This article is mentioned in:

Introduction

According to the latest global cancer data, there are expected to be 2.3 million new cases of breast cancer worldwide by 2022, accounting for 11.6% of all cancer cases. In 157 countries, breast cancer is the most common cancer among women (1). The incidence of bone metastasis (BM) in patients with breast cancer is ~8%, but can reach 30–85% in cases of advanced breast cancer (2). Breast cancer remains the leading cause of cancer-associated deaths among women. It is estimated that 666,000 women succumbed to breast cancer in 2022 worldwide, with metastatic disease being the main cause of death rather than the primary cancer (1,3). BM can disrupt bone metabolism, leading to bone-related events such as bone pain, pathological fractures, spinal cord compression, and hypercalcemia, which markedly affect the quality of life of patients and can even be life-threatening (4). A previous study showed that BM is a crucial factor affecting the prognosis of patients with breast cancer. The 5-year survival rate for patients with early-stage breast cancer without metastasis is as high as 90%, but once BM occurs, the 5-year survival rate drops to 10% (5). There is currently no cure for patients with breast cancer and BM; however, appropriate treatment can prolong survival and improve the quality of life of the patient. Therefore, it is important to accurately assess whether patients with breast cancer have BM (6).

X-ray imaging is routinely used to screen for bone disease, but is not effective for early BM detection because it only identifies lesions after a 30–50% loss of calcium (7). Bone scans are imaging techniques with high sensitivity but low specificity for the detection of bone lesions (8). Therefore, more sensitive and accurate methods are necessary to detect the BM associated with breast cancer earlier so that intervention can be initiated sooner, and thereby improve the survival time of the patient. Positron emission tomography/computed tomography (PET/CT) is an advanced diagnostic imaging technology that provides both metabolic information and precise anatomical localization. It has broad applications in the diagnosis, staging, location and treatment evaluation of various malignant tumors (9,10). Fluorine-18 fluorodeoxyglucose (18F-FDG) is a PET/CT tracer that is widely used for the diagnosis, staging and follow-up of patients with breast cancer due to its high diagnostic performance for lesions (11). A bone-specific radiotracer, 18F-sodium fluoride (18F-NaF), is effective in revealing changes in bone activity and has been widely used for the clinical detection of bone lesions (12,13). In patients with breast cancer and BM, metastases are predominantly osteolytic, but are osteogenic in 15–20% of cases (14–16). It has been shown that 18F-FDG is most sensitive in the detection of osteolytic metastases (17). Therefore, the present meta-analysis reviewed studies on the detection of BM in patients with breast cancer using PET/CT. The aim was to quantitatively evaluate and compare the diagnostic performance of 18F-FDG and 18F-NaF as PET/CT tracers in the detection of BM associated with breast cancer.

Patients and methods

Literature search to identify relevant studies

The present study was conducted in accordance with the Cochrane Collaboration's Systematic Review guidelines and Preferred Reporting Items for Systematic Reviews and Meta-Analyses requirements (18). The English literature on 18F-FDG PET/CT or 18F-NaF PET/CT in the detection of BM in breast cancer was retrieved from the PubMed (http://www.ncbi.nlm.nih.gov/pubmed) and Embase (https://www.embase.com/) databases. A systematic search was performed used multiple keywords: (‘PET/CT’ OR ‘PET-CT’ OR ‘positron emission tomography/computed tomography’ OR ‘positron emission tomography-computed tomography’) AND (‘breast cancer’ OR ‘breast carcinoma’ OR ‘mammary cancer’ OR ‘breast tumor;) AND (‘bone metastasis’ OR ‘skeletal metastases’ OR ‘osseous metastasis’) AND (‘18F-fluorodeoxyglucose’ OR ‘18F-FDG’ OR ‘18F-NaF’ OR ‘18F-fluoride’). The publication period was limited from January 1, 2000 to January 31, 2022. The final list of articles was supplemented by cross-checking the reference lists of all retrieved articles.

