Percutaneous CT-guided sternal biopsy: factors affecting the diagnostic yield

Abstract

Background

Sternal lesions are occasionally seen in clinical practice and their diagnosis can be important, especially for oncologic patients. However, percutaneous computed tomography (CT)-guided biopsy of sternal lesions is rarely performed.

Purpose

To assess the diagnostic yield of percutaneous CT-guided sternal biopsies and to analyze the factors that affect diagnostic yield.

Material and Methods

A retrospective review of 34 patients who underwent CT-guided sternal biopsy was carried out at a single institution. Pre-biopsy CT density, location, penetration length of biopsy needle, number of biopsy attempts, angle of needle approach, final diagnosis, and operator experience level were recorded. A biopsy was considered as diagnostic if it provided a confident pathologic result. All variables were compared using Chi-square tests.

Results

Twenty-two of the 34 (64.7%) biopsy procedures yielded a diagnostic sample and 12 (35.3%) were non-diagnostic. Eight participants in the non-diagnostic group were clinically diagnosed with inflammatory arthritis of the manubriosternal or costosternal joints. Longer penetration distance of the tumor by the biopsy needle showed higher diagnostic yield (P = 0.031). Osteoblastic lesions (P < 0.001), lesions in the manubriosternal joint (P = 0.018) and approaches using more obtuse angles (P = 0.009) were associated with significantly lower diagnostic yields. Malignancy in the final diagnosis led to a higher diagnostic yield than benign lesions (P < 0.001).

Conclusion

CT-guided percutaneous sternal biopsy has a relatively lower diagnostic yield. However, acute angle of needle approach may help increase biopsy success rate. Osteoblastic lesions and lesions in the manubriosternal joint tend to have lower diagnostic yield.

Keywords

Sternum computed tomography (CT)biopsy large-core needle image-guided biopsy

Introduction

The sternum, sternoclavicular, and manubriosternal joints are crucial structures of the anterior chest wall. Histopathologic analysis of sternal lesions of unknown origin is essential before treatment because it provides important information for further management. There are multiple benefits associated with percutaneous needle biopsies, including a three- to five-fold increase in cost-effectiveness compared with open surgical biopsy procedures, minimal post-procedure activity limitation, rapid recovery time, and significantly lower complication rates (1 –3).

For percutaneous needle biopsy procedures, mainly ultrasound (US), fluoroscopy, and computed tomography (CT) are used as guidance systems. US provides benefits, especially for superficial lesions, lytic lesions, or broken cortex with an extra-osseous component (4,5). These include real-time visualization of the needle tip and easy to manipulate an automatic biopsy gun for core needle biopsy (CNB). However, US has limitations in evaluating bone structure and deep bone lesions unless periosteal change developed (6).

Fine-needle aspiration cytology (FNAC) and CNB have been used as the primary diagnostic modalities to obtain adequate sample. FNAC is being used for initial diagnoses, as well as for recurrences and metastases. However, FNAC can have sampling errors attributed to low cellularity, inadequate sampling of the lesions, which are highly sclerotic, blastic, or cartilaginous, or contain cystic, bloody, necrotic materials (7 –9). Domanski et al. (10) demonstrated that FNAC and CNB used together increase the diagnostic accuracy of tissue samples.

Rimondi et al. (11) reported that of 2027 total percutaneous CT-guided biopsy cases, only 17 were performed at the sternum. To our knowledge, there have been no descriptive studies of diagnostic yield or of factors that could improve the technical success of percutaneous CT-guided sternal biopsy.

The objectives of this study were to assess the diagnostic yield of CT-guided sternal biopsy, to analyze the diagnostic spectrum, and to analyze the factors associated with non-diagnostic results.

