PET image denoising based on denoising diffusion probabilistic model.

PURPOSE: Due to various physical degradation factors and limited counts received, PET image quality needs further improvements. The denoising diffusion probabilistic model (DDPM) was a distribution learning-based model, which tried to transform a normal distribution into a specific data distribution based on iterative refinements. In this work, we proposed and evaluated different DDPM-based methods for PET image denoising. METHODS: Under the DDPM framework, one way to perform PET image denoising was to provide the PET image and/or the prior image as the input. Another way was to supply the prior image as the network input with the PET image included in the refinement steps, which could fit for scenarios of different noise levels. 150 brain [[Formula: see text]F]FDG datasets and 140 brain [[Formula: see text]F]MK-6240 (imaging neurofibrillary tangles deposition) datasets were utilized to evaluate the proposed DDPM-based methods. RESULTS: Quantification showed that the DDPM-based frameworks with PET information included generated better results than the nonlocal mean, Unet and generative adversarial network (GAN)-based denoising methods. Adding additional MR prior in the model helped achieved better performance and further reduced the uncertainty during image denoising. Solely relying on MR prior while ignoring the PET information resulted in large bias. Regional and surface quantification showed that employing MR prior as the network input while embedding PET image as a data-consistency constraint during inference achieved the best performance. CONCLUSION: DDPM-based PET image denoising is a flexible framework, which can efficiently utilize prior information and achieve better performance than the nonlocal mean, Unet and GAN-based denoising methods.

Solid Attenuation Components Attention Deep Learning Model to Predict Micropapillary and Solid Patterns in Lung Adenocarcinomas on Computed Tomography.

BACKGROUND: High-grade adenocarcinoma subtypes (micropapillary and solid) treated with sublobar resection have an unfavorable prognosis compared with those treated with lobectomy. We investigated the potential of incorporating solid attenuation component masks with deep learning in the prediction of high-grade components to optimize surgical strategy preoperatively. METHODS: A total of 502 patients with pathologically confirmed high-grade adenocarcinomas were retrospectively enrolled between 2016 and 2020. The SACs attention DL model was developed to apply solid-attenuation-component-like subregion masks (tumor area ≥ - 190 HU) to guide the DL model for predicting high-grade subtypes. The SACA-DL was assessed using 5-fold cross-validation and external validation in the training and testing sets, respectively. The performance, which was evaluated using the area under the curve (AUC), was compared between SACA-DL and the DL model without SACs attention (DLwoSACs), the prior radiomics model, or the model based on the consolidation/tumor (C/T) diameter ratio. RESULTS: We classified 313 and 189 patients into training and testing cohorts, respectively. The SACA-DL achieved an AUC of 0.91 for the cross-validation, which was significantly superior to those of the DLwoSACs (AUC = 0.88; P = 0.02), prior radiomics model (AUC = 0.85; P = 0.004), and C/T ratio (AUC = 0.84; P = 0.002). An AUC of 0.93 was achieved for external validation in the SACA-DL and was significantly better than those of the DLwoSACs (AUC = 0.89; P = 0.04), prior radiomics model (AUC = 0.85; P < 0.001), and C/T ratio (AUC = 0.85; P < 0.001). CONCLUSIONS: The combination of solid-attenuation-component-like subregion masks with the DL model is a promising approach for the preoperative prediction of high-grade adenocarcinoma subtypes.

Respiration-averaged CT versus standard CT attenuation map for correction of 18F-sodium fluoride uptake in coronary atherosclerotic lesions on hybrid PET/CT.

