Radiopharmaceutical therapy (RPT) is becoming a cornerstone of cancer treatments with the approval of 177Lu-PSMA (Pluvicto; Novartis), 177Lu-DOTATATE (Lutathera; Novartis), and 223Ra-dichloride (Xofigo; Bayer). An even larger number of treatments with different isotopes and biologic targets are being studied and promise a revolution in oncology care. However, the Food and Drug Administration's approval of 177Lu-PSMA RPT in the prechemotherapy space has raised concerns regarding access, logistics, supply, quality of care, and workforce challenges. Methods: To understand the current and future landscapes and to advocate for best patient care practices in the setting of prostate-specific membrane antigen-targeted RPT (PSMA RPT), the Society of Nuclear Medicine and Molecular Imaging Theranostics Leadership and Operations Group surveyed its members to explore the demand for PSMA RPT and the capacity and preparedness of the centers providing this therapy. Results: Survey results indicated that the glass is half-full, with nuclear medicine therapy centers meeting current challenges in various aspects of PSMA RPT including logistics, capacity, infrastructure, access, reimbursement, imaging, and clinical care. Eighty-six percent stated that they have a comprehensive consult with their patients, 83% reported that they ordered and managed their patient's initial and follow-up blood work, and 83% reported that reimbursement was satisfactory. Although 86% had no concerns managing their current volume, this dropped to 52% when queried about concerns regarding their capacity to handle PSMA RPT volume in the future, noting various challenges including staffing (physicians, technicians, and nurses), a limited number of treatment rooms, and limited access to SPECT/CT for posttherapy imaging. Conclusion: Survey results indicate the glass is half-full, with nuclear medicine therapy centers meeting current challenges in various aspects of PSMA RPT, including logistics, capacity, infrastructure, access, reimbursement, imaging, and clinical care. Future needs of nuclear medicine theranostics can be met with proper planning, vision, and effort.
Understanding where drug targets are expressed in the human body is essential for precision medicine, yet this information is difficult to obtain in living patients. Molecular imaging offers a non-invasive way to visualize target expression, but its application remains fragmented. We aim to develop a systematic framework to link approved drugs, their molecular targets, and existing imaging agents, with a focus on repurposing imaging strategies for clinical use. We integrate drug, target, and disease data from public databases and combine these with imaging probe annotations and transcriptomic data from more than 240,000 patient samples across multiple diseases. Co-expression analysis is used to identify candidate surrogate imaging targets for proteins that lack direct imaging agents. Statistical associations are assessed using correlation analysis with multiple testing correction. Here we show that existing imaging agents can be linked to 704 therapeutic targets across 1345 diseases. Nearly half of these targets are directly imageable, while the remainder can be connected to surrogate imaging targets through co-expression. In total, more than 4000 imaging agents are identified, enabling the systematic prioritization of candidate imaging strategies across diseases. This study provides a framework for repurposing molecular imaging agents to visualize drug targets in vivo. The approach expands the potential of imaging to guide patient selection and treatment monitoring, and highlights opportunities to translate existing but underused imaging agents into clinically actionable biomarkers. Choosing the right treatment often depends on whether a specific drug target is present in a patient’s body. This is usually checked using tissue samples, which can be invasive and only show part of the picture. Medical imaging can help by showing these targets throughout the whole body, but many imaging tools are not widely used. In this study, we brought together information on approved drugs, the targets they act on, and existing imaging methods. We also analysed gene activity data from over 240,000 patient samples to find links between drug targets and imaging markers. We found that many drug targets can already be seen in the body, either directly or through related markers. This means that existing imaging tools could potentially be reused to help doctors choose the best treatments and track how well they are working.
