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

Abstract

This article provides imaging recommendations for pediatric abdominal tumors that arise outside of the solid viscera. These tumors are rare in children and have been categorized in two groups: abdominal wall and peritoneal tumors (desmoid tumor and desmoplastic small round cell tumor) and tumors that arise from the gastrointestinal tract (gastrointestinal stromal tumor and gastrointestinal neuroendocrine tumor). Authors offer consensus recommendations for imaging assessment of these tumors at diagnosis, during follow-up, and when off-therapy.

1. Introduction

This article provides imaging recommendations for pediatric abdominal tumors that arise outside of the solid organs. Specifically, it covers abdominal desmoid tumor (DT) and desmoplastic small round cell tumor (DSRCT), both of which are mesenchymal tumors arising from the peritoneum or abdominal wall; and gastrointestinal neuroendocrine tumors (GI NET; “carcinoid tumor”) and gastrointestinal stromal tumor (GIST), both tumors that can arise from the gastrointestinal tract. This manuscript was funded in part by the National Clinical Trials Network Operations Center Grant U10CA180886. The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health

DT is an infiltrative, locally invasive fibroblastic mass that can present at any age, as firm, painful lesions prone to recurrence. DT is classified by location as abdominal wall, intraabdominal, or extraabdominal. The majority of pediatric cases are extraabdominal, but intraabdominal (e.g., mesentery or bowel wall,) has the highest mortality. The incidence in the overall population is about 3 per 1,000,000. DT has an association with familial adenomatous polyposis (FAP) and intra-abdominal DT is part of Gardner syndrome¹.

DSRCT is an aggressive sarcoma which originates from the peritoneum and carries a poor prognosis. It most commonly presents in the second or third decades with abdominal pain, emesis, and mass effect. DSRCT has a male predominance and incidence of about 0.3 per 1,000,000^2,3.

The incidence of GI NET in children is estimated to be 3 per 1,000,000⁴. GI NET comprise about one-third of all pediatric NET. Most of pediatric GI NET originate in the appendix, but they can less commonly arise from the stomach, liver, pancreas, or as unknown primary^5,6. While appendiceal NET are often discovered incidentally, low-grade, and considered cured by appendectomy alone, extra-appendiceal GI NET in children tend to have a higher grade and higher likelihood of metastasis (to liver or bone), recurrence, and carcinoid syndrome (flushing, labile blood pressure, diarrhea, and abdominal pain)^6–8.

GIST is the commonest sarcoma of the GI tract, but in children it has a reported incidence of only 0.02 per 1,000,000, with a three-fold higher incidence in girls. In contrast to adult GIST, pediatric GIST is more frequently associated with syndromes (neurofibromatosis type 1 (NF1); Carney-Stratakis syndrome; Carney triad of GIST, pulmonary chondroma and paraganglioma)^9,10. Also, pediatric GIST are often biologically different; pediatric GIST are typically “wild-type,” lacking mutations in c-kit or PDGFRA, as well as “SDH-deficient,” with functional loss of succinate dehydrogenase complex¹¹. In children, GIST most often originate in the stomach, and frequently present with anemia^10,12.

Management of the above cancers is challenging due to their rarity and may influenced by patient-specific factors (disease burden, histology, associated syndromes). Therefore, the approach for imaging can differ from more common abdominal tumors.

2. Staging

Given overall rarity of these tumors, no specific staging systems have been fully validated for DT, DSRCT, pediatric NET, or pediatric GIST, though some have been proposed, as detailed below.

Desmoid tumor

The main therapy for DT is often localized total resection. However, for DT in association with FAP, mesenteric DT is associated with higher post-operative morbidity, and therapy relies on expectant management. A prognostic staging system based on size, growth, and symptoms has been proposed^13,14.