BACKGROUND: To evaluate the impact of respiratory-averaged computed tomography attenuation correction (RACTAC) compared to standard single-phase computed tomography attenuation correction (CTAC) map, on the quantitative measures of coronary atherosclerotic lesions of 18F-sodium fluoride (18F-NaF) uptake in hybrid positron emission tomography and computed tomography (PET/CT). METHODS: This study comprised 23 patients who underwent 18F-NaF coronary PET in a hybrid PET/CT system. All patients had a standard single-phase CTAC obtained during free-breathing and a 4D cine-CT scan. From the cine-CT acquisition, RACTAC maps were obtained by averaging all images acquired over 5 seconds. PET reconstructions using either CTAC or RACTAC were compared. The quantitative impact of employing RACTAC was assessed using maximum target-to-background (TBRMAX) and coronary microcalcification activity (CMA). Statistical differences were analyzed using reproducibility coefficients and Bland-Altman plots. RESULTS: In 23 patients, we evaluated 34 coronary lesions using CTAC and RACTAC reconstructions. There was good agreement between CTAC and RACTAC for TBRMAX (median [Interquartile range]): CTAC = 1.65 [1.23 to 2.38], RACTAC = 1.63 [1.23 to 2.33], p = 0.55), with coefficient of reproducibility of 0.18, and CMA: CTAC = 0.10 [0 to 1.0], RACTAC = 0.15 [0 to 1.03], p = 0.55 with coefficient of reproducibility of 0.17 CONCLUSION: Respiratory-averaged and standard single-phase attenuation correction maps provide similar and reproducible methods of quantifying coronary 18F-NaF uptake on PET/CT.

Comparison of ejection fraction calculation between CT and SPECT at high heart rate: A dynamic cardiac phantom study.

The purpose of this study is to compare the ejection fraction (EF) calculation of CT and SPECT at high heart rate. A dynamic cardiac phantom with programmable end-systolic volume (ESV), end-diastolic volume (EDV), and heart rate was used to compare CT, which has high spatial resolution (< 1 mm) and modest temporal resolution of 175 msec, and SPECT, which has high temporal resolution of 16 bins per cardiac cycle but poor spatial resolution (> 1 cm) in EF, ESV, and EDV at the heart rates ≤ 100 bpm for EF = 30 (disease state) and EF = 60 (healthy state). EF calculations for SPECT were accurate in 2% for 40 to 100 bpm for both EF = 30 and EF = 60, and were not heart rate dependent although both ESV and EDV could be underestimated by 18-20%. EF calculations for CT were accurate in 2.2% for 40 and 60 bpm. Inaccuracy in EF calculations, ESV and EDV estimates increased when the heart rate or EF increased. SPECT was accurate for EF calculation for the heart rates ≤ 100 bpm and CT was accurate for the heart rates of ≤ 60 bpm. CT was less accurate for the high heart rates of 80 and 100 bpm, or high EF = 60.

Pitfalls in quantitative myocardial PET perfusion I: Myocardial partial volume correction.

BACKGROUND: PET quantitative myocardial perfusion requires correction for partial volume loss due to one-dimensional LV wall thickness smaller than scanner resolution. METHODS: We aimed to assess accuracy of risk stratification for death, MI, or revascularization after PET using partial volume corrections derived from two-dimensional ACR and three-dimensional NEMA phantoms for 3987 diagnostic rest-stress perfusion PETs and 187 MACE events. NEMA, ACR, and Tree phantoms were imaged with Rb-82 or F-18 for size-dependent partial volume loss. Perfusion and Coronary Flow Capacity were recalculated using different ACR- and NEMA-derived partial volume corrections compared by Kolmogorov-Smirnov statistics to standard perfusion metrics with established correlations with MACE. RESULTS: Partial volume corrections based on two-dimensional ACR rods (two equal radii) and three-dimensional NEMA spheres (three equal radii) over estimate partial volume corrections, quantitative perfusion, and Coronary Flow Capacity by 50% to 150% over perfusion metrics with one-dimensional partial volume correction, thereby substantially impairing correct risk stratification. CONCLUSIONS: ACR (2-dimensional) and NEMA (3-dimensional) phantoms overestimate partial volume corrections for 1-dimensional LV wall thickness and myocardial perfusion that are corrected with a simple equation that correlates with MACE for optimal risk stratification applicable to most PET-CT scanners for quantifying myocardial perfusion.

Optimization of reconstruction and quantification of motion-corrected coronary PET-CT.

