Abstract
Objectives:
To determine the differential diagnosis of musculoskeletal soft tissue masses in children referred to a specialist musculoskeletal oncology unit.
Methods:
All children (0–18 y) referred to a specialist musculoskeletal oncology unit over a 20-month period (September 2018–May 2020) were retrospectively reviewed. Demographic data and referral diagnoses were obtained from the electronic patient notes. MRI findings and histopathological results were recorded. The comparison of non-neoplastic, benign neoplastic and malignant diagnoses at the point of referral and final diagnosis was determined.
Results:
116 patients were included, 60 (51.7%) males and 56 (48.3%) females with mean age of 10.6 years (3 months–18 years). 69 (59.5%) patients were referred with a suspected sarcoma, 29 (25.0%) with a suspected benign tumour and 18 (15.5%) with a non-neoplastic lesion. A diagnosis was achieved by histological assessment in 61 (52.6%) cases, microbiological assessment in 3 (2.6%) or clinical and imaging assessment in 52 (44.8%). 67 (57.8%) cases had non-neoplastic pathology, 39 (33.6%) a benign tumour, 4 (3.4%) an intermediate-grade tumour, 6 (5.2%) a malignant tumour.
Conclusions:
Although over half of children referred to a specialist musculoskeletal oncology unit were suspected of having a soft tissue sarcoma at referral, only 5.2% were diagnosed with a malignant tumour.
Advances in knowledge:
Approximately, 6 of 69 (8.7%) children referred to a specialist musculoskeletal oncology unit with a suspected soft tissue sarcoma will have a malignant lesion. Most paediatric soft tissue masses are non-neoplastic, the commonest diagnosis being a vascular malformation.
Introduction
In England, soft tissue sarcoma (STS) is the fourth commonest cancer group diagnosis in children following (1) leukaemias, myeloproliferative and myelodysplastic diseases, (2) lymphoma and reticuloendothelial neoplasms and (3) central nervous system neoplasms.1 Between 2001 and 2015, there were 1369 new diagnoses of soft tissue and extraosseous sarcoma in children under 15 years of age living in England.1 Multiple studies and national statistics reflect the overall prevalence of benign and malignant soft tissue lesions in larger populations,1–3 but to the best of our knowledge, there is little published evaluating the prevalence of STS in a paediatric population referred to a specialist musculoskeletal oncology unit.
Diagnostic work-up includes a combination of clinical review, imaging, histological and microbiological assessment. Imaging strategies for the assessment of paediatric soft tissue tumours have been published,4 and multiple clinical and imaging review articles have also been written dealing with paediatric soft tissue tumours in general,5–7 benign soft tissue masses in children,8,9 paediatric STSs,10–12 as well as specific tumour types such as fat containing masses,13,14 vascular lesions,14,15 and various fibroblastic/myofibroblastic lesions.16,17 A study by Sanbri et al identified that rhabdomyosarcoma (RMS) accounted for 33% of limb STSs in children.18
The aim of the study was to determine the differential diagnosis of soft tissue masses in children referred to a tertiary musculoskeletal oncology unit, and the prevalence of malignant tumours in this referral group.
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.
This was a retrospective review of MRI studies of consecutive patients aged 0–18 years referred to a specialist musculoskeletal oncology unit for evaluation of a soft tissue mass over a period of 20 months (September 2018–May 2020). These patients had either been referred directly to the service from the family doctor through the national “2 week wait” sarcoma pathway or by a secondary care clinician. Clinical data collected included patient age and gender, lesion location and suspected referral diagnosis. If a diagnosis was not explicitly stated by the referring team, this was taken to mean a suspected STS given referral to the sarcoma unit. Patients were only included if MRI had been undertaken. Children referred after a diagnosis of STS elsewhere were excluded, as were children with intra-articular soft tissue tumours or recurrence of a previous tumour.
MRI was obtained either prior to referral (n = 93) or following referral for children who did not have an MRI study (n = 32; 7 at 1.5 T and 25 at 3 T). Referral MRI studies included varied combinations of imaging protocols, while MRI studies performed following referral comprised a combination of T1 weighted fast spin echo (T1W FSE), T2 weighted fast spin echo (T2W FSE), short tau inversion recovery, proton density-weighted fast spin echo (PDW FSE) and spectral attenuated inversion recovery sequences. All MRI studies were reviewed by a consultant musculoskeletal radiologist with >24 years’ experience of musculoskeletal tumour imaging, and an imaging diagnosis was either suggested or the lesion was classified as indeterminate.
