Stephans, Kevin L. MD*; Djemil, Toufik PhD*; Reddy, Chandana A. MS*; Gajdos, Stephen M. MS*; Kolar, Mathew MS*; Machuzak, Michael MD†; Mazzone, Peter MD†; Videtic, Gregory M. M. MD*
Surgical resection is the standard of care for stage I non-small cell lung cancer (NSCLC).1,2 A significant number (30–60%) of stage I patients, however, are medically inoperable secondary to comorbidities3,4 the most common of which is chronic obstructive pulmonary disease.5 Conventional radiotherapy of 60 to 70 Gy offered to inoperable patients results in high local failure rates and 5-year overall survivals of 15 to 30%.6 The literature, however, suggests that there exists a radiotherapy dose-response effect in NSCLC.6,7
Stereotactic body radiotherapy (SBRT) offers a novel means of safely delivering very high-radiation doses to stage I NSCLC. Onishi et al.8 reported a 5-year local control and overall survival rate of 84% and 54%, respectively, following SBRT to a median biologic equivalent dose of 100 Gy. Several series have demonstrated similar results for SBRT regimens ranging from 48 to 60 Gy in 3 to 10 fractions.4,9–18
The potential benefits of SBRT for cancer control in a vulnerable population should not be mitigated by toxicity, especially with reference to pulmonary function. Although most studies describe SBRT as well tolerated in these patients, there are reports of exacerbations of dyspnea,5 radiation pneumonitis,19 or airway obstruction14 after treatment. In the preliminary toxicity analysis of RTOG 0236, a recently completed phase II study, one episode of grade 4 (2%) and 7 episodes of grade 3 (13%) pulmonary/upper respiratory events were reported, many of which were decreases in pulmonary function test (PFT) values.20 Some studies on PFT changes describe minimal or transient alterations with SBRT.15,21–23 These contrasting reports prompted us to study further the effects of SBRT treatment and dosimetry on objective measures of pulmonary function in medically inoperable patients.
Patient and Tumor Characteristics
All data was collected within an institutional-review board approved registry. Between February 2004 and August 2007, 92 patients with clinical AJCC T1A or T1B lung cancer were treated with SBRT for 102 lesions (5 synchronous and 5 metachronous primaries). All patients were deemed medically inoperable by multidisciplinary evaluation. Evaluation included history and physical examination, contrast-enhanced computed tomography (CT) of the chest, [18F]fluorodeoxyglucose positron emission tomography, and PFTs including forced expiratory volume at 1 second (FEV1) and diffusion capacity to carbon monoxide (DLCO). Selection for SBRT was not limited by pulmonary status or tumor location. Charlson comorbidity index scores were retrospectively calculated from initial patient history24 with no points assigned for the lung cancer under treatment. Initial follow-up was 6 to 8 weeks after SBRT with same-day PFTs and chest CT. Thereafter, routine follow-up was every 3 months with CT imaging at each visit and PFTs twice yearly. Toxicity was assessed according to the Common Terminology Criteria for Adverse Events version 3.0.
Patients were simulated supine in a vacuum bag restriction system (BodyfixR, Elekta Inc, Stockholm, Sweden). An abdominal compression device was applied to reduce respiratory movement and adjusted under fluoroscopy. A 3-mm slice thickness planning axial CT scan was taken during quiet breathing, full inspiration and full expiration. Treatment plans were generated by BrainScan 5.31 (BrainLAB, Feldkirchen, Germany) planning software referenced to the free-breathing study. Patients were treated on a Novalis (BrainLAB, Feldkirchen, Germany) machine using orthogonal films and the ExacTrac (BrainLAB, Feldkirchen, Germany) stereotactic body system for positioning.
Gross tumor volume was delineated on each respiratory study using “lung window” setting. Critical structures including lungs, spinal cord, heart, brachial plexus, and esophagus were outlined and limited according to accepted standards. Lesions before March 2006 (n = 46) were preferentially treated to 50 Gy in 5 fractions, according to the principles of Uematsu et al.16 In March 2006, we adopted ROTG 0236 planning criteria25 and a prescription of 60 Gy in 3 fractions. Patients with central tumors, defined as within 2 cm of the tracheobronchial tree per RTOG 0236, were treated to 50 Gy in 5 fractions. Six patients with large planning target volumes directly adjacent to critical structures were treated to 50 Gy in 10 fractions. Measures of dose uniformity included conformality index defined as the maximum dose at any point divided by the prescription dose and heterogeneity index defined as the prescription isodose volume divided by the tumor volume.
