Detection of cerebral cavernous malformation associated developmental venous anomalies in gradient-echo and susceptibility-weighted magnetic resonance imaging: can we skip the contrast?

Purpose: Hemosiderin-sensitive MRI sequences are commonly utilized for microbleed identification in cerebral cavernous malformations (CCM). Efficacy of susceptibility-weighted (SWI) and gradient-echo T2*-weighted imaging (GRE-T2*) sequences in detecting CCM-associated developmental venous anomalies (DVA) remains uncertain.

Methods: We conducted a retrospective review of our institutional CCM database. Inclusion criteria comprised baseline characteristics and MRI datasets involving contrast-enhanced T1 (CE-T1), T2, and SWI or GRE-T2* sequences. The presence of CCM-related DVA was determined utilizing CE-T1 imaging. A subgroup of 200 patients, evenly distributed with or without DVA, underwent random selection for analysis, with 50 patients each having SWI and GRE-T2* imaging. The presence of DVA was evaluated by two blinded neuroradiologists based on SWI or GRE-T2* sequences. Interrater agreement, sensitivity, and specificity values for SWI and GRE-T2* sequences were analyzed.

Results: Imaging assessment demonstrated observer agreement in 76% of the SWI sequences (K = 0.51, p <0.001) and 82% of the GRE-T2* images (K = 0.39, p <0.001). While SWI sequences exhibited a sensitivity of 81.4% and a specificity of 60.6%, GRE-T2* sequences showed a sensitivity of 19.1% and a specificity of 97.5%. Misdiagnoses were frequent in small vessel DVAs, whereas large vessel DVAs were associated with higher diagnostic accuracy in both SWI and GRE-T2* sequences (p =0.002).

Conclusion: Evaluation of SWI and GRE-T2* sequences for assessing CCM-related DVA appears less effective than the routine use of CE-T1 sequences. The use of contrast agents still appears necessary for detailed diagnostics and appropriate surgery planning.