基于光子计数CT的脑动静脉畸形低剂量四维CT血管成像方案的辐射剂量与图像质量评估OA

Objective To investigate the feasibility of reducing radiation dose in brain arteriovenous malformation(AVM)imaging using a photon-counting CT based four-dimensional CT angiography(4D-CTA)protocol,and to evaluate the diagnostic image quality under the low-dose protocol.Methods A total of 16 patients with brain AVM admitted to the Department of Neurosurgery,Peking Union Medical College Hospital between May 2024 and November 2025 for planned interventional embolization,and required preoperative dynamic 4D-CTA for further observation of vascular structures,were prospectively and consecutively enrolled.General patient data were collected,including sex,age,and clinical symptoms(seizures,focal neurological deficits[motor impairment,sensory impairment,aphasia,etc.],non-specific symptoms[headache,vomiting,etc.],or no obvious symptoms).Imaging data of brain AVM vascular structures on preoperative DSA included the location of the AVM nidus(supratentorial or infratentorial),maximum diameter,source of feeding arteries(anterior circulation,posterior circulation,anterior and posterior circulation),and type of venous drainage(superficial or deep venous drainage).All patients underwent 4D-CTA examinations on a photon-counting CT,and three sets of perfusion parameter maps(cerebral blood flow[CBF],cerebral blood volume[CBV],and mean transit time[MTT])were reconstructed.The first 8 patients underwent the standard protocol(the preset scanning protocol of the photon-counting CT system),while the subsequent 8 patients underwent the low-dose protocol(the combined applications to reduce exposure[CARE]keV image quality target was set to the"vascular optimized"mode,dynamic scanning was initiated with a 4-second delay after contrast injection,and the temporal sampling interval was extended to 1.5 seconds,reducing the total acquisition time to 41.9 seconds and the number of dynamic time points to 21,primarily covering the critical time window from the arterial phase to venous drainage).Two radiologists independently evaluated the image quality of both protocols,including the signal-to-noise ratio(SNR)and contrast-to-noise ratio(CNR)of 4D-CTA and 4D-CTA-derived perfusion parameter maps(CBF,CBV,MTT),and compared them with the brain AVM vascular structures shown on DSA.A 4-point scale was used to assess the subjective image quality of the CBF,CBV,and MTT perfusion maps.Inter-observer agreement between the two radiologists for image quality assessment was tested using intraclass correlation coefficient(ICC)or Cohen's Kappa analysis.Results(1)A total of 16 patients with brain AVM were included(9 males,7 females;age range 9-65 years,mean age[38±15]years),with 8 in the standard protocol group and 8 in the low-dose protocol group.There were no statistically significant differences between the standard protocol and low-dose protocol groups in sex,age,clinical symptoms,AVM location,maximum diameter,source of feeding arteries,and type of venous drainage(all P＞0.05).(2)The tube current in the standard protocol group ranged from 140 mA to 190mA,with a median of 168(156,173)mA,while that in the low-dose group ranged from 7 mA to 48 mA,with a median of 42(41,44)mA;the difference was statistically significant(P=0.002).The radiation dose in the standard protocol group ranged from 4.8 mSv to 9.2 mSv,with a median of 8.4(7.0,8.9)mSv,whereas that in the low-dose protocol group ranged from 0.3 mSv to 1.8 mSv,with a median of 1.7(1.5,1.8)mSv,showing a statistically significant difference(P＜0.01).(3)The two radiologists reached substantial or almost perfect agreement in classifying AVM location,source of feeding arteries,and type of venous drainage(inter-observer agreement rates:16/16,14/16,and 15/16,respectively;Kappa values:1.000,0.667,and0.871,respectively),and showed excellent agreement in measuring the maximum diameter of AVM(ICC=0.994).The measurements of SNR and CNR for 4D-CTA and perfusion parameter maps(CBF,CBV,MTT)showed good or excellent agreement(ICC ranging from 0.762 to 0.972).The subjective image quality scores for CBV and CBF perfusion maps demonstrated almost perfect and substantial agreement,respectively(Kappa values:0.828 and 0.700),while the subjective scores for MTT perfusion maps showed moderate agreement(Kappa=0.478).(4)There were no statistically significant differences in the SNR and CNR of 4D-CTA images and CBF,CBV,and MTT perfusion maps between the standard and low-dose protocol groups(all P＞0.05).The 4D-CTA images from both protocols clearly displayed structural features such as AVM location,maximum diameter,source of feeding arteries,and type of venous drainage,which were consistent with the DSA findings.The subjective image quality score of the 4D-CTA-based MTT perfusion maps in the low-dose protocol group was lower than that in the standard protocol group(3.0[2.0,3.0]vs.3.5[3.0,4.0]),with a statistically significant difference(P=0.045).There were no statistically significant differences in the subjective scores of CBF and CBV perfusion maps between the two groups(both P＞0.05).Conclusions The photon-counting CT based low-dose 4D-CTA protocol can provide anatomical and perfusion parameter images of brain AVM with quality comparable to that of the standard protocol.The findings of this study warrant further validation in larger cohorts.