Image-Guided Percutaneous Biopsy for the Diagnosis of Musculoskeletal Tumors: An Analysis of 698 Cases and a Literature Review

Purpose: Incisional biopsy is still used to obtain material for a straightforward histological diagnosis in musculoskeletal lesions. The aim of our study was to evaluate 1) the percentage of diagnostic procedures, 2) diagnostic accuracy, and 3) the incidence of complications after imaging-guided percutaneous biopsy


Introduction
Diagnosis for patients with musculoskeletal tumors is based on history, clinical, radiological, and histological examination. Most frequent mistakes in treatment are due to delay in diagnosis or misdiagnosis, affecting the patient's survival and the possibility of limb salvage [1,2]. When clinical and radiologic features do not allow to distinguish with certainty about biological behaviour or in patients with a history of malignancy, histological diagnosis is fundamental [1][2][3][4][5][6]. It is even more critical in sites in which salvage surgery is demanding, such as pelvis or limb reconstructions, where it is essential to preserve as much tissue as possible to allow soft tissue coverage and function [7][8][9][10]. Consequently, it is mandatory to always perform a biopsy before treating musculoskeletal lesions, with very few exceptions.
The biopsy is the last step of staging, and it is a compromise between the need to have significant tissue and the need to avoid local or systemic contamination [3][4][5][6]. Different procedures such as incisional biopsy (IB), fine-needle aspiration (FNA), or percutaneous biopsy (PCB) were used with specific advantages and disadvantages . IB has high diagnostic accuracy but also complications up to 17% [11-14, 21, 24]. FNA has limited application in musculoskeletal oncology because of the need for architectural evaluation: it should only be used for diagnosis of recurrence [12][13][14][15][16][17][18]. PCB has few complications (less than 7%), even if its diagnostic accuracy is lower than IB (range, 76-97%) . PCB could be performed CT, ultrasound or fluoroscopy guidance, in order to target tumor area [14, 18, 20, 21, 26-28, 31, 34-36, 38, 40-44]. In the last years, new types of PCB guided were proposed like MRI-guided biopsy, PET/CT-guided biopsy, or radionuclide-guided biopsy; however, these techniques are costly and are not used routinely in clinical practice [45][46][47]. The aim of this study was to review our experience with imaging-guided PCB for musculoskeletal lesions in order to define if this could be defined as the gold standard for diagnosis, evaluating 1) percentage of diagnostic procedures, 2) diagnostic accuracy and 3) incidence of complications.

Case Series
This is a retrospective series of imaging-guided PCB performed at our institution between January 2016 and September 2019. PCB was performed in all cases under imaging guidance: using 14-gauge (1, 62 mm) semi-automatic tru-cut needle guided by ultrasound in soft tissue lesions (322 cases), or with 8-gauge (3, 26 mm) core needle in bone lesions guided by fluoroscopy (376 cases). Sites are summarized in (Tables 1 & 2). For each patient, we reviewed data regarding sex, age, site, history and imaging, expected diagnosis, type of biopsy, complications, histopathology report of biopsy and final histopathology report in those subsequently treated. All PCBs were performed after accurate patient history, clinical and radiological evaluations according to 3 main principles. First, a biopsy was located along the surgical approach for future limb-salvage resection in order to be able to remove en-bloc the needle track (potentially contaminated by tumor cells) at the time of definitive resection. Second, the biopsy was performed avoiding contamination of compartments not involved by tumor to prevent cell dissemination that could determine the necessity of major tissue removal during further surgery. Third, areas of the lesion where biopsy was performed were chosen according to preoperative MRI or CT scan, selecting zones with contrast enhancement, avoiding necrotic tissue, including capsule or pseudocapsule. A mean of 7 samples was performed (range, 6-10), and the material was transferred to our Pathological Institute (and to microbiological evaluation when osteomyelitis is suspected). Pathologists specialized in orthopaedic oncology evaluated all specimens for the nature of the lesion and specific histological diagnosis. We classified specimens as "diagnostic" (sufficient material) or "nondiagnostic" (insufficient material), according to the histological diagnosis. Moreover, we compared histopathological results from the biopsy with the definitive one in patients subsequently treated surgically. In these cases, we also evaluated Sensitivity, Specificity, Positive Predictive Value (PPV), Negative Predictive Value (NPV), and Diagnostic Accuracy.
In soft tissue, PCB was diagnostic in 304 cases (304/322, 94.4%). In non-diagnostic cases, the biopsy was repeated as PCB in 14 cases and as IB in 4 cases (tumor close to vessels and nerves). In the 14 repeated PCB, diagnostic specimens were obtained in 11, while 3 required further IB. At least 315 PCBs were diagnostic (98%). Diagnoses of soft tissue lesions were reported in (

