Radiation-induced lung injury

INTRODUCTION Radiation-induced lung injury was first described in 1889, soon after the development of roentgenograms. The distinction between two separate types of radiation-induced lung injury, radiation pneumonitis and radiation fibrosis, was made in 1925. Both types of lung injury are observed today in patients who have undergone thoracic irradiation for the treatment of lung, breast, or hematologic malignancies. Radiation-induced damage to normal lung parenchyma remains the dose-limiting factor in chest radiotherapy, and can involve other structures within the thorax in addition to the lungs

A large body of literature describes the histopathologic, biochemical, kinetic, physiologic, and molecular responses of lung cells to ionizing radiation. However, the clinical diagnosis of radiation-induced lung injury is often complicated by the presence of other conditions, including malignancy, infection, and cardiogenic pulmonary edema Radiation-induced lung injury will be reviewed here. The cardiac effects of therapeutic radiation are discussed separately.

Factors that affect the reaction to radiation Many factors affect the development of radiation-induced lung disease.

  Radiobiological tissue architecture of the lung Traditionally, the lung parenchyma has been considered to be composed of functional subunits in a so-called parallel architecture, meaning that compromise to any given subvolume is likely equivalent in effect regardless of the location within the lung. In an organ with this type of parallel architecture, the mean dose to the organ is expected to be a robust predictor of complications, since the degree of overall injury is expected to be proportional to mean dose. Indeed, numerous clinical reports have correlated mean dose to lung with the risk of pneumonitis.

However, observations into the use of stereotactic body radiation therapy (SBRT) for lung tumors have suggested that a purely parallel view of lung architecture is likely an incomplete perspective. SBRT involves the carefully targeted administration of ultra-high doses within a few individual treatments and has the potential to provide greater biological effect than conventionally fractionated treatment. Observations from early clinical investigations have led to the recognition that SBRT-induced fibrosis in the vicinity of large central airways can cause obstruction to distal airflow, thus impeding ventilation distal to the occlusion. Thus, lung architecture may be considered to have a component of "serial" structure, in the sense that at least in selected circumstances, injury to one portion in sequence can cause notable effects downstream.

  Volume of lung irradiated The risk of radiation-induced injury is directly related to the volume of irradiated lung. In patients with breast cancer, for example, the risk of transient lung inflammation following adjuvant chest wall irradiation is approximately 5 percent. The risk is higher with increasing lung volume in the tangential fields, treatment to the supraclavicular, axillary apex, and internal mammary regions, and the use of concurrent compared to sequential chemotherapy (8.8 versus 1.3 percent in one series). In one report, when tangential beam irradiation was utilized following breast-conserving surgery, pneumonitis was observed only when >10 percent of the lung was irradiated

  Dosage of radiation The dose of radiation delivered to the lung is a critical factor in determining if injury will occur. As noted above, mean dose can be a predictor of the risk of radiation pneumonitis, as can the V20, defined as the volume of normal lung (total lung volume minus planning target volume for radiotherapy) that receives more than 20 Gy. One group analyzed the dose-volume histogram (DVH) and other clinical information in a series of 99 patients treated definitively for inoperable non-small cell lung cancer (NSCLC). On multivariate analysis the V20 was the single independent predictor of the development of Grade 2 or higher pneumonitis, meaning the need for steroids, diuretics, or oxygen. For V20 <22 percent, no pneumonitis was observed. For V20 from 22 to 31 percent, there was a 16 percent rate of grade 2-4 toxicity: Higher rated were observed with higher V20, including several instances of grade 5 (fatal) pneumonitis in patients with V20 > 35 percent. For this reason both the Radiation Therapy Oncology Group (RTOG) and Southwest Oncology Group (SWOG) have specified upper limits of V20 in the range of 30 to 35 percent for patients treated on recent group protocols.

  Time-dose factor A high cumulative dose of radiation was also associated with an increased risk of lung injury in the above study. Twice daily fractionation of the dose appears to reduce the risk compared to administration of the same total daily dose as a single fraction

  Concurrent chemotherapy Several chemotherapeutic agents are known sensitizers to radiotherapy, including doxorubicin, taxanes, dactinomycin, bleomycin, cyclophosphamide, vincristine, mitomycin, and recombinant interferon-alpha  Patients receiving these drugs are at a higher risk of developing radiation-induced lung injury. In addition, several of the drugs themselves are associated with lung injury.

