Wei Fu

Spread through air spaces (STAS) is associated with recurrence and unfavorable outcomes in lung adenocarcinoma, yet reliable preoperative identification remains challenging. This multicenter retrospective study includes 212 patients with 219 tumors across three institutions and develops a transformer‐based multimodal framework integrating eight complementary 2.5D PET and CT representations for preoperative STAS risk assessment. The model achieves an area under the receiver operating characteristic curve of 0.822 (95% CI, 0.725–0.920) in the independent external test cohort, with higher discrimination than the evaluated single‐modality and conventional fusion approaches. Beyond STAS classification, the continuous model‐derived risk score is associated with progression‐free survival after adjustment for pathological T stage, N stage, and maximum tumor diameter (HR, 4.24; 95% CI, 1.47–12.25; P = 0.008), although the limited number of progression events warrants cautious interpretation. Exploratory transcriptomic analyses across two public cohorts further nominate candidate programs involving cell adhesion, vesicle trafficking, complement activity, and metabolic remodeling. Together, these findings highlight the potential of multimodal PET/CT fusion to integrate morphological and metabolic information for noninvasive STAS risk stratification and support further prospective multicenter evaluation toward individualized surgical decision support.

Purpose This prospective study evaluated the clinical performance of [¹⁸F]AlF-NOTA-GL01 PET/CT, examined the relationship between tracer uptake and tumor FAP expression, and performed head-to-head comparisons with [¹⁸F]FDG and [¹⁸F]FAPI-42 PET/CT in exploratory subset. Methods Between August 2025 and March 2026, 106 patients completed [¹⁸F]AlF-NOTA-GL01 PET/CT and paired-tracer PET/CT; 103 underwent paired [¹⁸F]FDG PET/CT and 3 underwent paired [¹⁸F]FAPI-42 PET/CT. Lesion SUVmax and tumor-to-background ratio (TBR) were compared on a lesion basis. In tumors with available histology, FAP expression was quantified by immunohistochemistry and correlated with lesion SUVmax. Patient- and lesion-based diagnostic performance was assessed against a composite reference standard. Results The final cohort comprised 106 patients (55 men, 51 women; mean age, 61.3 ± 9.0 years), including 102 malignant tumors and 4 benign lesions. Lesion [¹⁸F]AlF-NOTA-GL01 SUVmax correlated strongly with tumor FAP expression (r = 0.754, P < 0.001) and increased across low, intermediate, and high FAP-expression tumors. In patient-based analysis, [¹⁸F]AlF-NOTA-GL01 showed higher sensitivity than [¹⁸F]FDG (98.9% vs. 90.8%, P = 0.039), with comparable specificity (62.5% vs. 62.5%, P = 1.000). In lesion-based analysis, [¹⁸F]AlF-NOTA-GL01 outperformed [¹⁸F]FDG in sensitivity (96.4% vs. 71.3%, P < 0.001), specificity (50.6% vs. 24.7%, P = 0.003), and accuracy (90.5% vs. 65.3%, P < 0.001). In an exploratory subset, [¹⁸F]AlF-NOTA-GL01 showed lower normal-organ background and higher lesion-to-background ratios than [¹⁸F]FAPI-42. Conclusion [¹⁸F]AlF-NOTA-GL01 uptake was associated with tumor FAP expression and demonstrated higher lesion-based sensitivity and accuracy than [¹⁸F]FDG in this single-center cohort, although specificity remained modest. These findings support further clinical development of [¹⁸F]AlF-NOTA-GL01 as a broadly applicable, scalable FAP-targeted PET agent.

Background: ¹⁸ F-NaF PET/CT is widely used in prostate cancer for detecting skeletal metastases, but the examination also contains structural and metabolic information potentially relevant to bone health assessment. Objective: This study evaluated feasibility of this dual-parameter approach for opportunistic bone health assessment in prostate cancer. Design: This was a retrospective cross-sectional. Methods: A total of 105 men with prostate cancer without bone metastasis underwent ¹⁸ F-NaF PET/CT. Stand-alone diagnostic CT-L1 was available in 76 patients and was used as an external structural reference for feasibility analysis and gray-zone stratification. Patient-level analyses used mean lumbar SUV _avg and mean PET/CT-derived trabecular attenuation across L1-L4. Vertebra-level analyses used all available same-level SUV _avg and same-level PET/CT-derived trabecular attenuation measurements. Patient-level metabolic-structural discordance on ¹⁸ F-NaF PET/CT was assessed. Results: PET/CT-derived L1 trabecular attenuation agreed strongly with stand-alone diagnostic CT-L1, with a correlation coefficient of 0.929 and an intraclass correlation coefficient of 0.916. In the paired diagnostic CT, cross-validated discrimination for low diagnostic CT-L1 thresholds was already high with low-dose CT alone, and adding SUV did not improve performance. However, among 41 patients in the stand-alone diagnostic CT gray zone of 110–160 HU, 19 (46.3%) were classified as low ¹⁸ F-NaF uptake. At the patient level, mean SUV _avg across L1-L4 correlated positively with mean PET/CT-derived trabecular attenuation across L1-L4 (r = 0.521, pavg (95% CI: 0.058 to 0.110; p18 F-NaF PET/CT may feasibly support opportunistic bone health assessment by integrating low-dose CT-derived trabecular attenuation and ¹⁸ F-NaF uptake within the same examination. The clinical contribution of ¹⁸ F-NaF uptake appears to lie more in refining gray-zone stratification and identifying metabolically discordant patients. These feasibility findings are hypothesis-generating and require prospective validation against DXA and fracture endpoints before clinical use.

