We recently reported bilateral lymphoid pneumonitis in four patients with clinical and radiologic radiation pneumonitis after unilateral radiotherapy for breast carcinoma [4]. Bronchoalveolar lavage showed an intense alveolar lymphocytosis not only in the irradiated lung but also in the contralateral lung. We hypothesized that a lymphocyte-mediated hypersensitivity phenomenon may account for the generalized pathophysiologic changes observed in clinical radiation pneumonitis.

Our prospective study was designed to investigate the bilateral pulmonary effects of strictly unilateral radiotherapy. This group consisted of two of the women prospectively studied who developed clinical radiation pneumonitis and five more patients with clinical radiation pneumonitis who were not part of the prospective patient group and who were studied only once. Three of these patients have been previously reported [4].

Methods

Top Methods Results Discussion Author & Article Info References

The prospective study group consisted of 17 women referred for immediate post-operative radiotherapy for breast cancer. Their median age was 51 years (range, 34 to 78 years). Two patients were current smokers and two were former smokers. Previous breast surgery consisted of either total or partial mastectomy plus axillary dissection. Premenopausal women (n = 7) received adjuvant CMFP chemotherapy (chlorambucil, methotrexate, 5-fluorouracil, prednisolone) and postmenopausal women received tamoxifen, 20 mg daily. Chemotherapy was given for two cycles before radiotherapy and for four more cycles after radiotherapy. Two of these patients and five others who developed clinical pneumonitis after receiving radiotherapy were compared with the 15 patients studied prospectively who did not develop clinical changes after radiotherapy.

Radiotherapy Technique

All patients received a direct anterior field to the supraclavicular fossa angled to the ipsilateral axilla with the medial edge at midline. The chest wall or residual breast tissue was treated with a pair of medial-lateral opposed tangential beams from the midline anteriorly to the posterior axillary line. A separate internal mammary chain field was not used in any patient. The axilla was only treated if no surgical clearance had been done or if extensive extranodal disease existed (11 patients). A posterior boost was used on alternate days. Doses used were 50 Gy in 2 Gy fractions to the supraclavicular fossa, 45 Gy in 1.8 Gy fractions to the breast followed by a direct boost to the primary tumor site of 15 Gy in 8 fractions, or 50 Gy in 2 Gy fractions to the chest wall. The axilla, if treated, received a midpoint dose of 50 Gy over 5 weeks. All treatments were given 5 days per week, treating all fields each time using a Cobalt-60 machine (Theratron 80 AECL, Ottawa, Canada) at a source skin distance of 80 cm. Using this approach, there was no possibility of significant irradiation of the uninvolved lung.

Study Protocol

All patients gave informed consent for the study. A detailed medical history and examination were obtained before patients entered the study. At entry into the study the following studies were performed: physical examination, chest radiograph, gallium lung scan, respiratory function tests, and bronchoalveolar lavage. These tests were repeated 4 to 6 weeks after the completion of radiotherapy, resulting in a 10- to 12-week interval between the two sets of investigations.

Bronchoalveolar lavage was done under local anesthesia with a fiber-optic instrument. The bronchoscope was wedged into three separate subsegmental bronchi in turn, two segments from the right lung and one from the left. One hundred milliliters of room-temperature normal saline was instilled into each segment in four 25-mL aliquots, each being aspirated immediately after instillation. Cells from an unprocessed sample of recovered fluid were deposited on a millipore filter and treated as previously described [5]. The cells were washed, fixed, and stained with hematoxylin and eosin and for non-specific esterase. Differential cell counts were made using the millipore filter method, which produces a more accurate and generally higher estimate of lymphocyte numbers than does the more commonly used cytocentrifuge preparation [6, 7]. Total cell counts were estimated on a hemocytometer.

