-
基于CT和内置黄金基准标志物(gold fiducial marker,GFM)的前列腺癌图像引导放疗(image-guided radiotherapy,IGRT)已被证明显著提高了前列腺癌的治疗精准度和效果[1-2]。但是,由于CT图像的软组织分辨率较低,在一定程度上影响了靶区勾画的准确度;而MRI图像凭借其优良的软组织分辨率已成为取代CT的最佳选择,相关研究也已成为热点[3-5]。
目前,关于MRI在前列腺癌IGRT中的应用研究主要集中在CT与MRI图像的融合[5],以及利用MRI图像生成合成CT图像[6-7]。基于MRI图像的GFM辨识和定位的研究极少,仅有的研究主要集中在基于MRI图像的GFM自动辨识和定位。尽管在理想状态下GFM自动辨识和定位具有较高的阳性率(84%~96%)和准确度,但在临床实践中,钙化和出血点的存在对其造成了较大的影响[8-12]。而手动辨识和定位在CT图像的辅助下受钙化和出血点的影响较小。目前尚无关于单独基于MRI图像的GFM手动辨识和定位的相关研究。因此,本研究评价和量化了单独基于2种MRI序列法图像的GFM手动辨识和定位的可行性及其临床价值,并进一步优化基于MRI的前列腺癌IGRT。
-
纳入2019年6月至2021年6月于唐山市人民医院接受治疗的16例前列腺癌患者进行前瞻性研究,年龄(58.5±4.1)岁,其中,7例Gleason评分[13]≤6分、4例7分、5例≥8分。纳入标准:(1)接受外照射放疗;(2)身体质量指数为18~25 kg/m2。排除标准:(1)前列腺严重钙化或出血;(2)严重直肠息肉或痔疮。本研究获得了唐山市人民医院医学伦理委员会批准(批号:20190129),所有患者均于治疗前签署了知情同意书。
-
所有患者于放疗前均排空膀胱及直肠,应用荷兰Philips 公司iU22彩色多普勒超声诊断仪(端扫式9-4 V直肠探头,探头频率为2~10 MHz),经直肠超声引导术(对直肠肛管进行局麻并建立静脉通道,探头插入直肠并抵近前列腺)[14]和细针穿刺抽吸术(采用5 ml注射器和18 G针头随超声探头进入直肠,经过直前壁穿刺入前列腺体)将3枚比利时IBA Visicoil helical线性GFM(高5.0 mm,底面直径1.0 mm)置入靶区不同层面,且保持间距≥15 mm。GFM编号规则:沿头脚方向依次编号为1、2、3,若在同一层面则沿左右方向再次编号。
-
对GFM位置稳定且身体置于定位状态下真空塑形垫上的患者,在统一定位条件下,应用荷兰Philips 公司 Big Bore大孔径CT采集盆腔轴位CT平扫图像,扫描参数:管电压120 kV、管电流300 mA、层厚3 mm、采集矩阵512×512,分辨率0.98 mm×0.98 mm。CT图像用于标志物间距(inter-marker distance,IMD)计算参考和辅助GFM辨识。
-
对GFM位置稳定且身体置于定位状态下真空塑形垫上的患者,在统一定位条件下,应用美国GE 公司 SIGNA Voyager 1.5T MRI并通过3种序列采集盆腔轴位图像。(1)平衡稳态自由进动序列(balanced steady-state free precession,bSSFP):该序列最先获取;(2)T1加权双相位扰相梯度回波序列(spoiled gradient-recalled echo,SPGR):该序列在bSSFP后立即获取(最大时差≤5 min);(3)双相位梯度回波序列(gradient-recalled echo,GRE):该序列在检查结束前获取(于SPGR后获取,最小时差≥15 min)。3种序列的详细参数见表1。
MRI序列 TE1/TE2/TR(ms) 翻转角度(°) 视野(mm×mm ×mm) 采集矩阵 重建矩阵 重建体素
(mm×mm ×mm)带宽
(Hz)采集时间
(min)bSSFP 1.98/−/3.96 40 250×250×90 252×234×90 512×512×90 0.5×0.5×1.0 945 4.5 SPGR 1.40/2.70/4.40 10 467×467×300 312×314×200 320×320×200 1×1×1 1078 4.0 GRE 1.40/2.70/4.60 10 449×449×90 376×376×75 400×400×75 1.1×1.1×1.2 1142 2.0 注:bSSFP为平衡稳态自由进动序列;SPGR为扰相梯度回波序列;GRE为梯度回波序列;MRI为磁共振成像;TE为回波时间;TR为重复时间;−表示无此项数据 表 1 bSSFP、SPGR和GRE 3种MRI序列的成像参数
Table 1. Imaging parameters of three MRI scanning sequences: balanced steady-state free precession, spoiled gradient-recalled echo and gradient-recalled echo
-
由5名具有5年以上工作经验的放疗医师对每例患者bSSFP序列同一层面的轴位图像进行GFM手动辨识和定位,即单独序列(single sequence,SS)法。(1)单个GFM中心(CsGFM):排除假阳性辨识的GFM,计算每枚正确辨识的GFM坐标均值,获取其CsGFM坐标,再计算5名放疗医师获取的该枚GFM的CsGFM坐标均值。应用SPSS 19.0软件中Bland-Altman分析法评价5名放疗医师单独计算的GFM中心坐标同所有GFM的CsGFM>坐标均值在左右、头脚、腹背3个方向上的95%一致性界限(limit of agreement,LoA)。(2)3枚GFM空间质心(CmGFM):利用CsGFM坐标计算3枚GFM CmGFM坐标,再计算5名放疗医师获取的该例患者的CmGFM坐标均值。采用Bland-Altman分析法评价5名放疗医师单独计算的3枚GFM空间质心坐标同所有患者的CmGFM坐标均值在左右、头脚、腹背3个方向上的95%LoA,计算95%LoA最大范围,即上界与下界差值的绝对值。
