Bone is the third most common site of metastasis for malignant tumors after lung and liver, especially in osteophilic tumors, such as prostate, lung, breast, and thyroid cancer (1). It has been reported that the relative incidence of bone metastases varies between tumors, 65%–75% in prostate cancer, 65%–75% in breast cancer, 30%–40% in lung cancer, 40% in bladder cancer, 20%–25% in renal cell carcinoma, and 14%–45% in melanoma (2). Early identification of bone metastases in patients with malignant tumors is essential for staging and appropriate treatment selection. Whole-body bone scan (WBS) is one of the most commonly used methods to identify bone metastases early (3). Although WBS reveals a high sensitivity to detect bone lesions, it has specificity and spatial resolution limitations. With the addition of single-photon emission computed tomography (SPECT) and single-photon emission computed tomography/computed tomography (SPECT/CT) to WBS, the specificity limitations have been gradually improved. For instance, 62% of the lesions undetermined by WBS could be identified by SPECT (4) and 85%–92% of the lesions undetermined by SPECT could be identified by SPECT/CT (5). However, there are still difficulties in making a definitive diagnosis for some bone lesions in SPECT/CT (such as SPECT-positive/CT occult lesions, that is, focal abnormal radioactivity concentration on SPECT images in the absence of an identifiable anatomic lesion on CT images, or SPECT-positive/CT positive lesions, that is, focal abnormal radioactivity concentration on SPECT images with an identifiable anatomic lesion on CT images, but the diagnosis cannot be confirmed), especially in solitary lesions, which require further examination and are more expensive (6).

Quantitative analysis, widely used in positron emission tomography/computed tomography (PET/CT), has unique advantages such as objectivity, accuracy, and reproducibility compared with traditional qualitative analysis (7). In 2020, Zeintl et al. found that current commercially available SPECT/CT techniques can perform ^99mTc SPECT quantitative imaging with reasonable accuracy in both model (error <3.6%) and patient (error <1.1%) studies (8). Furthermore, with the improvement of SPECT/CT in terms of attenuation correction, scatter correction, and detection performance in recent years, SPECT/CT quantitative analysis is gradually being used in clinical practice. The study performed by Arvola et al. showed strong correlations between the SUV values measured for ¹⁸F-NaF PET/CT and ^99mTc HDP SPECT/CT, thus indicating the feasibility of using SPECT/CT SUVs in clinical practice (9). Yamane et al. also demonstrated that SUVmax of quantitative SPECT/CT has good repeatability and could be a reliable indicator in patient management (10).

Although the accuracy of SUVmax might be affected by equipment hardware, acquisition conditions, and reconstruction methods (11–13), many clinical studies have demonstrated the potential to serve as a promising biomarker of osteoblastic metabolism of SUVmax of the quantification bone SPECT/CT (14–17). Zhang et al.’s (14), Tabotta et al.’s (16), and Kuji et al.’s (17) studies already showed that the SUVmax of bone quantitative SPECT/CT is helpful to the differential diagnosis of benign and malignant bone lesions in prostate, breast, lung, liver, and thyroid cancer patients. Based on previous studies, our study attempts to determine the cutoff value of SUVmax to the differential diagnosis of benign and malignant bone lesions through retrospective analysis and then prospectively investigate the diagnostic value of the SUVmax in the evaluation of solitary undetermined bone lesions.

With the improvement and clinical application of quantitative SPECT/CT, it has been proved that SUVmax has good repeatability and great potential in the field of bone imaging and can contribute to better patient care and management (10), such as differential diagnosis of benign and malignant bone lesions, and monitor the progression of bone diseases and the response to treatment, especially in the differential diagnosis (20–23).

