Nuclear Medicine Physicians Research Articles

Improved Accuracy and Reliability of PRIMARY Scoring Using Delayed [68Ga] Ga-PSMA PET/CT Imaging.

Delayed [68Ga]Ga-prostate-specific membrane antigen (PSMA) positron emission tomography/computed tomography (PET/CT) images show reduced PSMA uptake in benign lesions and increased PSMA uptake in malignant lesions. This study investigated the efficacy of PRIMARY scoring on [68Ga]Ga-PSMA PET/CT images at standard versus delayed time points and assessed the potential added value of delayed imaging in PRIMARY scoring. A total of 140 patients with biopsy results of International Society of Urological Pathology grade groups (ISUP) 1-2 who had standard (median 60 min) and delayed images (median 138 min) with [68Ga]Ga-PSMA PET/CT before radical prostatectomy were included. Results were confirmed in pathological reports. For diagnostic parameters, two experienced nuclear medicine physicians, who were blinded to clinical data, independently reviewed the images, and a third physician provided consensus in cases of disagreement. PRIMARY scoring was also conducted by four nuclear medicine physicians on both images, with a 1-month interval between assessments for intraobserver agreement analyses. The percentage of lesions scored as 1-2 in PRIMARY scoring decreased from 29% to 10% in delayed images compared with standard images, whereas lesions scored as 3-5 increased from 71% to 90%. Additionally, agreement between two experienced nuclear medicine physicians regarding scoring was 66% for standard imaging and 77% for delayed imaging. The number of patients with PRIMARY score 5 increased from 31 to 46 in delayed imaging. All patients were confirmed to have clinically significant prostate cancer (csPCa). Furthermore, no csPCa of ISUP grade 3 or higher was detected in patients with a delayed PRIMARY score (dPRIMARY). The sensitivity of standard PRIMARY scoring was 71%, which increased to 92% with dPRIMARY scoring, with a consistent positive predictive value of 87% for both. Intraobserver agreement Cohen's kappa values for all observers were higher for delayed images than for standard images. Inter-observer agreement, assessed by Fleiss kappa, was 0.47 and 0.52 for standard images in rounds 1 and 2, respectively, and 0.61 and 0.72 for delayed images, respectively. Decreased background activity and increased primary tumor uptake in delayed images improved differentiation between primary tumors and benign lesions, leading to better primary tumor identification. Enhanced reliability was also observed in both intraobserver and interobserver assessments of delayed images.

A Machine Learning Model Based on Radiomic Features as a Tool to Identify Active Giant Cell Arteritis on [18F]FDG-PET Images During Follow-Up

Objective: To investigate the feasibility of a machine learning (ML) model based on radiomic features to identify active giant cell arteritis (GCA) in the aorta and differentiate it from atherosclerosis in follow-up [18F]FDG-PET/CT images for therapy monitoring. Methods: To train the ML model, 64 [18F]FDG-PET scans of 34 patients with proven GCA and 34 control subjects with type 2 diabetes mellitus were retrospectively included. The aorta was delineated into the ascending, arch, descending, and abdominal aorta. From each segment, 95 features were extracted. All segments were randomly split into a training/validation (n = 192; 80%) and test set (n = 46; 20%). In total, 441 ML models were trained, using combinations of seven feature selection methods, seven classifiers, and nine different numbers of features. The performance was assessed by area under the curve (AUC). The best performing ML model was compared to the clinical report of nuclear medicine physicians in 19 follow-up scans (7 active GCA, 12 inactive GCA). For explainability, an occlusion map was created to illustrate the important regions of the aorta for the decision of the ML model. Results: The ten-feature model with ANOVA as the feature selector and random forest classifier demonstrated the highest performance (AUC = 0.92 ± 0.01). Compared with the clinical report, this model showed a higher PPV (0.83 vs. 0.80), NPV (0.85 vs. 0.79), and accuracy (0.84 vs. 0.79) in the detection of active GCA in follow-up scans. Conclusions: The current radiomics ML model was able to identify active GCA and differentiate GCA from atherosclerosis in follow-up [18F]FDG-PET/CT scans. This demonstrates the potential of the ML model as a monitoring tool in challenging [18F]FDG-PET scans of GCA patients.

Leveraging small voxel with optimal acquisition time for [18F]mFBG total-body PET/CT imaging in pediatric patients with neuroblastoma: a preliminary study.

