Article Text
Abstract
Introduction To study the predictive value of the ‘spot sign’ in predicting spontaneous intracerebral haemorrhage (ICH) expansion.
Method PubMed, Medline, conference proceedings and manuscript references up to June 2017 were searched for studies reporting “computed tomography angiography” and “spot sign” or “intracerebral haemorrhage” and “spot sign”. The positive (PLR) and negative (NLR) likelihood ratios were calculated using MetaDisc. The PLR needs to be ≥5 and NLR ≤0.1 for a test to be clinically useful.
Results There were 36 studies describing 6888 patients. This current analysis includes 23 studies that were not used in previous meta-analyses. The frequency of haematoma growth was 21%. The PLR and NLR for the CTA spot sign indicating haematoma growth were 4.25 (95%CI 3.52–5.13) and 0.48 (95%CI 0.41–0.56) respectively. The sensitivity analyses show a plateau in PLR and NLR of 4.3 and 0.48. The heterogeneity chi square was 214.36–250.24 (p<0.001) and inconsistency I2 was 83.7%–86%.
Conclusion The PLR for the CTA spot sign did not reach the desired threshold of 5 to be clinically useful. The heterogeneity around this estimate adds to uncertainty. The high NLR suggest that the absence of the CTA spot sign does not guarantee a lack of ICH growth.