Material and Methods

This retrospective study was approved by our Institutional Review Board and the need for informed consent was waived. We performed a search of the pathology database for patients who underwent a sternal biopsy from January 1998 to February 2016. A total of 34 consecutive participants who underwent CT-guided sternal biopsy were identified and included in this study. Because this study was focused on CT-guided biopsy, biopsies performed via US guidance were excluded. We reviewed each patient’s electronic medical record for the following data: (i) demographic data; (ii) data related to CT-guided biopsy, including penetration length of the tumor by the biopsy needle, lesion location, number of needle penetrations, angle of needle approach, pre-procedural CT density, attending physician’s years of experience, and pathologic result; and (iii) final pathologic or clinical outcome, as indicated by results from pathologic examinations (image-guided biopsy, surgical biopsy) or by clinical confirmation with follow-up. Number of needle penetration means the number of different entry site or different angle with the same entry site we have selected. Pre-procedural CT density was based on visual inspection of the biopsied region and classified as osteoblastic or osteolytic (12). An osteoblastic lesion was defined when more than 50% of its volume was denser than the surrounding normal bone. An osteolytic lesion was defined when more than 50% of its volume was radiolucent compared to the surrounding normal bone. Each lesion was divided into two groups according to whether the penetration length of the tumor by the biopsy needle was ≥10 mm or <10 mm. Lesion locations were classified as occurring in the body, the manubriosternal joint, or the manubrium. We defined the angle of needle approach as the angle between the transverse axis of the sternum and the biopsy needle and we divided cases by whether the angle was ≥45° or <45°.

CT-guided biopsies were performed by ten musculoskeletal radiologists who had a range of 1–15 years of experience. All procedures were performed on five CT scanners. From 1998 to 2002, a one-slice spiral-CT scanner (Genesis HiSpeed RP; GE Healthcare, Milwaukee, WI, USA) was used. From 2003 to 2006, examinations were performed on a four-slice scanner (LightSpeed Qx/i; GE Healthcare). In 2007, an eight-slice scanner (LightSpeed Ultra, GE Healthcare) was used. Since November 2008, a 16-slice scanner (Aquilion; Toshiba Medical Systems, Otawara, Japan) was used, and since 2012 a 64-slice scanner (LightSpeed VCT, GE Healthcare) has been alternatively used alongside the 16-slice scanner. Immediately before biopsy, blood tests for coagulopathy (prothrombin time, activated partial thromboplastin time, international normalized ratio) and a complete blood count were performed. All patients’ vital signs were continuously monitored during the procedure with a pulse oximeter and an automated blood-pressure cuff. Scout images were obtained (slice thickness, 2–5 mm; scan length, 10–38 cm depending on the region of interest) with radio-opaque markers on the skin. After choosing the skin entry site, 1% lidocaine for local anesthesia was applied. To avoid any necrotic or cystic areas, all biopsies were preceded by pre-procedural magnetic resonance imaging (MRI) or positron emission tomography-computed tomography (PET-CT) review. When an area had low signal intensity on T1-weighted (T1W) images, high signal intensity on T2-weighted (T2W) images, showed no contrast enhancement, unilocular, and had irregular margin, it was regarded as a necrotic region (13). If the 18F labeled fluorodeoxyglucose (18F-FDG) uptake was higher than that of the mediastinum or liver, it was regarded as a viable target region (14). Following preliminary axial CT scanning, the most appropriate cross-section image for planning the ideal route for needle positioning into the lesion was selected. When the biopsy needle entered the outer cortex, close attention was required to prevent needle slippage. All biopsies were performed with a coaxial bone biopsy system (14-guage Bonopty; AprioMed, Uppsala, Sweden) (Fig. 1). For osteolytic lesions, a semi-automated biopsy gun (18-gauge Gun biopsy needle; M.I. Tech, Seoul, Republic of Korea) was used within the Bonopty needle cannula. CT scanning was repeated to confirm biopsy needle placement in the lesion. A minimum of one and a maximum of three cores were obtained for each biopsy. The core was judged to be sufficient depends on the subjective quality of the material obtained as assessed by visual inspection, and each radiologist’s personal experience and preference (15).

Fig. 1.

A 62-year-old woman presented with a sternal metastasis from breast cancer. (a) CT scan shows an osteoblastic lesion in the sternal body and enlarged left internal mammary lymph node. (b) Axial CT image shows a 15-gauge needle with the tip in the sternal lesion. A transverse oblique approach through the sternum was used.

