Stereotactic ablative radiotherapy: A potentially curable approach to early stage multiple primary lung cancer
Joe Y. Chang MD, PhD,
Department of Radiation Oncology, The University of Texas MD Anderson Cancer Center, Houston, Texas
Corresponding author: Joe Y. Chang, MD, PhD, Department of Radiation Oncology, Unit 97, The University of Texas MD Anderson Cancer Center, 1515 Holcombe Boulevard, Houston, TX 77030; Fax: (713) 563-2331; firstname.lastname@example.org
We thank all members of the Thoracic Radiation Oncology section in the Division of Radiation Oncology for their help and Ms. Christine Wogan for editorial assistance.
Surgical resection has been the standard treatment for early stage multiple primary lung cancer (MPLC). However, a significant proportion of patients with MPLC cannot undergo surgery. For this report, the authors explored the role of stereotactic ablative radiotherapy (SABR) for patients with MPLC.
Patients with MPLC who received SABR (50 grays [Gy] in 4 fractions or 70 Gy in 10 fractions) for the second tumor were reviewed. Four-dimensional, computed tomography-based, planning/volumetric image-guided treatment was used for all patients. Treatment outcomes/toxicities were analyzed.
For the 101 patients who received SABR, at a median follow-up of 36 months and with a median overall survival (OS) of 46 months, the 2-year and 4-year in-field local control rates were 97.4% and 95.7%, respectively. The 2-year and 4-year OS rates were 73.2% and 47.5%, respectively; and the progression-free survival (PFS) rates were 67% and 58%, respectively. Patients who had metachronous tumors had better OS and PFS than patients who had synchronous tumors (2-year OS: 80.6% metachronous vs 61.5% synchronous; 4-year OS: 52.7% vs 39.7%, respectively; P = .047; 2-year PFS: 84.7% vs 49.4%, respectively; 4-year PFS: 75.6% vs 30.4%, respectively; P = .0001). For patients who either underwent surgery or received SABR for an index tumor, the incidence of grade ≥3 radiation pneumonitis was 3% (2 of 71 patients); however, this increased to 17% (5 of 30 patients) for those who received conventional radiotherapy for an index tumor. Other grade ≥3 toxicities included grade 3 chest wall pain (3 of 101 patients; 3%) and grade 3 skin toxicity (1 of 101 patients; 1%).
The incidence of multiple primary lung cancer (MPLC) has been steadily increasing, presumably because of improving patient survival from effective treatment of the first primary lung tumor coupled with earlier detection by multislice, spiral computerized tomography (CT) and positron emission tomography (PET). Currently no guidelines are available on recommended treatments for MPLC. For early stage tumors, surgical resection has produced encouraging results, with the potential for cure[2-5]; however, a significant proportion of patients with MPLC are ineligible for surgery because of limited cardiopulmonary reserves or other medical conditions. Stereotactic ablative radiotherapy (SABR), also called stereotactic body radiotherapy (SBRT), has emerged as a novel radiation modality with significant applications for early stage lung cancer. The feasibility, safety, and efficacy of SABR for this purpose have been established by retrospective and prospective studies published over the past decade,[7-10] and this technique has become a new standard of care for patients with medically inoperable stage I nonsmall cell lung cancer (NSCLC). However, the application of SABR for multiple early stage primary lung tumors has been described only rarely.[11-13] The objective of this study was to explore the role of SABR for patients with MPLC.
MATERIALS AND METHODS
Diagnostic Criteria and Patient Selection
We used criteria modified from Martini and Melamed (Table 1) to define MPLC, which included both synchronous MPLC and metachronous MPLC. All patients with NSCLC who received treatment with SABR at our institution between October 2004 and December 2010 had been registered prospectively and were retrospectively reviewed. All eligible patients based on the definition of MPLC were identified and analyzed. The diagnosis of MPLC was originally made when 2 dominate lung nodules had been identified per patient. For patients with synchronous MPLC, the tumor that was most advanced based on clinical TNM stage was considered the “index” tumor; for patients with metachronous MPLC, the first tumor detected was considered the index tumor, and the others were considered second tumors. The histology of both the index and second tumors had to be confirmed pathologically; all second tumors had to have been treated with SABR regardless of how the index tumor had been treated, and the biologic effective dose (BED) of SABR had to be >100 grays (Gy) using a linear-quadratic model with an α:β ratio of 10. This study was approved by the Institutional Review Board.
