Internal dosimetry in radiopharmaceutical therapy (RPT) traditionally prioritizes methodological optimization driven by physical dose accuracy. However, even recommended multiple-time-point (MTP) dosimetry remains subject to uncertainties related to limited sampling schedules, pharmacokinetic, and curve-fitting modeling assumptions. In patients with metastatic castration-resistant prostate cancer (mCRPC) treated with [177Lu]Lu-PSMA-617 RPT, we explored an outcome-driven dosimetry optimization strategy by comparing MTP and single-time-point (STP) dosimetry, and identifying optimal time-points (TPs) for Hänscheid approximation based on therapy outcomes. Clinical and image data from 50 patients were retrospectively analyzed. Transient treatment-emergent adverse events (TEAEs) and prostate-specific antigen (PSA) response were monitored following CTCAE v5.0 and PCWG3. Cycle-level mean absorbed doses (MTPDtox, and 1d-, 2d-, 3d-STPDtox based on single SPECT acquisitions at 1-, 2-, 3-day(s) p.i.) were computed for organs-at-risk and whole-body tumors. Additionally, cumulative absorbed doses (STPDcum MTPDcum) were derived for whole-body tumors. Bone marrow STP dosimetry correlated significantly with anaemia grading (1d-STPDtox: Spearman’s ρ=0.35;2d-STPDtox: Spearman’s ρ=0.41; 3d-STPDtox: Spearman’s ρ=0.46; all padj<0.001), aligning with MTPDtox (Spearman’s ρ=0.43, p<0.001). Williams’ F-test confirmed no significant difference in correlation strength between STPDtox MTPDtox derived correlations at any TPs. For PSA response, both MTPDcum (Spearman’s ρ=-0.26, padj<0.05) and STPDcum (1d- 2d-STPDcum: Spearman’s ρ=-0.34, padj<0.001) showed significant correlations, without statistically significant differences between STPDcum MTPDcum derived correlations. The outcome-driven TP selection for Hänscheid-based STP dosimetry converges with the physics-based choice within 2-day p.i., the clinically driven approach offers an alternative and complementary strategy for dosimetry development and may facilitate its translation into clinical practice.

Purpose:Oncological patients regularly undergo PET/CT re-staging, which requires a report that outlines their current disease status and highlights relevant changes compared to the previous PET/CT. Large language models (LLMs) may be helpful with documentation in the future. This study is a pilot on LLM performance, focusing on test-retest stability and reproducibility. Methods:Three textbook melanoma follow-up cases of increasing complexity (involving one to eight organs) were selected. From standardized text-only prompts (no imaging data), follow-up reports were written by GPT-4o, Claude Sonnet 4 (each producing three independent revisions), and three nuclear medicine residents. This yielded nine reports per case (27 in total). Six blinded nuclear medicine experts (three internal, three external) performed test-retest evaluations of report quality and authorship identification. Results:The cosine similarity analysis revealed high intra-case coherence (mean: 0.599-0.727) regardless of authorship. The external human readers consistently rated reports higher than the internal human readers. The LLM-generated reports received comparable or superior ratings to human reports, with Claude achieving the highest external reader scores (mean 0.926, standard deviation 0.263, on a 0-1 scale). Human performance declined with case complexity, while Claude, in particular, improved. The external readers significantly preferred the LLM impressions (Fisher's exact test, p = 0.005). Neither the human nor LLM readers reliably identified authorship (balanced accuracy 0.343-0.500). Conclusion:In this pilot, blinded expert evaluation demonstrated that current LLMs can generate reports for melanoma [18F]fluorodeoxyglucose PET/CT of comparable quality to human-authored reports from text prompts in this study. High test-retest stability was obtained. Larger future studies will be required to confirm these findings.