All specimens obtained for histologic assessment were fixed with formalin and were analyzed by bone and soft-tissue sarcoma pathologists. In patients with suspected infection, a separate dry sample was sent for microbiological analysis, including an acid-fast bacilli culture for tuberculosis. The radiologic report and nursing notes were reviewed to determine whether any procedure-related complications occurred.

A biopsy was considered diagnostic if the needle biopsy contained adequate tissue and demonstrated histologic evidence of neoplastic tissue or positive tissue culture. If the biopsy specimen could not suggest a specific diagnosis, it was classified as non-diagnostic. For the non-diagnostic results, final diagnoses were confirmed by surgical biopsy, repeat percutaneous biopsy, response to medical treatment, or clinical and/or radiologic follow-up over three or more months. Diagnostic yield was defined as the number of diagnostic CT-guided biopsies divided by the total number of CT-guided biopsies.

Statistical analyses were performed using SPSS, version 22.0 (SPSS Inc., Chicago, IL, USA). Chi-square tests were used to compare the non-diagnostic biopsy rates according to lesion size, location, number of needle penetrations, angle of needle approach, pre-procedural CT density, attending physician’s years of experience, and pathologic result. Bonferroni’s correction was used to adjust the P values for multiple comparisons of the diagnostic yield only if the overall test was statistically significant. P values < 0.05 were considered statistically significant.

Results

The sample included 12 men and 22 women (mean age, 53.2 years; age range, 29–71 years). Thirty sternal lesions (88.2%, 30/34) were intra-osseous with no involvement of the associated soft tissue. The other four lesions (11.8%, 4/34) had an extra-osseous component. The diagnostic yield of CT-guided percutaneous sternal biopsies was 64.7% (22/34). There were 12 non-diagnostic biopsies, yielding an overall non-diagnostic rate of 35.3% (12/34).

In the diagnostic group, 21 of 22 (95.5%) lesions were malignant and one (4.5%) lesion was benign (Table 1). The most common malignant lesions were metastases (n = 16), followed by multiple myelomas/plasmacytomas (n = 5, Fig. 2). Thirteen of the 16 (81.3%) biopsy-proven metastatic lesions were confirmed to be single sternal metastases by a combination of MRI, CT, PET-CT, and whole-body bone scans. Three of 16 (18.7%) lesions had multiple bone metastases, but the sternal lesions were most accessible for percutaneous biopsy. One lesion had adjacent manubrium and costosternal junction involvement. Another lesion, originating from a primary thymic carcinoma, showed direct invasion of the sternum. In this case, the sternum was preferred for biopsy compared with the mediastinal mass. One benign lesion was confirmed to be an acute osteomyelitis and was detected from a small number of Gram-positive cocci.

Table 1.

Percutaneous needle biopsy histology results (n = 22).

Pathologic results	Cases (n)
Metastasis	16
Invasive ductal carcinoma of the breast	10
Clear cell renal cell carcinoma	1
Hepatocellular carcinoma	1
Duodenal adenocarcinoma	1
Papillary thyroid carcinoma	1
Lung adenocarcinoma	1
Thymic carcinoma	1
Multiple myeloma	5
Acute osteomyelitis	1

Fig. 2.

A 49-year-old woman presented with tenderness on anterior chest. (a) Whole body bone scan demonstrated increased uptake in the sternal body. (b) Axial CT image shows osteolytic lesion with endosteal scalloping in right sternum body. (c) Correlated coronal CT image shows similar imaging feature. (d) Oblique axial reconstructed CT image shows a biopsy needle within an osteolytic lesion. Pathology revealed plasma cell myeloma.

In the non-diagnostic group, eight participants were clinically diagnosed with inflammatory arthritides after CT-guided biopsy; these cases were localized in the manubriosternal joints or costosternal junction (Table 2). The patients had a history of malignancy and annual follow-up images showed newly developed osseous lesions without symptoms. Among these cases, two were diagnosed with SAPHO (synovitis, acne, pustulosis, hyperostosis, and osteitis) syndrome and corresponding biopsies were taken from the sternal body adjacent to the manubriosternal joint (Fig. 3). One patient with a history of a renal pelvis tumor had a sternal lesion that was diagnosed on whole-body bone scan with new uptake, but CT-guided biopsy revealed marrow fibrosis and did not suggest a specific diagnosis (Fig. 4). Over five years of follow-up with medication, the patient’s symptoms waxed and waned. When comparing the patient’s initial MRI with one from five years later, we observed that bony erosion and bone marrow edema had increased in multiple costosternal and costomanubrial joints, which means aggravation of costochondritis. The other five non-diagnostic cases were treated with NSAIDs or opioid analgesics for periods ranging from three months to six years.