Accurate reporting in nuclear medicine is essential for clinical decision-making. Trainees often generate preliminary reports with variable quality, and artificial intelligence (AI) tools such as ChatGPT-4o may enhance report clarity and accuracy, particularly in the impression section of the report. This study aimed to evaluate and compare the quality of positron emission tomography / computed tomography report impression sections generated by trainees and by the AI chatbot ChatGPT-4o (OpenAI), focusing on correctness, clarity, completeness, organization, use of diagnostic certainty terminology, and physician satisfaction. The impression sections of 200 positron emission tomography/ computed tomography reports generated by trainees and AI (100 reports each) were compared. The AI generated the impressions based on the stem, clinical history, and technical sections of the trainee-generated reports. Reports were blindly rated by 3 nuclear medicine physicians. Survey questions, including Likert-scale questions, assessed correctness, use of certainty terms, clarity, completeness, organization, and satisfaction. Statistical analyses included Fisher exact test, 2-tailed t tests, chi-square tests, ANOVA, and effect sizes (Cohen d). Thematic analysis was performed on free-text comments. AI-generated impressions were rated as correct in 92% of reports (mean 0.92, SD 0.27) and trainee-generated impressions rated as correct in 91% of reports (mean 0.91, SD 0.29; P=.80, negligible effect size). AI-generated impressions included certainty terms more frequently than trainee-generated impressions (94% vs 81%, respectively; P=.005; χ²2=6.58, small effect size). AI also used higher-level certainty terms more frequently than trainees (mean 4.23, SD 1.25 vs mean 3.69, SD 1.93 on a 5-point scale; P=.02). Clarity was high in both groups (AI: mean 4.49, SD 0.70, and 89.8% of reports rated as clear vs trainees: mean 4.35, SD 0.85, and 87.0% of reports rated as clear; P=.20). Completeness was significantly higher for AI-generated impressions than trainee-generated impressions (mean 4.52, SD 0.76, and 90.4% complete vs mean 4.09, SD 1.01, and 81.8% complete, respectively; P<.001, small-to-medium effect size). Organization was similar between the groups (mean 4.22, SD 0.91 vs mean 4.39, SD 0.74; P=.15). Satisfaction ratings were also compared, with AI-generated impressions achieving a mean score of 4.31 (SD 0.77) and 86.2% satisfaction, compared with a mean score of 4.14 (SD 0.91) and 82.8% satisfaction for trainee-generated impressions (P=.16). Thematic analysis showed that trainee-generated impressions were more frequently criticized for accuracy (74% vs 19%; P<.001) and actionability (20% vs 4.8%; P=.02), whereas AI-generated impressions were more frequently criticized for excessive length (33% vs 0%; P<.001). These findings suggest that the AI chatbot ChatGPT-4o can serve as a valuable adjunct in nuclear medicine reporting, particularly at the resident level, by improving decisiveness and completeness while complementing trainee education and clinical oversight.
Deep learning (DL) has shown promise in enabling attenuation correction (AC) for SPECT myocardial perfusion imaging (MPI) without relying on anatomical information or CT-derived attenuation maps (ATMs). We introduce a novel reconstruction-informed and multidomain (RIMD) DL framework utilizing both multi-input reconstruction and dual-domain supervision for AC in SPECT MPI. Our method incorporates multi-input non-AC (NAC) images as input to the DL model and employs a combined loss function that optimizes performance in both the ATM and AC domains. A dataset of 1058 SPECT/CT MPI scans using 99mTc-Sestamibi from two centers was used for training (934 cases) and an external test set (124 cases). SPECT projections were reconstructed into AC and NAC images using three reconstruction settings of the Ordered Subset - Expectation Maximization (OSEM) algorithm. SwinUnetR model was trained in a consistent 5-fold cross-validation framework with normalized NAC images as input. Our proposed method, as an indirect strategy, uses multi-input NAC images (incorporating all three images with different reconstruction settings) trained using a combination of ATM loss (between predicted and true ATMs) and AC loss (between AC images reconstructed from predicted ATMs and reference AC images). Evaluation included voxel-wise and region-wise metrics for both ATM and AC domains, 17-segment polar map analysis, and an organ-specific analysis. Clinical validation was also performed on part of the external dataset. Our proposed approach significantly outperformed direct and indirect methods in ablation comparison. This model yielded mean relative absolute error percentage (MRAE%) values of 25.02 ± 23 (internal) and 26.31 ± 14 (external) for ATMs, and 11.72 ± 6.3 (internal) and 19.31 ± 4.9 (external) for AC SPECT images. Organ-wise analysis showed region-wise MRAE% of 9.29 ± 6.5 (internal) and 17.28 ± 17 (external) in the ATM domain, and 4.51 ± 4.3 (internal) and 9.53 ± 6.6 (external) in the AC domain. Polar map analysis across 17 segments showed MRAE% of 5.04 ± 4.6 (internal) and 10.88 ± 7.3 (external). Clinical validation demonstrated high agreement between DLAC and CTAC images (ICC = 0.98), with physicians unable to distinguish between them (F1 score = 0.40), and no significant difference in diagnostic accuracy (DLAC: 0.63, CTAC: 0.70; p = 0.37). This study demonstrated that our proposed RIMD method utilizing multiple OSEM reconstruction inputs and jointly optimizing ATM and AC losses substantially improved model performance in the indirect strategy. The indirect method consistently outperformed the direct approach, and our model generalized well on external data, showing strong agreement with SPECT CTAC images in both quantitative and qualitative assessments. Preliminary clinical evaluation suggested comparable interpretability between DLAC and CTAC under controlled validation conditions.
