Background Template-based PET metrics quantify Alzheimer disease (AD) amyloid-β (Aβ) and tau burden but compress whole-brain data into a single scalar, overlooking disease heterogeneity and sometimes causing imaging-clinical discordance. Artificial intelligence (AI) approaches capture richer patterns but often lack biologic interpretability. Purpose To develop and validate an interpretable deep-learning framework that separates AD-specific abnormalities from physiologic uptake using pathophysiologic constraints, generating a clinically meaningful AI biomarker. Materials and Methods In this retrospective study, Aβ and tau PET scans from the Alzheimer's Disease Neuroimaging Initiative, Australian Imaging Biomarkers and Lifestyle study, Global Alzheimer's Association Interactive Network, and the authors' center were analyzed. An adversarial decomposition learning (ADL) network generated voxel-level pathologic maps and an AD adversarial decomposition (ADAD) score. Discriminatory performance for clinical AD versus cognitively normal individuals was evaluated using the area under the curve (AUC). Clinical relevance was assessed with cognitive, hippocampal volume, cerebrospinal fluid (CSF), and neuropathologic measures using longitudinal mixed-effects models and Spearman correlations. Results The study included 7457 Aβ PET scans from 3595 patients (median age, 71.4 years; IQR, 65.7-77.0 years; 1637 female patients) and 1894 tau PET scans from 1127 patients (median age, 72.0 years; IQR, 66.9-78.5 years; 545 female patients). External testing AUCs were 0.94 (95% CI: 0.89, 0.98) for Aβ and 0.98 (95% CI: 0.95, 1.00) for tau. ADL generated interpretable pathologic attribution maps that correlated with expert rankings (Aβ and tau, Spearman ρ = 0.79 and 0.63, respectively). Although Centiloid and CenTauRz showed numerically higher correlations with postmortem neuropathologic structure and stronger associations with CSF biomarkers, the ADAD score demonstrated independent baseline and longitudinal associations with cognitive outcomes and hippocampal atrophy after adjustment. Conclusion Pathophysiologic-constrained ADL provided interpretable, personalized pathologic maps and an AI-derived ADAD score that more closely linked PET pathologic abnormalities with multimodal clinical measures. © RSNA, 2026 Supplemental material is available for this article.

Supplementary Figure S4. Comparison of quantitative region-of-interest (ROI) data in [68Ga]Ga-XH-06 immunoPET imaging of Hep3B and SK-HEP-1 subcutaneous xenograft models.

AbstractPurpose: The study aims to compare the diagnostic performance of the novel melanin-targeted [18F]-N-(2-(diethylamino)ethyl)-5-(2-(2-(2-fluoroethoxy)ethoxy)ethoxy)picolinamide ([18F]PFPN) positron emission tomography (PET) and [18F]fluorodeoxyglucose ([18F]FDG) PET in mucosal melanoma, evaluate their impact on clinical staging, and assess correlations between imaging metrics and molecular markers. Experimental Design: This prospective study enrolled 65 participants with histologically confirmed mucosal melanoma from February 2021 to January 2025. All participants underwent both [18F]FDG PET and [18F]PFPN PET within 1 week. Lesion- and participant-based analyses compared detection sensitivity, false-positive rate, and staging concordance. Quantitative PET parameters were analyzed, and correlations with HMB45, SOX10, MelanA, S100, and mutation status (BRAF, KIT, NRAS) were evaluated using nonparametric tests and correlation analysis with Bonferroni correction. Decision curve analysis was used to evaluate clinical benefit. Results: Sixty-five participants were included. [18F]PFPN PET showed higher lesion-based sensitivity than [18F]FDG PET [363/399 (91%) vs. 332/399 (83.2%)] and no false positives [0/363 (0%) vs. 4/336 (1.2%)]. The normalized maximum standardized uptake value was significantly higher for [18F]PFPN across all lesion types (P < 0.05). PFPN-based staging was more consistent with clinical staging (6.2% vs. 18.5% discordant cases). [18F]PFPN uptake showed significant positive correlations with HMB45 and SOX10 expression, whereas [18F]FDG parameters showed no such associations. Conclusions: [18F]PFPN PET outperforms [18F]FDG PET in lesion detection and clinical staging in mucosal melanoma, especially for liver and bone metastases. Its association with melanin differentiation markers may support its use in personalized imaging strategies.

