A retrospective study on diagnostic yield,tumor types,and complications of CT-guided bone biopsies

Abstract

Background

CT-guided percutaneous bone biopsy is a minimally invasive and effective procedure for evaluating and diagnosing bone lesions.

Purpose

To evaluate the diagnostic yield, tumor types, and complication rates of CT-guided percutaneous bone biopsy procedures at a single tertiary institution.

Materials and Methods

This retrospective study analyzed 508 biopsy procedures performed on 473 patients between March 2019 and March 2024. Patients were identified through the RIS/PACS system at Aarhus University Hospital. Data on diagnostic yield, tumor types, and complications were retrospectively reviewed and collected from electronic health and histopathology records. Complications were classified according to the Society of Interventional Radiology (SIR) and the Cardiovascular and Interventional Radiological Society of Europe (CIRSE) classification systems.

Results

The most common findings were metastatic lesions (n = 234), with breast cancer being the most frequent primary tumor. The procedures resulted in an overall diagnostic yield of 88.7% and a total complication rate of 5.5%.

Conclusion

This study demonstrates that CT-guided percutaneous bone biopsy is a reliable method with a high diagnostic yield and a low complication rate. These findings support its continued role as a key diagnostic tool in the clinical management of bone lesions.

Keywords

bone biopsy CT-guided complications

Introduction

Bone lesions encompass a wide range of conditions, including both benign and malignant types. Benign lesions can be infectious, inflammatory, or neoplastic, while malignant lesions include primary bone tumors and metastases.¹ Notably, bone is the most common site for metastatic cancer involvement,² making bone metastases the most frequent type of malignant bone tumor.³

Although the exact incidence of bone metastases remains unknown, autopsy studies indicate that up to 75% of patients with advanced breast and prostate cancers—the malignancies most prone to bone metastasis—show evidence of bone involvement post-mortem.⁴ Bone metastases are often localized to the axial skeleton, including the skull, sternum, ribs, vertebral column, pelvis, and the proximal humerus and femur.^2,3

Diagnostic imaging and biopsy are essential for evaluating and diagnosing bone lesions. Clinical indications for bone biopsy include differentiating benign from malignant lesions, assessing suspicious metastatic lesions, guiding treatment decisions, or monitoring therapeutic response.^5,6

In the last two decades, advancements in imaging guidance modalities have led to a significant increase in the use of needle guidance, and the computed tomography (CT)-guided percutaneous approach has become the initial procedure of choice for most cases, although open biopsy is still considered the reference standard.^7,8 Open bone biopsy requires an incision, an operative suite, and frequently general anesthesia,⁹ whereas percutaneous CT-guided bone biopsies are minimally invasive, can be performed under local anesthesia, and have a low complication rate.^7,8 Furthermore, a large proportion of bone lesions identified on magnetic resonance imaging (MRI) and CT are confined within the bone (purely intraosseous) or are deep-seated.¹⁰ For such lesions, open biopsy can be technically challenging, requiring a large access route through healthy cortical bone, which increases the risk of post-biopsy fracture and morbidity.¹⁰ CT-guided bone biopsy is therefore often preferred, as it is able to precisely localize the lesion, yielding a high level of accuracy, and obviating the need for the more risky and invasive open surgical biopsy in most patients.¹⁰ Recent studies report a 68–96% diagnostic yield and a complication rate below 5%.^7,8

While the overall effectiveness of CT-guided percutaneous bone biopsies is well documented, data specific to individual institutions are valuable for local quality assurance and contribute to the broader understanding of procedural variation.

This study aimed to assess the data regarding diagnostic yield, tumor types, and complications associated with CT-guided percutaneous bone biopsies conducted at a single institution.

Materials and methods

Patient inclusion and data collection

This retrospective, observational study adhered to the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) statement.¹¹ It was approved by the National Data Protection Agency and the Institutional Review Board (Journal Number: 1-49-72-4-18). The study adhered to the principles outlined in the Helsinki Declaration.

