AI Transformers for Radiation Dose Reduction in Serial Whole-Body PET Scans.

Purpose: To develop a deep learning approach that enables ultra-low-dose, 1% of the standard clinical dosage (3 MBq/kg), ultrafast whole-body PET reconstruction in cancer imaging. Materials and Methods: In this Health Insurance Portability and Accountability Act-compliant study, serial fluorine 18-labeled fluorodeoxyglucose PET/MRI scans of pediatric patients with lymphoma were retrospectively collected from two cross-continental medical centers between July 2015 and March 2020. Global similarity between baseline and follow-up scans was used to develop Masked-LMCTrans, a longitudinal multimodality coattentional convolutional neural network (CNN) transformer that provides interaction and joint reasoning between serial PET/MRI scans from the same patient. Image quality of the reconstructed ultra-low-dose PET was evaluated in comparison with a simulated standard 1% PET image. The performance of Masked-LMCTrans was compared with that of CNNs with pure convolution operations (classic U-Net family), and the effect of different CNN encoders on feature representation was assessed. Statistical differences in the structural similarity index measure (SSIM), peak signal-to-noise ratio (PSNR), and visual information fidelity (VIF) were assessed by two-sample testing with the Wilcoxon signed rank t test. Results: The study included 21 patients (mean age, 15 years ± 7 [SD]; 12 female) in the primary cohort and 10 patients (mean age, 13 years ± 4; six female) in the external test cohort. Masked-LMCTrans-reconstructed follow-up PET images demonstrated significantly less noise and more detailed structure compared with simulated 1% extremely ultra-low-dose PET images. SSIM, PSNR, and VIF were significantly higher for Masked-LMCTrans-reconstructed PET (P < .001), with improvements of 15.8%, 23.4%, and 186%, respectively. Conclusion: Masked-LMCTrans achieved high image quality reconstruction of 1% low-dose whole-body PET images.Keywords: Pediatrics, PET, Convolutional Neural Network (CNN), Dose Reduction Supplemental material is available for this article. © RSNA, 2023.

Improved Detection of Bone Metastases in Children and Young Adults with Ferumoxytol-enhanced MRI.

Purpose To evaluate if ferumoxytol can improve the detection of bone marrow metastases at diffusion-weighted (DW) MRI in pediatric and young adult patients with cancer. Materials and Methods In this secondary analysis of a prospective institutional review board-approved study (ClinicalTrials.gov identifier NCT01542879), 26 children and young adults (age range: 2-25 years; 18 males) underwent unenhanced or ferumoxytol-enhanced whole-body DW MRI between 2015 and 2020. Two reviewers determined the presence of bone marrow metastases using a Likert scale. One additional reviewer measured signal-to-noise ratios (SNRs) and tumor-to-bone marrow contrast. Fluorine 18 (18F) fluorodeoxyglucose (FDG) PET and follow-up chest CT, abdominal and pelvic CT, and standard (non-ferumoxytol enhanced) MRI served as the reference standard. Results of different experimental groups were compared using generalized estimation equations, Wilcoxon rank sum test, and Wilcoxon signed rank test. Results The SNR of normal bone marrow was significantly lower at ferumoxytol-enhanced MRI compared with unenhanced MRI at baseline (21.380 ± 19.878 vs 102.621 ± 94.346, respectively; P = .03) and after chemotherapy (20.026 ± 7.664 vs 54.110 ± 48.022, respectively; P = .006). This led to an increased tumor-to-marrow contrast on ferumoxytol-enhanced MRI scans compared with unenhanced MRI scans at baseline (1397.474 ± 938.576 vs 665.364 ± 440.576, respectively; P = .07) and after chemotherapy (1099.205 ± 864.604 vs 500.758 ± 439.975, respectively; P = .007). Accordingly, the sensitivity and diagnostic accuracy for detecting bone marrow metastases were 96% (94 of 98) and 99% (293 of 297), respectively, with the use of ferumoxytol-enhanced MRI compared with 83% (106 of 127) and 95% (369 of 390) with the use of unenhanced MRI. Conclusion Use of ferumoxytol helped improve the detection of bone marrow metastases in children and young adults with cancer. Keywords: Pediatrics, Molecular Imaging-Cancer, Molecular Imaging-Nanoparticles, MR-Diffusion Weighted Imaging, MR Imaging, Skeletal-Appendicular, Skeletal-Axial, Bone Marrow, Comparative Studies, Cancer Imaging, Ferumoxytol, USPIO © RSNA, 2023 ClinicalTrials.gov registration no. NCT01542879 See also the commentary by Holter-Chakrabarty and Glover in this issue.

