Image-guided Percutaneous Bone Biopsy with a Simulated Van Sonnenberg Removable Hub System

Objective: To retrospectively examine the success and complication rates associated with image-guided percutaneous bone biopsy with a simulated Van Sonnenberg removable hub system. Materials and Methods: During a 3.5-year period, 27 bone lesions at different anatomic locations with an indication for biopsy based on plain film, computed tomography (CT) and/or magnetic resonance imaging (MRI) findings were determined, and a total of 28 image-guided (fluoroscopy or CT) percutaneous biopsies were performed using a simulated Van Sonnenberg-removable hub system. This technique entailed the use of a cutout Chiba needle hub that performed as a guide for the insertion of a larger needle. Either core and aspiration biopsy or core biopsy alone was utilized. Results: The procedure yielded diagnostic material 89% of the cases (48% infection, 22% benign lesions, and 19% malignant lesions). Combined use of core and aspiration biopsy resulted in a higher diagnostic accuracy as compared to core biopsy alone. No false positive or false negative diagnoses were observed. No serious complications such as neurological deficits, bleeding, or organ injury were observed. Conclusion: The simulated Van Sonnenberg removable hub system provides a useful technique for percutaneous bone biopsies and is particulary suitable for deep seated (such as vertebral) lesions with its ability to facilitate the accessibility of the lesion with its built-in guidance needle. The procedure is safe in light of the literature data. Sonuç: " Van Sonnenberg removable hub " sistemi simülasyonu, sağ-ladığı klavuz iğne yöntemi sayesinde özellikle vertebra gibi derin yer-leşimli kemik lezyonlarına ulaşımı kolaylaştıran uygun ve kullanışlı bir yöntem olup literatür bilgileri ışığında perkütan kemik biyopsilerinde bu tekniğin kullanımı güvenilirdir.


Introduction
Image-guided percutaneous biopsy represents an important component of the management of bone and soft tissue lesions.Imaging studies may fail to differenti-ate between infectious and tumoral processes, as well as between benign and malignant lesions, and appropriate management strategies can only be established with proper histopathological examination.Providing similar diagnostic accuracy with surgical biopsies, percutaneous biopsies have now become a routine radiological procedure with lower complication rates [1,2].
Potential complications of percutaneous bone biopsy are directly related to the anatomic location of the lesion and needle type used.In addition to general complications such as pain, bleeding, or infection, specific complications related to the characteristics of the biopsy site may also occur including pneumothorax, neurological injury, and pathologic fractures.
Van Sonnenberg removable hub system is a modified coaxial biopsy needle system particularly suitable for smallsized intra-thoracic or intra-abdominal lesions that aims to increase diagnostic accuracy and safety [3,4].With this system one can perform as many fine needle aspiration biopsies as required after advancing the 19-gauge cannula into or alongside a thoracic or abdominal lesion, and it also allows fine needle aspiration biopsies or core biopsies with 20-or 21-gauge tru-cut needles.The needle system consists of a 19-gauge outer cannula, a 22-gauge inner cannula that contains the "removable hub", and the 23-gauge stylet within the inner cannula [5].Previous studies have suggested that the use of this modified coaxial needle system under computed tomography CT-guidance for percutaneous biopsies of small, hard-to-access mediastinal, hilar or pulmonary lesions results in increased safety of the procedure [6].Similarly, in another study, tru-cut biopsies performed with this system were found to significantly increase diagnostic yield in lung lesions (up to 100% in malignant lesions) with significantly reduced complication rate.Subsequent to the first report of the Van Sonnenberg removable system, a similar technique described the use of a cut-out Chiba needle hub that performed as a guide for the insertion of a larger needle [6].To the best of our knowledge, there is no report on the use of the original or simulated Van Sonnenberg removable hub system for bone biopsies.We aimed to investigate the success and complication rates associated with image-guided percutaneous bone biopsy with a simulated Van Sonnenberg removable hub system.