The study included all CT-guided bone biopsies (n = 508) performed at Aarhus University Hospital (AUH) in Aarhus, Denmark, on 473 patients over the 5-year period between March 2019 and March 2024. Patients and biopsies were identified by examining all bone biopsies conducted in the radiology department via the RIS/PACS system.

Data on diagnostic yield, tumor type, and complications following the biopsies were retrospectively reviewed and collected from the patients’ electronic health and histopathology records. While there were no formal exclusion criteria, some cases were omitted because the procedure was canceled due to the patient’s decline in health, death, or missing the appointment before the biopsy.

Outcome measures

Diagnostic yield was calculated by dividing the number of biopsies with a conclusive histopathological diagnosis by the total number of biopsies performed. This metric was chosen over “diagnostic accuracy” because the retrospective study design did not provide a systematic reference standard (e.g., surgical follow-up) for all patients, which is required to calculate sensitivity and specificity.

Diagnostic yield = \frac{N u m b e r o f c o n c l u s i v e d i a g n o s i s}{T o t a l n u m b e r o f b i o p s i e s p e r f o r m e d}

Biopsies were deemed inconclusive if they lacked sufficient material for a histopathological diagnosis.

Biopsy procedure

Biopsy indications were determined through individual evaluations and, when applicable, a multidisciplinary team consensus. Biopsies were performed under local anesthesia using 10–20 mL of Lidocaine (10 mg/mL) by trained radiologists with CT guidance. The procedures utilized either the Madison™ Bone Biopsy System or the Bonopty® 14G system. The radiologist conducting the procedure selected the biopsy device, decided on the number of samples (at least two), and determined the approach to access the lesion. The biopsy samples were promptly sent for histopathological examination. Patients were monitored for 30 minutes after the procedure and could be discharged if no complications arose. Patients were advised to contact their general practitioner or the emergency doctor service if late complications occurred.

Classification of complications

Complications were documented during the procedure and up to 30 days after the biopsy and were defined as any symptoms or conditions potentially linked to the procedure.^{5,6,8,12–14} The Society of Interventional Radiology (SIR)’s Standards of Practice Complication Classification System¹⁵ and the Cardiovascular and Interventional Radiological Society of Europe (CIRSE)’s Classification System for Complications¹⁶ were used to classify complications. The SIR classification system categorizes complications as major or minor,¹⁵ while the CIRSE classification system grades complications from 1 to 6.¹⁶

Statistical analysis

Continuous variables were described as mean with range. Categorical variables were described with percentages and frequencies.

Results

Patient characteristics

A total of 508 CT-guided percutaneous bone biopsy procedures were performed on 473 patients. Patient characteristics and lesion localization are presented in Table 1. A total of 33 patients underwent more than one procedure; 31 patients had two procedures, and two patients had three procedures (Figure 1).

Table 1.

Patient characteristics and lesion localization.

Characteristics	Total, n (%)
Patient Demographics (N=473)
Sex
o Male	250 (52.9)
o Female	223 (47.1)
Age (years)
o Mean 65.5 (range 13–90)
Lesion Localization (N=508)
Region
o Pelvis	371 (73.0)
o Sacrum	35 (6.9)
o Sternum	25 (4.9)
o Upper extremity	25 (4.9)
o Lower extremity	20 (3.9)
o Ribs	14 (2.8)
o Scapula	14 (2.8)
o Clavicle	4 (0.8)

Figure 1.

Flowchart illustrating the exclusion of patients and data collection.

Tumor Type: Histopathology

The predominant tumor types were metastatic lesions (46.1%). This was followed by non-malignant findings (24.4%), which included histopathological descriptions such as “normal tissue,” “no malignancy,” “non-specific reactive changes,” and “inflammation.” In 10.6% (n = 54) of the biopsies, the samples were insufficient for a histopathological diagnosis (non-diagnostic), resulting in the classification “inconclusive.” The histopathological results can be found in Table 2.

Table 2.

General histopathological results.