Abstract This single-center subgroup analysis of a prospective trial (NCT04571086) at University Hospital Essen investigated tumor uptake and detection rates of [ 68 Ga]Ga-FAPI-46- versus [ 18 F]FDG PET in triple-negative breast cancer (TNBC) patients undergoing initial or follow-up staging. Lesions were recorded across anatomical regions, with detection efficacy, uptake values, tumor-to-liver ratios, and tumor volumes assessed. 22 patients were included (initial- n = 10, follow-up staging n = 12). No significant difference in region-based detection rates was observed between [ 18 F]FDG- (93%, 53/57 regions) and [ 68 Ga]Ga-FAPI-46 PET (93%, 53/57 regions) (McNemar p = 1.0). Three patients (25%) at follow-up staging showed an average tumor SUV mean ≥5 on [ 68 Ga]Ga-FAPI-46 PET; 2/3 showed SUV max >10 in most lesions. Mean tumor volume was higher for [ 68 Ga]Ga-FAPI-46, though not statistically significant at initial- (50.1 ± 120.8 mL vs. 11.9 ± 18.5 mL) and follow-up staging (102.2 ± 122.3 mL vs. 79.0 ± 90.4 mL). [⁶⁸Ga]Ga-FAPI-46 PET demonstrated comparable detection in TNBC; high tumor uptake in a subset, supporting its theranostic potential.

Despite significant progress in targeted cancer therapies and conventional imaging methodologies, the effective detection and treatment of solid tumours remain a major clinical challenge. This is thought to be caused by the complexity and heterogeneity found in the tumour microenvironment (TME), which significantly effects drug delivery and therapeutic response. Different levels of fibrosis, varying immune-cell infiltration, and disorganized vasculature form barriers for therapeutic approaches. However, in the next decade, radiotheranostics, defined here as the combined use of matched diagnostic and therapeutic radiopharmaceuticals, could present a targeted and flexible strategy for addressing some of the challenges caused by the TME. By combining molecular imaging with therapeutic delivery, it enables the in vivo visualization of TME features and the selective treatment of tumour and stromal compartments. This provides the unique opportunity to target tumour regions resistant to conventional therapies, including those shaped by (extracellular matrix) ECM stiffness, immune infiltration, or hypoxia. However, new strategies are needed to identify targets and evaluate their efficacy for more precise therapies. In this review, we will discuss why radiotheranostics is an ideal field for advancing the therapeutic approaches to solid tumours by incorporating the growing understanding of the TME. We will discuss how key microenvironmental features affect radiotracer distribution and treatment outcomes. We will highlight emerging tools including ECM- and immune-targeted imaging, patient-derived organoids, and organ-on-chip models which will be instrumental in developing physiologically relevant radiopharmaceutical therapies. Finally, we will discuss how spatial/single-cell transcriptomic approaches can support target discovery and allow for patient outcome assessment, with the aim of integrating microenvironment-aware insights into the development of novel radiotheranostic agents.

This study compared quantitative accuracy and image quality of post-radioembolization selective internal radiation therapy (SIRT) 90Y PET/CT using a long axial field-of-view (LAFOV) and standard axial field-of-view (SAFOV) system, focusing on absolute (ABS) versus relative (RS) scatter correction. Additionally, variance modeling and logistic regression were performed to define injected-activity thresholds ensuring reliable quantification at different acquisition durations. Phantom experiments were conducted on Biograph Vision Quadra (LAFOV) and Vision 600 (SAFOV) PET/CT scanners using NEMA IQ phantoms filled with 90Y activity. Reconstructions employed OSEM + TOF + PSF with ABS and RS scatter correction at 50 min (both scanners) and 10- and 1-minute (only for LAFOV). Metrics included NEMA recovery coefficients (RC) and background variability (BV). 44 post-SIRT patients (median age 70 y) were analyzed for image-based quantification of injected activity using whole-liver VOIs, across multiple durations (20, 15, 10, 5, 1 min). Metrics included Bland–Altman analysis, equivalence testing (± 10

