The patterns of failure of early stage NSCLC after SBRT have been reported by multiple institutions, however no institution to date has specifically reported the patterns of failure after SBRT for BAC [2, 3, 5, 6, 8, 9]. Ongoing prospective trials exploring SBRT for early stage NSCLC (RTOG 0618, RTOG 0915) have excluded BAC [23, 24]. Thus, reported outcomes from these studies cannot be used to infer utility of SBRT for BAC.
A large SEER analysis found that patients with bronchioloalveolar carcinoma that underwent lobectomy had improved overall survival compared to those that underwent wedge resection or segmentectomy . However, prospective studies have found that wedge resection is potentially curative for patients with non-invasive BAC [29–31].
The pattern of failure after surgical removal of BAC was evaluated in a series of 93 patients with Stage I-IV histologically confirmed BAC that underwent wedge resection, lobectomy or pneumonectomy for their disease. Any lung recurrence was reported in 39 (42%) patients. Thirteen (33%) of the recurrences were multiple lung nodules, eight (21%) as a single nodule or mass, four as an infiltrating or pneumonic mass, four as miliary lesions, four as mixed alveolar and nodular opacity, four as lymphangitic spreading, and one each as a pleural effusion or rib destruction .
It is estimated that the risk of developing a second lung cancer in patients who survived resection of a non-small-cell lung cancer is approximately 1%-2% per patient per year . Unless the patient’s second lung cancer is histologically unique it is difficult to ascertain whether the second lung cancer is a metastasis of the first cancer or a de novo cancer. This issue caused uncertainty in coding the patterns of failure after SBRT in our study. We elected to code multiple synchronous recurrences in the lung outside of the field of treatment as pulmonary metastases and a single new mass/nodule outside of the field of treatment as a second lung cancer.
We found that there is a trend towards an increased rate of metastases after SBRT in patient with a radiologic or histologic diagnosis of BAC compared to those with a diagnosis of non-BAC NSCLC (Figure 3C). This finding may be confounded by the known propensity for BAC to spread diffusely throughout the lungs. When considering only pulmonary metastases, there was no significant difference between patients with BAC compared to those with non-BAC NSCLC, but the results were limited by a smaller sample size.
As described above, patients with BAC were treated in an identical fashion as other patients with early stage NSCLC (Figure 1). We had no local failures amongst patients with BAC, and 12% rate of local failure (n = 12) in patients with non-BAC NSCLC which is consistent with the rates of local failure seen in SBRT series from other institutions [2–6, 8, 9]. There is a concern for marginal recurrence of BAC after less extensive local therapy such as a wedge resection instead of standard of care lobectomy. As mentioned previously a number of surgical series have found excellent local control rates after wedge resection for non-invasive BAC [29–31]. However, the local control rates after wedge resection for invasive Stage I NSCLC including those with a BAC component are not as favorable. A randomized trial of lobectomy versus a limited resection (wedge or segmentectomy) for T1 N0 NSCLC found increased rates of locoregional recurrence (17% vs 6%) and any recurrence (34% vs 26%) in patients who received a limited resection compared to those receiving lobectomy. These higher rates of recurrence led to a 50% increased rate of death with cancer and a 30% increase in overall death rate .
SBRT for localized BAC provides excellent local control and has similar patterns of failure vs. non-BAC NSCLC. Increased risk of distant metastases in this series is likely confounded by interpretation of new BAC versus true pulmonary metastases as implied by the similar rates of extrapulmonary failure. Patients with BAC should be considered for enrollment on ongoing and future prospective studies of SBRT.