A gallium lung scan was done 48 hours after the intravenous injection of 5 mCi of gallium-67 citrate. Using published methods and computerized data acquisition, a lung/liver gallium uptake index was calculated for the upper and lower regions of each lung, and a mean value derived [8, 9]. Tests of spirometry, lung volumes, and diffusing capacity were done for each patient (Minato AS-600 spirometer, Osaka, Japan; PK Morgan Transfer Test, Chatham, UK). Results were expressed as percent of predicted mean value using the reference equations suggested by Cotes [10].

An extensive analysis of lymphocyte subsets of bronchoalveolar lavage cells was done on one nonsmoking patient with florid clinically apparent radiation pneumonitis, and 2 controls before irradiation. Bronchoalveolar lavage cells at approximately 1 x 10^8/mL were incubated with fluorochrome-conjugated monoclonal antibody (mAb) for 5 minutes at 25 °C, then treated with Q-Prep reagents (Coulter Electronics, Hialeah, Florida), and fixed using paraformaldehyde. A standard panel of antibodies to lymphocyte subsets and activation markers was used at the manufacturers' recommended concentrations.

Two-color analysis of 2 to 10 x 10³ lymphocytes (gated on forward angle and side scatter; gating checked using CD14-RD1/CD45-FITC mAb) was done using an EPICS PROFILE flow cytometer (Coulter). Only one sample from a control contained enough CD4+ cells to allow analysis of markers within the CD4 subset.

Statistical Methods

Each patient served as her own control. Results before and after treatment, on the same individuals, were compared using the paired Student t-tests, whereas results comparing different groups were analyzed using independent t-tests. Unit sum constraints on the bronchoalveolar lavage cell data (that is, macrophage, lymphocyte and neutrophil proportions must add up to one) were taken into account by not analyzing one component (neutrophil) of the sum.

Results

Top Methods Results Discussion Author & Article Info References

Patients with lavage leukocyte and differential leukocyte counts obtained from the irradiated side of the chest, together with results of lung function and gallium scans, for all 17 participants are given in Table 1. The percentage of lavage lymphocytes was significantly elevated after irradiation (34.5% compared with 46.8%; P = 0.01) with 75% (13 of the 17 patients) developing a lymphocytosis. Although the absolute number of lymphocytes recovered was also increased (6.3 versus 12.6 million; P = 0.12), it was not statistically significant. The lymphocytic changes were accompanied by a reciprocal decrease in the percentage of lavage macrophages (61.4% to 49.4%; P = 0.015) and a small drop in the vital capacity (102.5% compared with 95.5%; P = 0.04). Lavage material, recovered and quantitated separately from both lungs, indicated a similar degree of lymphocytosis in the irradiated and non-irradiated lungs (Table 2). When the analysis of subgroups was done, no observable effect was found for surgical axillary dissection versus additional radiotherapy or of chemotherapy versus tamoxifen on the degree of lymphocytosis, although numbers in these subgroups were small. These seven patients with clinical radiation pneumonitis were older than the 15 who did not develop; the syndrome (67.4 ± 6.2 years versus 51.4 ± 9 years, P = 0.004).

A comparison of studies before radiotherapy with studies after radiotherapy in these two groups is given in Table 3. There was a marked increase in total lavage cell numbers (32.9 versus 83.7 million, respectively; P = 0.005). This appeared to be almost entirely attributable to the absolute increase in the lymphocyte count (9.4 versus 35.2 million, respectively; P = 0.005). The percentage of lymphocytes was also markedly increased (43.5% versus 67.6%, respectively; P = 0.005). Increased gallium uptake within the lung was seen only in the clinical pneumonitis group (3.4 versus 9.0, respectively; P < 0.001). In those with clinical pneumonitis, both lavage and gallium lung scan changes were identical on irradiated and unirradiated sides of the chest (see Table 2). Similarly, there was no significant difference in lavage findings between the irradiated and unirradiated sides of the chest in those without clinical pneumonitis (percentage lymphocytes, 43.5 ± 17.2 versus 39.2 ± 18.6 [P > 0.2]), respectively; and total lavage cell numbers (32.9 ± 40.9 versus 23.6 ± 23.3 [P > 0.2], respectively). After irradiation, vital capacity was also reduced (96.9% versus 76.7%, respectively; P = 0.04), as was diffusing capacity (91.4% to 72.6%, respectively; P = 0.003).