Objective: Radioactive iodine-131 (I-131) is a standard therapy for Graves’ disease (GD), yet predicting therapeutic efficacy remains challenging due to complex biological variables. Conventional linear models often fail to capture the dynamic dose-response relationship between thyroid size and treatment outcomes. This study aimed to characterize the non-linear relationship between thyroid weight and I-131 efficacy and identify critical clinical thresholds influencing therapeutic success. Methods: A retrospective multicenter cohort study was conducted involving 888 records, of whom 612 GD patients received initial I-131 treatment between 2018 and 2024. Thyroid weight was estimated using thyroid scintigraphy. Restricted cubic splines (RCS) were employed to model potential non-linear associations, and piecewise threshold analysis was utilized to determine optimal inflection points regarding the correlation between thyroid weight and treatment non-remission. Results: Among 612 patients (mean age 41.2 ± 12.4 years; 73.9% female), 250 (40.8%) experienced non-remission after initial I-131 therapy. RCS analysis revealed a significant non-linear association between thyroid weight and treatment outcome in both unadjusted (P-nonlinear = 0.034) and DAG-confounder-adjusted models including study center (P-nonlinear = 0.032). In multivariable DAG-based logistic regression, thyroid weight (adjusted OR = 1.025, 95% CI: 1.017-1.034), female gender (adjusted OR = 1.67, 95% CI: 1.09-2.56), disease duration >2 years (adjusted OR = 1.97, 95% CI: 1.30-2.99), and study center (adjusted OR = 0.62, 95% CI: 0.42-0.93) were independent predictors. Piecewise analysis identified an exploratory inflection point at 46 g (adjusted LRT: P = 0.030). Mediation analysis demonstrated that RAIU at 3 h was a modest partial mediator (Sobel Z = 2.43, P = 0.015; mediated proportion: 5.8%). Conclusion: Thyroid weight demonstrates a significant non-linear association with I-131 treatment efficacy in GD, with an exploratory inflection point at approximately 46 g. The association is robust to DAG-based confounder adjustment including study center and is only modestly mediated by radioiodine uptake. These findings support individualized RAI dosing strategies incorporating thyroid weight thresholds.

Objectives Assessing tumor viability after locoregional therapy in hepatocellular carcinoma remains challenging on multiphasic computed tomography or magnetic resonance imaging. We evaluated the diagnostic performance of ¹⁸ F-FAPI-42 PET/CT for post-treatment viability assessment and compared it with multiphasic CT/MRI. Methods In this single-center retrospective study, 35 patients with hepatocellular carcinoma who underwent locoregional therapy and paired ¹⁸ F-FAPI-42 PET/CT and multiphasic CT/MRI between January 2022 and May 2025 were included. The reference standard was histopathology or at least 3 months of clinical follow-up. Diagnostic performance was assessed on patient-based and lesion-based analyses. Semiquantitative PET parameters and effects on staging were also evaluated. Results On patient-based analysis, ¹⁸ F-FAPI-42 PET/CT showed higher accuracy (88.6% vs 77.1%), sensitivity (90.5% vs 76.2%), and specificity (85.7% vs 78.6%) than multiphasic CT/MRI. On lesion-based analysis, ¹⁸ F-FAPI-42 PET/CT also performed better, with accuracy of 88.6% versus 77.3%, sensitivity of 86.4% versus 77.3%, and specificity of 90.9% versus 77.3%. Viable tumors showed higher SUV _max , SUV _mean , and tumor-to-background ratio than non-viable lesions (all P < 0.001). Visual PET/CT assessment achieved the highest area under the curve (0.88; 95% CI, 0.75–0.96). PET/CT also changed nodal staging in 1 patient, metastatic staging in 3 patients, and identified additional intrahepatic metastases in 11 patients. Conclusion ¹⁸ F-FAPI-42 PET/CT showed better overall performance than multiphasic CT/MRI for evaluating tumor viability after locoregional therapy in hepatocellular carcinoma and provided additional staging information that may influence management.