Low-count whole-body PET/MRI restoration: an evaluation of dose reduction spectrum and five state-of-the-art artificial intelligence models.

PURPOSE: To provide a holistic and complete comparison of the five most advanced AI models in the augmentation of low-dose 18F-FDG PET data over the entire dose reduction spectrum. METHODS: In this multicenter study, five AI models were investigated for restoring low-count whole-body PET/MRI, covering convolutional benchmarks - U-Net, enhanced deep super-resolution network (EDSR), generative adversarial network (GAN) - and the most cutting-edge image reconstruction transformer models in computer vision to date - Swin transformer image restoration network (SwinIR) and EDSR-ViT (vision transformer). The models were evaluated against six groups of count levels representing the simulated 75%, 50%, 25%, 12.5%, 6.25%, and 1% (extremely ultra-low-count) of the clinical standard 3 MBq/kg 18F-FDG dose. The comparisons were performed upon two independent cohorts - (1) a primary cohort from Stanford University and (2) a cross-continental external validation cohort from Tübingen University - in order to ensure the findings are generalizable. A total of 476 original count and simulated low-count whole-body PET/MRI scans were incorporated into this analysis. RESULTS: For low-count PET restoration on the primary cohort, the mean structural similarity index (SSIM) scores for dose 6.25% were 0.898 (95% CI, 0.887-0.910) for EDSR, 0.893 (0.881-0.905) for EDSR-ViT, 0.873 (0.859-0.887) for GAN, 0.885 (0.873-0.898) for U-Net, and 0.910 (0.900-0.920) for SwinIR. In continuation, SwinIR and U-Net's performances were also discreetly evaluated at each simulated radiotracer dose levels. Using the primary Stanford cohort, the mean diagnostic image quality (DIQ; 5-point Likert scale) scores of SwinIR restoration were 5 (SD, 0) for dose 75%, 4.50 (0.535) for dose 50%, 3.75 (0.463) for dose 25%, 3.25 (0.463) for dose 12.5%, 4 (0.926) for dose 6.25%, and 2.5 (0.534) for dose 1%. CONCLUSION: Compared to low-count PET images, with near-to or nondiagnostic images at higher dose reduction levels (up to 6.25%), both SwinIR and U-Net significantly improve the diagnostic quality of PET images. A radiotracer dose reduction to 1% of the current clinical standard radiotracer dose is out of scope for current AI techniques.

Detection of bone marrow metastases in children and young adults with solid cancers with diffusion-weighted MRI.

OBJECTIVE: To compare the diagnostic accuracy of diffusion-weighted (DW)-MRI with b-values of 50 s/mm2 and 800 s/mm2 for the detection of bone marrow metastases in children and young adults with solid malignancies. METHODS: In an institutional review board-approved prospective study, we performed 51 whole-body DW-MRI scans in 19 children and young adults (14 males, 5 females; age range: 1-25 years) with metastasized cancers before (n = 19 scans) and after (n = 32 scans) chemotherapy. Two readers determined the presence of focal bone marrow lesions in 10 anatomical areas. A third reader measured ADC and SNR of focal lesions and normal marrow. Simultaneously acquired 18F-FDG-PET scans served as the standard of reference. Data of b = 50 s/mm2 and 800 s/mm2 images were compared with the Wilcoxon signed-rank test. Inter-reader agreement was evaluated with weighted kappa statistics. RESULTS: The SNR of bone marrow metastases was significantly higher compared to normal bone marrow on b = 50 s/mm2 (mean ± SD: 978.436 ± 1239.436 vs. 108.881 ± 109.813, p < 0.001) and b = 800 s/mm2 DW-MRI (499.638 ± 612.721 vs. 86.280 ± 89.120; p < 0.001). On 30 out of 32 post-treatment DW-MRI scans, reconverted marrow demonstrated low signal with low ADC values (0.385 × 10-3 ± 0.168 × 10-3mm2/s). The same number of metastases (556/588; 94.6%; p > 0.99) was detected on b = 50 s/mm2 and 800 s/mm2 images. However, both normal marrow and metastases exhibited low signals on ADC maps, limiting the ability to delineate metastases. The inter-reader agreement was substantial, with a weighted kappa of 0.783 and 0.778, respectively. CONCLUSION: Bone marrow metastases in children and young adults can be equally well detected on b = 50 s/mm2 and 800 s/mm2 images, but ADC values can be misleading.