The primary end point was any change in pulmonary function measured from the time of treatment to the latest available measurement at least 6 months post SBRT. PFT changes were assessed with a paired sign test. Kendall's rank correlations, Mann-Whitney tests, and Kruskal-Wallis tests were used to correlate patient and treatment related variables to individual PFT changes. Mann-Whitney tests were used to compare time to cardiac compared with other causes of death, FEV1% by cause of death (cardiac versus all others), as well as correlation of Charlson scores by the FEV1% median value. Fisher exact test was used to compare cause of death to FEV1% above or below median. Overall survival by FEV1% and DLCO cutoffs was analyzed using Kaplan-Meier curves, and the log-rank test was used to determine whether a statistically significant difference was present among patient groups. Univariate and multivariate analysis for overall survival were performed using Cox Proportional Hazards regression. Statistical analyses were performed using StatView 5.0 (SAS Institute, Cary, NC) and p value <0.05 was considered statistically significant.
Characteristics of the 92 patients are shown in Table 1. The most common reason for medical inoperability was predicted postresection FEV1 less than 0.8 liter. Reasons for inoperability and prevalence of select comorbidities are listed in Table 2. Median follow-up was 18.4 months (range, 1.7–48 months) and median time to most recent PFTs was 10.4 months. Local control was 98% and no patient died of isolated local failure. Median overall survival was 18 months (range, 1.9–48 months). Median pretreatment FEV1 was 1.21 liter (range, 0.41–2.72 liter) and median % predicted FEV1 (FEV1%) was 50% (range, 15–138%). Median DLCO was 56.5% of predicted (range, 14–143%) with higher values seen in patients with cardiac and vascular comorbidities. Dosimetric description of SBRT treatments is provided in Table 3. A total of 68 patients had both baseline and ≥6 month post treatment FEV1 data while 41 patients had both baseline and ≥6 month posttreatment DLCO data. There was no significant change in any measure of pulmonary function following SBRT. The mean change in FEV1 was −0.05 liter (range, −0.98 to +1.29 liter, SD 0.37 liter; p = 0.22) representing −1.88% predicted baseline FEV1 (range, −33 to + 43%, SD 12%; p = 0.62). DLCO declined 2.59% of predicted (range, −37 to +33%, SD 15%; p = 0.27).
Although the mean change in pulmonary function was not significant for the group as a whole, 10% of patients experienced an absolute FEV1% decline of at least 14.8% predicted while another 10% experienced an increase of at least 12.7%. Additionally, 10% of patients experienced an absolute DLCO decline of at least 19.4% of predicted while 10% experienced an increase of at least 18%. The distribution of changes in FEV1% and DLCO are shown in Figures 1, 2. There were no differences in patient or treatment characteristics between the 20% of patients with largest increase in FEV1% compared with 20% with largest decline in FEV1%. Two patients experienced grade 2 radiation pneumonitis, both resolved with oral steroids. There were no cases of grade 3 (requiring or increased supplemental oxygen) pneumonitis.
Given individual PFT variations we examined the effects of radiation dose and dose distribution upon pulmonary function, as in Table 4. Increasing conformality index, V5 and V10 were correlated with greater declines in FEV1% (p = 0.033, p = 0.0036, p = 0.025, respectively), however, had low Tau values (Tau = 0.180, Tau = 0.244, Tau = 0.188) suggesting that the magnitude of correlation was small. No other variables were correlated with FEV1% decline notably including total dose, age, involved lobe, or central versus peripheral location. Specifically regarding tumor location, the median change in FEV1 for all central lesions was +0.06 liter or +3% predicted FEV1% (range, −18 to + 16%) while the median change in FEV1 for peripheral lesions was −0.08 liter or −3% predicted FEV1% (range, −33 to +43%). This difference was not statistically significant for central versus peripheral location for FEV1 (p = 0.55) or FEV1% (p = 0.37). Comparison of DLCO changes was not done as only two patients with central lesions had pre and post SBRT DLCO. No patient or treatment related variable was predictive of DLCO changes (Table 5), though age (p = 0.085) and continued smoking (p = 0.095) approached significance.
To examine the tolerability of SBRT in patients with varying degrees of lung function we divided patients into quartiles based on baseline PFTs. Patients with the lowest FEV1% values had better overall survival regardless of whether patients were divided by the lowest quartile, below the median (Figure 3), or below the 75th percentile (p = 0.041, p = 0.0046, and p = 0.020, respectively). When patients were divided based on baseline DLCO there was no survival difference for the lowest quartile, below the median (Figure 4) or below the 75th percentile (p = 0.55, p = 0.44, p = 0.73, respectively).
