
浏览全部资源
扫码关注微信
1.重庆市红十字会医院/江北区人民医院放射科,重庆 400020
2.重庆医科大学附属第一医院放射科,重庆 400016
3.达州市中心医院放射科,达州 635000
万 露,Email:1171036701@qq.com,研究方向:MRI成像技术。
郁 斌,Email:c.benjamin@163.com。
收稿:2025-04-13,
修回:2025-07-13,
录用:2025-08-12,
纸质出版:2026-10-28
移动端阅览
万露,张志伟,陈翼等.深度学习重建优化前列腺小视野DWI图像质量及其对表观扩散系数分布特性的影响研究[J].重庆医科大学学报,2026,51(10):1513-1520.
Wan Lu,Zhang Zhiwei,Chen Yi,et al.Value of deep learning-based reconstruction in optimizing the image quality of focus-field-of-view diffusion weighted imaging for the prostate and its influence on the distribution characteristics of apparent diffusion coefficient[J].Journal of Chongqing Medical University,2026,51(10):1513-1520.
万露,张志伟,陈翼等.深度学习重建优化前列腺小视野DWI图像质量及其对表观扩散系数分布特性的影响研究[J].重庆医科大学学报,2026,51(10):1513-1520. DOI: 10.13406/j.cnki.cyxb.003931.
Wan Lu,Zhang Zhiwei,Chen Yi,et al.Value of deep learning-based reconstruction in optimizing the image quality of focus-field-of-view diffusion weighted imaging for the prostate and its influence on the distribution characteristics of apparent diffusion coefficient[J].Journal of Chongqing Medical University,2026,51(10):1513-1520. DOI: 10.13406/j.cnki.cyxb.003931.
目的
2
探讨深度学习重建(deep learning-reconstruction,DLR)优化前列腺小视野DWI(focus field-of-view diffusion-weighted imaging,FDWI)图像质量及其对表观扩散系数(apparent diffusion coefficients,ADC)分布特性的影响研究。
方法
2
回顾性纳入60例前列腺磁共振检查的患者影像学资料,对比DLR、常规重建FDWI病变区域、正常区域图像的信噪比(signal-to-noise ratio,SNR)、对比噪声比(contrast-to-noise ratio,CNR)和ADC。让2位医生采用双盲法对2种重建FDWI序列的噪声、磁敏感伪影和整体图像质量分别使用5分法评分,对比2位医生评分一致性;将ADC值高阶统计量偏度与峰度进行统计分析,验证其提升图像质量的同时保护数据分布完整性的技术优势。
结果
2
DL-FDWI序列病变区域和正常区域的SNR:[17.49(11.13,22.54)、17.34(11.44,26.56)]及CNR:[10.86(8.06,15.70)、9.32(5.82,13.11)]均明显高于常规-FDWI序列SNR:[12.88(10.56,18.14)、14.59(9.65,19.27)]及CNR:[8.87(7.11,11.64)、7.82(5.00,11.41)],差异均具有统计学意义(
P
<
0.001)。2位医生在噪声和整体图像质量方面DL-FDWI评分均高于常规-FDWI,差异均有统计学意义(
P
<
0.001),在磁敏感伪影方面DL-FDWI序列和常规-FDWI图像评分(
P
=0.637和
P
=0.655)差异无统计学意义。2位医生噪声评分Kappa值为0.924,强相关性;磁敏感伪影和整体图像质量评分Kappa值分别为0.778,0.736,有相关性。ADC值(×10⁻³ mm²/s):病变区域:DL[0.86(0.69,1.06)] vs. 常规 [0.86(0.69,1.06)];正常区域:DL[1.52(1.40,1.71)] vs. 常规[1.52(1.39,1.71)],偏度:病变区域:DL[-0.04(-0.67,0.53)] vs. 常规[-0.06(-0.35,0.26)];正常区域:DL[-0.03(-0.24,0.30)] vs. 常规[-0.17(-0.47,0.34)],峰度:病变区域:DL[-0.52(-0.82,0.25)] vs. 常规[-0.44(-0.83,0.12)];正常区域:DL[-0.19±0.79] vs. 常规[-0.23±0.78],DL-FDWI与常规-FDWI图像在以上测量结果中差异均无统计学意义(
P
>
0.05
)。
结论
2
DLR在不改变ADC值分布特性的前提下,明显提升FDWI图像SNR与CNR,其“提质保真”特性为基于ADC值分布的前列腺癌异质性分析提供了可靠的影像学基础。
Objective
2
To investigate the value of deep learning-based reconstruction(DLR) in optimizing the image quality of focus-field-of-view diffusion weighted imaging(FDWI) of the prostate and its influence on the distribution characteristics of apparent diffusion coefficients(ADC).