Discussion
Obtaining a specific histological diagnosis is mandatory before the treatment of musculoskeletal lesions [1][2][3][4][5][6]. Different types of biopsy could be used with related advantages and disadvantages [3][4][5][6]. PCB is reported to be a safe procedure burdened by few minor complications only; however, its diagnostic accuracy is lower than with IB due to the limited material obtainable, especially in soft tissue lesions usually presenting necrotic tissue, that makes anatomopathological diagnosis more complicated (Table 5) . For this reason, incisional biopsy is still used in some Specialized Centers.
The aim of our study was to review our experience with PCB for the diagnosis of musculoskeletal lesions in order to evaluate if this type of biopsy could be defined as the gold standard. All PCBs were imagingguided to provide significant/adequate samples containing vital tumor cells targeting the correct tumor area based on preoperative studies (CT or MRI with contrast).  This study had some limitations. First, this is a retrospective nonrandomized series, in which we perform PCB for bone as well as soft tissue lesions. However, PCB was performed with a tru-cut needle guided by ultrasound in all soft tissue lesions and with a core needle guided by fluoroscopy in all bone lesions. Second, PCB was performed to detect tumor as well as to exclude it. Consequently, not all patients underwent surgery, according to the treatment required by histological diagnosis, limiting our series and statistical power of analysis. However, non-surgically treated patients were periodically followed in the outpatient clinic, and no cases of misdiagnosis were observed. Third, diagnoses were heterogeneous, precluding evaluation of diagnostic accuracy related to histotypes. However, these kinds of tumors are rare, and we do not preclude further study with more specific analysis.
Our results seem to be better than those reported in the literature obtaining only 2.7% and 5.6% of non-diagnostic samples, respectively in bone (we used 8 Gauge trocar) and soft tissue (we performed at least 6 samples). Moreover, targeting with CT the proper area of the lesion to obtain vital tumor cells could be useful. Mitsuyoshi et al. [18] compared CT guided and fluoroscopy-guided PCB in 163 patients and report a decrease of non-diagnostic results with CT compared with fluoroscopy (7% vs. 14%), concluding that CT-guidance is more efficacious, thanks to high contrast resolution and spatial definition. However, since CT is more expensive and exposes to a high dose of radiation, its use should be limited.
Also, the diagnostic accuracy is usually higher for bone tumors than for soft-tissue masses, ranging from 80% to 100% and from 76% -100%, respectively [13-15, 18, 26-28, 35-38, 40-42]. The diagnosis of bone lesions can be aided by imaging tools, whereas soft-tissue lesions have more tumor necrosis and more different diagnoses [13]. In our series, diagnostic accuracy was determined by comparing histopathological results from PCB and subsequent tumor resection, in patients that subsequently were surgically treated. We identified an overall diagnostic accordance of 98% in bone lesions, and 95% in soft tissue tumors.
Complications are frequent (up to 17%) when IB is performed [6,11,13,14,18,23]. Mankin et al., were the first in 1982 to report the percentages of complications that occurred following an open biopsy in 329 patients [6]. In 17% of cases, there were hematomas, infections, surgical wound dehiscence, and dispersion of tumor cells in adjoining tissues. In 8.5% of patients, it was found that this procedure negatively affected the prognosis. Finally, it emerged that 4.5% of the patients underwent amputation following the open biopsy [6]. Conversely, complications after PCB are infrequent (less than 2%) for both bone and soft tissue lesions and are usually minor complications, such as biopsy tract infection, post-procedural bleeding, transient paresis [14,15,[26][27][28][35][36][37][38][39]. According to the literature, in our series, we did not encounter any complications related to PCB.
In conclusion, PCB is a safe, minimally invasive, and cost-effective technique for the diagnosis of bone and soft tissue lesions. Indications should be carefully evaluated by an experienced orthopaedic oncologist concerning the suspected entity, size, and location of the lesion to avoid incorrect or deficient results. Targeting the proper area of the lesion to obtain vital tumor cells is mandatory; it is based on pre-biopsy careful imaging evaluation (especially MRI) and imaging-guidance during the procedure. Obtaining multiple samples from different areas of the tumor is essential to have more representative specimens. IB should be reserved for "difficult" cases or after previous non-diagnostic PCBs.