The advent of total-body PET/CT presents an opportunity for significant advancements in imaging of neuroblastoma with [18F]meta-fluorobenzylguanidine ([18F]mFBG). Small voxel imaging has proven to have better lesion detectability but need enough radioactivity counts. This study aims to balance shortened acquisition times and small voxel reconstruction to keep sufficient image quality and diagnostic confidence on [18F]mFBG total-body PET for neuroblastoma. We retrospectively enrolled 33 pediatric patients with neuroblastoma who underwent 37 [18F]mFBG total-body uEXPLORER PET/CT scans of 10-min duration. PET images were reconstructed with varying acquisition times (0.5-10min) and three matrix sizes (192 × 192, 512 × 512 and 1024 × 1024). The subjective (scored on a 5-point scale) and objective image quality (signal-to-noise ratio, SNR) of all the sets of reconstructed images were analyzed by nuclear medicine physicians. For indeterminate lesions identified in the group of 192 × 192 matrix with the 10-min scan (G192-10), diagnostic confidence was further evaluated in images reconstructed with the 512 × 512 and 1024 × 1024 matrices (G512 and G1024). Of the 33 patients with 37 [18F]mFBG PET/CT scans, 17 patients with 20 scans had positive [18F]mFBG PET/CT findings. Sufficient subjective image quality was achieved with at least 2-min acquisition of 192 × 192 matrix and 4-min acquisition of 512 × 512 matrix (with all scores ≥ 3). SNR increased with longer acquisition times for the same voxel size, while decreased as voxel size shrunk. Although the Curie and SIOPEN scores remained consistent across G192, G512, and G1024-10 groups, the G512 groups with at least 2-min acquisition and G1024-10 showed significantly higher confidence scores for characterizing indeterminate lesions on the G192-10 images, with almost all indeterminate lesions being rated as very confident. A matrix of 512 × 512 with a minimum of 4-min acquisition on [18F]mFBG total-body PET/CT is recommended for sufficient image quality and improved diagnostic confidence, particularly in detecting indeterminate lesions.

Impact of tissue-independent positron range correction on [68Ga]Ga-DOTATOC and [68Ga]Ga-PSMA PET image reconstructions: a patient data study.

The positron range effect can impair PET image quality of Gallium-68 (68Ga). A positron range correction (PRC) can be applied to reduce this effect. In this study, the effect ofa tissue-independent PRC for 68Ga was investigated on patient data. PET/CT data (40 patients: [68Ga]Ga-DOTATOC or [68Ga]Ga-PSMA) were reconstructed using Q.Clear reconstruction algorithm. Two reconstructions were performed per patient, Q.Clear with and without PRC. SUVmax and contrast-to-noise ratio (CNR) values per lesion were compared between PRC and non-PRC images. Five experienced nuclear medicine physicians reviewed the images and chose the preferred reconstruction based on the image quality, lesion detectability, and diagnostic confidence. A total of 155 lesions were identified. The PRC resulted in statistically significant increase of the SUVmax and CNR for soft tissue lesions (6.4%, p < 0.001; 8.6%, p < 0.001), bone lesions (14.6%, p < 0.001; 12.5%, p < 0.001), and lung lesions (3.6%, p = 0.010; 6.3%, p = 0.001). This effect was most prominent in small lesions (SUVmax: 12.0%, p < 0.001, and CNR: 13.0%, p < 0.001). Similar or better image quality, lesion detectability, and diagnostic confidence was achieved in PRC images compared to the non-PRC images as those assessed by the expert readers. A tissue-independent PRC increased the SUVmax and CNR in soft tissue, bone, and lung lesions with a larger effect for the small lesions. Visual assessment demonstrated similar or better image quality, lesion detectability, and diagnostic confidence in PRC images compared to the non-PRC images.

Empowering PET imaging reporting with retrieval-augmented large language models and reading reports database: a pilot single center study.

The potential of Large Language Models (LLMs) in enhancing a variety of natural language tasks in clinical fields includes medical imaging reporting. This pilot study examines the efficacy of a retrieval-augmented generation (RAG) LLM system considering zero-shot learning capability of LLMs, integrated with a comprehensive database of PET reading reports, in improving reference to prior reports and decision making. We developed a custom LLM framework with retrieval capabilities, leveraging a database of over 10 years of PET imaging reports from a single center. The system uses vector space embedding to facilitate similarity-based retrieval. Queries prompt the system to generate context-based answers and identify similar cases or differential diagnoses. From routine clinical PET readings, experienced nuclear medicine physicians evaluated the performance of system in terms of the relevance of queried similar cases and the appropriateness score of suggested potential diagnoses. The system efficiently organized embedded vectors from PET reports, showing that imaging reports were accurately clustered within the embedded vector space according to the diagnosis or PET study type. Based on this system, a proof-of-concept chatbot was developed and showed the framework's potential in referencing reports of previous similar cases and identifying exemplary cases for various purposes. From routine clinical PET readings, 84.1% of the cases retrieved relevant similar cases, as agreed upon by all three readers. Using the RAG system, the appropriateness score of the suggested potential diagnoses was significantly better than that of the LLM without RAG. Additionally, it demonstrated the capability to offer differential diagnoses, leveraging the vast database to enhance the completeness and precision of generated reports. The integration of RAG LLM with a large database of PET imaging reports suggests the potential to support clinical practice of nuclear medicine imaging reading by various tasks of AI including finding similar cases and deriving potential diagnoses from them. This study underscores the potential of advanced AI tools in transforming medical imaging reporting practices.

