基于模型的迭代重建对低剂量肺部CT条件下计算机辅助检测系统气道分析的影响

Effect of Model-based Iterative Reconstruction on Computer-aided Detection for Quantitative Analysis of Airway Tree in Low-dose Chest CT

目的比较基于模型的迭代重建(MBIR)和自适应迭代重建(ASiR)对低剂量肺部CT条件下计算机辅助检测(CAD)气道分析性能的影响。资料与方法收集2016年7-10月于陕西中医药大学附属医院1个月内接受2次胸部CT平扫的受检者45例。初检采用常规辐射剂量扫描,肺算法ASiR重建;复查采用低辐射剂量扫描,肺算法ASiR和MBIR重建层厚0.625 mm图像。应用自动CT定量方法(Dexin-FACT)自动提取支气管树,选取并比较右肺中叶支气管长度。2名放射科医师以常规剂量ASiR重建为基础,对低剂量条件下ASiR和MBIR自动提取的支气管树结构连续性、清晰度采用盲法进行5分法主观评分。结果低剂量组有效辐射剂量降低约70%。重建算法和辐射剂量可影响CAD检测右肺中叶支气管长度。自动提取MBIR图像支气管长于常规剂量ASiR(P<0.05),但低剂量短于常规剂量ASiR(P<0.05)。重建算法可影响低剂量肺部CT条件下CAD自动提取支气管树的主观噪声、连续性、清晰度和末梢显示。低剂量MBIR优于常规剂量ASiR,但低剂量ASiR重建图像差于常规剂量ASiR(P<0.05)。结论 MBIR可通过改善图像质量提高CAD气道分析性能,进而可减少约70%的辐射剂量,为胸部CT降低辐射剂量提供了一种新的途径。

Purpose To evaluate the effect of model-based iterative reconstruction(MBIR)and adaptive statistical iterative reconstruction(ASiR)on the performance of computer-aided detection(CAD)for quantitative analysis of airway in low-dose chest CT.Materials and Methods Forty-five patients were included who were scanned for pulmonary disease follow-up using a clinical MDCT system in 1 months.The initial examination was performed at routine dose and the radiation dose level was reduced by about 30%in reexamination.Data of routine dose were reconstructed with ASiR,while those of low-dose were reconstructed with both ASiR and MBIR.Airway dimensions from three reconstruction algorithm images were measured using automated quantitative software that was designed to segment and quantify the bronchial tree,and a skeletonization algorithm was used to extract the center-line of airway trees automatically.For each patients and reconstruction algorithm,the right middle lobe bronchus was chosen as a representative.The bronchial length of the matched airways was measured by this scheme.Two radiologists used a semiquantitative 5 scale to subjectively assess image quality of airway trees about low-dose images reconstructed with ASiR and MBIR.Results Median effective radiation dose in reduced dose group was decreased by 70%than that in routine dose group.Reconstruction algorithms and radiation doses could influence CAD to detect the length of the right middle lobe bronchus.The auto-extracted bronchus in MBIR images were longer than routine-dose ASiR(P<0.05),but they were shorter in low dose than routine-dose ASiR(P<0.05).The reconstruction algorithm could influence the subjective noise,continuity,sharpness,and distal display of the auto-extracted bronchial tree from the CAD in low-dose chest CT.Low-dose MBIR was superior to routine-dose ASiR,but low-dose ASiR reconstruction images were inferior to routine-dose ASiR(P<0.05).Conclusion The quantification accuracy of airway is strongly influenced by reconstruction algorithm.MBIR can improve the performance of CAD airway analysis by improving image quality,which can reduce radiation dose by about 70%.This provides a new way to reduce radiation dose in chest CT.