Synthetic dual-energy CT for MRI-only based proton therapy treatment planning using label-GAN

Abstract

MRI-only treatment planning is highly desirable in the current proton radiation therapy workflow due to its appealing advantages such as bypassing MR-CT co-registration, avoiding x-ray CT exposure dose and reduced medical cost. However, MRI alone cannot provide stopping power ratio (SPR) information for dose calculations. Given that dual energy CT (DECT) can estimate SPR with higher accuracy than conventional single energy CT, we propose a deep learning-based method in this study to generate synthetic DECT (sDECT) from MRI to calculate SPR. Since the contrast difference between high-energy and low-energy CT (LECT) is important, and in order to accurately model this difference, we propose a novel label generative adversarial network-based model which can not only discriminate the realism of sDECT but also differentiate high-energy CT (HECT) and LECT from DECT. A cohort of 57 head-and-neck cancer patients with DECT and MRI pairs were used to validate the performance of the proposed framework. The results of sDECT and its derived SPR maps were compared with clinical DECT and the corresponding SPR, respectively. The mean absolute error for synthetic LECT and HECT were 79.98 ± 18.11 HU and 80.15 ± 16.27 HU, respectively. The corresponding SPR maps generated from sDECT showed a normalized mean absolute error as 5.22% ± 1.23%. By comparing with the traditional Cycle GANs, our proposed method significantly improves the accuracy of sDECT. The results indicate that on our dataset, the sDECT image form MRI is close to planning DECT, and thus shows promising potential for generating SPR maps for proton therapy.

Figure 1.

Schematic flow chart of the proposed MRI-based DECT synthesis method.

Figure 2.

Network architecture used in figure 1.

Figure 3.

Comparison between LECT and synthetic LECT of a representative patient. From left to right: MR image, CT image, synthetic LECT image, HU difference between CT and synthetic LECT images. Display window: [−1000, 1000] HU for CT images.

Figure 4.

Comparison between HECT and synthetic HECT of a representative patient. From left to right: MR image, CT image, synthetic HECT image, HU difference between CT and synthetic HECT images. Display window: [−1000, 1000] HU for CT images.

Figure 5.

Comparison between HECT and synthetic HECT of a representative patient. From left to right: MR image, CT image, synthetic HECT image, HU difference between CT and synthetic HECT images. Display window: [−1000, 1000] HU for CT images

Figure 6.

Comparison between the SPR by DECT and SPR by sDECT of a representative patient. From left to right: MR image, SPR image by DECT, SPR image by sDECT, and SPR difference between the SPR by DECT and sDECT. Bottom: profiles of red lines in SPR by DECT and sDECT images. Display window: [0, 3] for SPR map.

Figure 7.

Scatter plots of SPR MAE with respect to ground truth SPR of all voxels. (a) is from patient’s data that has relatively lower SPR MAE,(b)is from patient’s data that has relatively higher SPR MAE.

Figure 8.

The boxplots of DVH metrics difference (dose of using DECT-based SPR minus dose of using sDECT-based SPR) of CTV, left parotid, right parotid, spinal cord, and oral cavity. The tops and bottoms of each box are the 25th and 75th percentiles of the samples, respectively. The distances between the tops and bottoms interquartile ranges. The line in the middle of each box is the sample median. The whiskers are the lines extending above and below each box. The whiskers are drawn from the ends of the interquartile ranges to the furthest observations within the whisker length. Observations that are beyond the whisker length are marked as outliers. The outlier is a value that is more than 1.5 times the interquartile range away from the top or bottom of the box. The outliers beyond the box were plotted in red markers.

Figure 9.

Comparison of dose distribution and DVH. (a)–(c)show the dose distribution of plan using DECT-based SPR. (d)–(f)show the dose distribution of plan using sDECT-based SPR. (g)shows the cumulative DVH, where solid lines are for DECT plan, and dash lines are for sDECT plan.