Cardiac sympathetic imaging with iodine-labeled metaiodobenzylguanidine (MIBG) is clinically relevant for assessing autonomic dysfunction and arrhythmic risk in heart failure. The development of positron-emitting 124I-MIBG and new-generation PET/CT systems, particularly long axial field-of-view (LAFOV) scanners, may improve image quality and regional defect assessment while enabling protocol optimization. This study compared simulated cardiac 124I-MIBG image quality and lesion detectability across three PET/CT generations, analog standard axial field-of-view (SAFOV), digital SAFOV, and LAFOV, and explored the potential for acquisition-time and activity reduction. An anthropomorphic torso phantom with cardiac insert was prepared to simulate physiologic 124I-MIBG biodistribution. Two configurations were studied: a pathological model with a transmural defect (TD) and a non-transmural defect (NTD), and a normal model with homogeneous myocardial uptake. Acquisitions were performed on three PET/CT systems: analog SAFOV (Biograph mCT), digital SAFOV (Biograph Vision Edge 600), and LAFOV (Biograph Vision Quadra). List-mode data were acquired for 10 min and reconstructed to simulate 1-, 5-, and 10-min acquisitions, yielding nine datasets. Qualitative assessment included lesion detectability and overall image quality scored on 5-point Likert scales by two experienced readers. Quantitative analysis included signal-to-noise ratio (SNR) and contrast-to-noise ratio (CNR). Exploratory pairwise Wald-type comparisons were performed using propagated measurement uncertainty. LAFOV PET/CT showed the best qualitative and quantitative performance overall. Lesion detectability was highest on LAFOV at all acquisition times, with digital SAFOV intermediate and analog SAFOV lowest. Image quality likewise favored LAFOV, especially at 5 and 10 min. In the pathological phantom, SNR increased from 25.9 to 75.1 on LAFOV, from 18.7 to 41.8 on digital SAFOV, and from 15.5 to 17.1 on analog SAFOV. At 10 min, exploratory uncorrected pairwise comparisons showed nominally higher SNR on LAFOV than on analog SAFOV PET/CT (75.1 vs 17.1, p = 0.028). For the transmural defect, CNR also increased most on LAFOV, with a nominally higher value versus analog SAFOV at 10 min (67.6 vs 14.5, p = 0.044). By contrast, for the low-contrast non-transmural defect, CNR was numerically higher on LAFOV at longer acquisition times, but the absolute inter-scanner differences were smaller and did not reach statistical significance. LAFOV PET/CT provides superior image quality and lesion detectability for simulated cardiac 124I-MIBG imaging and supports shorter acquisitions and potentially lower administered activity. However, the advantage was most pronounced for high-contrast transmural defects, whereas the benefit for low-contrast non-transmural defects was less robust. Reliable detection of subtle partial innervation defects remains challenging and requires dedicated clinical validation.
The differentiation between benign and malignant persistent pulmonary ground-glass nodules (GGNs) remains challenging, and the relative value of radiomics handcrafted features and deep features derived from 18F-FDG PET/CT in this setting requires further comparison. This study aimed to develop and validate diagnostic models using radiomics handcrafted features and deep features extracted from 18F-FDG PET/CT for differentiating benign and malignant persistent pulmonary GGNs. Data from 173 patients (184 GGNs) across three PET/CT centers were retrospectively analyzed. Patients underwent 18F-FDG PET/CT and breath-hold chest CT, with diagnoses confirmed by pathology or follow-up. Models were developed using clinical features, handcrafted features, and deep features extracted via pretrained convolutional neural networks (VGG19 and ResNet50). The SUTAH dataset was used for model training and validation, and CZ2PH and CZCH datasets were used as external test sets. Single-modality and dual-modality diagnostic models were developed based on clinical/conventional imaging features, radiomics handcrafted features, and deep features. Deep features were extracted from PET and CT images using the pretrained convolutional neural networks VGG19 and ResNet50. Model performance was evaluated using the AUC and its 95% CI, while decision curve analysis (DCA) and the Brier score were used to assess net benefit and probability prediction error, respectively. In the internal validation set, the CT_HF and PET_HF models achieved AUCs of 0.889 (95% CI: 0.756-0.964) and 0.903 (95% CI: 0.774-0.972), respectively, which were numerically higher than that of the reference model (AUC = 0.864, 95% CI: 0.725-0.949). In the external test set, the PET_ResNet50/CT_VGG19 combined model achieved the highest AUC of 0.927 (95% CI: 0.802-0.984), compared with 0.824 (95% CI: 0.676-0.924) for the reference model and 0.884 (95% CI: 0.747-0.962) for PET_ResNet50; however, the differences among the three models were not statistically significant (P = 0.101-0.464). DCA showed that this combined model had a relatively high net benefit when the threshold probability exceeded 0.5, and its Brier score in the external test set was 0.122, which was comparable to that of the reference model. The integrated ¹⁸F-FDG PET/CT model with PET_ResNet50 and CT_VGG19 deep features achieved favorable discrimination and modest net clinical benefit in the external test set. However, its probability calibration and generalizability await validation on larger multicenter datasets due to insufficient benign nodules in the external test set.