The subsequent diagnostic pathway was determined in the sarcoma multidisciplinary team (MDT) meeting and comprised clinical and imaging diagnosis, with subsequent referral to another specialist unit or back to the original referrer. This was typically the case for children diagnosed with a vascular malformation. All other lesions were either managed with imaged-guided core needle biopsy (IGCNB), surgical resection or conservative treatment with clinical and imaging follow-up when the lesion was not concerning for a STS. All specimens from IGCNB or surgical resection were analysed by experienced soft tissue tumour consultant histopathologists, and the final histological diagnosis was recorded. Lesions were then classified according to the final diagnosis into four distinct categories: non-neoplastic, benign neoplastic, intermediate-grade neoplastic and malignant neoplastic according to the World Health Organisation (WHO) Classification of Tumours of Soft Tissue 2020.19 With regards to vascular anomalies, according to the International Society for the Study of Vascular Anomalies Classification 2018 (ISSVA), vascular anomalies are sub divided into either vascular tumours (benign or malignant) or vascular malformations (non-neoplastic).20 For example, a haemangioma was classified as a benign vascular tumour, whilst a vascular malformation was classified as a non-neoplastic lesion.
Statistical analysis
Descriptive statistics were used to determine the differential diagnoses and compare the referral diagnoses with final diagnoses. Relationship between gender and final diagnosis was compared using the χ2 test, while relationship between age and final diagnosis was assessed using the analysis of variance (ANOVA).
Results
During the time period of the study, 125 referrals for soft tissue masses were made in patients aged 18 or younger for whom MRI studies were available. However, at the time of data analysis, nine of these had no final diagnosis and were therefore not considered further. There were 60 (51.7%) males and 56 (48.3%) females with mean age of 10.6 years (3 months–18 years). Table 1 demonstrates the age and gender distribution for the four histological classification groups, although due to small numbers the intermediate lesions were combined with benign tumours. This showed that male sex was commoner for STS and that STS occurred in a younger age group. However, neither of these findings reached statistical significance (p = 0.71 and 0.13 respectively). 30 (25.9%) patients were referred through the national “2 week wait” sarcoma pathway, whilst 86 (74.1%) patients were referred by a secondary care clinician (n = 86; 74.1%). Mean maximal tumour dimension was 42.5 mm (range 3–144 mm). 50 lesions (43.1%) arose superficial to the fascia, while 66 (56.9%) were located deep to the fascia. Table 2 provides details of lesion location. 69 (59.5%) patients were referred with a suspected sarcoma, 29 (25.0%) with a suspected benign tumour and 18 (15.5%) with a non-neoplastic lesion. Following review of MRI studies by the senior consultant, an imaging diagnosis was suggested in 75 cases (64.7%), while 41 cases (35.3%) were considered indeterminate in nature (Table 3). Based on MRI, 28 (24.1%) children were thought to have a vascular malformation (Figures 1 and 2). These were usually followed-up with an ultrasound examination typically demonstrating a compressible lobular anechoic mass with variable flow demonstrated within the anechoic components (Figures 1 and 2). This allowed a confident imaging diagnosis with no need for IGCNB in 26 cases. In two cases, the radiologist performing the compression ultrasound was not convinced of the diagnosis of vascular malformation, and therefore IGCNB was performed confirming the diagnosis.
Table 1.
Age and gender differentiation for 116 patients with final diagnoses
| Variable | Non-neoplastic | Benign/Intermediate neoplastic | Malignant neoplastic | p value |
|---|---|---|---|---|
| Gender: Male - n (%) | 34 (51%) | 22 (52.4%) | 4 (33%) | 0.71 |
| Female - n (%) | 33 (49%) | 20 (47.6%) | 2 (67%) | |
| Age – Mean ± SD (range) |
11.1 ± 4.1 (2, 18) |
10.0 ± 5.1 (0.25, 18) |
7.6 ± 5.5 (0.4, 13) |
0.13 |
SD, standard deviation.
Table 2.
Lesion location for 116 paediatric soft tissue masses
| Location | Subdivision | No. of lesions (%) |
|---|---|---|
| Upper limb | Shoulder girdle | 6 (5.2%) |
| Arm | 6 (5.2%) | |
| Elbow | 4 (3.4%) | |
| Forearm | 5 (4.3%) | |
| Wrist/hand | 14 (12.1%) | |
| Lower limb | Pelvis, groin, buttock | 7 (6.0%) |
| Thigh | 11 (9.5%) | |
| Knee | 18 (15.5%) | |
| Calf | 14 (12.1%) | |
| Ankle/foot | 19 (16.4%) | |
| Trunk | Trunk/chest/abdominal wall | 12 (10.3%) |
All tumours arose within the extra articular soft tissues.
Table 3.