Methods
2
A retrospective analysis was performed for the imaging data of 60 patients undergoing prostate MRI,and signal-to-noise ratio(SNR),contrast-to-noise ratio(CNR),and ADC of the lesions and normal regions were compared between DLR-optimized FDWI(DL-FDWI) and conventionally reconstructed(FDWI). Based on a double-blind design,two radiologists independently determined the scores of noise,susceptibility artifacts,and overall image quality of the two reconstructed FDWI sequences using a 5-point scoring system,and inter-rater agreement was assessed. Higher-order statistical metrics(skewness and kurtosis) of ADC were analyzed to validate the technical advantage of DL-FDWI in protecting data integrity while improving image quality.
Results
2
For both lesion and normal regions,the SNR[17.49(11.13,22.54) and 17.34(11.44,26.56)] and CNR[10.86(8.06,15.70) and 9.32(5.82,13.11)] of the DL-FDWI sequences were significantly higher than those of the conventional-FDWI sequences[SNR:12.88(10.56,18.14) and 14.59(9.65,19.27);CNR: 8.87(7.11,11.64) and 7.82(5.00,11.41)],with all differences being statistically significant(
P
<
0.001).Both radiologists assigned significantly higher scores to the DL-FDWI images compared to conventional-FDWI images for noise and overall image quality(
P
<
0.001 for both). However,there were no statistically significant differences in the scores for susceptibility artifacts between the DL-FDWI and conventional-FDWI images(
P
=0.637 and
P
=0.655 for the two radiologists,respectively).Inter-reader agreement,assessed by Cohen’s kappa coefficient,was almost perfect for noise scores(k=0.924). Agreement for susceptibility artifact scores and overall image quality scores was substantial,with kappa values of 0.778 and 0.736,respectively.ADC values(×10⁻³ mm²/s):Lesion region: DL[0.86(0.69,1.06)] vs conventional [0.86(0.69,1.06)],Normal region: DL[1.52(1.40,1.71)] vs conventional [1.52(1.39,1.71)],Skewness:Lesion region:DL[-0.04(-0.67,0.53)] vs conventional[-0.06(-0.35,0.26)],Normal region: DL [-0.03(-0.24,0.30)] vs conventional [-0.17(-0.47,0.34)],Kurtosis:Lesion region: DL[-0.52(-0.82,0.25)] vs conventional [-0.44(-0.83,0.12)],Normal region: DL [-0.19±0.79] vs conventional [-0.23±0.78].No statistically significant differences(
P
>
0.05) were observed in any of these ADC values,skewness,or kurtosis measurements between DL-FDWI and conventional-FDWI images.
Conclusion
2
DLR significantly enhances the SNR and CNR of FDWI images without altering the distribution characteristics of ADC,and its features of “quality enhancement and fidelity preservation” provide a reliable foundation for the ADC-based heterogeneity analysis of prostate cancer.
汪 洋 , 宦 怡 , 任 静 . 前列腺影像报告与数据系统2.1版(PI-RADS v2.1):解读与比较 [J ] . 国际医学放射学杂志 , 2019 , 42 ( 6 ): 712 - 716 .
Wang Y , Huan Y , Ren J . Interpretation of Prostate Imaging-Reporting and Data System version 2.1(PI-RADS v2.1) and comparison with version 2.0 [J ] . Int J Med Radiol , 2019 , 42 ( 6 ): 712 - 716 .