Study selection and quality assessment

Two reviewers independently screened all titles and read abstracts. The full text of the selected articles was reviewed to determine eligibility. Data extraction and evaluation were performed independently by two authors, with disputes resolved by a third reviewer. Studies included in the meta-analysis met all of the following criteria: i) Patients of any age with breast cancer at any stage of disease, regardless of treatment status; ii) 18F-FDG PET/CT or 18F-NaF PET/CT used in the imaging and characterization of BM in patients with breast cancer; iii) histopathological findings or CT, magnetic resonance imaging (MRI) or clinical follow-up over 6 months included as reference standards; iv) a 2×2 contingency table could be constructed using directly extracted data or by the calculation of true positive (TP), false positive (FP), false negative (FN) and true negative (TN) values based on the sensitivity, specificity, and positive and negative prediction values provided in the article. Exclusion criteria were: i) Studies with <10 patients with breast cancer; ii) studies where the PET/CT tracer was not 18F-FDG or 18F-NaF; iii) studies with multiple published data or subsets of data; iv) case reports, letters, editorials, reviews, animal studies, in vitro studies and studies without original data; v) studies presenting results from different imaging modalities jointly, or those in which it was not possible to distinguish between the test performance assessments of individual imaging modalities.

The QUADAS-2 tool was used for the quality assessment of diagnostic accuracy, covering four key areas: Patient selection, index tests, reference standards and the flow and timing of patients through the study (19).

Data extraction

Data extraction was performed independently by two investigators. For each relevant study, the following data were collected: i) Basic information such as the first author, publication year, country and sample size; ii) patient age, patient selection (continuous or non-continuous) and clinical background; iii) study design information; iv) examination results, including the numbers of TP, FP, TN and FN cases; v) parameters of the CT techniques used for 18F-FDG PET or 18F-NaF PET/CT. If there was a dispute between the reviewers, a third researcher evaluated all discordant items until a consensus was reached.

Statistical analysis

Stata software version 14.0 (StataCorp LP) was used to perform the statistical analysis. The diagnostic performance of 18F-FDG PET/CT and 18F-NaF PET/CT in the detection of BM in breast cancer was evaluated using specificity, sensitivity, positive likelihood ratio (PLR), negative likelihood ratio (NLR), diagnostic odds ratio (DOR) and summary receiver operating characteristic (SROC) curves based on TP, FP, FN, and TN values extracted from the included studies. The area under the curve (AUC) and 95% confidence intervals (CIs) were calculated. Analyses were performed using the DerSimonan-Laird method, a random-effects model, to calculate weighted mean pooled sensitivity, specificity, PLR, NLR and DOR and their corresponding 95% CIs. Variability was assessed graphically by plotting metrics with 95% CIs for each study separately in a forest plot. Values of pooled PLR >10 and DOR >100 indicate that a positive test result helps to confirm the presence of BM, while pooled NLR values <0.1 indicate that a negative test result helps to exclude BM (20). Hierarchical logistic regression models were used to estimate the sensitivity and specificity of the included studies. For each study included in a forest plot, the corresponding 95% CIs were shown to graphically represent the index being measured. Heterogeneity among the studies was assessed using Cochran's Q test and Higgins I2 test (20). In Cochran's Q test, P<0.05 indicated the presence of heterogeneity. The degree of heterogeneity was assessed using the following criteria: An inconsistency index (I2) <50% indicated low heterogeneity; an I2 of 50–80% indicated moderate heterogeneity; and an I2 >80% indicated high heterogeneity. Subgroup analyses for 18F-FDG PET/CT were performed based on study sample size, mean patient age, study design type, attenuation correction, minimum scan slice thickness, imaging system supplier and whether the study was patient- or lesion-based. Publication bias was assessed using a funnel plot and Deek's asymmetry test for both 18F-FDG PET/CT and 18F-NaF PET/CT (21). The potential publication bias was estimated using Egger's quantitative test. A two-sample Z-test was used to evaluate the difference in diagnostic performance between the two methods for the detection of BM in breast cancer, with P<0.05 considered to indicate a statistically significant result.

Results

Eligible studies and quality assessment

A literature search identified357 potentially relevant articles. After the exclusion of 107 duplicates, the screening of titles and abstracts led to the exclusion of a further 211 articles for being reviews or guidelines (n=15), conference papers (n=20), animal studies (n=14) or on irrelevant topics (n=91), or due to the full text not being available (n=71). After reading the full texts of the remaining 39 articles, 25 articles were excluded due to not being published in English (n=1), lacking the data to construct a 2×2 contingency table (n=6), or not being relevant to the area of interest (n=18). Finally, 14 articles on the diagnostic performance of 18F-FDG or 18F-NaF PET/CT in breast cancer BM met the criteria for inclusion in the present meta-analysis. The identification and selection process for the studies is shown in Fig. 1.