The in vivo processes underlying neural stem cell (NSC) transplantation for ischemic stroke remain unclear, primarily due to the lack of effective in vivo monitoring and evaluation approaches. Therefore, we systematically investigated the in vivo mechanisms of NSC transplantation by positron emission tomography (PET) on a rat model with ischemic stroke. NSCs were generated from human induced pluripotent stem cells (hiPSCs), and their neuronal differentiation potential was assessed in vitro. After transplantation into rats with photothrombotic stroke, neurological function was evaluated using the cylinder test and forelimb placing test. 18F-FDG, 18F-SynVesT-1 and 18F-DPA-714 PET were used to monitor brain metabolism, synaptic density, and glial reactivity in vivo. And correlations among the results of in vivo experiments were analyzed. Neuronal differentiation, maturation and synaptic connection of transplanted NSCs were investigated, while changes of glial reactivity were explored. NSCs derived from hiPSCs exhibited high neuronal differentiation potential in vitro. After NSC transplantation, neurological function of rats with ischemic stroke was recovered, while transplanted NSCs differentiated into neurons and established synaptic connections in the rat brain. In vivo brain metabolism and synaptic density were both restored, revealed by 18F-FDG and 18F-SynVesT-1 PET imaging and highly correlated with behavioral recovery. In vivo glial reactivity was higher at 8 weeks after NSC transplantation in 18F-DPA-714 PET imaging, mainly induced by microglial M2 polarization, which promotes neurological function recovery. Our findings suggest PET molecular imaging can systematically monitor the brain metabolism, synaptic density and glial reactivity, providing effective in vivo monitoring and evaluation approaches for NSC transplantation in ischemic stroke.
Background: Neuroendocrine tumors-are relatively rare but increasingly diagnosed malignancies originating from diffuse neuroendocrine cells, most commonly affecting the gastroenteropancreatic system. Due to their long asymptomatic development and low incidence, pose a diagnostic and therapeutic challenge for physicians. Recently, the role of nuclear medicine has been growing not only in the diagnostic stage but also in treatment. Systemic radionuclide therapy using somatostatin analogs labelled with the radioisotope lutetium-177 is becoming increasingly common in patients with advanced-stage disease. Currently, most patients receive a standard activity of therapeutic radiopharmaceuticals. Recent clinical studies provide increasing evidence of a close relationship between the absorbed radiation dose in pathological lesions and the therapeutic effect of radioisotope therapy. Internal dosimetry is used to measure the doses of ionising radiation absorbed by the patient after administration of the radiopharmaceutical. The lack of individual internal dosimetry prior to therapy means that only a small fraction of patients receive optimal doses of radioactivity, which is markedly different from external beam radiotherapy planning. Methods: A narrative literature review was conducted using the PubMed/MEDLINE and Embase databases, focusing primarily on publications from the last years. The search strategy included combinations of keywords related to peptide receptor radionuclide therapy and dosimetry, such as "Lutetium-177", "neuroendocrine tumors", "dosimetry", "PRRT", "systemic radionuclide therapy" and "artificial intelligence". Particular emphasis was placed on recent prospective clinical studies, multicenter investigations, systematic reviews and consensus documents published by major nuclear medicine societies, including the European Association of Nuclear Medicine (EANM) and the Society of Nuclear Medicine and Molecular Imaging (SNMMI). Seminal earlier publications considered essential for understanding the development of dosimetry concepts and clinical implementation were also included. Results: This study confirms the existence of a clinically significant dose-response relationship in 177Lu-PRRT. Higher absorbed doses to tumour lesions are associated with longer progression-free survival. The lack of individualized internal dosimetry prior to therapy means that only a small proportion of patients receive optimal radiation doses. Simplified dosimetric approaches with a reduced number of imaging time points, together with emerging artificial intelligence-based tools, appear promising for reducing the complexity of the dosimetry process. Conclusions: The aim of this study was to analyse the current literature on the role of internal dosimetry in the treatment of neuroendocrine tumors using the radioisotope lutetium-177. Available data support the clinical relevance of individualized dosimetry and highlight its potential to optimize both therapeutic efficacy and treatment safety.
Immunoglobulin (Ig) G4-related disease (IgG4-RD) is commonly treated with glucocorticoids and B-cell depletion, but cumulative toxicity and relapse underscore the need for alternative approaches. We evaluated the efficacy, safety, and immunological effects of Bruton's tyrosine kinase (BTK) inhibition with zanubrutinib in active IgG4-RD. In this phase 2, open-label, proof-of-concept trial, 10 participants with lacrimal and submandibular gland IgG4-RD received zanubrutinib 80 mg twice daily for up to 24 weeks without glucocorticoid induction or background immunosuppression. The primary endpoint was change in lacrimal and submandibular gland volume at week 24 by blinded fluorodeoxyglucose positron emission tomography/magnetic resonance imaging (evaluable n = 8). Secondary endpoints included metabolic imaging parameters, clinical disease activity, serologic biomarkers, and safety. Single-cell RNA sequencing with immune repertoire profiling was performed to define cellular mechanisms. At week 24, mean gland volume decreased by 46.7% (lacrimal) and 29.9% (submandibular) (both P = .008), with concordant reductions in total lesion glycolysis (-91.6 g; P = .05) and total metabolic lesion volume (-20.7 cm³; P = .05). Clinical disease activity improved (IgG4-RD Responder Index -6.0 points; P = .01), alongside reductions in serum IgG4 (-417 mg/dL; P = .008) and circulating plasmablasts. Imaging and serologic changes were strongly correlated. Single-cell analyses demonstrated treatment-associated modulation of B-cell transcriptional programmes, reductions in IgG4-skewed plasmablasts, and attenuation of cytotoxic CD4⁺ T cells. Adverse events were predominantly mild; 1 serious event (COVID-19) occurred off treatment. Zanubrutinib monotherapy produced substantial imaging-defined and clinical improvements in active glandular IgG4-RD without glucocorticoid induction. BTK inhibition was associated with modulation of B-cell differentiation states and downstream immune programmes, supporting its development as a steroid-sparing, non-B-cell-depleting therapeutic strategy.