Diagnostic value of 99mTc-MDP single-photon emission computed tomography/computed tomography in medication-related osteonecrosis of the jaw: A single-center retrospective study of 39 patients.

The focus of this study was to assess the diagnostic effectiveness of 99mTc-MDP single-photon emission computed tomography/computed tomography (SPECT/CT) in pinpointing medication-related osteonecrosis of the jaw (MRONJ). We gathered a total of 39 patients who were suspected of having MRONJ and displayed radionuclide concentration in the jaws on whole-body bone scans. These patients were subjected to simultaneous SPECT/CT fusion imaging of the jaws during delayed imaging. A consistent team of nuclear medicine physicians and oral and maxillofacial surgeons carried out diagnostic evaluations using SPECT/CT image analysis and dental specialist examinations, and the diagnoses were documented. Based on the SPECT/CT image analysis and dental specialist examination outcomes, we calculated the sensitivity, specificity, accuracy, positive predictive value, and negative predictive value of 99mTc-MDP SPECT/CT in diagnosing suspected MRONJ. Further, we enumerated the number and proportion of patients at each stage. Our findings suggest that 99mTc-MDP SPECT/CT demonstrates high diagnostic precision in patients suspected of MRONJ who display radionuclide concentration in the jaw on whole-body bone scans. Moreover, it proves beneficial in identifying the early stages of MRONJ.

Impact of 68Ga-FAPI positron emission tomography/computed tomography on staging and tumor management in patients with gastric cancer

PurposeTo evaluate the added value of additional 68Ga-FAPI PET/CT following CT for primary staging, detection of postoperative recurrence, and management of gastric cancer patients.MethodsWe retrospectively included patients with gastric cancers who underwent contrast-enhanced computed tomography (ceCT), followed by 68Ga-FAPI PET/CT within 30 days. 68Ga-FAPI PET/CT was performed for initial staging or detection of postoperative recurrence. Two nuclear medicine physicians and a radiologist independently decided on imaging-based staging. Pre-68Ga-FAPI PET/CT treatment decisions were made by a simulated tumor board and post-68Ga-FAPI PET/CT decisions were extracted from medical records. We evaluated the impact of 68Ga-FAPI PET/CT with inconsistent new findings based on the initial findings from ceCT and the resulting changes in treatment strategies.ResultsWe included 112 patients, 84 for initial staging and 28 for detection of postoperative recurrence. Compared to CT, 29 new findings in 24 patients were diagnosed as, or ruled out, cancer involvement on 68Ga-FAPI PET/CT. Among the 112 patients, 21 patients (18.8%) experienced changes in stage or postoperative recurrence. Among patients for initial staging, 14 had stage changes, with 10 being upstaged and 4 being downstaged. Among patients for detection of postoperative recurrence, 7 more patients were diagnosed with tumor recurrence. New findings of 68Ga-FAPI PET/CT led to treatment change in 20/112 (17.9%) patients, which was deemed of major change in 19 patients and minor change in 1 patient.Conclusions68Ga-FAPI PET/CT is valuable for precise staging and detection of postoperative recurrence of gastric cancers, and has the potential to influence management.

Borderline Findings in O-(2-[18F]-Fluoroethyl)-l-Tyrosine PET of Patients with Suspected Glioma Relapse: Role in Clinical Practice.

One of the most common clinical indications for amino acid PET using the tracer O-(2-[18F]-fluoroethyl)-l-tyrosine (18F-FET) is the differentiation of tumor relapse from treatment-related changes in patients with gliomas. A subset of patients may present with an uptake of 18F-FET close to recommended threshold values. The goal of this study was to investigate the frequency of borderline cases and the role of quantitative 18F-FET PET parameters in this situation. Methods: We retrospectively identified 439 patients with pretreated gliomas who underwent 18F-FET PET for suspected tumor relapse and in whom the final diagnoses were confirmed by histopathology (n = 175) or clinical course (n = 264). Two experienced nuclear medicine physicians, masked to the final diagnoses, evaluated visually the PET scans by consensus. The findings were classified into 3 categories: clearly positive findings, borderline findings, or clearly negative findings. The diagnostic performance of established 18F-FET PET parameters (i.e., tumor-to-brain ratio [TBR], time-to-peak ratio, slope, intercept) was evaluated separately for these 3 groups using receiver operating characteristics analyses. Results: In the visual analysis, 18F-FET uptake was classified as clearly negative in 67 patients (15%), clearly positive in 234 patients (53%), and borderline in 136 patients (31%), with averaged mean TBR values of 1.5, 2.3, and 1.9, respectively. Receiver operating characteristics analysis showed a high accuracy for TBR values in patients rated as clearly positive or negative in visual rating (area under curve [AUC], 0.84-0.86), whereas the diagnostic performance of TBR values in borderline cases according to visual analysis was significantly lower (AUC, <0.60). Using TBR values ± 10% above or below the cutoff values increased the AUC by approximately 10% (AUC, 0.82-0.84). Conclusion: A considerable number of patients may present with borderline findings in 18F-FET PET. In these patients, quantitative parameters should be used with caution for decision-making. The use of TBR values above or below the range of the cutoff values ±10% may increase the reliability of quantitative parameters to differentiate between tumor relapse and treatment-related changes.

