The increasing use of zirconium-89 ( 89 Zr) labeled radiopharmaceuticals has highlighted a lack of guidance regarding post-administration radiation restriction advice. This study aimed to estimate cumulative radiation exposure to close contacts following 37 MBq 89 Zr administration and to evaluate the need for radiation restriction advice. Published dose rate measurements following administration of 89 Zr-girentuximab were used to model cumulative exposure using physical decay only. Contact modeling incorporating realistic social and domestic patterns was adapted from established methodologies. Radiation restriction durations required to maintain exposure below the 300 µSv public dose constraint were determined iteratively for several contact groups. The same methodology was applied to fluorine-18 fluorodeoxyglucose ( 18 F-FDG) at 400 MBq for comparison. Without restrictions, all contact groups exceeded the 300 µSv dose constraint. Estimated cumulative doses were highest for partners (4552 µSv) and infants under 2 years (4516 µSv). Restriction durations to reduce cumulative dose below the constraint were 14 days for partners and infants, 11 days for children aged 2–5 years, 8 days for children aged 5–11 years, 3 days for work colleagues, and 2 days for other adults. No restrictions were necessary for 18 F-FDG administration. Restriction durations following 89 Zr-girentuximab administration were longer than anticipated from administered activity or dose rate measurements alone. By contrast, 18 F-FDG required no restrictions, highlighting the importance of physical half-life and emission energy in determining cumulative contact exposure. These findings highlight the importance of contact modeling and provide a practical framework for developing evidence-based radiation restriction advice following 89 Zr imaging.
Background: Undifferentiated pleomorphic sarcoma (UPS) is the most common adult soft tissue sarcoma subtype, with heterogeneous survival despite optimal surgery. We evaluated whether the preoperative maximum standardized uptake value (SUVmax) on 18 F-fluorodeoxyglucose PET/computed tomography ( 18 F-FDG PET/CT) is independently associated with overall survival (OS) in surgically resected UPS. Methods This retrospective cohort study included 79 patients with localized UPS who underwent curative resection between 2015 and 2023. Preoperative SUVmax, tumor grade (French Federation of Cancer Centers Sarcoma Group), tumor size, and Ki-67 index were assessed. Cox proportional hazards regression was used to identify independent prognostic factors for OS. Results The median age was 59 (range: 18–92) years; 53.2% of patients were male, and 59.5% of tumors were high grade. The median tumor size, Ki-67 index, and SUVmax were 60 mm, 60%, and 10, respectively. At a median follow-up of 58.9 months, median OS was not reached. In univariable analysis, tumor size ( P = 0.007), Ki-67 ( P = 0.001), SUVmax [hazard ratio: 1.08 per unit, 95% confidence interval (CI): 1.04–1.10, P = 0.001], and high grade ( P = 0.003) were associated with worse OS. In multivariable analysis, only SUVmax remained independently associated with OS (adjusted hazard ratio: 1.06 per unit, 95% CI: 1.02–1.10, P = 0.002), whereas grade, size, and Ki-67 did not retain significance. Conclusion Preoperative 18 F-FDG PET/CT SUV max is independently associated with OS in surgically resected UPS and provides prognostic information beyond conventional clinicopathological parameters.
Large language models (LLMs) are rapidly transforming clinical documentation workflow. This scoping review evaluates the emerging role of LLMs in nuclear medicine and molecular imaging, as recent literature suggests increasing evaluation of newer LLM architectures alongside GPT-4. We conducted a scoping review following Preferred Reporting Items for Systematic reviews and Meta-Analyses extension for Scoping Reviews guidelines. A comprehensive search of PubMed, Lens, and Google Scholar was performed from January 2023 to January 2026. Studies were included if they applied LLMs to nuclear medicine tasks (PET/computed tomography, single-photon emission computed tomography, and scintigraphy). General radiology studies were only included if separate nuclear medicine data were extractable. A total of 14 studies were included. While the GPT-4 family remains the most studied, recent literature includes increasing evaluation of non-GPT models, with several studies reporting comparable performance in specific tasks. In data extraction tasks, such as tumor–node–metastasis staging for breast cancer, Claude 3.5 Sonnet and DeepSeek have demonstrated performance comparable or superior to GPT-4o. The applications are primarily categorized into generative reporting (drafting impressions) and data extraction (structuring unstructured text). However, limitations persist: LLMs struggle with dynamic curve analysis (e.g. renal scintigraphy) and visual interpretation. Safety remains a concern, with one study reporting an 8% rate of potentially harmful responses in patient education scenarios. LLMs are evolving from experimental chatbots into capable ‘Clinical copilots’ that can significantly reduce administrative burden through automated reporting and data structuring. However, because of risks of hallucinations and domain-specific limitations in dynamic imaging, human oversight remains non-negotiable. Future research must focus on multimodal integration and multicenter prospective validation.
Objective The preoperative differentiation of clear cell renal cell carcinoma (ccRCC) from other renal masses remains challenging. Carbonic anhydrase IX (CAIX) is a highly specific target for noninvasive ccRCC imaging. We conducted a systematic review and meta-analysis to determine the diagnostic accuracy of CAIX-targeted PET/computed tomography (CT) for differentiating ccRCC from non-ccRCC renal masses. Methods A comprehensive search of PubMed, Embase, Scopus, and Registers (up to 30 September 2025) was supplemented by a bibliographic review. Studies evaluating CAIX-PET/CT against histopathologically were included. Of 528 records, five studies ( n = 544 patients) were included. Pooled sensitivity and specificity were synthesized using a univariate fixed-effects model; predictive values, likelihood ratios, and diagnostic odds ratio were also calculated. Results For the detection of histopathologically confirmed ccRCC (vs. non-ccRCC renal masses), the pooled sensitivity was 85.5% [95% confidence interval (CI): 81.6–88.7] and specificity of 86.5 (95% CI: 80.4–90.9). The hierarchical summary receiver operating characteristic model yielded an area under the curve of 0.894. Exploratory subgroup analysis suggested a potential difference in diagnostic performance between the tracer subtypes. Conclusion This review provides low-certainty evidence that CAIX-targeted PET/CT is a highly accurate, noninvasive tool for differentiating ccRCC from non-ccRCC renal masses in presurgical populations. Clinically, its high positive predictive value (94.6%) strongly ‘rules in’ disease, while its modest negative predictive value (~69%) limits its ability to ‘rule out’ malignancy. Further studies are warranted to validate these findings in broader clinical cohorts and for small-molecule agents.