Histological diagnosis	Total, n (%)
Benign primary neoplasms	1 (0.2)
Malignant primary neoplasms including hematopoietic neoplasms	36 (7.1)
Metastasis	234 (46.1)
Suspicious for malignancy	9 (1.8)
Carcinoma of unknown primary site	20 (3.9)
Absence of malignancy	124 (24.4)
Inconclusive	54 (10.6)
Missing histopathological description	30 (5.9)

Malignant primary neoplasms, including hematopoietic neoplasms, were observed in 7.1% (n = 36) of cases. The histopathological diagnoses are presented in Table 3.

Table 3.

Malignant primary neoplasms, including hematopoietic neoplasms.

Histological diagnosis	Total, n (%)
Myeloma/Plasmacytoma	15 (3.0)
Lymphoma	13 (2.6)
Chordoma	2 (0.4)
Ewing sarcoma	2 (0.4)
Osteosarcoma	2 (0.4)
Chondrosarcoma	1 (0.2)
Leukemia	1 (0.2)

Of the metastatic lesions (46.1%), the most common primary tumors included breast cancer (18.7%), prostate cancer (7.3%), and lung cancer (5.9%). Table 4 presents an overview of the primary tumors associated with the metastatic neoplasms.

Table 4.

Primary tumors of metastatic neoplasms.

Primary tumor of the metastatic neoplasms	Total, n (%)
Breast	95 (18.7)
Prostate	37 (7.3)
Lung	30 (5.9)
Bladder	25 (4.9)
Colon	10 (2.0)
Kidney	9 (1.8)
Sarcoma (unspecified)	6 (1.2)
Skin	4 (0.8)
Stomach	3 (0.6)
Neuroendocrine tumor	3 (0.6)
Leiomyosarcoma	3 (0.6)
Esophagus	2 (0.4)
Liposarcoma	1 (0.2)
Thyroid	1 (0.2)
Pituitary gland	1 (0.2)
Cervix	1 (0.2)
Neuronal tumor	1 (0.2)
Liver	1 (0.2)
Angiosarcoma	1 (0.2)

Diagnostic yield

Of the 508 biopsies, 30 were excluded from the pathologist’s evaluation due to their inclusion in a cancer genome research project. Consequently, 478 biopsies were examined by a pathologist. Among these, 54 biopsies were inconclusive and therefore inadequate for a histological diagnosis. In total, the CT-guided percutaneous bone biopsies provided a histological diagnosis in 424 out of the 478 examined biopsies, resulting in a diagnostic yield of 88.7% (Figure 2).

Figure 2.

Flowchart illustrating the calculation of diagnostic yield. *440 patients had one biopsy, and 33 patients had >1 biopsy (31 had two biopsies and 2 had three biopsies).

Among the 35 repeated biopsies, 13 were excluded from the “repeat” category as they were performed due to lesion progression observed on imaging or for research purposes. Therefore, 22 biopsies were classified as genuine repeat biopsies—15 of which were due to inconclusive results, and seven of which were performed because additional material was deemed necessary, even though a diagnosis had been made with the first biopsy.

Of the 54 inconclusive biopsies, 15 were repeated using the CT-guided percutaneous technique, while two were re-evaluated through a surgical (open) biopsy. Consequently, 37 inconclusive biopsies were not reassessed.

Complications

Complications related to biopsies were seen in 5.5% of cases (n = 28). This comprised 21 instances of pain at the site of the procedure, five cases of suspected infections, and two occurrences of vasovagal response during the procedure.

Out of the 21 patients who had post-procedural pain, seven did not need treatment, while 14 required opioids for pain management. All five patients with suspected biopsy-related infections required antibiotics, and four were admitted to the hospital. The two patients who experienced vasovagal responses during the procedure did not need therapy or extra observation.

According to the SIR classification, 85.7% (n = 24) were minor (grade A: n = 9; grade B: n = 15) and 14.3% (n = 4) were major (grade C: n = 4). Using the CIRSE system, 32.1% (n = 9) were grade 1, 53.6% (n = 15) were grade 2, and 14.3% (n = 4) were grade 3. No complications reached CIRSE grade 4 or 5 (permanent sequelae or death). The findings are summarized in Table 5.