Figure 1. Flow chart of the study selection process. 18F, fluorine-18; FDG, fluorodeoxyglucose; NaF, sodium fluoride; PET/CT, positron emission tomography/computed tomography.

A total of 14 articles (15,22–34) were included in the study. These comprised 8 studies on 18F-FDG PET/CT, 5 studies on 18F-NaF PET/CT, and 1 study on both, including a total of 919 patients and 2,054 lesions. The sample sizes in the studies ranged from 20 to 150 patients, with mean ages ranging from 43.8 to 64 years. All 14 articles were published between 2010 and 2019, and comprised 8 prospective studies and 6 retrospective studies. Among these, 3 studies included patients with breast cancer who had previously received treatment, 3 studies included patients newly diagnosed with breast cancer who were clinically suspected of having BM, and 8 studies included both treated and newly diagnosed patients. The baseline characteristics of each study are presented in Table I, and the PET/CT parameters used in each study are presented in Table SI. The quality of each study was assessed using the QUADAS-2 tool. This assessment revealed that all studies met at least 5 of the 7 reference criteria, which included 4 items associated with the risk of bias, namely patient selection, index test, reference standard, and flow and timing, and 3 items associated with application concerns, namely patient selection, index test and reference standard; therefore, they were considered satisfactory (35). With regard to patient selection, 5 studies (22,24,25,29,32) were considered high-risk for reference standards, as only imaging and follow-up results were used as the reference standards. Additionally, one study had only a 2-month follow-up period (24), which was also considered high-risk. The risk of bias for flow and timing was unclear in all studies because the time interval between the index test and the reference standard was not reported. The results of the QUADAS-2 assessment are shown in Table SII.

Table I. Clinical characteristics and diagnostic results of the detection of bone metastases reported in each eligible study.

Table I.

Clinical characteristics and diagnostic results of the detection of bone metastases reported in each eligible study.

A, 18F-FDG PET/CT
				Study design					Patient-based analysis				Lesion-based analysis
First author, year	Country	No. of patients	No. of lesions				Clinical setting	Mean age (range), years									Reference standard	(Refs.)
				Prospective	Multicenter	Consecutive			TP	FP	FN	TN	TP	FP	FN	TN
Koizumi et al, 2019	Japan	120	-	No	No	Yes	New +	-	34	8	1	77	-	-	-	-	HP follow-up >6 months	(30)
							treated
Caglar et al, 2016	Turkey	150	-	No	No	Yes	New + treated	52 (27–85)	84	1	1	64	-	-	-	-	HP; follow-up >10 months	(26)
Hahn et al, 2011	Germany	28	129	No	No	Yes	New + treated	57.5 (35–78)	7	0	1	20	67	5	3	54	MRI follow-up	(22)
Damle et al, 2013	India	72	-	Yes	No	Yes	Treated	52 (30–77)	25	1	9	37	-	-	-	-	HP; consensus from MRI/thin-slice CECT/skeletal radiograph findings	(15)
Al-Muqbel, 2017	Jordan	35	-	No	No	Yes	Treated	48.1	25	0	9	1	-	-	-	-	Staging or follow-up 18F-FDG-PET/CT.	(25)
Botsikas et al, 2019	Switzerland	80	175	Yes	No	Yes	New + treated	48	6	0	3	71	18	0	8	149	HP; follow-up >12 months	(34)
Rager et al, 2018	Switzerland	25	109	No	Yes	Yes	New	55 (38–82)	10	0	2	13	43	0	48	18	Follow-up >21 months	(24)
Heusner et al, 2010	Germany	20	-	Yes	No	Yes	New + treated	54.5 (25.4–78.2)	7	0	0	13	-	-	-	-	Consensus from MRI and bone scan	(28)
Teke et al, 2015	Turkey	-	496	No	No	-	New	44.5 (28–81)	-	-	-	-	141	2	10	343	Follow-up >6 months	(32)
B, 18F-NaF PET/CT
				Study design					Patient-based analysis				Lesion-based analysis
First author, year	Country	No. of patients	No. of lesions				Clinical setting	Mean age (range), years									Reference standard	(Refs.)
				Prospective	Multicenter	Consecutive			TP	FP	FN	TN	TP	FP	FN	TN
Yoon et al, 2013	Korea	-	119	Yes	No	-	New + treated	55.6	-	-	-	-	49	36	3	31	HP; follow-up >12 months	(33)
Broos et al, 2018	Netherlands	118	-	Yes	No	-	New +	64	50	6	2	60	-	-	-	-	Follow-up >6 months	(23)
							treated
Damle et al, 2013	India	72	-	Yes	No	Yes	Treated	52 (30–77)	34	11	0	27	-	-	-	-	HP; consensus from MRI/thin-slice CECT/skeletal radiograph findings	(15)
Passah et al, 2017	India	-	199	Yes	No	-	New	43.8	-	-	-	-	178	0	0	21	99mTc-MDP skeletal scintigraphy	(29)
Abikhzer et al, 2016	UK	41	284	Yes	No	-	New + treated	58 (30–75)	21	3	0	17	73	6	7	198	HP; follow-up >33 months	(31)
Piccardo et al, 2012	Italy	39	662	Yes	-	-	Treated	60	27	0	0	12	491	11	51	109	Follow-up >12 months	(27)