This study aims to evaluate the effectiveness of deep learning algorithms in simulating standard acquisition time images from shortened acquisition times i

Fibroblast activation protein (FAP) is highly expressed in cancer-associated fibroblasts and represents an attractive target for PET imaging. Here, we report [18F]AlF-NOTA-GL01, an improved FAPI-46 derivative bearing an N-methyl-benzenesulfonic acid side chain and labeled via an Al[18F]-NOTA platform. [18F]AlF-NOTA-GL01 was obtained with high radiochemical purity (>95%), high FAP binding affinity (IC50 = 2.94 nM), and high hydrophilicity (LogD = −3.06). In A549-hFAP xenografts, PET imaging showed high tumor uptake (7.45 ± 2.34%ID/g at 60 min), rapid renal clearance, and low hepatobiliary background, with strong concordance between PET-derived uptake and ex vivo biodistribution. In a first-in-human study, physiologic uptake decreased over time, whereas tumor uptake persisted to 120 min. The effective dose was 0.03 ± 0.01 mSv/MBq. Malignant lesions exhibited higher target-to-background ratios on [18F]AlF-NOTA-GL01 PET/CT than on [18F]FDG and [18F]FAPI-42. Collectively, these findings support the clinical translation potential of [18F]AlF-NOTA-GL01, which combines high affinity with excellent imaging contrast.

Rationale: Persistent midline radioiodine uptake after thyroid remnant ablation can mimic metastatic disease and lead to unnecessary re-treatment. We report a case of persistent iodine-131 ( ¹³¹ I) uptake in the thyroglossal duct region with undetectable thyroglobulin (Tg). Patient concerns: A 45-year-old woman was found to have thyroid nodules on routine ultrasound, and further evaluation supported papillary thyroid carcinoma. Diagnoses: After total thyroidectomy and central neck dissection, post-ablation and follow-up ¹³¹ I imaging demonstrated persistent focal uptake in the midline neck. Single photon emission computed tomography, computed tomography (SPECT/CT) localized the activity to the thyroglossal duct tract, while repeated serum Tg remained undetectable and no structural lesion was identified, supporting a benign thyroglossal duct remnant. Interventions: The patient underwent total thyroidectomy with central neck dissection, followed by 100 mCi ¹³¹ I remnant ablation after thyroid hormone withdrawal. She then received levothyroxine for thyroid-stimulating hormone suppression. Six months later, levothyroxine was withdrawn again for 4 weeks, and a low-dose diagnostic ¹³¹ I scan (3 mCi) with SPECT/CT was performed. Outcomes: Diagnostic ¹³¹ I imaging at 6 months again demonstrated persistent uptake at the same thyroglossal duct site, with no uptake in the thyroid bed, undetectable Tg, and no structural evidence of recurrent disease. The patient was managed conservatively with thyroid-stimulating hormone suppression and periodic biochemical and ultrasound follow-up. Lessons: For isolated midline ¹³¹ I uptake after ablation, SPECT/CT localization along the thyroglossal duct tract combined with undetectable Tg supports a benign thyroglossal duct remnant and favors observation with structured follow-up over empiric repeat ablation.

Glomus tumors (glomangiomas) are rare mesenchymal neoplasms derived from modified smooth muscle cells of the glomus body. Since they most commonly occur in the subungual regions of the extremities, primary involvement of the spine is exceedingly rare, particularly in atypical variants with borderline biological behavior. A 61-year-old man presented with a long-standing history of intermittent back pain that had recently worsened. Magnetic resonance imaging revealed a solid lesion involving the T10-T11 vertebral region with intermediate T1 signal intensity, mildly hypointense T2 signal, intralesional hemorrhagic foci, and marked heterogeneous enhancement causing spinal canal stenosis. ¹⁸F-FDG PET/CT demonstrated mild to moderate focal radiotracer uptake (SUVmax 5.6) corresponding to the spinal lesion, without evidence of other FDG-avid lesions suggestive of a primary malignancy or metastatic disease. Surgical resection was primarily indicated based on progressive symptoms and MRI findings; PET/CT findings did not alter the surgical decision but provided additional staging information, and histopathological examination established the diagnosis of an atypical glomus tumor. This case highlights the complementary role of MRI and ¹⁸F-FDG PET/CT in the evaluation of rare spinal glomus tumors. ¹⁸F-FDG PET/CT may provide supportive metabolic information and help assess the presence of additional FDG-avid lesions in atypical presentations.