BACKGROUND: Coronary PET shows promise in the detection of high-risk atherosclerosis, but there remains a need to optimize imaging and reconstruction techniques. We investigated the impact of reconstruction parameters and cardiac motion-correction in 18F Sodium Fluoride (18F-NaF) PET. METHODS: Twenty-two patients underwent 18F-NaF PET within 22 days of an acute coronary syndrome. Optimal reconstruction parameters were determined in a subgroup of six patients. Motion-correction was performed on ECG-gated data of all patients with optimal reconstruction. Tracer uptake was quantified in culprit and reference lesions by computing signal-to-noise ratio (SNR) in diastolic, summed, and motion-corrected images. RESULTS: Reconstruction using 24 subsets, 4 iterations, point-spread-function modelling, time of flight, and 5-mm post-filtering provided the highest median SNR (31.5) compared to 4 iterations 0-mm (22.5), 8 iterations 0-mm (21.1), and 8 iterations 5-mm (25.6; all P < .05). Motion-correction improved SNR of culprit lesions (n = 33) (24.5[19.9-31.5]) compared to diastolic (15.7[12.4-18.1]; P < .001) and summed data (22.1[18.9-29.2]; P < .001). Motion-correction increased the SNR difference between culprit and reference lesions (10.9[6.3-12.6]) compared to diastolic (6.2[3.6-10.3]; P = .001) and summed data (7.1 [4.8-11.6]; P = .001). CONCLUSIONS: The number of iterations and extent of post-filtering has marked effects on coronary 18F-NaF PET quantification. Cardiac motion-correction improves discrimination between culprit and reference lesions.

Potential effects of low-dose average CT on cardiac implantable electronic devices.

BACKGROUND: Average CT has been shown to be more accurate than conventional helical CT in quantitation of the PET data. The risk of CT irradiation of a cardiac implantable electronic device (CIED) causing an adverse event is low and is generally outweighed by the clinical benefit of a medically indicated examination. However, irradiation of CIED over one breath cycle in cine CT scan for average CT could impose risks on a patient who is pacing dependent. The purpose of this study was to demonstrate that low-dose average CT can be safe for CIED. METHODS: A Medtronic CIED of model Protecta VR was submerged in a saline bath for a series of 4-s cine CT scans on a GE CT scanner programmed to deliver a 2-cm-wide radiation at a dose rate of 0.9 to 41.2 mGy/s to the CIED. The number of over-sensings was recorded as the interference of radiation to the CIED. RESULTS: Dose rates ≥ 1.9 mGy/s caused over-sensing. The higher the dose rate, the more over-sensings there were. The lowest dose rate of 0.9 mGy/s did not cause any over-sensing. CONCLUSIONS: Low-dose average CT at 0.9 mGy/s can be safe for a CIED patient who is pacing dependent.

Comparison of enhancement quantification from virtual unenhanced images to true unenhanced images in multiphase renal Dual-Energy computed tomography: A phantom study.

Multiphase computed tomography (CT) exams are a commonly used imaging technique for the diagnosis of renal lesions and involve the acquisition of a true unenhanced (TUE) series followed by one or more postcontrast series. The difference in CT number of the mass in pre- and postcontrast images is used to quantify enhancement, which is an important criterion used for diagnosis. This study sought to assess the feasibility of replacing TUE images with virtual unenhanced (VUE) images derived from Dual-Energy CT datasets in renal CT exams. Eliminating TUE image acquisition could reduce patient dose and improve clinical efficiency. A rapid kVp-switching CT scanner was used to assess enhancement accuracy when using VUE compared to TUE images as the baseline for enhancement calculations across a wide range of clinical scenarios simulated in a phantom study. Three phantoms were constructed to simulate small, medium, and large patients, each with varying lesion size and location. Nonenhancing cystic lesions were simulated using distilled water. Intermediate (10-20 HU [Hounsfield units]) and positively enhancing masses (≥20 HU) were simulated by filling the spherical inserts in each phantom with varied levels of iodinated contrast mixed with a blood surrogate. The results were analyzed using Bayesian hierarchical models. Posterior probabilities were used to classify enhancement measured using VUE compared to TUE images as significantly less, not significantly different, or significantly higher. Enhancement measured using TUE images was considered the ground truth in this study. For simulation of nonenhancing renal lesions, enhancement values were not significantly different when using VUE versus TUE images, with posterior probabilities ranging from 0.23-0.56 across all phantom sizes and an associated specificity of 100%. However, for simulation of intermediate and positively enhancing lesions significant differences were observed, with posterior probabilities < 0.05, indicating significantly lower measured enhancement when using VUE versus TUE images. Positively enhancing masses were categorized accurately, with a sensitivity of 91.2%, when using VUE images as the baseline. For all scenarios where iodine was present, VUE-based enhancement measurements classified lesions with a sensitivity of 43.2%, a specificity of 100%, and an accuracy of 78.1%. Enhancement calculated using VUE images proved to be feasible for classifying nonenhancing and highly enhancing lesions. However, differences in measured enhancement for simulation of intermediately enhancing lesions demonstrated that replacement of TUE with VUE images may not be advisable for renal CT exams.