The Pancreatic Hypermetabolic Lesions of Lung Carcinoma Patients in F-18-Fluorodeoxyglucose PET-CT: Correlation with Pathology

Aim: The ratio and results of the hypermetabolic lesions of the pancreas in lung carcinoma patients in FDG PET-CT which is rare, wasn’t sufficiently evaluated previously in the literature. The aim of this study was to determine these lesions’ pathologic correlations. Materials and Methods: The F-18 FDG PET/CT images of the 516 patients with Lung Carcinoma was evaluated retrospectively by an Experienced Nuclear Medicine Physician and correlation of Pathology results were analyzed. Results: The ratio of the pancreatic hypermetabolic lesions were 7% (36 patients). Pathology results revealed small cell lung carcinoma metastasis in 23, adenocarcinoma in 8, squamous cell carcinoma in 1 and atypical carcinoid tumor in 1 and other pathology types in rest of the patients in primary Lung Carcinoma. Five of the pancreatic lesions pathology results could be obtained and three were small cell lung carcinoma, one of them was adenocarcinoma and the squamous cell carcinoma in other one. Conclusion: The hypermetabolic pancreatic lesions in Lung Carcinoma patients in F-18 FDG PET/CT should be evaluated and considered carefully. Since unexpected pathology results could be achieved.

Prospective Head-to-Head Comparison of 18F-PSMA PET/CT and 18F-NaF PET/CT for Assessing Bone Metastases in 160 Patients with Newly Diagnosed High-Risk Prostate Cancer.

Prostate-specific membrane antigen (PSMA) PET/CT is increasingly used for primary staging in prostate cancer (PC), mainly because of its improved accuracy in detecting lymph node metastases compared with conventional imaging. However, the diagnostic benefit of PSMA PET/CT for detecting bone metastases is less well established. This study compares the diagnostic accuracy of 18F-PSMA PET/CT and 18F-NaF PET/CT for detecting bone metastases in patients newly diagnosed with PC. Methods: This prospective study included patients with histologically confirmed high-risk PC. All participants were referred from the department of urology to 18F-NaF PET/CT and underwent 18F-PSMA PET/CT within 3 weeks. Images were reviewed by 2 nuclear medicine physicians unaware of the results of the other imaging modality. Presence or absence of bone metastases and number of metastatic lesions were recorded. A reference standard was established at the patient level based on agreement between the 2 imaging modalities. In cases of concordance, both modalities were deemed correct. In cases of discordance, additional follow-up scans were performed. Diagnostic performance metrics, including sensitivity, specificity, and accuracy, were calculated. Results: In total, 160 participants were included. Sensitivity, specificity, and accuracy for detecting bone metastases at the patient level were 0.98, 0.99, and 0.99, respectively, for 18F-PSMA PET/CT, and 0.91, 1.00, and 0.97, respectively, for 18F-NaF PET/CT. No significant differences were found. The concordance rate of bone metastases between 18F-NaF and 18F-PSMA PET/CT at the patient level was observed in 154 patients (96.3%). 18F-PSMA PET/CT tended to identify more bone metastases per patient than 18F-NaF PET/CT. Conclusion: Both 18F-NaF and 18F-PSMA PET/CT exhibit high diagnostic accuracy for detecting bone metastases in newly diagnosed high-risk PC patients. 18F-PSMA PET/CT may detect additional metastatic lesions compared with 18F-NaF PET/CT. Subsequent 18F-NaF PET/CT may be redundant if no bone metastases are found on 18F-PSMA PET/CT.