Concurrent rather than sequential chemotherapy appears to increase the risk, especially in women undergoing anthracycline-based adjuvant chemoradiotherapy for breast cancer. In one report, the risk of pneumonitis in women treated with a supraclavicular field and sequential versus concurrent chemotherapy was 1.3 versus 9 percent, respectively. Concurrent anthracycline-based chemotherapy and radiation are generally avoided in the treatment of breast cancer.

It is unclear whether sequential administration of paclitaxel and RT diminishes the risk of radiation pneumonitis as compared to concurrent treatment. Women who receive taxanes as a component of their adjuvant therapy for breast cancer may need to have a lower volume of lung included in the radiation field. This topic is addressed in detail elsewhere.

  Other factors Prior thoracic irradiation, volume loss due to lung collapse, younger age, smoking history, poor pretreatment performance status, poor pretreatment lung function, chronic obstructive pulmonary disease, female sex, and steroid withdrawal during radiotherapy have also been associated with an increased risk of radiation pneumonitis

Smoking has been reported as a risk-modifying factor in some, but not all, studies. A 1998 retrospective study compared the incidence of clinical radiation pneumonitis in active smokers and nonsmokers, among 405 women who underwent radiotherapy for treatment of breast or esophageal cancer The authors found that none of the subjects who were active cigarette smokers developed clinical pneumonitis following irradiation.

These findings are intriguing because other inflammatory lung diseases with a predominance of lymphocytes, such as hypersensitivity pneumonitis, are also uncommon in smokers.

Although several studies suggest that concurrent rather than sequential use of tamoxifen increases the rate of pulmonary fibrosis, this has not been a consistent finding, and in general, higher rates of symptomatic pneumonitis have not been seen.

PATHOLOGY The pathologic and clinical changes in the lung following irradiation may be divided into five phases

  bullet The immediate phase begins within hours to days following radiation exposure, and is generally asymptomatic. It is characterized by hyperemic, congested mucosa with leukocytic infiltration and increased capillary permeability, resulting in pulmonary edema. An exudative alveolitis follows, accompanied by tracheal bronchial hypersecretion and degenerative changes in the alveolar epithelium and endothelium. Type I alveolar epithelial cells (pneumocytes) are sloughed, and alveolar surfactant levels are increased

During the next phase (the latent phase), thick secretions accumulate due to an increase in the number of goblet cells combined with ciliary dysfunction.

The third phase (acute exudative phase) is clinically referred to as radiation pneumonitis. It occurs three to twelve weeks following exposure and consists of sloughing of endothelial and epithelial cells, with narrowing of the pulmonary capillaries and microvascular thrombosis. Hyaline membranes form as a result of alveolar pneumocyte desquamation and leakage of a fibrin-rich exudate into the alveoli. Giant cells may be seen along the endothelium, and type II pneumocytes become hyperplastic with marked atypia.

In the fourth phase (intermediate phase), there may be resolution of the alveolar exudate and dissolution of the hyaline membranes, or there may be collagen deposition by fibroblasts, which results in thickening of the interstitium. Fibroblasts, probably of bone marrow origin, migrate into and proliferate within the alveolar walls and spaces

A final phase consists of fibrosis. It may be evident as early as six months following irradiation, and can progress over years. There is an increase in the number of myofibroblasts within the interstitium and alveolar spaces, along with an increase in collagen. The anatomic narrowing of alveolar spaces results in diminishing lung volume; vascular subintimal fibrosis and distortion cause a loss of capillaries. Traction bronchiectasis, complicated by chronic infection, can develop.

A late histopathologic appearance consistent with bronchiolitis obliterans organizing pneumonia has also been reported, particularly when injury is seen in the contralateral lung

EPIDEMIOLOGY The incidence of radiation pneumonitis varies depending upon the particular regimen used and upon the radiation field. In addition, there is a discrepancy between the frequency of clinically apparent pneumonitis and radiographic evidence of lung disease. One review, for example, noted the following frequencies of radiation-induced lung injury

  bullet In patients with breast cancer, clinical pneumonitis occurred in 0 to 10 percent, while radiographic abnormalities were present in 27 to 40 percent.

