The effects of mismatch between SPECT and CT images on quantitative activity estimation - A simulation study

Enhancing voxel-based dosimetry accuracy with an unsupervised deep learning approach for hybrid medical image registration

BACKGROUND

Deformable registration is required to generate a time-integrated activity (TIA) map which is essential for voxel-based dosimetry. The conventional iterative registration algorithm using anatomical images (e.g., computed tomography (CT)) could result in registration errors in functional images (e.g., single photon emission computed tomography (SPECT) or positron emission tomography (PET)). Various deep learning-based registration tools have been proposed, but studies specifically focused on the registration of serial hybrid images were not found.

PURPOSE

In this study, we introduce CoRX-NET, a novel unsupervised deep learning network designed for deformable registration of hybrid medical images. The CoRX-NET structure is based on the Swin-transformer (ST), allowing for the representation of complex spatial connections in images. Its self-attention mechanism aids in the effective exchange and integration of information across diverse image regions. To augment the amalgamation of SPECT and CT features, cross-stitch layers have been integrated into the network.

METHODS

Two different 177 Lu DOTATATE SPECT/CT datasets were acquired at different medical centers. 22 sets from Seoul National University and 14 sets from Sunway Medical Centre are used for training/internal validation and external validation respectively. The CoRX-NET architecture builds upon the ST, enabling the modeling of intricate spatial relationships within images. To further enhance the fusion of SPECT and CT features, cross-stitch layers have been incorporated within the network. The network takes a pair of SPECT/CT images (e.g., fixed and moving images) and generates a deformed SPECT/CT image. The performance of the network was compared with Elastix and TransMorph using L1 loss and structural similarity index measure (SSIM) of CT, SSIM of normalized SPECT, and local normalized cross correlation (LNCC) of SPECT as metrics. The voxel-wise root mean square errors (RMSE) of TIA were compared among the different methods.

RESULTS

The ablation study revealed that cross-stitch layers improved SPECT/CT registration performance. The cross-stitch layers notably enhance SSIM (internal validation: 0.9614 vs. 0.9653, external validation: 0.9159 vs. 0.9189) and LNCC of normalized SPECT images (internal validation: 0.7512 vs. 0.7670, external validation: 0.8027 vs. 0.8027). CoRX-NET with the cross-stitch layer achieved superior performance metrics compared to Elastix and TransMorph, except for CT SSIM in the external dataset. When qualitatively analyzed for both internal and external validation cases, CoRX-NET consistently demonstrated superior SPECT registration results. In addition, CoRX-NET accomplished SPECT/CT image registration in less than 6 s, whereas Elastix required approximately 50 s using the same PC's CPU. When employing CoRX-NET, it was observed that the voxel-wise RMSE values for TIA were approximately 27% lower for the kidney and 33% lower for the tumor, compared to when Elastix was used.

CONCLUSION

This study represents a major advancement in achieving precise SPECT/CT registration using an unsupervised deep learning network. It outperforms conventional methods like Elastix and TransMorph, reducing uncertainties in TIA maps for more accurate dose assessments.

Generative adversarial network-based attenuation correction for 99mTc-TRODAT-1 brain SPECT

Background

Attenuation correction (AC) is an important correction method to improve the quantification accuracy of dopamine transporter (DAT) single photon emission computed tomography (SPECT). Chang's method was developed for AC (Chang-AC) when CT-based AC was not available, assuming uniform attenuation coefficients inside the body contour. This study aims to evaluate Chang-AC and different deep learning (DL)-based AC approaches on 99mTc-TRODAT-1 brain SPECT using clinical patient data on two different scanners.

Methods

Two hundred and sixty patients who underwent 99mTc-TRODAT-1 SPECT/CT scans from two different scanners (scanner A and scanner B) were retrospectively recruited. The ordered-subset expectation-maximization (OS-EM) method reconstructed 120 projections with dual-energy scatter correction, with or without CT-AC. We implemented a 3D conditional generative adversarial network (cGAN) for the indirect deep learning-based attenuation correction (DL-ACμ) and direct deep learning-based attenuation correction (DL-AC) methods, estimating attenuation maps (μ-maps) and attenuation-corrected SPECT images from non-attenuation-corrected (NAC) SPECT, respectively. We further applied cross-scanner training (cross-scanner indirect deep learning-based attenuation correction [cull-ACμ] and cross-scanner direct deep learning-based attenuation correction [call-AC]) and merged the datasets from two scanners for ensemble training (ensemble indirect deep learning-based attenuation correction [eDL-ACμ] and ensemble direct deep learning-based attenuation correction [eDL-AC]). The estimated μ-maps from (c/e)DL-ACμ were then used in reconstruction for AC purposes. Chang's method was also implemented for comparison. Normalized mean square error (NMSE), structural similarity index (SSIM), specific uptake ratio (SUR), and asymmetry index (%ASI) of the striatum were calculated for different AC methods.

