Introduction

99mTc-MDP whole-body bone scintigraphy is the most common method for diagnosing bone metastasis of malignant tumors. It mainly reflects information on bone blood flow and metabolism, and in this way detects whether there is tumor metastasis in the whole-body bone [1,2,3,4]. Compared with other imaging methods, it has good diagnostic efficiency. Local SPECT/CT imaging based on whole-body bone scintigraphy can not only confirm the osteolytic or osteogenic changes of bone metastasis, but also effectively identify bone degenerative lesions, fractures and other primary benign bone lesions. Compared with whole-body bone scintigraphy and SPECT, the accuracy of the diagnosis of tumor bone metastasis is improved [5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,20,21,22]. The combination of whole-body bone scintigraphy and local SPECT/CT is the most succinct one-stop method for screening and diagnosing tumor bone metastasis. Although SPECT/CT has been widely used in the diagnosis of tumor bone metastasis, there is still a lack of uniform diagnostic criteria for total bone imaging and local SPECT/CT examination for bone metastasis.

This retrospective study followed cases that underwent SPECT/CT with equivocal skeletal lesions detected by whole-body bone scintigraphy in the past 10 years since the introduction of SPECT/CT in the Department of Nuclear Medicine, Zhongshan Hospital, Fudan University. The final diagnosis of each skeletal lesion was based on pathological confirmation after biopsy or surgery within 4 weeks. Firstly, we obtained the data of bone metastasis and analyzed the characteristics of whole-body bone scintigraphy and SPECT/CT of bone metastases, summarized the whole-body bone scintigraphy combined with SPECT signs and CT signs, and established a grading diagnostic criterion for bone metastases by whole-body bone scintigraphy combined with SPECT/CT. Then we used this standard to independently read the whole-body bone scintigraphy images and SPECT/CT fusion images by two reviewers, with the pathological diagnosis considered as the “gold standard”, with the diagnostic efficacy of the bone metastases and the diagnostic consistency between the two reviewers analyzed to verify the criterion.

Materials and methods

Patients

From April 2008 to November 2017, 43,717 patients underwent bone scintigraphy. Of them, 24,720 patients underwent SPECT/CT with equivocal skeletal lesions in the Department of Nuclear Medicine, Zhongshan Hospital, Fudan University. This study was approved by the Institutional Review Board of Zhongshan Hospital, Fudan University. All patients provided written informed consent prior to enrollment in the study.

Case inclusion criteria and exclusion criteria

Inclusion criteria

At least one of the positive skeletal lesions found in the range of SPECT/CT imaging was diagnosed by biopsy or surgery. All of the lesions included in the study were confirmed by pathological diagnosis.

Exclusion criteria

The patient had no history of malignant tumors, furthermore the pathological diagnosis of positive lesions in SPECT/CT was not a bone metastasis; the previous pathology was clearly diagnosed as a history of primary bone tumor or tumor bone metastasis; previous imaging studies diagnosed tumor bone metastases for which therapeutic intervention was given; the lesion site has a clear surgery history of primary bone tumor or other bone lesions.

Selected cases

A total of 120 patients were included (Table 1). Of them, 97 cases had a history of malignant tumors before bone scintigraphy (24 hepatocellular carcinoma; 22 lung cancer; 12 breast cancer; 12 colorectal cancer; 6 kidney cancer; 4 thyroid cancer; 3 esophageal cancer; 3 prostate cancer; 2 bladder cancer; nasopharyngeal carcinoma, lung cancer with nasopharyngeal carcinoma, gastric cancer, scalp squamous cell carcinoma, endometrial cancer, pancreatic cancer, retroperitoneal malignant pheochromocytoma, brain malignant glioma, and nasal rhabdomyosarcoma, one each), confirmed as bone metastases, primary bone tumors or other bone lesions by bone pathology; another 23 cases had no history of malignant tumors before bone scintigraphy confirmed as tumor bone metastasis (2 hepatocellular carcinoma; 6 lung cancer; 3 prostate cancer; breast cancer, kidney cancer, thyroid cancer, cholangiocarcinoma, and thymic carcinoma one each; 7 with bone metastases with unclear primary sites).

Table 1 Types and numbers of primary tumors in 120 cases
Full size table

Bone scintigraphy

Whole-body planar scintigraphy was performed approximately 3 to 6 h after intravenous injection of approximately 1110 MBq (30 mCi) Tc-99m-MDP with a three-detector gamma camera (Prism-IRIX, Marconi Medical Systems, Cleveland, OH, USA) or two-detector gamma camera [Philips Precedence with 16-slice diagnostic CT (Philips, Medical System, Bothell, WI, USA) or Symbia Intevo with 16-slice diagnostic CT (Siemens Healthcare Ltd, Germany)] equipped with a low-energy, high-resolution, parallel-hole collimator. For planar whole-body scintigraphy acquisition, counts from the 20% energy windows at 140 keV were acquired in a 256 × 1024 matrix.