In patients with lung cancer, clinical pneumonitis occurred in 5 to 15 percent of patients, while radiographic abnormalities were present in 66 percent. It is unclear to what extent the latter were due to irradiation versus tumor.

In patients with mediastinal lymphoma, clinical pneumonitis did not occur, although radiographic abnormalities were present in 60 to 92 percent.

The incidence of radiation pneumonitis may differ when protocols employing non-external beam radiation are used. As an example, in a study of 46 patients who received high dose endobronchial brachytherapy for malignant airway obstruction, radiation pneumonitis was only detected in those subjects who also received external beam therapy. A second study of 80 patients with inoperable hepatic tumors who underwent 90-Yttrium microsphere irradiation found that 5 developed radiation pneumonitis. The occurrence of pneumonitis correlated with the degree of pulmonary shunting of the microspheres: pneumonitis occurred in five of nine patients with at least a 13 percent shunt to the pulmonary circulation (as assessed by 99m technetium-labeled macroaggregated albumin), compared to no patients with a shunt below 13 percent.

CLINICAL MANIFESTATIONS Symptoms of radiation pneumonitis usually have an insidious onset and include the following:

  bullet An early nonproductive cough.
Dyspnea may only occur with exertion, or may be described as an inability to take a deep breath.
Fever is usually low grade, but can be more pronounced in severe cases.
Chest pain may be pleuritic or substernal and can represent pleuritis, esophageal pathology, or rib fracture.
Malaise and weight loss may be observed.

Physical signs include the following:

  bullet Crackles or a pleural rub may be heard; in some cases auscultation is normal.
Dullness to percussion may be detected as a result of a small pleural effusion; this occurs in about 10 percent of patients. Effusions often cause no symptoms and may spontaneously remit. In contrast to malignant effusions, radiation-induced effusions do not increase in size after a period of observed stability.
Skin erythema may outline the radiation port but is not predictive of the occurrence or the severity of radiation pneumonitis.
Tachypnea, cyanosis, or signs of pulmonary hypertension may be seen in more advanced cases.


DIAGNOSIS The diagnosis of radiation pneumonitis is complicated by the fact that the inflammatory lung injury may or may not produce symptoms. In addition, pulmonary symptoms must be differentiated from those of infectious, cardiac (pericarditis), or esophageal (esophagitis, tracheoesophageal fistula) origin

Laboratory studies No commonly employed laboratory test predicts the development of radiation pneumonitis. A low-grade polymorphonuclear leukocytosis is often present, and the sedimentation rate, serum LDH, and C-reactive protein may be modestly elevated, but these findings are nonspecific.

Serum KL-6, a sialylated carbohydrate epitope that is highly expressed in bronchial epithelial cells and type II pneumocytes, was found to be elevated in six patients with lung cancer who developed radiation pneumonitis compared to six patients who did not. Although KL-6 is not specific for radiation pneumonitis, it is usually not elevated in patients with bacterial pneumonia.

Further analyses on larger numbers of patients will be needed to assess the sensitivity, specificity, and clinical utility of serum KL-6 and TGF-b1 concentrations as predictors of radiation pneumonitis.

Chest imaging studies Chest radiographic abnormalities following thoracic irradiation need to be distinguished from other pulmonary diseases, such as infection, lymphangitic or direct extension of tumor, drug-induced pneumonitis, hemorrhage, and cardiogenic edema

  bullet Chest roentgenograms may be normal in symptomatic subjects during the subacute phase of radiation pneumonitis.
Perivascular haziness is an early radiation-induced abnormality on chest roentgenogram, often progressing to patchy alveolar filling densities.
Radiographs taken during the chronic phase of radiation pneumonitis may show volume loss with coarse reticular or dense opacities.
A straight line effect, which does not conform to anatomical units but rather to the confines of the radiation port, is often seen and is virtually diagnostic of radiation-induced lung injury.
Small pleural effusions and rib fractures may be seen, but lymphadenopathy does not occur.