Mechanoporation enables rapid and efficient radiolabeling of stem cells for PET imaging.

Regenerative medicine uses the patient own stem cells to regenerate damaged tissues. Molecular imaging techniques are commonly used to image the transplanted cells, either right after surgery or at a later time. However, few techniques are fast or straightforward enough to label cells intraoperatively. Adipose tissue-derived stem cells (ADSCs) were harvested from knee joints of minipigs. The cells were labeled with PET contrast agent by flowing mechanoporation using a microfluidic device. While flowing through a series of microchannels, cells are compressed repeatedly by micro-ridges, which open transient pores in their membranes and induce convective transport, intended to facilitate the transport of 68Ga-labeled and lipid-coated mesoporous nanoparticles (MSNs) into the cells. This process enables cells to be labeled in a matter of seconds. Cells labeled with this approach were then implanted into cartilage defects, and the implant was imaged using positron emission tomography (PET) post-surgery. The microfluidic device can efficiently label millions of cells with 68Ga-labeled MSNs in as little as 15 min. The method achieved labeling efficiency greater than 5 Bq/cell on average, comparable to 30 min-long passive co-incubation with 68Ga-MSNs, but with improved biocompatibility due to the reduced exposure to ionizing radiation. Labeling time could also be accelerated by increasing throughput through more parallel channels. Finally, as a proof of concept, ADSCs were labeled with 68Ga-MSNs and quantitatively assessed using clinical PET/MR in a mock transplant operation in pig knee joints. MSN-assisted mechanoporation is a rapid, effective and straightforward approach to label cells with 68Ga. Given its high efficiency, this labeling method can be used to track small cells populations without significant effects on viability. The system is applicable to a variety of cell tracking studies for cancer therapy, regenerative therapy, and immunotherapy.

Diagnostic Accuracy of 2-[18F]FDG-PET and whole-body DW-MRI for the detection of bone marrow metastases in children and young adults.

OBJECTIVES: To compare the diagnostic accuracy of 2-[18F]fluoro-2-deoxy-D-glucose-enhanced positron emission tomography (2-[18F]FDG-PET) and diffusion-weighted magnetic resonance imaging (DW-MRI) for the detection of bone marrow metastases in children and young adults with solid malignancies. METHODS: In this cross-sectional single-center institutional review board-approved study, we investigated twenty-three children and young adults (mean age, 16.8 years ± 5.1 [standard deviation]; age range, 7-25 years; 16 males, 7 females) with 925 bone marrow metastases who underwent 66 simultaneous 2-[18F]FDG-PET and DW-MRI scans including 23 baseline scans and 43 follow-up scans after chemotherapy between May 2015 and July 2020. Four reviewers evaluated all foci of bone marrow metastasis on 2-[18F]FDG-PET and DW-MRI to assess concordance and measured the tumor-to-bone marrow contrast. Results were assessed with a one-sample Wilcoxon test and generalized estimation equation. Bone marrow biopsies and follow-up imaging served as the standard of reference. RESULTS: The reviewers detected 884 (884/925, 95.5%) bone marrow metastases on 2-[18F]FDG-PET and 893 (893/925, 96.5%) bone marrow metastases on DW-MRI. We found different "blind spots" for 2-[18F]FDG-PET and MRI: 2-[18F]FDG-PET missed subcentimeter lesions while DW-MRI missed lesions in small bones. Sensitivity and specificity were 91.0% and 100% for 18F-FDG-PET, 89.1% and 100.0% for DW-MRI, and 100.0% and 100.0% for combined modalities, respectively. The diagnostic accuracy of combined 2-[18F]FDG-PET/MRI (100.0%) was significantly higher compared to either 2-[18F]FDG-PET (96.9%, p < 0.001) or DW-MRI (96.3%, p < 0.001). CONCLUSIONS: Both 2-[18F]FDG-PET and DW-MRI can miss bone marrow metastases. The combination of both imaging techniques detected significantly more lesions than either technique alone. KEY POINTS: • DW-MRI and 2-[18F]FDG-PET have different strengths and limitations for the detection of bone marrow metastases in children and young adults with solid tumors. • Both modalities can miss bone marrow metastases, although the "blind spot" of each modality is different. • A combined PET/MR imaging approach will achieve maximum sensitivity and specificity for the detection of bone marrow metastases in children with solid tumors.

