Authors: Nahyun Celina Jo, Girish S. Shroff, Jitesh Ahuja, Rishi Agrawal, Melissa C. Price, Carol C. Wu
Categories: Thoracic Imaging, COVID-19-related complications, Chemotherapy complication, Immunotherapy, Radiation therapy, Radiation-induced pneumonitis
Source: Korean Journal of Radiology
Radiation recall pneumonitis is an inflammatory reaction of previously radiated lung parenchyma triggered by systemic pharmacological agents (such as chemotherapy and immunotherapy) or vaccination. Patients present with non-specific symptoms such as cough, shortness of breath, or hypoxia soon after the initiation of medication or vaccination. Careful assessment of the patient’s history, including the thoracic radiation treatment plan and timing of the initiation of the triggering agent, in conjunction with CT findings, contribute to the diagnosis. Once a diagnosis is established, treatment includes cessation of the causative medication and/or initiation of steroid therapy. Differentiating this relatively rare entity from other common post-therapeutic complications in oncology patients, such as recurrent malignancy, infection, or medication-induced pneumonitis, is essential for guiding downstream clinical management.
Keywords: Radiation-induced pneumonitis, Radiation therapy, Immunotherapy, COVID-19-related complications, Chemotherapy complication
Radiotherapy (RT) is commonly used to treat primary lung cancer and other thoracic malignancies. It is also effective in controlling intrathoracic metastatic disease in selected patients. After completion of RT, patients often undergo follow-up chest CT to evaluate treatment response, complications, and recurrent malignancy. It is important for radiologists to understand the evolving post-RT imaging findings to differentiate the expected changes from progressive/recurrent disease, infection, or medication-induced pneumonitis [1,2], as the subsequent evaluation and management differ significantly [3].
Radiation recall pneumonitis (RRP) is a subtype of radiation-induced pneumonitis, in which an inflammatory reaction in previously irradiated lung parenchyma is triggered and worsened by an inciting agent. RRP has been reported to be more commonly associated with immunotherapy [4,5,6,7,8] than conventional chemotherapy agents. Vaccination has emerged as a potential trigger for RRP. RRP can be diagnosed by using a combination of imaging findings and a history of use of known triggering agents after thoracic RT. Radiologists play a key role in detecting and diagnosing RRP. In this pictorial essay, we demonstrate the imaging manifestations of RRP and provide tips for differentiating this entity from its potential mimics in oncology patients.
The pathophysiology of RRP remains unclear. Hypotheses of underlying abnormalities include RT-induced damage to stem cells, reduction of the inflammatory response threshold, and/or increase in vascular permeability, resulting in local accumulation of causative medications post-RT and elevated focal lung toxicity [4,9,10].
Several offending agents of RRP have been described, including chemotherapeutic agents such as taxanes, anthracyclines, gemcitabine, and oxaliplatin [11,12,13,14,15]. With the introduction of targeted and immune therapies in cancer treatment, agents such as erlotinib [16], everolimus [17], sunitinib [18], or nivolumab [5,19] have been identified as causes for RRP. Recently, the COVID-19 vaccination was reported as a novel triggering event [20,21].
In a study of 80 patients treated with programmed death-1 (PD-1)/PD ligand-1 (PD-L1) inhibitors for advanced lung cancer who had previously received thoracic RT, 15 patients (18.8%) were found to have RRP. Results showed that smoking history, presence/severity of emphysema, tumor histology, prior chemotherapy, and treatment goals (palliative vs. curative) were not statistically significant risk factors for RRP development. RRP was observed in patients who underwent conventional and stereotactic body RT [22]. Immune checkpoint inhibitor (ICI)-related pneumonitis was associated with RRP in one-third of patients and developed concomitantly or after RRP.
Typically, RRP occurs soon after the initiation of the inciting agent, but may also occur after several courses of treatment [23]. In a recent study by Cousin et al. [22], the median time between ICI initiation and RRP occurrence was 61 days (range, 4–520 days). RRP symptoms are non-specific and may include shortness of breath, coughing, chest tightness, chest pain, and low-grade fever. A bronchoalveolar lavage may reveal an elevated lymphocyte count. Management involves cessation of triggering agents, steroid therapy, and supportive care. No standard steroid dose has been established for the treatment of RRP. When compared to “typical” radiation pneumonitis, RRP resolves more quickly and without residual scarring. Rechallenge with the inciting agent does not always elicit a recurrent reaction [24].
In patients who undergo thoracic RT for primary or metastatic malignancies, consolidative and ground-glass opacities often develop within six months of RT and evolve over time, with a decrease in the extent of pulmonary opacities and an increase in the associated volume loss, architectural distortion, and traction bronchiectasis. These imaging findings frequently stabilize two years post-RT [25]. RRP should be considered when ground-glass and/or consolidative opacities that conform to the radiation treatment plan develop or increase months or even years after the completion of RT. Comparison of the location of pulmonary opacities to the radiation dosimetry map and subsequent correlation with the initiation of new inciting agents is essential for clinically diagnosing RRP (Figs. 1, 2, 3).
