The safety of primary surgical excision of small deep indeterminate musculoskeletal soft tissue masses

Abstract

Objective:

To determine the suitability of primary excision of small indeterminate deep soft tissue masses presenting to a tertiary musculoskeletal oncology service.

Methods and materials:

Review of all patients referred to a specialist musculoskeletal oncology service over a 20-month period with a deep indeterminate soft tissue mass by non-contrast MRI criteria that was recommended for primary surgical excision due to relatively small size (<30 mm). Data collected included age, gender, site and maximal size of the lesion, and final histological diagnosis for excised lesions.

Results:

85 patients were included, mean lesion size being 12 mm (range 5–29 mm). Primary surgical resection had been undertaken in 69 cases (81.2%) by the conclusion of data collection, 36 males and 33 females with mean age of 45.6 years (range 11–80 years). Of these, 11 cases (15.9%) were non-neoplastic, 53 (76.8%) were benign, 1 (1.4%) was intermediate grade, while 4 (5.8%) were malignant including 3 synovial sarcomas. Two of these were treated with re-excision of the tumour bed showing no residual disease, with no evidence of local recurrence at a mean of 10.7 months post-excision.

Conclusion:

Primary surgical excision of small deep soft tissue masses that are indeterminate by non-contrast MRI criteria is considered a safe procedure when undertaken in a specialist musculoskeletal oncology service, with only 4 of 69 cases (5.8%) being malignant.

Advances in knowledge:

Small indeterminate deep soft tissue masses can safely be treated with primary excision in the setting of a specialist musculoskeletal oncology service.

Introduction

The differential diagnosis for patients presenting with a soft tissue mass (STM) is wide. The latest edition of the World Health Organisation (WHO) Classification of Tumours of Soft Tissue 2020 highlights the broad range of differential diagnoses of a STM.¹

The clinical and imaging differentiation between benign and malignant STMs is challenging, although many non-neoplastic and benign lesions do have classical MRI features which may allow confident primary surgical excision without the need for pre-operative needle biopsy.^2–4 However, despite the use of state-of-the-art MRI with dedicated tumour protocols as well as functional MRI techniques,^5,6 a significant proportion of STMs cannot be characterised accurately. These require image-guided core needle biopsy (IGCNB), as advised by UK and European Guidelines.^7,8 Accurate histological diagnosis and staging of tumours preceded and followed by discussion within a sarcoma multidisciplinary team (MDT) meeting is critical in planning optimal management, including surgery.⁷

Ultrasound-guided needle biopsy is established as a safe and accurate technique for the investigation of STMs,^9,10 and biopsy of small (<20 mm) tumours arising in either the superficial or deep compartments is also reliable for achieving a diagnosis.¹¹ However, the safety of percutaneous biopsy of small masses in terms of possible tumour spread, biopsy tract seeding, and late local recurrence is still somewhat controversial.^12–14 In addition, in our experience, tissue yield from biopsy of lesions <15 mm can be less reliable, and this has also been reported by other authors.^11,15 For this reason, it is our preference to recommend primary surgical excision for small (<30 mm) STMs irrespective of their location in relation to the deep fascia. However, this must be undertaken in a specialist sarcoma service, since it is well-known that unplanned excision of small sarcomas which were thought to be benign is associated with a high rate of residual disease.¹⁶

The safe primary excision of small, indeterminate superficial soft tissue masses undertaken in a specialist sarcoma service has been reported.¹⁷ The purpose of the current study is to assess whether primary surgical excision of indeterminate small deep STMs is also a safe practice within a specialist sarcoma service.

Methods and materials

The study was approved by the local Research and Innovation Centre of The Institute of Orthopaedics under the Integrated Research Application System number 262826, with no requirement for informed patient consent.

The study was a retrospective review of all patients referred to a Specialist Musculoskeletal Sarcoma Service between July 2018 and April 2020. The patient data have been collected prospectively following each MDT meeting, so no patients were missed. Included patients fulfilled the following criteria: 1) the STM was located deep to the deep peripheral fascia surrounding muscle¹⁸ ; 2) was not a simple lipoma; 3) was not intra-articular; 4) was recommended for primary surgical excision at the MDT meeting due to indeterminate MRI appearances and small size (<30 mm) such that IGCNB was considered difficult/unsafe. All lesions of the fingers and toes were considered to be deep in location. This decision was made by a Consultant Musculoskeletal Radiologist with >25 years’ experience of musculoskeletal tumour imaging but was also agreed by the treating surgeons. Studies had been re-reviewed by a second Consultant Musculoskeletal Radiologist with 13 years’ experience of tumour imaging at the termination of data collection. Data collected included patient age, gender, location of the lesion, and maximal tumour dimension.

