Lung Abnormalities at Multimodality Imaging after Radiation Therapy for Non–Small Cell Lung Cancer

LEARNING OBJECTIVES FOR TEST 4

After reading this article and taking the test, the reader will be able to:

•.	Recognize the CT features of radiation-induced lung injury at various intervals after 3D CRT and SBRT for non–small cell lung cancer.
•.	Describe the relation between the CT pattern of radiation-induced lung injury and the radiation portals and doses used.
•.	Differentiate radiation-induced lung injury from residual or recurrent lung cancer at PET/CT after 3D CRT or SBRT.

Introduction

Radiation therapy remains a valuable treatment modality for non–small cell lung cancer, despite advances in chemotherapy (1,2). Radiation therapy may be used with various goals, depending on the disease stage. Patients with medically nonoperable stage I–II disease may be candidates for radiation therapy with curative intent (3). For resected or resectable stage IIIA non–small cell lung cancer with mediastinal nodal involvement (N2), radiation therapy is used as an adjuvant treatment to improve local control (4). In locally advanced unresectable stage III non–small cell lung cancer (ie, stages IIIA N2 and IIIB), radiation therapy, often combined with chemotherapy, is the primary local treatment, although long-term outcomes are generally poor because of local and systemic recurrences (5). Radiation therapy is also used with palliative intent in patients with locally advanced or metastatic lung cancer, to relieve pain and symptoms and preserve quality of life (6).

Evidence suggests that higher doses of radiation improve local tumor control (2,7). Local tumor control is important to prevent metastatic dissemination and prolong survival. Traditional radiation techniques result in the delivery of therapeutic radiation doses to a broad area, beyond the primary tumor margins (8). Because a greater quantity of normal lung tissue is included in the area treated, the risk of side effects increases and the maximum radiation dose that can be used is limited (8,9).

In efforts to reduce the toxic effects and improve the results of radiation therapy, innovative techniques such as three-dimensional (3D) conformal radiation therapy (CRT) and stereotactic body radiation therapy (SBRT) were developed over the past decade (10). In both CRT and SBRT, multiple beams are used to generate a dose distribution that conforms tightly to the target volume (8,11). SBRT, in particular, is used to deliver hypofractionated high-dose tumoricidal x-rays to stage I non–small cell lung cancer without irradiating regional lymph nodes (12). Both techniques allow delivery of the maximum radiation dose to the entire tumor volume while minimizing the exposure of normal tissues to radiation (8,12). The more recently developed intensity-modulated radiation therapy and four-dimensional radiation therapy techniques have brought further improvements (increased conformality of the external beam to the target, and reduction in errors due to patient positioning and respiration-related motion, respectively) (10). In helical tomotherapy, one offshoot of intensity-modulated radiation therapy, the geometry of a helical CT scanner is integrated with a linear accelerator to deliver highly conformal radiation doses in an intensity-modulated distribution (13). Intensity-modulated and four-dimensional radiation therapies have not yet been shown definitively to improve outcomes (10), and 3D CRT and SBRT are still the most widely used techniques for clinical treatment of non–small cell lung cancer.

Depending on patient-specific factors, radiation dose, and tumor site, radiation-induced lung injury may occur (14). Two distinct clinical, pathologic, and radiologic phases of radiation-induced lung injury are recognized: an early phase of radiation pneumonitis, which usually occurs between 1 and 6 months after the completion of radiation therapy (11,15–17); and a later phase of chronic radiation fibrosis, which usually occurs between 6 and 12 months after the completion of radiation therapy (8,14,15). Radiologic manifestations of lung injury after conventional radiation therapy (typical manifestations) usually correspond to the margins of the irradiated field. The complex portal configurations in 3D CRT and SBRT result in lung abnormalities that differ from these typical manifestations in regard to their morphologic characteristics, extent, distribution, and location (atypical manifestations) (14,15). The main differential diagnoses include various conditions that may occur in patients treated for lung cancer, such as infections, lymphangitic carcinomatosis, local recurrence of malignancy, and radiation-induced tumors. Awareness of the therapeutic technique and radiation portals used and the associated possible radiologic pattern of radiation-induced lung injury is required to accurately interpret findings of lung abnormalities at follow-up imaging in patients treated with 3D CRT and SBRT.