Turkbey B , Rosenkrantz AB , Haider MA , et al . Prostate imaging reporting and data system version 2.1:2019 update of prostate imaging reporting and data system version 2 [J ] . Eur Urol , 2019 , 76 ( 3 ): 340 - 351 .
Choi MH , Lee YJ , Jung SE . Tracking changes in clinical practice patterns following prebiopsy biparametric prostate MRI [J ] . Acad Radiol , 2020 , 27 ( 9 ): 1255 - 1260 .
Schiavina R , Chessa F , Borghesi M , et al . State-of-the-art imaging techniques in the management of preoperative staging and re-staging of prostate cancer [J ] . Int J Urol , 2019 , 26 ( 1 ): 18 - 30 .
Peng Y , Li Z , Tang H , et al . Comparison of reduced field-of-view diffusion-weighted imaging(DWI) and conventional DWI techniques in the assessment of rectal carcinoma at 3.0 T:Image quality and histological T staging [J ] . Magn Reson Imaging , 2018 , 47 ( 4 ): 967 - 975 .
Hu L , Wei LM , Wang SH , et al . Better lesion conspicuity translates into improved prostate cancer detection:comparison of non-parallel-transmission-zoomed-DWI with conventional-DWI [J ] . Abdom Radiol , 2021 , 46 ( 12 ): 5659 - 5668 .
Zhao WJ , Ju SY , Yang HY , et al . Improved value of multiplexed sensitivity encoding DWI with reversed polarity gradients in diagnosing prostate cancer:a comparison study with single-shot DWI and MUSE DWI [J ] . Acad Radiol , 2024 , 31 ( 3 ): 909 - 920 .
Li L , Wang L , Deng M , et al . Feasibility study of 3-T DWI of the prostate:readout-segmented versus single-shot echo-planar imaging [J ] . AJR Am J Roentgenol , 2015 , 205 ( 1 ): 70 - 76 .
Bai Y , Pei Y , Liu WV , et al . MRI:evaluating the application of FOCUS-MUSE diffusion-weighted imaging in the pancreas in comparison with FOCUS,MUSE,and single-shot DWIs [J ] . J Magn Reson Imaging , 2023 , 57 ( 4 ): 1156 - 1171 .
方 姝 , 吴梦雄 , 陈 乾 , 等 . 深度学习在肝脏屏气T2加权成像图像质量评价中的应用研究 [J ] . 磁共振成像 , 2023 , 14 ( 5 ): 31 - 35
Fang S , Wu MX , Chen Q , et al . Application Research of Deep Learning in Image Quality Assessment of Liver Breath-Hold T2-Weighted Imaging [J ] . Magnetic Resonance Imaging , 2023 , 14 ( 5 ): 31 - 35 .
Yokota Y , Takeda C , Kidoh M , et al . Effects of deep learning reconstruction technique in high-resolution non-contrast magnetic resonance coronary angiography at a 3-tesla machine [J ] . Can Assoc Radiol J , 2021 , 72 ( 1 ): 120 - 127 .
饶显锋 , 杨舒文 , 陈 静 , 等 . 基于深度学习重建超快速扫描方案用于颈椎MR检查 [J ] . 中国医学影像技术 , 2024 , 40 ( 6 ): 843 - 847 .
Rao XF , Yang SW , Chen J , et al . Application of deep learning reconstruction-based ultra-fast scanning protocol for cervical spine MR imaging [J ] . Chin J Med Imaging Technol , 2024 , 40 ( 6 ): 843 - 847 .
Qiu DF , Zhang SX , Liu Y , et al . Super-resolution reconstruction of knee magnetic resonance imaging based on deep learning [J ] . Comput Meth Programs Biomed , 2020 , 187 : 105059 .
Kim EH , Choi MH , Lee YJ , et al . Deep learning-accelerated T2-weighted imaging of the prostate:impact of further acceleration with lower spatial resolution on image quality [J ] . Eur J Radiol , 2021 , 145 : 110012 .