To investigate potential factors influencing the survival of patients with higher baseline FEV1%, we calculated the mean survival by cause of death (Table 6). Survival was significantly shorter for patients dying of cardiac disease compared with other causes (median 9.2 versus 19.0 months, p = 0.028). All cardiac deaths occurred in patients with FEV1% at or above the median. However, patients with FEV1% above the median also died twice as often from noncardiac causes, and the median FEV1% of patients dying of cardiac death was not significantly higher than patients dying of other causes (65% versus 58%, respectively, p = 0.28).
Charlson comorbidity index correlated with overall survival on univariate analysis (p = 0.0032, hazard ratio [HR] = 1.31). Patients with FEV1% above the median, however, did not have statistically higher Charlson scores and the median for both groups was 3, p = 0.58. A modified Charlson score including additional points for modern cardiac risk factors of known coronary artery disease, hypertension, hyperlipidemia, and body mass index ≥30 was also tested, but was not found to be a stronger predictor of survival (p = 0.0065, HR = 1.27) than the standard Charlson score. Patients with FEV1% above median did not have statistically higher modified Charlson scores than those below median (p = 0.23).
In addition, we analyzed the effects of age, stage (IA versus IB), KPS, gender, active smoking, FEV1% (both as a continuous variable and ≥ median), and DLCO on overall survival. Stage IB (p = 0.0061, HR = 2.83) and FEV1% (p = 0.0059, HR = 1.023 per % predicted and p = 0.0063, HR = 3.40 for FEV1% ≥ median) were correlated with overall survival.
Two separate multivariate analyses were conducted given that absolute FEV1% and FEV1% ≥ median are dependent upon each other. In the first, both FEV1% as a continuous variable (p = 0.012, HR = 1.021 per %) and stage (p = 0.031, HR = 2.51) significantly predicted for survival whereas Charlson score approached significance (p = 0.066, HR= 1.22). In the second, FEV1% ≥ median (p = 0.0065, HR = 3.44) and stage (p = 0.044, HR = 2.48) significantly predicted for survival whereas Charlson score did not (p = 0.13, HR = 1.18).
This study on the effects of SBRT on pulmonary function in medically compromised patients found (1) no significant long-term change in FEV1, FEV1%, or DLCO after treatment and (2) no grade 3 or higher clinical toxicities. This is supported by other series: An Indiana University dose escalation study in SBRT demonstrated an initially minimal decline in PFTs followed by a return to baseline,15 while patients treated on their phase II protocol demonstrated no change in FEV1 with a small decline in DLCO of 1.11 mg/min/mm Hg/y.21 Ohashi described no change in FEV1 and an increase in DLCO in a small population of Japanese patients.22 Two abstracts with a total of 48 patients reported larger magnitude median declines in FEV1% and DLCO, however, were not noted to be statistically significant.26,27 The Indiana series and ours are the largest reported with details on PFT changes at this point and suggest minimal changes which may be comparable to those expected by aging. In addition we performed detailed analysis looking at the relationship between morbidity and mortality patterns and looked for associations with the PFT changes.
To investigate individual variations in PFTs we examined their correlation with patient, tumor, and dosimetric parameters. Despite large variations in biologic equivalent dose of the 3 treatment regimens in our population there was no effect of dose on FEV1% (p = 0.95) or DLCO (p = 0.30) suggesting that altering dose for patients with poor pulmonary function is not necessary. In addition, there was no difference between upper and lower lobe tumors or central versus peripheral location. While frequent grade 3 pulmonary complications have been reported for central tumors treated to 60 to 66 Gy,14 we treated central lesions to 50 Gy and found no differences in PFT changes compared with peripheral lesions. Japanese investigators have also safely treated central lesions to doses <60 Gy.8,16 We did find higher conformality index, V5 and V10 were associated with larger declines in FEV1%. Correlation scores however were quite low (largest was Tau of 0.244) suggesting this effect is not dramatic. This is of interest given the correlation of exceptionally high rates of radiation pneumonitis (29% grade 2, 12% grade 5) in a small series of Japanese patients treated with a high conformality index.19 Conformality Index is defined as prescription isodose volume divided by planning target volume (essentially tumor volume plus set-up error) and represents the amount of normal tissue treated to encompass the target tissue in a given treatment plan. Care should be taken to minimize the volume of normal lung treated both to high and low doses. Our median conformality index of 1.48 is relatively low and may have been an important factor in minimizing the risk of pneumonitis.19 Likewise, the RTOG 0236 protocol suggests conformality index should be kept below 1.2 except in small tumors (<2.5 cm axial or <1.5 cm craniocaudal tumor volume dimension)25 where this can not be accomplished because of minimum field size restrictions. Most series describe a clinical pneumonitis risk of less than 5%.4,8–10,15,16,20,22,23,28
Significant individual variations in PFT parameters were observed, however, there was no significant change for the population as a whole. This is consistent with a previous study in which worsening dyspnea scores were noted at some point after SBRT in 40% of patients.5 The authors, however, noted a high prevalence of dyspnea at baseline (64%) and large interindividual variability in the onset and duration of dyspnea aggravation with no relation to timing of SBRT. In our study, we found no correlation of PFT changes to radiation dose, tumor volume, or tumor location. Likewise, there was no difference in patient or treatment characteristics between the patients with the largest increases and those with the largest decreases in FEV1%. Large individual patient fluctuations in PFT's (which we observed in both the positive and negative direction at nearly equal magnitude) may thus be more related to fluctuations in the patient's underlying comorbid medical illness than to treatment with SBRT.