[i] ¹⁸F, fluorine-18; FDG, fluorodeoxyglucose; PET/CT, positron emission tomography/computed tomography; TP, true positive; FP, false positive; FM, false negative; TN, true negative; HP, histopathological; MRI, magnetic resonance imaging; CECT, contrast-enhanced CT; ^99mTc-MDP, technetium-99m methylene diphosphonate; NaF, sodium fluoride.

Diagnostic accuracy

The 9 studies using the 18F-FDG PET/CT method had sensitivities ranging from 0.47 (95% CI, 0.37–0.58) to 1.0 (95% CI, 0.59–1.00) for the identification of breast cancer BM, and specificities ranging from 0.91 (95% CI, 0.82–0.96) to 1.0 (95% CI, 0.98–1.00). The pooled sensitivity and specificity of 18F-FDG PET/CT for the identification of BM derived from breast cancer were 0.88 (95% CI, 0.76–0.94) and 0.99 (95% CI, 0.97–1.00), respectively, as shown in Fig. 2. In addition, Cochran's Q test and Higgins I2 test indicated high heterogeneity in sensitivity (Q, 168.81, P≤0.01; I2, 93.48) and moderate heterogeneity in specificity (Q, 44.38, P≤0.01; I2, 75.21). The 6 studies describing the use of 18F-NaF PET/CT in the detection of breast cancer BM had sensitivities ranging from 0.91 (95% CI, 0.83–0.96) to 1.00 (95% CI, 0.84–1.00) and specificities ranging from 0.46 (95% CI, 0.34–0.59) to 1.00 (95% CI, 0.74–1.00). The pooled sensitivity and specificity were 0.98 (95% CI, 0.92–1.00) and 0.91 [95% CI, 0.76–0.97), respectively, as shown in Fig. 3. Cochran's Q test and Higgins I2 test also showed high heterogeneity in sensitivity (Q, 70.87, P≤0.01; I2, 90.12) and specificity (Q, 228.86, P≤0.01; I2, 96.94) among the studies. The pooled PLR and NLR for 18F-FDG PET/CT were 129.2 (95% CI, 27.1–616.4) and 0.13 (95% CI, 0.06–0.25), respectively. For 18F-NaF PET/CT, the pooled PLR and NLR were 10.9 (95% CI, 3.8–31.5) and 0.02 (95% CI, 0.01–0.1), respectively. The pooled DOR for 18F-FDG PET/CT in the diagnosis of breast cancer BM was 1,028 (95% CI, 244–4,330), while for 18F-NaF PET/CT, the pooled DOR was 489 (95% CI, 65–3,654), as shown in Table II. These data suggest that a positive result from FDG testing helps confirm the presence of BM, while a negative result from NaF testing helps to rule out BM. No significant difference in the DOR between 18F-FDG PET/CT and 18F-NaF PET/CT was detected. The area under the SROC curves for 18F-FDG PET/CT and 18F-NaF PET/CT were both 0.99 (95% CI, 0.98–1.00), as shown in Fig. 4.

Figure 2. Pooled sensitivity and specificity for fluorine-18 fluorodeoxyglucose positron emission tomography/computed tomography. Patient- and lesion-based analyses were separately included in the pooled analysis; *indicates lesion-based analysis. CI, confidence interval; df, degrees of freedom.