Introduction Single-time-point (STP) dosimetry is practical for clinical routine but remains limited by population-based kinetic assumptions and strict late imaging time points. This study introduced a deep learning (DL) method for voxel-wise adjustment of STP dosimetry in 177Lu-PSMA radiopharmaceutical therapy (RPT), aiming to improve accuracy and scan-time flexibility. Methods A total of 317 post-therapeutic SPECT/CT scans from 21 patients with metastatic castration-resistant prostate cancer (mCRPC) receiving multiple cycles of [177Lu]Lu-PSMA-617 RPT were included. A decomposition-based DL model was developed to predict a voxel-wise scaling factor map from single-time-point SPECT/CT. The predicted map was multiplied with the quantified SPECT and a learned polynomial of imaging time to generate the corrected TIA map. Ground-truth TIA maps were generated using voxel-wise integrals of time-activity curves fitted by sums of exponential functions. Patient-wise five-fold cross-validation and comparative experiments were performed. Generalization was evaluated using an external dataset of 23 patients with mCRPC receiving [177Lu]Lu-PSMA-I&T RPT. Results Across all test data (n = 317), the average voxel-wise normalized root-mean-square errors were within 15.0% for all evaluated organs and tumors, with a global structural similarity of 0.9691. Using 24h SPECT/CT (n = 81), the method achieved mean absolute percentage errors of 17.9% ± 16.5% for kidneys, 15.8% ± 13.4% for liver, 21.0% ± 28.8% for spleen, 20.6% ± 18.2% for bone marrow, and 26.5% ± 19.1% for tumors. The MAPEs generally decreased with later imaging, reaching 10.3% ± 9.1% (kidneys), 10.2% ± 7.8% (liver), 21.3% ± 12.9% (spleen), 11.1% ± 8.7% (bone marrow), and 13.3% ± 11.2% (tumors) at 72h (n = 49). On the external dataset, the lowest MAPEs were 18.6% ± 12.0% for liver and 23.7% ± 20.9% for spleen at 24h, and 12.6% ± 14.4% for kidneys and 20.2% ± 12.2% for bone marrow at 48h. Conclusion The decomposition-based DL method enabled voxel-wise correction of STP dosimetry at time points including 24h, 48h, and 72h post injection, improved scan-time flexibility, and demonstrated the technical feasibility of instant STP dosimetry.

e13657 Background: The clinical application of Large Language Models (LLMs) in oncology is currently limited by opaque reasoning and the potential for "hallucinations," which pose safety risks for Multidisciplinary Tumor Boards (MTBs). Neuroendocrine tumors (NETs) specifically require precise interpretation of complex guidelines and recent primary evidence from clinical trials that are not already implemented in the guidelines. To address this need, we present PRISM (Personalized Recommendations via Integrated Synthesis & Modeling), a multi-agent LLM architecture designed to improve the quality, traceability, and reproducibility of therapy recommendations in NET care through a transparent, self-correcting workflow. Methods: PRISM employs a structured seven-stage fully automated, LLM-based workflow decomposing clinical reasoning into explicit steps, including guideline selection (e.g. ENETS & ESMO), initial patient case analysis, and search for matching trial evidence. A validation agent autonomously verifies matched evidence against existence (e.g., trial numbers), triggering correction loops to correct hallucinations or citation errors. In a feasibility study of 15 complex NET cases, we compared PRISM against a standard single-pass LLM (baseline). Performance was assessed by an expert using a structured framework evaluating clinical quality and safety given the patient’s organ functions, together with an overall quality rating (1–10) and implementation willingness (Yes/Maybe/No). Results: PRISM autonomously identified an average of 12 clinical trials per patient with potentially relevant primary evidence for treatment options, from which 66.7% were used in the treatment recommendations. The validation agent rejected 93% (14/15) of initial drafts due to hallucinations or citation errors. Autonomous re-processing (mean 2.6 iterations) led to full pass of 73% (11/15) of cases. In a blinded head-to-head comparison, the baseline model produced unsafe recommendations in 13% of cases, leading to hard implementation rejections (“No”) and expert-identified hallucinations. In contrast, PRISM reduced unsafe outputs to 6.7% and eliminated all hard rejections, achieving 100% implementation potential (60% “Yes”, 40% “Maybe”). PRISM achieved higher consistency in output quality (range: 5–9 vs. 1–10). Conclusions: PRISM addresses key limitations of standard LLMs by replacing opaque generation with a transparent, self-correcting validation loop. It systematically identifies relevant guidelines and primary evidence from clinical trials. In this feasibility study, the system ensured trustworthiness by enforcing strict evidence traceability. These findings support the prospective evaluation of multi-agent LLM systems as decision-support tools for treatment decisions in oncology, especially to bridge the expertise gap in non-specialized settings.