Table 1. Criteria for Diagnosis of Second Primary Lung Tumora
No. of Patients
Table reproduced, with modifications, from Martini N, Melamed M. Multiple primary lung cancers. J Thorac Cardiovasc Surg. 1975;70:606–612.
No patient had extrapulmonary or common lymphatic carcinoma at the time of diagnosis.
Synchronous tumors (diagnosed within 6 mo)
Same histology; second tumor in different lobe or lungb
Techniques for patient immobilization and SABR treatment planning have been described in detail in our previous reports.[9, 15-17] Briefly, 4-dimensional (4D) CT images were obtained to take tumor motion into consideration. Dose was prescribed according to the planning target volume (PTV), and >95% coverage was intended if normal tissue constraints were met. Patients received 6 to 12 coplanar or noncoplanar 6-megavolt photon beams using Pinnacle calculation algorithms with heterogeneity correction.
For patients who had previously received radiotherapy to the chest, markings on prior tattoos were used to identify the areas of previous radiotherapy, and fused simulation images were used to generate a composite plan. If possible, SABR beam angles should avoid previously irradiated areas (except for ipsilateral peripheral lung), particularly in the esophagus, bronchial tree, brachial plexus, major vessels, heart, spinal cord, and chest wall. The dose-volume constraints used for critical structures were consistent with our previously published guidelines.[9, 15-17] No violation of constraints for the spinal cord, esophagus, or brachial plexus were allowed; constraints for other normal tissues were judged on the basis of PTV coverage. Typically, when the tumor was close to a critical structure, a compromise in PTV coverage was considered acceptable, but internal gross target volume coverage still had to be maintained. Patients who had lesions that were very close to or abutting critical structures and whose normal tissue dose-volume constraints for 50 Gy in 4 fractions could not be met, even with compromise of PTV coverage, received 70 Gy in 10 fractions. For patients who had received radiotherapy (conventional radiotherapy or SABR) for the previous lung tumor, composite plans were generated,[16, 17] and adjustments of SABR planning were made to limit the radiation dose to critical structures on an individual basis. When both the index tumor and the second tumor were treated with SABR, the composite plan still had to meet with dose-volume constraints for single SABR. Because there is no practical approach for converting and calculating the cumulative BED with different radiotherapy fractionations, BEDs with different fractionations were not taken into consideration when those who had received prior radiotherapy were treated with conventional fractionation and the composite plans were generated. However, both the SABR plan and the composite plan were reviewed and compared separately. In general, the SABR plan needs to meet with the dose-volume constraints of SABR.[9, 15] Composite plans should meet with the dose-volume constraints of conventional fractionation, as we published previously. When these guidelines can not be met, the case should be presented to a quality-assurance meeting for discussion before the treatment plan is finalized. When evaluating the potential risk of radiation pneumonitis for combined conventional fractionation and SABR, a prediction model was developed and applied. Patients were positioned each day by using either CT-on-rails or cone-beam CT systems. Treatments were delivered on contiguous days but not on Saturdays or Sundays.
Patients underwent chest CT scanning every 3 months for 2 years after SABR and then every 6 months for another 3 years. PET scans were recommended at 3 to 12 months after SABR. Toxic effects were scored according to version 4 of the National Cancer Institute Common Terminology Criteria for Adverse Effects. Tumor control was evaluated based on findings from both PET and CT images. Local recurrence, defined as progressive CT soft-tissue abnormalities over time that corresponded to avid areas on PET/CT images >6 months after SABR with maximum standard uptake value >5, as we described previously, or a positive biopsy anywhere in the involved lobe was scored in 1 of 2 ways, as either an “in-field failure” (occurring in the area inside the PTV) or an “involved lobe failure” (within the involved lobe but outside the PTV).
