Patch-based generative adversarial neural network models for head and neck MR-only planning
Patch-based generative adversarial neural network models for head and neck MR-only planning
复制标题
DOI：
10.1002/mp.13927
复制
发表时间：
2019-12-25
期刊：
影响因子：
3.8
通讯作者：
Tyagi, Neelam
中科院分区：
文献类型：
--
作者：
Klages, Peter;Benslimane, Ilyes;Tyagi, Neelam
目的评价pix 2 pix和CycleGAN，并评估多种组合策略对头颈部（HN）癌症患者磁共振（MR）治疗计划中基于贴片的合成计算机断层扫描（sCT）生成准确性的影响。材料和方法回顾性分析了在我们机构接受治疗的HN癌症患者的23对变形配准的CT和mDixon FFE MR数据集，以通过pix 2 pix和CycleGAN模型评估基于补丁的sCT准确性。为了测试重叠sCT补丁对估计的影响，我们（a）训练三个正交视图的模型以观察空间背景的影响，（B）我们通过使用每历元数据增强来增加有效集大小，以及（c）我们评估了三种不同方法的性能，用于组合不同补丁重叠参数的重叠Hounsfield单位（HU）估计。23例病例中的12例对应于先前用于基于图谱的sCT生成的策划数据集，并用于采用留二交叉验证进行训练。8个病例用于独立测试，包括以前未见过的图像特征，如融合的椎骨、小的突出骨和大到足以使正常身体轮廓变形的肿瘤。我们分析了MR图像预处理（包括直方图标准化和强度裁剪）对sCT生成精度的影响。在另外三个病例中测试了mDixon对比度（同相vs水）差异的影响。使用计划CT和sCT图像之间的平均绝对误差（MAE）和平均误差（ME）（HU）评价sCT生成准确度。评价了独立测试集中所有临床相关结构的剂量准确度，并评价了与计划CT图像相关的数字重建X线片（DRR）。结果使用pix 2 pix和CycleGAN的整个HN区域的交叉验证MAE分别为66.9 +/- 7.3 vs 82.3 +/- 6.4 HU。在具有额外伪影和先前不可见图像特征的独立测试集上，pix 2 pix和CycleGAN的全HN区域MAE分别为94.0 +/- 10.6和102.9 +/- 14.7 HU。对于具有不同组织对比度的患者（水mDixon MR图像），pix 2 pix和CycleGAN的MAE分别增加至122.1 +/- 6.3和132.8 +/- 5.5 HU。我们的研究结果表明，与单视图非重叠补丁结果相比，在每个体素处组合重叠sCT估计减少了MAE和ME。在我们的独立测试集中，PTV和所有临床相关结构（包括具有图像伪影的结构）的绝对百分比平均/最大剂量误差为2%或更低。计划CT和sCT之间的定量DRR比较显示骨区域位置与基于剂量测定和MAE的准确性的一致性，沿着来自sCT的DRR之间的相似性，表明pix 2 pix和CycleGAN是用于HN癌症的仅MR治疗计划的有前景的方法。我们的重叠补丁的HU估计的方法研究还表明，重叠补丁的组合变换估计是一种潜在的方法，以减少生成错误，同时也提供了一种工具，可能估计MR到CT任意模型变换的不确定性。然而，由于患者样本量小，需要进一步研究。
Purpose To evaluate pix2pix and CycleGAN and to assess the effects of multiple combination strategies on accuracy for patch-based synthetic computed tomography (sCT) generation for magnetic resonance (MR)-only treatment planning in head and neck (HN) cancer patients. Materials and methods Twenty-three deformably registered pairs of CT and mDixon FFE MR datasets from HN cancer patients treated at our institution were retrospectively analyzed to evaluate patch-based sCT accuracy via the pix2pix and CycleGAN models. To test effects of overlapping sCT patches on estimations, we (a) trained the models for three orthogonal views to observe the effects of spatial context, (b) we increased effective set size by using per-epoch data augmentation, and (c) we evaluated the performance of three different approaches for combining overlapping Hounsfield unit (HU) estimations for varied patch overlap parameters. Twelve of twenty-three cases corresponded to a curated dataset previously used for atlas-based sCT generation and were used for training with leave-two-out cross-validation. Eight cases were used for independent testing and included previously unseen image features such as fused vertebrae, a small protruding bone, and tumors large enough to deform normal body contours. We analyzed the impact of MR image preprocessing including histogram standardization and intensity clipping on sCT generation accuracy. Effects of mDixon contrast (in-phase vs water) differences were tested with three additional cases. The sCT generation accuracy was evaluated using mean absolute error (MAE) and mean error (ME) in HU between the plan CT and sCT images. Dosimetric accuracy was evaluated for all clinically relevant structures in the independent testing set and digitally reconstructed radiographs (DRRs) were evaluated with respect to the plan CT images. Results The cross-validated MAEs for the whole-HN region using pix2pix and CycleGAN were 66.9 +/- 7.3 vs 82.3 +/- 6.4 HU, respectively. On the independent testing set with additional artifacts and previously unseen image features, whole-HN region MAEs were 94.0 +/- 10.6 and 102.9 +/- 14.7 HU for pix2pix and CycleGAN, respectively. For patients with different tissue contrast (water mDixon MR images), the MAEs increased to 122.1 +/- 6.3 and 132.8 +/- 5.5 HU for pix2pix and CycleGAN, respectively. Our results suggest that combining overlapping sCT estimations at each voxel reduced both MAE and ME compared to single-view non-overlapping patch results. Absolute percent mean/max dose errors were 2% or less for the PTV and all clinically relevant structures in our independent testing set, including structures with image artifacts. Quantitative DRR comparison between planning CTs and sCTs showed agreement of bony region positions to The dosimetric and MAE based accuracy, along with the similarity between DRRs from sCTs, indicate that pix2pix and CycleGAN are promising methods for MR-only treatment planning for HN cancer. Our methods investigated for overlapping patch-based HU estimations also indicate that combining transformation estimations of overlapping patches is a potential method to reduce generation errors while also providing a tool to potentially estimate the MR to CT aleatoric model transformation uncertainty. However, because of small patient sample sizes, further studies are required.