The majority of patients had undergone MRI studies prior to referral, which included various combinations of T₁ weighted turbo spin echo (T₁W TSE), short tau inversion recovery (STIR), T₂ weighted fast spin echo (T₂W FSE), proton density weighted fast spin echo (PDW FSE) and axial spectral attenuated inversion recovery (SPAIR) sequences in the sagittal, axial or coronal planes. The remainder of the patients had imaging based on our local tumour imaging protocol, including STIR, T₁W TSE, T₂W FSE, PD FSE in sagittal, axial and coronal planes. Contrast enhancement was not utilised in any of the cases. Based on review of the MRI studies, a specific diagnosis was suggested, or the lesion was classed as indeterminate. Indeterminate lesions were typically of intermediate signal intensity (SI) on T₁W TSE, T₂W FSE and PDW FSE sequences, and mildly hyperintense on fat suppressed sequences with no specific morphological characteristics.

Final histological diagnosis was obtained by review of histopathology reports for those cases that had undergone surgical excision by the time of final data analysis. Tumours were classified as non-neoplastic, benign, intermediate or malignant based on the 2020 WHO Classification of Soft Tissue and Bone Tumours.¹ For patients who had not had their tumours excised, any follow-up imaging studies and the last available clinic note were reviewed to determine the reason.

Results

A total of 85 patients fulfilled the inclusion criteria. Mean maximal lesion size based on MRI was 12 mm (range 5–29 mm), 70 cases (82.4%) measuring up to 20 mm in maximal dimension. The remaining cases were all <30 mm in maximal dimension but were included since other factors such as small axial dimension or location were also felt to contribute to potential difficulty of IGCNB. The commonest lesion locations were the wrist/hand (n = 25; 29.4%) and ankle/foot (n = 20; 23.5%). Primary surgical resection had been undertaken in 69 cases (81.2%) by the conclusion of data collection, 36 males and 33 females with mean age of 45.6 years (range 11–80 years). The location of the 69 excised lesions is presented in Table 1.

Table 1.

Anatomical distribution of 69 excised lesions

Body region	Subdivision	No. of lesions (%)
Upper limb	Shoulder girdle	2 (2.9%)
	Arm	4 (5.8%)
	Elbow	2 (2.9%)
	Forearm	5 (7.2%)
	Wrist & hand	19 (27.5%)
Lower limb	Pelvis, groin, buttock	0 (0%)
	Thigh	5 (7.2%)
	Knee	2 (2.9%)
	Calf	4 (5.8%)
	Ankle & foot	17 (24.6%)
Trunk	Chest wall/trunk	7 (10.1%)
	Abdominal wall	2 (2.9%)

Of the 69 lesions for which a histological diagnosis was available, 11 cases (15.9%) were non-neoplastic (Figure 1), 53 (76.8%) were benign (Figures 2 and 3), 1 (1.4%) was intermediate grade (Figure 4), while 4 (5.8%) were malignant (Figures 5–7). The final diagnoses are presented in Table 2. There was a wide range of differential diagnoses, but schwannoma, angioleiomyoma and tenosynovial giant cell tumour (tGCT) were the three commonest lesions, accounting for 32 (46.4%) of the all excised cases, and 60.4% of benign tumours.

Figure 1.

A 38-year-old male with a small soft tissue mass adjacent to the middle finger proximal interphalangeal joint. (a) Coronal T₁W TSE and (b) axial T₂W FSE MR images show a lobular intermediate SI lesion (arrows) which measured 21 mm in maximal dimension. The imaging features were considered indeterminate. Histology confirmed a diagnosis of tophaceous gout.I, signal intensity; T₁W TSE, T₁ weighted turbo spin echo.

Figure 2.

A 37-year-old male with a small soft tissue mass in the right thigh. (a) Coronal T₁W TSE, (b) sagittal T₂W FSE and (c) axial PDW FSE MR images show a lobular intermediate SI lesion with a hypointense margin (arrows) within the vastus lateralis muscle. The imaging features were considered indeterminate. Histology confirmed a diagnosis of nodular fasciitis. PDW, proton density weighted; SI, signal intensity; T₁W FSE, T₁ weighted fast spin echo.

Figure 3.

A 30-year-old male with a small mass in the right hand. (a) Coronal T₁W SE and (b) axial STIR MR images showing a small oval mass (arrows) located in the palm overlying the flexor tendons. Histology confirmed a diagnosis of fibroma of tendon sheath. STIR, short tau inversion recover; T₁W SE, T₁ weighted spin echo.