Desmoplastic small round cell tumor

For DSRCT, a proposed staging system incorporates intraoperative and imaging assessment of peritoneal tumor burden, and imaging evaluation for liver and extra-abdominal metastases¹⁴. Metastasis to the liver or chest is common, and in cases of widespread disease, metastasis to head, neck, or bone can be seen^3,15.

Gastrointestinal neuroendocrine tumor

The tumor-node-metastasis (TNM) staging system by the American Joint Committee on Cancer has been validated to determine disease-specific survival for NET in adults^16,17. TNM criteria for NET depend on site of origin, with differing “tumor” criteria for small bowel, appendiceal, and colorectal sites in adults¹⁸.

The TNM staging system could be considered for pediatric extra-appendiceal GI NET; for appendiceal NET in children, however, staging is non-contributory to management unless symptomatically indicated ^6,19 (GRADE B; SOR 2.20). Prognosis for pediatric appendiceal NET after resection has been very favorable in observational studies and meta-analyses, regardless of common prognostic markers (tumor size, nodal or mesenteric involvement)^6,19,20.

Gastrointestinal stromal tumor

The TNM staging classification correlates with disease progression for adult GIST²¹ and might be considered for use in pediatric non-“wild-type” GIST but has not been validated for pediatric, familial or syndromic GIST^6,21. GIST in children is commonly “wild-type,” tends to be indolent with a longer course, and prognostic factors continue to be studied for this group^22,23. While no formal staging system exists, imaging assessment of primary disease and metastatic extent (most frequently liver, lymph nodes, peritoneum) is useful for surgical planning, medical therapy decisions, and ongoing surveillance^10,22.

Authors recommend that management of DT, DSRCT, and GIST be based on patient-specific factors such as tumor burden, histology, metastases, and associated syndromes (GRADE D; SOR 1.60).

3. Imaging modalities for diagnosis

Desmoid tumor

CT is often the first imaging modality for DT, due to technical ease, and it is particularly useful for assessing urgent complications related to an intraabdominal DT²⁴. MRI has superior contrast resolution, valuable for characterizing tumor content and the relationship to surrounding tissues. Varying degrees of fibrosis from fibroblast proliferation in DT results in T2 shortening (intermediate to low signal on T2-weighted images), whereas T2 prolongation is potentially associated with faster DT growth^13,25. MRI depiction of infiltrative growth patterns of DT can assist in surgical planning¹³. If accessible, MRI is the study of choice for initial assessment of DT (GRADE B; SOR 1.60). However, bowel and respiratory motion artifact may degrade evaluation of intraabdominal mesenteric DT, and CT may be preferred in such cases¹³.

18-fluoro-deoxyglucose positron-emission tomography (¹⁸FDG-PET) has limited utility for DT assessment (GRADE B; SOR 1.80). Although DT can be multifocal within a region, they do not metastasize, so whole-body imaging is not necessary. Furthermore, DT are not very metabolically active (maximum standardized uptake value, SUV, ≤4.8)¹³.

Desmoplastic small round cell tumor

At diagnosis, DSRCT presents on CT as multiple or confluent peritoneal masses, heterogeneously enhancing and centrally hypoattenuating, commonly with ascites, diffuse peritoneal thickening and visceral metastases^15,26,27. Most of the published imaging experience of this rare tumor has been with CT^3,14,15,28. It should be noted that MRI provides superior contrast resolution and diffusion-weighted imaging (DWI), which can increase conspicuity of peritoneal and abdominal visceral lesions^14,15,28. Authors recommend that MRI, where available, be the preferred initial modality for DSRCT, with contrast-enhanced CT of the chest to evaluate for thoracic metastases (GRADE B; SOR 2.40).

¹⁸FDG PET has been increasingly used for DSRCT, as its high FDG avidity can aid detection of disease extent^3,15,29,30, and is recommended as the functional imaging modality to assess DSRCT burden at diagnosis (GRADE B; SOR 1.40).