Accurate local staging of primary prostate cancer (PCa) is crucial for guiding therapeutic strategies. Current imaging methods, including MRI and PET/CT, may have variable accuracy in detecting key disease features. This head-to-head study compared the diagnostic performance of prostate-specific membrane antigen (PSMA), gastrin-releasing peptide receptor (GRPR) PET/CT, multiparametric MRI (mpMRI), and combined PET/CT plus MRI for local staging of intermediate-risk and high-risk PCa, along with their prognostic significance. Methods: In this retrospective analysis, patients with intermediate-risk or high-risk PCa underwent mpMRI, [68Ga]Ga-PSMA-617 PET/CT, and [68Ga]Ga-RM26 (GRPR-targeted) PET/CT before radical prostatectomy. Imaging findings were compared with whole-mount histopathology for local T stage, bilateral intraprostatic disease, extraprostatic extension, and seminal vesicle invasion. The prognostic value for predicting biochemical recurrence-free survival was assessed. Results: Among 81 eligible men, PSMA PET/CT showed higher overall accuracy than GRPR PET/CT (56% vs. 36%, P = 0.011) and improved detection of bilateral intraprostatic disease compared with mpMRI (72% vs. 54%, P = 0.024). In the pure acinar adenocarcinoma subgroup, PSMA PET/CT outperformed both mpMRI and GRPR PET/CT for overall accuracy (58% vs. 39% and 34%, P = 0.029 and 0.005, respectively). The combined PSMA PET/CT plus mpMRI further enhanced staging accuracy compared with mpMRI alone (61% vs. 41%, P = 0.002). Additionally, a local stage T3a or greater based on PSMA PET/CT plus mpMRI was an independent predictor of biochemical recurrence-free survival (hazard ratio, 4.277; P < 0.001), surpassing conventional clinicopathologic factors. Conclusion: PSMA PET/CT, especially when combined with mpMRI, offers superior accuracy in local staging and provides incremental prognostic value beyond standard clinicopathological parameters. Incorporating PET/MRI-derived local staging into clinical decision-making may improve patient stratification, guide surgical or focal therapy strategies, and ultimately enhance patient outcomes.

Objectives This study investigates the potential anatomical independence of PET imaging by assessing the feasibility of PET-only multi-organ segmentation using a deep learning-based approach. Methods In this retrospective study, the model was trained on a dataset of non-corrected PET images acquired from two commercially available total-body PET/CT scanners across two different centers. Ground-truth labels were generated using Computed Tomography (CT) images as input to TotalSegmentator and subsequently validated by two board-certified physicians, each with over a decade of experience. A U-Net-like model was trained on the non-corrected PET images (n = 938) and validated on two internal (n = 68) and nine external (n = 382) test datasets from diverse scanners, sites, and tracers. Results The model was designed to segment 17 anatomical structures; we report performance for nine clinically relevant organs: brain, heart, kidneys, liver, lungs, pancreas, spleen, thyroid gland, and urinary bladder. Across the two internal test sets, the mean Dice score was 0.828 (95% CI, 0.816–0.839) for scans acquired on the Biograph Vision Quadra (Siemens Healthineers) in Bern (n = 34) and the uExplorer (United Imaging) in Shanghai (n = 34). On external datasets, the model achieved a mean Dice of 0.787 (95% CI, 0.782–0.791) across five cross-scanner cohorts and 0.726 (95% CI, 0.698–0.755) across four cross-tracer cohorts. Conclusion Despite its potentially controversial nature, this methodology may broaden the application of PET imaging by avoiding redundant CT scans, which is particularly beneficial in multi-tracer studies or longitudinal monitoring.