Optimized reconstruction algorithms can enhance the image quality and signal-to-noise ratio (SNR) of positron emission tomography (PET) images. This study aimed to compare the impacts of the ordered subset expectation maximization (OSEM) and regularized expectation maximization image reconstruction (HYPER Iterative) algorithms on the image quality and epileptogenic zone (EZ) detection sensitivity in fluorine-18-DPA-714 (18F-DPA-714) images of drug-resistant focal epilepsy patients. Drug-resistant focal epilepsy patients who underwent presurgical brain 18F-DPA-714 PET imaging between May 2023 and July 2024 were retrospectively included. PET images were reconstructed using the OSEM (number of iterations ranging from 2 to 6) and HYPER Iterative algorithms [penalization factors (β) of 0.3, 0.5, and 0.8]. Image quality was evaluated by subjective (overall image quality, image noise, and EZ conspicuity) and quantitative assessment [mean standardized uptake value (SUVmean), SUVmax, contrast-to-noise ratio (CNR), SNR, and standardized uptake value ratio (SUVr)]. The sensitivity of detecting EZs in optimal PET images reconstructed by the OSEM and HYPER Iterative algorithms was compared, with histopathology as the gold standard. Thirty patients were included. The subjective assessment revealed that the HYPER Iterative algorithms outperformed OSEM in overall image quality, with HY0.8 achieving the highest score. Among OSEM algorithms, O5 had the top score and outperformed O6 significantly. Image noises scores were higher in HY0.5 and HY0.8 than in HY0.3 and OSEM algorithms. EZ conspicuity scores were better for HYPER Iteratives, O5, and O6 than for O2 - O4. Quantitative assessment revealed that the SNR was higher in O2, HY0.5, and HY0.8 than in O6, with HY0.8 significantly higher SNR than O5. The SUVr was higher in O6 and HY0.3 compared with O2. No significant differences were observed in SUVmean, SUVmax, mean standardized uptake value of cerebellar gray matter (SUVb) or CNR across all groups. For EZ detection, O5 and HY0.8 were optimal for the OSEM and HYPER Iterative algorithms, respectively. HY0.8 achieved a numerically 13.3% higher sensitivity for EZ detection than O5(80.0% vs. 66.7%), though this difference did not reach statistical significance (p = 0.125). The HYPER Iterative algorithm is capable of reducing image noise, enhancing EZ conspicuity, and improving the overall image quality in 18F-DPA-714 PET for patients with drug-resistant focal epilepsy, further increasing visual interpretability. A penalization factor of 0.8 results in a numerical improvement in the detection sensitivity of EZ, thus facilitating readers' confidence in detecting EZs. The online version contains supplementary material available at 10.1007/s13139-025-00981-7.
Metaiodobenzylguanidine (MIBG) scintigraphy is an established imaging modality for evaluating neuroendocrine tumors, particularly pheochromocytoma and neuroblastoma; however, false-positive uptake remains a significant diagnostic challenge. This systematic review aimed to summarize reported pitfalls in MIBG scintigraphy, focusing on false-positive findings. PubMed and Scopus were searched for studies published between 1980 and November 2024 that reported false-positive MIBG uptake, including case reports, case series, and observational studies. Study selection followed predefined criteria, with additional articles identified through reference screening. Risk of bias was assessed using the Quality Assessment of Diagnostic Accuracy Studies-2 (QUADAS-2) tool for diagnostic accuracy studies and the CAse REport (CARE) checklist for case reports and case series. Ninety-one studies were included, comprising 170 documented false-positive cases. False-positive uptake most frequently involved the abdomen and pelvis (57.6%), followed by soft tissues and the skeletal system (31.8%), chest (7.1%), and head and neck regions (3.5%). Reported etiologies included non-tumoral conditions, benign and malignant tumors, and physiological or variant uptake patterns. The adrenal glands were the most common site of false-positive uptake, substantially complicating the diagnosis of pheochromocytoma. Awareness of these pitfalls, combined with careful clinical correlation, hybrid imaging, and complementary diagnostic modalities, is essential to improve diagnostic accuracy and reduce misinterpretation.