In Part I of this study, we demonstrated that SUVmax of malignant bone lesions (26.49 ± 12.63) is significantly higher than benign bone lesions (13.92 ± 7.16) and normal vertebrae (6.97 ± 1.52) after excluding the influence of age, sex, height, weight, BMI, injection dose, and waiting time, which can aid in clinical decision-making and improve patient outcomes, and is consistent with previous studies (14, 23, 24). A previous study by Gurkan et al. who used BS for different types of metastases showed no significant difference (p > 0.05) in mean ROImax (ratio of largest lesion to normal bone count on BS) for osteolytic lesions (5.33 ± 3.60), osteoblastic lesions (6.42 ± 4.22), and mixed lesions (6.32 ± 4.03) (25). Similar results were obtained in our study. However, it was in contrast to the study by Li et al. which only included lung cancer patients and showed that SUVmax was significantly higher (P < 0.05) in osteoblastic (29.16 ± 16.63) and mixed (26.62 ± 14.97) lesions than in osteolytic (15.79 ± 5.57) and CT-negative (16.51 ± 6.93) lesions (23). It is possible that the inclusion of different primary tumors contributed to the inconsistent results, and the heterogeneity of different primary tumors should be taken into account in subsequent studies.

Furthermore, an ROC curve analysis was performed to assess the diagnostic accuracy of SUVmax in differentiating benign bone lesions from malignant bone lesions. In our study, the SUVmax cutoff value for distinguishing malignant bone lesions from benign bone lesions was 18.07 and had a high AUC value of 0.84 (95%CI: 0.80–0.88), which is similar to the values registered by Rohani et al. (0.84 vs.0.874), and also with comparable sensitivity (75.00% vs. 73.8%) and specificity (81.70% vs. 85.4%), the differences in the data gathered from the two studies are represented by the cutoff value of ≥18.07 determined by our data and ≥20 from Rohani et al. (21). A previous study also showed a quite similar SUVmax cutoff, with Gherghe et al.’s study that included only breast cancer having an SUVmax threshold of 16.6 with a higher AUC value of 0.974 (95% CI 0.95–0.98) (22). Moreover, the study performed by Lin et al., which included 252 lung cancer bone metastases (mean SUVmax 23.85 ± 14.34) and 140 benign bone lesions (mean SUVmax 9.67 ± 7.47), showed a SUVmax cutoff value of 11.10 for distinguishing bone metastases from osteoarthritic lesions, with an AUC of 0.9097 (95% CI: 0.88–0.94), a sensitivity of 87.70%, and a specificity of 80.71% (23). In general, the SUVmax has preferable diagnostic efficacy in differentiating benign and malignant bone lesions. The reason for the lower SUVmax cutoff value in this study than Rohani et al. may be the lower mean SUVmax of bone metastases in patients with a variety of cancers than in patients only with prostate cancer; the SUVmax of malignant bone lesions in our study was 26.49 ± 12.63, compared with 36.64 ± 24.83 by Rohani et al. Compared with Lin et al. (23) and Gherghe et al. (22), the higher mean SUVmax of benign bone lesions in our data (9.67 ± 7.47 vs. 10.26 ± 4.67 vs.13.92 ± 7.16) was mainly attributed to the higher SUVmax cutoff value, which was due to a high representation of fractures in our study (56 lesions with a high SUVmax of 23.23 ± 14.92) in contrast to only three lesions in the study by Lin et al. and no fracture lesions by Gherghe et al. Of course, the SUVmax in our study differs from other studies, not only because of the different patients and lesions enrolled but also because of some other factors, including instruments, acquisition parameters, reconstruction algorithm, correction technology, and postinjection acquisition time (26).

Moreover, we prospectively used the SUVmax cutoff value of 18.07 as a reference criterion to validate the diagnostic efficacy in 49 solitary undetermined lesions of Part II. To the best of our knowledge, this is the first study to validate the diagnostic efficacy of SUVmax of quantitative SPECT/CT in bone lesions. Our results showed preferable diagnostic efficiency with a sensitivity of 82.35% and a specificity of 93.75%. A significantly higher specificity resulted from validation sets in Part II compared with experimental sets in Part I. Our study provides compelling evidence for using SUVmax of quantitative SPECT/CT imaging as a reliable diagnostic tool for bone lesions.