SPECT/spiral CT

Based on the findings of undetermined foci from planar whole-body scintigraphy, the patients underwent SPECT/CT scanning. Each SPECT/CT scan was acquired with the undetermined foci included in the center of the field of view. For SPECT acquisition, counts from the 15% energy window at 140 keV were acquired in a 64 × 64 matrix. A CT scan with the same field of view as the SPECT scan was acquired. The SPECT raw data were reconstructed into transaxial, coronal, and sagittal slices using reconstruction software (Astonish: Philips San Jose, CA, USA or Symbia: Siemens Healthcare Ltd, Germany). For all fused images, the accuracy of the matching of internal anatomic landmarks visible on both CT and SPECT was checked. Both SPECT and CT were performed with the patients stably lying supine and breathing shallowly. No misregistration exceeding 2 mm was found.

Analysis of characteristics of bone metastases

A total of 117 bone metastases were obtained from 100 patients with tumor bone metastasis. 60.7% (71/117) of lesions in the spine (including cervical, thoracic, and lumbar spine), 8.5% (10/117) in the thorax (including rib, clavicle, sternum, and scapula), 18.8% (22/117) in the pelvis (including hip, sacrum, and sacral iliac area), and 12.0% (14/117) in the limbs.

Whole-body bone scintigraphy and SPECT of 117 bone metastases showed 33.3% (39/117) for radioactive concentration in three or more sites except for the symmetrical part of the extremities, 1.7% (2/117) for ‘doughnut sign’ radioactive concentration, and 65.0% (76/117) for focal radioactive concentrated foci (49 cases of spinal lesions located in the pedicle and vertebral body, 2 cases of rib lesions elongated along the ribs, 25 cases located in other parts).

CT imaging of 117 bone metastases showed 6.0% (7/117) for osteogenic bone destruction, 16.2% (19/117) for mixed bone destruction, and 77.8% (91/117) for osteolytic bone destruction (including 29 cases with soft tissue masses and 24 cases with pathological fractures).

Establishment of whole-body bone scintigraphy combined with SPECT sign classification in nuclear medicine and CT sign classification

According to the characteristics of the above bone metastases and related literature [18,19,20,21,22] on the diagnosis of tumor bone metastasis by whole-body bone scintigraphy and SPECT/CT imaging, we summarized the signs of whole-body bone scintigraphy combined with SPECT in nuclear medicine and CT signs of corresponding parts.

Signs of whole-body bone scintigraphy combined with SPECT

The signs are: (1) “double track sign” and “string beads sign”, the concentration of joints in the limbs, chest, costal and sacroiliac, the concentration of borderline junction of sternum, and the radioactive sparsity of corresponding bones in radiotherapy areas; (2) focal radioactivity concentration or sparsity; (2A) spinal lesions adjacent to the edge of the vertebral body or protruding outside the outline of the vertebral body and vertebral facet joints, rib lesions being round, quasi-circular or short strip, but vertical to the long axis of the ribs; lesions of the metaphyseal of the long bone involving the articular surface; (2B) spinal lesions located within the vertebral body and pedicles, and rib lesions elongated along the ribs; (3) radioactivity concentration of three or more sites except for symmetrical sites of the limbs; (4) “doughnut sign” and “super scintigraphy”.

CT signs of the corresponding parts

The signs are: (1) degenerative changes such as hyperosteogeny, osteosclerosis, Schmorl’s node, bone island, compression of vertebrae, non-pathological fracture, increased or decreased bone density with clear focus in bone boundary, and intramedullary bone density changes unassociated with cortical bone destruction; (2) osteogenic bone destruction; (3) mixed bone destruction; 4, osteolytic bone destruction; (4A) osteolytic destruction of the spine invading intervertebral disc, or bone destruction of the metaphyseal of the long bone involving the articular surface with or without dead bone formation; (4B) associated with soft tissue masses; (4C) associated with pathological fractures.

Then we established the diagnostic criterion for bone metastases using whole-body bone scintigraphy combined with local SPECT/CT. The grading diagnostic criterion is classified into five grades (Table 2).