There are a few case reports of radiographic abnormalities attributed to radiation pneumonitis outside of the port, and even in the contralateral lung. This may be due to radiation scatter, lymphatic obstruction, or immunologic (hypersensitivity-like) mechanisms.

Chest computed tomography (CT) is more sensitive than the chest roentgenogram in detecting subtle lung injury following radiation treatment. The nonanatomic straight edge effect described above is also evident on CT. However, most patients with radiation-induced pulmonary injury do not require a chest CT to make the diagnosis; this test should be reserved for subjects in whom the diagnosis is uncertain.

Nuclear medicine studies Gallium scanning has been used to evaluate the lungs of patients with pulmonary fibrosis. However, this study adds little to the assessment, since other causes of abnormal parenchymal densities in the immunocompromised host (infection, tumor) also take up gallium. In addition, analysis of regions along the midline is hampered by uptake in the sternum and spine.

Indium may have several advantages over other radiographic modalities employed for the evaluation of radiation pneumonitis, including better resolution than ventilation-perfusion (V/Q) scanning and lack of interference from mediastinal structures. Indium scans also provide the option of single photon emission computed tomography (SPECT) imaging, which can be used to assess superficial versus deep lung injury and quantify abnormal radiographic areas.

Scanning using indium-111-pentreotide, a somatostatin analogue, was studied in 10 subjects with suspected radiation-induced lung injury [54]. Indium scans were positive in all the symptomatic patients and correlated with abnormal areas seen on chest roentgenogram and V/Q scans. Indium-111-scanning is currently employed only as a research tool; further studies will be necessary to evaluate the utility of this imaging modality in the diagnosis and management of patients with radiation pneumonitis.

Pulmonary function tests Pulmonary function testing generally demonstrates a reduction in lung volumes (TLC, FVC, RV), diffusing capacity, and lung compliance in patients with radiation-induced lung injury. Tidal volumes are also decreased, and the respiratory rate may be elevated. Airway resistance is usually normal or slightly elevated. As with other causes of fibrotic lung disease, there is a normal or low arterial partial pressure of oxygen and carbon dioxide at rest, with an increased alveolar-arterial oxygen gradient. In early or mild cases, abnormal oxygenation may only be manifest when arterial blood gases are drawn during exercise. (

The diffusing capacity for carbon monoxide (DLCO or transfer factor) is usually depressed in patients with radiation-induced lung damage, but this finding is nonspecific. One trial suggested that failure of the DLCO to increase from the nadir value following myeloablative chemotherapy was more closely associated with the risk of progressive pulmonary dysfunction during subsequent irradiation than other parameters of lung function

Bronchoscopy Bronchoalveolar lavage fluid (BALF) findings are not specific, usually showing an increased number of leukocytes (predominantly lymphocytes). The majority of BAL lymphocytes post-irradiation are CD4+. Lymphocyte numbers are increased in both the irradiated and nonirradiated lung. The number of neutrophils, eosinophils, and macrophages may also be increased. In addition, there are more activated lymphocytes in the BALF of patients after irradiation than in controls. Finally, there is an increase in BALF total cell count, percentage of lymphocytes, and ICAM-1 positive T-cells in irradiated subjects with abnormal chest roentgenograms compared to those with normal chest films.

Tissue specimens are only occasionally required in the evaluation of patients suspected to have radiation pneumonitis. Transbronchial and transthoracic needle biopsy specimens are too small to establish a diagnosis, but may be useful for ruling out infection or lymphangitic spread of tumor in cases that are clinically atypical for radiation-induced lung injury.

TREATMENT There are no prospective controlled studies evaluating the efficacy of therapies for radiation pneumonitis in humans. Nevertheless, many experts recommend the use of corticosteroids for symptomatic patients with a subacute onset of radiation lung injury

Corticosteroids Prednisone (at least 60 mg/day) is generally given for two weeks, with a gradual taper over three to twelve weeks, although the guidelines for tapering are poorly defined. Corticosteroids may be effective in the treatment of radiation-associated bronchiolitis obliterans organizing pneumonia; however, symptoms and radiographic abnormalities tend to recur with discontinuation of therapy

