Stereotactic body radiotherapy (SBRT) for high-risk central pulmonary metastases

Metastatic lesions of the central pulmonary tree negatively impact a patient’s quality of life and present complex radiobiological challenges for the treating radiation oncologist. Given their sensitive location these lesions can provoke severe cough, shortness of breath, and/or hemoptysis. Although chemotherapy is typically first-line treatment for metastatic disease it is unlikely to provide durable local control for such lung metastases [4]. Radiotherapy offers a very effective treatment option for both palliation and local control. SBRT has emerged as an ideal treatment modality given the excellent local control rates observed in the definitive treatment of early stage inoperable NSCLC [10]. Particularly important in the metastatic setting, SBRT offers a shorter treatment length yielding a minimal break if any of systemic therapy, which is still the backbone of treatment for the metastatic patient. SBRT for lung metastases has been previously shown to yield effective local control with minimal toxicity [17]. Nevertheless, Timmerman et al. reported higher rates of treatment-related toxicity in central relative to peripheral primary NSCLC tumors in the definitive setting [5]. Moreover, Baumann et al. demonstrated inferior rates of local control in central relative to peripheral tumors [18]. Conversely, recent accumulating retrospective data of admittedly heterogeneous dose and fractionation schedules seems to indicate the danger of irradiation to this central region may not have as clinically significant an impact as previously postulated [13, 19‐21]. Broadly speaking treatment-related toxicities can be seen in several anatomic locations including: mainstem bronchus, lung parenchyma, esophagus, brachial plexus and vagus nerve, and chest wall [8]. Many of these toxicities can be limited by placing dosimetric constraints on adjacent organs, however central tumors are often intricately involved with mainstem bronchi making achieving these constraints more difficult.

A European analysis of factors contributing to the efficacy of SBRT for inoperable stage I NSCLC found that central lung tumors were both more likely to fail locally and to develop atelectasis as a treatment-related toxicity [18]. Likewise, the most common late toxicity observed in our patient population was atelectasis. Interestingly, Joyner et al. reported major airway wall thickening evident on follow up CT scans as a common observation in their cohort of central lung tumors treated with SBRT [11]. Although the pathophysiology of this atelectasis is not firmly established in the literature, it is hypothesized to be secondary to dose-dependent radiation-induced damage of the bronchial wall [8]. This dose-dependence is consistent with our observation of maximum point dose to the major bronchus as the most significant predictor of late grade 2 or higher atelectasis. Given the hypothesized etiology of bronchial wall damage it may be prudent to consider the major bronchus as a serial organ at risk and make concerted efforts to avoid hot spot placement in this region. As toxicity and radiological data accumulates an evolution in our understanding of the bronchial wall as a dose limiting structure is paramount.

It is important to note that not all central pulmonary lesions are created equally. Hamamoto et al. demonstrated that metastatic lung lesions rather than NSCLC primary lesions tend to have higher rates of local recurrence [22]. Additionally, Milano et al. reported lower rates of local control in pulmonary metastases of gastrointestinal origin [23]. Similarly, Singh et al. reported outcomes of 5-fraction SBRT to a higher total dose (50 Gy) for treatment of oligometastatic thoracic disease and found all 5 patients with local failures were those with colorectal primaries [24]. Although, in the present study there was no significant difference in duration of local control or overall survival between metastases of GI and non-GI origin (p = 0.12 and p = 0.23, respectively), there was a trend towards increased local progression in the GI subset. Consequently, it is crucial to note that tumors with lower α/β may necessitate altered fractionation schemes to achieve durable local control.

In a similar study, Oshiro et al. reported their single institutional experience in a retrospective review of 21 patients who all received linac-based SBRT to the pulmonary hilum [7]. In this study, both metastatic and primary hilar lung tumors were treated to a median dose of 50 Gy in 5 fractions (range of 25 to 60 Gy in 1 to 13 fractions). Hilar metastases in this study were defined as within 2 cm of the mainstem bronchus. Reported local control rates of 74.3 and 59.6 % were observed at 1 and 2 years, respectively. There were no reported grade 2 or above acute toxicities. Observed late grade 3 or above toxicities included 1 chronic dyspnea requiring supplemental oxygen, 1 recurrent bronchial obstruction causing intractable cough, and 1 fatal hemoptysis. Of note, the one reported fatal hemoptysis occurred in a patient who was previously treated twice with thoracic radiation. Notwithstanding the fact that tumors in our study were intricately involved with the central airway rather than just within 2 cm, reported grade 3 or higher late toxicities were less frequent. Additionally, the 12-month local control rate of 74.3 % is similar to that reported in our study (70.1 %) despite the lower median dose delivered to our cohort.