Tumor heterogeneity remains a major challenge in oncology, influencing prognosis and treatment response. Novel PET-derived spatial radiomic features, including the normalized hotspot-to-centroid (NHOC) and normalized hotspot-to-perimeter (NHOP) distances, aim to quantify intratumoral metabolic heterogeneity and spatial distribution of metabolic activity. These biomarkers may provide biologically interpretable indicators of tumor aggressiveness and patient outcomes. The aim of this study was to systematically evaluate the prognostic and predictive performance of NHOC and NHOP derived from PET imaging across different tumor types. A systematic review and meta-analysis were conducted according to PRISMA guidelines. PubMed/MEDLINE and Google Scholar were searched up to March 2026. Studies evaluating NHOC or NHOP extracted from PET imaging in cancer patients were included. Hazard ratios (HRs) for overall survival (OS) and progression-free survival (PFS) were pooled using random-effects models. Ten studies involving 1,496 patients were included. Elevated NHOC was statistically significantly associated with worse OS (HR of 2.313, 95% CI [1.418, 3.775], p = 0.001). Subgroup analysis showed a strong association in lung cancers (2.864, 95% CI [2.018, 4.063], p < 0.001), while results in breast cancer and glioma were inconclusive due to limited data. Decrease in NHOP demonstrated a non-significant trend toward poorer PFS (HR = 2.92, 95% CI: 0.71-11.92). Predictive models based on NHOC showed moderate performance (AUC up to 0.77). PET-derived spatial biomarkers, particularly NHOC, show promise as prognostic indicators of tumor aggressiveness and survival outcomes. Larger multicenter studies with standardized imaging protocols are needed to validate their clinical utility.
Despite advancements in therapeutic cancer vaccines, clinical translation has been hindered by limited efficacy, with Sipuleucel-T remaining the only FDA-approved therapeutic cancer vaccine to date. However, recent advances in personalized mRNA vaccines, such as Moderna's mRNA-4157 and BioNTech's autogene cevumeran, have demonstrated significant reductions in recurrence risk and improved survival across several cancer types, renewing optimism in the field. Personalized cancer vaccines leverage patient-specific tumor antigens to initiate potent and targeted immune responses. This review outlines various classes of personalized vaccines, including DNA-, mRNA-, peptide-, dendritic cell-, and whole-cell-based platforms, and examines the immunological challenges they face, such as tumor heterogeneity, immunosuppressive microenvironments, and inadequate immune memory. To address these limitations, both conventional and nanotechnology-enhanced delivery systems have been developed. Notably, nanovaccines constructed from lipid-polymer hybrids, biomimetic membranes, and stimulus-responsive materials enable codelivery of neoantigens and immunostimulatory agonists, promoting enhanced lymph node targeting, dendritic cell activation, and antigen cross-presentation. Furthermore, biomimetic formulations incorporating autologous tumor membranes preserve native antigenic diversity and allow dynamic adaptation to evolving tumors. When integrated with artificial intelligence for antigen selection and multiomics for patient stratification, these platforms accelerate vaccine design and improve precision. Combination regimens with immune checkpoint inhibitors or other agents further potentiate efficacy and promote durable antitumor immunity. Increasing clinical evidence, especially in melanoma and pancreatic cancer, underscores the potential of these strategies to induce long-term protection and reduce recurrence. Overall, next-generation personalized cancer vaccines are advancing the transition from reactive treatment to proactive, precision-controlled cancer immunotherapy.
Head and neck cancer (HNC) threatens communication through its impact on voice and speech. The neural systems linking depressive symptoms with perceived voice handicap remain poorly characterized. We examined whether these symptom domains show dissociable associations with regional brain metabolism. In this cross-sectional 18F-FDG PET, we studied 63 HNC patients following diagnosis. Regional glucose metabolism (standardized uptake value ratios) was quantified in a priori regions of interest: Broca's area, Wernicke's area, left and right insula, and bilateral hippocampus. Depressive symptoms (Zung Self-Rating Depression Scale) and perceived voice handicap (Voice Handicap Index) were assessed. Spearman correlations with false discovery rate correction, partial correlations, and unique variance analyses were performed. Depression and voice handicap were strongly correlated (ρ = 0.64, p < 0.001) and exhibited partially dissociable metabolic correlates. Depressive symptoms were associated with reduced metabolism in Broca's area (ρ = -0.33, pFDR = 0.041) and higher metabolism in the left insula (ρ = 0.36, pFDR = 0.039), with graded insular elevation in moderate-severe depression (+12.5%, p = 0.008). These associations remained significant after age/sex adjustment and nominally significant after tumor-site adjustment. In exploratory analyses, voice handicap showed a negative association with hippocampal metabolism that did not survive FDR (ρ = -0.28, pFDR = 0.137) but reached significance after covariate adjustment (ρ = -0.34, p = 0.016). Depression and voice handicap in HNC show partially dissociable associations with regional brain metabolism despite clinical co-occurrence. Routine clinical imaging may be leveraged to generate hypotheses for psychosomatic and rehabilitation research.