Ascorbic Acid and Iron Supplement Treatment Improves Stem Cell-Mediated Cartilage Regeneration in a Minipig Model.

BACKGROUND: The transplantation of mesenchymal stem cells (MSCs) into cartilage defects has led to variable cartilage repair outcomes. Previous in vitro studies have shown that ascorbic acid and reduced iron independently can improve the chondrogenic differentiation of MSCs. However, the combined effect of ascorbic acid and iron supplementation on MSC differentiation has not been investigated. PURPOSE: To investigate the combined in vivo effects of ascorbic acid and a US Food and Drug Administration (FDA)-approved iron supplement on MSC-mediated cartilage repair in mature Göttingen minipigs. STUDY DESIGN: Controlled laboratory study. METHODS: We pretreated bone marrow-derived MSCs with ascorbic acid and the FDA-approved iron supplement ferumoxytol and then transplanted the MSCs into full-thickness cartilage defects in the distal femurs of Göttingen minipigs. Untreated cartilage defects served as negative controls. We evaluated the cartilage repair site with magnetic resonance imaging at 4 and 12 weeks after MSC implantation, followed by histological examination and immunofluorescence staining at 12 weeks. RESULTS: Ascorbic acid plus iron-pretreated MSCs demonstrated a significantly better MOCART (magnetic resonance observation of cartilage repair tissue) score (73.8 ± 15.5), better macroscopic cartilage regeneration score according to the International Cartilage Repair Society (8.6 ± 2.0), better Pineda score (2.9 ± 0.8), and larger amount of collagen type II (28,469 ± 21,313) compared with untreated controls (41.3 ± 2.5, 1.8 ± 2.9, 12.8 ± 1.9, and 905 ± 1326, respectively). The obtained scores were also better than scores previously reported in the same animal model for MSC implants without ascorbic acid. CONCLUSION: Pretreatment of MSCs with ascorbic acid and an FDA-approved iron supplement improved the chondrogenesis of MSCs and led to hyaline-like cartilage regeneration in the knee joints of minipigs. CLINICAL RELEVANCE: Ascorbic acid and iron supplements are immediately clinically applicable. Thus, these results, in principle, could be translated into clinical applications.

Brain iron deposition after Ferumoxytol-enhanced MRI: A study of Porcine Brains.