分别使用SPGR和GRE序列同一层面的轴位图像进行基于bSSFP序列图像的GFM手动辨识和定位,即组合序列(combined sequence,CS)法。95%LoA最大范围的计算方法与上述SS法相同。
-
排除假阳性辨识的GFM,利用CsGFM坐标均值和同一坐标系下CT图像上GFM中心坐标,通过下述公式分别计算SS法和CS法图像与CT图像GFM的IMD。
$ {\rm{IMD}} = \sqrt{{({{X}}_{\mathrm{M}\mathrm{R}} - {{X}}_{\mathrm{C}\mathrm{T}})}^{2} + {({{Y}}_{\mathrm{M}\mathrm{R}} - {{Y}}_{\mathrm{C}\mathrm{T}})}^{2} + {({{Z}}_{\mathrm{M}\mathrm{R}} - {{Z}}_{\mathrm{C}\mathrm{T}})}^{2}} \quad $ 式中,XMR、YMR、ZMR分别为MRI图像上GFM中心左右、头脚、腹背方向坐标,XCT、YCT、ZCT分别为CT图像上GFM中心左右、头脚、腹背方向坐标。应用SPSS 19.0软件对2种序列法与CT定位的比较结果进行统计学分析。符合正态分布的计量资料以
$\bar x \pm s $ 表示,组间比较采用配对样本t检验(方差齐)。P<0.05为差异有统计学意义。
基于MRI的前列腺癌放疗中标志物手动辨识与定位评价
Manual identification and localization evaluation of marker in prostate cancer radiotherapy based on magnetic resonance imaging
-
摘要:
目的 评价基于MRI的前列腺癌放疗中黄金基准标志物(GFM)手动辨识和定位方法的效能。 方法 纳入2019年6月至2021年6月于唐山市人民医院接受治疗的16例前列腺癌患者进行前瞻性研究,年龄(58.5±4.1)岁,每例患者均置入了3枚GFM后接受放疗定位CT扫描,然后采用单独序列(SS)法和组合序列(CS)法对所有患者进行MRI检查。由5名放疗医师分别对2种序列法获得的所有图像进行GFM手动辨识和定位。计算单个GFM中心(CsGFM)坐标均值和3枚GFM空间质心(CmGFM)坐标均值,并采用Bland-Altman分析法分别评价2种序列法定位的一致性。通过与CT定位比较并计算GFM的标志物间距(IMD),评价2种序列法定位的准确度,组间比较采用配对样本t检验。 结果 SS法GFM辨识阳性率为79.17%(38/48),CS法为93.75%(45/48)。(1)一致性结果。SS法中,每名放疗医师计算的GFM中心坐标同所有GFM的CsGFM坐标均值在左右、头脚、腹背3个方向上的95%一致性界限(LoA)范围分别为−1.46~0.97 mm、−1.06~1.73 mm、−1.96~1.12 mm;CS法中分别为−0.79~1.09 mm、0.10~1.47 mm、−0.87~1.40 mm。SS法中,每名放疗医师计算的3枚GFM空间质心坐标同所有患者的CmGFM坐标均值在左右、头脚、腹背3个方向上的95%LoA范围分别为−1.38~0.94 mm、−1.60~1.07 mm、−1.07~1.75 mm;CS法中分别为−0.57~0.76 mm、−0.71~0.98 mm、−1.16~0.76 mm。(2)准确度结果。与CT定位比较,SS法中IMD为(0.59±0.39) mm,显著大于CS法中的(0.32±0.17) mm,且差异有统计学意义(t=−1.89,P=0.027)。 结论 基于MRI的2种GFM手动辨识和定位方法的效能均可满足临床要求,其中CS法GFM手动辨识的阳性率、定位一致性和准确度均优于SS法。 Abstract:Objective To evaluate the effectiveness of manual identification and localization methods of gold fiducial marker (GFM) in prostate cancer radiotherapy based on MRI. Methods Sixteen patients with prostate cancer who were treated in Tangshan People's Hospital from June 2019 to June 2021 were included in this prospective study. The age of patients was (58.5±4.1) years. All the patients received radiotherapy localization CT scan after the implantation of three GFMs. Single-sequence (SS) and combined sequence (CS) method were used for MRI scanning for all the patients. All GFMs on the images obtained by the two sequencing methods were manually identified and located by five radiologists. The average coordinates of a single GFM center (CsGFM) and