Suspected MRI diagnosis for 75 lesions classified into non-neoplastic, neoplastic benign, intermediate-grade neoplastic and neoplastic malignant (ordered by frequency and number in parenthesis)
| Non-neoplastic | Benign | Intermediate | Malignant |
|---|---|---|---|
| Vascular malformation (28) | Tenosynovial giant cell tumour (11) | Lipofibromatosis (1) | Rhabdomysosarcoma (2) |
Trauma (5)
|
Lipoblastoma (4) | Desmoid-type fibromatosis (1) | Malignant peripheral nerve sheath tumour (1) |
| Fat necrosis (3) | Nodular fasciitis (1) | ||
| Foreign body granuloma (1) | Schwannoma (1) | ||
| Myositis ossificans (3) | Atypical lipomatous tumour (1) | ||
| Tumoral calcinosis (3) | Lipoma (1) | ||
Bursitis (2)
|
|||
| Baker’s cyst (1) | |||
| Chronic abscess (1) | |||
| Ganglion cyst (1) | |||
| Lymph node (1) | |||
| Meniscal cyst (1) | |||
| Pyomyositis (1) |
Figure 1.
A 12-year-old girl with a swelling in the volar ulnar-side-of the forearm. (a) Axial T1W FSE MR image demonstrates a lobular intermediate SI mass with peripheral fat (arrow) adjacent to the ulna. (b) Coronal STIR MR image shows a hyperintense lesion with no surrounding reactive change (arrow) containing two small hypointense foci consistent with pleboliths. An imaging diagnosis of vascular malformation was made. (c) Longitudinal ultrasound study shows a lobular hypoechoic mass (arrows) which reduces in size on compression. SI, signal intensity; STIR, short tau inversion recovery; T1W FSE, T1 weighted fast spin echo.
Figure 2.
A 15-year-old girl with a swelling in the dorsal aspect of the forearm. (a) Axial T1W FSE MR image demonstrates a lobular oval intermediate SI mass (arrow) in the extensor compartment musculature. (b) Sagittal T2W FSE MR image shows an elongated intermediate SI lesion with no surrounding reactive change (arrow). An imaging diagnosis of vascular malformation was made. (c) Longitudinal ultrasound study shows a highly vascular mass (arrows) which flattens on compression. SI, signal intensity; T1W FSE, T1 weighted fast spin echo.
IGCNB was undertaken in 34 (29.3%) patients, typically for indeterminate lesions > 2 cm in maximal dimension. Surgical resection was undertaken in 47 (40.5%) patients, either as a primary procedure for small lesions or those with classical imaging appearances (n = 29; 25.0%) or following IGCNB result (n = 18; 15.5%). In total, 61 (52.6%) patients had a histological diagnosis, while a further 3 (2.6%) had a diagnosis based on a combination of MRI and microbiology [abscess (n = 2) and pyomyositis (n = 1)] (Figure 3). In 52 (44.8%) patients, the final diagnosis was based on imaging findings combined with clinical examination/follow-up.
Figure 3.
A 6-year-old girl with pain and swelling in the right groin, referred as a possible STS. (a) Coronal T1W FSE MR image demonstrates an oval intermediate SI mass with a mildly hyperintense rim consistent with a “penumbra” sign (arrow) suggestive of an abscess in the right adductor musculature. (b) Coronal STIR MR image shows the lesion (arrow) to be associated with extensive surrounding muscle oedema (arrowheads). (c) Axial T2W FSE MR image shows the mass with a fluid SI centre (arrow) and muscle oedema (arrowhead). An imaging diagnosis of muscle abscess was made and pyomyositis was confirmed on aspiration. SI, signal intensity; STIR, short tau inversion recovery; STS, soft tissue sarcoma; T1W FSE, T1 weighted fast spin echo.
A non-neoplastic lesion was diagnosed in 67 (57.8%) patients, a benign tumour in 39 (33.6%), an intermediate-grade tumour in 4 (3.4%) and a STS in 6 (5.2%) children. 23 children (19.8%) were referred to a further subspecialty service, most commonly vascular interventional for management of a vascular malformation (n = 18) and occasionally to dermatology. The final diagnoses for 116 children are listed Table 4. The commonest non-neoplastic lesion was a vascular malformation (n = 31; 26.7%) (Figures 1, 2, 4 and 5). The commonest benign lesion was tenosynovial giant cell tumour (tGCT) (n = 10; 8.6%) (Figure 6), followed by nodular fasciitis (Figure 7), lipoma/lipoblastoma and benign nerve sheath tumours (Figure 8). The intermediate-grade lesions were desmoid-type fibromatosis, an angiomatoid fibrous histiocytoma (Figure 9) and lipofibromatosis. The malignant tumours included rhabdomyosarcoma (Figure 10), low-grade spindle cell sarcoma, malignant peripheral nerve sheath tumour, epithelioid haemangioendothelioma (Figure 11), Ewing sarcoma (Figure 12) and a low-grade fibromyxoid sarcoma.
Table 4.