A detailed analysis of lymphocyte subsets was undertaken in a limited number of patients, two controls, and one patient with clinical radiation pneumonitis (Figure 1). Virtually all lymphocytes recovered in the patient sample were identified as T lymphocytes (CD2⁺, CD3⁺). Further, the number of T lymphocytes recovered was dramatically increased compared to samples from two preradiotherapy controls (23.1 versus 1.8 million). No B lymphocytes (CD20⁺) and few NK cells (CD56⁺) were observed. In the patient, almost all T cells (75% versus 30%; 18.0 versus 1.4 million) were of the CD4⁺ helper subset. There was no increase in the percentage CD8⁺ T-suppressor cells in the patient (18% versus 15%), although there was an increase in absolute number (4.4 versus 0.48 million). By contrast with the CD4⁺ cells of the control, which all expressed CD45RO (99% and 2.6 million), in the patient, these CD4⁺ cells nearly all expressed both CD45RA (94% and 16.9 million) and CD45RO (95% and 17.1 million), suggesting that the cells had very recently been activated. Other markers of early T-cell activation were also highly expressed on CD4⁺ from the patient sample compared with the control, including HLA-DR (83% and 14.9 million versus 53% and 1.4 million), CD38 (56% and 10.1 million versus 4% and 1.4 million), and CD69 (75% and 13.5 million versus undetectable). There was very little detectable interleukin-2 receptor (CD25) expression on either patient or control cells. Sometimes, however, expression of this receptor is at quite low levels, and more sensitive flow cytometric techniques may be needed to quantify this marker accurately. Very similar changes in the percentage of CD8⁺ cells expressing these activation markers was also noted (data not shown). However, because of the relatively small number of CD8⁺ cells present, in absolute terms this number was quite small (data not shown).

View larger version (34K): [in this window] [in a new window]   Figure 1. Flow cytometric analysis. Comparison of lavage lymphocytes from one patient with clinical radiation pneumonitis and two controls. These cells have been labeled with control monoclonal antibody, and antibodies to all T cells (CD3), helper T cells (CD4), suppressor/cytotoxic T cells (CD8), NK cells (CD56), two isoforms of tyrosine phosphatase (CD45RA and CD 45RO), and T-cell activation markers HLA DR, CD25 (interleukin-2 receptor), CD38, and CD69.

Discussion

Top Methods Results Discussion Author & Article Info References

Our prospective study provides the first strong evidence for the presence of a severe bilateral lymphoid interstitial process in patients with clinical pneumonitis after strictly unilateral thoracic irradiation. This process was suggested by a marked increase in the absolute number and percentage of lymphocytes recovered in the bronchoalveolar lavage as well as the presence of bilateral increased gallium uptake within the lung and decreased lung volumes. However—more importantly—it also shows a similar but less intense reaction in the lavage of asymptomatic patients who received radiotherapy. These findings suggest that there is a spectrum of lung-tissue reactivity to radiation. This bilateral process was subclinical and self-limiting in most patients but may have progressed to an overt clinical syndrome in a few patients.