Figure 3. Pooled sensitivity and specificity for fluorine-18 positron emission tomography/computed tomography. Patient- and lesion-based analyses were separately included in the pooled analysis; *indicates lesion-based analysis. CI, confidence interval; df, degrees of freedom.

Figure 4. SROC curves of the diagnostic performance of different PET/CT imaging agents. SROC curves for (A) 18F-fluorodeoxyglucose PET/CT and (B) 18F-sodium fluoride PET/CT in the diagnosis of breast cancer bone metastases. SROC, summary receiver operating characteristic; PET/CT, positron emission tomography/computed tomography; 18F, fluorine-18; AUC, area under the curve; Sens, sensitivity; Spec, specificity.

Table II. Summary of the diagnostic performance characteristics of 18F-FDG and 18F-NaF PET/CT in breast cancer bone metastases.

Table II.

Summary of the diagnostic performance characteristics of 18F-FDG and 18F-NaF PET/CT in breast cancer bone metastases.

	18F-FDG PET/CT		18F-NaF PET/CT
Parameter	Estimate	95% CI	Estimate	95% CI
Sensitivity	0.88	0.76, 0.94	0.98	0.92, 1.00
Specificity	0.99	0.97, 1.00	0.91	0.76, 0.97
Positive likelihood ratio	129.2	27.1, 616.4	10.9	3.8, 31.5
Negative likelihood ratio	0.13	0.06, 0.25	0.02	0.01, 0.1
Diagnostic odds ratio	1,028	244, 4,330	489	65, 3,654
AUC	0.99	0.98, 1.00	0.99	0.98, 1.00

[i] ¹⁸F, fluorine-18; FDG, fluorodeoxyglucose; NaF, sodium fluoride; PET/CT, positron emission tomography/computed tomography; CI, confidence interval; AUC, area under the curve.

Publication bias

Deek's funnel plots for publication bias in the studies on 18F-FDG PET/CT and 18F-NaF PET/CT are shown in Fig. 5. The statistical significance of the slope coefficient for 18F-FDG PET/CT (P=0.02) is suggestive of publication bias. However, the slope coefficient for 18F-NaF PET/CT lacked significance (P=0.37), indicating a low possibility of publication bias. When analyzed using Egger's test, both 18F-FDG PET/CT and 18F-NaF PET/CT exhibited no evidence of publication bias (P=0.187 and P=0.123, respectively; Fig. 6).

Figure 5. Deeks' funnel plots for publication bias in the studies of different PET/CT imaging agents. Deeks' funnel plots for (A) 18F-fluorodeoxyglucose PET/CT and (B) 18F-sodium fluoride PET/CT with P-values for funnel plot asymmetry. PET/CT, positron emission tomography/computed tomography; 18F, fluorine-18; ESS, effective sample size.

Figure 6. Egger's publication bias plots for the studies of different PET/CT imaging agents. Egger's plots for (A) 18F-fluorodeoxyglucose PET/CT and (B) 18F-sodium fluoride PET/CT. PET/CT, positron emission tomography/computed tomography; 18F, fluorine-18; CI, confidence interval; SND, standard normal deviate.

Exploration of heterogeneity

The results of the meta-regression analysis are shown in Table III. Eight studies reported patient-based results for the performance of 18F-FDG PET/CT in the diagnosis of breast cancer BM, with a sensitivity of 0.89 (95% CI, 0.80–0.99) and a specificity of 0.99 (95% CI, 0.98–1.00). Four studies, including 909 lesions, were lesion-based, with a sensitivity of 0.84 (95% CI, 0.67–1.00) and a specificity of 1.00 (95% CI, 0.98–1.00). The 7 studies in which the mean age of the patients was ≥50 years had a sensitivity of 0.88 (95% CI, 0.78–0.99) and specificity of 0.98 (95% CI, 0.96–1.00), while the 4 studies in which the mean age of the patients was <50 years had a sensitivity of 0.80 (95% CI, 0.60–0.99) and specificity of 1.00 (95% CI, 0.99–1.00). Patient-based analysis, mean patient age, slice thickness and imaging system supplier were not found to be responsible for the between-study heterogeneity (P>0.05). However, study design, sample size, attenuation correction value and the different imaging system supplier were identified as sources of heterogeneity in the diagnostic performance of 18F-FDG PET/CT in breast cancer BM (P<0.05). Due to the small number of studies on 18F-NaF PET/CT, it was not possible to perform a further subgroup analysis to identify the causes of heterogeneity.