three GFM space centroids (CmGFM) were calculated. The consistency of the two sequencing methods was evaluated using Bland-Altman method. Sequencing accuracy was determined by comparison with CT localization and calculation of the imter-marker distance (IMD) of GFM. The difference between groups was analyzed by paired sample t test. Results The positive rate of GFM identification by SS method was 79.17% (38/48), and that of CS method was 93.75% (45/48). (1) In SS method, the range of 95% limit of agrecment (LoA) between the GFM central coordinate calculated by each radiologist and the mean value of CsGFM coordinates in left-right, superior-inferior, and anterior-posterior directions were −1.46–0.97, −1.06–1.73, and −1.96–1.12 mm, respectively; in CS method, the values were −0.79–1.09, 0.10–1.47, and −0.87–1.40 mm, respectively. In SS method, the 95%LoA of the three GFM space centroid coordinates calculated by each radiologist and the mean value of CmGFM coordinates in left-right, superior-inferior, and anterior-posterior directions were −1.38–0.94, −1.60–1.07, and −1.07–1.75 mm, respectively; in CS method, the values were −0.57–0.76, −0.71–0.98, and −1.16–0.76 mm, respectively. (2) For accuracy, compared with CT localization, the IMD in SS method was (0.59±0.39) mm, which was significantly higher than that in CS method (0.32±0.17) mm. The difference was statistically significant (t=−1.89, P=0.027). Conclusion The effectiveness of two manual identification and localization methods of GFM based on MRI can meet the clinical requirement. The positive rate, localization consistency and accuracy of CS method are superior to those of SS method. -
Key words:
- Prostate neoplasms /
- Magnetic resonance imaging /
- Radiotherapy /
- Fiducial markers
-
表 1 bSSFP、SPGR和GRE 3种MRI序列的成像参数
Table 1. Imaging parameters of three MRI scanning sequences: balanced steady-state free precession, spoiled gradient-recalled echo and gradient-recalled echo
MRI序列 TE1/TE2/TR(ms) 翻转角度(°) 视野(mm×mm ×mm) 采集矩阵 重建矩阵 重建体素
(mm×mm ×mm)带宽
(Hz)采集时间
(min)bSSFP 1.98/−/3.96 40 250×250×90 252×234×90 512×512×90 0.5×0.5×1.0 945 4.5 SPGR 1.40/2.70/4.40 10 467×467×300 312×314×200 320×320×200 1×1×1 1078 4.0 GRE 1.40/2.70/4.60 10 449×449×90 376×376×75 400×400×75 1.1×1.1×1.2 1142 2.0 注:bSSFP为平衡稳态自由进动序列;SPGR为扰相梯度回波序列;GRE为梯度回波序列;MRI为磁共振成像;TE为回波时间;TR为重复时间;−表示无此项数据 -
[1] 田龙, 闫洁诚, 胡逸民, 等. 前列腺癌容积旋转调强放疗中标志物可探测性研究[J]. 国际放射医学核医学杂志, 2021, 45(12): 767−772. DOI: 10.3760/cma.j.cn121381-202102024-00123.
