The Diagnostic Accuracy of <sup>18</sup>F-NaF PET/CT versus <sup>99m</sup>Tc Bone Scintigraphy for Detection of Skeletal Metastases: A Systematic Review and Meta-Analysis

Objectives:To compare the diagnostic accuracy of 18F-sodium fluoride (NaF) PET/CT (Positron Emission Tomography/Computed Tomography) and 99mTc(Technitium-99m) bone scintigraphy for detecting skeletal metastases across malignancies using head-to-head, patient-level studies. Methods:A systematic search of PubMed, Embase, Scopus, Web of Science, and Cochrane Library was conducted up to March 2025 following PRISMA 2020 guidelines. Eligible studies directly compared NaF PET/CT and 99mTc scintigraphy in the same patients and provided extractable 2×2 data. Quality was assessed with QUADAS-2 (quality assessment of diagnostic accuracy studies). Pooled sensitivity, specificity, and diagnostic odds ratios (DORs) were calculated using a bivariate random-effects model. Narrative synthesis was performed for studies without full 2×2 tables. Results:Six studies met inclusion, with four eligible for meta-analysis (prostate, breast, thyroid, renal, and nasopharyngeal cancers; n=468 patients). Pooled sensitivity and specificity of NaF PET/CT were 0.96 (95% CI, 0.91–0.99) and 0.93 (95% CI, 0.88–0.97), respectively, compared with 0.72 (95% CI, 0.63–0.80) and 0.81 (95% CI, 0.71–0.89) for 99mTc bone scintigraphy. The pooled DOR for NaF PET/CT was 342.1 versus 23.5 for bone scintigraphy. Heterogeneity was low-to-moderate. Narrative synthesis of two additional studies confirmed consistent superiority of NaF PET/CT. Conclusions:NaF PET/CT demonstrates significantly higher sensitivity, specificity, and overall diagnostic accuracy than 99mTc bone scintigraphy for detecting skeletal metastases across malignancies. Supported by multiple head-to-head studies and meta-analyses, NaF PET/CT is well positioned to replace bone scintigraphy as the reference standard in oncologic practice. Future work should assess cost-effectiveness, multicancer prospective validation, and integration with PET/MRI platforms. Introduction Bone metastases represent a frequent and serious complication of advanced malignancies, especially prostate, breast, lung, thyroid, and renal cancers, and are associated with significant morbidity, pain, and skeletal-related events that negatively impact survival and quality of life (1). Accurate and early detection of skeletal metastases is therefore essential for appropriate staging, prognostication, and treatment planning. For decades, 99mTc-methylene diphosphonate (MDP) bone scintigraphy has served as the reference standard for detecting osseous metastases. While widely available and cost-effective, its diagnostic performance is hampered by limited spatial resolution and suboptimal specificity, particularly in differentiating malignant from benign degenerative or inflammatory changes (2). The addition of SPECT or SPECT/CT has improved localization, but sensitivity remains modest for small or early lesions (2, 5). NaF PET/CT has emerged as a powerful alternative owing to its favorable pharmacokinetics, rapid blood clearance, and high affinity for hydroxyapatite crystals at sites of active bone remodeling (3, 4). These properties provide markedly higher bone-to-background ratios and superior image quality compared with 99mTc-based tracers. The integration of PET technology further enhances lesion detection through higher spatial resolution, tomographic capability, and quantitative assessment. Multiple head-to-head studies have demonstrated the superiority of NaF PET/CT over conventional scintigraphy. Even-Sapir et al. (5) showed significantly higher sensitivity in patients with high-risk prostate cancer, while Ota et al. (7) reported similar findings in differentiated thyroid carcinoma. Abikhzer et al. (8) extended these observations to breast cancer, and Gerety et al. (9) found that NaF PET/CT detected all lesions in renal cell carcinoma compared with only 29% by bone scintigraphy. More recently, Wang et al. (6) confirmed the superior diagnostic accuracy of NaF PET/CT in nasopharyngeal carcinoma. Jambor et al. (14) further validated its performance against MRI, demonstrating NaF PET/CT as the most accurate modality in high-risk prostate cancer. Early studies by Schirrmeister et al. (20) also highlighted its potential in breast cancer patients. Systematic reviews and meta-analyses corroborate these findings. Sheikhbahaei et al. (1) and Perera et al. (16) demonstrated pooled sensitivities exceeding 90% and specificities above 95% for NaF PET/CT in prostate cancer, markedly outperforming bone scintigraphy. Evangelista et al. (18) and Fan et al. (12) extended these observations across multiple malignancies, confirming consistently superior diagnostic accuracy. Reviews by Langsteger et al. (11), Mick et al. (10), Bastawrous et al. (4), and Lindenberg et al. (15) emphasize NaF PET/CT’s role in routine oncologic imaging, while Beheshti et al. (12) published EANM (European Association of Nuclear Medicine) guidelines formalizing technical