Table 5.

Complication grading.

Complication type	Detail	n	SIR grade	CIRSE grade
Pain	No treatment needed	7	Minor, A (no therapy)	1 (no treatment)
Pain	Required opioids	14	Minor, B (nominal therapy)	2 (minor therapy)
Vasovagal	No treatment/observation	2	Minor, A (no therapy)	1 (no treatment)
Infection	Antibiotics (no admission)	1	Minor, B (nominal therapy)	2 (minor therapy)
Infection	Antibiotics + admission	4	Major, C (minor hosp. <48h)	3 (hospitalization)
TOTAL		28

SIR: Society of Interventional Radiology; CIRSE: Cardiovascular and Interventional Radiological Society of Europe.

Discussion

This study shows that CT-guided percutaneous bone biopsies have a high diagnostic yield and a favorable safety profile. An overall diagnostic yield of 88.7% validates the efficacy of this minimally invasive method.

In our study, the most prevalent tumor type was metastatic lesions, followed by benign conditions, consistent with recent findings in comparable studies.^13,14,17,18 Although studies by Maciel et al. and Rimondi et al. reported varying frequencies of benign lesions and primary malignant neoplasms,^19,20 they all consistently identified breast, prostate, and lung cancer as the most common primary tumors associated with metastatic lesions.^{13,14,17–20}

The prevalence of metastatic tumors relative to other types can be influenced by factors such as geographic differences in disease incidence, demographic variations, clinical symptom assessments, biopsy criteria, candidate selection, and biopsy method (core needle vs surgical). However, our study lacks specific data on these underreported factors, which potentially affect histopathological analysis. The criteria for bone biopsies are not universally standardized; instead, they depend on case-by-case evaluations or multidisciplinary team discussions, resulting in a diverse array of tumor types being sampled.

The efficacy of imaging-guided core needle biopsies has historically been debated, primarily due to concerns that smaller or lower-quality samples might compromise diagnostic yield compared to surgical biopsies.²¹ While open biopsy remains the reference standard, the lack of real-time guidance can make locating deep-seated lesions difficult, often necessitating a more invasive approach. Consequently, this typically requires general anesthesia and is associated with an increased risk of morbidity.¹⁰

Despite concerns regarding sample size, CT-guided percutaneous bone biopsies have demonstrated high diagnostic performance that rivals surgical intervention, with reported accuracies typically ranging from 68–96%.^5,8 While some studies, such as those by Rimondi et al. and Harris et al. reported diagnostic accuracies of approximately 78%,^14,20 others have found significantly higher rates. Monfardini et al. and Rehm et al. reported accuracies closer to 94%,^17,18 while Baffour et al. and Maciel et al. achieved accuracies of 96.8% and 98.4%, respectively.^13,19 These findings support the use of CT-guided percutaneous bone biopsies as a robust and minimally invasive alternative to the more complex reference standard.

Our study found a diagnostic yield of 88.7%, and among the 54 inconclusive biopsies, 15 underwent a new biopsy, and 14 produced a diagnosis. This aligns with previous findings, as Rimondi et al. reported a “77.3% accuracy rate” for primary CT-guided biopsies and 83.3% for repeated biopsies, with a “final diagnostic accuracy of 94%.”²⁰ This suggests that our diagnostic yield could have been higher if all inconclusive biopsies had been repeated, as repeated procedures tend to refine the technique. In our study, the primary reasons for not conducting repeat biopsies were advanced cancer, a low expected diagnostic yield as assessed by the pathologist, and clinical judgment—all factors that could have affected the diagnostic yield. Rimondi et al. did not analyze whether the difference between primary and repeat biopsy accuracy was statistically significant,²⁰ indicating an area for future research.