FAP-targeted PET radiotracers have shown considerable promise across multiple malignancies, offering improved tumour detection and characterisation compared with conventional imaging. Nephro-urological diseases pose particular diagnostic challenges, as both standard imaging and currently available PET tracers are limited by nonspecific uptake and high background signal related to renal excretion. Given the prominent role of tumour-associated fibroblasts in several nephro-urological cancers, FAP-targeted PET may help address key limitations of current staging and response assessment in this setting. Moreover, upregulated FAP expression is also observed in a range of benign nephro-urological conditions, underpinning growing interest in FAP-targeted PET beyond oncology. This review summarizes the biology of FAP expression in the nephro-urological tract and synthesizes emerging evidence on oncological and non-oncological applications of FAP-targeted PET, with a focus on refining lesion characterization, guiding personalised management, and identifying novel theranostic opportunities in nephro-urology. Taken together, current evidence suggests that FAP-targeted PET may become a valuable complement to existing imaging strategies in nephro-urology, although its standalone clinical utility remains uncertain and requires confirmation in larger, prospective, disease-specific studies before routine adoption.

Fibroblast activation protein inhibitor (FAPI) imaging has become a promising approach in musculoskeletal oncology, with growing evidence supporting its application in both primary sarcomas and skeletal metastatic disease. Fibroblast activation protein (FAP) is expressed by cancer-associated fibroblasts and by tumor cells themselves in several sarcoma subtypes, while demonstrating limited expression in normal adult tissues. This biological profile has supported the development of radiolabeled FAP inhibitors, which provide high tumor-to-background contrast on PET imaging and enable theranostic applications with FAP-targeting radioligands. In sarcoma, available studies demonstrate higher FAPI uptake and improved lesion detection compared with [18F]FDG PET across most histological subtypes, with particularly notable advantages in low- and intermediate-grade tumors. However, FAPI uptake varies according to histological subtype. In metastatic bone disease, FAPI PET detects more lesions than [18F]FDG PET, particularly osteolytic and marrow-based metastases. The benefit is greatest in tumor entities with low [18F]FDG uptake, though at the cost of reduced specificity, as benign skeletal and fibroinflammatory processes also take up the tracer. As FAPI and [18F]FDG reflect different tumor biology with distinct strengths, their roles are increasingly regarded as complementary rather than competing. Beyond imaging, FAPI PET facilitates patient selection for FAP-targeted radioligand therapy, with early clinical experience demonstrating feasibility and encouraging preliminary evidence of disease control. Prospective studies are needed to define its clinical impact, establish standardized interpretation and response criteria, and determine the role of FAPI-based imaging and therapy in the management of musculoskeletal malignancies.

The clinical use of CD3–CD20–directed bispecific antibodies (BsAbs) has significantly improved outcomes of patients with relapsed/refractory LBCL. However, up to 50

An 80-year-old man with hypercalcemia and hyperparathyroidism was suspected of having a parathyroid adenoma, which was localized on [ 18 F]F-choline-PET/CT. Still, after surgical removal of the suspected adenoma with an intraoperative parathyroid hormone decrease of >50%, severe hypercalcemia recurred quickly. A repeated [ 18 F]F-choline-PET/CT raised suspicion of a single contralateral parathyroid adenoma or multiglandular disease. Given the inconclusive findings, a histological reassessment of the suspected parathyroid adenoma showed the presence of a singular vessel invasion, which alerted the diagnosis to parathyroid carcinoma. An additionally performed [ 18 F]FDG-PET/CT revealed a liver lesion and MRI confirmed multiple liver metastases, which could not be seen on [ 18 F]F-choline-PET/CT.