Our 2-year local control rate of 57.4 % is notably inferior when compared broadly with that seen in SBRT-treated central tumors elsewhere in the literature (64 to 95 %) [21]. This is likely a reflection of the following factors: (1) SBRT treatment of metastatic pulmonary lesions demonstrates inferior local control relative to primary NSCLC, (2) the high-risk nature of these tumors in comparison to standard-risk or RTOG defined central tumors, and (3) the lower BED employed in their treatment [9, 21, 22, 25]. Total BED may be the root cause of this inferior local control. A very large (613 patient) retrospective German analysis of centrally located stage I NSCLC found significantly lower BEDs were used to treat central versus peripheral tumors (72.0 versus 84.4 Gy, respectively) and this resulted in a notably lower 3-year freedom from local progression rate (52 versus 84 %, respectively) [25]. There is an indication from a small 7 patient Stanford cohort of “ultra-central” tumors that these high-risk lesions may be successfully treated with higher doses (median 12.5 Gy × 4) to achieve excellent 2-year local control of 100 % without significant increase in toxicity [13]. Of note, 6 of the 7 ultra-central tumors were of primary rather than metastatic histology and would be expected to have higher local control rates compared to metastatic lesions. Overall, data regarding the safety and efficacy of SBRT to high-risk central tumors is scant, and additional research is essential to determine if dose escalation is safe and can achieve superior local control.

As our understanding of the optimal SBRT treatment for central pulmonary lesions evolves, recent data suggests treatment of these lesions may not be as dangerous as previously suspected. Mangona et al. reported the results of a propensity score matched-pair analysis of 158 central and peripheral lung tumors (primary or metastatic) treated with SBRT (range of 48 to 60 Gy in 4 to 5 fractions) [19]. Central tumors were defined as within 2 cm of the proximal bronchial tree or PTV touching the mediastinum. Interestingly, there was no observed significant difference in any 2-year reported adverse event based on tumor location. Moreover, the 2-year incidence of grade 4 and 5 adverse events was exceedingly low at <1 and 0 %, respectively. Park et al. retrospectively investigated a cohort of 251 patients with similarly defined central NSCLC primary tumors treated with SBRT [20]. Although patients in the central lesion cohort had significantly larger tumors and were of older age, multivariate analysis found central tumor location was not associated with inferior toxicity, local control, or overall survival. Finally, Davis et al. published a 111 patient cohort with central primary or metastatic lung tumors treated with SBRT and found no grade 3 or higher acute or late toxicities [21]. Nevertheless, they did report a lower 2-year rate of local control in metastatic versus primary lung lesions (69.8 versus 76.4 %) as would be expected. Additional prospective data with appreciable follow-up will be necessary to confirm if these intriguing modern retrospective reports of minimal SBRT-related toxicity are reproducible.

Endobronchial lesions may represent the highest of high-risk pulmonary metastatic lesions. Survival after diagnosis of endobronchial tumors is generally poor and in our admittedly very small cohort these lesions were shown to have significantly lower overall survival relative to non-endobronchial high-risk central lesions [26]. Overall, diagnosis of endobronchial metastatic tumors is quite rare with autopsy series reporting a 2 % incidence in solid malignancies [27]. These lesions are most commonly secondary to lung cancers, but can also arise from metastatic breast, colon, or renal cell carcinoma [26‐28]. Much of the literature surrounding management and clinical outcomes of endobronchial lesions is limited to case reports and retrospective reviews of small cohorts. There is some data to suggest these lesions respond poorly to systemic therapy and can lead to meager median survivals as low as 9 months [29]. Nevertheless, these lesions tend to be quite symptomatic prior to treatment and significant mitigation of symptoms can be achieved with local therapy [29].

Our study reports the efficacy of a 5-fraction SBRT treatment regimen for high-risk central pulmonary metastases. Of the patients who had symptomatic lesions prior to treatment the majority (64 %) had effective mitigation of their symptoms. Relatively durable tumor control was observed with a 12-month local control of 70.1 %, which translated into a 1-year overall survival of 75 %. Late grade 2 or higher atelectasis was the most common treatment-related toxicity and correlated significantly with a increased maximum point dose to the major bronchus. Furthermore, those patients with invasive endobronchial lesions demonstrated a significantly lower overall survival and a trend towards shorter local control. Limitations of the present study include the small patient population, the heterogeneity of metastatic histologies, and its retrospective analysis. As evidence accumulates regarding the lower than expected toxicity of SBRT treatment of central lung tumors the investigation of dose escalation for these high-risk lesions seems warranted in attempt to achieve superior local control closer to that seen in standard-risk central lung tumors. Reports from the RTOG 0813 and EORTC 22113-0813 Lungtech trials exploring the proper SBRT fractionation schedule for central primary lung tumors may be extrapolated to help address the optimal fractionation schedule for precarious central pulmonary metastatic lesions.