In recent studies, diagnostic accuracy ranges from 74% to 98% for CNB of bone and soft tissue tumours
[8–13]. In line with these results, in our study we identified an overall diagnostic accuracy of 92.9% in soft tissue tumours and 98% in bony lesions, although many authors also included metastatic tumours, usually yielding higher accuracy than more heterogeneous sarcomas
[14, 15]. However, Sung et al. found no significant differences regarding diagnostic accuracy between CNB of heterogeneous and homogeneous bone tumours
. The correct histopathological diagnosis compared to the resection specimen was obtained in 84.2% of soft tissue masses and 93.9% of bony lesions in our study, confirming previously published results
[8, 9, 17].
Comparing open biopsy and CNB in suspected malignant bone tumours, we identified slightly superior results for CNB. However, considering the relatively small number of cases included in our series, the results should presumably be almost equal in a larger patient sample. CNB of soft tissue tumours revealed inferior results compared to open biopsy in the current study. These findings correlate well with recent literature, indicating that very heterogeneous tumours including liposarcoma, angiosarcoma and synovial sarcoma are, amongst others, potentially difficult to diagnose by CNB
[9, 16, 17]. In this context, Kasraeian et al. found significantly deficient results regarding specific diagnosis in a prospective study comparing open biopsy with CNB in soft tissue tumours
. Favourable results were obtained, and there was only one non-diagnostic sample in our study (7.6%). However, this is presumably attributed to our relatively small patient sample.
To prevent lower diagnostic accuracy due to insufficient or inadequate sampling in CNB, McCarthy proposed a full discussion between orthopaedic surgeons, radiologists and pathologists to identify suitable tumour regions prior to CNB
. Additionally, the procedure should be conducted by a well-trained interventional radiologist or orthopaedic surgeon
[4, 7, 19, 20]. Regardless of meeting either recommendation, one CNB sample was insufficient in our series emphasizing the importance of harvesting tissue from multiple passes within the tumour.
A major disadvantage of CNB, apart from insufficient sampling, is the fixation of the tissue in formalin, as pathologic analysis is mostly limited to histology and immunohistochemistry, and frozen samples are needed for recent molecular diagnostic approaches such as real-time PCR (rt-PCR) or microarray analysis
[21, 22]. Furthermore, the acquisition of additional tissue for a cryobank or research purposes is limited with CNB.
On the other hand, CNB is associated with less morbidity and fewer complications compared to open biopsy. In recent studies, complication rates for open biopsy ranging from 0% to 17% are quoted
[4, 7, 12, 18, 19]. Core needle biopsy, in contrast, has a reported complication rate of 0% to 7.4%
[7, 8, 11, 12, 18, 23, 24]. Most commonly, haematoma, bleeding and infection are mentioned
[12, 25, 26]. In our series there was no significant complication associated with either biopsy technique. However, it has to be noted that all open biopsies were performed according to the sarcoma principles with intensive haemostasis and application of a redon drain
Overall, we identified good results for both biopsy techniques. However, we have to note some limitations of our study mainly due to the retrospective design. First, we were not able to compare the pathological grading of biopsy samples to the resection specimens, because grading was not performed in every case for two main reasons. One problem was possible false negative results if a low-grade area within these very heterogeneous tumours was sampled. Furthermore, grading of tumour samples harvested after neoadjuvant chemotherapy may be misleading and thus, is rarely conducted. A subsequent grading for this study was not possible in these cases. Second, we performed only one kind of biopsy technique in each patient. The indication for either procedure was made by a well-trained senior orthopaedic surgeon based on the clinical and radiological findings as well as his experience. Accordingly, there is a possible bias in favour of core needle biopsy. Third, we investigated a quite heterogeneous study population in terms of tumour entities. Although we only included patients with suspected sarcomas, great differences remain regarding histopathological analysis. Furthermore, the diagnostic standard by which both techniques were judged was based on complete surgical resection and the final clinical diagnosis of the orthopaedic oncologist. Although this is a good measure, a possibly wrong diagnosis could influence our results.