Dynamic long-axial-field-of-view (LAFOV) PET imaging offers unprecedented opportunities for quantitative assessment of tracer kinetics across the entire body. This review discusses the core technical challenges posed by LAFOV datasets in parametric image generation and introduces methodological developments from a statistical perspective, including arterial input function strategies, classical and flexible kinetic models (compartment models, spectral analysis, adiabatic approximation to the tissue homogeneity, and non-parametric models), graphical techniques (Patlak, Logan, and their variants), and emerging directions such as direct parametric reconstruction, dimension reduction, and deep learning. These methodologies are further linked to dynamic clinical protocols designed to shorten scanning times, enable multi-tracer injections, support low-dose imaging, and open avenues for novel tracer and drug development. Finally, we summarize the most recent software packages, particularly those tailored for LAFOV PET parametric imaging. These advances indicate that reliable parametric imaging holds promise for broader clinical adoption, grounded in robust modeling, multi-center validation, and ongoing software advancements.

Patient-specific dosimetry is currently a clinical need to evaluate lesion and organs at risk evolution in radiopharmaceutical therapy (RPT). Conventional dosimetry protocols are often time and/or computationally intensive, which dampers the applicability or real personalized dosimetry. Deep learning solutions for time-integrated activity to dose conversion present alternatives to costly Monte Carlo simulations while not relying on generic anthropomorphic models that are agnostic of the patient's anatomy. Artificial intelligence-enabled segmentation strategies support the evolution of personalized, image-guided RPT planning and monitoring. Quantification of radiopharmaceutical uptake and response at the lesion level enable clinicians to assess therapeutic efficacy and adapt treatment accordingly.

We investigated whether prompt changes in prostate-specific-antigen (PSA) levels within two days after the first cycle of prostate-specific-membrane-antigen radioligand therapy (PSMA-RLT) with [177Lu]Lu-PSMA-617 predicted treatment response and mean survival. In a retrospective study of 76 metastatic castration resistant prostate cancer (mCRPC) patients, we evaluated pretreatment PSA-values and their relative changes in PSA (dPSA) two days later. We tested for correlations between dPSA with long-term biochemical response (BCR) to treatment, using a priori criteria for relevant PSA decrease (dPSA < -10

Measuring the orthopositronium (oPs) lifetime could provide diagnostic information about the tissue microenvironment that goes beyond standard positron emission tomography (PET) imaging. In this study, three subjects received a dose of 148.8 MBq [68Ga]-Ga-DOTA-TOC, 159.7 MBq [68Ga]Ga-PSMA-11 and 420.7 MBq [82Rb]Cl. In addition to the standard protocol, the three subjects were scanned for 20, 40 and 10 minutes with a single-crystal interaction acquisition mode on a Biograph Vision Quadra (Siemens Healthineers) PET/CT that allows for an off-line three-photon event selection. Through a Bayesian fitting procedure we determined the oPs lifetime’s marginalized posterior distribution for selected organs. The methodology is extended to a hierarchical model in order to investigate possible common oPs lifetime distributions of the heart chambers in the [82Rb]Cl scan. Most interestingly, the mean values of the right heart chambers were higher than in the left heart chambers of the subject that received [82Rb]Cl: the 68% HDI of the atria are [1.15 ns, 1.72 ns] (left) and [1.46 ns, 1.99 ns] (right) with mean values 1.50 ns and 1.76 ns, respectively. For the ventricles we obtained [1.22 ns, 1.60 ns] (left) and [1.69 ns, 2.18 ns] (right) with mean values 1.44 ns and 1.96 ns. This might signal the different oxygenation levels of venous and arterial blood. This study demonstrates that in vivo oPs lifetime measurements on a commercial LAFOV PET/CT system are feasible at the organ level with 68Ga and [82Rb]Cl. Nevertheless, count statistics of three-photon events remains a major challenge for in vivo oPs lifetime measurements.