Supplementary Figure S6. Quantitative ROI data in [68Ga]Ga-XH-06 immunoPET imaging of Hep3B orthotopic xenograft models.

To evaluate the diagnostic efficacy of 68Ga-PSMA PET/MRI-derived parameters in prostate cancer and their significance in risk stratification. A retrospective study was performed on patients who underwent ⁶⁸Ga-PSMA PET/MRI from November 2020 to April 2024, all with histopathological confirmation obtained from biopsy or prostatectomy. Parameters analyzed included SUVmax, PSMA tumor volume at a 40

PURPOSE:This study aims to develop and evaluate a novel glypican-3 (GPC3)-targeted, single-chain variable fragment-based PET radiotracer for noninvasive assessment of GPC3 expression and precise diagnosis of hepatocellular carcinoma (HCC) in both preclinical models and a first-in-human clinical study. EXPERIMENTAL DESIGN:The novel anti-GPC3 single-chain variable fragment, namely, XH-06, was labeled with gallium-68 to obtain [68Ga]Ga-XH-06. Cell uptake assays, small-animal PET imaging, and biodistribution studies were performed to evaluate its targeting ability. In the first-in-human study, eight patients with suspected HCC underwent [68Ga]Ga-XH-06 PET/MRI. Radiotracer uptake in tumors and normal tissues was quantified, and tumor-to-blood ratios and tumor-to-liver ratios were calculated. RESULTS:[68Ga]Ga-XH-06 was synthesized with high radiochemical purity and exhibited specific uptake and efficient internalization in GPC3-positive cells. In subcutaneous and orthotopic animal models, [68Ga]Ga-XH-06 effectively visualized HCC tumors. Eight patients underwent [68Ga]Ga-XH-06 PET/MRI scans, and no adverse events were observed. The radiotracer successfully detected HCC lesions, including subcentimeter tumors, with high imaging contrast. Among seven patients with HCC, the median maximum standardized uptake value of the lesions was 16.9 (range, 8.2-51.8), the median tumor-to-liver ratio was 6.2 (range, 2.8-24.7), and the median tumor-to-blood ratio was 16.6 (range, 3.0-57.6) at 2.5 hours after injection. CONCLUSIONS:[68Ga]Ga-XH-06 is clinically promising for detecting GPC3-positive HCC with a favorable safety profile. Further investigation is warranted to validate the clinical value of GPC3-targeted PET imaging in HCC management. See related commentary by Seo and Flavell, p. 445.

Semi-quantitative positron emission tomography (PET) analysis, particularly Centiloid and CenTauRz scaling, is essential for Alzheimer’s disease (AD) research and diagnosis. However, standard quantification workflows often depend on structural MRI for spatial normalization (SN) or rely on computationally intensive software, limiting clinical accessibility. In this retrospective, multi-center study (3539 patients; 6535 scans; 2005–2025), we compiled data across 7 modalities and 13 tracers to develop and validate the Deep Cascaded Cerebral Calculator (DCCC). This fully automated, PET-only framework employs cascaded CNN-based rigid/affine and VoxelMorph-based elastic registration modules for rapid SN. We benchmarked DCCC against the standard MRI-guided SPM12 pipeline and other PET-only tools using meta region-of-interest (ROI) standard uptake value ratio (SUVr) and correlation analyses. DCCC achieved a mean absolute relative SUVr error of 1.34±0.59% and a voxelwise Pearson correlation of 0.96±0.02, demonstrating robust generalization to unseen tracers and modalities including neuroinflammation and methionine metabolism imaging, with superior consistency compared to conventional template-based PET-only methods. Centiloid and CenTauRz estimates were highly accurate (R²>0.97) with a processing speed of 1.22±0.64 s per image. We further demonstrated DCCC’s utility across 3 scenarios: (1) longitudinal tracking, where it identified a distinct low-Centiloid AD subgroup; (2) deep learning preprocessing, yielding classification AUCs comparable to standard methods (P = 0.36); and (3) exploratory clinical support, where DCCC-derived metrics were adopted in 79% Aβ and 61% tau cases and were associated with changes in interpretation and increased agreement with reference labels in a multi-reader survey. Collectively, DCCC provides accurate, PET-only standardization, facilitating harmonized biomarker estimation without MRI and enabling large-scale, tracer-agnostic analyses in AD neuroimaging. A free standalone command-line interface program and a 3D Slicer plugin are provided.