Final diagnoses for 116 lesions classified into non-neoplastic, neoplastic benign, intermediate-grade neoplastic and neoplastic malignant (ordered by frequency and number in parentheses)
| Non-neoplastic | Benign | Intermediate | Malignant |
|---|---|---|---|
| Vascular malformation (31) | Tenosynovial giant cell tumour (10) | Angiomatoid fibrous histiocytoma (1) | Ewing sarcoma (1) |
Trauma-related (4)
|
Lipoblastoma (3) | Lipofibromatosis (1) | Epitheloid haemangioendothelioma (1) |
| Deep granuloma annulare (3) | Lipoma (3) | Desmoid-type fibromatosis (2) | Low-grade fibromyxoid sarcoma (1) |
| Epidermal cyst (3) | Nodular fasciitis (3) | Low-grade spindle cell sarcoma (1) | |
| Fat necrosis (3) | Schwannoma (3) | Malignant peripheral nerve sheath tumour (1) | |
| Foreign body granuloma (3) | Benign fibrous histiocytoma (2) | Rhabdomysosarcoma (1) | |
| Myositis ossificans (3) | Pilomatrixoma (2) | ||
| Tumoral calcinosis (3) | Calcifying aponeurotic fibroma (1) | ||
Bursitis (2)
|
Castleman’s disease (1) | ||
| Baker’s cyst (1) | Enchondroma protuberans (1) | ||
| Bland osteocartilaginous lesion (accessory bone) (1) | Fibroblastic tumour (1) | ||
| Chronic abscess (1) | Fibroma of the tendon sheath (1) | ||
| Fibrous lesion: NEOM (1) | Fibromatosis-inclusion body (1) | ||
| Ganglion cyst (1) | Glomangiopericytoma (1) | ||
| Heterotopic ossification (1) | Leiomyoma (1) | ||
| Lymph node (1) | Myofibroma (1) | ||
| Meniscal cyst (1) | Myxoma (1) | ||
| Necrobiotic type granulomata (1) | Myxopapillary ependymoma (1) | ||
| Indeterminate NEOM (1) | Neurofibroma (1) | ||
| Pyomyositis (1) | Perineurioma (1) | ||
| Rheumatoid nodule (1) |
NEOM, No evidence of malignancy.
Figure 4.
A 3-year-old girl with a painless swelling in the left lower back. (a) Sagittal T1W FSE MR image demonstrates a large lobular intermediate SI mass with peripheral fat (arrow) in the left posterior abdominal wall musculature. (b) Sagittal T2W FSE MR image shows the lesion (arrows) containing multiple hypointense foci consistent with phleboliths (arrowheads). (c) Axial fat suppressed T2W FSE MR image shows the mass (arrow) to contain fluid levels (arrowhead). An imaging diagnosis of slow-flow vascular malformation was made. SI, signal intensity; T1W FSE, T1 weighted fast spin echo.
Figure 5.
A 17-year-old girl presenting with swelling of the medial distal calf. (a) Coronal T1W FSE and (b) axial STIR MR image show an extensive reticular abnormality in the subcutaneous tissues (arrows) consistent with a lymphatic malformation. STIR, short tau inversion recovery; T1W FSE, T1 weighted fast spin echo.
Figure 6.
A 16-year-old girl presenting with a mass in the posterolateral aspect of the left upper forearm. (a) Coronal T1W FSE, (b) sagittal T2W FSE and (c) axial PDW FSE MR images demonstrate a well-defined oval mass (arrows) in the subcutaneous fat adjacent to the radial head. The patient was referred with a possible STS and the MRI findings were considered to be indeterminate. A diagnosis of tenosynovial giant cell tumour was made following IGCNB. IGCNB, imaged-guided core needle biopsy; PDW, proton density-weighted; STIR, short tau inversion recovery; STS, soft tissue sarcoma; T1W FSE, T1 weighted fast spin echo.
Figure 7.
A 14-year-old boy presenting with a mass protruding from the lateral aspect of the left distal thigh. (a) Coronal T1W FSE and (b) axial non-contrast SPAIR MR images demonstrate a well-defined exophytic mass (arrows) in the subcutaneous fat. The patient was referred with a possible STS and the MRI findings were considered to be indeterminate. A diagnosis of nodular fasciitis was made following IGCNB. IGCNB, imaged-guided core needle biopsy; PDW, proton density-weighted; SPAIR, spectral attenuated inversion recovery; STIR, short tau inversion recovery; STS, soft tissue sarcoma; T1W FSE, T1 weighted fast spin echo.
Figure 8.
