Imaging of Pediatric Extremity Soft Tissue Tumors: A COG Diagnostic Imaging Committee/SPR Oncology Committee White Paper

Abstract

Pediatric soft tissue tumors of the extremity include rhabdomyosarcoma and non-rhabdomyosarcoma neoplasms. This manuscript provides consensus-based imaging recommendations for imaging evaluation at diagnosis, during treatment, and following completion of therapy for patients with a soft tissue tumor of the extremity.

Introduction

Pediatric soft tissue tumors of the extremity include rhabdomyosarcoma (RMS) and non-rhabdomyosarcoma soft tissue sarcomas (NRSTS). The 2020 World Health Organization classification further subdivides NRSTS into 140 distinct entities according to their resemblance to mature non-neoplastic tissue.¹ In children, RMS is the most common soft tissue neoplasm and the third most frequent extracranial solid tumor with approximately 400 cases diagnosed each year in the United States. Most soft tissue sarcomas present before age 9 with extremity tumors predominantly observed in adolescent males.² Although typically sporadic, association between RMS and cancer predisposition syndromes including Li-Fraumeni, Costello, DICER1 and neurofibromatosis type 1 have been documented.³

Extremity soft tissue sarcomas typically present as an enlarging, non-tender mass. Compression of adjacent structures may cause limb swelling or neuropathy. Given their propensity to spread along facial planes, tumors may be extensive at presentation and show early dissemination to regional lymph nodes.⁴ Patients are often initially evaluated with ultrasound, although diagnosis is ultimately made by excisional or percutaneous biopsy. There are two major histologic subtypes of RMS, embryonal and alveolar, and in extremity tumors, the latter is most common and can be associated with a chromosomal translocation resulting in the fusion of PAX3-7 and FOXO1 genes. The presence of fusion-positive alveolar RMS may predict more aggressive clinical behavior and is important for risk stratification in ongoing clinical trials.⁵ Spindle cell/sclerosing and pleiomorphic RMS rarely manifest in the limbs. In addition to the standard laboratory tests and physical examination, imaging plays a vital role in diagnosis, during treatment, and following completion of therapy (Table 1).

Table 1:

Advantages and disadvantages of each modality for the evaluation of rhabdomyosarcoma and non-rhabdomyosarcoma soft tissue sarcomas of the extremity.

Modality	Timepoints	Advantages	Disadvantages
Ultrasound	Screening	Characterize the mass as cystic, solid, and/or vascular. Assess relationship to vascular structures. Sedation is not needed	Low specificity for malignancy Large or deep lesions may not be well characterized No data to support use during response or surveillance timepoints
MRI	Diagnosis, follow-up, surveillance	No ionizing radiation Superior soft tissue resolution	Size response may not predict outcome Sedation/anesthesia may be needed for young patients Role in surveillance is not yet validated
Chest CT	Staging, follow-up, surveillance	High sensitivity to detect pulmonary metastases Rapid image acquisition	Exposure to ionizing radiation Not used to assess primary tumor
PET/CT or PET/MRI	Staging, follow-up, surveillance	Whole-body imaging High sensitivity for non-pulmonary metastases May identify occult lesions Has replaced bone scintigraphy	Exposure to ionizing radiation Low sensitivity for pulmonary metastases May not detect regional lymph node involvement if borderline Role in response assessment and surveillance are not yet validated

Imaging at Diagnosis

Ultrasound is recommended as the initial imaging modality for a child with a soft tissue mass in the extremity (Grade D, SOR 1.81, strong recommendation).⁶

Primary Tumor

Magnetic resonance imaging (MRI) with intravenous (IV) gadolinium is the modality of choice to characterize and locally stage soft-tissue tumors (Grade B, SOR 1.09, very strong recommendation).^6-9 Table 2 outlines recommended imaging parameters for MRI. Initial imaging for both suspected RMS and NRSTS is similar given the overlap in both clinical and radiological features at presentation. For extremity lesions, the field-of-view should include the entire local-regional nodal basin to assess for draining nodal involvement, which assists in surgical and radiation planning and in prognostication. To date, imaging is unable to differentiate between embryonal and alveolar RMS or fusion status. Computed tomography (CT) of the extremity is not recommended for initial evaluation unless as a supplement to assess osseous invasion, for surgical planning, or if the patient has a contraindication to MRI.