Ultra-fast [18F]florbetapir PET imaging using the uMI Panorama PET/CT system

BackgroundThere is a need for faster amyloid PET scans to reduce patients’ discomfort, minimize movement artifacts, and increase throughput. The recently introduced uMI Panorama PET/CT system featuring enhanced spatial resolution and sub-200ps TOF offers the potential for shorter scan duration without sacrificing image quality or efficacy to detect Aβ deposition. The study aims to establish a faster acquisition protocol for [18F]florbetapir PET imaging using digital PET/CT scanner uMI Panorama, while ensuring adequate image quality and amyloid-β (Aβ) detectability comparable to the standard 10-minute scan.MethodsThirty-eight participants (29 Aβ positive and 9 Aβ negative) from a prospective dementia cohort at Peking Union Medical University Hospital underwent routine [18F]florbetapir PET scans using the uMI Panorama PET/CT scanner and a T1-weighted brain MRI scan. List-mode PET data were reconstructed into durations of 10 min, 2 min, 1 min, 45 s, and 30 s (G10min, G2min, G1min, G45s, G30s). Two trained nuclear medicine physicians independently evaluated the image quality using a 5-point scale and provided binary diagnosis. Standardized uptake value ratios (SUVr) of the composite cortex (frontal, lateral parietal, lateral temporal, and cingulate cortices) were calculated to discriminate Aβ status and coefficient of variation assessed objective image quality. Comparisons of image quality and Aβ detectability between various fast scan groups and G10min group were conducted.ResultsThe subjective image quality evaluation and Aβ detectability results from the two physicians showed both good intra-reader and inter-reader agreements (Cohen’s kappa coefficient: 0.759-1.000). The subjective and objective image qualities of the G2min scans were comparable to the G10min scans, whereas adequate image quality was achieved with the G1min and G45s scans (5-point score ≥ 3). Subjective visual diagnosis by two physicians yielded consistent accuracy for G10min, G2min, and G1min groups, but lower specificity for G45s and G30s groups. The objective detection of Aβ status by cortex SUVr across all scan durations maintained perfect discriminatory efficiency and relatively high effect size (Hedge’s G: 2.48–2.54).ConclusionsA 1-min ultra-fast scan is feasible for [18F]florbetapir PET imaging using uMI Panorama PET/CT, while maintaining adequate image quality and Aβ diagnostic efficiency.Clinical trial registrationNCT05023564. Registered September 2022 https://clinicaltrials.gov/search?term=NCT05023564.

Three-dimensional spatial localization and volume estimation of prostate tumors using 18F-PSMA-1007 PET/CT versus multiparametric MRI.

Fluorine-18 prostate-specific membraneantigen-1007 positron emission tomography/computed tomography (18F-PSMA-1007 PET/CT) has been shown to be superior to multiparametric magnetic resonance imaging (MRI) for the locoregional staging of intermediate-risk and high-risk prostate tumors. This study aims to evaluate whether it is also superior in estimating tumor parameters, such as three-dimensional spatial localization and volume. 134 participants underwent 18F-PSMA-1007 PET/CT and MRI prior to radical prostatectomy as part of the validating paired-cohort Next Generation Trial (NCT05141760). MRI, 18F-PSMA-1007 PET/CT, and final pathology were independently assessed by blinded radiologists, nuclear medicine physicians, and pathologists, respectively. Individual tumor nodules were measured in three dimensions and cognitively registered to 38 segment prostate diagrams as per PI-RADSv2.1. Correct spatial localization was compared using McNemar test and estimation of tumor volumes were compared using linear regression and partial F-test. 286 tumor nodules were identified by final histopathology. 18F-PSMA-1007 PET/CT was superior to MRI for correct localization (186 [65.0%] vs 134 [46.9%], p < 0.001) and tumor volume estimation (R2 = 0.545 vs 0.431, p < 0.001). Larger tumors and higher Gleason Grade Group (GGG) were associated with correct localization by 18F-PSMA-1007 PET/CT (OR = 2.05, p < 0.001 for tumor volume and OR = 4.92, p < 0.01 for ≥ GGG3) and MRI (OR = 1.81, p < 0.001 for tumor volume and OR = 11.67, p < 0.001 for ≥ GGG3). 18F-PSMA-1007 PET/CT outperforms MRI for determination of three-dimensional spatial localization and volume of prostate tumors. These findings support the use of 18F-PSMA-1007 PET/CT prior to definitive treatment of localized prostate cancers.

Clinical validity and reproducibility of a visual rating scale for cingulate island sign in a real-world memory clinic: An FDG-PET/MRI study.

Brain [18F]FDG-PET is a supportive biomarker for cognitive impairment in Lewy bodies disease (LBD) showing reduced occipital metabolism and presence of the cingulate island sign (CIS), a relative preservation of posterior cingulate cortex (PCC) metabolism compared with precuneus and cuneus. We assess validation, clinical utility, and reproducibility of a qualitative visual CIS scale in the differential diagnosis with Alzheimer's disease (AD) in a memory clinic setting. Sixty-seven patients were studied: 36 LBD, of whom 30 with dementia (DLB) and 6 with mild cognitive impairment (MCI-LB), and 31 AD (20 typical and 11 atypical presentations). They underwent FDG-PET/MRI scans and were followed for at least 24 months. The visual CIS rating scale was scored by a nuclear medicine physician and a neurologist independently. Qualitative CIS scores were validated with ROI-based semiquantitative FDG analysis. Mean CIS scores were 1.84 ± 1.69 for LBD and 0.9 ± 1.24 for AD (p = 0.001). With a cutoff CIS score ≥2, sensitivity was 0.56, and specificity 0.81 (accuracy 0.67). Positive CIS in patients with AD was due to atypical presentations. Negative CIS in LBD was due to (i) normal FDG-PET in MCI-LB or (ii) marked hypometabolism of both the PCC and cuneus. Visual CIS scores correlated with FDG-uptake (r = 0.45; p < 0.001) and held a high inter-specialists concordance. The visual CIS scale can be successfully scored by different specialists. Lower sensitivity is expected in cases of MCI-LB or dementia due to mixed LBD/AD changes. Specificity may be influenced by the inclusion of atypical AD cases.

