Spinal metastatic patients treated in the Department of Orthopedics of the Cancer Hospital of the Chinese Academy of Medical Science between February 2021 and February 2022 were retrospectively enrolled, and 47 patients were included in this analysis. All patients gave written informed consent, and the study was approved by the ethics board committee of the hospital.

The demographic and clinical information of patients were obtained from electronic medical records, including age, sex, diagnosis, and involved vertebrae. The visual analog scale (VAS) score (0 = no pain, 10 = severe pain) was recorded to evaluate pain intensity. The Spinal Instability Neoplastic Score (SINS) was calculated per published guidelines (27). Plain radiographs and CT scans of the corresponding vertebrae were obtained before and after the operation. If patients received systemic treatment (chemotherapy, targeted therapy, immunotherapy, endocrinotherapy, bone-protecting agents, and so on) or radiation before surgery was also recorded.

The enrollment criteria included the following: i) diagnosis of metastatic cancer; ii) clinical and imaging evidence (MRI or CT) of vertebral metastases in the cervical, thoracic, lumbar or sacral segments; iii) pulmonary CT before and after the operation; iv) expected survival time >3 months; and v) PVP combined with RFA.

Contraindications for the procedure included: i) clinical signs of spinal cord compression or cauda equina syndrome; ii) fractures with epidural involvement and contact with spinal cord or nerve roots; iii) the lesions close to the vital structures such as nerves, spinal cord, blood vessels; and iv) local infection at the puncture site or septicemia. Relative contraindications included vertebral body height reduced more than 75%, and transient chemotherapy−induced hematologic anomalies, including leukopenia (<2.5×10⁹/L), thrombocytopenia (<100.0×10⁹/L) and elevated international normalized ratio >1.5.

The patients were placed in the prone position with conscious sedation. A puncture trocar (Zhongshan Shiyitang Medical Equipment Co., Ltd, China) was inserted from the vertebral pedicle to the anterior one-third of the vertebral body under the guidance of C-arm fluoroscopy (Siemens Healthcare, Munich, Germany). The trocar was removed, and a biopsy device (STERYLAB, Italy) was placed to obtain the bone fragments for pathology. Then, a monopole RFA electrode (17G) (MedSphere Shanghai, China) was inserted through the cannula. RFA was conducted for 10 min at a temperature ranging from 80°C to 100°C. The electrode was removed, and prepared high-viscosity polymethyl methacrylate (PMMA) bone cement (Weigao Medical GmbH, China) was injected into the vertebral body under intermittent fluoroscopic examination from the lateral plane. Injection was stopped when substantial resistance was met, or when the PMMA cement reached the posterior margin of the vertebral body, or cement extravasation was identified through fluoroscopy ( Supplementary Figure 1 ). For multi-level cases, single cement kit was used for each level. 40mg methylprednisolone was given before radiofrequency ablation.

The patients were sent back to the ward after the operation and made to lie in bed for 6 hours, then could move freely, but a spinal brace was advised. Vital signs as well as sensory motor functions of the lower limbs were closely monitored. Plain radiographs and CT scans were performed before discharge.

The patients were followed up every 3 months for at least 9 months. Pulmonary CT scans were not necessary unless the patient complained of pulmonary-related symptoms. However, patients with spinal metastasis are always hospitalized several times for radiotherapy, chemotherapy, or other treatments, and pulmonary CT scans are essential for hospitalization. PCE was confirmed by comparing pre- and postoperative pulmonary images.

Statistical analysis was performed with Prism 8 software (GraphPad Software, San Diego, CA, USA). All measurement data are described as the mean and standard deviation (SD). Demographic data in both groups were compared using the chi-square test for qualitative data and the unpaired t test for quantitative data. Multiple logistic regression analysis was carried out to identify risk factors that were significantly related to PCE resulting from cement leakage. A p value < 0.05 was considered statistically significant.