After a brief technical description of 3D radiation therapies, the article surveys the spectrum of lung abnormalities that may be seen at CT after treatment with 3D CRT or SBRT and discusses the role of positron emission tomography (PET) performed with fluorine 18 fluorodeoxyglucose (FDG) in distinguishing between radiation therapy–related lung parenchymal abnormalities and residual or recurrent lung cancer. Specific topics discussed include the clinical and pathologic features of radiation-induced lung injury, the CT appearances of 3D CRT- and SBRT-induced lung injury, the differential diagnosis of posttreatment lung injury, and the role of PET/CT.

Three-dimensional Radiation Therapies

In conventional radiation therapy, the daily radiation dose is delivered to the tumor in two parallel beams with opposed orientations (eg, anteroposterior and posteroanterior beams, with or without oblique angulation), usually totaling 2 Gy per field combination per day (8). Because of the limited beam orientation, relatively large volumes of normal tissues adjacent to the treatment field (including the mediastinum, chest wall, and adjacent lung) often are irradiated. Doses of more than 60 Gy, which are desirable to improve tumor local control, have not been regularly administered with this technique because of the high risk of radiation-induced injury to these surrounding tissues (8,15).

In 3D CRT, a 3D image reconstructed from CT data is used to determine the target volume to be irradiated. In lung cancer treatment planning, the target volume generally includes an additional 1- to 2-cm margin around the tumor edge and a 1-cm margin around regional lymph nodes to compensate for respiratory motion and variations in set-up between daily treatment sessions (8,15). A computer planning system designs beam arrangements with various orientations depending on the 3D configuration of the tumor (Fig 1). Generally, multiple coplanar and noncoplanar radiation fields are used. The dose per day (fraction) is usually less than 2 Gy per field combination. This daily dose is administered over 6 to 7 weeks (the conventional fractionation schedule) (8). The total exposure per fraction is distributed among multiple beams so that normal anatomic structures are exposed to a subtherapeutic dose, while the target volume receives the maximal radiation dose (8).

In both techniques, multiple coplanar and noncoplanar beams oriented in various directions are used to generate a dose distribution that conforms to the target volume. However, in SBRT, a steeper gradient is generated between the periphery of the planned target volume (high-dose area) and normal adjacent structures (low-dose areas) because the tumors treated with SBRT are smaller than those treated with 3D CRT and the beam focus is therefore more narrowly circumscribed (11,16). Thus, SBRT has a strong antitumoral effect, while minimizing radiation injury to normal tissues. In SBRT for lung cancer, a hypofractionated scheme is typically used to deliver a total dose of 50–60 Gy, administered in three to five fractions over a period of 1–2 weeks, with a high ablative per-fraction dose (10–20 Gy per fraction) and a resultant reduction in overall treatment time. However, there is no consensus in the literature about the best dose fractionation schedule for SBRT (12). The small size of tumors treated allows the delivery of a larger dose per fraction so that better local tumor control can be achieved (10,12). In particular, SBRT is highly effective in the treatment of nonoperable stage I non–small cell lung cancer, with reported local control rates of 80% to 100% (19).