Gastrointestinal neuroendocrine tumor

CT of the abdomen and pelvis has shown high sensitivity for GI NET and allows for assessment of bowel and mesentery^4,18,19. As pediatric appendiceal NET is considered cured by appendectomy, no imaging is indicated beyond initial preoperative detection in the otherwise asymptomatic patient^6,19,20.

For extra-appendiceal GI NET, anatomic imaging can be performed by CT or MRI with protocols targeted to the suspected primary site (e.g., pancreas, liver, bowel)¹⁹. For suspected small bowel GI NET, CT or MRI performed with enterography technique should be considered, with negative enteric contrast and in arterial and portal venous phases (Table 2 and 3)¹⁸. The enterography technique aids in detection of small bowel tumor, and its multi-phase nature also increases sensitivity for liver metastases, as they may be occult on portal venous phase alone¹⁹. MRI (Table 2), while more vulnerable to motion (and therefore less sensitive for small mesenteric tumors), is likely superior to CT in detecting metastases, such as in liver and marrow, so is the favored anatomic imaging modality for extra-appendiceal pediatric GI NET^18,19 (GRADE B; SOR 2.20).

Table 2.

Parameters for MRI of the abdomen/pelvis for abdominal tumor assessment.

Plane(s)	Sequence	Contrast phase	Coverage	Required/optional?	Comment
Axial, coronal, sagittal	T2-weighed fast spin echo	N/A	Abdomen and pelvis	Required	Use breath-hold or respiratory trigger
Axial	T2-weighted with fat saturation, radial sampling	N/A	Abdomen and pelvis	Required	Robust for patient motion in X&Y planes; Use respiratory trigger.
Axial	Diffusion-weighted imaging	N/A	Abdomen and pelvis	Required	b values 600-1000 s/mm2;Ref: 40
Axial	T1-weighted gradient echo chemical shift imaging	N/A	Abdomen and pelvis	Optional	Used to assess intra-cellular lipid
Coronal	Steady-state free precession	N/A	Abdomen and pelvis	Optional	For bowel tumors (enterography)
Axial (or coronal)	T1-weighted gradient echo with fat saturation	Pre-contrast, arterial (~10 s), portal venous (~70 s), equilibrium (~120 s)	Abdomen and pelvis	Required	Need breath-hold and power injection; post-minus-pre-contrast subtraction is useful41. Slice thickness no less than 5 mm, with little to no gap, preferred for staging. For enterography, consider IV glucagon (motion artifact may impact diagnostic utility more than in inflammatory bowel disease).

For tumors arising from the gastrointestinal tract, administration of enterography contrast by mouth or nasogastric tube is needed, 20 mL/kg given in two parts 1 hour prior to scan. Glucagon dose is 0.5 mg if weight is <25 kg and 1 mg if weight is >25 kg, given over 5 minutes before post-contrast sequence.

N/A: not applicable; ~: approximately; s: seconds.

Table 3.

Parameters for whole-body MRI, as described for surveillance of desmoplastic small round cell tumor.

Plane(s)	Sequence	Contrast phase	Coverage	Required/optional?	Comment
Coronal	Short-tau inversion recovery	N/A	Head to toe	Required
Coronal	T1-weighted fast spin echo	N/A	Head to toe	Required
Axial	T2-weighted with fat saturation	N/A	Chest, abdomen, and pelvis	Required	If anomaly is identified, additional sequences may be necessary for characterization
Axial	Diffusion-weighted imaging	N/A	Chest, abdomen, and pelvis	Optional	b values 600-1000 s/mm2;Ref: 40

