Deep learning-based dynamic PET parametric Ki image generation from lung static PET

Feasibility exploration of myocardial blood flow synthesis from a simulated static myocardial computed tomography perfusion via a deep neural network

BACKGROUND

 Myocardial blood flow (MBF) provides important diagnostic information for myocardial ischemia. However, dynamic computed tomography perfusion (CTP) needed for MBF involves multiple exposures, leading to high radiation doses.

OBJECTIVES

 This study investigated synthesizing MBF from simulated static myocardial CTP to explore dose reduction potential, bypassing the traditional dynamic input function.

METHODS

 The study included 253 subjects with intermediate-to-high pretest probabilities of obstructive coronary artery disease (CAD). MBF was reconstructed from dynamic myocardial CTP. A deep neural network (DNN) converted simulated static CTP into synthetic MBF. Beyond the usual image quality evaluation, the synthetic MBF was segmented and a clinical functional assessment was conducted, with quantitative analysis for consistency and correlation.

RESULTS

 Synthetic MBF closely matched the referenced MBF, with an average structure similarity (SSIM) of 0.87. ROC analysis of ischemic segments showed an area under curve (AUC) of 0.915 for synthetic MBF. This method can theoretically reduce the radiation dose for MBF significantly, provided satisfactory static CTP is obtained, reducing reliance on high time resolution of dynamic CTP.

CONCLUSIONS

 The proposed method is feasible, with satisfactory clinical functionality of synthetic MBF. Further investigation and validation are needed to confirm actual dose reduction in clinical settings.

Self-supervised neural network for Patlak-based parametric imaging in dynamic [18F]FDG total-body PET

PURPOSE

The objective of this study is to generate reliable Ki parametric images from a shortened [18F]FDG total-body PET for clinical applications using a self-supervised neural network algorithm.

METHODS

We proposed a self-supervised neural network algorithm with Patlak graphical analysis (SN-Patlak) to generate Ki images from shortened dynamic [18F]FDG PET without 60-min full-dynamic PET-based training. The algorithm deeply integrates neural network architecture with a Patlak method, employing the fitting error of the Patlak plot as the neural network's loss function. As the 0-60 min blood time activity curve (TAC) required by the standard Patlak plot is unobtainable from shortened dynamic PET scans, a population-based "normalized time" (integral-to-instantaneous blood concentration ratio) was used for the linear fitting of Patlak plot of t* to 60 min, and the modified Patlak plot equation was then incorporated into the neural network. Ki images were generated by minimizing the difference between the input layer (measured tissue-to-blood concentration ratios) and the output layer (predicted tissue-to-blood concentration ratios). The effects of t* (20 to 50 min post injection) on the Ki images generated from the SN-Patlak and standard Patlak was evaluated using the normalized mean square error (NMSE), and Pearson's correlation coefficient (Pearson's r).

RESULTS

The Ki images generated by the SN-Patlak are robust to the dynamic PET scan duration, and the Ki images generated by the SN-Patlak from just a 10-minute (50-60 min post-injection) dynamic [18F]FDG total-body PET scan are comparable to those generated by the standard Patlak method from 40-min (20-60 min post injection) with NMSE = 0.15 ± 0.03 and Pearson's r = 0.93 ± 0.01.

CONCLUSIONS

The SN-Patlak parametric imaging algorithm is robust and reliable for quantification of 10-min dynamic [18F]FDG total-body PET.

Total-Body PET/CT: A Role of Artificial Intelligence?

The purpose of this paper is to provide an overview of the cutting-edge applications of artificial intelligence (AI) technology in total-body positron emission tomography/computed tomography (PET/CT) scanning technology and its profound impact on the field of medical imaging. The introduction of total-body PET/CT scanners marked a major breakthrough in medical imaging, as their superior sensitivity and ultralong axial fields of view allowed for high-quality PET images of the entire body to be obtained in a single scan, greatly enhancing the efficiency and accuracy of diagnoses. However, this advancement is accompanied by the challenges of increasing data volumes and data complexity levels, which pose severe challenges for traditional image processing and analysis methods. Given the excellent ability of AI technology to process massive and high-dimensional data, the combination of AI technology and ultrasensitive PET/CT can be considered a complementary match, opening a new path for rapidly improving the efficiency of the PET-based medical diagnosis process. Recently, AI technology has demonstrated extraordinary potential in several key areas related to total-body PET/CT, including radiation dose reductions, dynamic parametric imaging refinements, quantitative analysis accuracy improvements, and significant image quality enhancements. The accelerated adoption of AI in clinical practice is of particular interest and is directly driven by the rapid progress made by AI technologies in terms of interpretability; i.e., the decision-making processes of algorithms and models have become more transparent and understandable. In the future, we believe that AI technology will fundamentally reshape the use of PET/CT, not only playing a more critical role in clinical diagnoses but also facilitating the customization and implementation of personalized healthcare solutions, providing patients with safer, more accurate, and more efficient healthcare experiences.