Table III. Meta-regression analysis results for fluorine-18 fluorodeoxyglucose positon emission tomography/computed tomography in the detection of bone metastases in patients with breast cancer.

Table III.

Meta-regression analysis results for fluorine-18 fluorodeoxyglucose positon emission tomography/computed tomography in the detection of bone metastases in patients with breast cancer.

Parameter	No. of studies	Sensitivity (95% CI)	P-value	Specificity	P-value
Basis			0.88		0.21
Patient	8	0.89 (0.80–0.99)		0.99 (0.98–1.00)
Lesion	4	0.84 (0.67–1.00)		1.00 (0.98–1.00)
Design			0.13		<0.01
Prospective	4	0.78 (0.55–1.00)		1.00 (0.99–1.00)
Retrospective	8	0.91 (0.83–0.98)		0.99 (0.97–1.00)
Mean age, years			0.73		0.38
≥50	7	0.88 (0.78–0.99)		0.98 (0.96–1.00)
<50	4	0.80 (0.60–0.99)		1.00 (0.99–1.00)
Sample no.			0.65		<0.01
>50	4	0.91 (0.81–1.00)		0.98 (0.95–1.00)
≤50	4	0.87 (0.71–1.00)		1.00 (1.00–1.00)
Vendor			0.82		0.07
GE Healthcare	4	0.92 (0.81–1.00)		1.00 (0.99–1.00)
Siemens Healthineers	6	0.89 (0.77–1.00)		1.00 (0.99–1.00)
AC			0.85		0.02
Yes	7	0.89 (0.78–1.00)		1.00 (0.98–1.00)
No	5	0.86 (0.72–1.00)		0.99 (0.98–1.00)
Slice thickness, mm			0.55		0.08
≥4	4	0.90 (0.76–1.00)		0.99 (0.95–1.00)
<4	4	0.83 (0.64–1.00)		0.99 (0.95–1.00)

[i] CI, confidence interval; AC, attenuation correction.

Discussion

In the present meta-analysis, covering 919 patients and 2,054 lesions from 14 studies, the diagnostic performance of 18F-NaF PET/CT and 18F-FDG PET/CT was compared in the detection of breast cancer BM. The results indicate that 18F-NaF PET/CT is more sensitive than 18F-FDG PET/CT for the detection of BM in patients with breast cancer (98 vs. 88%), while 18F-FDG PET/CT is more specific than 18F-NaF PET/CT for this purpose (99 vs. 91%). However, these differences are not statistically significant, suggesting that both tracers have a good diagnostic performance, with both having an AUC of 0.99 (95% CI, 0.98–1.00) when used in PET/CT imaging for the detection of BM associated with breast cancer.

18F-FDG PET/CT is a sensitive molecular imaging method that is able to diagnose BM by detecting the increased uptake of FDG in metastatic cancer cells (36). Previous meta-analyses have shown that 18F-FDG PET/CT has high diagnostic performance in the identification of lymph node metastasis, staging, evaluation of treatment efficacy and assessment of the prognosis of patients with breast cancer after chemotherapy (37–41). The present meta-analysis, which included 9 studies on 18F-FDG PET/CT with 530 patients and 909 lesions, showed that 18F-FDG PET/CT has a good diagnostic performance. As an osteophytic tracer, 18F-NaF offers the advantageous features of high and rapid bone uptake accompanied by very rapid blood clearance. This results in a high bone-to-background ratio in a short time and allows areas of altered skeletal activity to be displayed, which makes it an increasingly favored agent for use in the detection of bone lesions (42). Previous studies have shown that 18F-NaF PET/CT can accurately detect BM in malignant tumors such as non-small cell lung cancer, breast cancer and prostate cancer. In particular, it is useful for assessing the extent of BM and aiding in treatment decisions, making it a good tool for the early and accurate detection of BM (15,43). The present meta-analysis, which included 6 studies on 18F-NaF PET/CT, showed that 18F-NaF PET/CT is more sensitive but less specific than 18F-FDG PET/CT in the detection of breast cancer BM. This lower specificity may be due to benign diseases also being able to cause new bone formation and increase NaF uptake, which can create false positives (26). Moreover, it is notable that in addition to showing high accuracy in the detection of BM, 18F-FDG is also highly accurate in the identification of distant organ tissue metastasis and lymph node metastasis (38,39). The results of the present meta-analysis indicate that 18F-FDG and 18F-NaF have comparable accuracy in the detection of BM in patients with breast cancer. Therefore, it is suggested that 18F-FDG should be considered first in clinical practice, and additional 18F-NaF examinations may not be necessary. Previous studies revealed that 18F-FDG PET/CT is more useful than bone imaging for the detection of osteolytic BM, and that it more accurately detects pure bone marrow metastases, particularly fast-growing lesions (41,44), while it is not recommended for detecting blastic BM (45). For BM with low 18F-FDG intake, 18F-NaF PET/CT has been shown to be a better choice due to greater sensitivity (46). Although the current study indicates that 18F-FDG and 18F-NaF have similar diagnostic value, the choice of imaging agent may differ according to the clinical situation.