A 16-year-old boy presenting with a mass in the posteromedial aspect of the left calf. (a) Coronal T1W FSE, (b) sagittal T2W FSE and (c) axial fat suppressed T2W FSE MR images demonstrate a well-defined lobular mass (arrows) in the medial aspect of soleus. The patient was referred with a possible STS, but the MRI findings were considered typical of a peripheral nerve sheath tumour due to the “target” sign on the T2W images. Primary resection was undertaken confirming a diagnosis of neurofibroma. STS, soft tissue sarcoma; T1W FSE, T1 weighted fast spin echo.
Figure 9.
A 12-year-old boy presenting with a possible STS in the anterior left proximal thigh. (a) Sagittal T1W FSE and (b) axial PDW FSE MR images demonstrate a well-defined heterogeneous mass with a profoundly hypointense capsule and evidence of sub acute haemorrhage (arrows) though to be consistent with a sub acute/chronic haematoma. The patient was treated conservatively but the mass grew considerably on follow-up MRI. IGCNB was undertaken, confirming a diagnosis of angiomatoid fibrous histiocytoma. IGCNB, imaged-guided core needle biopsy; PDW, proton density-weighted; STS, soft tissue sarcoma; T1W FSE, T1 weighted fast spin echo.
Figure 10.
A 2-year-old boy presenting with a mass in the posterior right calf. (a) Axial T1W FSE and (b) sagittal STIR MR images demonstrate a lobular mass (arrows) in the posterior calf. Note also popliteal lymphadenopathy (arrowhead-b). (c) Coronal STIR MR image of the pelvis demonstrates additional right common iliac lymphadenopathy (arrow). The child was referred with a possible STS and the MRI findings were considered most likely those of rhabdomyosarcoma, which was confirmed on IGCNB. IGCNB, imaged-guided core needle biopsy; PDW, proton density-weighted; STIR, short tau inversion recovery; STS, soft tissue sarcoma; T1W FSE, T1 weighted fast spin echo.
Figure 11.
A 14-year-old boy presenting with painful swelling of the right calf. (a) Coronal T1W FSE, (b) sagittal T2W FSE and (c) axial non-contrast SPAIR MR images show a poorly defined mass (arrows) in the calf (arrows) with prominent reactive muscle changes (arrowheads). IGCNB confirmed a diagnosis of epithelioid haemangioendothelioma. IGCNB, imaged-guided core needle biopsy; SPAIR, spectral attenuated inversion recovery; T1W FSE, T1 weighted fast spin echo.
Figure 12.
An 11-year-old girl presenting with a small subdermal mass in the anterolateral aspect of the left proximal thigh. (a) Axial T1W FSE and (b) coronal STIR MR images demonstrate an indeterminate mass (arrows) which was considered to be benign due to its size and location. Primary excision was undertaken revealing a diagnosis of Ewing sarcoma. STIR, short tau inversion recovery; T1W FSE, T1 weighted fast spin echo.
Of the 41 (35.3 %) cases that had an indeterminate MRI appearance, 23 (56.1%) underwent IGCNB and 26 (63.4%) underwent surgical resection. 4 (9.8%) children with indeterminate MRI findings were subsequently found to have STS, while 14 (34.1 %) were found to have a non-neoplastic lesion (Table 5).
Table 5.
Final diagnosis for 41 cases with indeterminate MRI imaging findings, classified into non-neoplastic, neoplastic benign, intermediate-grade neoplastic and neoplastic malignant (ordered by frequency and number in parentheses)
| Non-neoplastic | Benign | Intermediate | Malignant |
|---|---|---|---|
| Deep granuloma annulare (3) | Tenosynovial giant cell tumour (3) | Desmoid-type fibromatosis (1) | Ewing sarcoma (1) |
| Vascular malformation (3) | Nodular fasciitis (3) | Epitheloid haemangioendothelioma (1) | |
| Epidermal cyst (2) | Schwannoma (3) | Low-grade fibromyxoid sarcoma (1) | |
| Fibrous lesion: NEOM (1) | Benign fibrous histiocytoma (2) | Low-grade spindle cell sarcoma (1) | |
| Foreign body granuloma (2) | Pilomatrixoma (2) | ||
| Necrobiotic-type granulomata (1) | Calcifying aponeurotic fibroma (1) | ||
| Indeterminate NEOM (1) | Fibroblastic tumour (1) | ||
| Rheumatoid nodule (1) | Fibroma of the tendon sheath (1) | ||
| Glomangiopericytoma (1) | |||
| Leiomyoma (1) | |||
| Myofibroma (1) | |||
| Myxoma (1) | |||
| Myxopapillary ependymoma (1) | |||
| Perineurioma (1) |
NEOM, No evidence of malignancy.
Six patients with indeterminate MRI findings remain without a diagnosis: in two cases, the lump resolved clinically and the other four are awaiting further follow-up.