Bidirectional dynamic frame prediction network for total-body [68Ga]Ga-PSMA-11 and [68Ga]Ga-FAPI-04 PET images

PURPOSE

Total-body dynamic positron emission tomography (PET) imaging with total-body coverage and ultrahigh sensitivity has played an important role in accurate tracer kinetic analyses in physiology, biochemistry, and pharmacology. However, dynamic PET scans typically entail prolonged durations ([Formula: see text]60 minutes), potentially causing patient discomfort and resulting in artifacts in the final images. Therefore, we propose a dynamic frame prediction method for total-body PET imaging via deep learning technology to reduce the required scanning time.

METHODS

On the basis of total-body dynamic PET data acquired from 13 subjects who received [68Ga]Ga-FAPI-04 (68Ga-FAPI) and 24 subjects who received [68Ga]Ga-PSMA-11 (68Ga-PSMA), we propose a bidirectional dynamic frame prediction network that uses the initial and final 10 min of PET imaging data (frames 1-6 and frames 25-30, respectively) as inputs. The peak signal-to-noise ratio (PSNR) and structural similarity index measure (SSIM) were employed as evaluation metrics for an image quality assessment. Moreover, we calculated parametric images (68Ga-FAPI: [Formula: see text], 68Ga-PSMA: [Formula: see text]) based on the supplemented sequence data to observe the quantitative accuracy of our approach. Regions of interest (ROIs) and statistical analyses were utilized to evaluate the performance of the model.

RESULTS

Both the visual and quantitative results illustrate the effectiveness of our approach. The generated dynamic PET images yielded PSNRs of 36.056 ± 0.709 dB for the 68Ga-PSMA group and 33.779 ± 0.760 dB for the 68Ga-FAPI group. Additionally, the SSIM reached 0.935 ± 0.006 for the 68Ga-FAPI group and 0.922 ± 0.009 for the 68Ga-PSMA group. By conducting a quantitative analysis on the parametric images, we obtained PSNRs of 36.155 ± 4.813 dB (68Ga-PSMA, [Formula: see text]) and 43.150 ± 4.102 dB (68Ga-FAPI, [Formula: see text]). The obtained SSIM values were 0.932 ± 0.041 (68Ga-PSMA) and 0.980 ± 0.011 (68Ga-FAPI). The ROI analysis conducted on our generated dynamic PET sequences also revealed that our method can accurately predict temporal voxel intensity changes, maintaining overall visual consistency with the ground truth.

CONCLUSION

In this work, we propose a bidirectional dynamic frame prediction network for total-body 68Ga-PSMA and 68Ga-FAPI PET imaging with a reduced scan duration. Visual and quantitative analyses demonstrated that our approach performed well when it was used to predict one-hour dynamic PET images. https://github.com/OPMZZZ/BDF-NET .

Advancements of non‐invasive imaging technologies for the diagnosis and staging of liver fibrosis: Present and future

Liver fibrosis is a reparative response triggered by liver injury. Non‐invasive assessment and staging of liver fibrosis in patients with chronic liver disease are of paramount importance, as treatment strategies and prognoses depend significantly on the degree of fibrosis. Although liver fibrosis has traditionally been staged through invasive liver biopsy, this method is prone to sampling errors, particularly when biopsy sizes are inadequate. Consequently, there is an urgent clinical need for an alternative to biopsy, one that ensures precise, sensitive, and non‐invasive diagnosis and staging of liver fibrosis. Non‐invasive imaging assessments have assumed a pivotal role in clinical practice, enjoying growing popularity and acceptance due to their potential for diagnosing, staging, and monitoring liver fibrosis. In this comprehensive review, we first delved into the current landscape of non‐invasive imaging technologies, assessing their accuracy and the transformative impact they have had on the diagnosis and management of liver fibrosis in both clinical practice and animal models. Additionally, we provided an in‐depth exploration of recent advancements in ultrasound imaging, computed tomography imaging, magnetic resonance imaging, nuclear medicine imaging, radiomics, and artificial intelligence within the field of liver fibrosis research. We summarized the key concepts, advantages, limitations, and diagnostic performance of each technique. Finally, we discussed the challenges associated with clinical implementation and offer our perspective on advancing the field, hoping to provide alternative directions for the future research.