Tian L, Yan JC, Hu YM, et al. Research on detectability of markers in volume-modulated arc therapy for prostate cancer[J]. Int J Radiat Med Nucl Med, 2021, 45(12): 767−772. DOI: 10.3760/cma.j.cn121381-202102024-00123.[2] 田龙, 闫洁诚, 李明辉, 等. 利用多次采集计划CT和锥形束CT评价前列腺癌靶区运动相关性[J]. 中国医学物理学杂志, 2021, 38(2): 172−177. DOI: 10.3969/j.issn.1005-202X.2021.02.009.
Tian L, Yan JC, Li MH, et al. Evaluating the correlation between pre-and in-treatment target displacements in prostate cancer radiotherapy by multiple acquisition planning CT and cone-beam CT[J]. Chin J Med Phys, 2021, 38(2): 172−177. DOI: 10.3969/j.issn.1005-202X.2021.02.009.[3] 姜玲, 周海中. 前列腺特异性膜抗原核素显像及治疗在分化型甲状腺癌中的应用进展[J]. 国际放射医学核医学杂志, 2022, 46(8): 497−501. DOI: 10.3760/cma.j.cn121381-202109013-00205.
Jiang L, Zhou HZ. Application progress of PSMA in the imaging and treatment of differentiated thyroid cancer[J]. Int J Radiat Med Nucl Med, 2022, 46(8): 497−501. DOI: 10.3760/cma.j.cn121381-202109013-00205.[4] 高金龙, 王海峰, 李娜, 等. 自MR-T2图像勾画肝细胞癌患者放疗靶区精度研究[J]. 实用肝脏病杂志, 2021, 24(4): 565−568. DOI: 10.3969/j.issn.1672-5069.2021.04.028.
Gao JL, Wang HF, Li N, et al. Target delineation accuracy of tumors for radiotherapy based on MR-T2 image in patients with hepatocellular carcinoma[J]. J Pract Hepatol, 2021, 24(4): 565−568. DOI: 10.3969/j.issn.1672-5069.2021.04.028.[5] 柏正璐, 李军, 田书畅. CT-MR图像融合在脑胶质瘤术后放疗中的应用[J]. 中国医疗设备, 2020, 35(12): 20−23. DOI: 10.3969/j.issn.1674-1633.2020.12.006.