Fig. 1 Radiation recall pneumonitis. A 57-year-old female with left upper lobe squamous cell carcinoma initially treated with left pneumonectomy. Subsequently, the mediastinal nodal recurrence was treated with chemoradiation. A: The RT plan depicts areas of the right lung that received scattered radiation. B: CT a year after RT showed a small right lower lobe opacity consistent with post-radiation changes. Subsequently, the patient was placed on immunotherapy with durvalumab and returned with cough and shortness of breath. C: CT 13 months after initiation of durvalumab showed new and increased opacities conforming to the areas of radiated lung. D: Follow-up CT after a course of steroid and cessation of durvalumab showed resolution of imaging findings. RT = radiotherapy
Fig. 2 Radiation recall pneumonitis. A 62-year-old male with stage III squamous cell carcinoma was initially treated with chemoradiation. A: The radiation dosimetry map showed the distribution of radiation in the right lung. B: CT two years post-radiotherapy showed stabilized radiation fibrosis in the right lung. Subsequently, the patient was placed on pembrolizumab, abraxane, and carboplatin followed by atezolizumab for subcarinal nodal disease. The patient developed a cough and shortness of breath. C: A chest CT obtained six months after initiation of atezolizumab demonstrated new and increased opacities within the previously radiated areas of the right lung (within the most peripheral and lowest radiation level encompassed by the purple line on the dosimetry map in A). D: The symptoms and the imaging findings improved with slow steroid taper.
Fig. 3 Radiation recall pneumonitis. A 72-year-old male with stage IV lung adenocarcinoma was treated with chemoradiation. A: The radiation dosimetry map showed the distribution of radiation in both lungs. B: CT three years after completion of radiation therapy showed a stable post-radiation opacity in the right lung. Pembrolizumab was added to maintenance chemotherapy. C: A month after addition of pembrolizumab, CT showed increased consolidation in the right lower lobe and new opacities in the medial left upper lobe conforming to the areas of radiation, thought to be due to radiation recall pneumonitis. Imaging findings and symptoms improved after discontinuation of pembrolizumab and a course of steroid therapy.
Patients with cancer often receive a variety of systemic therapeutic agents, including conventional chemotherapy, targeted therapy, and/or immunotherapy during the course of their disease. Many of these agents have been associated with lung injury or pneumonitis in various patterns, such as organizing pneumonia or non-specific interstitial pneumonitis [26]. A sarcoid-like reaction, another pattern of pneumonitis, may be observed with any systemic cancer therapy and has been observed more frequently in recent years in association with immunotherapy. Histopathologically, sarcoid-like reactions are identical to those of sarcoidosis in the presence of non-caseating granulomas. The clinical symptoms of medication-induced pneumonitis are often non-specific and similar to those of RRP.
Pneumonitis related to systemic therapy tends to be bilateral, with multilobar involvement. Unlike in RRP, the areas of consolidation, ground-glass opacities, and/or nodules on CT are not limited to the RT treatment plan (Fig. 4). On imaging, sarcoid-like reactions typically manifest as hilar or mediastinal lymphadenopathies. Upper lobe-predominant peribronchovascular and subpleural opacities and nodules may also be seen [27,28,29,30,31]. However, in patients receiving multi-modality therapies, the distinction between RRP and pneumonitis or sarcoid-like reactions can sometimes be difficult if CT findings have an asymmetrical or limited distribution (Fig. 5) [32], or if the patient has pre-existing lung disease or pulmonary lymphangitic carcinomatosis. Furthermore, RRP and medication-induced pneumonitis can coexist on CT scans (Fig. 6) [32]. Other ancillary findings such as pleural effusions and reactive hilar and mediastinal lymphadenopathies can occur in both RRP and medication-induced pneumonitis [31,33]. In other cases, these findings may mimic cancer progression or infection [32]. Tissue sampling and/or follow-up imaging may be required to differentiate these entities.
Fig. 4 Radiation pneumonitis and immunotherapy-related pneumonitis. A 60-year-old male with stage IV right upper lobe lung adenocarcinoma was treated with chemotherapy and nivolumab for two years followed by radiation therapy of the right upper lobe primary tumor and re-initiation of nivolumab. A: The radiation dosimetry map showed the distribution of radiation in the right lung. B: CT seven months following re-initiation of nivolumab showed stable right upper lobe consolidative and left upper lobe ground glass opacities consistent with post-radiation changes. C: The same CT showed new bilateral multifocal nodular opacities outside of radiotherapy zone thought to represent immunotherapy-related pneumonitis of organizing pneumonia pattern. D: Nivolumab was discontinued and the bilateral lower lobe pulmonary opacities improved on the follow-up CT.
