensitivity pneumonitis, non-specific interstitial pneumonia and acute respiratory distress syndrome.⁹ Diagnostic bronchoscopy with lung biopsy may be important for excluding competing diagnoses.

Treatments of drug-induced pneumonitis are selected based on the severity of symptoms.¹⁵ Clinical or diagnostic observations are often sufficient for asymptomatic patients (grade 1), as discontinuation of the drug with close monitor of radiographic progression will often resolve the cases. For patients who are symptomatic and have limited activity of daily living or 25%–50% of the lung parenchyma involved (grade 2), prednisone 1–2 mg/kg/day is given and tapered by 5–10 mg/week over four to 6 weeks. Possible infections should also be considered if there are any clinical indications. In more severe pneumonitis with or without life-threatening respiratory compromise (grade 3 and grade 4), patients often require hospitalisation and intravenous methylprednisolone 1–2 mg/kg/day. Infliximab, mycophenolate mofetil, intravenous immune globulin or cyclophosphamide is also added if there is no improvement 48 hours post-methylprednisolone treatment. Checkpoint inhibitors can be resumed for grade 1 and grade 2 pneumonitis. However, the drugs are permanently discontinued for severe and life-threatening pneumonitis (grade 3 and 4).

Drug-induced pneumonitis may reoccur with or without immunotherapy rechallenge. In a clinical report, 11 out of 43 patients with checkpoint inhibitor-induced pneumonitis experienced recurrent pneumonitis during drug holding and/or corticosteroid therapy after initial clinical improvement.¹⁰ Twelve patients received immunotherapy again after complete resolution of pneumonitis. Nine patients could tolerate rechallenge and three developed recurrent pneumonitis of the same grade compared with previous ones. Importantly, studies showed that some patients continued benefiting from the previous therapy even if they had previous adverse events due to checkpoint inhibitors.^{16 17} Therefore, a decision to restart checkpoint inhibitor relies on the severity of previous episodes, the overall status of cancer and the availability of alternative treatments. Additional studies need to be done to determine whether shorter durations of immunotherapy treatment would jeopardise the extent of survival benefit.

Learning points.

• Checkpoint inhibitor-induced pneumonitis may occur in 3%–5% of treated patients and is a diagnosis of exclusion.

• Prednisone is the current preferred treatment given typically over a 4-week to 6-week course.

• Pneumonitis may recur with or without checkpoint inhibitor rechallenge therapy. However, some patients may tolerate rechallenge therapy without developing pneumonitis.