Clinical and Pathologic Features of Radiation-induced Lung Injury

Radiation-induced lung injury generally manifests with two distinct well-known clinical and pathologic phases: an early phase of transient radiation pneumonitis and a later phase of chronic radiation fibrosis. The reference point for dating lung changes is the time of completion of radiation therapy (8,14,15,20). Radiation pneumonitis induced by 3D CRT or SBRT usually occurs within the first 6 months after completion of treatment, whereas radiation fibrosis typically occurs at 6–12 months after completion of treatment (11,15–17). Signs and symptoms of radiation pneumonitis may include dyspnea, a cough, a low-grade fever, and chest discomfort, all of which may resolve spontaneously over several weeks, whereas the dyspnea and persistent dry cough in radiation fibrosis are progressive and may be accompanied by signs and symptoms of cor pulmonale. However, fibrosis is more often asymptomatic and may be seen at radiologic imaging without any antecedent radiation pneumonitis (19). Clinically significant radiation pneumonitis develops in 13%–37% of patients who undergo 3D CRT with a radical radiation dose (70 Gy) for treatment of lung cancer (18,21). Steroid treatment that is started in early-stage pneumonitis might prevent the subsequent development of pulmonary fibrosis (22). Symptomatic pulmonary disease from toxic effects of SBRT is less common (4%) (11), presumably because a smaller lung volume is exposed to a high dose of radiation (23).

In radiation-induced lung injury, as in other forms of diffuse alveolar damage, three sequential pathologic phases are distinguished: an acute exudative phase, an organizing or proliferative phase, and a chronic fibrotic phase. The first two phases correspond to radiation pneumonitis, being characterized by more or less organized infiltration of macrophages, and the third phase corresponds to radiation fibrosis with progressive collagen deposition and fibrotic changes (14,20).

Many variables contribute to the toxic effects of radiation on the lungs (14,20,22). These variables include patient-specific factors such as lung performance status, the presence or absence of preexisting lung disease, and the severity of any pulmonary function impairment. In addition, a number of chemotherapeutic agents (eg, actinomycin D, adriamycin, bleomycin, and busulfan) potentiate the effects of radiation. A tumor location near the mediastinum or nerves also may result in more debilitating radiation-induced injuries (21). The use of SBRT to treat centrally located lung tumors should be weighed carefully against the risk of serious complications following the administration of high-dose radiation with a hypofractionated schedule (12).

Treatment-related factors that influence the degree of radiation damage to normal tissue include the total radiation dose, fractionation, and dose rate, as well as the irradiated volume and beam arrangement (14,22). The correlation between the radiation dose and the prevalence of radiation damage is not linear, but the latter generally increases above a certain dose threshold. Pulmonary damage rarely occurs after total radiation doses of less than 20 Gy (subcritical doses), whereas it commonly occurs after doses of 30–40 Gy and almost always occurs after doses of more than 40 Gy (14,24). A more protracted fractionation schedule reduces the biologic effects of radiation on lung tissue (14).

Some studies have investigated a hypothetical dose-volume relationship affecting the likelihood of pulmonary injury after 3D CRT and SBRT (9,23). Since the extent of pulmonary injury has not always correlated with the dose distribution (isodose curve), the minimal dose in the area demonstrating pulmonary change has been taken into consideration. It was reported that the minimal dose to the injured lung is related to the percentage lung volume that receives more than 20 Gy (ie, the V20). The higher the V20, the lower the minimal dose at which damage occurs (9,23). According to Graham et al (9), the V20 should be less than 25% to achieve an incidence of radiation pneumonitis estimated at 0%–4%. Akoi et al (23) reported a minimal lung dose (range, 16–36 Gy; median, 24 Gy) to the area demonstrating pulmonary injury after SBRT. Nevertheless, there is a lack of data regarding the relationship between the probability of severe radiology-induced changes and the dose-volume distribution used in 3D radiation therapy (11).

CT Appearances of Radiation-induced Lung Injury

Manifestations of radiation-induced changes in the lung after conventional radiation therapy have been well described in the literature by Libshitz et al (25–27) and Ikezoe et al (28,29). These authors classified the CT appearances of radiation-induced lung injury without considering the time interval after the completion of radiation therapy. Nowadays, it is largely accepted that radiation-induced lung changes should be classified radiologically as either early or late phase with regard to the time interval after the end of treatment, because this classification better corresponds to the clinical and pathologic aspects of radiation-induced lung abnormalities (8,14,15,20).