N/A: not applicable; s: seconds

Functional imaging with radiotracers that target somatostatin receptors is a mainstay for diagnosis and staging/restaging of extra-appendiceal GI NET^8,19. ¹¹¹In-octreotide imaging (with whole-body planar imaging and single-photon-emission computed tomography (SPECT)) has been a primary modality for decades. Its sensitivity for GI NET of 80-100% is higher than for NET from other sites (e.g., pancreas)^8,31, although its sensitivity for marrow metastases is relatively low (60%) and may require supplemental imaging¹⁹. Recently, however, somatostatin receptor PET imaging, with ⁶⁸Ga-DOTA-TATE (tetraxetan(0), tyrosine(3)-octreotide), an octreotide analog with higher affinity for somatostatin receptors, has shown substantially greater sensitivity for primary and metastatic NET, with a 2-3 fold higher spatial resolution than octreotide scanning^4,32–34. Somatostatin receptor PET is likely also lower in radiation dose^35–37. We therefore recommend ⁶⁸Ga-DOTA-TATE or DOTA-TOC whole-body PET as the functional imaging modality for extra-appendiceal GI NET where available or via referral if possible³⁷ (GRADE B; SOR 2.33). ¹⁸FDG PET, while less sensitive for well-differentiated and lower grade NET, may be appropriate for NET that is moderately or poorly differentiated or that is not somatostatin receptor-avid^8,19,31. ¹²³I-meta-iodobenzylguanidine (MIBG) has lower sensitivity for GI NET than octreotide scan or somatostatin receptor PET but may be considered on a case-specific basis if other functional imaging is negative or if ¹³¹I-MIBG therapy is planned^8,18,19,31.

Gastrointestinal stromal tumor

CT is the primary modality for diagnosis of GIST in adults, but given expected long-term need for multiple scans, MRI is favored in children to reduce radiation exposure when undergoing primary tumor characterization and metastatic evaluation^10,38 (GRADE B; SOR 2.00). Enterography technique can be applied to CT or MRI if small bowel involvement is expected, such as in NF1^10,11,39. ¹⁸FDG PET should also be considered at diagnosis to assess extent of disease and metastases^10,38 (GRADE B; SOR 1.40), to help determine need for systemic therapy^10,22. ¹⁸FDG-PET may also show associated masses of NF1, Carney-Stratakis syndrome or the Carney triad¹⁰.

Advantages and disadvantages of imaging modalities for diagnosis, and for follow-up and surveillance, discussed herein are presented in Table 1, and imaging parameters for various modalities are presented in Tables 2–5.

Table 1.

Advantages and disadvantages of modalities used for imaging pediatric abdominal tumors.

Imaging modality	Timepoints	Advantages	Disadvantages
CT abdomen/pelvis with IV and oral contrast	Diagnosis; follow-up; surveillance	Accessible; fast; less susceptible to motion; no sedation needed	Lower contrast resolution and fewer biomarker possibilities than MRI; uses ionizing radiation
MRI abdomen/pelvis with-and-without contrast	Diagnosis; follow-up; surveillance	Superior contrast resolution with potential for biomarkers; assess for distant metastases	Need for sedation at younger developmental ages; more susceptible to motion; less accessible than CT in many pediatric centers
18FDG PET-CT/MRI┼	Diagnosis; follow-up	Assesses tumor functionality and extent	Longer scan time; needs sedation at younger developmental age; uses ionizing radiation; limited utility in tumors with low FDG avidity (DT, well-differentiated NET)
CT chest with IV contrast	Diagnosis; follow-up	Accessible; fast; superior tissue contrast	Uses ionizing radiation. For routine follow-up, may forego IV contrast.
111In-octreotide SPECT-CT	Diagnosis	Good affinity for somatostatin receptors; wider availability than PET	Longer duration and lower sensitivity than 68Ga-somatostatin receptor PET
68Ga-DOTA-TATE/-DOTA-TOC PET-CT/MRI┼	Diagnosis	Strong affinity and specificity for NET	Limited experience at many pediatric centers

^*for surveillance of desmoplastic small round cell tumor.

^┼PET-MRI is not widely available.