Table 2.

Final diagnoses in patients with initial non-diagnostic needle biopsies (n = 12).

Results	Cases (n)
Metastasis		4
Breast	2
Lung	2
Inflammatory arthritides		8
SAPHO syndrome	2
Costochondritis	1
Not specified	5

SAPHO, synovitis, acne, pustulosis, hyperostosis, and osteitis.

Fig. 3.

A 60-year-old woman presented with anterior chest wall pain. (a) Axial CT image shows a biopsy needle within a sclerotic lesion in the manubriosternal joint. (b) Sagittal reformatted CT images of the thorax show sclerosis, erosions, and hyperostosis in the sternoxiphoid and manubriosternal joints and the discovertebral joint of mid-thoracic spine (arrows), which are highly suggestive of SAPHO syndrome.

Fig. 4.

A 29-year-old man with a renal pelvis tumor presented with anterior chest wall pain. The bone scan illustrates increased uptake (not shown) in the sternum. (a) Fat-suppressed T1W coronal MRI after contrast enhancement shows patchy enhancement in the periphery of the sternal body adjacent to the costosternal junction. (b) Axial CT image shows a biopsy needle within a sclerotic lesion in the costosternal junction. (c) Post-contrast coronal MRI five years after initial imaging shows significantly increased erosion and joint-space widening with bone marrow enhancement of multiple costosternal junctions. The patient was clinically diagnosed with costochondritis.

Of the four participants whose initial biopsies were non-diagnostic, one was revealed to have a metastasis after US-guided biopsy in the extra-osseous component four months later. The other three participants were revealed to have metastases after undergoing annual follow-up imaging for the following criteria: (i) previous biopsied lesion grew notwithstanding chemo-radiotherapy or without treatment; (ii) new lesion developed that looks similar imaging feature of previous biopsied lesion; and (iii) increasing metabolic activities in PET-CT.

Penetration length by the biopsy needle had a statistically significant higher diagnostic yield (82.4%, 14/17, P = 0.031). Osteoblastic lesions had a statistically significant higher non-diagnostic rate (71.4%, 10/14, P < 0.001) than osteolytic lesions (Table 3). Manubriosternal joint lesion (70.0%, 7/10, P = 0.018), an obtuse angle (>45°) of needle approach (55.6%, 10/18, P = 0.009), and having a benign lesion as the final diagnosis (88.9%, 8/9, P < 0.001) had significantly higher non-diagnostic rates. There were no statistically significant differences in the non-diagnostic rates when stratified by the attending physician’s years of experience (P = 0.320), and number of needle penetrations (P = 0.638). The diagnostic yield between the manubrium and manubriosternal joint was significantly different after a Bonferroni correction (P = 0.027) (Table 4).

Table 3.

Analyses for diagnostic yield.

Variables	Non-diagnostic	Diagnostic	P value
CT density
Blastic	10 (71.4%)	4 (28.6%)	<0.001
Lytic	2 (10%)	18 (90%)	<0.001
Location
Body	4 (26.7%)	11 (73.3%)	0.018
Manubriosternal joint	7 (70.0%)	3 (30.0%)
Manubrium	1 (11.1%)	8 (88.9%)
Experience of the operator (years)
<2	4 (50.0%)	4 (50.0%)	0.320
≥2	8 (30.8%)	18 (69.2%)	0.320
Attempts (n)
1	9 (33.3%)	18 (66.7%)	0.638
≥2	3 (42.9%)	4 (57.1%)	0.638
Final diagnoses
Benign	8 (88.9%)	1 (11.1%)	<0.001
Malignant	4 (16.0%)	21 (84.0%)	<0.001
Angle of the needle approach (°)
<45	2 (12.5%)	14 (87.5%)	0.009
≥45	10 (55.6%)	8 (44.4%)	0.009
Penetration length by the biopsy needle (mm)
<10	9 (52.9%)	8 (47.1%)	0.031
≥10	3 (17.6%)	14 (82.4%)	0.031

Table 4.