Manifestations of radiation-induced lung injury after 3D CRT and SBRT may be similar to those of conventional therapy with regard to the timeline and radiologic findings in early and late phases. Nevertheless, lung abnormalities after SBRT do not usually occur before 2–3 months after completion of therapy, because of the high dose per fraction (11,17). It is well recognized that as the radiation dose per fraction increases, the probability of late phase injury grows (17).

Immediately after treatment, tumor shrinkage might be observed on CT scans without any other findings in the surrounding parenchyma. The early phase of radiation pneumonitis (1–6 months after completion of therapy) manifests radiologically as ground-glass opacities, consolidation, or both, usually in the irradiated lung (Figs 2, 3). Occasionally, at the time of radiation pneumonitis, an ipsilateral pleural effusion associated with atelectasis of the lung may develop (15,24). Although the opacities of radiation pneumonitis may resolve gradually over 6 months without radiologic sequelae when the injury to the lung is limited, in cases of more severe changes, a progression to fibrosis usually occurs (20). The late phase of radiation fibrosis (6–12 months after completion of therapy) appears radiologically as a well-defined area of volume loss with a linear scar or consolidation, parenchymal distortion, and traction bronchiectasis that conforms to the treatment portals (Fig 2). Radiation fibrosis may stabilize or may continue to evolve for as long as 24 months. Shrinkage of the region of fibrotic consolidation or a more sharply defined demarcation between normal and irradiated lung parenchyma may occur (Fig 3). Occasionally, these findings are associated with ipsilateral displacement of the mediastinum and adjacent pleural thickening or effusion.

After 3D CRT, radiation pneumonitis may manifest as focal or nodular ground-glass opacity, consolidation, or both, with the findings usually limited to the area immediately surrounding the treated tumor (8,15). Nevertheless, there is evidence in the literature of CT findings of radiation pneumonitis appearing farther from the tumor site, in both lungs, although still delimited by the radiation portals (Fig 4) (15,30). Abnormalities in lung volumes irradiated at doses under 20 Gy also have been observed (30).

Lung abnormalities after SBRT are not usually seen at sites remote from the target volume (16). Because of the steeper gradient between the periphery of the target volume (high-dose area) and normal adjacent tissue (low-dose area), radiographic changes after SBRT occur within the high-dose region (which encompasses the tumor and a 3D margin of normal tissue) and typically conform to the shape of the tumor (17). CT findings of early radiation pneumonitis after SBRT have been classified, on the basis of the system developed by Ikezoe et al (28,29), into four patterns: (a) diffuse consolidation (Fig 3), (b) diffuse ground-glass opacities, (c) patchy consolidation and ground-glass opacities (Fig 5), and (d) patchy ground-glass opacities (11,17). The findings are defined as diffuse or patchy if lung abnormalities completely fill or do not completely fill the high-dose region, as previously defined.

Radiologic imaging manifestations of radiation fibrosis after both 3D CRT and SBRT have been classified according to one of three patterns described as modified conventional, masslike, or scarlike (8,17). The modified conventional pattern of radiation fibrosis consists of a well-defined consolidation with volume loss and traction bronchiectasis. This pattern is described as “modified conventional” because it closely resembles the pattern of fibrosis seen after conventional radiation therapy. However, conventional radiation fibrosis generally involves the entirety of irradiated lung tissue, from anterior to posterior pleural surface, whereas modified conventional radiation fibrosis is less extensive (Fig 6) (8).

When consolidation with traction bronchiectasis is focal and confined to a 2-cm margin around the original tumor (the region corresponding to the maximal isodose curve delivered), the CT appearance is that of a masslike area larger than the original tumor, and this pattern therefore is described as masslike (Fig 7). In the modified conventional and masslike patterns of radiation fibrosis, consolidations may change shape and location during the 1st year of follow-up because fibrosis causes deformity of the lung, with displacement of the region of change toward or away from the hilum (Fig 7) (16,17).