Patient-reported lung symptoms as an early signal of impending radiation pneumonitis in patients with non-small cell lung cancer treated with chemoradiation: an observational study.

PURPOSE: Clinician ratings of concurrent chemoradiation (CRT)-induced radiation pneumonitis (RP) in patients with non-small cell lung cancer (NSCLC) are based on both imaging and patient-reported lung symptoms. We compared the value of patient-reported outcomes versus normal-lung uptake of 18F-fluoro-2-deoxyglucose in positron emission computed tomography (FDG PET/CT) during the last week of treatment, for indicating the development of grade ≥ 2 RP within 4 months of CRT completion. METHODS: 132 patients with NSCLC-reported RP-related symptoms (coughing, shortness of breath) repeatedly using the validated MD Anderson Symptom Inventory lung cancer module. Of these patients, 68 had FDG PET/CT scans that were analyzed for normal-lung mean standardized FDG uptake values (SUVmean) before, during, and up to 4 months after CRT. Clinicians rated RP using CTCAE version 3. Logistic regression models examined potential predictors for developing CTCAE RP ≥ 2. RESULTS: For the entire sample, patient-rated RP-related symptoms during the last week of CRT correlated with clinically meaningful CTCAE RP ≥ 2 post-CRT (OR 2.74, 95% CI 1.25-5.99, P = 0.012), controlled for sex, age, mean lung radiation dose, comorbidity, and baseline symptoms. Moderate/severe patient-rated RP-related symptom score (≥ 4 on a 0-10 scale, P = 0.001) and normal-lung FDG uptake (SUVmean > 0.78, P = 0.002) in last week of CRT were equally strong predictors of post-CRT CTCAE RP ≥ 2 (C-index = 0.78, 0.77). CONCLUSIONS: During the last week of CRT, routine assessment of moderate-to-severe RP-related symptoms provides a simple way to identify patients with NSCLC who may be at risk for developing significant post-CRT RP, especially when PET/CT images of normal-lung FDG uptake are not available.

Feasibility of using respiration-averaged MR images for attenuation correction of cardiac PET/MR imaging.

Cardiac imaging is a promising application for combined PET/MR imaging. However, current MR imaging protocols for whole-body attenuation correction can produce spatial mismatch between PET and MR-derived attenuation data owing to a disparity between the two modalities' imaging speeds. We assessed the feasibility of using a respiration-averaged MR (AMR) method for attenuation correction of cardiac PET data in PET/MR images. First, to demonstrate the feasibility of motion imaging with MR, we used a 3T MR system and a two-dimensional fast spoiled gradient-recalled echo (SPGR) sequence to obtain AMR images ofa moving phantom. Then, we used the same sequence to obtain AMR images of a patient's thorax under free-breathing conditions. MR images were converted into PET attenuation maps using a three-class tissue segmentation method with two sets of predetermined CT numbers, one calculated from the patient-specific (PS) CT images and the other from a reference group (RG) containing 54 patient CT datasets. The MR-derived attenuation images were then used for attenuation correction of the cardiac PET data, which were compared to the PET data corrected with average CT (ACT) images. In the myocardium, the voxel-by-voxel differences and the differences in mean slice activity between the AMR-corrected PET data and the ACT-corrected PET data were found to be small (less than 7%). The use of AMR-derived attenuation images in place of ACT images for attenuation correction did not affect the summed stress score. These results demonstrate the feasibility of using the proposed SPGR-based MR imaging protocol to obtain patient AMR images and using those images for cardiac PET attenuation correction. Additional studies with more clinical data are warranted to further evaluate the method.

Evaluation of 4D CT acquisition methods designed to reduce artifacts.