Patient Population
During a 3.5-year period, we performed 28 image-guided biopsies for 27 lesions in a total of 27 patients referred to our unit from other centers.Of these subjects, 16 were male and 11 were female, with an age range between 38 and 79 years (mean age, 60.4 years).

Patient Selection
We obtained institutional review board approval before beginning the study.We carefully examined patients referred to our clinic together with plain film, CT and/or MRI findings.Before the biopsy procedure, we obtained a history of bleeding diathesis from each patient, and evaluated coagulation parameters including platelet count, partial thromboplastin time (aPTT) and prothrombin time (PTZ).After explaining the benefits and risks of the procedure, we performed biopsy in those subjects who provided a written consent.

Exclusion Criteria
Exclusion criteria included the following: a history of bleeding diathesis or thrombocytopenia (<50.000mm 3 ), absence of percutaneous access to the lesion, inability to tolerate the procedure due to general health status, requirement for general anesthesia, and absence of written informed consent.

Lesion Characteristics
Based on CT image characteristics, we classified the lesions in four subgroups as follows: lytic (with a density lower than that of the normal medullary bone); sclerotic (density higher than that of the normal medullary bone); mixed; and normal (pathological MRI signal change only in the bone marrow).Table 1 shows the distribution of the lesions in terms of their anatomic location and CT characteristics.Accordingly, the most frequent site of disease involvement was the lumbar vertebrae (12/27, 44%) and most of the lesions were lytic (20/27, 74%).Also, 11 of the lesions had accompanying soft tissue involvement (11/27, 40%).

Patient Preparation
After performing comprehensive radiological assessments, we determined the appropriate imaging modality and access window.We based the latter decision on the adjacent anatomic structures, lesion type and extent.We opened an intravenous access route to premedicate the patient with pheniramine maleate (Avil, Sandoz medical, Kocaeli, Turkey) 2 mL, metochlopramide (Metpamid, Recordati medical, Istanbul, Turkey) 2 mL, and ranitidine hydrochloride (Ulcuran, Abfar medical, İstanbul, Turkey) after routine control of the blood pressure and heart rate.Also, we administered a narcotic analgesic, pehthidine hydrochloride (Aldolan, Liba medical, Istanbul, Turkey) intramuscularly (1 mg/kg) to sub-jects with intact cortical bone.We monitored the patients in our unit for 24 hours for the occurrence of possible early complications.