However, there are still two false-positive lesions and three false-negative lesions. The SUVmax of the two false-positive lesions was 55.74 and 77.93, respectively. There were CT occult lesion and no history of trauma at the rib, which were finally diagnosed as occult fractures according to the follow-up CT imaging. It is well known that fracture (including occult fractures, stress fractures, insufficiency fractures), which could lead to a high MDP concentration due to increased local bone metabolism and perifocal hematoma and necrosis, as well as calcification, is one of the crucial causes of false positives in SPECT/CT (27), and our study’s initial findings revealed that despite the SUVmax of fractures being smaller than that of malignant lesions, with a statistical difference, there is still considerable overlap. In clinical practice, when a solitary undetermined bone lesion with high SUVmax occurs in a vulnerable site (such as ribs), the possibility of fractures (particularly occult fractures and insufficiency fractures) must be excluded by combining a patient history of trauma and short-term imaging follow-up (around 2 weeks later) (28). Another three false-negative lesions had the SUVmax of 8.43 (rib), 14.67 (ilium), and 15.67 (ilium), respectively, and two lesions located at ilium were CT occult lesion, one lesion located at the rib with identifiable anatomical changes, which were finally diagnosed as metastases. The possible reason leading to a low SUVmax was the small volume (<1 cm³) of the false-negative lesions. The small volume of the lesion and the low regional bone-blood flow lead to low MDP concentration and low SUVmax; at the same time, due to the partial volume effect, it further leads to lower SUVmax (13), also given the fact that bone uptake of MDP reflects both bone formation and bone turnover, as the overall turnover of bone represents the collective activities of osteoblasts and osteoclasts. Therefore, the possible reason is that the change of bone metabolic activity is in different stages in early CT occult metastasis lesions. It may also be related to the final development of osteoblastic metastasis or osteolytic metastasis, which is related to the dominant position of osteoblastic cell or osteoclastic cells (29). Based on the above, it may also explain why the two CT occult lesions in the ilium presented as false negatives.

The main limitations of our study are as follows: (i) The final diagnosis of bone lesions was mainly based on the gold standard of at least 12 months follow-up, lacking histopathological confirmation. (ii) The study was retrospective, and the data were collected from medical records, which may have introduced some selection bias and limited the scope of the study, such as low proportion of patients with osteophilic tumors such as prostate, breast, and thyroid cancers in the study. (iii) Due to the limited sample size, the heterogeneity of primary tumors was not included in the study and a follow-up project is required. (iiii) Finally, the SUVmax cutoff value used in our study was based on a single-center experience and may not be applied to other institutions or populations. Further studies are needed to validate the optimal SUVmax cutoff value for differentiating malignant from benign bone lesions.

Our study suggests that the SUVmax of quantitative SPECT/CT is valuable in evaluating solitary undetermined bone lesions. Using a cutoff SUVmax value of 18.07, quantitative SPECT/CT demonstrated high sensitivity, specificity, and accuracy in differentiating malignant from benign bone lesions. Future studies should address these limitations and further investigate the diagnostic utility of quantitative SPECT/CT in more extensive, multicenter cohorts with a broader range of bone lesions.

The original contributions presented in the study are included in the article/supplementary material. Further inquiries can be directed to the corresponding author.

The studies involving humans were approved by The Ethics Committee of the National Cancer Center/National Clinical Research Center for Cancer/Cancer Hospital and Shenzhen Hospital, Chinese Academy of Medical Sciences and Peking Union Medical College (JS2023-15). The studies were conducted in accordance with the local legislation and institutional requirements. Written informed consent for participation was not required from the participants or the participants’ legal guardians/next of kin in accordance with the national legislation and institutional requirements.

FD and YL designed the project. XW and YZ analyzed data. ML, TL, and SH contributed to SPECT/CT scans. MZ and RW contributed to image processing. FD wrote the manuscript. YL revised the manuscript. All authors contributed to the article and approved the submitted version.