Four-dimensional computed tomography (4D CT) is used to account for respiratory motion in radiation treatment planning, but artifacts resulting from the acquisition and postprocessing limit its accuracy. We investigated the efficacy of three experimental 4D CT acquisition methods to reduce artifacts in a prospective institutional review board approved study. Eighteen thoracic patients scheduled to undergo radiation therapy received standard clinical 4D CT scans followed by each of the alternative 4D CT acquisitions: 1) data oversampling, 2) beam gating with breathing irregularities, and 3) rescanning the clinical acquisition acquired during irregular breathing. Relative values of a validated correlation-based artifact metric (CM) determined the best acquisition method per patient. Each 4D CT was processed by an extended phase sorting approach that optimizes the quantitative artifact metric (CM sorting). The clinical acquisitions were also postprocessed by phase sorting for artifact comparison of our current clinical implementation with the experimental methods. The oversampling acquisition achieved the lowest artifact presence among all acquisitions, achieving a 27% reduction from the current clinical 4D CT implementation (95% confidence interval = 34-20). The rescan method presented a significantly higher artifact presence from the clinical acquisition (37%; p < 0.002), the gating acquisition (26%; p < 0.005), and the oversampling acquisition (31%; p < 0.001), while the data lacked evidence of a significant difference between the clinical, gating, and oversampling methods. The oversampling acquisition reduced artifact presence from the current clinical 4D CT implementation to the largest degree and provided the simplest and most reproducible implementation. The rescan acquisition increased artifact presence significantly, compared to all acquisitions, and suffered from combination of data from independent scans over which large internal anatomic shifts occurred.

The prognostic value of interim positron emission tomography scan in patients with classical Hodgkin lymphoma.

The prognostic value of interim positron emission tomography (PET) was evaluated after 2 cycles of doxorubicin, bleomycin, vinblastin and dacarbazine in classical Hodgkin lymphoma patients (n = 229), based on Deauville criteria. In early stage non-bulky disease, bulky stage II disease, advanced stage low International Prognostic Score (IPS ≤2) and advanced stage (IPS ≥3), 3-year progression-free survival rates in PET2-negative vs. PET2-positive groups were 95·9% vs. 76·9% (P < 0·0018), 83·3% vs. 20·0% (P = 0·017), 77·0% vs. 30·0% (P < 0·001) and 71·0% vs. 44·4%(P = 0·155), respectively. The outcome after positive PET2 was better than previously reported. The results from non-randomized studies of PET2-guided therapy would be valuable with careful interpretation.

Defining internal target volume using positron emission tomography for radiation therapy planning of moving lung tumors.

Substantial disagreement exists over appropriate PET segmentation techniques for non-small cell lung cancer. Currently, no segmentation algorithm explicitly considers tumor motion in determining tumor borders. We developed an automatic PET segmentation model as a function of target volume, motion extent, and source-to-background ratio (the VMSBR model). The purpose of this work was to apply the VMSBR model and six other segmentation algorithms to a sample of lung tumors. PET and 4D CT were performed in the same imaging session for 23 patients (24 tumors) for radiation therapy planning. Internal target volumes (ITVs) were autosegmented on maximum intensity projection (MIP) of cine CT. ITVs were delineated on PET using the following methods: 15%, 35%, and 42% of maximum activity concentration, standardized uptake value (SUV) of 2.5 g/mL, 15% of mean activity concentration plus background, a linear function of mean SUV, and the VMSBR model. Predicted threshold values from each method were compared to measured optimal threshold values, and resulting volume magnitudes were compared to cine-CT-derived ITV. Correlation between predicted and measured threshold values ranged from slopes of 0.29 for the simplest single-threshold techniques to 0.90 for the VMSBR technique. R2 values ranged from 0.07 for the simplest single-threshold techniques to 0.86 for the VMSBR technique. The VMSBR segmentation technique that included volume, motion, and source-to-background ratio, produced accurate ITVs in patients when compared with cine-CT-derived ITV.

Assessment of a quantitative metric for 4D CT artifact evaluation by observer consensus.

The benefits of four-dimensional computed tomography (4D CT) are limited by the presence of artifacts that remain difficult to quantify. A correlation-based metric previously proposed for ciné 4D CT artifact identification was further validated as an independent artifact evaluator by using a novel qualitative assessment featuring a group of observers reaching a consensus decision on artifact location and magnitude. The consensus group evaluated ten ciné 4D CT scans for artifacts over each breathing phase of coronal lung views assuming one artifact per couch location. Each artifact was assigned a magnitude score of 1-5, 1 indicating lowest severity and 5 indicating highest severity. Consensus group results served as the ground truth for assessment of the correlation metric. The ten patients were split into two cohorts; cohort 1 generated an artifact identification threshold derived from receiver operating characteristic analysis using the Youden Index, while cohort 2 generated sensitivity and specificity values from application of the artifact threshold. The Pearson correlation coefficient was calculated between the correlation metric values and the consensus group scores for both cohorts. The average sensitivity and specificity values found with application of the artifact threshold were 0.703 and 0.476, respectively. The correlation coefficients of artifact magnitudes for cohort 1 and 2 were 0.80 and 0.61, respectively, (p < 0.001 for both); these correlation coefficients included a few scans with only two of the five possible magnitude scores. Artifact incidence was associated with breathing phase (p < 0.002), with presentation less likely near maximum exhale. Overall, the correlation metric allowed accurate and automated artifact identification. The consensus group evaluation resulted in efficient qualitative scoring, reduced interobserver variation, and provided consistent identification of artifact location and magnitudes.