Multiple comparisons with location for diagnostic yield.

P value
Variables	Raw	Bonferroni corrected
Body vs. manubriosternal joint	0.032	0.096
Body vs. manubrium	0.364	1.000
Manubriosternal joint vs. manubrium	0.009	0.027

P values are derived from Chi-square tests using the Bonferroni correction.

One procedure-related complication developed, which was a small hematoma in the anterior mediastinum. The hematoma did not increase in size and had disappeared by the time a follow-up CT scan was performed the next day.

Discussion

Abnormalities in sternal lesions are commonly seen in clinical practice; however, they are a wide spectrum of congenital variants and pathologic conditions (16). A systematic approach for clinical assessment and imaging evaluation of these lesions is essential for correct diagnosis. However, for cases with diagnostic uncertainty, tissue diagnosis should be performed after open surgical or percutaneous needle biopsy (15). There are many studies comparing the diagnostic yield between FNAC and CNB. Reported diagnostic accuracy of FNAC has varied greatly, in the range of 67–99% (9,17). Adequate sampling with a thick intact cortex is usually not easy. Dense sclerotic lesions such as osteoblastoma and osteoid osteoma as well as cartilaginous tumors frequently showed insufficient specimens. Despite these limitations, the diagnostic accuracy for bone metastasis is high (18). Thus, some studies indicated good results when FNAC was performed concurrently with CNB (10,19). On the contrary, Yang et al. (8) and Handa et al. (20) suggested that CNB has a higher diagnostic accuracy than FNAC, including the nature of the tumor, histologic type, and grading.

US-guided biopsy for musculoskeletal lesions has been shown to be advantageous for osteolytic lesions with a large extra-osseous component or soft tissue mass in superficial location (4,5). It showed higher diagnostic yield in case of osteosarcoma, metastasis, and malignant round-cell tumor (4). However, using US in intra-osseous lesions is still controversial.

This study evaluated the diagnostic yield of CT-guided sternal biopsy in patients who were referred to a tertiary-care hospital showing a diagnostic yield of 64.7%. The 35.3% non-diagnostic rate was slightly higher than that reported by previous studies (21,22). Compared to other studies, our study population had a large proportion (88.2%) of lesions that lacked an extra-osseous component, which may have contributed to a lower diagnostic yield (15). Also, most previous studies have reported on percutaneous biopsies in the entire musculoskeletal system, but did not focus on sternal lesions. This study showed that biopsy targets located in the manubriosternal joint had a higher non-diagnostic biopsy rate than those in other locations. Lesions in the manubriosternal junction were associated with a high proportion of inflammatory arthritides (70%, 7/10). Jurik et al. (23) reported radiographic sclerotic changes in the sternoclavicular joint with swelling and tenderness even after long-term clinical follow-up. Histopathological findings from these lesions were non-specific, which suggests that inflammatory arthritides is difficult to confirm with percutaneous needle biopsy. In our study, real incidence of false negative results which means non-diagnostic biopsy with clinical usefulness was 11.8% (4/34) and the sum of true positive and true negative results was 88.2% (30/34). Awareness of the high non-diagnostic rate of manubriosternal joint biopsies may be helpful for predicting potential outcomes from the biopsy.