Our complication rate of 5.5% aligns with the rates reported in related studies. Rimondi et al. reported a complication rate of 1.1%,²⁰ Maciel et al. and Harris et al. 1.6%,^14,19 Rehm et al. 2.8%,¹⁸ Monfardini et al. 3.5%,¹⁷ and Baffour et al. 5.6%.¹³ However, the definition of a complication varies across studies. For instance, not all studies consider procedural site pain a complication. In our study, as in those by Monfardini et al. and Baffour et al., complications were classified according to the SIR complication classification system,¹⁵ which may explain the relatively higher complication rates in these studies. Several studies did not use a formal classification system^14,18–20; only Rehm et al. considered pain a complication.¹⁸ Pain significantly contributed to the complication rates in both the literature and our study, influencing overall rates. Therefore, implementing a standardized classification system, such as SIR or CIRSE, is essential for accurately documenting complications. Excluding pain could lead to an underestimation of the actual complication rate.

In our study, outpatients were discharged 30 minutes post-biopsy if no signs of complications were present. Before leaving the hospital, patients were instructed to contact their general practitioner or emergency services if late-onset complications occurred. Several studies did not provide exact information on how complications were detected.^14,19,20 Rhem et al. observed patients for 6 hours post-biopsy.¹⁸ Monfardini et al. performed an immediate post-biopsy CT scan to detect complications, followed by 2 hours of bedrest before discharge.¹⁷ Baffour et al. observed patients for at least 1 hour post-biopsy, with a follow-up phone call 24–72 hours later to detect possible complications.¹³ With ongoing follow-up, this study noted increased complication rates, all linked to pain at the procedural site.¹³ This suggests that complications, especially minor ones, may be underreported without rigorous follow-up. Study population characteristics may also affect complication rates, as bone lesions that are difficult to access may carry a higher risk of complications. Similarly, biopsies performed on patients with significant comorbidities may lead to higher complication rates. Standardized criteria for selecting biopsy candidates could enhance study comparability, but differences in healthcare systems make this challenging.

This study has several strengths, including a significant sample of 473 patients and 508 biopsies. Participants were obtained from the RIS/PACS system, ensuring that all individuals who underwent a CT-guided percutaneous bone biopsy during the specified timeframe were included. Additionally, a complication classification system enhanced data comparability and reliability by minimizing the chances of underreporting complications.

Nonetheless, there were significant limitations. Primary among them is the risk of selection bias, as the decision to biopsy was driven by individual clinical evaluations. Consequently, lesions with obvious benign radiographic features may have been managed conservatively without biopsy, potentially skewing our cohort toward a higher prevalence of malignancy or diagnostically complex cases. Additionally, the retrospective design relied exclusively on information from patients’ electronic health and histopathology records.

Furthermore, we have consistently utilized the term “diagnostic yield” rather than “diagnostic accuracy” to describe our results. As the assessment of diagnostic accuracy, including sensitivity and specificity, requires a direct comparison between a target test and a reference standard, our retrospective study design and the lack of a systematic reference standard for all included patients preclude such an analysis. To maintain terminological stringency and avoid overstating the statistical findings, “diagnostic yield” was deemed the most appropriate metric to reflect the proportion of biopsies that resulted in a definitive diagnosis.

Moreover, complications were tracked only for 30 days post-biopsy, whereas other studies monitored patients for up to a year, which might influence the reported complication rate. Finally, the lack of a dedicated follow-up protocol to identify late-onset complications may have contributed to an underestimation of the actual complication rate, as patients were instructed to reach out to their general practitioners for late complications, and this information might not have been accessible for hospital review.

In conclusion, this study shows that CT-guided percutaneous bone biopsies have a high diagnostic yield and a complication rate in line with previous studies.

Footnotes

ORCID iDs

Maria Tækker

Flemming Kromann Nielsen

Author contributions

1. LH, OG, and FKN contributed as guarantors of integrity of the entire study.

2. LH, OG, and FKN contributed to study concepts and design.

3. LH and FKN contributed to literature research.

4. LH, NLO, JA, MT, and FKN contributed to clinical studies.

5. LH, NLO, JA, MT, and FKN contributed to experimental studies/data analysis and statistical analysis.

6. LH, OG and FKN contributed to manuscript preparation.

7. LH, NLO, JA, MT, OG and FKN contributed to manuscript editing.

Funding

The authors received no financial support for the research, authorship, and/or publication of this article.

