Up to Four-Year Outcomes of Single-Fraction Stereotactic Radiosurgery for Small Brain Metastases: A Competing-Risk Analysis.
Single-fraction stereotactic radiosurgery (SRS) is recommended for small brain metastases (< 2 cm). However, long-term outcomes accounting for competing risks-particularly death, which precludes observation of local treatment failure (LTF)-remain limited.
We retrospectively analyzed cumulative incidences of LTF, defined as tumor regrowth within the planning target volume, following 18 Gy single-fraction SRS (July 2014-December 2021) in patients with brain metastases. Patients with prior radiotherapy or surgery to the index lesion were excluded. Cumulative incidences of LTF, death without prior LTF, overall survival (OS), and progression-free survival (PFS) were estimated. Outcomes were compared between radioresistant (melanoma, sarcoma, renal cell carcinoma) and radiosensitive tumors. Subdistribution hazard ratios (SDHRs) and cause-specific hazard ratios (CSHRs) identified risk factors for each endpoint.
Seventy-two patients (median age 62 years, IQR 55-68) with median tumor volume 0.3 cc (IQR 0.2-0.7) were followed for a median of 57.2 months (IQR 49.3-85.4). Thirty-three tumors were radioresistant and 39 radiosensitive. At 36 months, cumulative LTF was 62% (95% CI, 42%-77%) versus 31% (95% CI, 17%-46%) in radioresistant versus radiosensitive tumors (p = 0.011), whereas death without prior LTF was 19% (95% CI, 7%-34%) versus 46% (95% CI, 30%-61%) (p = 0.003). OS was 42% (95% CI, 28%-63%) and 36% (95% CI, 24%-55%), and PFS was 19% (95% CI, 9%-40%) and 23% (95% CI, 13%-41%), respectively. Radioresistant histology was associated with higher LTF hazard (CSHR 3.19; SDHR 3.91) and lower hazards of death without prior LTF (CSHR 0.37; SDHR 0.24), with no significant effect on OS or PFS. The 36-month cumulative incidence of radiation necrosis was 10% (95% CI, 4%-18%).
Local control after 18 Gy SRS is poor in radioresistant brain metastases. Higher LTF incidence is amplified by lower competing mortality, highlighting the interplay between tumor biology and survival. These findings support competing risk-based analyses and consideration of alternative dose-escalation for high-risk tumors.
We retrospectively analyzed cumulative incidences of LTF, defined as tumor regrowth within the planning target volume, following 18 Gy single-fraction SRS (July 2014-December 2021) in patients with brain metastases. Patients with prior radiotherapy or surgery to the index lesion were excluded. Cumulative incidences of LTF, death without prior LTF, overall survival (OS), and progression-free survival (PFS) were estimated. Outcomes were compared between radioresistant (melanoma, sarcoma, renal cell carcinoma) and radiosensitive tumors. Subdistribution hazard ratios (SDHRs) and cause-specific hazard ratios (CSHRs) identified risk factors for each endpoint.
Seventy-two patients (median age 62 years, IQR 55-68) with median tumor volume 0.3 cc (IQR 0.2-0.7) were followed for a median of 57.2 months (IQR 49.3-85.4). Thirty-three tumors were radioresistant and 39 radiosensitive. At 36 months, cumulative LTF was 62% (95% CI, 42%-77%) versus 31% (95% CI, 17%-46%) in radioresistant versus radiosensitive tumors (p = 0.011), whereas death without prior LTF was 19% (95% CI, 7%-34%) versus 46% (95% CI, 30%-61%) (p = 0.003). OS was 42% (95% CI, 28%-63%) and 36% (95% CI, 24%-55%), and PFS was 19% (95% CI, 9%-40%) and 23% (95% CI, 13%-41%), respectively. Radioresistant histology was associated with higher LTF hazard (CSHR 3.19; SDHR 3.91) and lower hazards of death without prior LTF (CSHR 0.37; SDHR 0.24), with no significant effect on OS or PFS. The 36-month cumulative incidence of radiation necrosis was 10% (95% CI, 4%-18%).
Local control after 18 Gy SRS is poor in radioresistant brain metastases. Higher LTF incidence is amplified by lower competing mortality, highlighting the interplay between tumor biology and survival. These findings support competing risk-based analyses and consideration of alternative dose-escalation for high-risk tumors.