We aimed to compare various imaging-based response criteria in men with metastatic castration-resistant prostate cancer (mCRPC) treated with [177Lu]Lu-Prostate-specific membrane antigen radioligand therapy (LuPSMA). This retrospective study included 84 men who received a median of 4 [177Lu]PSMA cycles (IQR 2–5) and median of 24.3 GBq (IQR 14.9–32.9 GBq) at the Department of Nuclear Medicine at University Hospital Essen between March 2019 and May 2022. Response assessments were conducted comparing baseline PET/CT and PET/CT at 6–8 weeks after second cycle of LuPSMA using multiple criteria: Response Evaluation Criteria in Solid Tumors (RECIST) 1.1, the adapted Prostate Cancer Working Group Criteria 4 (aPCWG4) without follow-up confirmation, Positron Emission Tomography Response Criteria in Solid Tumors (PERCIST), the PSMA PET Progression (PPP), and Response Evaluation Criteria in PSMA-Imaging 1.0 (RECIP) with visual assessment or different quantitative volumetry methods (qPSMA, SUV ≥ 4). Responses were categorized as progressive disease (PD) or non-PD. The primary endpoint was the prognostic significance of these response criteria for overall survival, evaluated via Cox regression analysis. Harrell`s C-index was used for concordance of different imaging-based criteria and survival. A total of 34 (40.5

BACKGROUND:Fibroblast activation protein (FAP)-targeted tracers have emerged as promising agents for breast cancer imaging, with recent studies demonstrating PET performance comparable to or surpassing, that of [18F]FDG. Nevertheless, data comparing [68Ga]Ga-FAPI-46 and [18F]FDG uptake in hormone receptor and/or HER2-positive breast cancer (luminal-like vs HER2-positive) remain scarce. Aim of this study was to investigate the diagnostic performance of [68Ga]Ga-FAPI-46 versus [18F]FDG PET/CT in patients with hormone-receptor and/or HER2-positive breast cancer, and to evaluate the uptake of both tracers stratified by molecular subtypes (luminal-like vs HER2-positive). A sub-analysis of a prospective observational trial (NCT04571086) was conducted. Patients with histologically confirmed, hormone receptor- and/or HER2-positive breast cancer who underwent whole-body [68Ga]Ga-FAPI-46 and [18F]FDG PET/CT in the same week for initial staging or follow-up were included. [68Ga]Ga-FAPI-46 or [18F]FDG PET-positive lesions were defined as visually increased lesion uptake compared to adjacent organ background. Semi-automatic segmentation was performed to determine SUVmax, SUVpeak, TLRpeak, total number of lesions, total tumour volume, and total tumour SUV mean. Data were compared between molecular subtypes (luminal-like vs HER2-positive). RESULTS:Thirteen patients were included. Overall, the detection performance was comparable between [68Ga]Ga-FAPI-46 and [18F]FDG PET. The semi-quantitative analysis showed comparable mean uptake values in breast cancer lesions on [68Ga]Ga-FAPI-46 and [18F]FDG PET/CT (SUVmax: 13.4 vs. 12.9; TLRpeak: 5.6 vs. 4.5) and revealed no significant differences in the median lesion count (4.5 vs. 5), mean total tumour volume (71.5 vs. 73.2 mL), or mean total tumour SUVmean (5.4 vs. 5.4). No substantial differences between molecular subtypes (luminal-like vs. HER2-positive) were observed. CONCLUSION:In this small exploratory cohort, comparable uptake patterns in [68Ga]Ga-FAPI-46- and [18F]FDG-positive breast cancer lesions were observed across subtypes, underscoring the potential of [68Ga]Ga-FAPI-46 as a versatile imaging tool. Future studies in larger cohorts are warranted to explore the potential of FAP-targeted theranostics in different breast cancer subtypes. TRIAL REGISTRATION:68-Ga-FAPI-PET for Tumor Detection: A Prospective Observational Trial, NCT04571086, 09-15-2020, https://clinicaltrials.gov/study/NCT04571086 .