Of the 30 cases referred from primary care, 19 (63.3%) lesions were non-neoplastic while 11 36.7%) were benign. Of the 86 cases referred from secondary care, 48 (55.8%) lesions were non-neoplastic while 30 (34.9%) were benign, 2 (2.3%) were intermediate-grade and 6 (7.0%) were malignant.
Figures 13 and 14 demonstrate the differences between referral diagnosis and final diagnosis.
Figure 13.
Breakdown of different diagnoses based on clinical concern or imaging findings at the time of referral (59.5% malignant, 25.0% benign and 15.5% non-neoplastic).
Figure 14.
Breakdown of final diagnoses, divided into 5.2% malignant, 3.4% intermediate, 33.6% benign and 57.8% non-neoplastic diagnosis.
Discussion
STS in children are uncommon, with only ~91 new cases/year diagnosed in England in children <15 years of age, representing ~6% of childhood cancers.1 An ultrasound study of 135 superficial soft tissue masses in children identified only 18 malignant lesions, the commonest diagnosis being lymphoma and only 1 of which was STS (rhabdomyosarcoma).19 However, paediatric soft tissue masses are a relatively common presentation in clinical practice and differentiating between non-neoplastic, benign neoplasms or malignant lesions is of paramount importance.
The current study describes the differential diagnosis of 116 consecutive soft tissue masses referred to a specialist musculoskeletal sarcoma unit in patients aged 0–18 years. Of the 69 (59.5%) patients with a referral diagnosis of possible STS, only 6 (8.7%) children were diagnosed with a sarcoma while 4 further children were diagnosed with an intermediate-grade tumour according to the 2020 WHO Classification of Soft Tissue and Bone Tumours.20 Otherwise, 67 (57.8%) were diagnosed with a non-neoplastic lesion and 39 (33.6%) with a benign neoplasm.
Overall, the single commonest diagnosis was a vascular malformation (26.7%), a non-neoplastic lesion according to the ISSVA Classification.21 Based on the clinical and imaging features of vascular malformations, all referred cases were classified as capillary-type, venous-type or less commonly lymphatic-type malformations.15,22,23 Compression ultrasound and the use of Doppler are valuable adjuncts to the diagnosis of a vascular malformation suspected on MRI. Capillary-type vascular malformations may be difficult to identify with ultrasound, particularly if intramuscular. Venous-type vascular malformations appear as anechoic/hypoechoic spaces which are easily compressible by the ultrasound probe, while macrocystic lymphatic malformations appear similar but although deformable by the ultrasound probe do not compress completely.24,25 Compression ultrasound allowed a confident diagnosis of vascular malformation in 26 of 28 cases suspected at MRI. Of the remaining cases for which an MRI diagnosis was suggested, this was consistent with the final diagnosis in 39, including 10 fatty lesions, 6 cases of tGCT, 3 cases of myositis ossificans, 3 cases of tumoral calcinosis and 2 subacute haematomas. In seven patients, the final diagnosis differed from the suggested MRI diagnosis. Five of these had been diagnosed on MRI as tGCT due to very low T2W SI, but represented a variety of non-neoplastic and benign lesions. One child with a large mass in the popliteal fossa was thought to have a rhabdomyosarcoma, but IGCNB yielded a diagnosis of Castleman’s disease. The most significant mis-diagnosis was in a young boy with a thigh mass that had initial imaging features suggestive of a chronic haematoma. However, due to an increase in size of the lesion at clinical follow-up a repeat MRI and IGCNB were undertaken, yielding a diagnosis of angiomatoid fibrous histiocytoma. Of the 75 patients with a suggested MRI diagnosis, only 10 required IGCNB and 2 patients with a suspected infection underwent ultrasound-guided aspiration. 15 patients went straight to primary resection, all bar one having a benign diagnosis. The child with MPNST developing on a background of a plexiform neurofibroma in NF1 was also treated with primary resection due to the strong suspicion of malignant change based on clinical and MRI findings. These results suggest that review of MRI by an experienced musculoskeletal tumour radiologist within the MDT setting is highly accurate for deciding which patients can be treated conservatively, which should undergo IGCNB and which can be safely treated with primary resection. A flow diagram for the possible management of a child with a STM on MRI is presented in Figure 15.
Figure 15.
Proposed flow-chart for the management of a paediatric soft tissue mass seen at MRI. MDT, multidisciplinary-disciplinary team.
Of the 41 children with an indeterminate MRI diagnosis, 25 underwent IGCNB while 13 went to primary resection due to small tumour size. 12 of these 13 cases were either non-neoplastic or benign lesions. Regarding the remaining five children with STS, a diagnosis of rhabdomyosarcoma was suggested based on MRI in a 2-year-old by with a large calf mass and popliteal and iliac lymphadenopathy, this confirmed on IGCNB of the calf lesion. The other four tumours had indeterminate MRI features. One child had a small subdermal mass in the left thigh which was treated with primary resection but turned out to be a Ewing sarcoma. She underwent re-excision of the tumour bed which showed no evidence of residual tumour. The remaining three cases were diagnosed with IGCNB and then underwent surgical resection.