Kim M , Kim SH , Hong S , et al . Evaluation of extra-prostatic extension on deep learning-reconstructed high-resolution thin-slice T2-weighted images in patients with prostate cancer [J ] . Cancers(Basel) , 2024 , 16 ( 2 ): 413 .
Filli L , Ghafoor S , Kenkel D , et al . Simultaneous multi-slice readout-segmented echo planar imaging for accelerated diffusion-weighted imaging of the breast [J ] . Eur J Radiol , 2016 , 85 ( 1 ): 274 - 278 .
Shanbhogue K , Tong A , Smereka P , et al . Accelerated single-shot T2-weighted fat-suppressed(FS) MRI of the liver with deep learning–based image reconstruction:qualitative and quantitative comparison of image quality with conventional T2-weighted FS sequence [J ] . Eur Radiol , 2021 , 31 ( 11 ): 8447 - 8457 .
Liu W , Patil D , Howard DH , et al . Impact of prebiopsy magnetic resonance imaging of the prostate on cancer detection and treatment patterns [J ] . Urol Oncol Semin Orig Investig , 2019 , 37 ( 3 ): 181 - 188 .
Kidoh M , Shinoda K , Kitajima M , et al . Deep learning based noise reduction for brain MR imaging:tests on phantoms and healthy volunteers [J ] . Magn Reson Med Sci , 2020 , 19 ( 3 ): 195 - 206 .
Ueda T , Ohno Y , Yamamoto K , et al . Deep learning reconstruction of diffusion-weighted MRI improves image quality for prostatic imaging [J ] . Radiology , 2022 , 303 ( 2 ): 373 - 381 .
王绎忱 , 张馨心 , 胡满仓 , 等 . 深度学习重建技术在优化前列腺磁共振T2加权成像扫描时间和图像质量中的应用价值 [J ] . 磁共振成像 , 2023 , 14 ( 5 ): 48 - 52
Wang YC , Zhang XX , Hu MC , et al . Application Value of Deep Learning Reconstruction Technology in Optimizing Scan Time and Image Quality of Prostate MRI T2-Weighted Imaging [J ] . Magnetic Resonance Imaging , 2023 , 14 ( 5 ): 48 - 52
可 赞 , 李 亮 , 宋鑫洋 , 等 . 基于深度学习重建技术改善前列腺T2WI图像质量的研究 [J ] . 磁共振成像 , 2023 , 14 ( 05 ): 41 - 47 .
Ke Z , Li L , Song XY , et al . Research on Improving Prostate T2WI Image Quality Based on Deep Learning Reconstruction Technology [J ] . Magnetic Resonance Imaging , 2023 , 14 ( 05 ): 41 - 47 .
邓戈文 . 基于深度学习的MRI重建方法研究 [D/OL ] . 济南:山东中医药大学 ,2024-06-01[ 2025-08-09 ] .
Deng GW . Research on MRI reconstruction methods based on deep learning [D/OL ] . Jinan:Shandong Univ Tradit Chin Med , 2024-06-01 [ 2025-08-09 ] .
Hectors SJ , Semaan S , Semaan S , et al . Advanced diffusion-weighted imaging modeling for prostate cancer characterization:correlation with quantitative histopathologic tumor tissue composition-a hypothesis-generating study [J ] . Radiology , 2018 , 286 ( 3 ): 918 - 928 .
Liu Y , Wang X , Cui Y , et al . Comparative study of monoexponential,intravoxel incoherent motion,kurtosis,and IVIM-kurtosis models for the diagnosis and aggressiveness assessment of prostate cancer [J ] . Front Oncol , 2020 , 10 : 1763 .
Li C , Yu L , Jiang Y , et al . The histogram analysis of intravoxel incoherent motion-kurtosis model in the diagnosis and grading of prostate cancer-a preliminary study [J ] . Front Oncol , 2021 , 11 : 604428 .
0
浏览量
3
下载量
0
CSCD
关联资源
相关文章
相关作者
相关机构
京公网安备11010802024621