Continuous variables were summarized using descriptive statistics. Categorical variables were tabulated by frequency and percentage. The follow-up and survival times were computed from the initiation of local treatment (surgery or radiotherapy) for any of the tumors in the synchronous MPLC group or the second tumor in the metachronous MPLC group. Survival functions were calculated using Kaplan-Meier estimates, and the log-rank test was used to assess the equality of the curves. Local control rates for the patients who received SABR were calculated from the starting date of SABR for individual tumors, also using Kaplan-Meier method. Probability (P) values < .05 were considered statistically significant, and statistical tests were based on a 2-sided significance level. Statistical analyses were performed using the Statistical Package for Social Sciences (SPSS) software package (SPSS, Chicago, Ill).
Patient Demographics and Tumor Characteristics
Examinations for staging synchronous MPLC or restaging multiple MPLC included pathologic confirmation by biopsy or cytology and chest CT imaging in all patients, 18F-fluorodeoxyglucose-PET scanning in 98 patients (97%), and brain CT/magnetic resonance imaging in 87 patients (86%). All patients were evaluated by a multidisciplinary thoracic oncology group, which included thoracic surgeons, medical oncologists, radiation oncologists, radiologists, and pathologists. In total, 101 patients with MPLC were eligible and were included in the analyses. In all, 130 tumors (29 index tumors and 101 second tumors) were treated with SABR with dose regimens of 50 Gy in 4 fractions (for 120 tumors) or 70 Gy in 7 fractions (for 10 tumors). Numbers of patients in each diagnostic category are presented in Table 1. The median patient age at the time of MPLC diagnosis was 72 years (range, 50-90 years). Age at diagnosis was similar for those with synchronous MPLC (median, 72 years; range, 50-87 years) and those with metachronous MPLC (median, 72.5 years; range, 51-90 years). Slightly more patients were men (n = 58) than women (n = 43).
The reasons for the use of SABR instead of surgery included concomitant medical problems (79 patients) and patient preference (22 patients). Tumor characteristics are listed in Table 2. The median interval between diagnoses of the 2 tumors (index and the secondary) for the entire group was 15 months (range, 0-210 months)—1 month (range, 0-6) for synchronous MLPCs and 33 months (range, 7-210) for metachronous MPLCs. Among all 202 tumors, 188 (93%) were diagnosed as NSCLC, and 14 (7%) were diagnosed neuroendocrine carcinomas (7 small cell carcinomas, 2 large cell neuroendocrine carcinomas, and 5 carcinoids). Tumor histology was similar for the index and the secondary tumors in 68 patients (67%) and was different in 33 patients (33%). Patients with small cell lung cancer as the index tumor received treatment with conventional radiotherapy. There were no tumors with small cell lung cancer histology among the second tumors. The index and second tumors were both in the ipsilateral lung in 24 patients (24%) and were in both lungs in 77 patients (76%). According to the 7th edition (2010) of the American Joint Committee on Cancer staging system, all second tumors were stage I; of the index tumors, 71 (70%) were stage I, 16 (16%) were stage II, and 14 (14%) were stage III. Among the index tumors, 47 patients had undergone surgery (lobectomy in 42 patients, wedge resection in 3 patients, sleeve resection in 1 patient, and pneumonectomy in 1patient), 25 patients had received definitive conventional radiotherapy, 5 patients had received postoperative conventional radiotherapy at median dose of 63 Gy (range, 45-70 Gy), and 29 patients had received treatment with SABR (50 Gy in 4 fractions for 25 patients and 70 Gy in 10 fractions for 4 patients). Among the 30 patients who had received conventional radiotherapy, 13 had stage III index tumors, 9 had stage II index tumors, and 8 had stage I index tumors. All of the second tumors were treated using SABR; most patients (n = 95) received 50 Gy in 4 fractions, and the other 6 patients received 70 Gy in 10 fractions. Twenty of 39 patients who had synchronous MPLC had received chemotherapy. Of the 62 patients who had metachronous MPLC, 28 patients had received chemotherapy for the index tumor, and only 5 patients had received chemotherapy for the second tumor.