CT: computed tomography; MRI: magnetic resonance imaging; NET: neuroendocrine tumor; SPECT: single photon-emission computed tomography; ¹⁸FDG, 18-fluoro-deoxygluocse; ⁶⁸Ga-DOTA-TATE: 68-gallium-tetraxetan(0), tyrosine(3)-octreotide; DOTA-TOC: tetraxetan(0), phenylaniline(1), tyrosine(2)-octreotide

Table 5.

Parameters for functional imaging for abdominal tumor assessment.

Study	Patient preparation	Radio-pharmaceutical	Dose range	Delivery	Time from dose to imaging	Image acquisition	Comment
111In-octreotide planar imaging, with abdomen/pelvis SPECT-CT	NPO overnight; Oral laxative the day before. Discuss holding somatostatin analogues with oncologist35.	111In–DTPA-octreotide	5 MBq/kg	IV	4 hours and 24 hours. A 48 hour scan may be needed if there is substantial bowel activity.	Whole body anterior and posterior planar images, abdomen/pelvis SPECT-CT	Lateral planar image of head-neck may improve detection of cervical lymph node metastases. Fusion to CT is recommended. Consider IV contrast for localization CT to facilitate tumor and vessel depiction.
18FDG PET-CT/MRI┼	NPO and no IV dextrose 4-6 h prior; or exercise 48 hours prior Blood glucose <150 −200 mg/dL Brown fat warming protocol per local practice	18FDG	3.7 MBq/kg, min 37 Mbq and max 370 MBq	IV	1 hour	Whole body, 1 hour duration	Consider β blockers or fentanyl in young children to suppress uptake by brown fat. Consider IV contrast for localization CT to facilitate tumor and vessel depiction.
68Ga-DOTA-TATE/DOTA-TOC PET-CT/MRI┼	Oral laxative the day before. Use of somatostatin analogs needs to be assessed. * Void before image acquisition.	68Ga-DOTA-TATE	2 MBq/kg, max 200 MBq	IV	1 hour	Whole body, 1 hour duration	IV furosemide may be given. Consider IV contrast for the localization CT to facilitate tumor and vessel depiction.

^┼PET-MRI is not widely available.

SPECT: single photon-emission computed tomography; CT: computed tomography; ¹⁸FDG, 18-fluoro-deoxygluocse; ⁶⁸Ga-DOTA-TATE: 68-gallium-tetraxetan(0), tyrosine(3)-octreotate; DOTA-TOC: tetraxetan(0), phenylalanine(1), tyrosine(3)-octreotide; PET: positron-emission tomography; MRI: magnetic resonance imaging; NPO: non per os; IV: intravenous; Bq: Becquerel

^*Discuss holding somatostatin analogues with oncologist³⁵. Short-acting agents can be used up to 24 hours prior to injection. If patient taking a long-acting agent, scan should be performed just prior to next dose

(Please note: MIBG imaging protocol is not included here as it is uncommonly used for pediatric GI NET and more extensively covered in discussions of neuroblastoma imaging.)

4. Follow-up imaging

Desmoid tumor

For follow-up imaging of DT, as with initial staging, MRI is preferred over CT for short- and long-term follow-up of DT¹⁵ (GRADE B; SOR 1.80), to assess for progression, treatment response, and any tumor- or treatment-related complications, noting that MRI characteristics of DT (signal on T2-weigthed images, enhancement pattern, signal on DWI) may correlate with treatment response^13,42,43.

After chemoradiation, follow-up imaging is recommended at 6-12 weeks, and then 3-6 month intervals. Follow-up imaging every 3-6 months is also generally recommended for patients managed conservatively⁴³. Later follow-up intervals can be determined depending on symptoms, imaging biomarkers, and established growth rate^13,43

Desmoplastic small round cell tumor

Conventional follow-up anatomic imaging for DSRCT consists of CT of the chest, abdomen, and pelvis. MRI of the abdomen and pelvis, given its superior contrast resolution and potential for imaging biomarkers, should be considered as an alternative to CT for follow-up of DSRCT (GRADE B; SOR 1.80). Given the aggressiveness of the tumor, ¹⁸FDG PET and contrast-enhanced chest CT are also recommended during follow-up imaging^{3,14,15,30,44} (GRADE C; SOR 1.80). The follow-up imaging interval can be personalized depending on tumor progression, metastatic extent, success of cytoreductive surgery, and feasibility of complex imaging modalities such as MRI and PET.