Automated Measurement of Effective Radiation Dose by 18F-Fluorodeoxyglucose Positron Emission Tomography/Computed Tomography.

Calculating the radiation dose from CT in 18F-PET/CT examinations poses a significant challenge. The objective of this study is to develop a deep learning-based automated program that standardizes the measurement of radiation doses. The torso CT was segmented into six distinct regions using TotalSegmentator. An automated program was employed to extract the necessary information and calculate the effective dose (ED) of PET/CT. The accuracy of our automated program was verified by comparing the EDs calculated by the program with those determined by a nuclear medicine physician (n = 30). Additionally, we compared the EDs obtained from an older PET/CT scanner with those from a newer PET/CT scanner (n = 42). The CT ED calculated by the automated program was not significantly different from that calculated by the nuclear medicine physician (3.67 ± 0.61 mSv and 3.62 ± 0.60 mSv, respectively, p = 0.7623). Similarly, the total ED showed no significant difference between the two calculation methods (8.10 ± 1.40 mSv and 8.05 ± 1.39 mSv, respectively, p = 0.8957). A very strong correlation was observed in both the CT ED and total ED between the two measurements (r2 = 0.9981 and 0.9996, respectively). The automated program showed excellent repeatability and reproducibility. When comparing the older and newer PET/CT scanners, the PET ED was significantly lower in the newer scanner than in the older scanner (4.39 ± 0.91 mSv and 6.00 ± 1.17 mSv, respectively, p < 0.0001). Consequently, the total ED was significantly lower in the newer scanner than in the older scanner (8.22 ± 1.53 mSv and 9.65 ± 1.34 mSv, respectively, p < 0.0001). We successfully developed an automated program for calculating the ED of torso 18F-PET/CT. By integrating a deep learning model, the program effectively eliminated inter-operator variability.

Uncertainty-aware automatic TNM staging classification for [18F] Fluorodeoxyglucose PET-CT reports for lung cancer utilising transformer-based language models and multi-task learning

Background[18F] Fluorodeoxyglucose (FDG) PET-CT is a clinical imaging modality widely used in diagnosing and staging lung cancer. The clinical findings of PET-CT studies are contained within free text reports, which can currently only be categorised by experts manually reading them. Pre-trained transformer-based language models (PLMs) have shown success in extracting complex linguistic features from text. Accordingly, we developed a multi-task ‘TNMu’ classifier to classify the presence/absence of tumour, node, metastasis (‘TNM’) findings (as defined by The Eight Edition of TNM Staging for Lung Cancer). This is combined with an uncertainty classification task (‘u’) to account for studies with ambiguous TNM status.Methods2498 reports were annotated by a nuclear medicine physician and split into train, validation, and test datasets. For additional evaluation an external dataset (n = 461 reports) was created, and annotated by two nuclear medicine physicians with agreement reached on all examples. We trained and evaluated eleven publicly available PLMs to determine which is most effective for PET-CT reports, and compared multi-task, single task and traditional machine learning approaches.ResultsWe find that a multi-task approach with GatorTron as PLM achieves the best performance, with an overall accuracy (all four tasks correct) of 84% and a Hamming loss of 0.05 on the internal test dataset, and 79% and 0.07 on the external test dataset. Performance on the individual TNM tasks approached expert performance with macro average F1 scores of 0.91, 0.95 and 0.90 respectively on external data. For uncertainty an F1 of 0.77 is achieved.ConclusionsOur ‘TNMu’ classifier successfully extracts TNM staging information from internal and external PET-CT reports. We concluded that multi-task approaches result in the best performance, and better computational efficiency over single task PLM approaches. We believe these models can improve PET-CT services by assisting in auditing, creating research cohorts, and developing decision support systems. Our approach to handling uncertainty represents a novel first step but has room for further refinement.