Biopsy Procedure
We performed fluoroscopy-or CT-guided biopsy in 9 and 19 procedures, respectively.We determined the access window such that it would lead to minimal soft tissue distance to the lesion and minimal injury.Depending on the access window we used, we maintained the patients in prone or supine position during the procedure.For vertebral lesions, we preferred posterolateral, transpedincular, and transcostovertebral approaches.Table 2 shows lesion levels and preferred access routes.Transpedincular approach was most frequently used (13/18, 72%) (Figure 2).Posterolateral approach was used in only cervical and lumbar vertebrae in a limited number of cases (4/18, 22%) (Figure 3).Transcostovertebral route was used only for one patient, who had a lesion in the thoracic vertebra (Figure 4).
When using fluoroscopy, we confirmed the direction of the needle in two different projections.In CT-guided procedures, we determined the site of the lesion using anteroposterior or lateral scanograms.Then, we obtained a set of 3-5 mm crosssectional images at the level of the lesion in order to determine the most appropriate entry point for the safest possible access route.We marked the skin surface using CT lights and a metallic marker.We calculated the distance between the entry point and the lesion as well as the access angle.
We cleansed the entry point using povidone iodide (Batidex, Tosel medical, Ankara, Turkey) and covered the procedure area with sterile drapes.We gave prilocaine (Citanest, Astra Zeneca medical, Istanbul, Turkey) with a 25-gauge needle for the local anesthesia of the skin and subcutaneous tissues.We made a small skin incision for the entry point.We advanced a 22-gauge Chiba needle with a 23-gauge stylet up to the periosteum (Figure 5a).After the 23-gauge stylet was removed, we administered local anesthetic to the periosteum and its periphery at the biopsy site.We cut off the hub of the 22-gauge cannula with scissors in order to let it function as a guide thereafter.Using this guide, we advanced the 11-gauge bone marrow biopsy needle until reaching the cortical bone (Figure 5b), and then checked its position with axial crosssectional CT images passing through the level of the tip of the needle.If required, we altered the direction of the biopsy needle and removed the guide needle.Upon reaching the cortical bone, we created holes to fix the needle in the cortex by means of clockwise and counterclockwise maneuvers.At this point, we cautiously advanced the needle with gentle hammer strokes (Figure 5c) to avoid possible complications, particularly in areas with higher bone density.During the procedure, we used intermittent axial cross-sectional CT images to check the position of the needle.We removed the inner cannula of the 11-gauge needle when we reached the lesion.Thereafter, we sent a 16-gauge semi-automatic biopsy needle or a 22-gauge Chiba needle through the 11-gauge   cannula as necessary to obtain adequate sample material for core and aspiration biopsy.In lesions with adequate bone density, we advanced bone marrow biopsy needle with rotational movements until reaching the posterior margin of the lesion.When we achieved adequate depth of the needle tip, we separated it from the surrounding bony tissue by back and forth movements.We placed osseous and/or soft tissue materials obtained under core biopsy with tru-cut needle or bone marrow biopsy needle in formalin solution and sent to the pathology department.On the other hand, we promptly sent aspiration biopsy samples to the pathology department following smear preparation and air drying.If there was a suspicion of infection based on clinical or radiological findings, microbiological tests on appropriate samples were also performed.Following the procedure, we monitored patients in our unit for 24 hours.We obtained a chest x-ray to rule out pneumothorax following thoracic biopsies.We determined the success of the procedure as the provision of sufficient material that allows the pathologist to reach a diagnosis.

Details of Biopsies
We individually determined the type of the biopsy needle and imaging modality depending on the lesion characteristics.Accordingly, we performed core biopsy alone in 22 of the 28 procedures (79%), and a combination of core and fine needle aspiration biopsies in 6 (21%).For aspiration biopsies, we used a 22-gauge Chiba needle with 10 mL injectors.For core biopsies, we used an 11-gauge bone biopsy needle and a 16-gauge semi-automatic tru-cut needle.
Of the 22 core biopsies, we performed 17 (17/28, 61%) using bone marrow biopsy needle, and the remaining five (18%) using bone marrow biopsy needle and tru-cut needle in combination.In the other six procedures, we performed core and aspiration biopsies together.We used bone marrow biopsy needle and Chiba needle for four of these latter procedures (4/28, 14%), and bone marrow biopsy needle, tru-cut needle, and Chiba needle together for two (2/28, 7%).The radiologist determined the number of tissue samples to be obtained in each procedure.We made a special effort to obtain larger number of samples when a more challenging histopathological examination was anticipated on the basis of lesion characteristics and/or inadequate biopsy material.The complete duration of the biopsy procedure, including

Results of Pathological Examination
In 24 of the 27 lesions (89%) we could obtain adequate material that allowed cytologic, microbiologic, and/or histopathologic diagnosis to be made.In one of these 24 lesions, a diagnosis was possible only after a second biopsy was performed due to non-specific findings in the initial procedure.In three lesions (11%) a diagnosis could not be achieved due to inadequate material.There were no false positive or negative results.
Table 3 shows the number, location, and histopathology of the lesions.Infection was the most frequent diag-nosis (13/27, 48%), with 6 of these (6/27, 22%) exhibiting a chronic granulomatous process.Although histopathology suggested tuberculosis in four cases, acid-fast bacteria could not be observed microscopically and grown in culture in any of these materials.All of these patients responded favorably to anti-tuberculous treatment.In one subject, Brucella tube agglutination was positive and relevant antimicrobial treatment was administered.In the remaining case with a chronic granulomatous infection, the causative microorganism proved to be S. aureus following surgery.A diagnosis of hydatid cyst was made in the iliac wing in a single patient.Other six patients (6/27, 22%) had non-specific inflammation.Of the six benign lesions (6/27, 22%), three were due to