protocols and clinical applications. Despite robust evidence, most prior reviews included lesion-based analyses or heterogeneous designs, limiting their direct applicability to clinical decision-making. To address these limitations, the present systematic review and meta-analysis focuses strictly on patient-level, head-to-head comparisons of NaF PET/CT versus 99mTc bone scintigraphy, providing pooled estimates of sensitivity and specificity across multiple cancer types. Methods Literature Search Strategy We conducted a systematic literature search of PubMed, Embase, Scopus, and Web of Science from inception through March 2025 to identify studies comparing NaF PET or PET/CT with 99mTc-based bone scintigraphy (planar or SPECT) for detection of skeletal metastases. Search terms combined synonyms for “sodium fluoride,” “PET/CT,” “bone scintigraphy,” “skeletal metastasis,” and cancer-specific keywords (e.g., prostate, breast, lung, thyroid, nasopharyngeal, renal cell carcinoma). Reference lists of retrieved ar-ticles and relevant reviews were also screened. The review process followed PRISMA-DTA guidelines, and a flow diagram is shown in Figure 1. Eligibility Criteria Studies were included if they: - Enrolled patients with histologically confirmed or clinically suspected malignancy at risk of bone metastasis; - Performed both NaF PET/CT (or PET alone) and 99mTc bone scintigraphy in the same patient cohort or in contemporaneous comparative cohorts; - Reported sufficient per-patient diagnostic data to construct a 2×2 contingency table (true positive, false positive, true negative, false negative). Studies were excluded: - if they only reported lesion-level data; - did not include a direct comparison, or duplicated previously published cohorts; - Studies without extractable patient-level data were retained for narrative synthesis only. Data Extraction and Quality Assessment Two reviewers independently extracted study characteristics, patient demographics, technical details of imaging protocols, and diagnostic accuracy outcomes (Table 1). For eligible studies, 2×2 contingency data were extracted to calculate sensitivity and specificity for each modality. If raw 2×2 data were unavailable but sufficient secondary data were provided (e.g., reported sensitivity and denominators), reconstructed values were derived and flagged accordingly. Studies lacking such reconstruction were included in the narrative-only table. Risk of bias was assessed using the QUADAS-2 tool, evaluating domains of patient selection, index test, comparator test, reference standard, and flow/timing. Statistical Analysis Sensitivity and specificity with 95% confidence intervals (CIs) were calculated using the Clopper–Pearson exact method. Forest plots were generated for each modality (Figures 2–3). Pooled estimates were obtained using a random-effects model (DerSimonian–Laird) applied on the logit scale. Summary results are provided in Table 2. A ROC scatter plot was constructed for NaF PET/CT (3) to illustrate study-level performance distribution. Narrative-only studies were synthesized descriptively to contextualize findings in additional cancer populations. Results The initial search identified 750 records (732 from databases and 18 from other sources). After removal of duplicates, 600 records were screened by title and abstract, of which 40 full-text articles were assessed. Ultimately, 7 studies were included in the quantitative synthesis and 2 were retained for narrative-only analysis (Figure 1). The 7 quantitative studies comprised 476 patients across diverse malignancies: prostate (two studies), lung (two studies), nasopharyngeal carcinoma (one study), differentiated thyroid carcinoma (one study), and breast cancer (one reconstructed dataset). Five were head-to-head comparisons, and two were reconstructed data-sets based on published accuracy values (Table 1). The two narrative-only studies evaluated renal cell carcinoma and a multicenter mixed cohort of prostate and breast cancer. Diagnostic accuracy results are presented in Table 2. Across individual studies, NaF PET/CT demonstrated consistently higher sensitivity than 99mTc bone scintigraphy. For instance, in prostate cancer cohorts, NaF PET/CT achieved sensitivities of 0.95–1.00 compared with 0.60–0.75 for bone scintigraphy. Similar patterns were observed in lung and nasopharyngeal carcinoma studies. Specificity was also generally higher for NaF PET/CT (typically 0.90–1.00) compared with bone scintigraphy (0.70–0.90), although greater variability was noted when equivocal scintigraphy scans were classified as positive. When pooled using a random-effects model, 18F-NaF PET/CT achieved a sensitivity of 0.96 (95% CI 0.91–0.99) and specificity of 0.93 (95% CI 0.88–0.97), whereas bone scintigraphy achieved a pooled sensitivity of 0.72 (95% CI 0.63–0.80) and specificity of 0.81 (95% CI 0.71–0.89). Forest plots illustrate the per-study accuracy estimates for sensitivity and specificity of both modalities (Figures 2 & 3), while the ROC scatter plot demonstrates that NaF PET/CT re