Neuroinflammation is a key factor contributing to cognitive decline in Alzheimer’s disease (AD). This study aims to investigate the mechanistic associations among neuroinflammation, glymphatic dysfunction, tau pathology, and cognitive decline in AD spectrum. The study included 355 participants from the Alzheimer’s Disease Neuroimaging Initiative (ADNI) and a supportive cohort of 59 individuals from Wuhan Union Hospital (WHUH). Tau pathology was quantified using 18F-AV1451 positron emission tomography (PET). Glymphatic function was estimated through diffusion tensor image analysis along the perivascular space (DTI-ALPS). Neuroinflammation was assessed via plasma glial fibrillary acidic protein (GFAP) in two cohorts and translocator protein (TSPO) PET imaging with 18F-DPA-714 in supportive cohort. Correlation analyses and mediation models were employed to evaluate the directional relationships among tau deposition, inflammation, glymphatic function, and cognition. Higher levels of inflammation were significantly associated with lower DTI-ALPS index (β = −0.171, P = 0.046), which in turn was associated with higher tau burden (β = 0.162, P = 0.010). Path analysis revealed significant indirect associations linking neuroinflammation to cognitive performance through glymphatic dysfunction and tau pathology, with total indirect effects of − 0.165 (95

The analysis of abnormalities in various brain regions requires combining metabolic data from PET with anatomical segmentation from MR due to the relatively low resolution of PET images. For brain MR segmentation, dealing with the intricate morphological characteristics of the Central Nervous System (CNS) represents a fundamental challenge in medical image analysis. While traditional methods such as Atlas-based techniques have been widely used, recent advancements in deep learning (DL) have significantly transformed the field. In this study, we automatically segmented two representative brain areas from hybrid PET/MR scans of twelve PD and three MSA patients using both Atlas- and DL-based methods. Then we compared the Standardized Uptake Values (SUVs) and accuracy of segmentation (DSC, MSD and HD) in the corresponding regions, with manual segmentation used as the ground truth for comparison. The Atlas-based brain segmentation relied on an atlas template from the automated anatomical labeling atlas, which includes 70 segmented regions labeled from 1 to 70. On the other hand, DL-based brain segmentation utilized a 3D transformer model based on MONAI framework for whole brain segmentation, employing a pre-trained model for inferring the whole brain with 133 structures on T1-weighted MR images. Both methods involved the use of SPM12 for the 3D affine registration of T1-weighted MR images to the MNI space. Manual segmentation was carried out by a clinical neuroimaging expert using ITK-SNAP, focusing on two cerebral nuclei containing four sub-regions (left Caudate, right Caudate, left Putamen, and right Putamen). The results of this study indicate that the DL-based method produced superior segmentation accuracy compared to the Atlas-based method. However, there were no significant differences in the SUVmax and SUVmean across the different methods for the segmentation of Caudate and Putamen. Despite being more accurate for Caudate and Putamen segmentation, the DL-based method had little effect on the calculation of their SUVs in hybrid PET/MR scans, as compared to the widely used Atlas-based method. This comparison and evaluation could potentially extend to other structures based on the algorithms.