Recent evidence of gadolinium deposition in the brain has raised safety concerns. Iron oxide nanoparticles are re-emerging as promising alternative MR contrast agents, because the iron core can be metabolized. However, long-term follow up studies of the brain after intravenous iron oxide administration have not been reported thus far. In this study, we investigated, if intravenously administered ferumoxytol nanoparticles are deposited in porcine brains. Methods: In an animal care and use committee-approved prospective case-control study, ten Göttingen minipigs received either intravenous ferumoxytol injections at a dose of 5 mg Fe/kg (n=4) or remained untreated (n=6). Nine to twelve months later, pigs were sacrificed and the brains of all pigs underwent ex vivo MRI at 7T with T2 and T2*-weighted sequences. MRI scans were evaluated by measuring R2* values (R2*=1000/T2*) of the bilateral caudate nucleus, lentiform nucleus, thalamus, dentate nucleus, and choroid plexus. Pig brains were sectioned and stained with Prussian blue and evaluated for iron deposition using a semiquantitative scoring system. Data of ferumoxytol exposed and unexposed groups were compared with an unpaired t-test and a Mann-Whitney U test. Results: T2 and T2* signal of the different brain regions was not visually different between ferumoxytol exposed and unexposed controls. There were no significant differences in R2* values of the different brain regions in the ferumoxytol exposed group compared to controls (p>0.05). Prussian blue stains of the same brain regions, scored according to a semiquantitative score, were not significantly different either between the ferumoxytol exposed group and unexposed controls (p>0.05). Conclusions: Our study shows that intravenous ferumoxytol doses of 5-10 mg Fe/kg do not lead to iron deposition in the brain of pigs. We suggest iron oxide nanoparticles as a promising alternative for gadolinium-enhanced MRI.

Differentiation of benign and malignant lymph nodes in pediatric patients on ferumoxytol-enhanced PET/MRI.

The composition of lymph nodes in pediatric patients is different from that in adults. Most notably, normal lymph nodes in children contain less macrophages. Therefore, previously described biodistributions of iron oxide nanoparticles in benign and malignant lymph nodes of adult patients may not apply to children. The purpose of our study was to evaluate if the iron supplement ferumoxytol improves the differentiation of benign and malignant lymph nodes in pediatric cancer patients on 18F-FDG PET/MRI. Methods: We conducted a prospective clinical trial from May 2015 to December 2018 to investigate the value of ferumoxytol nanoparticles for staging of children with cancer with 18F-FDG PET/MRI. Ferumoxytol is an FDA-approved iron supplement for the treatment of anemia and has been used "off-label" as an MRI contrast agent in this study. Forty-two children (7-18 years, 29 male, 13 female) received a 18F-FDG PET/MRI at 2 (n=20) or 24 hours (h) (n=22) after intravenous injection of ferumoxytol (dose 5 mg Fe/kg). The morphology of benign and malignant lymph nodes on ferumoxytol-enhanced T2-FSE sequences at 2 and 24 h were compared using a linear regression analysis. In addition, ADCmean-values, SUV-ratio (SUVmax lesion/SUVmean liver) and R2*-relaxation rate of benign and malignant lymph nodes were compared with a Mann-Whitney-U test. The accuracy of different criteria was assessed with a receiver operating characteristics (ROC) curve. Follow-up imaging for at least 6 months served as the standard of reference. Results: We examined a total of 613 lymph nodes, of which 464 (75.7%) were benign and 149 (24.3%) were malignant. On ferumoxytol-enhanced T2-FSE images, benign lymph nodes showed a hypointense hilum and hyperintense parenchyma, while malignant lymph nodes showed no discernible hilum. This pattern was not significantly different at 2 h and 24 h postcontrast (p=0.82). Benign and malignant lymph nodes showed significantly different ferumoxytol enhancement patterns, ADCmean values of 1578 and 852 x10-6 mm2/s, mean SUV-ratios of 0.5 and 2.8, and mean R2*-relaxation rate of 127.8 and 84.4 Hertz (Hz), respectively (all p<0.001). The accuracy of ADCmean, SUV-ratio and pattern (area under the curve (AUC): 0.99; 0.98; 0.97, respectively) was not significantly different (p=0.07). Compared to these three parameters, the accuracy of R2* was significantly lower (AUC: 0.93; p=0.001). Conclusion: Lymph nodes in children show different ferumoxytol-enhancement patterns on MRI than previously reported for adult patients. We found high accuracy (>90%) of ADCmean, SUV-ratio, pattern, and R2* measurements for the characterization of benign and malignant lymph nodes in children. Ferumoxytol nanoparticle accumulation at the hilum can be used to diagnose a benign lymph node. In the future, the delivery of clinically applicable nanoparticles to the hilum of benign lymph nodes could be harnessed to deliver theranostic drugs for immune cell priming.