Figure 4.

A 37-year-old female with a small mass in the right abdominal wall. (a) Sagittal T₂W FSE and (b) axial SPAIR MR images show a fairly well-defined oval mass (arrows) within the right rectus abdominis muscle. The imaging features were considered indeterminate. Histology confirmed a diagnosis of desmoid-type fibromatosis. SPAIR, spectral attenuated inversion recovery; T₂W FSE, T₂ weighted fast spin echo.

Figure 5.

A 52-year-old male with a small mass in the right foot. (a) Sagittal T₁W TSE and (b) coronal T₂W FSE MR images show a poorly defined low SI mass (arrows) within the deep flexor muscles. The imaging features were considered indeterminate. Histology confirmed a diagnosis of monophasic synovial sarcoma, Grade 2 fibrous-type. SI, signal intensity; T₁W TSE, T₁ weighted turbo spin echo;T₂W FSE, T₂ weighted fast spin echo.

Figure 6.

A 29-year-old female with a small soft tissue mass in the hand deep to the palmar fascia. (a) Sagittal T₁W TSE, (b) coronal T₂W FSE and (c) axial SPAIR MR images show a lobular intermediate SI lesion with (arrows) with poorly defined infiltrative margins and associated muscle oedema (arrowheads-c). The imaging features were considered indeterminate. Histology confirmed a diagnosis of bi-phasic synovial sarcoma Grade 2. SI, signal intensity; SPAIR, spectral attenuated inversion recovery; T₁W TSE, T₁ weighted turbo spin echo; T₂W FSE, T₂ weighted fast spin echo.

Figure 7.

A 48-year-old female with a small soft tissue mass in the extensor compartment of the proximal forearm. (a) Coronal T₁W TSE MR image shows an elongated intermediate SI lesion (arrow) with an associated ‘split fat’ sign (arrowheads) suggestive of a benign peripheral nerve sheath tumour. (b) Axial SPAIR MR image shows a mildly hyperintense lesion (arrow) with no surrounding reactive changes. (c) Axial SPAIR MR-image 8 months later shows clear growth of the lesion (arrow). The imaging features were considered indeterminate. Histology confirmed a diagnosis of fibrous-type monophasic synovial sarcoma Grade 2. SI, signal intensity; SPAIR, spectral attenuated inversion recovery; T₁W TSE, T₁ weighted turbo spin echo.

Table 2.

Histological diagnoses for 66 excised lesions categorised into non-neoplastic, benign, intermediate and malignant (alphabetical order and number in parentheses)

Non-neoplastic (11)	Benign (53)	Intermediate (1)	Malignant (4)
Bland osteocartilaginous lesion [accessory bone] (1)	Angioleiomyoma (7)	Desmoid-type fibromatosis (1)	Synovial sarcoma (3)
Degenerate skeletal muscle (1)	Angiolipoma (1)		Epithelioid haemangioendothelioma (1)
Epidermal cyst (1)	Angiomyxoma -superficial (1)
Gout (1)	Atypical fibrous histiocytoma (1)
Non-specific intravascular thrombosis (1)	Desmoplastic fibroblastoma (1)
Scar-like tissue with fat necrosis (3)	Endometriosis (2)
Reparative process (1)	Fibroma of tendon sheath (3)
Vascular malformation (1)	Haemangioma (3)
Bursa (1)	Intramuscular nodular fasciitis (4)
	Myopericytoma (1)
	Neurofibroma (1)
	Perineurioma (1)
	Periosteal chondroma (1)
	Schwannoma (18)
	Superficial palmar fibromatosis (1)
	Tenosynovial giant cell tumour (7)

Three of the malignant lesions were synovial sarcomas, these located in the foot, hand and proximal forearm. The lesion in the foot occurred in a 51-year-old male and measured 23 mm in maximal dimension (Figure 5), showing no change in size between MRI studies performed at a 4 month interval. It had particularly low SI on T₂W FSE sequences suggestive of a fibrous tumour, with no surrounding reactive change. Histology confirmed a diagnosis of monophasic fibrous-type synovial sarcoma Grade 2. Re-resection of the tumour bed revealed no evidence of residual tumour.

The lesion in the palm of the hand occurred in a 29-year-old female and showed mild increase in size on MRI studies obtained 7 months apart, the lesion measuring 12 × 8 x 6 mm immediately prior to excision. It was also associated with poorly defined margins and reactive oedema in the surrounding muscles (Figure 6). Histology revealed bi-phasic synovial sarcoma Grade 2. Post-operative MRI showed no evidence of macroscopic residual disease and the patient is under active surveillance.