Gastrointestinal neuroendocrine tumor

Close follow-up for non-appendiceal NET in children is recommended, as recurrence and metastatic spread can be unpredictable, especially in those with larger primary tumor size^8,45. For extra-appendiceal GI NET, CT or MRI may be obtained for follow-up, with initial scan at 3-6 months¹⁸. For reasons discussed above, authors recommend MRI for follow-up of extra-appendiceal GI NET (GRADE B; SOR 2.20). Somatostatin receptor imaging is not routinely indicated for follow-up but could be considered if primary or metastatic lesions were poorly seen on anatomic imaging, or to evaluate concerning clinical or anatomic imaging findings^18,46 (GRADE B; SOR 1.80).

Gastrointestinal stromal tumor

If complete resection with negative surgical margins is achieved for non-metastatic GIST, a post-operative MRI followed by semi-annual/annual abdominal ultrasound and chest radiograph can be considered for long-term follow-up¹⁰. For patients undergoing adjuvant therapy for incomplete resection or metastatic disease, imaging follow-up with MRI of the abdomen/pelvis during the first few years can be performed¹⁰. It should be noted that during tyrosine kinase inhibitor (TKI) therapy, anatomic imaging (CT or MRI) may underestimate response of GIST, as there is a 1-2 month delay in decrease in tumor volume of responding lesions. A decrease in glucose uptake is a marker of response in this period for TKI therapy, and ¹⁸FDG PET is therefore recommended if early assessment is needed for patients treated with TKI, often those with large burden of disease^10,47 (GRADE B; SOR 1.40). Contrast-enhanced ultrasound (CEUS) has been suggested as a follow-up modality for these patients¹⁰. However, due to limited potential for detection of small metastases and a lack of uniform experience with CEUS across centers, we do not currently recommend CEUS for these patients (GRADE D; SOR 1.60). MRI is recommended to assess response over longer periods (≥3 months), and imaging features that suggest response are decreased enhancement and decreased diffusion restriction¹⁰ (GRADE B; SOR 1.40).

5. Response assessment

Response criteria for these relatively rare types of pediatric abdominal tumors have not been validated in children, but considerations regarding various response criteria for each tumor type are discussed below.

Desmoid tumor

Response Evaluation Criteria in Solid Tumors (RECIST 1.1) using a single long-axis measurement is commonly used for response assessment of tumors in adults but should be used cautiously for DT as DT can demonstrate a greater size reduction in its short axis. Additionally, treatment can decrease the active tumor component prior to any decrease in size, and conversely relapse can present as increase in the active component before increase in size⁴⁸. Imaging biomarkers (e.g., size, CT attenuation, signal on T2-weighted MRI, enhancement) have been shown in some studies to correlate with DT therapy response, so various modified response criteria (e.g., Choi criteria based on CT attenuation and nodular enhancement, or modified RECIST (mRECIST) based on arterial enhancement pattern) have been used to incorporate such markers, although none have been fully validated^13,48. Choi criteria or mRECIST may be most appropriate for response assessment in DT, with additional consideration of tumor T2 characteristics (GRADE C; SOR 2.20).