Table 2 The grading diagnostic criterion for bone metastases using whole-body bone scintigraphy combined with local SPECT/CT
Full size table

Image analyses

All images were independently interpreted by two experienced nuclear medicine physicians who have both CT and nuclear medicine physician qualifications and did not participate in the summary or analysis of the characteristics of the above-mentioned bone metastasis lesions. The reviewers were aware of the patients’ gender, age, any history of histologically confirmed extraskeletal malignancy and its location and type, but unaware of the results of other imaging modalities and laboratory tests. Two reviewers read the images separately according to the diagnostic criterion.

Each reviewer performed a separate reading of the whole-body bone image and SPECT/CT fusion image on the display of each workstation. According to the criterion proposed in this study, the lesions were judged as benign (Grade 1), likely benign (Grade 2), likely bone metastasis (Grade 3), highly likely bone metastasis (Grade 4), and bone metastasis (Grade 5).

Statistical analysis

The data were computed using SPSS 22.0 software (SPSS Inc., Chicago, IL, USA). With pathological diagnosis as the “gold standard”, the diagnostic efficacy of bone metastases and the diagnostic consistency between the two reviewers were analyzed to verify the feasibility of the criterion. Grade 1 and grade 2 were classified as non-tumor bone metastases, while grade 3, grade 4, and grade 5 were classified as tumor bone metastases. Diagnostic effectiveness was expressed in terms of sensitivity, specificity, positive predictive value, negative predictive value, and accuracy. The consistency between observers was evaluated by the Kappa test. The kappa values ranged from − 1 (complete disagreement) to + 1 (perfect agreement). Kappa scores were classified as follows: 0, chance agreement; < 0.20, poor agreement; 0.21–0.40, fair agreement; 0.41–0.60, moderate agreement; 0.61–0.80, substantial agreement; and 0.81–1.00, almost perfect agreement [10]. P < 0.05 was considered statistically significant.

Results

Pathological diagnosis of all lesions

A total of 141 lesions in 120 cases were pathologically diagnosed. There were 117 bone metastases in 100 patients with bone metastases, 21 benign bone lesions in 17 patients, 2 patients had single primary malignant bone tumor lesions, and 1 patient had a single primary bone tumor intermediate lesion (Table 3).

Table 3 Pathological diagnosis of 141 lesions in 100 patients
Full size table

Diagnostic efficacy of whole-body bone scintigraphy combined with SPECT/CT between two reviewers

Two reviewers analyzed the results of grading diagnosis and pathological diagnosis of the whole-body bone scintigraphy combined with SPECT/CT (Table 4). With the pathological diagnosis considered as the “gold standard”, the diagnosis of whole-body bone scintigraphy combined with SPECT/CT is benign (grade 1), possibly benign (grade 2) classified as non-tumor bone metastasis, likely bone metastasis (grade 3), highly likely bone metastasis (grade 4), and bone metastasis (grade 5) classified as tumor bone metastasis. The accuracy of the diagnosis of the two reviewers was 90.1% and 92.9%, respectively, sensitivity was both 100%, specificity was 41.7% and 58.3%, respectively, positive predictive value was 89.3% and 92.1%, respectively, and negative predictive value was both 100%.

Table 4 Comparison of diagnosis criterion of whole-body bone imaging combined with SPECT/CT and pathological diagnosis
Full size table

Consistency evaluation of graded diagnosis between two reviewers

The distribution of grading diagnosis results of 141 lesions by whole-body imaging and SPECT/CT of the two reviewers is shown in Table 5. The κ value of the two reviewers’ whole-body bone imaging combined with SPECT/CT grading diagnostic consistency test was 0.919 (P < 0.001), indicating that the consistency of the two reviewers was very high using this method.

Table 5 Distribution of diagnosis criterion results of whole-body bone imaging combined with SPECT/CT of two reviewers
Full size table

The cases of inconsistent diagnosis between the two reviewers or inconsistent with pathological diagnosis

There were 17 lesions in 14 patients inconsistent between the two reviewers or inconsistent with pathological diagnosis (Table 6). Because the signs of SPECT and CT are not specific, 12 benign skeletal lesions in ten patients with malignant tumors were misdiagnosed as metastasis, likely metastasis or highly likely metastasis by at least one reviewer (cases 1–4 and cases 6–11 in Table 6), including five non-specific inflammatory lesions (cases 1–4 in Table 6, Fig. 1), three tuberculosis lesions (case 5 and case 6 in Table 6), one humerus bone tissue hyperplasia (case 8 in Table 6), one nucleus pulposus of lumbar degeneration (case 9 in Table 6), one thoracic vertebral lymphangioma (case 10 in Table 6), and one thoracic plasma cell proliferative lesion (case 11 in Table 6). A case of lumbar Langerhans cell histiocytosis in a patient with rectal cancer which showed osteolytic bone destruction on CT image and radioactive concentration located within the vertebral body and the pedicles on SPECT image was misdiagnosed as highly probable bone metastasis by both reviewers (case 12 in Table 6, Fig. 2). A case of pubic osteosarcoma lesion in a patient with thyroid cancer which showed bone destruction on CT image and radioactive concentration located in the lesion on SPECT image was misdiagnosed as a likely bone metastasis by both reviewers (case 13 in Table 6). A case of thoracic plasmacytoma in a patient with colon cancer which showed bone destruction and vertebral compression on CT image and radioactive concentration located in the vertebral body and the pedicles on SPECT image was misdiagnosed as highly likely bone metastasis by both reviewers (case 14 in Table 6, Fig. 3). Two endplate inflammations of the lumbar spine in a patient with colon cancer were diagnosed as benign and possibly benign by both reviewers, respectively, and the results of the diagnosis were inconsistent (case 5 in Table 6).