The lesion in the proximal forearm occurred in a 49-year-old lady. The initial MRI study obtained 9 months prior to excision demonstrated a ‘split fat’ sign on the coronal T₁W TSE sequence (Figure 7a) raising the possibility of a benign peripheral nerve sheath tumour. However, it had clearly grown on an MRI study obtained 1 month prior to excision, measuring 13 × 8 x 6 mm in dimension (Figure 7b and c). Mild oedema was seen at the distal pole of the lesion. Histology confirmed a diagnosis of monophasic fibrous-type synovial sarcoma Grade 2. No residual tumour was identified on re-excision of the tumour bed.

In the patient with epithelioid haemangioendothelioma, the tumour was located to the cubital fossa in intimate relation to the cubital vein (Figure 8), with a maximum dimension of 14 mm which would have presented a significant challenge to IGCNB. Excision biopsy was performed with clear margins, and the patient did not require any further treatment.

Figure 8.

A 48-year-old female with a small painful mass in the right cubital fossa. (a) Sagittal T₂W FSE and (b) axial PDW FSE MR images show an oval mass (arrows) closely related to the cubital vein and median nerve. The imaging features were considered indeterminate. Histology confirmed a diagnosis of epithelioid haemangioendothelioma. PDW, proton density weighted; T₁W FSE, T₁ weighted fast spin echo.

Of the 16 patients for whom no excision had been undertaken, 6 are still awaiting excision and 1 is awaiting up-to-date MR imaging. In three cases, repeat MRI showed either a stable appearance or slight reduction in lesion size, so these patients are being treated conservatively. In three cases, the patients chose to not proceed to excision, while in one case ultrasound examination of the lesion confirmed a slow-flow vascular malformation. In one case, the patient was referred back to the initial hospital, while the last case was lost to follow-up. Mean clinical follow-up from the date of first MRI was 13 months (range 1–24 months).

Discussion

MRI characterisation of STMS is an important part of the diagnostic process, being especially useful in selected pathological entities such as lipomatous tumours,¹⁹ benign nerve sheath tumours,²⁰ and vascular malformations.²¹ Ultrasound can play a complementary role in diagnosis, being most useful when assessing more superficial lesions, especially those superficial to the fascia.²² It is useful in distinguishing cystic and solid masses, e.g. in the case of ganglia and epidermoid cysts²³ and is also excellent for the confirmation of a vascular malformation,²⁴ and in the evaluation of the neural origin of peripheral nerve sheath tumours.²⁵ More recently, shear wave elastography has been utilised for the differentiation of benign from malignant STMs, but provides no further useful information compared to MRI.²⁶

Lesions with classical imaging appearances may be safely treated without diagnostic needle biopsy. These include benign peripheral nerve sheath tumours (PNSTs). Desai described primary excision of 442 suspected benign PNSTs based on clinical and MRI findings using intraoperative frozen section to confirm the diagnosis, finding that the suggested diagnosis was correct in all but five cases (98.9%).²⁷ A similar approach can be adopted with suspected tGCT, the MRI features of which have also been well-described. Its location in close relation to a tendon, well-defined lobular morphology, and typically profound low SI on T₂W FSE and gradient echo sequences due to haemosiderin deposition being very characteristic features.^28–30

However, in the setting of indeterminate STMs, the standard approach to diagnosis involves IGCNB, usually with ultrasound-guidance which has been shown to have high diagnostic accuracy and is reliable for the differentiation between benign and malignant lesions.^9–11 In our practice, there are some situations where IGCNB is not considered to be ideal, e.g. if the STM is relatively small (<30 mm), too deep and therefore difficult to target, or small and closely related to major neurovascular structures. We have adopted this practice due to the unknown risks of IGCNB in respect of the potential for spreading tumour cells into adjacent muscle compartments, thereby possibly resulting in incomplete resection and/or late local recurrence. Such lesions are usually treated with primary surgical excision.

The risk of malignancy in a small deep indeterminate STM is difficult to quantify. Studies looking at differentiating features between benign and malignant soft tissue tumours commonly use 5 cm as the cut-off value,^31,32 which does not help in determining the risk of malignancy in lesions measuring <30 mm in maximal dimension. A study by Lee et al³³ identified 22% of STMs measuring <30 mm as being malignant, although it is unclear if these were superficial, deep or a combination of both. Data does exist with regards small indeterminate superficial STMs. Khoo et al identified approximately 7% of 58 indeterminate superficial STMs < 30 mm in size as being malignant.¹⁷ In a study of small (<20 mm) indeterminate STMs by Pham et al,³⁴ 21% of 42 trunk and extremity lesions were malignant, of which 10 cases were due to melanoma and only 2 lesions were located deep to the fascia.