AI‐powered automated analysis of bone scans: A survey

AbstractAs one of the key techniques of artificial intelligence, deep learning has emerged as an effective approach for analysing medical images. Various imaging techniques including the planar bone scintigraphy, single photon emission computed tomography and PET can be used to evaluate, in vivo, bone conditions. The introduction of deep learning techniques especially the convolutional neural networks can significantly improve diagnosis accuracy and efficiency of nuclear medicine physicians. Focusing on bone scans acquired by various nuclear medicine imaging techniques, his paper reviews existing work on deep learning‐based classification, segmentation and object detection of bone scans. Specifically, an overview of existing work about research objective is presented, deep learning models are adopted, and main results are achieved. Research challenges and directions for developing automated analysis of bone scans with deep learning techniques are then discussed.

18F-FDG PET/CT in Inflammation and Infection: Procedural Guideline by the Korean Society of Nuclear Medicine.

This guideline outlines the use of 18F-fluoro-2-deoxyglucose positron emission tomography / computed tomography for the diagnosis and management of infectious and inflammatory diseases. It provides detailed recommendations for healthcare providers on patient preparation, imaging procedures, and the interpretation of results. Adapted from international standards and tailored to local clinical practices, the guideline emphasizes safety, quality control, and effective use of the technology in various conditions, including spinal infections, diabetic foot, osteomyelitis, vasculitis, and cardiac inflammation. The aim is to assist nuclear medicine professionals in delivering accurate diagnoses and improving patient outcomes while allowing flexibility to adapt to individual patient needs, technological advancements, and evolving medical knowledge. This document is a comprehensive resource for enhancing the quality and safety of 18F-fluoro-2-deoxyglucose positron emission tomography / computed tomography for the evaluation of infectious and inflammatory diseases. The Korean Society of Nuclear Medicine (KSNM) was established in 1961 to promote the clinical and technological advancement of nuclear medicine in South Korea, with members that include nuclear medicine physicians and associated scientists. The KSNM regularly formulates and revises procedural guidelines for nuclear medicine examinations to enhance the field and improve the quality of patient care. These guidelines are designed to support healthcare professionals in providing appropriate medical care to patients. However, they are not immutable rules or mandatory requirements for conducting examinations.Therefore, KSNM states that these guidelines should not be used in legal actions challenging a healthcare professional's medical decisions. The ultimate judgment regarding specific procedures or appropriate measures should be made by nuclear medicine physicians, considering the unique circumstances of each case. Deviation from these guidelines does not imply substandard medical practice. Rather, reasonable judgments differing from the guidelines can be made based on the patient's condition, available resources, and advancements in knowledge or technology. Due to the diversity and complexity of patients, it is often challenging to predict the most appropriate diagnostic and accurate therapeutic responses. Thus, adherence to these guidelines does not always guarantee an exact diagnosis or successful outcomes.The purpose of this guideline is to assist healthcare providers in making reasonable decisions and conducting effective and safe examinations based on current medical knowledge, available resources, and patient needs when performing 18F-fluoro-2-deoxyglucose (18F-FDG) positron emission tomography (PET)/computed tomography (CT) examinations for infectious/inflammatory diseases.

Manifestation of Adult children of patients having Alzheimer’s dementia

AbstractBackgroundIndividuals with his parents having Alzheimer’s dementia (AD) have a 4 to 10 times higher risk of developing AD compared to those without family history. The increasing risks of developing AD may result from several underlying mechanisms to be clarified for further strategies of management. We are going to report the demographic and biological characteristics of adult children (AC) of AD parents cohort with its clinical significances in Taiwan.MethodAC of AD patients age from 50‐79 years old were recruited from the neurological outpatient department of Kaohsiung Municipal Ta‐Tung Hospital, Taiwan. The paired individuals, adult children with his/her parents were received annual following up in clinical, neuropsychological, and neuroimaging including PET/CT with 18F‐Florbetaben (FBB) imaging session, using DiscoveryTM MI DR PET/CT system (General Electric Medical Systems, Waukesha, Wisconsin, USA) examinations and assays of blood for genetic and protein analyses. This study was approved by Kaohsiung Medical University Hospital Institutional Review Board. Written informed consent was obtained from all participants before including in this study.Result63 AC with age 59.3 ± 5.8 years, female predominant (76.2%), baseline MMSE 28.3 ± 1.8, CASI 94.7 ± 3.4, and having at least one of APOE e4 allele 34.9%, have recruited. 19 voluntary participants (25.4%) completed 18F‐FBB PET. 9 of 19 (47.4%) were classified as amyloid‐positive based on visual assessment, rated by a senior board‐certificated nuclear medicine physician with experience in interpretation of 18F‐FBB PET who evaluated all images according to the guideline of 18F‐FBB PET interpretation, blind to all clinical information.ConclusionAC of AD patients have higher proportion of having APOE4 allele and higher proportion of having amyloid PET (+). It may imply AC does increase the risk of developing AD, and AC with amyloid PET (+) could be preclinical AD.

Diagnostic Accuracy and Value of CXCR4-targeted PET/MRI Using 68Ga-Pentixafor for Tumor Localization in Cushing Disease.