Desmoplastic small round cell tumor

The high recurrence rate and poor prognosis of DSRCT stress the importance of reliable response assessment. DSRCT response evaluation typically relies on subjective assessment of tumor burden on CT/MRI or quantitative size assessment per RECIST or European Pediatric Soft Tissue Sarcoma Study Group (3D-EpSSG) criteria. Such anatomic criteria may not always conform well to DSRCT, as most patients present with widespread, multifocal disease^14,15,28. Additionally, as with DT, a fibrotic stromal component can obscure an actual tumor response, leading to underestimation of response³. Therefore, incorporation of imaging biomarkers, including DWI, which reflects tumor cellularity, and dynamic contrast-enhanced (DCE)-MRI, may provide more detailed qualitative and quantitative information about DSRCT treatment response^3,44. Finally, tumor metabolic activity on follow-up ¹⁸FDG PET has been used as a biomarker in DSRCT; it may precede change in tumor size on CT or MRI, clarify inconclusive anatomic findings, and provide prognostication for survival and remission ^3,15. While not validated for DSRCT, response assessment based on FDG avidity may be performed in addition to analysis of tumor biomarkers on MRI (GRADE D; SOR 2.00).

Gastrointestinal neuroendocrine tumor

The use of biomarkers from DWI and dynamic contrast-enhanced MRI have been reported in GIST response assessment, but mRECIST is most often used in adults^49,50. While it has not been validated for pediatric GI NET, mRECIST can be considered for response assessment in children (GRADE C; SOR 2.20). These response criteria may be applied to evaluate liver metastases, particularly after tumor embolization or ablation. Somatostatin receptor PET can assist in response assessment, but to avoid potential confusion regarding progression, it should be compared to prior somatostatin receptor PET as its sensitivity can be higher than other conventional imaging modalities³⁷.

Gastrointestinal stromal tumor

Response criteria for GIST have not been validated in children. Morphologic criteria such as RECIST can be inaccurate during TKI therapy, and response may be better captured by metabolic changes and cystic remodeling¹⁰. Response criteria that focus on tumor metabolism (e.g., PET Response Criteria in Solid Tumors, for early response assessment on PET) or tumor composition (e.g., degree of diffusion restriction on MRI, or Choi criteria based on CT) as well as size, may be considered for pediatric GIST¹⁰ (GRADE C; SOR 2.40).

6. Off-therapy imaging

Desmoid tumor

For off-therapy imaging of DT, MRI remains the recommended modality (GRADE B; SOR 1.40), though CT can be considered for certain cases. Imaging frequency is not established and may be based on individual disease progression/stability.

Desmoplastic small round cell tumor

Although there are no studies validating off-therapy imaging in DSRCT, the high recurrence rate and poor prognosis support continued surveillance. A combination of anatomic imaging and whole-body ¹⁸FDG PET is recommended³ (GRADE D; SOR 2.17). Imaging frequency is not established and may be guided by expected rate of progression and symptoms.

Gastrointestinal neuroendocrine tumor

GI NET outside the appendix has a substantial risk of recurrence and close surveillance with CT or MRI is recommended, at 6-12 month intervals for at least 7 years^18,45. Given the large number of scans, authors recommend MRI for follow-up of extra-appendiceal GI NET (GRADE B; SOR 2.00). Appendiceal NET do not require long-term surveillance after resection unless symptomatically indicated.

Gastrointestinal stromal tumor

The goal of surveillance in pediatric GIST is the timely detection of progression. Pediatric wild-type GIST demonstrates a relatively indolent course with low overall mortality, but frequent progression can be seen²³. The frequency of surveillance has not been standardized for pediatric GIST, and the strategy may be tailored with consideration of risk factors (e.g., tumor genetic profile, size, grade, metastases)^12,23,47,51. As extra-abdominal metastases are uncommon in pediatric GIST, MRI of the abdomen/pelvis for early follow-up (at 6-month intervals for 3 years, e.g.) and abdominal ultrasound thereafter for long-term follow-up has been suggested, noting that patients with Carney-Stratakis syndrome or the Carney triad may need life-long surveillance for risk of later local recurrence¹⁰ (GRADE C; SOR 2.40).

