This analysis represents a single-center experience in treating oligometastatic lung lesions with curative intended SFRS and fSBRT. The 1-, 2-year LC and OS rates for the entire cohort were 82%, 70% and 84%, 71%, respectively. Our findings are comparable with the current findings in the literature (Tab. 7) (7-15).
It has been postulated by others that a BED > 100 Gy, smaller tumor size, shorter interval between diagnosis and treatment of metastases are favorable prognostic factors influencing the local control of lung metastases after SBRT (8, 16-18). The existing data on fractionation schedules as well as dosage of SBRT for lung metastases is limited by retrospective nature. Therefore, no standardized treatment regimens are yet available. The primary results of TROG 13.01 SAFRON II Phase II study which compares SFRS to fSBRT for lung metastases are expected soon (19). Meanwhile, fractionation schedules for small (T1-T2, N0, M0) NSCLC tumors were investigated in several Phase 2 trials. RTOG 0618 trial reported 54 Gy delivered in 3 fractions for operable peripheral tumors to be safe and effective (LC at 4-years 96%, with 8% grade 3 adverse events) (20). Timmerman et al. concluded that dose escalation up to 3 x 20/21 Gy for central lesions in inoperable patients should not be applied due to high toxicity rates (21). Furthermore, Videtic and colleagues compared grade ≥ 3 adverse event rates after SBRT with 34 Gy in 1 fraction vs. 48 Gy in 4 fractions for inoperable NSCLC. Analyzing a cohort of 94 patients they found that SFRS with 34 Gy is superior compared to fSBRT regarding safety (10.3 % vs. 13.3%) and 1-year LC rates (97% vs 92.7%) (22). According to our data, small lung metastases (median PTV ≤9.9 cm³) might safely be treated with SFRS applying 24-26 Gy at the surrounding 70% isodose (median Dmax of 53 Gy and a median BED of 81 Gy) with excellent 1- and 2-year LC rates of 89% and 83%. This finding suggests that a BED <100 Gy using SFRS might be sufficient for durable control in some lung metastases. Previous studies rarely investigated SFRS regimes for different tumor volumes especially for small lesions. In accordance with our data, Filippi and colleagues reported excellent 1- and 2-year LC rates (93.4% and 88.1% respectively) after SFRS with 26 Gy (80% surrounding isodose) for lung metastases with a median tumor diameters of 17 mm (7-38 mm) in 67 patients with 90 lung lesions mainly from NSCLC and CRC primaries (14). Siva et al. observed no difference in LC, OS and distant progression between 65 patients with mainly NSCLC or CRC treated with either SFRS (26 Gy for peripheral LM or 18 Gy for central LM) or multifractionated SBRT (48Gy/4 and 50Gy/5 for peripheral targets or 50Gy/5 for central targets) for 1 to 3 lung oligometastases (23). Randomized, prospective studies are needed to determine which fractionation schedule is optimal in terms of therapy outcomes, treatment related costs as well as patient’s compliance.
Recently, Hong et al. developed a prognostic tool for the discrimination of oligometastatic patients with extracranial lesions who benefit most from SBRT. The authors found that primaries from breast, kidney and prostate cancer lead to long-term survival with 3-year OS rates of 75% (24). Furthermore, a retrospective analysis of 700 patients with lung metastases found that breast cancer and CRC were positive prognostic factors for superior OS (8). Newly, Sharma and colleagues reported 38% lower mortality risk in patients with lung oligometastases from CRC compared with non-CRC histology. This OS improvement might be explained by the broad arsenal of effective systemic therapeutic agents and radical ablative treatment for the primary tumor while breast cancer is associated with less aggressive tumor biology (10). Our cohort consisted of patients with 10 different tumor entities with the leading diagnosis of CRC. There was no significant OS difference regarding the primary cancer most likely due to the heterogeneous cohort and small sample size. Nonetheless, our data suggest that metastases from CRC have a higher risk of local failure compared to other histologies. One and 2-year LC rates for metastases from CRC vs. non-CRC were 59% and 46% vs. 90% and 80%, respectively. Recently, systematic review and meta-analysis have been published in which the prognostic value of CRC histology after SBRT for lung lesions was investigated. Analysis of 1920 patients (619 with CRC, 1301 non-CRC) showed that LC was significantly inferior in the CRC group (p <0.00001). Furthermore, the dose escalation (BED >130 Gy) was associated with decreased local recurrences (p <0.00001) (25). Ahmed et al. found that lung metastases from rectal carcinoma, soft-tissue sarcoma, renal cell carcinoma and melanoma are related with increased radio-resistance and thus reached significantly worse LC after SBRT. The authors recommend a BED beyond 100 Gy for radio-resistant lung metastases in order to improve LC rates (26). In the present study, 75.9% pulmonary metastases from CRC originated from primary rectal tumors. The median BED for relapsed pulmonary lesions from rectal carcinoma was 87.5 Gy (range: 56-124.8). We hypothesize that median BED less than 100 Gy for radio-resistant metastases might be associated with high local relapse rates in patients with rectal cancer observed in our cohort.