Bai ZL, Li J, Tian SC. Application of CT-MR image fusion in postoperative radiotherapy for gliomas[J]. China Med Devices, 2020, 35(12): 20−23. DOI: 10.3969/j.issn.1674-1633.2020.12.006.[6] Persson E, Gustafsson C, Nordström F, et al. MR-OPERA: a multicenter/multivendor validation of magnetic resonance imaging-only prostate treatment planning using synthetic computed tomography images[J]. Int J Radiat Oncol Biol Phys, 2017, 99(3): 692−700. DOI: 10.1016/j.ijrobp.2017.06.006. [7] Tyagi N, Fontenla S, Zhang J, et al. Dosimetric and workflow evaluation of first commercial synthetic CT software for clinical use in pelvis[J]. Phys Med Biol, 2017, 62(8): 2961−2975. DOI: 10.1088/1361-6560/aa5452. [8] Ghose S, Mitra J, Rivest-Hénault D, et al. MRI-alone radiation therapy planning for prostate cancer: Automatic fiducial marker detection[J]. Med Phys, 2016, 43(5): 2218−2228. DOI: 10.1118/1.4944871. [9] Fernandes CD, Dinh CV, Steggerda MJ, et al. Prostate fiducial marker detection with the use of multi-parametric magnetic resonance imaging[J]. Phys Imag Radiat Oncol, 2017, 1: 14−20. DOI: 10.1016/j.phro.2017.02.001. [10] Sanders JW, Venkatesan AM, Levitt CA, et al. Fully balanced SSFP without an endorectal coil for postimplant QA of MRI-Assisted Radiosurgery (MARS) of prostate cancer: a prospective study[J]. Int J Radiat Oncol Biol Phys, 2021, 109(2): 614−625. DOI: 10.1016/j.ijrobp.2020.09.040. [11] Gustafsson C, Korhonen J, Persson E, et al. Registration free automatic identification of gold fiducial markers in MRI target delineation images for prostate radiotherapy[J]. Med Phys, 2017, 44(11): 5563−5574. DOI: 10.13140/RG.2.2.22225.43367. [12] Maspero M, Van Den Berg CAT, Zijlstra F, et al. Evaluation of an automatic MR-based gold fiducial marker localisation method for MR-only prostate radiotherapy[J]. Phys Med Biol, 2017, 62(20): 7981−8002. DOI: 10.1088/1361-6560/aa875f. [13] 苑克慧, 胡蓉蓉, 姚树林, 等. 巨大多房前列腺囊腺瘤18F-FDG PET/CT显像一例[J]. 国际放射医学核医学杂志, 2021, 45(5): 338−341. DOI: 10.3760/cma.j.cn121381-202006034-00059.
Yuan KH, Hu RR, Yao SL, et al. 18F-FDG PET/CT imaging of giant multilocular prostatic cystadenoma: a case report[J]. Int J Radiat Med Nucl Med, 2021, 45(5): 338−341. DOI: 10.3760/cma.j.cn121381-202006034-00059.[14] 田龙, 范学武, 许蕊, 等. 对比T2WI与T2*WI用于前列腺癌放射治疗计划的效果[J]. 中国介入影像与治疗学, 2022, 19(4): 239−244. DOI: 10.13929/j.issn.1672-8475.2022.04.011.
Tian L, Fan XW, Xu R, et al. Comparison on T2WI and T2*WI for radiotherapy planning of prostate cancer[J]. Chin J Interv Imaging Ther, 2022, 19(4): 239−244. DOI: 10.13929/j.issn.1672-8475.2022.04.011.[15] 牛向欣, 赵明娟, 田龙, 等. C臂X线辅助前列腺癌内置金标志物辨识的可行性研究[J]. 中国医疗设备, 2022, 37(6): 85−88. DOI: 10.3969/j.issn.1674-1633.2022.06.020.
Niu XX, Zhao MJ, Tian L, et al. Feasibility study of C-arm X-ray-assisted identification of implanted gold fiducial marker in prostate cancer[J]. China Med Devices, 2022, 37(6): 85−88. DOI: 10.3969/j.issn.1674-1633.2022.06.020.[16] 梁晓秋, 曹凌玲, 陈溢旭. 经直肠超声造影引导前列腺穿刺活检诊断前列腺癌[J]. 中国介入影像与治疗学, 2020, 17(2): 93−97. DOI: 10.13929/j.issn.1672-8475.2020.02.008.
Liang XQ, Cao LL, Chen YX. Contrast-enhanced transrectal ultrasonography in guiding prostate biopsy for diagnosis of prostate cancer[J]. Chin J Interv Imaging Ther, 2020, 17(2): 93−97. DOI: 10.13929/j.issn.1672-8475.2020.02.008.[17] Nyholm T, Nyberg M, Karlsson MG, et al. Systematisation of spatial uncertainties for comparison between a MR and a CT-based radiotherapy workflow for prostate treatments[J/OL]. Radiat Oncol, 2009, 4: 54[2021-11-01]. https://doi.org/10.1186/1748-717x-4-54. DOI: 10.1186/1748-717X-4-54. [18] Kassim I, Joosten H, Barnhoorn JC, et al. Implications of artefacts reduction in the planning CT originating from implanted fiducial markers[J]. Med Dosim, 2011, 36(2): 119−125. DOI: 10.1016/j.meddos.2010.02.002.