The results of our study are consistent with those from previous studies, which have shown that osteoblastic lesions and benign diagnoses have a higher non-diagnostic rate than osteolytic lesions and malignant diagnoses (12,21,24). In contrast, Omura et al. (25) reported no significant correlation between diagnostic accuracy and bone lesion matrix. However, they did find that intermediate-to-high grade neoplasms had significantly higher diagnostic yields compared with low-grade neoplasms. Nouh et al. (26) have suggested that the diagnostic yield is higher from sclerotic and mixed lesion cases compared with osteolytic lesions. Their patients underwent percutaneous biopsy at least six times, but it is difficult to apply this type of biopsy at the sternum, which is close to vital organs and is a relatively thin bone. Our study did find significantly higher non-diagnostic rates with an obtuse angle-to-needle approach, which may indicate that it is important to obtain sufficient material to improve the diagnostic yield (Fig. 5). Our results are also consistent with those of a previous study that showed no significant changes in diagnostic yield depending on the amount of attending physician experience (22), contrary to Wu et al., who found that the number of needle penetrations could influence the diagnostic yield (12). Wu et al. concluded that the optimal number of cores needed for a successful biopsy was more than three cores from bone lesions and more than four cores from soft-tissue lesions. When the initial sample volume obtained was sufficient, the operators in our study did not obtain another core, but when the initial attempt yielded an insufficient volume, more attempts were made.

Fig. 5.

A 40-year-old woman presented with breast cancer. (a) Fat-suppressed T1W coronal MRI after contrast enhancement revealed a peripheral rim-enhancing lesion in the sternal body. (b) Axial CT image shows a biopsy needle using a direct perpendicular approach into the sternal lesion, which did not allow for adequate tissue retrieval. After surgical excisional biopsy, pathology revealed a metastatic carcinoma originating from the breast.

In the diagnostic group, 95.5% (21/22) of sternal lesions were malignancies. A previous study reported a 92.3% incidence of malignant disease in surgical sternal resections (27). This relatively high incidence of malignancy corresponds closely to our results. However, a selection bias cannot be excluded due to the high proportion of difficult cases seen at a tertiary referral center.

The main limitation of the study was its retrospective design. Consequently, there was a lack of uniformity in the biopsy procedures among our cases. Second, among the clinically diagnosed cases of inflammatory arthritides, we could not obtain a definite diagnosis for five patients. These patients’ diagnoses were dependent on symptom improvement after medication. Most diseases involving the sternoclavicular and manubriosternal joints are systematic; thus, careful investigation that includes patients’ symptoms, family history, and medication is necessary. Furthermore, surgery is rarely indicated, except in cases with infection or with refractory conditions that cannot be managed non-operatively (28). Third, all the diagnosis could not be confirmed by surgical resection after percutaneous biopsy. However, the majority of sternal lesions were confirmed to be metastases or multiple myelomas, which did not require surgery as the initial treatment of choice (29,30). Fourth, there is a lack of quantification of the quality and volume of the sample. Instead of quantification, we used the penetration length of the biopsy needle as a variable. Fifth, not all of the patients underwent PET-CT and MRI before CT-guided biopsy.

In conclusion, CT-guided percutaneous sternal biopsy has a relatively low diagnostic yield but, considering certain conditions related to pre-procedural CT density, lesion location, or angle of needle approach, it may be helpful for increasing the biopsy success rate for some cases.

Footnotes

Declaration of Conflicting Interests

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

Funding

The author(s) received no financial support for the research, authorship, and/or publication of this article.

References

Choi

Davis

Blankenbaker

. Percutaneous musculoskeletal biopsy. Semin Roentgenol 2004; 39: 114–128.

Espinosa

Jamadar

Jacobson

. CT-guided biopsy of bone: a radiologist’s perspective. Am J Roentgenol 2008; 190: W283–289.

Welch

. Percutaneous image-guided biopsy of the musculoskeletal system. Tech Vasc Interv Radiol 2011; 14: 110–117.

Saifuddin

Mitchell

Burnett

. Ultrasound-guided needle biopsy of primary bone tumours. J Bone Joint Surg Br 2000; 82: 50–54.

Gil-Sanchez

Marco-Domenech

Irurzun-Lopez

. Ultrasound-guided skeletal biopsies. Skeletal Radiol 2001; 30: 615–619.

Chao

Lin

Huang

. Color Doppler ultrasonographic evaluation of osteomyelitis in children. J Ultrasound Med 1999; 18: 729–734. quiz 735–726.

Ayala

Zornosa

. Primary bone tumors: percutaneous needle biopsy. Radiologic-pathologic study of 222 biopsies. Radiology 1983; 149: 675–679.