Amyloid PET/CT is increasingly used in the diagnostic evaluation of cognitive disorders; however, its real-world impact on patient and caregiver experiences remains insufficiently characterised. This study examined how amyloid PET results influence perceived diagnosis, emotional responses, behavioural adaptations, and daily life organisation. We conducted a retrospective-prospective monocentric mixed-methods study integrating clinical data from the CLEMENS registry, routine amyloid PET interpretations, and a structured telephone survey that included open-ended questions. Adults who underwent amyloid PET between 2015 and 2024 were contacted, and therapeutic representatives were interviewed when patients were unable to participate. Quantitative outcomes were compared between PET-positive and PET-negative individuals using Fisher's exact and nonparametric tests. Qualitative responses were analysed thematically and integrated with quantitative findings. Twenty-one individuals completed the survey (12 PET-positive and 9 PET-negative individuals). PET-positive individuals were slightly older (72.4 vs. 68.9 years) and had lower MoCA scores (19.4 vs. 22.2) than PET-negative individuals. They exhibited major neurocognitive disorders (5/12 vs. 1/9) and behavioural adaptations (73% vs. 33%) more frequently. Emotional responses differed markedly by PET status: depressive mood (64% vs. 12.5%; p = 0.059) and shock/distress (42% vs. 22%) were more frequent among PET-positive individuals, whereas relief was more common among PET-negative individuals (50% vs. 18%). Qualitative themes indicated that PET-positive results were perceived as turning points associated with loss of autonomy and future uncertainty, while PET-negative results commonly provided reassurance and stability. Amyloid PET/CT influences not only diagnostic understanding but also the emotional, behavioural, and practical experiences of patients and representatives. PET-positive results are associated with a greater psychosocial burden, whereas PET-negative results provide reassurance. As amyloid PET becomes central to diagnostic and therapeutic pathways, patient-centred communication and supportive counselling will be essential.
Lutetium-177-PSMA-617 (Lu-PSMA) radioligand therapy (RLT) is established in metastatic castration-resistant prostate cancer (mCRPC), with regulatory approvals based on the VISION and TheraP trials. Subsequent trials have extended the evidence to taxane-naive mCRPC (PSMAfore) and demonstrated that combining Lu-PSMA with enzalutamide yields a significant overall survival benefit over enzalutamide alone (ENZA-p). However, higher and more homogeneous PSMA expression in treatment-naive disease, combined with lower tumor burden and preserved bone marrow reserve, provides a biological rationale for deploying RLT earlier in the disease course. In metastatic hormone-sensitive prostate cancer (mHSPC), the Phase III PSMAddition trial reported improved radiographic progression-free survival when Lu-PSMA was added to standard androgen deprivation therapy (ADT) plus androgen receptor pathway inhibitor (ARPI), and the Phase II UpFrontPSMA trial demonstrated enhanced biochemical responses with Lu-PSMA induction before docetaxel. In oligometastatic and oligorecurrent disease, the BULLSEYE and LUNAR trials have shown progression-free survival benefits, raising the possibility of deferring androgen deprivation therapy and its associated morbidity. Meanwhile, next-generation radionuclides, including actinium-225 (WARMTH) and the dual beta-Auger emitter terbium-161 (VIOLET), are entering clinical development to address the radiobiological limitations of Lutetium-177. This review synthesizes the evidence for PSMA-targeted radioligand therapy across the prostate cancer disease continuum and discusses patient selection, treatment sequencing, and the access and cost-effectiveness considerations that will shape adoption in earlier disease settings.
The cadmium-zinc-telluride (CZT) camera has generated renewed interest in single-photon emission computed tomography (SPECT)-myocardial flow reserve (MFR) assessments in patients with coronary artery disease (CAD). However, discussion regarding the relationship between MFR and conventional indices or their role in improving CAD diagnosis is limited. This study evaluated how left ventricular (LV) phase analysis indices-such as bandwidth, phase standard deviation (SD), and entropy-calculated from myocardial perfusion imaging (MPI), contribute to improving CAD diagnosis by comparing them with MFR. We retrospectively analyzed the clinical images of 60 patients (42 men and 18 women; mean age, 66.8 ± 10.0 years) who underwent dynamic CZT SPECT for suspected CAD. Receiver operating characteristic (ROC) and multivariate ROC analyses were used to assess the improvement in CAD diagnosis when using MPI (summed stress score [SSS]) alone, MPI combined with MFR, or MPI combined with LV phase analysis indices (stress indices). The area under the ROC curve (AUC) for SSS using MPI alone was 0.87, whereas that for SSS combined with MFR was 0.93, confirming a significant improvement in CAD diagnosis (p = 0.04). The AUCs for bandwidth, phase SD, and entropy combined with SSS were 0.93, 0.91, and 0.91, respectively, with no significant differences when compared individually. However, combining all LV phase indices yielded a significant improvement, with an AUC of 0.94 (p = 0.04). A similar diagnostic accuracy for CAD was obtained by combining the SSS from MPI with either the MFR or the three phase indices.
