Our research found that the changes in HU value of ASPECTS were related to the clinical prognosis of stroke patients with MCAO treated by IVT. HU difference was the independent risk factor for prognosis of stroke patients with MCAO treated by IVT. The changes in HU value can be used to predict the clinical prognosis of stroke patients with MCAO treated by IVT. HU value measurement can help clinicians make quick decisions and select potential beneficiaries of IVT.
Our research found that the HU ratio of the good prognosis group was higher than that of the poor prognosis group, and the HU difference was lower than that of the poor prognosis group. This may be because the decrease in HU value is significantly related to the increase in water content of ischemic brain tissue. It was reported that increasing the water content of the cerebral hemisphere by 1% will cause the HU value to decrease by 1.8 HU. Water uptake content of infarcted brain tissue was related to the degree of infarction[8, 9]. Our research found the HU ratio was negatively correlated with the infarct volume, and the HU difference was positively correlated with the infarct volume. Smaller HU ratio and larger HU difference were related to larger infarct volume. HU value is also related to the state of cerebral blood flow. Kucinski et al[10] reported that within 4.5 hours after the onset of stroke, the decrease in HU value was related to the decrease in cerebral blood flow and volume. In the hyperacute phase of stroke, the HU value changes very slightly. Compared with CTP and MRI, NCCT is difficult to accurately delineate the infarct range and measure the HU value. However, for stroke centers without advanced imaging techniques such as CTP and MRI, NCCT is still the preferred imaging examination before thrombolytic therapy. In order to overcome this difficulty, we selected 10 ASPECT regions to measure HU values in stroke patients with MCAO. For the hyperacute phase of stroke, NCCT can detect brain tissue swelling and gray matter lesions in time. NCCT has the advantages of fast imaging speed, low price and wide popularity. NCCT does not require contrast injection and will not cause adverse reactions. Our study also found that early image signs of NCCT (hyperdense middle cerebral artery sign, lenticular nucleus obscuration) were associated with the prognosis of thrombolytic patients, which is consistent with previous studies[11-13].
The ASPECTS is a scoring system for evaluating early ischemic changes in the blood supply area of the middle cerebral artery by NCCT. It is essentially based on visually low attenuation. In addition, the ASPECTS is time-dependent, and a study shown that the reliability of ASPECTS is moderate at stroke onset time<90 min, good at 90-180 min, and excellent at >180 min. The HU value is based on the attenuation coefficient of water to calculate the HU value of each tissue[14], which is quantitative in nature, and the measurement of HU value is independent of the subjective interpretation of any observer. Compared with the ASPECTS, the reliability and consistency of the HU value are greatly improved. In this study, two patients had the same ASPECTS, but their HU values changed differently, and their clinical outcomes were different (see Fig.3, Fig.4). The ROC curve showed that the AUC of the HU ratio was 0.743, when the cutoff value was 10, it had a higher sensitivity of 80.56%. The AUC of the HU difference was 0.833, when the cutoff value was 13.97, it had a higher specificity of 93.75%. The results indicated that the HU ratio and HU difference had important predictive value for the prognosis of stroke patients treated by IVT. For stroke patients with the same ASPECTS, the HU value can be used for a second evaluation to improve the accuracy of diagnosis. In the future, more research is needed to perfect this algorithm to improve the predictive value of HU. Our study has some limitations: ①The influence from the occlusion position on the patient was not taken into account, for stroke patients with a distant occlusion position, the infarct size was too small, and the HU value did not change significantly. ②Previous study shown that the weight of each area of ASPECTS was different[15]. We did not weight each area when caculated the HU ratio or HU difference. ③The sample size was small. ④No subgroup analysis based on ASPECTS. ⑤Ten ASPECTS regions were manually outlined and maybe there is an deviation.