According to our data, N0 stage and long DMFS from the diagnosis of primary tumor seems to play an important role in predicting OS for patients with lung oligometastases. Therefore, assessment of these potentially prognostic factors might be useful for selecting patients with stage IV cancer for curative SBRT. In the present study, patients with N0 and DMFS ≥12 months had excellent 1- and 2-year OS rates of 100% and 75%. In addition, a trend for better OS was observed in patients with smaller primary tumors (≤T2). Our results are in agreement with a meta-analysis including 757 patients with oligometastatic NSCLC. After performing recursive partitioning analysis, the authors identified 3 risk groups regarding to N status and DMFS: low risk (N0 and DMFS ≥2 months), intermediate risk (N0 and DMFS ≤2) and high risk (N+ and DMFS ≤2 months). There was a significant 1- and 2-years OS difference among the groups 88.4% and 66.3% vs. 76.2% and 57.4% vs. 53.6% and 34.1%, respectively (p <0.001). Moreover, a higher T-stage predicted inferior OS in univariate analysis. However, T-stage could not be established as independent prognostic factor in a multivariate analysis (27). The evidence of other retrospective series support the data that shorter DMFS as well as synchronous metastases are associated with more aggressive tumor behavior and thereby decreased OS rates in patients with pulmonary metastases (28-30). The present findings are consistent with results from the study of Inoue et al. which demonstrated DMFS ≥12 to be a positive prognostic factor for OS among the patients with extracranial (lung, adrenal glands) and intracranial oligometastases (31). Cao et al. reported superior OS rates in patients with pulmonary metastases after mastectomy with DMFS ≥18 (32). In contrast to most published data, a recent retrospective study in a cohort of 206 patients with lung oligometastases identified synchronous metastasis as an independent prognostic factor for longer OS. (10) The possible explanation for the conflicting results named by the authors was a selection bias. ew prospective studies showed that patients even with initial stage IV NSCLC reach longer PFS or OS after ablative therapy for all active lesions and benefit from local therapy despite the advanced disease (33-36).
For the definition of oligometastatic disease, the number of metastases considerably varies between 1-5 lesions in 1-3 affected organs (37-39). Presence of multiple metastases is usually reported as one of the most important negative predictive factors for OS (8, 39-41). Although, a correlation between number of lesions and OS was not observed in the current study, our data demonstrated significantly longer PFS in the patient cohort with maximum 3 metastases (2-year PFS rates of 26.1 % vs. 0 %, p=0.002). Similar findings were observed in a retrospective analysis of 66 patients with CRC where multiple lung lesions were associated with worse PFS (15). Furthermore, one surgical study of 615 patients with lung metastases from CRC identified 3 metastases as a cut-off number for selection of operable patients (42). Nonetheless, other studies failed to demonstrate any relationship between number of metastases and survival outcomes (16) (43). The strict number of lesions alone might be an insufficient factor to predict survival outcomes since the metastases might vary significantly in size (44). Therefore, we assessed TTD and found a trend towards improvement in PFS for TTD≤ 5 cm vs. TTD>5 cm (1-year PFS 40% vs. 14%, p=0.073). Although our finding was not significant, TTD might be an important factor for discrimination of the patients with oligometastases. In a retrospective study on stage IV gastric cancer the metastatic tumor diameter ≥ 55 mm was found to be an adverse prognostic factor for OS after systemic therapy (45). Further studies are needed to examine the effect of TTD and number of metastases on outcome.
Masahiko et al. investigated the risk of rib fracture after SBRT (54–56 Gy in 9–7 fractions) for peripheral lung tumors. The 4-year probability of rip fracture was 47.7% when Dmax for the ribs was 54 Gy or more. In our cohort we did not observe any rib fractures. This might be explained by retrospective nature of the study and median follow-up that less than 4 years (46).
The major limitation of our study is its retrospective design with inhomogeneous primary tumor types. Different dose calculation algorithms were used. Treatment plan calculations with Ray-Tracing, Pencil Beam or Monte Carlo dose algorithms can produce large differences in dose for targets and organs at risk near lung tissue. A correction factor (e.g. 1.2) which should provide approximate equivalence between Ray-Tracing and Monte Carlo dose calculation algorithm was not used. Furthermore, multiple metastases in the same patient were treated with different fractionation regimens, so the assessment of prognostic value of SFRS vs. fSBRT for survival outcomes (OS, PFS, DMFS) was not possible. Nevertheless, our analysis provides valuable data on fractionation regimes and prognostic factors for OS and PFS treating lung metastases in oligometastic patient cohort, irrespective of a relatively small sample size and dispersed follow up periods.