The median follow-up interval for all patients was 36 months (range, 3-80 months), and it was 48 months (range, 21-80 months) for the patients who were alive at the time of analysis. When this analysis was undertaken, 31 patients were alive and free of lung cancer, and 18 patients were alive with disease. Among the 52 deaths, 28 were related to lung cancer. For all patients, the median survival was 46 months (95% confidence interval [CI], 35-57 months), and the 2-year and 4-year overall survival (OS) rates were 73.2% (95% CI, 64.6%-81.8%) and 47.5% (95% CI, 36.7%-58.3%) (Fig. 1), respectively. The 2-year OS rate was 80.6% for those with metachronous MPLC and 61.5% for those with synchronous lesions; and the corresponding 4-year OS rates were 52.7% and 39.7, respectively (P = .047) (Fig. 1). The 2-year and 4-year progression-free survival (PFS) rates for all patients were 71.3% (95% CI, 62.1%-80.5%) and 58% (95% CI, 46%-70%), respectively (Fig. 2). PFS rates were higher for patients with metachronous tumors than for patients with synchronous tumors (2-year PFS, 84.7% metachronous vs 49.4% synchronous; 4-year PFS, 75.6% vs 30.4%; P = .0001) (Fig. 2). For patients who had index and second tumors with the same histology (indicating that the second lesion may have been a satellite, a metastasis, or a recurrent lesion), the 2-year and 4-year OS rates were 76.4% and 51.2%, respectively, which were no different from the OS rates for patients who had tumors with different histology (2-year OS, 66.7%; 4-year OS, 40.5%; P = .406). The 2-year and 4-year OS rates for patients who had both tumors classified as stage I were 76.1% and 55.2%, respectively, which were better than the rates for patients who had index tumors of higher stage (2-year OS, 66.7%; 4-year OS, 26.6%; P = .049) (Fig. 3).
Local Control of Tumors Treated With SABR and Patterns of Failure
All 130 tumors that were treated with SABR were stage I. Four of those tumors (3%) had recurred within the PTV (in-field failure), and 4 (3%) had recurred within the same lobe but outside the PTV (involved-lobe failure). The 2-year and 4-year in-field local control rates for these tumors were 97.4% and 95.7%, respectively.
Among the patients who experienced treatment failure after previous treatment (surgery or radiotherapy) and current SABR, the first site of cumulative failure was local recurrence in 12 of 101 patients (11.9%), regional lymph node recurrence in 15 of 101 patients (14.9%), and distant metastasis in 12 of 101 patients (11.9%). It is noteworthy that no difference was observed in the patterns of failure between patients who had stage I index tumors and those who had higher stage index tumors. The presence of a third primary tumor appeared to be more prevalent among those with stage I index tumors (15.5% vs 6.7% for those with higher stage index tumors; data not shown). Among the 13 patients with third primary lung tumors (defined according to our MPLC criteria), 7 received retreatment with SABR, 1 received proton therapy, 1 received systemic therapy, and 4 did not receive treatment at our hospital because of other comorbidities.
The common SABR-associated toxicities are listed in Table 3. To compare the distribution of toxicities, we classified all patients into 3 groups according to the treatment approach used for the index tumor: an SABR group (n = 29), a surgery group (n = 42 [excluding the patients who received postoperative radiotherapy]), and a conventional radiotherapy group (n = 30; including patients who received definitive and postoperative conventional radiotherapy) (Table 3). It is noteworthy that the incidence of grade 2 radiation pneumonitis (RP) was lower in group who had index tumors treated with SABR compared with those who underwent surgery and received conventional radiotherapy, indicating that lung volume loss caused by surgery or conventional radiotherapy may have a negative impact in RP. One patient (3%) in the SABR group, 1 patient (2%) in the surgery group, and 5 patients (17%) in the conventional radiotherapy group experienced grade ≥3 RP. All 5 patients who had grade ≥3 RP in the conventional radiotherapy group had index tumors that were higher than stage I, suggesting that the large conventional radiotherapy volumes used to treat these higher stage tumors may have contributed to the higher incidence of RP. The patient who experienced the sole grade 5 event (pneumonitis) had received concurrent chemoradiation therapy plus consolidation chemotherapy for a stage IIIA index tumor, and grade 2 RP had developed after that. Four years later, the patient received SABR to the contralaterally located second tumor. Two months after SABR, this patient developed first RP and then an overwhelming pneumonia that resulted in respiratory failure and death. Because this event may have been precipitated by the RP and need for prednisone, it was scored as a related event. This patients' forced expiratory volume in 1 second before SBRT had been 33%, and the composite dose-volume parameters were a mean lung dose of 22 Gy; percentage of lung volume that received at least 5 Gy (V5), 81%; and lung V20, 39%.