Results

The NMSE for Chang's method, DL-ACμ, DL-AC, cDL-ACμ, cDL-AC, eDL-ACμ, and eDL-AC is 0.0406 ± 0.0445, 0.0059 ± 0.0035, 0.0099 ± 0.0066, 0.0253 ± 0.0102, 0.0369 ± 0.0124, 0.0098 ± 0.0035, and 0.0162 ± 0.0118 for scanner A and 0.0579 ± 0.0146, 0.0055 ± 0.0034, 0.0063 ± 0.0028, 0.0235 ± 0.0085, 0.0349 ± 0.0086, 0.0115 ± 0.0062, and 0.0117 ± 0.0038 for scanner B, respectively. The SUR and %ASI results for DL-ACμ are closer to CT-AC, Followed by DL-AC, eDL-ACμ, cDL-ACμ, cDL-AC, eDL-AC, Chang's method, and NAC.

Conclusion

All DL-based AC methods are superior to Chang-AC. DL-ACμ is superior to DL-AC. Scanner-specific training is superior to cross-scanner and ensemble training. DL-based AC methods are feasible and robust for 99mTc-TRODAT-1 brain SPECT.

SPECT and CT misregistration reduction in [99mTc]Tc-MAA SPECT/CT for precision liver radioembolization treatment planning

PURPOSE

Respiration and body movement induce misregistration between static [99mTc]Tc-MAA SPECT and CT, causing lung shunting fraction (LSF) and tumor-to-normal liver ratio (TNR) errors for 90Y radioembolization planning. We aim to alleviate the misregistration between [99mTc]Tc-MAA SPECT and CT using two registration schemes on simulation and clinical data.

METHODS

In the simulation study, 70 XCAT phantoms were modeled. The SIMIND Monte Carlo program and OS-EM algorithm were used for projection generation and reconstruction, respectively. Low-dose CT (LDCT) at end-inspiration was simulated for attenuation correction (AC), lungs and liver segmentation, while contrast-enhanced CT (CECT) was simulated for tumor and perfused liver segmentation. In the clinical study, 16 patient data including [99mTc]Tc-MAA SPECT/LDCT and CECT with observed SPECT and CT mismatch were analyzed. Two liver-based registration schemes were studied: SPECT registered to LDCT/CECT and vice versa. Mean count density (MCD) of different volumes-of-interest (VOIs), normalized mutual information (NMI), LSF, TNR, and maximum injected activity (MIA) based on the partition model before and after registration were compared. Wilcoxon signed-rank test was performed.

RESULTS

In the simulation study, compared to before registration, registrations significantly reduced estimation errors of MCD of all VOIs, LSF (Scheme 1: - 100.28%, Scheme 2: - 101.59%), and TNR (Scheme 1: - 7.00%, Scheme 2: - 5.67%), as well as MIA (Scheme 1: - 3.22%, Scheme 2: - 2.40%). In the clinical study, Scheme 1 reduced 33.68% LSF and increased 14.75% TNR, while Scheme 2 reduced 38.88% LSF and increased 6.28% TNR compared to before registration. One patient may change from 90Y radioembolization untreatable to treatable and other patients may change the MIA up to 25% after registration. NMI between SPECT and CT was significantly increased after registrations in both studies.

CONCLUSION

Registration between static [99mTc]Tc-MAA SPECT and corresponding CTs is feasible to reduce their spatial mismatch and improve dosimetric estimation. The improvement of LSF is larger than TNR. Our method can potentially improve patient selection and personalized treatment planning for liver radioembolization.