Table 2:

Suggested MRI protocol for evaluation of a soft tissue tumor in the extremity at diagnosis and during follow-up. STIR=short tau inversion recovery FS=fat-suppressed, TSE=turbo spin echo

Plane	Sequence	Contrast	Field of View	Required
Coronal	STIR	Pre-Contrast	Field-of-view should include the draining nodal basin. If tumor is in the lower extremity, include both limbs on the pre-contrast coronal T1 and T2 sequences.	Yes
Coronal	T1 TSE
Sagittal	T2 FS
Axial	T2 FS
Axial	T1 TSE
Axial	T1 FS
Axial	T1 FS	Post-contrast
Coronal	T1 FS
Axial	Diffusion			Optional

Metastatic work-up:

Chest CT should be performed in all patients at diagnosis to assess for pulmonary metastases (Grade B, SOR 1.36, very strong recommendation).^8-10 Table 3 outlines recommended imaging parameters for chest CT. The lungs are the most frequent site of metastatic disease and, therefore, a CT is required at the time of diagnosis to assess the pulmonary parenchyma. The sensitivity for CT in the detection of pulmonary metastases exceeds 90%.¹¹ This examination may be acquired either separately or as part of a PET/CT study, if the latter is optimized to be of diagnostic quality.¹² A non-contrast CT chest is sufficient for detection of pulmonary metastases; however, IV contrast is indicated if mediastinal, chest wall, lymph node, or great vessel involvement is suspected based on clinical or prior imaging findings.¹⁰ Potential diagnostic criteria for defining pulmonary metastases and indeterminate lung nodules have been proposed.¹³ For tumors confined to their site of origin (T1), which have only a 0.5% risk of metastasis, it may be reasonable to evaluate the chest with radiographs alone. Many clinical trials; however, require chest CT for enrollment.¹⁴

Table 3:

Suggest chest CT protocol for evaluation of metastatic disease at diagnosis and during follow-up

Coverage	Slice thickness	Contrast phase	Reformat planes/types/reconstruction kernel
Chest	≤5 mm. Thin slices not required.	Optional	Coronal and sagittal reformat planes Axial lung kernel Maximum intensity projections are recommended

2-deoxy-2-[F-18]fluoro-deoxyglucose (FDG) positron emission tomography PET/CT or PET/MRI is recommended in the initial evaluation of RMS and NRMST (Grade B, SOR 1.36, very strong recommendation).^15-17 Table 4 outlines recommended imaging parameters for PET. Primary RMS tumors are generally FDG-avid, and the standard uptake value (SUV) has been shown to have prognostic utility.¹⁸ In addition, whole-body FDG-PET may identify occult lesions, not detected by conventional imaging, and has been shown to change management in 18% patients with RMS.¹⁷

Table 4:

Suggested FDG-PET protocol for the evaluation of metastatic disease at diagnosis and during follow-up. NPO = nothing per os.

Patient Preparation	Radiopharmaceutical	Dose Range	Time to imaging	Field of view
NPO and no intravenous dextrose 4-6 hours prior. Blood glucose < 150 -200 mg/dL Brown fat warming protocol +/or medication per local practice	2-deoxy-2-[F-18]fluoro-deoxyglucose	0.1 mCi/kg (min 1 mCi, max 10 mCi)	60-minute uptake phase	Whole body (vertex to toes)

Regional nodal involvement is more common in RMS presenting in the limbs than elsewhere and is an independent prognostic factor for patients with alveolar RMS.¹⁹ Lymph node evaluation is therefore critically important and may prompt treatment with radiation therapy. Clinically involved lymph nodes are defined as those that are palpable, are greater than 1.5 cm in short axis on CT or MRI or are ¹⁸F-FDG-avid. If lymph nodes are borderline in terms of size or degree of metabolic activity, FDG-PET alone is neither highly specific nor sensitive for metastatic involvement. Therefore, for non-clinically involved nodes, lymph node sampling either by excisional or percutaneous core needle biopsy is strongly recommended.^20,21 Fine needle aspiration should not be performed due to the potential for false negatives. Sentinel nodal evaluation may be performed using either methylene blue injection or technetium-99m sulfur colloid at the site of primary tumor and is preferred over random lymph node sampling.²⁰ The role of sentinel lymph node biopsy in NRSTS, however, is less certain.²²