The incidence of chest wall pain and rib fracture was higher in the group that had index tumors treated by SABR compared with surgery and conventional radiotherapy (Table 3). However, these data comprised the cumulative incidence from 2 courses of SABR for the index tumor and the secondary tumor. We did not count rib fractures caused by surgical procedures. There was a correlation between rib fracture and chest wall pain: 20 of 24 patients with rib fracture developed grade 1 or 2 (n = 18) or grade 3 (n = 2) chest wall pain.
In the current study, 130 stage I MPLC tumors were treated with SABR. The 4-year in-field local control rate of 95.7% was satisfactory and consistent with previous reports of SABR for stage I single primary lung cancers.[7-10] Radiation dose is crucial for local tumor control, as reported previously, and a BED of >100 Gy to the target volume is needed to achieve optimal local control. In our current study, all tumors had received SABR with a BED >100 Gy (112.5 Gy at 50 Gy in 4 fractions and 119 Gy at 70 Gy in 10 fractions). There was 1 patient with MPLC who received 40 Gy in 4 fractions (BED = 80 Gy) who was not included in the current study but was reported previously. This patient developed an in-field local recurrence, and we have not used that regimen since our previous publication. Most important, the OS and PFS rates at 4 years were 47.5% and 58%, respectively, for all patients and 52.7% and 75.6%, respectively, for patients who had metachronous lesions. These findings suggest that SABR may be able to cure up to 58% of all MPLCs and 75.6% of metachronous MPLCs.
The diagnosis of MPLC in routine practice is usually not easy, and there is a potential conflict between diagnoses of MPLC and oligometastasis. A difference in histology in the primary and secondary tumors is considered a reliable indicator of MPLC; however, if both tumors have the same or similar histology, then it can be difficult to distinguish a second primary carcinoma from a recurrent, metastatic, or satellite lesion arising from the first tumor. Clinically, patients who have tumors that meet the MPLC criteria listed in Table 1 often are considered to have metastatic disease and are offered palliative therapy, such as chemotherapy, which would be expected to have a negative effect on outcomes. In our study, no difference in survival was observed between patients who had tumors with the same or different histologies. Thus, at least some second tumors that had the same histology as the first tumor actually were secondary primary carcinomas and, as such, warranted active local treatment with curative intent.
For patients with stage I NSCLC, similar outcomes after SABR and surgery treatments have been reported by several groups,[22-25] although none of those studies involved data from randomized clinical trials. For MPLC treated by surgical resection, Aziz et al reported a median OS of 40 months and a 5-year survival of rate 38%. Rice et al analyzed a cohort from a prospective study and demonstrated that, even when both the index tumor and the second tumor were stage I, the median survival after surgical resection was 49.2 months. With a median OS of 46 months and an OS rate of 47.5% at 4 years (estimated OS, 37.9% at 5 years) in our study, the effectiveness of SABR appears to be comparable to that of surgery for early MPLC.[2-5] However, a more detail analysis, such as a case-control or propensity-matched study, should be performed to validate our observation. It is noteworthy that the OS for patients who had both a stage I index tumor and a stage I second primary tumor was similar to the OS for patients who had a single, pure stage I primary lung cancer. These results further suggest that MPLC should be treated with curative intent rather than palliation as metastatic disease. Because patients with MPLC have other significant medical issues that may preclude surgery, SABR provides an attractive alternative approach for a potentially curative treatment.
