Time-delayed contrast-enhanced MRI improves detection of brain metastases and apparent treatment volumes.

OBJECTIVE Contrast-enhanced MRI is the preeminent diagnostic test for brain metastasis (BM). Detection of BMs for stereotactic radiosurgery (SRS) planning may improve with a time delay following administration of a high-relaxivity agent for 1.5-T and 3-T imaging systems. Metastasis detection with time-delayed MRI was evaluated in this study. METHODS Fifty-three volumetric MRI studies from 38 patients undergoing SRS for BMs were evaluated. All studies used 0.1-mmol/kg gadobenate dimeglumine (MultiHance; Bracco Diagnostics) immediately after injection, followed by 2 more axial T1-weighted sequences after 5-minute intervals (final image acquisition commenced 15 minutes after contrast injection). Two studies were motion limited and excluded. Two hundred eighty-seven BMs were identified. The studies were randomized and examined separately by 3 radiologists, who were blinded to the temporal sequence. Each radiologist recorded the number of BMs detected per scan. A Wilcoxon signed-rank test compared BM numbers between scans. One radiologist determined the scan on which BMs were best defined. All confirmed, visible tumors were contoured using iPlan RT treatment planning software on each of the 3 MRI data sets. A linear mixed model was used to analyze volume changes. RESULTS The interclass correlations for Scans 1, 2, and 3 were 0.7392, 0.7951, and 0.7290, respectively, demonstrating excellent interrater reliability. At least 1 new lesion was detected in the second scan as compared with the first in 35.3% of subjects (95% CI 22.4%-49.9%). The increase in BM numbers between Scans 1 and 2 ranged from 1 to 10. At least 1 new lesion was detected in the third scan as compared with the second in 21.6% of subjects (95% CI 11.3%-35.3%). The increase in BM numbers between Scans 2 and 3 ranged from 1 to 9. Between Scans 1 and 3, additional tumors were seen on 43.1% of scans (increase ranged from 1 to 14). The median increase in tumor number for all comparisons was 1. There was a significant increase in number of BMs detected from Scan 1 to Scan 2 (p < 0.0367) and from Scan 1 to Scan 3 (p < 0.0264). In 34 of the 51 subjects (66.7%), the radiologist selected the third scan as the one providing the clearest tumor definition. There was an average 25.4% increase in BM volume between Scans 1 and 2 (p < 0.0001) and a 9% increase in BM volume between Scans 2 and 3 (p = 0.0001). CONCLUSIONS In patients who are being prepared for SRS of BMs, delayed MRI after contrast injection revealed more targets that needed treatment. In addition, apparent treatment volumes increased with a time delay. To avoid missing tumors that could be treated at the time of planned SRS and resultant "treatment failures," the authors recommend that postcontrast MR images be acquired between 10 and 15 minutes after injection in patients undergoing SRS for treatment of BMs.