Approximately 10-20% of all patients with RMS present with metastatic disease to other sites and alveolar RMS, in particular, has a propensity to metastasize to unusual locations including the pancreas and soft tissues.^23,24 PET/CT has a reported 95-100% sensitivity and 80-100% specificity for detection of non-pulmonary distant metastases.^12,25 Due to the superiority in detecting cortical and marrow lesions, FDG-PET has supplanted technetium-99m bone scintigraphy in evaluating skeletal metastases.^12,14

The sensitivity of FDG-PET for the identification of lung metastases is lower and many small nodules may be undetected due to the limited spatial resolution of PET.^26-28 Therefore, FDG-PET should be interpreted with a diagnostic quality CT obtained either simultaneously with the PET or separately as a dedicated chest CT.²⁹ Data on PET/MR continue to emerge and advantages over PET/CT include lower radiation doses and superior delineation of local disease, thereby potentially providing a more comprehensive assessment in a single imaging session.^12,27,30,31

Tumor Staging

Staging of pediatric RMS is performed according to the pre-treatment site-modified tumor-node-metastasis (TNM) staging system used by the Soft Tissue Sarcoma Committee of the Children's Oncology Group (Grade A, SOR 1.55, strong recommendation). ^14,32,33 The TNM system sub-divides patients based on favorable and unfavorable sites (all primary extremity sites are unfavorable), presence or absence of invasion into surrounding structures, tumor size > or ≤5 cm in longest dimension, and regional lymph node status. The TNM stage is then combined with the surgical and histopathologic grouping system, known as Clinical Group, to assign patients as low, intermediate, or high risk of treatment failure.

Staging of pediatric NRMST is performed according to the TNM staging system (Grade C, SOR 2.0, moderate recommendation). While NRSTS represent a diverse group of tumor types, trials specifically launched to evaluate them, such as Children’s Oncology Group ARST1321 and ARST0332, have used both a streamlined risk and stage assignment, defined according to the presence of metastatic disease, primary tumor size, extent of the surgical resection, and histologic grade.³⁴ Characteristics such as primary tumor size >5 cm or the presence of metastasis are strongly associated with a poor 5-year event free survival and decreased overall survival. Like RMS, NRSTS tumors are also grouped into low, intermediate, and high risk.

Follow-up Imaging

For both RMS and NRMST, MRI with IV gadolinium of the affected extremity is recommended to assess tumor response to treatment (Grade B, SOR 1.45, very strong recommendation).^35,36 MRI should be performed using similar technical parameters and the same field-of-view as at the time of diagnosis. If regional nodes were not involved at diagnosis, focus can be paid to the primary tumor alone.¹² If pulmonary metastases were present, a chest CT should be acquired to evaluate the response to therapy (Grade C, SOR 1.36, very strong recommendation). As previously stated, IV contrast is recommended only if mediastinal nodes, chest wall, or vascular involvement were present. For patients with RMS who had indeterminate lung nodules at diagnosis, these may not obligate follow-up; however, institutional and protocol requirements may vary.^24,37 Follow-up FDG-PET imaging is unnecessary for children without baseline metastatic disease. For those with metastases at presentation, the role of FDG-PET to assess response to therapy is not yet validated for prognostication. If performed, imaging parameters should be identical to those used at diagnosis to facilitate an accurate comparison.

Early response assessment can be performed after 2-3 cycles of chemotherapy, although timing will vary based on clinical trial protocols. Disease reassessment is also recommended before local control, at the end of induction chemotherapy, after maintenance therapy, and at the completion of treatment.¹²

Tumor Response Criteria

We recommend using 1-D measurements per RECIST 1.1 criteria for the primary tumor and metastatic disease response assessment (Grade C, SOR 1.91, strong recommendation). ^34,38-41 Tumor length can be measured several ways, most often by choosing the longest diameter in any plane as per the Response Assessment Criteria in Solid Tumors (RECIST) version 1.1.⁴² Often, pediatric soft tissue sarcomas are irregular in shape and exhibit a non-uniform response to therapy, which has raised concerns regarding the accuracy of a unidimensional measurement and possible underestimation of tumor response.³⁵ To overcome this limitation, some investigators use 3-D volumetric measurements of the primary tumor, though, this too has not been a reliable marker of prognosis.⁴¹ Studies comparing the two methods of measurement are limited by significant interobserver variability, which has led to a lack of standardization in assessment.^38,39 The method of tumor measurement may ultimately depend on treatment protocols or institutional practice patterns. It is important that the method and technique for measurements stay consistent throughout all imaging time points.