Malignant epithelial
Rib (1) 1 tumor with no differentiation between a primary/secondary origin degenerative changes and three represented benign bone tumors.In five patients with a malignancy (5/27, 19%), lesions consisted of primary and/or secondary bone tumors.
The diagnosis was surgically confirmed in only five (5/24, 21%) of the 24 lesions with a histological diagnosis.These included hydatid cyst, spondylodiscitis (non-specific and due to S. aureus), giant cell tumor, and squamous cell lung cancer.In one case with inflammatory findings and suspicion of non-specific spondylodiscitis, the causative organism proved to be S. aureus.In one case with no clear histopathological distinction between primary or secondary malignancy, a diagnosis of squamous cell lung cancer could be made following surgery.In the remaining three cases, there was a complete concordance between histopathology and post-surgical results.
In those patients who did not undergo surgery, we used laboratory follow-up and the response to non-surgical treatment to verify the accuracy of biopsy results (Table 4).Figure 6 shows prototypical flow diagram of a diagnostic accuracy.In three patients with no definitive diagnosis, the material was inadequate for a specific diagnosis.Consequent diagnoses included one case of non-Hodgkin lymphoma and one case of multiple myeloma.In another patient with a preliminary diagnosis of metastasis, no other findings could be obtained due to patient's refusal to further diagnostic workup.In all these patients, the lesion was in the lumbar region.Two were lytic lesions, and pathologic MRI signal change was present in only one.A core biopsy under CT guidance was performed in all.Table 5 shows the results of histopathologic examination, CT findings, type of biopsy, needle types and the rate of successful diagnosis.

Complications
There were no serious complications such as bleeding, organ injury, or permanent neurologic deficit.However, in two patients temporary paraparesis that lasted for only one to two hours was observed.Biopsy procedures on these cases were performed by CT guidance and entailed vertebral lesions on both, one in a thoracic (T7) and the other a lumbar (L1) vertebra.The final diagnoses of these lesions were granulomatous inflammation and osteoporotic compression fracture, respectively.