PVP and RFA were performed for 47 patients with a total of 84 segments, including 34 thoracic vertebrae and 50 lumbar vertebrae ( Table 1 ). The patients were 19 males and 28 females with a mean age of 59.9 ± 1.6 years (range 34 to 81 years). The preoperative VAS score was 5.1 ± 0.3, and the preoperative SINS score was 9.6 ± 0.4. Among the 47 patients, 42 patients showed osteolytic lesions, and 5 patients exhibited osteoblastic lesions. Operations were performed on one segment in 27 patients, two segments in 11 patients, three segments in 7 patients, four segments in 1 patient, five segments in 2 patients, and six segments in 1 patient. 28 patients received systemic treatment before surgery, and 14 vertebral levels in 8 patients got pre-operative radiation treatment. Among the 84 vertebrae, pathologic compression fracture was found in 35 segments, and the other 49 segments had no compression fracture with obvious imaging abnormalities. The amount of cement injected per lesion ranged from 1.5 to 12 ml with a mean volume of 6.1 ± 0.2 ml. The postoperative pathological diagnoses of the spinal tumors confirmed that they were all metastatic tumors: 12 from lung cancer, 12 from breast cancer, 5 from kidney cancer, 4 from prostate cancer, 2 from bile duct cancer, 2 from thyroid cancer, 2 from gastric cancer, 1 from liver cancer, 1 from ovarian cancer, 1 from cervical cancer, 1 from esophageal cancer, 1 from soft tissue Ewing’s sarcoma and 3 from unknown malignancies ( Table 2 ).

According to the detection of postoperative pulmonary CT scans, the patients were divided into pulmonary cement embolism group (PCE group) and non-pulmonary cement embolism group (NPCE group). The presence of cement in the pulmonary arteries was identified in 11 patients (incidence rate 23.4%), including 5 males and 6 females, with a mean age of 55.7 ± 3.9 years. There were 36 patients in the NPCE group, including 14 males and 22 females, with a mean age of 60.5 ± 1.6 years ( Table 3 ).

All the patients in the PCE group were asymptomatic. No dyspnea, chest pain, cough, tachycardia, or hypoxia was observed. The patients were followed up regularly. There were no critical patients who needed surgical treatment or died. In all 11 patients, pulmonary emboli were located in the subsegmental or peripheral arteries, and no central pulmonary artery embolism was found ( Figure 1 ).

No significant differences were found in age, sex, preoperative VAS scores, SINS scores, or if got systemic treatment between the patients in the NPCE group and PCE group ( Table 3 ). Among the 11 PCE patients, 9 patients showed osteolytic destruction, and the other 2 patients showed osteoblastic lesions, which was not significantly different from the NPCE group. However, 4 patients (36.4%) had multiple segments (≥3) involved in the PCE group, while only 8.3% of patients (3 patients) had multiple segment involvement in the NPCE group, which was a significant difference (p=0.022).

We then set the individual vertebrae as the study object. Lesion location, cortex integrity, compression fracture, preoperative radiation, vein leakage, and PMMA injection volume in vertebrae of the two groups were analyzed ( Table 4 ). The 11 PCE patients had 25 segments involved in total, including 17 thoracic vertebrae and 8 lumbar vertebrae. The NPCE patients had 59 segments in total, including 17 thoracic vertebrae and 42 lumbar vertebrae. The results revealed that patients with thoracic vertebrae treated had a greater chance of developing pulmonary cement embolism (p=0.0008). Patients who had cement vein leakage during surgery also had a significantly increased risk compared with those who did not (p=0.017), especially when leakage occurred in the thoracic vertebrae (p<0.001). The two groups did not show significant differences in cortex integrity, compression fracture, of if got preoperative radiation.

The incidence of complications was closely related to the operators’ skills. The parameters related to technical characteristics including operative approach, PMMA volume per level, and operating time were analyzed. Among the 25 segments of PCE group, 80% were treated through unipedicular approach, while the proportion is 47.5% in NPCE group (p=0.006) ( Table 5 ). As the volume of thoracic vertebrae was smaller than lumbar vertebrae, we separated volume on injection reporting between thoracic and lumbar vertebrae. However there were no significant difference between NPCE and PCE groups. There was no difference as to operating time of two groups either.

Previous data showed that there was a high incidence of PCE if bone cement leaked into the paravertebral venous plexus. We next analyzed the risk factors related to vein leakage of cement. Among the 47 patients with 84 vertebrae, 37 vertebrae had vein leakage of bone cement (leakage group, 44%), and the other 47 segments did not have leakage (nonleakage group, 56%). Vertebra integrity, compression fracture, lesion type, location, and PMMA volume were analyzed ( Table 6 ). The results showed no significant differences in lesion type, location, or PMMA volume between the two groups. However, 75.7% of segments (28) in the leakage group showed a disrupted cortex, and only 46.8% of segments (22) in the nonleakage group showed a significant difference (p=0.008).