To examine the tolerability of SBRT in patients with varying degrees of baseline pulmonary function we divided patients into quartiles by FEV1% and DLCO. A similar analysis was recently reported by Indiana demonstrating the seemingly counter-intuitive result that patients with lower FEV1 values actually had better overall survival.21 They postulated this may be due to patients in the higher spectrum of FEV1 being more likely to be inoperable because of cardiac comorbidity.
We chose to divide patients by FEV1% rather than FEV1 to minimize potential confounding effects of patient's height, age, and gender. Patient's with lower FEV1% had statistically superior survival and FEV1% above median was as strong a predictor of overall survival (HR = 3.12) as T-stage (HR = 2.83). That no patient was refused treatment based on pulmonary function and the lowest quartile of FEV1% (from 15 to 36% predicted) had better survival than the group as a whole while there was no significant overall decline in PFTs suggests that SBRT appears to be safe even in those patients with the worst pulmonary function. Anecdotally, the patient with the worst baseline PFTs (FEV1 0.61 liter, FEV1% 15, DCLO 16%) was alive with no evidence of disease at last follow-up, 29 months after treatment.
Cardiac comorbidities seem to be at least in part responsible for the shorter survival of patients with FEV1% above median. All patients dying of cardiac disease had FEV1% at or above median and died statistically sooner than patients dying of other causes. In addition, there was no difference in patients' survival when divided into quartiles by DLCO in the Indiana study21 or ours. DLCO has been noted to decrease in patients with longstanding congestive heart failure29 and may already account for some comorbidities. Notably, in addition to having more cardiac deaths, patients with baseline FEV1% above median also had twice as many noncardiac deaths suggesting additional factors may contribute to survival differences. To investigate other medical comorbidities we used the Charlson index, a validated measure of assessing a patient's risk of dying in a given time period based on their medical comorbidities.24 Charlson index did correlate with survival in our population, however, was not higher in patients with FEV1% above the median, even when using an alternative modified Charlson index incorporating more modernly identified cardiac risk factors. Although cardiac disease is likely an important determinant of outcome it may not be the only critical factor in a population with significant selection bias such as medically inoperable patients. We plan to readdress this question with a competing risk analysis in the future when more patients and follow-up are available.
Our findings support the safety of SBRT in early stage NSCLC, even for those patients with extreme pulmonary comorbidities. No patient was denied treatment based on pulmonary function in our population and there was no significant decline in PFTs noted with treatment. Central lesions were safely treated with 50 Gy in 5 fractions. Patients even in the worst quartile of FEV1% and DLCO had survival equivalent or better than the group as a whole.