Based on available literature, we were unable to inform our treating surgeons of the likelihood that a small indeterminate deep STM is malignant when a decision is made to proceed to primary excision biopsy. Therefore, we decided to undertake the current study to determine the safety of this practice, the results suggesting that this is a safe option in the setting of a Specialist Musculoskeletal Oncology Service. Of the 66 indeterminate lesions, all but 5 (7.6%) were either non-neoplastic lesions or benign tumours. One case (1.5%) was classed as an intermediate-grade lesion with locally aggressive behaviour based on WHO 2020¹ (desmoid-type fibromatosis), while there were only 4 malignant lesions (6%), 3 cases of synovial sarcoma involving the foot, hand and proximal forearm, and 1 epithelioid haemangioendothelioma involving the elbow. All of these four lesions were small and less than 15 mm in maximal dimension, one measuring no more than 8 mm. Two of the synovial sarcomas were adherent to palmar and plantar fasciae respectively, while the epithelioid haemangioendothelioma was adherent to the cubital vein. All these factors made IGCNB both potentially unsafe as well as possibly giving poor biopsy yield.

Of the 16 patients who did not undergo resection, 3 tumours showed either stable appearances or slight reduction in size on follow-up MRI, while a further tumour was shown to be a slow-flow vascular malformation on ultrasound. Three further patients chose follow-up with no clinical evidence of tumour growth. Therefore, if these seven cases are also considered to represent non-neoplastic or benign lesions, the risk of a small deep STS is more likely to be in the region of 5%.

Of the three patients with synovial sarcoma, two underwent re-excision of the tumour bed as is recommended practice,¹⁷ and at a mean of 10.7 months follow-up (range 6–19 months), there is no clinical evidence of recurrent disease. In the patient with the epithelioid haemangioendothelioma, this was resected with clear margins requiring no further treatment and the patient remains clinically disease-free at 20 months. A recent multicentre study has suggested that unplanned excision of soft tissue sarcoma does not adversely affect long term outcome,³⁵ which lends further support to this treatment option.

A wide range of non-neoplastic and benign tumours were identified (Table 2), similar to previous studies of small indeterminate STMs which were predominantly in the superficial compartment.^18,35 The commonest benign tumour was schwannoma. As in the current study, a previous report on intramuscular PNSTs has documented the absence of classical features such as the entering and exiting nerve sign, and the target sign.³⁶ The ‘split fat sign’ or ‘fatty rind’ sign has also been described in a series of 50 intramuscular tumours, being identified in 37 cases, 36 of which were benign and most commonly PNSTs.^{37 38} Only a single sarcoma demonstrated this feature, similar to the synovial sarcoma in the current series. However, the lesion had a maximal dimension of 11 cm, as opposed to the lesion we described which had a maximal dimension of 13 mm. Synovial sarcoma presenting as a small indeterminate lesion on MRI has been previously reported, all patients having undergone primary excision due to the ‘benign’ appearance of the lesions.³⁸ In two of our three cases, interval MRI showed tumour growth in a relatively short time, and this should possibly have indicated the malignant nature of the lesion.

The study has several limitations. The decision-making process was based on the opinion of a single radiologist, who nevertheless had >25 years’ experience of musculoskeletal tumour imaging. The decision was also made in an MDT setting which allowed the treating surgeons to give an opinion as to the ease/suitability of primary resection for each of the cases. Contrast-enhanced studies were not included since the overlap between benign and malignant lesions is sufficiently great to limit its value in further characterisation.²³ Additional techniques including diffusion-weighted imaging (DWI) and dynamic post-contrast MRI (DCE-MRI) were not employed. DWI has been shown to increase the diagnostic confidence for identifying small malignant superficial STMs,³⁹ while DCE-MRI may also aid in the differentiation of indeterminate STMs.⁴⁰

In conclusion, the current study would suggest that the primary excision of small (<30 mm) deep STMs which are indeterminate in nature by non-contrast MRI criteria is a safe option, with only 4 of 66 histologically proven cases representing a high-grade sarcoma. However, the decision must be made within the setting of a specialist musculoskeletal oncology MDT setting, and the excision undertaken by specialist orthopaedic oncologists.