Yang

Damron

. Comparison of needle core biopsy and fine-needle aspiration for diagnostic accuracy in musculoskeletal lesions. Arch Pathol Lab Med 2004; 128: 759–764.

Stormby

Akerman

. Cytodiagnosis of bone lesions by means of fine-needle aspiration biopsy. Acta Cytol 1973; 17: 166–172.

10.

Domanski

Akerman

Carlen

. Core-needle biopsy performed by the cytopathologist: a technique to complement fine-needle aspiration of soft tissue and bone lesions. Cancer 2005; 105: 229–239.

11.

Rimondi

Rossi

Bartalena

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

12.

Goldsmith

Horwich

. Bone and soft-tissue lesions: what factors affect diagnostic yield of image-guided core-needle biopsy? Radiology 2008; 248: 962–970.

13.

Nishimura

Zhang

Ohkuma

. MR imaging of soft-tissue masses of the extraperitoneal spaces. Radiographics 2001; 21: 1141–1154.

14.

Adams

de Klerk

Heggelman

. Malignancy rate of biopsied suspicious bone lesions identified on FDG PET/CT. Eur J Nucl Med Mol Imaging 2016; 43: 1231–1238.

15.

Datir

Pechon

Saifuddin

. Imaging-guided percutaneous biopsy of pathologic fractures: a retrospective analysis of 129 cases. Am J Roentgenol 2009; 193: 504–508.

16.

Restrepo

Martinez

Lemos

. Imaging appearances of the sternum and sternoclavicular joints. Radiographics 2009; 29: 839–859.

17.

Söderlund

. Combined radiology and cytology in the diagnosis of bone lesions–a review of 399 cases. Acta Orthop Scand Suppl 2004; 75: 51–56.

18.

Layfield

. Cytologic diagnosis of osseous lesions: a review with emphasis on the diagnosis of primary neoplasms of bone. Diagn Cytopathol 2009; 37: 299–310.

19.

Koscick

Petersilge

Makley

. CT-guided fine needle aspiration and needle core biopsy of skeletal lesions. Complementary diagnostic techniques. Acta Cytol 1998; 42: 697–702.

20.

Handa

Bal

Mohan

. Fine needle aspiration cytology in the diagnosis of bone lesions. Cytopathology 2005; 16: 59–64.

21.

Yang

Frassica

Fayad

. Analysis of nondiagnostic results after image-guided needle biopsies of musculoskeletal lesions. Clin Orthop Relat Res 2010; 468: 3103–3111.

22.

Chang

Huang

Bredella

. Percutaneous CT-guided needle biopsies of musculoskeletal tumors: a 5-year analysis of non-diagnostic biopsies. Skeletal Radiol 2015; 44: 1795–1803.

23.

Jurik

. Noninflammatory sclerosis of the sternal end of the clavicle: a follow-up study and review of the literature. Skeletal Radiol 1994; 23: 373–378.

24.

Didolkar

Anderson

Hochman

. Image guided core needle biopsy of musculoskeletal lesions: are nondiagnostic results clinically useful? Clin Orthop Relat Res 2013; 471: 3601–3609.

25.

Omura

Motamedi

UyBico

. Revisiting CT-guided percutaneous core needle biopsy of musculoskeletal lesions: contributors to biopsy success. Am J Roentgenol 2011; 197: 457–461.

26.

Nouh

Abu Shady

. Initial CT-guided needle biopsy of extremity skeletal lesions: diagnostic performance and experience of a tertiary musculoskeletal center. Eur J Radiol 2014; 83: 360–365.

27.

Incarbone

Nava

Lequaglie

. Sternal resection for primary or secondary tumors. J Thorac Cardiovasc Surg 1997; 114: 93–99.

28.

Higginbotham

Kuhn

. Atraumatic disorders of the sternoclavicular joint. J Am Acad Orthop Surg 2005; 13: 138–145.

29.

Rochau

Jahn

Qerimi

. Decision-analytic modeling studies: An overview for clinicians using multiple myeloma as an example. Crit Rev Oncol Hematol 2015; 94: 164–178.

30.

Agarwal

Nayak

. Management of skeletal metastases: An orthopaedic surgeon’s guide. Indian J Orthop 2015; 49: 83–100.