Until recently, studies in humans have been limited by the lack of a suitable investigative tool. The advent of bronchoalveolar lavage has helped overcome this problem. The first well-documented bilateral study of pneumonitis using this method was published by our group [4] and provided strong evidence that radiation pneumonitis was not confined to the field of irradiation. We have previously reported four patients with severe bilateral radiation pneumonitis after strictly unilateral chest irradiation. They had a severe lymphoid interstitial pneumonitis, with a marked increase in the total cell count and percentage of lymphocytes to between 60% and 75% of the recovered cells. The lavage abnormality was present in both the irradiated and unirradiated sides of the chest. Two other reports have shown unilateral bronchoalveolar lavage findings in the irradiated segments of lung in patients with clinical pneumonitis [11, 12]. Both have shown the presence of a lymphocytic alveolitis. A further case report of lavage findings from a patient with radiologic unilateral pneumonitis after treatment for Hodgkin disease also showed equal lymphocytosis in both irradiated and unirradiated lung segments [13]. Other brief reports have suggested the more extensive nature of radiation pneumonitis. Case reports of bilateral radiologic changes in acute-phase pneumonitis after unilateral radiotherapy have been recorded [14, 15]. More recently, gallium lung scan studies have confirmed bilateral increased uptake in patients with acute pneumonitis after unilateral irradiation [16]. A further study in patients irradiated for unilateral lung cancer has shown increased bilateral pulmonary uptake in asymptomatic individuals [17]. Histologic studies have been limited by the invasive methods required to obtain samples, but a series of open-lung biopsies taken from within the field of irradiation in patients with clinical pneumonitis has suggested a lymphocytic infiltration as the dominant finding in acute-phase pneumonitis [18].

Current opinion suggests that two inter-related mechanisms are important in radiation-induced pulmonary damage. Microcapillary vascular damage occurs after radiation exposure, producing ischemic tissue injury that eventually results in fibrotic healing. In addition, damage to type I and II pneumocytes in the presence of increased vascular permeability results in altered surfactant production, atelectasis, reduced ventilation, and secondary vascular atrophy [19, 20]. These mechanisms may explain direct tissue damage from radiotherapy within the radiation treatment portal but do not adequately explain the degree of respiratory distress after irradiation of a relatively small volume of lung tissue, the sporadic nature of clinical pneumonitis, and the resolution of symptoms with time without significant sequelae in most patients. These mechanisms also cannot explain the pathologic findings on gallium lung scan and lavage that occur well outside the field of irradiation.

The degree of lymphocytosis observed in this study is found in few respiratory conditions and implies an immunologic process. The main differential diagnoses to consider are sarcoidosis and tuberculosis. No other clinical or radiologic features suggested these two diagnoses, and bronchoalveolar lavage fluid was sent for mycobacterial culture to further exclude the diagnosis of tuberculosis. Similar findings of lymphocytosis in bronchoalveolar lavage specimens have recently been reported in methotrexate-induced pneumonitis, which is now recognized as a hypersensitivity phenomenon [21, 22]. The spectrum of abnormalities found in radiation pneumonitis is also not without precedent in immunologic respiratory disorders. Similar bronchoalveolar lavage findings have been reported in asymptomatic patients heavily exposed to bird proteins or molds, only a small number of whom actually develop a clinically evident lymphocytic alveolitis [23, 24].

The mechanism for the development of the lymphocytic alveolitis that we reported is not clear. The way in which treatment was administered made it impossible for this lung to receive significant irradiation. Analysis of lymphocyte subsets in a patient with clinical pneumonitis and in two controls (see Figure 1) demonstrated dramatically increased numbers [percent and absolute] of recently activated T lymphocytes, especially CD4⁺ helper cells, in the patient sample, compared with that usually found in the circulation and in other lavage samples. It is possible that direct tissue damage from radiotherapy, known to cause antigen release [25], overcomes the normal tolerance processes and produces sensitization of autoreactive lymphocyte clones, which then migrate to the lung and react with pulmonary tissues. A lymphocytopenia is known to occur after irradiation due to destruction of these cells in adjacent lymphoid structures or migration through the lung at the time of irradiation. This damage to lymphocytes may alter the important balance between lymphocyte subsets, which may aid this process. Recovery in lymphoid populations after radiotherapy may explain the spontaneous regression of symptoms and signs even in untreated patients who survive the acute disease. The progression to clinical pneumonitis seen in some patients may result from the interaction with genetic or environmental factors, such as the degree of lymphoid irradiation, age, and concurrent therapy.

If hypersensitivity is responsible for radiation pneumonitis, this may lead to strategies for minimizing complications and may also suggest ways that this important therapy can be used in higher doses and in combination with other agents that now produce unacceptable pulmonary toxicity.