As demonstrated in Figures 13 and 14, there was a large difference between suspected referral diagnosis and final diagnosis. From review of referral letters, it is estimated that ~60% of children were referred with a possible STS. This may be an overestimation, since as explained in the Materials and Methods any child with an indeterminate STM at the referral hospital was assumed to have been referred with STS due to referral to a sarcoma unit. The final diagnosis of STS was made in ~5% of cases which is clearly reassuring, but this is based on radiological assessment/management in a specialist service and should not deter referral of children with indeterminate soft tissue lesions seen in general practice/non-specialist services.
There are several limitations to the study. Due to the fact that children were referred to a specialist sarcoma service, there is very likely to be an inherent bias in our population towards more complex cases and less common diagnoses. Soft tissue lesions that are reliably diagnosed with either clinical or clinical and imaging correlation (e.g. lipomas and ganglion cysts) are less likely to be referred to the unit and will thus be under represented. Since most cases had been imaged prior to referral, there were a large variety of imaging protocols used. Additional techniques such diffusion-weighted imaging and dynamic post-contrast MRI (DCE-MRI) were not employed. Both may improve the ability to differentiate between benign and malignant soft tissue tumours in adults,26 but the use of intravenous contrast is best avoided in children and there is no evidence that the techniques reduce the requirement for IGCNB in indeterminate lesions.
In conclusion, this study has determined that the prevalence of STS in the paediatric population at the point of referral to a specialist musculoskeletal sarcoma unit is low, with ~5% of patients ultimately diagnosed with a malignant tumour. Just over half of patients were determined to have a non-neoplastic abnormality, whilst a third of the referral population were determined to have a benign neoplasm.
Footnotes
Acknowledgements: The authors would like to acknowledge Mr Paul Bassett for statistical input.
Contributor Information
Catriona Reid, Email: catriona.reid7@nhs.net.
Asif Saifuddin, Email: asif.saifuddin@nhs.net.
REFERENCES
- 1.Public Health England . Childhood cancer statistics, England annual report 2018. National cancer registration and analysis service. 2018. Available from: http://www.ncin.org.uk/view?rid=3715.
- 2.Kransdorf MJ. Benign soft-tissue tumors in a large referral population: distribution of specific diagnoses by age, sex, and location. AJR Am J Roentgenol 1995; 164: 395–402. doi: 10.2214/ajr.164.2.7839977 [DOI] [PubMed] [Google Scholar]
- 3.Kransdorf MJ. Malignant soft-tissue tumors in a large referral population: distribution of diagnoses by age, sex, and location. AJR Am J Roentgenol 1995; 164: 129–34. doi: 10.2214/ajr.164.1.7998525 [DOI] [PubMed] [Google Scholar]
- 4.Brisse HJ, Orbach D, Klijanienko J. Soft tissue tumours: imaging strategy. Pediatr Radiol 2010; 40: 1019–28. doi: 10.1007/s00247-010-1592-z [DOI] [PubMed] [Google Scholar]
- 5.Stein-Wexler R. Mr imaging of soft tissue masses in children. Magn Reson Imaging Clin N Am 2009; 17: 489–507. doi: 10.1016/j.mric.2009.03.009 [DOI] [PubMed] [Google Scholar]
- 6.Navarro OM. Soft tissue masses in children. Radiol Clin North Am 2011; 49: 1235–59. doi: 10.1016/j.rcl.2011.07.008 [DOI] [PubMed] [Google Scholar]
- 7.Ahlawat S, Fayad LM. De novo assessment of pediatric musculoskeletal soft tissue tumors: beyond anatomic imaging. Pediatrics 2015; 136: e194–202. doi: 10.1542/peds.2014-2316 [DOI] [PubMed] [Google Scholar]
- 8.Navarro OM. Imaging of benign pediatric soft tissue tumors. Semin Musculoskelet Radiol 2009; 13: 196–209. doi: 10.1055/s-0029-1237689 [DOI] [PubMed] [Google Scholar]
- 9.Thacker MM. Benign soft tissue tumors in children. Orthop Clin North Am 2013; 44: 433–44. doi: 10.1016/j.ocl.2013.05.001 [DOI] [PubMed] [Google Scholar]