PURPOSE:This study aims to establish a prognostic prediction system for mucosal melanoma patients by integrating melanocyte-targeting PET with clinical parameters. PATIENTS AND METHODS:A retrospective study was conducted on primary mucosal melanoma patients who underwent [ 18 F]PFPN PET scans between January 2021 and April 2024. Researchers manually delineated all lesions on the [ 18 F]PFPN PET images and recorded the imaging features. Kaplan-Meier survival analysis, Cox regression, and stepwise regression were used to analyze prognosis and construct the prediction model. RESULTS:Fifty patients (mean age 60.5±8.8 y) were included, with a median follow-up of 13 months (1-39 mo) and an average survival of 14.62 months. Multivariate analysis showed that lower Whole-body Melanotic Tumor Volume (WBMTV), earlier stage, younger age, immunotherapy, and head and neck or vulvar subtypes were associated with longer overall survival (OS, P <0.05). A simplified prognostic scoring system was developed based on 5 variables: stage (not detected or I/II/III/IV: 0/5/10/15 points), age (<60/≥60: 0/10), WBMTV (<1.52/≥1.52: 0/10), immunotherapy (yes/no: 0/5), and subtype (head and neck or vulva/others: 0/10). Patients were stratified into low (0-19), intermediate (20-29), and high-risk (30-50) groups. The model's concordance index was 0.85, outperforming the clinical staging (0.76). OS declined progressively from low-risk to high-risk groups, with 1-year survival from 100% to 74.6% and 3-year survival from 100% to 0%. CONCLUSIONS:[ 18 F]PFPN PET provides an accurate assessment of mucosal melanoma burden, and a prognostic model combining [ 18 F]PFPN PET features and clinical data offers reliable stratification of patient outcomes.

PURPOSE:To evaluate the efficiency of 68Ga-FAPI-04 PET (PET/MRI or PET/CT) for N and M staging in gastric carcinoma and compare outcomes with histopathology and contrast-enhanced computed tomography (CECT). PATIENTS AND METHODS:Patients with gastric carcinoma who had undergone 68Ga-FAPI-04 PET/MRI or PET/CT before treatment were retrospectively enrolled. Histopathology post lymphadenectomy was the gold standard for N staging, while histopathology and follow-up data were the reference for overall outcomes. The diagnostic efficiency of 68Ga-FAPI-04 PET for detecting regional lymph node involvement and distant metastases was compared to that of CECT. RESULTS:Sixty-two patients were enrolled. In 18 patients who underwent 68Ga-FAPI-04 PET/MRI and lymphadenectomy, 532 lymph nodes were dissected. 68Ga-FAPI-04 PET/MRI showed similar sensitivity, specificity, and accuracy compared to CECT (28.3% vs. 23.2%, 99.8% vs. 99.3%, and 86.5% vs. 85.2%, all P > 0.05). Fifty-five patients had regional lymph node metastasis, 68Ga-FAPI-04 PET exhibited comparable diagnostic efficiency to CECT, with sensitivity of 83.6% versus 87.3%, specificity of 100% versus 85.7%, accuracy of 85.5% versus 87.1% (all P > 0.05). Excluding 3 patients with only abdominal CECT, 32 out of 59 patients had distant metastasis, with no significant differences in sensitivity, specificity, and accuracy between 68Ga-FAPI-04 PET and CECT based on patient (100% vs. 87.5%, 92.6% vs. 96.3%, and 96.6% vs. 91.5%, all P >0.05). Notably, 68Ga-FAPI-04 PET outperformed CECT in detecting peritoneal, distant lymph nodes, bone, liver, and ovarian metastases by visualizing more lesions or greater lesion extent. CONCLUSIONS:68Ga-FAPI-04 PET exhibits comparable diagnostic performance to CECT for patient-based N staging and M staging of gastric cancer. However, it surpasses CECT in visualizing distant metastases.