Table 6 Cases with inconsistent diagnosis between two physicians and cases with inconsistent pathological diagnosis
Full size table
Fig. 1
figure 1

Male, 53 years old, 7 months after cerebral glioma surgery, left knee pain without obvious cause for more than 1 month. Whole-body bone plane image shows abnormal radioactive concentration at the lower left femur (a). SPECT/CT image shows left lower femur bone destruction with abnormal radioactive concentration (b). Two reviewers diagnosed it as likely benign and likely bone metastases, respectively. The patient underwent left lower femoral lesion scraping and bone cement filling. The postoperative pathological diagnosis was inflammation

Full size image
Fig. 2
figure 2

Male, 70 years old, 7 years after rectal cancer surgery, low back pain with left lower extremity pain without obvious cause for more than 1 month. Whole-body bone plane image shows abnormal radioactive concentration at the fifth lumbar spine (a). SPECT/CT image shows fifth lumbar vertebral body and left accessory osteolytic bone destruction with abnormal radioactive concentration (b). Both reviewers diagnosed it as highly likely bone metastases. The patient underwent fifth lumbar spine tumor resection and internal fixation. The postoperative pathological diagnosis was Langerhans cell histiocytosis

Full size image
Fig. 3
figure 3

Male, 57 years old, 10 years after colon cancer surgery, low back pain without obvious cause for 2 months. Whole-body bone plane image shows 11th thoracic abnormal radioactive concentration (a). SPECT/CT image shows 11th thoracic vertebral body and left accessory osteolytic bone destruction with abnormal radioactive concentration (b). Both reviewers diagnosed it as highly likely bone metastases. The patient underwent fifth lumbar vertebral tumor resection and internal fixation. The postoperative pathological diagnosis was plasmacytoma

Full size image

Discussion

On the basis of bone planar imaging, the clinical application of SPECT/CT with diagnostic CT, through the anatomical and morphological diagnostic information provided by diagnostic CT images, achieves accurate anatomical position of lesions and improves the diagnostic accuracy of tumor bone metastasis.

Bone metastases can occur in any part of the body. The typical nuclear medicine images appear as multiple, scattered, varying in size, and different forms of radioactive accumulation. The most common sites are in the central axis bone such as the spine, ribs, and pelvis. For example, multiple asymmetric radioactive abnormalities throughout the body, radioactive abnormal distribution of ribs along the direction of the ribs or the involvement of the spine in the pedicle is considered to be a more reliable sign of metastasis. CT findings of tumor bone metastases are osteolytic, osteogenic or mixed bone destruction, which may be accompanied by soft tissue mass formation, pathological fractures, etc. Metastatic intertrabecular vertebral tumors that infiltrate the marrow space without trabecular bone alteration are not visible on CT or bone scintigraphy [23]. Some benign lesions also have some characteristic signs, such as “double track sign” and “bead sign” of whole-body imaging. Radioactive concentration of the limb joints, thoracic rib joint, acromioclavicular joint and sternal angle. Radioactive sparseness of bones within the radiotherapy area; the lesion of the spine is adjacent to the edge of the vertebral body of the intervertebral disc, protrudes beyond the contour of the vertebral body or at the facet joint of the vertebra. Rib lesions are round, oval, or short strips perpendicular to the long axis of the rib. On this basis, the classification of the whole-body bone scintigraphy and SPECT signs of skeletal lesions and the CT features of the corresponding sites were summarized. The grading diagnosis of tumor bone metastasis by whole-body bone scintigraphy combined with SPECT/CT was proposed. The diagnostic efficacy was verified in the study.