For all abdominal tumors discussed, late effects of treatment, such as liver lesions and secondary malignancy from radiation, may be characterized as needed by dedicated imaging during surveillance.

7. Imaging trends

Radiomics, the use of quantitative imaging markers to characterize lesions, may play increasingly crucial role in imaging of tumors, particularly sarcomatous lesions, like DSRCT or GIST. Time-activity curves can be obtained from DCE-MRI and will allow for a quantitative assessment of response, as enhancement as a function of time may be able to differentiate between residual sarcomatous tumor, responding tumor, and benign enhancing tissue^52,53. Apparent diffusion coefficient maps may have similar applications for sarcomas⁵⁴. For DT, lesion heterogeneity on MRI can be quantified with radiomics and may also lead to an earlier and more accurate response assessment⁴⁸.

Emerging radiotherapy agents and radiotracers might further improve imaging and treatment options for some of these rare pediatric abdominal tumors. For DSRCT, the novel intra-peritoneal immunotherapy with omburtamab (¹³¹I-8H9, a monoclonal antibody against CD276/BH73 labelled with a beta-emitting isotope) has yielded the prospect of imaging tumors with PET using ¹²⁴I-8H9 and assessing response with immune-related RECIST criteria^14,30. The use of ¹⁷⁷Lu-DOTA-TATE, a beta-emitter, has shown favorable results as a therapeutic agent for adults with NET, and may be increasingly applied in pediatrics^19,55. ⁶⁴Cu-DOTA-TATE may provide advantages over ⁶⁸Ga for imaging of GI NET⁵⁶.

Table 4.

Parameters for CT of the abdomen/pelvis for abdominal tumor assessment and CT of the chest for metastatic assessment (as described for desmoplastic small round cell tumor).

Study	Coverage	Slice thickness	Contrast phase	Reformat planes	Reconstruction	Comment
CT abdomen/pelvis with IV and oral contrast	Abdomen and pelvis	5 mm	Portal venous (60 s after injection begins). Add arterial phase (~10 s delay) for NET.	Sagittal and coronal	Soft tissue algorithm at 1-2 mm	Inject at 2.5 mL/s. Use enterography contrast for bowel tumors. For NET, inject at 5 mL/s.
CT chest with IV contrast	Clavicles through adrenal glands	3 mm	60 s after injection begins	Sagittal and coronal	Lung algorithm at 1-2 mm; axial MIP 10 mm slabs	For follow-up, may forego IV contrast

For tumors arising from the gastrointestinal tract, administration of enterography contrast by mouth or nasogastric tube is suggested, 20 mL/kg given in two parts 1 hour prior to scan. CT: computed tomography; IV: intravenous; s: seconds; NET: neuroendocrine tumor; MIP: maximum intensity projection

Abbreviations key

CT

computed tomography

DOTA-TOC

tetraxetan(0), phenylalanine(1), tyrosine(3)-octreotide

DSRCT

desmoplastic small round cell tumor

DT

desmoid tumor

DWI

diffusion-weighted imaging

FAP

familial adenomatous polyposis

¹⁸FDG-PET

18-fluoro-deoxyglucose positron-emission tomography

⁶⁸Ga-DOTA-TATE

68-gallium-tetraxetan(0), tyrosine(3)-octreotide

GI

gastrointestinal

GIST

gastrointestinal stromal tumor

MIBG

meta-iodobenzylguanidine

mRECIST

modified Response Evaluation Criteria in Solid Tumors

MRI

magnetic resonance imaging

NET

neuroendocrine tumor

NF1

neurofibromatosis type 1

PDGFRA

platelet-derived growth factor receptor alpha gene

RECIST

Response Evaluation Criteria in Solid Tumors

SDH

succinate dehydrogenase gene

SPECT

single photon emission computed tomography

SUV

standardized uptake value

TNM

tumor-node-metastasis

TKI

tyrosine kinase inhibitor