In our study, patients with synchronous disease experienced significantly reduced survival compared with those who had metachronous disease, a finding similar to what has been published elsewhere.[26-28] Although the possible explanation for this difference may lie in the impact of the stage of the index tumor on survival, in our cohort, the proportion of higher stage (II and III) index tumors was almost same for those with synchronous disease and those with metachronous disease. The median interval between diagnosis of the index tumor and the second tumors in patients with metachronous disease was 33 months compared with 1 month in patients with synchronous disease; this means that >50% of index tumors had been controlled for more than 33 months, indicating the inherently favorable nature of the disease. In addition, some of the synchronous MPLCs may have been metastatic disease from the index tumor. It is noteworthy that no significant difference in the pattern of failure was observed between patients who had stage I index tumors and those who had higher stage index tumors, indicating that some index tumors had been cured even when those tumors were higher than stage I. The incidence of third primary tumors appeared to be higher among patients with stage I index tumors (P value not significant), possibly indicating that patients with stage I index tumors have the highest chance of curative treatment, the longest survival, or both.
RP was the most common SABR-associated toxicity in this study. However, most grade ≥3 RP occurred in patients who had received conventional radiotherapy for their index tumors (5 of 30 patients; 17%), a rate similar to that in our previous studies of SABR for patients who received previous conventional thoracic radiotherapy, and a predictive model of RP has been reported.[14-17] All 5 of our patients who had received conventional radiotherapy and developed grade ≥3 RP had had stage I or higher index tumors (total, 22 patients), and none of the patients with stage I index tumors who had received with conventional radiotherapy (8 patients) experienced grade ≥3 RP. Treatment volumes were not available for all of the patients in our study, because some had been treated at outside hospitals several years before. However, considering the correlation between treatment volume and tumor stage, we assume that the larger treatment volumes may have contributed to the development of these side effects in addition to the use of conventional radiotherapy techniques. The rate of grade ≥3 RP in patients whose index tumors were treated with SABR or surgery (<3%) was similar to the rate among patients who had received SABR for single tumors.[7-9] These findings suggest that careful attention should be paid to patients with MPLC who have more advanced tumors and have received both SABR and conventional radiotherapy.
The rates and severity of other toxicities were considered acceptable and comparable to those in other studies of SABR for single lung tumors. Three patients in our study (3%) experienced grade 3 chest wall pain, a rate similar to that in the largest analysis of chest wall toxicity after SABR for stage I NSCLC to date, in which severe chest wall pain was observed in 2.2% of patients.[29-31] All 3 of the patients with grade 3 chest wall pain in the current study had received SABR for both tumors, and all 6 of their tumors were located close to the chest wall, which probably contributed to the development of severe chest wall pain. However, the pain in those patients probably was not related to dose overlap of the SABR fields on the chest wall, because the 2 tumors in these patients were located bilaterally. The factors that predict the occurrence of chest wall pain have been explored in previous studies.[29-31]
To our knowledge, the current study involves the largest number of patients with MPLC and the longest follow-up of patients with MPLC who received treatment with SABR. To date, only a few studies[11-13] have been evaluated the use of SABR for MPLC. Creach et al retrospectively analyzed 63 patients with MPLC who received SABR at least 1 tumor. The 2-year OS in that study was 58.5%, and no grade ≥3 toxicities were reported. Two other studies[12, 13] included only 10 patients each, and all of the lung tumors in those patients had been treated with SABR. Favorable therapeutic effects and minimal toxicity were reported. It is difficult to compare those studies with ours because of differences in patient selection criteria and length of follow-up. However, the promising treatment outcomes and the acceptable toxicity profile in reports from us and from others lead us to believe that SABR can be an effective alternative to surgery for patients with early stage MPLC tumors.
This study did have some limitations. First, the results are based on a group of patients that was selected retrospectively from our SABR program according to the criteria listed in Table 1. Second, we did not conduct more a detailed comparison between surgery and SABR to adjust for other variables. Finally, this study focused mostly on patients who had only 2 tumors, which may limit the generalizability of our findings to MPLCs that include more tumors. In summary, the current results indicate that: 1) SABR achieves an excellent long-term tumor control and promising PFS and OS in patients with early stage MPLC; 2) toxicity may occur but usually is within the scope of SABR in those with stage I disease; and 3) caution should be taken when integrating SABR with previous comprehensive radiotherapy for patients with stage II/III disease.
This research was supported in part by National Institutes of Health Clinical and Translational Science Award UL1 RR024148 and by National Cancer Institute Cancer Center Support Grant CA016672 to The University of Texas MD Anderson Cancer Center.