1. Smythe WR. Treatment of stage I non-small cell lung carcinoma. Chest
2. Spira A, Ettinger DS. Multidisciplinary management of lung cancer. N Engl J Med
3. McGarry RC, Song G, des Rosiers P, Timmerman R. Observation-only management of early stage, medically inoperable lung cancer: poor outcome. Chest
4. Nyman J, Johansson KA, Hulten U. Stereotactic hypofractionated radiotherapy for stage I non-small cell lung cancer–mature results for medically inoperable patients. Lung Cancer
5. Paludan M, Traberg Hansen A, Petersen J, Grau C, Høyer M. Aggravation of dyspnea in stage I non-small cell lung cancer patients following stereotactic body radiotherapy: is there a dose-volume dependency? Acta Oncol
6. Sibley GS. Radiotherapy for patients with medically inoperable stage I nonsmall cell lung carcinoma: smaller volumes and higher doses–a review. Cancer
7. Rosenzweig KE, Fox JL, Yorke E, et al. Results of a phase I dose-escalation study using three-dimensional conformal radiotherapy in the treatment of inoperable nonsmall cell lung carcinoma. Cancer
8. Onishi H, Shirato H, Nagata Y, et al. Hypofractionated stereotactic radiotherapy (HypoFXSRT) for stage I non-small cell lung cancer: updated results of 257 patients in a Japanese multi-institutional study. J Thorac Oncol
9. Baumann P, Nyman J, Lax I, et al. Factors important for efficacy of stereotactic body radiotherapy of medically inoperable stage I lung cancer. A retrospective analysis of patients treated in the Nordic countries. Acta Oncol
10. McGarry RC, Papiez L, Williams M, Whitford T, Timmerman RD. Stereotactic body radiation therapy of early-stage non-small-cell lung carcinoma: phase I study. Int J Radiat Oncol Biol Phys
11. Nagata Y, Takayama K, Matsuo Y, et al. Clinical outcomes of a phase I/II study of 48 Gy of stereotactic body radiotherapy in 4 fractions for primary lung cancer using a stereotactic body frame. Int J Radiat Oncol Biol Phys
12. Ng AW, Tung SY, Wong VY. Hypofractionated stereotactic radiotherapy for medically inoperable stage I non-small cell lung cancer–report on clinical outcome and dose to critical organs. Radiother Oncol
13. Onimaru R, Fujino M, Yamazaki K, et al. Steep dose-response relationship for stage I non-small-cell lung cancer using hypofractionated high-dose irradiation by real-time tumor-tracking radiotherapy. Int J Radiat Oncol Biol Phys
14. Timmerman R, McGarry R, Yiannoutsos C, et al. Excessive toxicity when treating central tumors in a phase II study of stereotactic body radiation therapy for medically inoperable early-stage lung cancer. J Clin Oncol
15. Timmerman R, Papiez L, McGarry R, et al. Extracranial stereotactic radioablation: results of a phase I study in medically inoperable stage I non-small cell lung cancer. Chest
16. Uematsu M, Shioda A, Suda A, et al. Computed tomography-guided frameless stereotactic radiotherapy for stage I non-small cell lung cancer: a 5-year experience. Int J Radiat Oncol Biol Phys
17. Xia T, Li H, Sun Q, et al. Promising clinical outcome of stereotactic body radiation therapy for patients with inoperable Stage I/II non-small-cell lung cancer. Int J Radiat Oncol Biol Phys
18. Zimmermann FB, Geinitz H, Schill S, et al. Stereotactic hypofractionated radiotherapy in stage I (T1–2 N0 M0) non-small-cell lung cancer (NSCLC). Acta Oncol
19. Yamashita H, Nakagawa K, Nakamura N, et al. Exceptionally high incidence of symptomatic grade 2–5 radiation pneumonitis after stereotactic radiation therapy for lung tumors. Radiat Oncol
20. Timmerman R, Paulas R, Galvin J, et al. Toxicity analysis of RTOG 0236 using stereotactic body radiation therapy to treat medically inoperable early stage lung cancer patients. Int J Radiat Oncol Biol Phys
21. Henderson M, McGarry R, Yiannoutsos C, et al. Baseline pulmonary function as a predictor for survival and decline in pulmonary function over time in patients undergoing stereotactic body radiotherapy for the treatment of stage I non-small-cell lung cancer. Int J Radiat Oncol Biol Phys
22. Ohashi T, Takeda A, Shigematsu N, et al. Differences in pulmonary function before vs. 1 year after hypofractionated stereotactic radiotherapy for small peripheral lung tumors. Int J Radiat Oncol Biol Phys
23. Onimaru R, Shirato H, Shimizu S, et al. Tolerance of organs at risk in small-volume, hypofractionated, image-guided radiotherapy for primary and metastatic lung cancers. Int J Radiat Oncol Biol Phys
24. Charlson ME, Pompei P, Ales KL, MacKenzie CR. A new method of classifying prognostic comorbidity in longitudinal studies: development and validation. J Chronic Dis
26. Bucci M, Chang J, McAleer M, et al. Pulmonary function tests before and after stereotactic radiation of the lung. Int J Radiat Oncol Biol Phys
27. Pratt D, El Naqa J, Bradley J. Pulmonary function testing after stereotactic body radiation therapy for medically inoperable patients with early stage NSCLC. Int J Radiat Oncol Biol Phys
28. Bradley J. Radiographic response and clinical toxicity following SBRT for stage I lung cancer. J Thorac Oncol
29. Puri S, Baker BL, Dutka DP, Oakley CM, Hughes JM, Cleland JG. Reduced alveolar-capillary membrane diffusing capacity in chronic heart failure. Its pathophysiological relevance and relationship to exercise performance. Circulation
30. McKee PA, Castelli WP, McNamara PM, et al. The natural history of congestive heart failure: the Framingham study. N Engl J Med