Clinical feasibility study of early 30-minute dynamic FDG-PET scanning protocol for patients with lung lesions

PURPOSE

This study aimed to evaluate the clinical feasibility of early 30-minute dynamic 2-deoxy-2-[18F]fluoro-D-glucose (18F-FDG) positron emission tomography (PET) scanning protocol for patients with lung lesions in comparison to the standard 65-minute dynamic FDG-PET scanning as a reference.

METHODS

Dynamic 18F-FDG PET images of 146 patients with 181 lung lesions (including 146 lesions confirmed by histology) were analyzed in this prospective study. Dynamic images were reconstructed into 28 frames with a specific temporal division protocol for the scan data acquired 65 min post-injection. Ki images and quantitative parameters Ki based on two different acquisition durations [the first 30 min (Ki-30 min) and 65 min (Ki-65 min)] were obtained by applying the irreversible two-tissue compartment model using in-house Matlab software. The two acquisition durations were compared for Ki image quality (including visual score analysis and number of lesions detected) and Ki value (including accuracy of Ki, the value of differential diagnosis of lung lesions and prediction of PD-L1 status) by Wilcoxon's rank sum test, Spearman's rank correlation analysis, receiver operating characteristic (ROC) curve, and the DeLong test. The significant testing level (alpha) was set to 0.05.

RESULTS

The quality of the Ki-30 min images was not significantly different from the Ki-65 min images based on visual score analysis (P > 0.05). In terms of Ki value, among 181 lesions, Ki-65 min was statistically higher than Ki-30 min (0.027 ± 0.017 ml/g/min vs. 0.026 ± 0.018 ml/g/min, P < 0.05), while a very high correlation was obtained between Ki-65 min and Ki-30 min (r = 0.977, P < 0.05). In the differential diagnosis of lung lesions, ROC analysis was performed on 146 histologically confirmed lesions, the area under the curve (AUC) of Ki-65 min, Ki-30 min, and SUVmax was 0.816, 0.816, and 0.709, respectively. According to the Delong test, no significant differences in the diagnostic accuracies were found between Ki-65 min and Ki-30 min (P > 0.05), while the diagnostic accuracies of Ki-65 min and Ki-30 min were both significantly higher than that of SUVmax (P < 0.05). In 73 (NSCLC) lesions with definite PD-L1 expression results, the Ki-65 min, Ki-30 min, and SUVmax in PD-L1 positivity were significantly higher than that in PD-L1 negativity (P < 0.05). And no significant differences in predicting PD-L1 positivity were found among Ki-65 min, Ki-30 min, and SUVmax (AUC = 0.704, 0.695, and 0.737, respectively, P > 0.05), according to the results of ROC analysis and Delong test.

CONCLUSIONS

This study indicates that an early 30-minute dynamic FDG-PET acquisition appears to be sufficient to provide quantitative images with good-quality and accurate Ki values for the assessment of lung lesions and prediction of PD-L1 expression. Protocols with a shortened early 30-minute acquisition time may be considered for patients who have difficulty with prolonged acquisitions to improve the efficiency of clinical acquisitions.

Short-axis PET image quality improvement based on a uEXPLORER total-body PET system through deep learning

PURPOSE

The axial field of view (AFOV) of a positron emission tomography (PET) scanner greatly affects the quality of PET images. Although a total-body PET scanner (uEXPLORER) with a large AFOV is more sensitive, it is more expensive and difficult to widely use. Therefore, we attempt to utilize high-quality images generated by uEXPLORER to optimize the quality of images from short-axis PET scanners through deep learning technology while controlling costs.

METHODS

The experiments were conducted using PET images of three anatomical locations (brain, lung, and abdomen) from 335 patients. To simulate PET images from different axes, two protocols were used to obtain PET image pairs (each patient was scanned once). For low-quality PET (LQ-PET) images with a 320-mm AFOV, we applied a 300-mm FOV for brain reconstruction and a 500-mm FOV for lung and abdomen reconstruction. For high-quality PET (HQ-PET) images, we applied a 1940-mm AFOV during the reconstruction process. A 3D Unet was utilized to learn the mapping relationship between LQ-PET and HQ-PET images. In addition, the peak signal-to-noise ratio (PSNR) and structural similarity index measure (SSIM) were employed to evaluate the model performance. Furthermore, two nuclear medicine doctors evaluated the image quality based on clinical readings.

