深度学习算法在前列腺TSE-T_(2)WI序列中的应用价值  

Application value of deep learning algorithm in prostate TSE-T_(2)WI sequence

在线阅读下载全文

作  者:武鹏飞[1] 张久楼 荣凡令[1] 武晨江[1] 周丹[2] 季丽雅 WU Pengfei;ZHANG Jiulou;RONG Fanling;WU Chenjiang;ZHOU Dan;JI Liya(Department of Radiology,First Affiliated Hospital of Nanjing Medical University,Nanjing 210029,China;Department of Radiology,BenQ Hospital Affiliated to Nanjing Medical University,Nanjing 210019,China)

机构地区:[1]南京医科大学第一附属医院放射科,江苏南京210029 [2]南京医科大学附属明基医院放射科,江苏南京210019

出  处:《中国中西医结合影像学杂志》2024年第6期715-718,共4页Chinese Imaging Journal of Integrated Traditional and Western Medicine

基  金:国家自然科学基金项目(62101278)。

摘  要:目的:探讨深度学习(DL)算法在改善前列腺快速自旋回波(TSE)T_(2)WI序列图像质量和加快扫描速度中的应用价值。方法:前瞻性收集54例前列腺疾病患者,均行前列腺常规T_(2)WI序列和联合DL的T_(2)WI(T_(2)WI_(DL))序列扫描。对2组图像质量进行定性、定量评估,定性评估包括:伪影、前列腺包膜清晰度、病灶的细节清晰度和整体图像质量;定量评估包括:以前列腺病灶和髂腰肌为ROI,测量并比较2组图像病灶SNR,以及病灶与髂腰肌的CNR。结果:T_(2)WI_(DL)序列在伪影、前列腺包膜清晰度、病灶的细节清晰度和整体图像质量的评分均优于常规T_(2)WI序列(均P<0.001)。T_(2)WI_(DL)序列图像的SNR_(病灶)和CNR_(病灶/肌肉)均优于常规T_(2)WI序列(均P<0.05)。结论:与常规T2WI序列相比,T_(2)WI_(DL)序列可提高图像质量,且扫描时间减少了96 s(约47%),在前列腺疾病的临床应用中有较大潜力。Objective:To explore the application value of deep learning(DL)algorithm in improving the image quality of the prostate TSE-T_(2)WI sequence and reducing scanning time.Methods:54 patients with prostate lesions were prospectively collected for prostate conventional T_(2)WI and DL-assisted T_(2)WI(T_(2)WI_(DL))scans.The image quality of two groups was evaluated qualitatively and quantitatively.The artifact,the clarity of the prostate envelope,detail clarity of the lesion and overall image quality were qualitatively assessed.The prostate lesions and iliopsoas were chosen as ROIs,the lesion SNR and CNR between lesions and iliopsoas of the 2 groups were measured and compared.Results:T_(2)WI_(DL)was superior to conventional T_(2)WI in the scores of artifact,clarity of the prostate envelope,detail clarity of the lesion and overall image quality(all P<0.001).All the lesion SNR and CNR of T_(2)WI_(DL)sequence were better than those of conventional T_(2)WI sequence(both P<0.05).Conclusions:Compared with conventional T_(2)WI sequence,T_(2)WI_(DL)sequence can improve image quality and reduce scanning time by 96 s,which has great potential for clinical application in prostate diseases.

关 键 词:前列腺 深度学习 磁共振成像 快速自旋回波 

分 类 号:R737.25[医药卫生—肿瘤] R445.2[医药卫生—临床医学]

 

参考文献:

正在载入数据...

 

二级参考文献:

正在载入数据...

 

耦合文献:

正在载入数据...

 

引证文献:

正在载入数据...

 

二级引证文献:

正在载入数据...

 

同被引文献:

正在载入数据...

 

相关期刊文献:

正在载入数据...

相关的主题
相关的作者对象
相关的机构对象