[From Xrays to PET/MR, and then? - Future imaging in pediatric radiology]. / Vom Röntgen zum PET/MRT, und dann? ­ Zukunftsweisende Bildgebung in der Kinderradiologie.

Significant changes can be expected in modern pediatric radiology. New imaging techniques are progressively added to basic modalities like Xrays and ultrasound. This essay summarizes recent advances and technical innovations in pediatric radiology, which are supposed to gain further importance in the future. Thus, CT dose reduction techniques including artificial intelligence as well as advances in the fields of magnetic resonance and molecular imaging are presented. KEY POINTS: · Technical innovations will lead to significant changes in pediatric radiology.. · CT dose reduction is crucial for pediatric patient collectives.. · New MR-techniques will lower the need for sedation and contrast media application.. · Functional MR-imaging might gain further importance in patients with chronic lung disease.. · Molecular imaging enables detection, characterization and quantification of molecular processes in tumors.. CITATION FORMAT: · Staatz G, Daldrup-Link HE, Herrmann J et al. From Xrays to PET/MR, and then? - Future imaging in pediatric radiology. Fortschr Röntgenstr 2019; 191: 357 - 366.

Bone marrow oedema predicts bone collapse in paediatric and adolescent leukaemia patients with corticosteroid-induced osteonecrosis.

OBJECTIVES: Corticosteroid treatment of paediatric leukaemia patients can lead to osteonecrosis (ON). We determined whether bone marrow oedema (BME) is an early sign of progressive ON and eventual bone collapse. METHODS: In a retrospective study, two radiologists reviewed MR imaging characteristics of 47 early stage epiphyseal ON in 15 paediatric and adolescent leukaemia patients. Associations between BME on initial imaging studies and subchondral fracture, disease progression and bone collapse were assessed by Cochran-Mantel-Haenszel tests. Differences in time to progression and bone collapse between lesions with and without oedema were assessed by log rank tests. RESULTS: Forty-seven occurrences of ON were located in weight bearing joints, with 77% occurring in the femur. Seventeen lesions progressed to collapse, two lesions worsened without collapse, and 28 remained stable or improved. BME was significantly associated with subchondral fracture (p = 0.0014), disease progression (p = 0.0015), and bone collapse (p < 0.001), with a sensitivity and specificity of 94% and 77%, respectively, for bone collapse. Time to progression for ON with oedema was 2.7 years (95% CI: 1.7-3.4); while the majority of no-oedema ON were stable (p = 0.0011). CONCLUSIONS: BME is an early sign of progressive ON and eventual bone collapse in paediatric and adolescent leukaemia patients. KEY POINTS: • Bone marrow oedema in corticosteroid-induced osteonecrosis predicts progression to bone collapse. • Bone marrow oedema is associated with subchondral fractures in corticosteroid-induced osteonecrosis. • Bone marrow oedema can be used to stratify patients to joint-preserving interventions. • Absence of bone marrow oedema can justify a "wait and watch" approach.

On the evaluation of segmentation editing tools.

Efficient segmentation editing tools are important components in the segmentation process, as no automatic methods exist that always generate sufficient results. Evaluating segmentation editing algorithms is challenging, because their quality depends on the user's subjective impression. So far, no established methods for an objective, comprehensive evaluation of such tools exist and, particularly, intermediate segmentation results are not taken into account. We discuss the evaluation of editing algorithms in the context of tumor segmentation in computed tomography. We propose a rating scheme to qualitatively measure the accuracy and efficiency of editing tools in user studies. In order to objectively summarize the overall quality, we propose two scores based on the subjective rating and the quantified segmentation quality over time. Finally, a simulation-based evaluation approach is discussed, which allows a more reproducible evaluation without the need for human input. This automated evaluation complements user studies, allowing a more convincing evaluation, particularly during development, where frequent user studies are not possible. The proposed methods have been used to evaluate two dedicated editing algorithms on 131 representative tumor segmentations. We show how the comparison of editing algorithms benefits from the proposed methods. Our results also show the correlation of the suggested quality score with the qualitative ratings.