The recommendation for prednisone use is based upon data derived from murine models of radiation pneumonitis that demonstrated a protective effect of corticosteroids. These drugs may reduce radiation pneumonitis via reduction in inflammation, inhibition of TNF-induced nitric oxide-mediated endothelial cell cytotoxicity, or a lymphotoxic mechanism. The prophylactic administration of corticosteroids, antibiotics, or heparin has not been beneficial in reducing the incidence of radiation pneumonitis

Pentoxifylline Pentoxifylline is a xanthine derivative that inhibits platelet aggregation and enhances microvascular blood flow; it also has immunomodulating and antiinflammatory properties that are probably mediated by inhibition of tumor necrosis factor (TNF) and interleukin 1. Pentoxifylline is beneficial for the treatment of radiation-induced fibrosis involving the skin and subcutaneous tissues, and also inhibits experimental bleomycin-induced pulmonary fibrosis in rats, possibly via its anti-TNFa effects.

A modest benefit for pentoxifylline in the prevention of radiation-induced lung toxicity was suggested in a trial in which 40 patients undergoing RT for breast or lung cancer were randomly assigned to pentoxifylline (400 mg three times daily) or placebo during treatment. During six months of follow-up, the number of patients with grade 2 or 3 pulmonary toxicity was significantly less in the pentoxifylline group (20 versus 50 percent). Four patients in the placebo group (30 percent) with grade 3 lung impairment required steroids and oxygen, while only one patient who received pentoxifylline (5 percent) had clinical impairment from grade 2 pulmonary toxicity. The pentoxifylline group had a significantly better diffusing capacity for carbon monoxide at both three and six months following therapy (73 versus 58, and 72 versus 57 percent at three and six months, respectively), but there were no significant differences in spirometry between the two groups. Although intriguing, these results require independent confirmation.

Inhibition of collagen synthesis Since excess collagen deposition is a key histopathologic feature of radiation fibrosis, drugs that inhibit collagen synthesis, such as colchicine, penicillamine, IFN-gamma, or pirfenidone, may have the potential to modify the progression of fibrosis. However, there are no controlled studies with these agents in humans with radiation-induced lung injury. Azathioprine and cyclosporin A were both effective in treating the symptoms of radiation pneumonitis in single case reports; these agents may be considered in patients who do not tolerate corticosteroids

Amifostine Amifostine, a cytoprotective agent that appears to shield normal tissues from the toxic effects of chemotherapy and radiotherapy, has been investigated as an agent that may reduce the incidence and severity of radiation pneumonitis. Approved as therapy for xerostomia in patients following neck irradiation, amifostine is a prodrug that is dephosphorylated by alkaline phosphatase in tissues to a pharmacologically-active free thiol metabolite, which can reduce the toxic effects of chemotherapy, and act as a scavenger of free radicals generated in tissues exposed to radiation.

Early evidence suggests that amifostine may decrease radiation-induced pulmonary injury without diminishing the therapeutic effect of radiation. This was demonstrated in a randomized controlled trial of radiation plus amifostine compared to radiation alone in 146 patients with locally advanced lung cancer. Amifostine therapy was associated with a significant decrease in pneumonitis (9 versus 43 percent) and also grade 3 esophagitis; in addition, patients receiving amifostine were less likely to have evidence of fibrosis or pneumonitis on chest CT scan (16 versus 49 percent)

However, these results have not been replicated. In a randomized study involving 243 patients receiving carboplatin, paclitaxel, and thoracic radiotherapy, amifostine did not reduce the rate of grade 3 esophagitis. Rates of pneumonitis in that trial have not been reported.

Captopril Captopril has been shown to reduce radiation-induced lung fibrosis in rats. One retrospective comparison of the incidence of irradiation-induced lung injury between subjects taking angiotensin converting enzyme (ACE) inhibitors and those who were not, failed to reveal a protective effect from the administration of ACE inhibitors. However, serum concentrations of ACE inhibitors used by subjects in the retrospective human study would be expected to be considerably less than those achieved in the animal model.

PROGNOSIS Significant improvements in the perfusion and ventilation of radiation-injured lung tissue may be noted from 3 to 18 months after radiation therapy. After 18 months, however, further significant improvement appears unusual