Imaging of non-traumatic intracerebral and intraventricular haemorrhage

The diagnostic yield and accuracy of CT angiography (CTA), and of additional MR imaging/angiography (MRI/MRA) and digital subtraction angiography (DSA) for detection of macrovascular causes of intracerebral haemorrhage are described. The reference standard was the best available evidence from all findings during follow-up (median duration of 450 days). We enrolled 298 patients, aged 18 to 70 years, excluding those older than 45 years of age with hypertension and ICH in basal ganglia, thalamus, or posterior fossa. A macrovascular cause was identified in 69 patients (23%). The overall yield of CTA and MRI/MRA for detection of a macrovascular cause was slightly higher than the yield of early CTA as a single modality (18 versus 17%), whereas the combination of CTA, MRI/MRA and DSA increased the overall yield to 23%. The positive predictive value (PPV) of CTA was 72%, of additional MRI/MRA 35%, and of additional DSA 100%. The standardised diagnostic workup failed to identify a cavernoma in one patient, which was identified by a repeated MRI. Predictors for a macrovascular cause were younger age, lobar or posterior fossa ICH location, and absence of signs of small vessel disease on non-contrast CT. The discriminative ability of the proposed prediction score was good (c statistic 0.83, 95%CI 0.78-0.88) for estimation of low (<5%), intermediate (5-15%) and high (>15%) risk of a macrovascular cause. Prediction charts were generated for patients aged 18-50 years old and patients aged 51-70. We conclude that CTA is an appropriate initial investigation and additional MRI/MRA may find cavernomas or alternative diagnoses, but DSA is indicated to find macrovascular causes undetected by CTA or MRI/MRA. The DIAGRAM prediction score may be helpful for clinical practice to estimate the probability of finding a macrovascular cause, and needs external validation. We also provide a case series and systematic review of the literature on the yield of angiographic examinations in patients with isolated intraventricular haemorrhage. In conclusion, the reported yield of DSA in isolated IVH varies considerably, which is probably related to confounding by indication. Data on the yield of CTA and MRI/MRA in patients with isolated IVH are urgently needed. Also a Markov decision-analytic model was developed to examine the cost-effectiveness of different diagnostic strategies to find macrovascular causes in patients with ICH. Adding MRI/MRA and DSA to CTA as single investigation in the diagnostic workup of patients with ICH does not improve health outcomes but does increase healthcare costs.