Purpose/objective: With the increasing number of patients successfully treated with stereotactic radiosurgery for brain metastases, decision making after therapy based on follow-up imaging findings becomes more and more important. Magnetic resonance imaging (MRI) is the most sensitive means for follow-up studies. The objective of this study was to investigate the treatment outcome of our radiosurgery program and to describe the response of brain metastases to contrast-enhanced MRI after linear accelerator (linac) stereotactic radiosurgery and identify factors to distinguish among local control and local failure.
Methods and materials: Using serial MRI, we followed the course of 87 brain metastases in 48 consecutive patients treated between September 1996 and November 1997 with linac-based radiosurgery with 15-MV photons. Treatment planning was performed on an MR data cube. For spherical metastases, radiosurgery was delivered using a 9 noncoplanar arc technique with circular-shaped collimators. For irregularly shaped targets, radiosurgery was delivered using a manually driven multi-leaf collimator with a leaf width of 1.5 mm projected to the isocenter. Median radiosurgery dose was 20 Gy prescribed to the 80% isodose. Together with whole brain radiotherapy (20 x 2 Gy, 5/w), a median radiosurgical dose of 15 Gy was delivered. Median follow-up was 8 (range 2--36) months. Factors influencing local control and survival rates were analyzed with respect to MRI response, and Kaplan-Meier curves were calculated.
Results: Actuarial local tumor control was 91% at one and two years. Patient survival at one and two years was 30% and 18%. Median survival was 9 months. During follow-up in 70 (81%) of the 87 treated metastases, the contrast-enhancing volumes on T1W images were stable or disappeared partly or completely. A transient enlargement of contrast-enhancing volumes was observed in 11 (12%) of the 87 lesions treated, while a progressive enlargement due to local treatment failure was observed in 6 (7%) of the 87 treated metastases. Younger age, early contrast onset after radiosurgery, and previous chemotherapy were associated with this transient enlargement of contrast-enhancing lesion volume.
Conclusions: Linac-based radiosurgery is an effective, noninvasive, and safe treatment option for patients with brain metastases. A marked enlargement of the contrast-enhancing volume on T(1)-weighted MR images after radiosurgery is a sensitive predictor for, but not equivalent with, local failure. In as many as two-thirds of the cases with contrast enlargement in MRI follow-up, the contrast enlargement is transient with no need for further treatment. While some MRI findings are more likely if transient enlargement is present, a clear decision cannot be made based on MRI, and ultimately the clinical status dictates further action.