Radiation dose reduction for CT lung cancer screening using ASIR and MBIR: a phantom study.

The purpose of this study was to reduce the radiation dosage associated with computed tomography (CT) lung cancer screening while maintaining overall diagnostic image quality and definition of ground-glass opacities (GGOs). A lung screening phantom and a multipurpose chest phantom were used to quantitatively assess the performance of two iterative image reconstruction algorithms (adaptive statistical iterative reconstruction (ASIR) and model-based iterative reconstruction (MBIR)) used in conjunction with reduced tube currents relative to a standard clinical lung cancer screening protocol (51 effective mAs (3.9 mGy) and filtered back-projection (FBP) reconstruction). To further assess the algorithms' performances, qualitative image analysis was conducted (in the form of a reader study) using the multipurpose chest phantom, which was implanted with GGOs of two densities. Our quantitative image analysis indicated that tube current, and thus radiation dose, could be reduced by 40% or 80% from ASIR or MBIR, respectively, compared with conventional FBP, while maintaining similar image noise magnitude and contrast-to-noise ratio. The qualitative portion of our study, which assessed reader preference, yielded similar results, indicating that dose could be reduced by 60% (to 20 effective mAs (1.6 mGy)) with either ASIR or MBIR, while maintaining GGO definition. Additionally, the readers' preferences (as indicated by their ratings) regarding overall image quality were equal or better (for a given dose) when using ASIR or MBIR, compared with FBP. In conclusion, combining ASIR or MBIR with reduced tube current may allow for lower doses while maintaining overall diagnostic image quality, as well as GGO definition, during CT lung cancer screening.

Comments on the paper "Data-driven gating (DDG)-based motion match for improved CTAC registration. EJNMMI Physics. 2024;11(1):42."

Misregistration between CT and PET in PET/CT is mainly caused by respiratory motion or irregular respiration during the CT scan in PET/CT. Other than repeat CT, repeat PET/CT, or data-driven gated (DDG) CT, there is no practical approach to mitigate the misregistration artifacts and subsequent CT attenuation correction (CTAC) of the PET data. DDG PET derives a respiratory motion model based on the multiple phases of PET images without hardware gating and it allows for a potential correction of the misregistration artifacts based on the respiratory motion model. The purpose of this commentary was to compare the recent two publications on matching the random phase of helical CT with one of the PET phases derived from the motion model of DDG PET and warping the misregistered helical CT for CTAC of and registration with PET or DDG PET. The two publications were similar in methodology. However, the data sets used for the comparison were different and could potentially impact their conclusions.

Data-driven gated positron emission tomography/computed tomography for radiotherapy.

Purpose: Software-based data-driven gated (DDG) positron emission tomography/computed tomography (PET/CT) has replaced hardware-based 4D PET/CT. The purpose of this article was to review DDG PET/CT, which could improve the accuracy of treatment response assessment, tumor motion evaluation, and target tumor contouring with whole-body (WB) PET/CT for radiotherapy (RT). Material and methods: This review covered the topics of 4D PET/CT with hardware gating, advancements in PET instrumentation, DDG PET, DDG CT, and DDG PET/CT based on a systematic literature review. It included a discussion of the large axial field-of-view (AFOV) PET detector and a review of the clinical results of DDG PET and DDG PET/CT. Results: DDG PET matched or outperformed 4D PET with hardware gating. DDG CT was more compatible with DDG PET than 4D CT, which required hardware gating. DDG CT could replace 4D CT for RT. DDG PET and DDG CT for DDG PET/CT can be incorporated in a WB PET/CT of less than 15 min scan time on a PET/CT scanner of at least 25 cm AFOV PET detector. Conclusions: DDG PET/CT could correct the misregistration and tumor motion artifacts in a WB PET/CT and provide the quantitative PET and tumor motion information of a registered PET/CT for RT.