Declaration of conflicting interests

The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

References

Eweje

Bao

, et al. Deep learning for classification of bone lesions on routine MRI. EBioMedicine 2021; 68: 1–9.

Coleman

Croucher

Padhani

, et al. Bone metastases. Nat Rev Dis Primers 2020; 6: 83.

Migliorini

Maffulli

Trivellas

, et al. Bone metastases: a comprehensive review of the literature. Mol Biol Rep 2020; 47: 6337–6345.

Chin

Kim

. Bone metastasis: concise overview. Fed Pract 2015; 32: 24–30.

Filippiadis

Mazioti

Kelekis

. Percutaneous, imaging-guided biopsy of bone metastases. Diagnostics 2018; 8: 1–6.

Spinnato

Colangeli

Rinaldi

, et al. Percutaneous CT-Guided bone biopsies: indications, feasibility and diagnostic yield in the different skeletal sites—from the skull to the toe. Diagnostics 2023; 13: 1–15.

Masood

Mallinson

Sheikh

, et al. Percutaneous bone biopsy. Tech Vasc Interv Radiol 2022; 25: 1–10.

Tomasian

Hillen

Jennings

. Bone biopsies: what radiologists need to know. Am J Roentgenol 2020; 215: 523–533.

Adams

Potter

Pitcher

, et al. Office-based core needle biopsy of bone and soft tissue malignancies: an accurate alternative to open biopsy with infrequent complications. Clin Orthop Relat Res 2010; 468: 2774–2780.

10.

Spinnato

Colangeli

Rinaldi

, et al. Percutaneous CT-Guided bone biopsies: indications, feasibility and diagnostic yield in the different skeletal sites-from the skull to the toe. Diagnostics 2023; 13: 2350.

11.

von Elm

Altman

Egger

, et al. The strengthening the reporting of observational studies in epidemiology (STROBE) statement: guidelines for reporting observational studies. Lancet 2007; 370: 573–577.

12.

Gupta

Wallace

Cardella

, et al. Quality improvement guidelines for percutaneous needle Biopsy. J Vasc Interv Radiol 2010; 21: 969–975.

13.

Baffour

Moynagh

Eiken

, et al. Effectiveness and safety of percutaneous CT-Guided rib Biopsy. J Vasc Interv Radiol 2019; 30: 82–86.

14.

Harris

Rajashekar

Sharma

, et al. Performance of computed tomography-guided spine biopsy for the diagnosis of malignancy and infection. Oper Neurosurg 2021; 21: 126–130.

15.

Omary

Bettmann

Cardella

, et al. Quality improvement guidelines for the reporting and archiving of interventional radiology procedures. J Vasc Interv Radiol 2003; 14: 293–295.

16.

Filippiadis

Binkert

Pellerin

, et al. Cirse quality assurance document and standards for classification of complications: the cirse classification system. Cardiovasc Interv Radiol 2017; 40: 1141–1146.

17.

Monfardini

Preda

Aurilio

, et al. CT-guided bone biopsy in cancer patients with suspected bone metastases: retrospective review of 308 procedures. Radiol Med 2014; 119: 852–860.

18.

Rehm

Veith

Akbar

, et al. CT-Guided percutaneous spine Biopsy in suspected infection or malignancy: a study of 214 patients. Röfo 2016; 188: 1156–1162.

19.

Maciel

MJS

Tyng

Barbosa

PNVP

, et al. Computed tomography-guided percutaneous biopsy of bone lesions: rate of diagnostic success and complications. Radiol Bras 2014; 47: 269–274.

20.

Rimondi

Rossi

Bartalena

, et al. Percutaneous CT-guided biopsy of the musculoskeletal system: results of 2027 cases. Eur J Radiol 2011; 77: 34–42.

21.

Vasilevska

Ivanoski

Samardziski

, et al. Percutaneous imaging-guided versus open musculoskeletal biopsy: concepts and controversies. Semin Musculoskelet Radiol 2020; 24: 667–675.