For patients with intermediate or high-risk RMS, change in primary tumor size, whether assessed by 1-D or 3-D measurements, is not a reliable predictor of patient outcome.^35-37 This is particularly true in the current era of targeted immunotherapy and local control options from interventional radiology. Studies have also shown that metabolic response is not a predictor of prognosis.⁴³ Therefore, in the absence of progression, other methods, such as changes in tumor vascularity, or cellularity may provide better methods of assessing response; however, clinical trials are needed to validate such criteria.

Imaging Off Therapy/Surveillance

The timeline for surveillance imaging may vary based on clinical trial guidelines or insurance payer reimbursement, but generally should occur every 3 months for the first year, every 4 months for the second and third year, and then every 6 months for the fourth year. Surveillance imaging includes an MRI of the primary tumor site and a chest CT if pulmonary metastases were present (Grade C, SOR 1.45, very strong recommendation).¹⁰ Given concerns regarding the use of gadolinium in children, IV contrast may not be necessary for post-operative surveillance in the absence of a T2 hyperintense mass.⁸ If there is documented stability on chest CT, typically over 2 years, then a chest radiograph can be considered for surveillance. Similar to response assessment, the role of screening FDG-PET after the completion of therapy has not been established.⁴⁴

Although off treatment surveillance for pediatric RMS remains standard of care, it is not without controversy.^44-46 Some studies have found no difference in overall survival between patients with relapse diagnosed by imaging and those identified clinically.^44,45 Hence, the role of routine imaging surveillance in these patients should be further investigated.

Future Quantitative Imaging Techniques

As soft-tissue sarcomas generally show restricted diffusion, reflecting increased cellular density, diffusion-weighted imaging may be a useful marker in the assessment of tumor response.^7,47 Some authors have found that the percent of tumor volume that exhibits both restricted diffusion on MRI and hypermetabolic activity on FDG-PET compared to whole tumor volume at time of diagnosis in children and young adults with RMS can be used as a prognostic biomarker for event-free survival, independent of RECIST-1.1-based response assessments or ADC values alone.⁴⁸

Dynamic contrast-enhanced (DCE) MRI is currently the preferred MRI-based method to assess perfusion within tumors for the purpose of identifying regions of necrosis, viability, granulation tissue, or fibrosis. DCE-MRI relies on acquiring images after injection of intravenous contrast material using a time-resolved MR angiography technique.⁷ Some authors have proposed the use of quantitative DCE-MRI to better reflect tumor biology and have shown its feasibility for early prediction of pathological response of soft-tissue sarcomas to neoadjuvant therapy.^49,50

Finally, radiomics analysis, which refers to the conversion and extraction of quantitative data contained in medical digital images with the purpose of creating models that can predict outcomes, can be done with a variety of imaging modalities.^51,52 Integrating demographic, clinical, genomic, and comorbidity data can expand this analysis. Of particular interest is the study by Banerjee et al that specifically investigated RMS in children and young adults up to 20 years of age.⁵³ They described a radiomics framework that enabled differentiation of the embryonal variant from the alveolar subtype by assessing the fusion of diffusion-weighted imaging and contrast-enhanced T1-weighted images.

Abbreviations

RMS

Rhabdomyosarcoma

NRSTS

Non-rhabdomyosarcoma soft tissue sarcomas

MRI

Magnetic resonance imaging

IV

Intravenous

CT

Computed tomography

FDG

2-deoxy-2-[F-18]fluoro-deoxyglucose

SUV

Standard uptake value

RECIST

Response Assessment Criteria in Solid Tumors

DCE

Dynamic contrast-enhanced