The kappa scores for the degree of agreement between the two reviewers using whole-body bone imaging combined with SPECT/CT grading diagnostic consistency test was 0.919 (P < 0.001), indicating that the consistency of the two reviewers was very good. Bone whole-body imaging and SPECT/CT imaging, under the same conditions of inspection equipment, collection conditions, image processing, etc., may lead to inconsistent diagnostic results because of differences in the level of clinical diagnosis between different reviewers and differences in the analysis of characteristics of different signs. In this study, the consistency between different reviewers using the same grading diagnostic criterion was very good, which indicated that there was no significant difference between the two reviewers for the whole-body bone imaging and SPECT signs and CT image recognition for grading diagnosis. Thus, the grading diagnostic criterion can minimize the difference in diagnostic results caused by inconsistent diagnostic levels of the reviewers.

The advantage of nuclear medicine is its high sensitivity and low specificity. Radioactive concentration may occur regardless of bone metastasis, primary bone tumor, inflammation, trauma, fracture, degenerative changes, etc. CT findings of bone metastases are osteolytic, osteogenic or mixed bone destruction, which may be accompanied by soft tissue mass formation, pathological fractures, etc. However, CT images of some bone inflammatory lesions or benign bone lesions can also manifest as bone destruction and pathological fractures. For example, some lesions of vertebral tuberculosis also manifest as osteolytic bone destruction involving the vertebral body or pedicle, and it is easy to be misdiagnosed as tumor bone metastasis when intervertebral disc invasion is not obvious. In this study, five non-specific inflammatory lesions in four patients with malignant tumors and three tuberculosis lesions in two patients were misdiagnosed as metastasis, highly likely metastasis or likely metastasis by at least one reviewer. Bone proliferative lesion of a breast cancer patient showed a soft tissue density in the medullary cavity with local radioactive concentration, which was misdiagnosed as likely metastasis by a reviewer. A Schmorl's nodule of the lumbar spine in an elderly patient with rectal cancer was misdiagnosed as highly likely metastasis by both reviewers. The thoracic hemangiolymphangioma of a liver cancer patient showed pathological fractures and was misdiagnosed as highly probable metastasis by both reviewers. The pathological fracture of thoracic vertebral hemangiolymphangioma after trauma was first considered to be bone metastasis due to the history of liver cancer. A thoracic plasma cell proliferative lesion of a lung cancer patient that showed osteolytic changes was misdiagnosed as likely metastasis by both reviewers. A lumbar Langerhans cell histiocytosis of a rectal cancer patient showed osteolytic bone destruction of vertebral bodies and pedicles with radioactive concentration, which was misdiagnosed as highly likely metastasis by both reviewers. A pubic osteosarcoma of a thyroid cancer patient showed bone destruction and soft tissue mass formation with radioactive concentration was misdiagnosed as likely metastasis by both reviewers. A single plasmacytoma of the thoracic spine of a colon cancer patient showed osteolytic bone destruction of the vertebral body and pedicle with peripheral osteogenic changes accompanied by radioactive concentration, which was misdiagnosed as highly likely metastasis by both reviewers. Langerhans cell histiocytosis, osteosarcoma and plasmacytoma belong to primary malignant bone tumor. Their nuclear medicine and CT signs lack specificity and can be expressed in varying degrees of bone destruction and radioactive concentration. Especially when patients have a history of malignant tumors, they are easily misdiagnosed as having bone metastasis. In addition, two endplate inflammations of the lumbar spine of a colon cancer patient were inconsistently diagnosed as benign and likely benign by the two reviewers.

When the manifestations of nuclear medicine and the CT morphological features of some benign lesions or primary bone tumors are atypical, especially in the case of bone destruction, it is still very difficult to identify bone metastases in patients with malignant tumors, and to a large extent it also depends on the clinical experience of the reviewers and the amount of medical history information and other related imaging data.

The 141 lesions in 120 patients included in this study were based on pathological diagnosis as the gold standard, making the data of this study more reliable. However, the current study has a few limitations. Many metastatic bone tumors and most benign bone lesions diagnosed by imaging are not recommended for surgical resection or biopsy, resulting in a limited number of cases included in the sample study. Moreover, due to the very small number of benign cases, most benign lesions obtained by pathological diagnosis are not well diagnosed by imaging or subjected to surgery until the bone lesion affects important functions or patient quality of life, introducing bias into the statistical data. Furthermore, these benign lesions did not reflect the proper population of benign lesions as encountered in clinics.

Conclusion

This study investigated the performance of SPECT/CT combined with whole-body bone scintigraphy and the diagnostic criterion for bone metastases to screen skeletal lesions in malignant tumor patients with high diagnostic accuracy and good consistency among different reviewers, but the specificity was still low even with SPECT/CT.