The current meta-analysis revealed heterogeneity in pooled sensitivity and specificity for the studies on both 18F-FDG PET/CT and 18F-NaF PET/CT. Subgroup analysis showed that study design, sample size and the use of attenuation correction were factors contributing to heterogeneity among the studies. Specifically, the specificity of retrospective studies was lower than that of prospective studies, possibly due to inherent bias in patient selection. In addition, studies with a sample size of <50 patients showed higher specificity, which may be due to the fact that a small sample size means that the diversity of the sample may be reduced. The specificity of studies using attenuation correction was higher than that of those without, likely due to improved image quality and clearer visualization of the lesions after attenuation correction (44). The meta-regression results indicated that lesion-based analysis was more specific than patient-based analysis, and that the specificity of studies with a mean patient age <50 years was greater than that of studies with a higher mean patient age, but these differences were not statistically significant. Slice thickness was not found to contribute to the heterogeneity between studies observed in the present meta-analysis. Meta-regression analysis of 18F-NaF PET/CT was not possible because only 6 studies met the inclusion criteria, and some data were not available.

The main limitation of the present meta-analysis is the limited number of eligible studies, particularly those on 18F-NaF PET/CT. During data extraction, it was found that two articles had inconsistencies in the reported TP, FP, FN and TN values, and their sensitivity and specificity; therefore, these studies were excluded (45,46). Additionally, heterogeneity in the assessment of diagnostic accuracy among the studies on 18F-FDG PET/CT and 18F-NaF PET/CT limits the quality of the meta-analysis. Histopathological validation was not available for BM in all patients; instead, imaging-based reference standards such as CT and MRI were used, which may increase clinical heterogeneity. However, as it is impractical and unethical to obtain histological evidence for all skeletal lesions, non-invasive imaging results that are not rigorously validated by histological examination are considered acceptable. Although the present study compared the diagnostic performance of 18F-NaF PET/CT and 18F-FDG PET/CT in the detection of breast cancer BM, limited data may affect the estimates of diagnostic efficacy. However, the diagnostic performance of these imaging techniques provides a reference for clinical practice and helps to avoid the subjective interpretation of results.

In conclusion, the current study shows that 18F-NaF PET/CT and 18F-FDG PET/CT are accurate methods for the detection of BM in patients with breast cancer, and are comparable in diagnostic accuracy. Moreover, it contributes to a more comprehensive understanding of the use of 18F-FDG and 18F-NaF as PET/CT imaging agents for the detection of BM in patients diagnosed with breast cancer, and may serve as a point of reference for patient care.

Supplementary Material

Supporting Data

Acknowledgements

Not applicable.

Funding

This study was funded by the National Natural Science Foundation of the People's Republic of China, NSFC (grant no. 82260353).

Availability of data and materials

The data generated in the present study may be requested from the corresponding author.

Authors' contributions

JC, XH and HH conceived the study and designed the structure of the manuscript. XH and JC were involved in the methodology, data validation, writing, reviewing and editing the study, and project supervision/CMS validated the data, carried out investigation, wrote, reviewed and edited the manuscript, and completed project supervision. JC, ZL, WY, DL and SL were involved in the conceptualization of the study, data visualization, supervision and project administration. XH and JC confirm the authenticity of all the raw data. All authors read and approved the final version of the manuscript.

Ethics approval and consent to participate

Not applicable.

Patient consent for publication

Not applicable.

Competing interests

The authors declare that they have no competing interests.

Glossary

Abbreviations

Abbreviations:

References