Background Alzheimer's disease (AD) is the most prevalent neurodegenerative disorder characterized by cognitive deficit and pathological accumulation of amyloid-β (Aβ) and tau proteins. The rodent models have contributed greatly to unravel AD pathogenesis, but these AD models have been shown a modest clinical translational effectiveness. Objective Therefore, developing mass-producible primate AD models is promising for more effective drug development. Methods Here, we constructed the AD monkey models by simultaneously infusing AAV-Tau and Aβ into different brain regions. Results The induced monkeys showed a durable cognitive impairment lasting for at least 10 months after the modeling. Simultaneously, the increased levels of total tau and hyperphosphorylated tau (pTau) at several AD-associated sites, and neurofilament light chains (NfL) with altered Aβ level were detected at different time points in cerebrospinal fluid and/or plasma by using MSD kits. The increased brain accumulation of Aβ and tau proteins was also detected by positron emission tomography/magnetic resonance imaging and immunohistochemical staining. The model monkeys also had significant glial activation; an indicator of inflammation commonly seen in the brains of AD patients. Conclusions Together, this study provides mass-producible monkey models showing durable AD-like hallmark pathologies (Aβ, tau, NfL, i.e., ATN) and cognitive deficits. As monkeys are genetically and metabolically the closest to humans, these models will offer more effective drug discovery and development for AD.

Gallium-68-labeled fibroblast activation protein inhibitor ([68Ga]Ga-FAPI) positron emission tomography (PET) has demonstrated excellent diagnostic performance in various malignancies, including gastric carcinoma. However, its prognostic utility is unclear. This study evaluates the prognostic value of [68Ga]Ga-FAPI-04 PET/MRI(CT) in gastric carcinoma. We retrospectively analyzed patients with gastric cancer who underwent [68Ga]Ga-FAPI-04 PET/MRI(CT) between June 2020 and June 2023. Semi-quantitative parameters, including maximum and mean standard uptake value (SUVmax, SUVmean), FAPI-avid tumor volume (FTV), total lesion FAP expression (TLF), tumor to background ratio (TBR), heterogeneity factor (HF) and coefficient of variation (CV) of the primary tumor were measured or calculated. Overall survival (OS) and progression-free survival (PFS) were obtained through follow-up. The relationships between disease prognosis and potential predictors were analyzed， and predictive models were established. Eighty-six patients (median age 59 years) were included. Thirty-five patients experienced disease progression, and 26 of them died. Univariable analysis revealed SUVmax, FTV, TLF, TBR, HF and CV were significant prognostic factors for both OS and PFS. In multivariate Cox regression analysis, a nomogram model for OS was established, incorporating body mass index (BMI) and CV as independent predictors. The time-dependent C-index of the nomogram model > 0.75 indicates good predictive performance. When predicting PFS, a stratified analysis was performed based on distant metastasis, FTV was an independent prognostic factor among patients without distant metastasis. CV and FTV, derived from [68Ga]Ga-FAPI-04 PET imaging, could serve as independent prognostic factor for OS and PFS in patients with gastric cancer, respectively.

Tumor-associated neovasculature and energy metabolism reprogramming serve as critical indicators of tumor proliferation, progression, invasion, and metastasis. This study conducted a head-to-head clinical investigation and comparison of [18F]FDG and [68Ga]Ga-HX01 PET by reflecting neovasculature and glucose metabolism in sarcoma patients, respectively. We reviewed the imaging data of sarcoma patients who underwent [68Ga]Ga-HX01 PET/MR and [18F]FDG PET/CT from June 29, 2022, to December 21, 2023. The two imaging modalities were performed on two separate days within one week of each other. A cohort of 21 patients with an average age of 45.81 ± 19.99 years were enrolled. The location, number and PET characteristics of all lesions were collected. The relationships between the two tracers were evaluated. Among the 21 patients, 4 underwent imaging for initial disease staging, while the remaining 17 were imaged to detect recurrences. Patient-based analysis revealed that [68Ga]Ga-HX01 PET/MR diagnostic performance was equivalent to [18F]FDG PET/CT in lesion detection (P = 1.0). The SUVmax value of [68Ga]Ga-HX01 (4.64 ± 1.90) was significantly lower than that of [18F]FDG (9.43 ± 6.17, P = 0.002) across all patients. In terms of lesion-based analysis, [68Ga]Ga-HX01 identified two additional lesions compared to [18F]FDG, though this difference was not statistically significant (94 vs. 92, P = 0.678). The SUVmax value for all lesions with [68Ga]Ga-HX01 (3.47 ± 1.68) was also lower than that with [18F]FDG (5.82 ± 4.81, P = 0.003). Notably, [68Ga]Ga-HX01 was preferred in patients receiving hematopoietic cytokines. [68Ga]Ga-HX01 PET offers comparable diagnostic efficacy to [18F]FDG PET/CT in sarcoma, with potential advantages in specific clinical scenarios. Larger cohorts are needed to validate these findings. NCT05490849 and NCT06416774.