Fig. 5 Sarcoid or sarcoid-like reaction. A 39-year-old female with left breast cancer was treated with chemotherapy, mastectomy, and radiation therapy two years prior, and also received monthly goserelin acetate implants. A-C: Chest CT showed new bilateral mediastinal and hilar adenopathy (A) as well as new subtle interlobular septal and subpleural nodularity limited to the left upper lobe (arrows in B, C). Fine-needle aspiration confirmed absence of metastatic disease and presence of granuloma suspicious for sarcoid or sarcoid-like reaction. The asymmetric and focal subpleural distribution of pulmonary findings could lead to consideration of radiation recall pneumonitis but the presence of symmetrical lymphadenopathy and perilymphatic distribution of nodules directed the radiologist to the correct diagnosis.
Fig. 6 RRP and immunotherapy-related pneumonitis. A 61-year-old female with vaginal squamous cell carcinoma post-pelvic exenteration four years prior, and biopsy-proven metastases in the right middle lobe and right hilar lymph node. A: The radiation dosimetry map showed distribution of radiation in the right lung and hilum. Concurrent chemotherapy was given. B: CT five months after completion of RT showed right perihilar opacities consistent with post-radiation changes. Pembrolizumab was initiated in the same month. C, D: CT three months after initiation of pembrolizumab demonstrated markedly increased right lung opacities conforming to the radiated region and compatible with RRP, and also new bilateral subpleural and peribronchovascular nodular opacities outside of the RT zone due to immunotherapy-related pneumonitis. Some of the nodular opacities exhibited the reversed-halo sign suggestive of the organizing pneumonia pattern of pneumonitis. RRP = radiation recall pneumonitis, RT = radiotherapy
Recurrent disease is a major differential consideration for new or increased opacities within the RT area. RRP was misinterpreted as tumoral progression (8/15 cases, 53.3%) or infection (1/15 cases, 6.7%) in the study by Cousin et al. [22]. When the location of the increased opacity corresponds to the location of the prior tumor, and especially when the opacity has a mass-like appearance, convex margins, and pre-existing air bronchograms (related to radiation fibrosis) are no longer present, recurrence should be strongly considered (Fig. 7). ^18^F-fludeoxyglucose (FDG) PET/CT is helpful, as it frequently demonstrates intense focal FDG uptake at the site of local tumor recurrence. It should be noted that pulmonary opacities related to radiation pneumonitis including RRP can also be metabolically active in the absence of recurrent disease [34]. In the absence of contraindications, tissue sampling should be performed to confirm the diagnosis of recurrent disease.
Fig. 7 Recurrence. A 72-year-old female patient with right lower lobe squamous cell carcinoma was treated with RT and nivolumab. A: The dosimetry map showed the original tumor and the RT plan. B: One year post-RT, CT showed right lower lobe opacity with associated volume loss and architectural distortion consistent with radiation fibrosis. C: Evaluation six months later showed enlargement of the right lower lobe opacity with subtle convexity of previously concave margins (arrow) and filling-in of the previous air-bronchogram concerning for recurrence. D: The opacity demonstrated avid focal fludeoxyglucose uptake in the subsequent PET/CT. A subsequent biopsy revealed recurrent disease. RT = radiotherapy
Infectious causes should always be considered when new or increasing pulmonary opacities are identified on imaging, particularly in patients with symptomatic cancer. When radiologists encounter new pulmonary opacities in patients who have previously received thoracic RT, a comparison of the dosimetry map and timing of RT is always helpful. If new pulmonary opacities are present outside the RT zone, radiation pneumonitis and RRP can be effectively excluded. However, if pulmonary opacities are ipsilateral to the radiated lung and RT dosimetry maps are not available for review, distinguishing between RRP and pneumonia based on CT alone can be difficult. Tree-in-bud opacities, when present, are suggestive of infection or aspiration, and are not typically observed in RRP. Clinical symptoms, such as fever, cough, and shortness of breath, can be seen in both pulmonary infection and RRP, and are not usually helpful. According to the American Society of Clinical Oncology clinical practice guideline for management of lung toxicities related to ICIs, infection should be excluded in symptomatic patients (grade 2 or higher) [33]. The infection workup includes nasal swabs, sputum culture and sensitivity tests, blood culture and sensitivity tests, and urine culture and sensitivity tests. In patients requiring hospitalization or >50% lung involvement on imaging (grade 3 or 4 toxicity), further investigation can be performed using bronchoscopy with bronchoalveolar lavage, with or without transbronchial biopsy.
RRP is an inflammatory reaction triggered by chemotherapy, immunotherapy, or vaccination within an area of previously radiated lung parenchyma. In patients with an appropriate clinical history, the correlation of CT findings with the radiation treatment plan to assess overlap can help suggest the diagnosis of RRP. In cases where imaging and clinical information are insufficient to draw a conclusion, a multidisciplinary approach should be considered to assess the need for further diagnostic studies such as bronchoalveolar lavage or biopsy to exclude diagnoses such as infection or recurrent malignancy.