Deep-learning-based estimation of attenuation map improves attenuation correction performance over direct attenuation estimation for myocardial perfusion SPECT

BACKGROUND

Deep learning (DL)-based attenuation correction (AC) is promising to improve myocardial perfusion (MP) SPECT. We aimed to optimize and compare the DL-based direct and indirect AC methods, with and without SPECT and CT mismatch.

METHODS

One hundred patients with different 99mTc-sestamibi activity distributions and anatomical variations were simulated by a population of XCAT phantoms. Additionally, 34 patients 99mTc-sestamibi stress/rest SPECT/CT scans were retrospectively recruited. Projections were reconstructed by OS-EM method with or without AC. Mismatch between SPECT and CT images was modeled. A 3D conditional generative adversarial network (cGAN) was optimized for two DL-based AC methods: (i) indirect approach, i.e., non-attenuation corrected (NAC) SPECT paired with the corresponding attenuation map for training. The projections were reconstructed with the DL-generated attenuation map for AC; (ii) direct approach, i.e., NAC SPECT paired with the corresponding AC SPECT for training to perform direct AC.

RESULTS

Mismatch between SPECT and CT degraded DL-based AC performance. The indirect approach is superior to direct approach for various physical and clinical indices, even with mismatch modeled.

CONCLUSION

DL-based estimation of attenuation map for AC is superior and more robust to direct generation of AC SPECT.

Validation of automated image co-registration integrated into in-house software for voxel-based internal dosimetry on single-photon emission computed tomography images

Objective

To develop an automated co-registration system and test its performance, with and without a fiducial marker, on single-photon emission computed tomography (SPECT) images.

Materials and Methods

Three SPECT/CT scans were acquired for each rotation of a Jaszczak phantom (to 0°, 5°, and 10° in relation to the bed axis), with and without a fiducial marker. Two rigid co-registration software packages-SPM12 and NMDose-coreg-were employed, and the percent root mean square error (%RMSE) was calculated in order to assess the quality of the co-registrations. Uniformity, contrast, and resolution were measured before and after co-registration. The NMDose-coreg software was employed to calculate the renal doses in 12 patients treated with 177Lu-DOTATATE, and we compared those with the values obtained with the Organ Level INternal Dose Assessment for EXponential Modeling (OLINDA/EXM) software.

Results

The use of a fiducial marker had no significant effect on the quality of co-registration on SPECT images, as measured by %RMSE (p = 0.40). After co-registration, uniformity, contrast, and resolution did not differ between the images acquired with fiducial markers and those acquired without. Preliminary clinical application showed mean total processing times of 9 ± 3 min/patient for NMDose-coreg and 64 ± 10 min/patient for OLINDA/EXM, with a strong correlation between the two, despite the lower renal doses obtained with NMDose-coreg.

Conclusion

The use of NMDose-coreg allows fast co-registration of SPECT images, with no loss of uniformity, contrast, or resolution. The use of a fiducial marker does not appear to increase the accuracy of co-registration on phantoms.

Voxel-S-Value based 3D treatment planning methods for Y-90 microspheres radioembolization based on Tc-99m-macroaggregated albumin SPECT/CT

Partition model (PM) for Y-90 microsphere radioembolization is limited in providing 3D dosimetrics. Voxel-S-Values (VSV) method has good agreement with Monte Carlo (MC) simulations for 3D absorbed dose conversion. We propose a new VSV method and compare its performance along with PM, MC and other VSV methods for Y-90 RE treatment planning based on Tc-99m MAA SPECT/CT. Twenty Tc-99m-MAA SPECT/CT patient data are retrospectively analyzed. Seven VSV methods are implemented: (1) local energy deposition; (2) liver kernel; (3) liver kernel and lung kernel; (4) liver kernel with density correction (LiKD); (5) liver kernel with center voxel scaling (LiCK); (6) liver kernel and lung kernel with density correction (LiLuKD); (7) proposed liver kernel with center voxel scaling and lung kernel with density correction (LiCKLuKD). Mean absorbed dose and maximum injected activity (MIA) obtained by PM and VSV are evaluated against MC results, and 3D dosimetrics generated by VSV are compared with MC. LiKD, LiCK, LiLuKD and LiCKLuKD have the smallest deviation in normal liver and tumors. LiLuKD and LiCKLuKD have the best performance in lungs. MIAs are similar by all methods. LiCKLuKD could provide MIA consistent with PM, and precise 3D dosimetrics for Y-90 RE treatment planning.