- 10.Loeb DM, Thornton K, Shokek O. Pediatric soft tissue sarcomas. Surg Clin North Am 2008; 88: 615–27. doi: 10.1016/j.suc.2008.03.008 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Stein-Wexler R. Pediatric soft tissue sarcomas. Semin Ultrasound CT MR 2011; 32: 470–88. doi: 10.1053/j.sult.2011.03.006 [DOI] [PubMed] [Google Scholar]
- 12.Kapoor G, Das K. Soft tissue sarcomas in children. Indian J Pediatr 2012; 79: 936–42. doi: 10.1007/s12098-011-0560-4 [DOI] [PubMed] [Google Scholar]
- 13.Sheybani EF, Eutsler EP, Navarro OM. Fat-containing soft-tissue masses in children. Pediatr Radiol 2016; 46: 1760–73. doi: 10.1007/s00247-016-3690-z [DOI] [PubMed] [Google Scholar]
- 14.Navarro OM, Laffan EE, Ngan B-Y. Pediatric soft-tissue tumors and pseudo-tumors: MR imaging features with pathologic correlation: Part 1. imaging approach, pseudotumors, vascular lesions, and adipocytic tumors. Radiographics 2009; 29: 887–906. doi: 10.1148/rg.293085168 [DOI] [PubMed] [Google Scholar]
- 15.Tomà P, Esposito F, Granata C, Paolantonio G, Terranova MC, Lo Re G, et al. Up-to-date imaging review of paediatric soft tissue vascular masses, focusing on sonography. Radiol Med 2019; 124: 935–45. doi: 10.1007/s11547-019-01050-8 [DOI] [PubMed] [Google Scholar]
- 16.Laffan EE, Ngan B-Y, Navarro OM. Pediatric soft-tissue tumors and pseudotumors: MR imaging features with pathologic correlation: part 2. tumors of fibroblastic/myofibroblastic, so-called fibrohistiocytic, muscular, lymphomatous, neurogenic, hair matrix, and uncertain origin. Radiographics 2009; 29: e36. doi: 10.1148/rg.e36 [DOI] [PubMed] [Google Scholar]
- 17.Sargar KM, Sheybani EF, Shenoy A, Aranake-Chrisinger J, Khanna G. Pediatric fibroblastic and myofibroblastic tumors: a pictorial review. Radiographics 2016; 36: 1195–214. doi: 10.1148/rg.2016150191 [DOI] [PubMed] [Google Scholar]
- 18.Sambri A, Bianchi G, Cucurnia I, Gambarotti M, Donati DM. Pediatric soft tissue sarcoma of the limbs: clinical outcome of 97 patients. Eur J Orthop Surg Traumatol 2018; 28: 1–7. doi: 10.1007/s00590-017-2019-4 [DOI] [PubMed] [Google Scholar]
- 19.Yamaguchi M, Takeuchi S, Matsuo S. Ultrasonic evaluation of pediatric superficial masses. J Clin Ultrasound 1987; 15: 107–13. doi: 10.1002/jcu.1870150204 [DOI] [PubMed] [Google Scholar]
- 20.WHO Classification of Tumours Editorial Board Soft tissue and bone tumours. Lyon (France). International agency for research on Cancer; (WHO classification of tumours series. 5th edition Volume 3 Lyon: IARC; 2020. https://publications.iarc.fr/588. [Google Scholar]
- 21.20th ISSVA Workshop. International society for the study of vascular anomalies classification for vascular anomalies. 2018. Available from: https://www.issva.org/UserFiles/file/ISSVA-Classification-2018.pdf
- 22.Steiner JE, Drolet BA. Classification of vascular anomalies: an update. Semin Intervent Radiol 2017; 34: 225–32. doi: 10.1055/s-0037-1604295 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Madani H, Farrant J, Chhaya N, Anwar I, Marmery H, Platts A, et al. Peripheral limb vascular malformations: an update of appropriate imaging and treatment options of a challenging condition. Br J Radiol 2015; 88: 20140406. doi: 10.1259/bjr.20140406 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.Behravesh S, Yakes W, Gupta N, Naidu S, Chong BW, Khademhosseini A, et al. Venous malformations: clinical diagnosis and treatment. Cardiovasc Diagn Ther 2016; 6: 557–69. doi: 10.21037/cdt.2016.11.10 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25.Esposito F, Ferrara D, Di Serafino M, Diplomatico M, Vezzali N, Giugliano AM, et al. Classification and ultrasound findings of vascular anomalies in pediatric age: the essential. J Ultrasound 2019; 22: 13–25. doi: 10.1007/s40477-018-0342-1 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26.Choi YJ, Lee IS, Song YS, Kim JI, Choi K-U, Song JW. Diagnostic performance of diffusion-weighted (DWI) and dynamic contrast-enhanced (DCE) MRI for the differentiation of benign from malignant soft-tissue tumors. J Magn Reson Imaging 2019; 50: 798–809. doi: 10.1002/jmri.26607 [DOI] [PubMed] [Google Scholar]