Background Gallium 68 (68Ga) pentixafor has emerged as a potential C-X-C chemokine receptor type 4 (CXCR4)-targeted radiotracer for neuroendocrine tumor, yet its application in Cushing disease remains uncertain. Purpose To assess the diagnostic accuracy and value of 68Ga-pentixafor PET/MRI in localizing adrenocorticotropic hormone (ACTH)-secreting pituitary tumors in Cushing disease. Materials and Methods A prospective single-center study was conducted from March 2023 to February 2024 in participants with Cushing disease scheduled for surgical pituitary tumor resection. All participants underwent preoperative 68Ga-pentixafor PET/MRI and contrast-enhanced MRI. Two radiologists and nuclear medicine physicians analyzed images. Surgical and histologic findings served as the reference standard. Diagnostic performance was compared using the McNemar test. The Wilcoxon signed rank test was used to compare the maximum standardized uptake value (SUVmax) between tumor and normal pituitary. Correlations between SUVmax and histopathologic or hormone characteristics were analyzed using the Spearman coefficient and logistic regression tests. Results A total of 43 participants (median age, 37 years [IQR, 31-49 years]; 35 female) were included, with 44 pituitary lesions identified after investigational imaging scans, 41 of which were confirmed as ACTH secreting. Sensitivity and diagnostic accuracy of 68Ga-pentixafor PET/MRI were 92.7% (38 of 41 lesions; 95% CI: 80.6, 97.5) and 88.6% (39 of 44 lesions; 95% CI: 76.0, 95.0), respectively, which were higher compared with that of contrast-enhanced MRI (78.0% [32 of 41 lesions; 95% CI: 63.3, 88.0] and 77.3% [34 of 44 lesions; 95% CI: 63.0, 87.2], respectively). Both techniques combined improved sensitivity to 100% (41 of 41 lesions; 95% CI: 91.4, 100.0) and accuracy to 95.5% (42 of 44 lesions; 95% CI: 84.9, 99.2) (P = .01). ACTH-secreting pituitary tumors exhibited a higher SUVmax than normal pituitary (3.9 vs 1.3, P < .001). The SUVmax exhibited a positive correlation with CXCR4 H-score (R = 0.5 [95% CI: 0.2, 0.7], P < .001) and ACTH levels (R = 0.4 [95% CI: 0.1, 0.6], P = .01). Conclusion 68Ga-pentixafor PET/MRI demonstrated high sensitivity in localizing ACTH-secreting pituitary tumors. 68Ga-pentixafor uptake was associated with CXCR4 expression and ACTH level. © RSNA, 2024 Supplemental material is available for this article.

AI-Based Noise-Reduction Filter for Whole-Body Planar Bone Scintigraphy Reliably Improves Low-Count Images.

Background/Objectives: Artificial intelligence (AI) is a promising tool for the enhancement of physician workflow and serves to further improve the efficiency of their diagnostic evaluations. This study aimed to assess the performance of an AI-based bone scan noise-reduction filter on noisy, low-count images in a routine clinical environment. Methods: The performance of the AI bone-scan filter (BS-AI filter) in question was retrospectively evaluated on 47 different patients' 99mTc-MDP bone scintigraphy image pairs (anterior- and posterior-view images), which were obtained in such a manner as to represent the diverse characteristics of the general patient population. The BS-AI filter was tested on artificially degraded noisy images-75, 50, and 25% of total counts-which were generated by binominal sampling. The AI-filtered and unfiltered images were concurrently appraised for image quality and contrast by three nuclear medicine physicians. It was also determined whether there was any difference between the lesions seen on the unfiltered and filtered images. For quantitative analysis, an automatic lesion detector (BS-AI annotator) was utilized as a segmentation algorithm. The total number of lesions and their locations as detected by the BS-AI annotator in the BS-AI-filtered low-count images was compared to the total-count filtered images. The total number of pixels labeled as lesions in the filtered low-count images in relation to the number of pixels in the total-count filtered images was also compared to ensure the filtering process did not change lesion sizes significantly. The comparison of pixel numbers was performed using the reduced-count filtered images that contained only those lesions that were detected in the total-count images. Results: Based on visual assessment, observers agreed that image contrast and quality were better in the BS-AI-filtered images, increasing their diagnostic confidence. Similarities in lesion numbers and sites detected by the BS-AI annotator compared to filtered total-count images were 89%, 83%, and 75% for images degraded to counts of 75%, 50%, and 25%, respectively. No significant difference was found in the number of annotated pixels between filtered images with different counts (p > 0.05). Conclusions: Our findings indicate that the BS-AI noise-reduction filter enhances image quality and contrast without loss of vital information. The implementation of this filter in routine diagnostic procedures reliably improves diagnostic confidence in low-count images and elicits a reduction in the administered dose or acquisition time by a minimum of 50% relative to the original dose or acquisition time.