The reversibility of early liver fibrosis highlights the need for improved early detection and monitoring techniques. Fibroblast activation protein (FAP) is a promising theranostics target significantly upregulated during fibrosis. This preclinical and preliminary clinical study investigated a FAP-targeted probe, gallium-68-labeled FAP inhibitor 04 ([68Ga]Ga-DOTA-FAPI-04), for its capability to visualize liver fibrosis. The preclinical study employed [68Ga]Ga-DOTA-FAPI-04 micro-positron emission tomography (PET)/computed tomography (CT) on carbon tetrachloride-induced mice model (n = 34) and olive oil-treated control group (n = 26), followed by validation of the probe's biodistribution. Hepatic uptake was correlated with fibrosis and inflammation levels, quantified through histology and serum assays. FAP and α-smooth muscle actin expression were determined by immunohistochemistry, as well as immunofluorescence. The subsequent clinical trial enrolled 26 patients with suspected or confirmed liver fibrosis to undergo [68Ga]Ga-DOTA-FAPI-04 PET/magnetic resonance imaging or PET/CT. Key endpoints included correlating [68Ga]Ga-DOTA-FAPI-04 uptake with histological inflammation grades and fibrosis stages, and evaluating its diagnostic and differential efficacy compared to established serum markers and liver stiffness measurement (LSM). [68Ga]Ga-DOTA-FAPI-04 mean uptake in mice livers was notably higher than in control mice, increasing from week 6 [0.70 ± 0.11 percentage injected dose per cubic centimeter ( https://clinicaltrials.gov/study/NCT04605939

Radiolabeled probes targeting prostate-specific membrane antigen (PSMA) have been used in prostate cancer. Moreover, PSMA is also overexpressed on neovessels in hepatocellular carcinoma (HCC). This study aimed to preliminarily evaluate the diagnostic effectiveness of [68Ga]Ga-PSMA-617 PET/MRI for HCC. Patients suspected of HCC were prospectively enrolled in this single-center study (NCT05006326, 2021-08-16) to perform [68Ga]Ga-PSMA-617 PET/MRI, along with contrast enhanced CT (ceCT) or ceMRI. The main suspicious intrahepatic lesions were resected and pathologically verified. Visual evaluation of [68Ga]Ga-PSMA-617 PET/MRI images was performed on all lesions. Maximum standard uptake value (SUVmax), mean standard uptake value (SUVmean), tumor-to-liver ratio (TLR), tumor-to-blood ratio (TBR), and tumor-to-parotid ratio (TPR) were measured or calculated. The diagnostic efficiency of different modalities was summarized. PSMA expression was evaluated by immunohistochemistry and the correlation of PSMA expression and [68Ga]Ga-PSMA-617 uptake in HCC primary tumors was quantitatively analyzed. A total of 12 patients (ten men and two women; mean age 58.75 ± 12.08 years) were included. Ten patients were diagnosed with HCC, 2 with intrahepatic cholangiocarcinoma (ICC), and 4 with hemangioma. The SUVmax, TLR, TBR, and TPR of HCC primary tumors were higher than those of ICC and hemangioma. The diagnostic accuracy of [68Ga]Ga-PSMA-617 PET/MRI for primary HCC was 82.4 https://clinicaltrials.gov/study/NCT05006326 .