The reliability of PET radiomic features is fundamentally constrained by inherent stochastic noise originating from Poisson-distributed photon-counting statistics and additional fluctuations introduced by electronic signal processing. To assess its impact, we repeated PET scans under identical acquisition and reconstruction conditions to evaluate the consistency of radiomic feature measurements and quantify the noise-induced variability as a repeatability measure. An American College of Radiology-accredited (ACR) phantom was utilized to evaluate measurement variability in PET radiomic features under two scenarios: uniform (Experiment 1) and non-uniform (Experiment 2) radiotracer distributions, employing 18F-FDG and 68Ga-PSMA tracers. A total of 93 radiomic features across six categories were extracted using PyRadiomics. Each experiment was repeated ten times, and measurement variability was quantified using the coefficient of variation (CV). Features were benchmarked and classified into three quality grades of repeatability: Grade A (CV < 10%), Grade B (10% ≤ CV < 20%), and Grade C (CV ≥ 20%). Experiment 1 identified 22 Grade A and 14 Grade C features, whereas Experiment 2 yielded 30 Grade A and 15 Grade C features. Notably, CV values exhibited substantial variability. In Experiment 1, CV ranged from 0.17% to 205% for 18F-FDG and from 0.14% to 140% for 68Ga-PSMA. Similarly, in Experiment 2, CV spanned from 0.16% to 230% for 18F-FDG and from 0.23% to 79% for 68Ga-PSMA. By isolating inherent stochastic noise, this study applied the coefficient of variation to evaluate measurement variability across repeated PET scans and benchmarked radiomic features into three repeatability quality categories. While some features demonstrated high noise resilience, others showed poor repeatability across repeated scans. These findings highlight the need to establish repeatability standards to guide the development of noise-robust features that support trustworthy imaging biomarkers for clinical applications. The online version contains supplementary material available at 10.1007/s13139-026-01008-5.
The current BTS guidelines recommend evaluation of suspicious pulmonary nodules using [18F]FDG-PET/CT imaging, followed by Herder model risk stratification. However, it is based on limited imaging features, which may limit diagnostic accuracy. This study aims to develop a PET/CT-based deep learning (DL) model for malignancy probability estimation (AITO-PETCT-MP) and compare its performance to the Herder model and clinician performance. In a single-center retrospective study, we collected 533 indeterminate pulmonary nodules (268 malignant) with a mean diameter of 18.4 mm (SD ± 12.1) in 436 patients. Histopathological malignancy confirmation or a minimum 2-year benign national cancer registry follow-up served as the reference standard. Model diagnostic performance was compared against the Herder model and seven clinicians in a reader study on a test set of 161 nodules (80 malignant). AITO-PETCT-MP achieved an AUC of 0.78 [95% CI: 0.70-0.85] compared to the Herder model: AUC = 0.73 [0.65-0.80] (non-inferiority: p = 0.005). On average, experienced clinicians achieved an AUC of 0.80 [0.75-0.85]. Stratifying into BTS follow-up categories, the Herder model referred more benign nodules for potential direct treatment (26/81) than AITO-PETCT-MP and clinicians (both 3/81), while AITO-PETCT-MP and clinicians assigned more malignant cases to CT surveillance instead of direct treatment. AITO-PETCT-MP demonstrated non-inferior performance to the guideline-recommended Herder model, while only using imaging data. Diagnostic performance fell in the performance range of seven clinicians. Differences in BTS follow-up recommendations between the Herder model and clinicians suggest a difference in patient management compared to current clinical practice. Question How well can an imaging-only deep learning model estimate pulmonary nodule malignancy probability on [18F]FDG-PET/CT compared to the established Herder model and expert clinicians? Findings The model (AITO-PETCT-MP) performed non-inferior to the Herder model (AUC 0.78 vs 0.73, p = 0.005) and comparably to seven expert readers (AUC 0.74-0.87). Clinical relevance BTS-based follow-up stratification showed the Herder model referred more benign nodules to potential direct treatment than clinicians and AITO-PETCT-MP, while assigning fewer malignant cases to surveillance. This suggests a difference between Herder recommendations and current clinical practice.
Accurate staging is paramount for the effective management of medullary thyroid carcinoma (MTC). Imaging cancer-associated fibroblasts utilizing Gallium-68-labeled fibroblast activation protein inhibitors (FAPI) presents a novel approach for molecular imaging across various cancers, including MTC. This study aimed to compare the diagnostic accuracy of FAP-targeted imaging (68Ga-RTX-1363S, abbreviated as 68Ga-FAPI) with the established imaging standard, 18F-DOPA, for the detection of metastases in patients with MTC. This retrospective study compared 18F-DOPA- and 68Ga-FAPI PET/CT imaging using per-patient and per-lesion analyses in patients with recurrent MTC. Imaging findings were correlated with morphological imaging (CT/MRI/sonography) or histopathology as gold standard. Quantitative assessment included a comparison of standardized uptake values (SUVs) and tumor-to-background ratios (TBRs). Nine patients (mean age 55 years, range 34-80 years) with 62 lesions were included. Compared to 18F-DOPA, 68Ga-FAPI showed significantly higher lesion detection rates for lymph node metastases (true-positive [TP] rate: 100% [8/8] vs. 50% [4/8]; TBR: 10.9 vs. 6.9), lung metastases (TP rate: 85.7% [12/14] vs. 42.9% [6/14]; TBR: 6.1 vs. 3.6), and liver metastases (TP rate: 100% [22/22] vs. 23% [5/22]; TBR: 24.1 vs. 2.7). In direct comparison, 68Ga-FAPI detected a higher number of metastases with higher TBR values. 68Ga-FAPI PET/CT holds promise for improving staging in MTC. However, these findings necessitate confirmation in a larger, prospective study.