The scarlike pattern of post-CRT or -SBRT radiation fibrosis consists of a linear opacity less than 1 cm wide that is associated with moderate to severe volume loss and that remains at the tumor site when the primary mass has completely or almost completely resolved (Fig 8). This pattern is markedly different from that of conventional radiation fibrosis.

The advent of multidetector CT, with its improved spatial and temporal resolution, has facilitated the identification and characterization of radiation-induced lung injuries in routine CT studies performed during follow-up of patients treated for lung cancer.

Differential Diagnosis

Careful differential diagnosis, especially with regard to the differentiation of evolving radiation fibrosis from recurrent tumors, is important for determining appropriate therapy.

In general, in patients who have received a radiation dose of more than 40 Gy over a period of at least 4 weeks for 3D CRT or 1–2 weeks for SBRT, the most likely cause of abnormalities within the radiation portals on radiologic images obtained within 6 months after cessation of radiation therapy is radiation pneumonitis (11,15–17).

A review of chest radiographs and CT scans obtained at the initiation of, during, and after therapy may aid in differentiating normal posttherapeutic changes from abnormal ones.

The presence of an infection should be considered a possibility if chest CT scans show pulmonary opacities before the completion of therapy or outside the radiation portal or if diffuse or bilateral lung abnormalities are present (Fig 9) (15,20). Other findings that are suggestive of an infection include centrilobular nodules with a “tree-in-bud” appearance associated with consolidation or cavitation, common features in cases of tuberculosis (31). Sometimes a superimposed infection occurs in an area of radiation-induced lung abnormality, and cavitation may appear (20,26). Since radiation pneumonitis normally follows a more indolent course than infection does, an abrupt onset of lung abnormalities is generally suggestive of infection (unless steroid therapy was recently discontinued, an event that might unmask latent radiation pneumonitis, or chemotherapeutic agents that potentiate radiation effects were used concomitantly), and appropriate diagnostic and therapeutic steps should be initiated (15,32).

Lymphangitic carcinomatosis may clinically mimic radiation-induced lung abnormality by causing dyspnea. However, the greater severity and more rapid progression of symptoms in lymphangitic carcinomatosis, along with the identification of specific CT findings such as smooth or nodular interlobular septal thickening, peribronchovascular interstitial thickening, pleural effusion, and mediastinal lymph nodes, often allow a confident diagnosis (Fig 10). Evidence of diffuse lung abnormalities outside the radiation portals, or bilateral distribution of specific CT findings, increases the level of confidence in a diagnosis of lymphangitic carcinomatosis (33).

Local tumor recurrences usually manifest within 2 years after treatment, depending on the initial tumor size, stage, histologic type, and treatment (34). It may be difficult to identify a recurrent tumor on CT scans obtained during the evolution of radiation fibrosis, particularly when a masslike pattern of fibrosis develops (8,15).

Parenchymal consolidation with a straight lateral margin and air bronchograms are typical CT features of stable radiation fibrosis. Alteration in the contour and dimensions of the fibrotic area, with the appearance of a homogeneous opacity without air bronchograms and with convex borders in the irradiated lung on CT scans, should arouse the suspicion that a local tumor recurrence is present (Fig 5) (35). In addition, filling-in of bronchi within a region of radiation fibrosis may represent a locally recurrent malignancy or superimposed infection (36). However, in patients who have undergone hypofractionated SBRT, Takeda et al (37) reported that these CT signs are not always indicative of tumor recurrence because the region affected by radiation fibrosis may continue to evolve as long as 2 years after treatment.

When other signs of tumor recurrence appear (eg, nodules outside the zone of radiation fibrosis, pleural effusion long after treatment completion, bone destruction, or mediastinal involvement), the diagnosis of local recurrence becomes easier.