RESULTS

The generated PET images of the brain, lung, and abdomen were quantitatively and qualitatively compatible with the HQ-PET images. In particular, our method achieved PSNR values of 35.41 ± 5.45 dB (p < 0.05), 33.77 ± 6.18 dB (p < 0.05), and 38.58 ± 7.28 dB (p < 0.05) for the three beds. The overall mean SSIM was greater than 0.94 for all patients who underwent testing. Moreover, the total subjective quality levels of the generated PET images for three beds were 3.74 ± 0.74, 3.69 ± 0.81, and 3.42 ± 0.99 (the highest possible score was 5, and the minimum score was 1) from two experienced nuclear medicine experts. Additionally, we evaluated the distribution of quantitative standard uptake values (SUV) in the region of interest (ROI). Both the SUV distribution and the peaks of the profile show that our results are consistent with the HQ-PET images, proving the superiority of our approach.

CONCLUSION

The findings demonstrate the potential of the proposed technique for improving the image quality of a PET scanner with a 320 mm or even shorter AFOV. Furthermore, this study explored the potential of utilizing uEXPLORER to achieve improved short-axis PET image quality at a limited economic cost, and computer-aided diagnosis systems that are related can help patients and radiologists.

Machine learning methods for tracer kinetic modelling

Tracer kinetic modelling based on dynamic PET is an important field of Nuclear Medicine for quantitative functional imaging. Yet, its implementation in clinical routine has been constrained by its complexity and computational costs. Machine learning poses an opportunity to improve modelling processes in terms of arterial input function prediction, the prediction of kinetic modelling parameters and model selection in both clinical and preclinical studies while reducing processing time. Moreover, it can help improving kinetic modelling data used in downstream tasks such as tumor detection. In this review, we introduce the basics of tracer kinetic modelling and present a literature review of original works and conference papers using machine learning methods in this field.

Quantitation of dynamic total-body PET imaging: recent developments and future perspectives

BACKGROUND

Positron emission tomography (PET) scanning is an important diagnostic imaging technique used in disease diagnosis, therapy planning, treatment monitoring, and medical research. The standardized uptake value (SUV) obtained at a single time frame has been widely employed in clinical practice. Well beyond this simple static measure, more detailed metabolic information can be recovered from dynamic PET scans, followed by the recovery of arterial input function and application of appropriate tracer kinetic models. Many efforts have been devoted to the development of quantitative techniques over the last couple of decades.

CHALLENGES

The advent of new-generation total-body PET scanners characterized by ultra-high sensitivity and long axial field of view, i.e., uEXPLORER (United Imaging Healthcare), PennPET Explorer (University of Pennsylvania), and Biograph Vision Quadra (Siemens Healthineers), further stimulates valuable inspiration to derive kinetics for multiple organs simultaneously. But some emerging issues also need to be addressed, e.g., the large-scale data size and organ-specific physiology. The direct implementation of classical methods for total-body PET imaging without proper validation may lead to less accurate results.

CONCLUSIONS

In this contribution, the published dynamic total-body PET datasets are outlined, and several challenges/opportunities for quantitation of such types of studies are presented. An overview of the basic equation, calculation of input function (based on blood sampling, image, population or mathematical model), and kinetic analysis encompassing parametric (compartmental model, graphical plot and spectral analysis) and non-parametric (B-spline and piece-wise basis elements) approaches is provided. The discussion mainly focuses on the feasibilities, recent developments, and future perspectives of these methodologies for a diverse-tissue environment.

Generation of Whole-Body FDG Parametric Ki Images from Static PET Images Using Deep Learning

FDG parametric Ki images show great advantage over static SUV images, due to the higher contrast and better accuracy in tracer uptake rate estimation. In this study, we explored the feasibility of generating synthetic Ki images from static SUV ratio (SUVR) images using three configurations of U-Nets with different sets of input and output image patches, which were the U-Nets with single input and single output (SISO), multiple inputs and single output (MISO), and single input and multiple outputs (SIMO). SUVR images were generated by averaging three 5-min dynamic SUV frames starting at 60 minutes post-injection, and then normalized by the mean SUV values in the blood pool. The corresponding ground truth Ki images were derived using Patlak graphical analysis with input functions from measurement of arterial blood samples. Even though the synthetic Ki values were not quantitatively accurate compared with ground truth, the linear regression analysis of joint histograms in the voxels of body regions showed that the mean R2 values were higher between U-Net prediction and ground truth (0.596, 0.580, 0.576 in SISO, MISO and SIMO), than that between SUVR and ground truth Ki (0.571). In terms of similarity metrics, the synthetic Ki images were closer to the ground truth Ki images (mean SSIM = 0.729, 0.704, 0.704 in SISO, MISO and MISO) than the input SUVR images (mean SSIM = 0.691). Therefore, it is feasible to use deep learning networks to estimate surrogate map of parametric Ki images from static SUVR images.