Dynamic CBCT imaging using prior model-free spatiotemporal implicit neural representation (PMF-STINR).

Objective. Dynamic cone-beam computed tomography (CBCT) can capture high-spatial-resolution, time-varying images for motion monitoring, patient setup, and adaptive planning of radiotherapy. However, dynamic CBCT reconstruction is an extremely ill-posed spatiotemporal inverse problem, as each CBCT volume in the dynamic sequence is only captured by one or a few x-ray projections, due to the slow gantry rotation speed and the fast anatomical motion (e.g. breathing).Approach. We developed a machine learning-based technique, prior-model-free spatiotemporal implicit neural representation (PMF-STINR), to reconstruct dynamic CBCTs from sequentially acquired x-ray projections. PMF-STINR employs a joint image reconstruction and registration approach to address the under-sampling challenge, enabling dynamic CBCT reconstruction from singular x-ray projections. Specifically, PMF-STINR uses spatial implicit neural representations to reconstruct a reference CBCT volume, and it applies temporal INR to represent the intra-scan dynamic motion of the reference CBCT to yield dynamic CBCTs. PMF-STINR couples the temporal INR with a learning-based B-spline motion model to capture time-varying deformable motion during the reconstruction. Compared with the previous methods, the spatial INR, the temporal INR, and the B-spline model of PMF-STINR are all learned on the fly during reconstruction in a one-shot fashion, without using any patient-specific prior knowledge or motion sorting/binning.Main results. PMF-STINR was evaluated via digital phantom simulations, physical phantom measurements, and a multi-institutional patient dataset featuring various imaging protocols (half-fan/full-fan, full sampling/sparse sampling, different energy and mAs settings, etc). The results showed that the one-shot learning-based PMF-STINR can accurately and robustly reconstruct dynamic CBCTs and capture highly irregular motion with high temporal (∼ 0.1 s) resolution and sub-millimeter accuracy.Significance. PMF-STINR can reconstruct dynamic CBCTs and solve the intra-scan motion from conventional 3D CBCT scans without using any prior anatomical/motion model or motion sorting/binning. It can be a promising tool for motion management by offering richer motion information than traditional 4D-CBCTs.

Real-time CBCT Imaging and Motion Tracking via a Single Arbitrarily-angled X-ray Projection by a Joint Dynamic Reconstruction and Motion Estimation (DREME) Framework.

Objective: Real-time cone-beam computed tomography (CBCT) provides instantaneous visualization of patient anatomy for image guidance, motion tracking, and online treatment adaptation in radiotherapy. While many real-time imaging and motion tracking methods leveraged patient-specific prior information to alleviate under-sampling challenges and meet the temporal constraint (< 500 ms), the prior information can be outdated and introduce biases, thus compromising the imaging and motion tracking accuracy. To address this challenge, we developed a framework (DREME) for real-time CBCT imaging and motion estimation, without relying on patient-specific prior knowledge. Approach: DREME incorporates a deep learning-based real-time CBCT imaging and motion estimation method into a dynamic CBCT reconstruction framework. The reconstruction framework reconstructs a dynamic sequence of CBCTs in a data-driven manner from a standard pre-treatment scan, without utilizing patient-specific knowledge. Meanwhile, a convolutional neural network-based motion encoder is jointly trained during the reconstruction to learn motion-related features relevant for real-time motion estimation, based on a single arbitrarily-angled x-ray projection. DREME was tested on digital phantom simulation and real patient studies. Main results: DREME accurately solved 3D respiration-induced anatomic motion in real time (~1.5 ms inference time for each x-ray projection). In the digital phantom study, it achieved an average lung tumor center-of-mass localization error of 1.2±0.9 mm (Mean±SD). In the patient study, it achieved a real-time tumor localization accuracy of 1.8±1.6 mm in the projection domain. Significance: DREME achieves CBCT and volumetric motion estimation in real time from a single x-ray projection at arbitrary angles, paving the way for future clinical applications in intra-fractional motion management. In addition, it can be used for dose tracking and treatment assessment, when combined with real-time dose calculation.