Radiation is a well-established carcinogen, particularly with regard to the induction of solid tumors (38). It must be assumed that in addition to genetic predisposition and age-related cancer risk, a successful curative radiation therapy of cancer may, in some cases, cause a new second cancer. The risk of developing a second lung cancer after undergoing irradiation of an initial tumor is difficult to assess because of a lack of large-scale studies including long-term survivors (38). An incidence rate of 2.4 per 100 patient-years is reported for second primary cancers among patients who have undergone combined chemo- and radiation therapy for locally advanced non–small cell lung cancer, with the most frequent sites of second tumors being the lung, esophagus, and stomach (39). The risk increases significantly with time after treatment, with the median time interval from the beginning of treatment to the diagnosis of the second primary tumor being approximately 9.6 years (95% confidence interval, 8.1–11.1 years). Lung cancer may arise within or at the edge of the irradiated area (39). The appearance of homogeneous consolidation or increased opacity within preexisting and stable lung abnormalities in the irradiated areas over a long interval should arouse the suspicion that a radiation-induced tumor may be present (Fig 11).

Chest CT does not always allow differentiation of residual or recurrent malignancy from radiation-induced lung injury, particularly early in the post–radiation therapy period, before the abnormalities have stabilized (Figs 12, 13) (15).

Role of PET/CT

FDG PET allows the differentiation of metabolically active tumor from inactive fibrosis after radiation therapy and may have a role in the evaluation of patients with radiation-induced lung changes (40–44).

Sensitivity of 100% and specificity of 92% were reported with the use of PET/CT to detect residual or recurrent tumor, in comparison with sensitivity of 71% and specificity of 95% with the use of CT alone (42,44). PET/CT may allow detection of a recurrent tumor even before it is manifested clinically (44). However, PET has demonstrated a lower specificity in this setting, one of its limitations being the frequent false-positive uptake of FDG soon after completion of radiation therapy.

Because radiation pneumonitis may lead to FDG uptake that mimics recurrent disease, PET should not be performed until at least 3 months after the completion of radiation therapy, to reduce the likelihood of false-positive results (Fig 14) (15). Persistent uptake associated with radiation pneumonitis occasionally lasts for 15 months after the end of therapy, and a biopsy may be necessary in such cases. In this context, the evaluation of sequential and semiquantitative PET scans and the calculation of standardized uptake values could provide additional information useful for distinguishing between the inflammatory effects of radiation and the recurrence of a tumor. However, the standardized uptake value is affected by a number of variables (eg, plasma glucose levels, tumor size, camera image resolution) (43), and an absolute value that allows differentiation between tumor recurrence and radiation-induced inflammatory changes has not been proved. In our experience, PET/CT is accurate and reliable for this purpose if performed when there is clinical or radiologic evidence suggestive of recurrence (Fig 5). In that context, increased FDG uptake in lesions is more likely to be indicative of residual or recurrent lung cancer, and additional diagnostic and interventional procedures (bronchoscopy, percutaneous needle-aspiration biopsy, open lung biopsy, thoracentesis) then are required. Because of the high negative predictive value of FDG PET, a finding of little or no FDG uptake is considered a definitive indication that no recurrent lung cancer is present and that CT follow-up alone is sufficient.

Summary

The appearance and evolution of lung abnormalities resulting from 3D CRT or SBRT for non–small cell lung cancer differ from the appearance and evolution of lung abnormalities induced by conventional radiation therapy. Since 3D CRT and SBRT are commonly used to treat lung cancer, knowledge of the full spectrum of CT manifestations that may occur after these therapies, especially in relation to the locations of radiation portals and time intervals from the completion of treatment, is useful for avoiding diagnostic errors. Further, PET/CT can help differentiate between radiation-induced lung abnormalities and residual or recurrent lung cancer. Thus, multimodality imaging after radiation therapy for lung cancer can provide important guidance for patient care.