Discussion
Our results suggest that a simulated Van Sonnenberg removable hub system provides a successful and safe alternative for image-guided percutaneous bone biopsies.The reported diagnostic yield of percutaneous biopsy of musculoskeletal lesions in the literature is 69%-90% [7][8][9][10][11][12][13][14].Our success rate (in terms of diagnostic material yield) of 89% corresponds to the high end of the reported data.
Unsuccessful biopsy may result from several factors such as failure in lesion biopsy, inadequate material, non-specific histology, necrosis or compression artifacts preventing accurate diagnosis (despite adequate material), and disagreement between biopsy results and clinical/radiological findings.Also, site of the lesion, quality of biopsy equipment, internal structure of the lesion (sclerotic, lytic), and histological type of the lesion (e.g., benign versus malignant, low-versus highgrade) may affect the success rates [10].
Technical issues such as inability to reach the lesion site or failure to enter the lesion may result in inadequate biopsy material.In this regard, generally lower rates of successful biopsy have been reported for vertebral lesions compared to lesions in extremities or pelvis [11,12].Similarly, in our study, success rate was higher for pelvic lesions (100%) than for vertebral lesions (83%).Specifically, among vertebral lesions, those in the thoracic level have been associated with lowest success rates [15].In contrast, a recent retrospective review has found remarkably high success rates for thoracic, lumbar, and sacral lesions as compared to those in the cervical region [14].In our study, lumbar lesions were associated with lower success rates in comparison with thoracic and cervical lesions.However, we believe that this was mainly due to the small sample size of our study.
The preferred method of biopsy and the biopsy equipment also play a role in the success rates.Wu et al. [13]  FNAB: fine needle aspiration biopsy; CT: computed tomography recommended obtaining higher number of samples during biopsy procedures, and they advised to obtain three samples in bone biopsies.In image-guided bone biopsies, fine aspiration biopsy or piece (core) biopsy may be used.Fine needle aspiration biopsies may help to differentiate benign lesions from metastatic ones.On the other hand, core biopsies are superior in terms of the ability to determine the cell type and tumor grade, which are required for the diagnosis of primary bone tumors [16].Reported success (in terms of diagnostic material yield) of percutaneous biopsies for bone or soft tissue lesions varies between 71% and 88% for fine needle aspiration [1,11,17] and between 70% and 90% for core biopsies [9,11,12,14,18].Some other studies suggest a higher diagnostic yield for the combined use of aspiration and core biopsies, with a complementary role suggested for these two approaches [19,20].In line with these suggestions, successful diagnosis was more common in lesions where core and fine needle aspiration were used together (100%) as compared to those where core biopsy was used alone (86%).
The internal structure and histological type are among the important factors that have an impact on the diagnostic success rate.For instance, a higher diagnostic success rate has been reported for lytic lesions than in sclerotic lesions [21,22].In our study, two of the three cases with an inconclusive biopsy result due to inadequate material were of lytic character, while the other had pathological signal change in MRI only.However, a comparison between different internal structure types in terms of diagnostic success rate is obviously not possible due to the small sample size in our study.Histologically, metastatic tumors have been reported to give a higher diagnostic yield as compared to primary bone tumors and infections [23].In a comprehensive study by Hau et al. [11] rates of successful diagnosis for primary malignant tumors was over 90%, approximately 90% for metastatic lesions, approximately 90% for benign lesions, and 50% for infections.Similarly, recent studies have reported lower rates of successful diagnosis for benign tumors, low-grade malignant tumors and infections as compared to primary malignant tumors or metastases [10,14].Non-specific histology and inability to grow bacteria in culture media in patients receiving antibiotic treatment are among the factors responsible for lower rates of diagnosis in infective lesions such as osteomyelitis and spondylodiscitis.Also in systemic malignant conditions such as Hodgkin's lymphoma, non-Hodgkin's lymphoma, and multiple myeloma histological diagnosis presents several challenges and requires immunohistochemical examination [14].
In our study, of the 12 cases with infection, six exhibited signs of chronic granulomatous inflammation, while four had findings suggestive of tuberculosis despite the absence of acid-fast bacilli and growth in culture.The remaining cases showed non-specific findings.In these cases, pathological agent could be identified on the basis of laboratory findings, response to medical treatment, and in one case by open surgical biopsy.In a single case, the causative organism could be determined only after open surgical biopsy.Of the three cases with no initial diagnosis, consequent diagnoses could be established in two (one non-Hodgkin lymphoma, one multiple myeloma).The initial failure of the percutaneous biopsy to diagnose these cases was explained on the basis of inadequate sample material and the need for further tests such as immunohistochemistry.
Complication rate of percutaneous bone biopsy is associated with the site of the lesion and type of the needle used, and its reported incidence is between 0% and 8% [1, 8, 9, 11-14, 17, 18, 21, 23-25].In addition to general complications such as infection, bleeding, and inadvertent spread of disease, specific complications including pneumothorax, neurologic deficit, and pathological fracture may occur.The occurrence of complications following percutaneous bone biopsy is usually within the first days.Recently, Huang et al. [26] reported on the delayed complications such as fever, pain, bruising/hematoma, or swelling that were encountered within 14 days following percutaneous CT-guided biopsy of bone and soft tissue lesions of the spine and extremities.None of these minor complications that are reportedly seen in up to 16% of cases (i.e., pain) altered patient management in the study by Huang et al. [26], which has the longest period for the follow-up of complications of percutaneous bone biopsy ever mentioned in the literature.Except for two cases with temporary sciatalgia, we did not observe any serious or persistent complications in our patients.However, since our patients were followed for only 24 hours after the biopsy due to our study design, we are not able to provide any data regarding later complications.
The advantage of the Van Sonnenberg removable hub system over other coaxial systems may stem from the use of thinner needles for guidance to enter the lesion, probably resulting in decreased risk of complications.Furthermore, manipulation of the guidance needle without injuring the adjacent tissues is possible while accessing small and deepseated lesions.In other coaxial systems, a direct entry into the lesion with 18-gauge or 19-gauge needle is required, which is particularly challenging for deep-seated smaller lesions.
Duration of the CT-guided percutaneous core needle biopsy procedure, including the pre-biopsy CT examination, for deep-seated musculoskeletal lesions varied between 15 and 60 minutes in a study performed by Puri et al. [12].Considering this single study in the literature where such duration information was given, the use of a Van Sonnenberg removable hub system in our study does not appear to cause a noticeable change in the duration of procedure.
Posterolateral approach is the most frequently preferred route of entry in vertebral lesions.On the other hand, the more recently described transpedincular approach can be safely used in thoracolumbar vertebral lesions, for which posterolateral approach may not be suitable.However, cortical bone is difficult to access in the peduncles.Another constraining factor is the difficulty in advancing the needle due to the thickness of the peduncles [27,28].Transpedincular approach was preferred in our osteoporotic patients, due to its safety and absence of the above-mentioned limitations.We also believe that it is safer to advance into the lesion using gentle strokes on a hammer than with plain hand maneuvers.
A major drawback of our study is its retrospective design.Also, the limited number of certain patient subgroups such as those with cervical lesions, those with sclerotic CT images, and those with histological findings suggestive of primary bone tumors is a limitation.The lack of follow-up beyond 24 hours following the biopsy procedure may also be regarded as a drawback of our study.These limit the value of our conclusions on the association between different techniques, imaging methods, and successful biopsy rates.Another limitation of our study is the lack of control groups with biopsies using other coaxial or non-coaxial techniques.Nevertheless, we did compare our study with other studies in the literature on such parameters as complication and success rates, and procedure time.
In conclusion, one of the most important advantages of the original or simulated Van Sonnenberg removable hub systems is their ability to allow combined use of core and aspiration biopsy due to their coaxial approach, increasing the diagnostic yield of the procedure, particularly for lytic lesions.We believe that a simulated Van Sonnenberg removable hub system may be chosen especially for the percutaneous biopsy of vertebral lesions, since it increases the accessibility of the deeper lesions and safety of the procedure, thanks to the presence of a guidance route.

Figure 2 .
Figure 2. a-d.A 69 year old woman who presented with back pain.T2-weighted (a), T1-weighted pre-contrast (b) and post-contrast (c) sagittal MR images show abnormal bone marrow signal intensity at T10-12 accompanied by a prevertebral soft tissue component (arrows, a).Histological diagnosis following biopsy by a transpedicular approach to the T11 mixed lytic-sclerotic lesion in prone position was plasma cell neoplasm (multiple myeloma) (d).

Figure 3 .
Figure 3. a-d.A 50-year-old male who presented with low back pain.STIR (a), T1-weighted pre-contrast (b) and post-contrast (c) sagittal MR images show complete involvement of the L4 vertebral body with loss of vertebral height, abnormal bone marrow signal and contrast enhancement.Histological diagnosis following biopsy by a posterolateral approach to the lytic L4 lesion in prone positionwas plasma cell neoplasm (plasmacytoma) (d).MR: magnetic resonance; STIR: short tau inversion recovery

Figure 4 .Figure 5
Figure 4. a-c.T7-8 granulomatous spondylodiscitis in a 45-year-old man.T2-weighted (a) and T1-weighted (b) sagittal MR images show a lesion involving most of the T7 and T8 vertebral bodies, with vertebral and intervertebral disc height loss and irregularity of the superior and inferior endplates.Axial CT fluoroscopic image (c) obtained in prone position shows the bone marrow biopsy needle inserted into the T7-8 disc and T7 inferior subchondral vertebral body via a right transcostovertebral approach.CT: computed tomography

Figure 6 .
Figure 6.Flow diagram, in prototype, of a diagnostic accuracy.