Physical Address
304 North Cardinal St.
Dorchester Center, MA 02124
Pediatrics imaging
Acknowledgments
This work was in part supported by a grant from the National Cancer Institute, grant number R01CA269231, and the Sarcoma Foundation of America. We thank Mehdi Khalighi, Dawn Holley, and Kim Halbert from the PET/MRI Metabolic Service Center for their assistance with the acquisition of PET/MRI scans at the Lucas Research Center at Stanford.
Introduction
Children and adolescents with cancer are referred to specific staging tests based on the primary tumor organ of origin and histology. Based on the accumulated evidence and published experience with 18 F-fluorodeoxyglucose ( 18 F-FDG) positron emission tomography (PET)/magnetic resonance imaging (MRI) studies ( ; ; ; ; ), this technology can add value in imaging pediatric lymphomas and sarcomas. These also represent the main cohort of pediatric cancer patients referred to 18 F-FDG PET/computed tomography (CT) ( ; ; ; ). Evidence in patients with other tumors is emerging and will also be discussed. Integrating MRI and 18 F-FDG PET in children with cancer is advantageous allowing local and whole-body (WB) staging in one session. The label “patient convenience” in this context might be underrated. One-stop exams can help decrease patient suffering and psychological distress while maximizing precious time outside the hospital.
Imaging protocol
Patient preparation
All patients must fast for at least 4 h before an 18 F-FDG PET/MRI scan. Directly before the 18 F-FDG radiotracer injection, we confirm that the blood glucose concentration is below 175 mg/dL.
Radiotracer
We typically inject 3 MBq/kg of 18 F-FDG; followed by a radiotracer uptake time of 45 min. Considering the time to place the patient in the PET/MRI scanner and apply the surface coils, the PET scan starts 60 min after injection. The effective dose from the 18 F-FDG injection can be calculated using the age-specific conversion factors published by the International Commission on radiological protection ( ). While the main radiotracer for pediatric PET/MRI is 18 F-FDG, a few alternate radiotracers can be advantageous for specific clinical questions and will be discussed ahead.
PET/MRI protocol
To set up a time-efficient protocol, whole body axial PET acquisition slabs (25 cm axial field-of-view; 3–4 min per slab) are our time-limiting step. Each bed position is subsequently “filled” with suitable MRI sequences. Different investigators suggested a wide variety of T1-and T2-weighted sequences for anatomical coregistration of 18 F-FDG PET data in children ( ; ; ; ; ). Pediatric tumors can be equally well delineated on Gd-enhanced (CE) T1-weighted scans and unenhanced T2-weighted scans ( ). However, intravenous contrast improved vessel and tumor delineation compared to unenhanced T1-weighted scans, especially when faster T1-weighted sequences were applied. We found CE T1-weighted gradient-echo scans particularly useful for accurate tumor measurements and surgical planning. Therefore, our WB staging protocol always includes CE breath-hold Liver Acquisition with Volume Acquisition (LAVA; 3D Fast Spoiled Gradient Echo) sequences for anatomical coregistration of PET data. These CE scans are acquired sequentially or simultaneously, depending on the contrast agent used (see below). In addition, we acquire axial in and out-of-phase T1-weighted LAVA sequences for attenuation correction and diffusion-weighted imaging (DWI) sequences simultaneously with PET. After completing the WB scan, we add an axial periodically rotated overlapping parallel lines with enhanced reconstruction (PROPELLER) sequence through the lungs to screen for lung nodules. Note that both MRI and PET have limited sensitivity for the detection of subcentimeter pulmonary nodules ( ). Finally, we apply a surface coil and add T1-and T2-weighted Fast Spin Echo (FSE) sequences in appropriate orientations for local staging of the primary tumor. PET data are reconstructed using a 3D time of flight iterative ordered subsets expectation-maximization algorithm (24 subsets, three iterations, temporal resolution =400 px, matrix 256 × 256: voxel size 2.8 × 2.8 × 2.8 mm), accounting for attenuation from coils and patient cradle.
MRI contrast agent
The MRI component of a PET/MRI is often enhanced with gadolinium (Gd)-chelates. When using gadolinium chelates, we perform (1) head-to-toe axial PET and DWI images, followed by Gd-injection, and (2) a second series of head-to-toe LAVA sequences. Considering the growing concern about Gd deposition in the brain, we have introduced the iron oxide nanoparticle compound ferumoxytol (Feraheme) as a contrast agent. Ferumoxytol is FDA-approved as an iron supplement for treating anemia in adult patients and can be used “off-label” as a contrast agent for MRI and PET/MRI. Ferumoxytol provides a long-lasting vascular enhancement for the entire duration of a WB scan and, therefore, can be integrated into the protocol (1) above, eliminating the need for Gd (2). Ferumoxytol enhanced MRI added value to 18 F-FDG PET scans by providing anatomical detail and facilitating the differentiation between reconverted hematopoietic marrow and bone marrow metastases ( ). Nanoparticle-enhanced 18 F-FDG PET/MRI might also improve tumor response assessments to immunotherapies ( ; ).
Lymphomas
Clinical background
Lymphomas comprise up to 10% of newly diagnosed cancers in children (0–14 years) and up to 23% of newly diagnosed cancers in adolescents (15–19 years). Lymphomas are classified as Hodgkin (HL) or non-Hodgkin lymphoma (NHL) ( ). HL has an overall incidence of 1.3 per 100,000 cases ( ), is the most commonly diagnosed cancer in adolescents ( ), and is twice as common in boys as in girls aged 0–14 years, (0.8 vs. 0.4 per 100,000 cases, respectively) ( ). Histological classification of HL includes nodular lymphocyte-predominant (5% of HL cases) and classic HL (95% of HL cases). Classic HL is subcategorized as nodular sclerosis, mixed cellularity, lymphocyte depleted, and lymphocyte-rich ( ). In contrast to HL, while the overall incidence rate of NHL is 1.3 per 100,000 cases, NHL has a higher incidence in male children and adolescents (1.8 M vs. 0.8 F per 100,000 cases, respectively) ( ). According to the World Health Organization (WHO) classification, NHL is phenotypically classified as B cell, including Burkitt lymphoma and diffuse large B-cell lymphoma (DLBCL), T cell, and natural killer (NK) cell NHL ( ). The overall survival rate of HL is significantly higher than NHL ( ).
Diagnosis
Early diagnosis and correct staging of lymphoma can improve optimized therapies and thereby maximize survival. HL staging is performed using the modified Ann Arbor staging system ( ), while the staging of NHL is based on the St. Jude classification ( ). The Lugano classification is simplified and applies to both initial staging and subsequent follow-up ( ). Imaging plays a critical role as the disease stage will determine the treatment ( ). 18 F-FDG PET/CT has been established as the standard technique for WB staging of lymphoma ( ). MRI, compared to CT, can potentially detect additional lesions in the case of involvement of bones, bone marrow, and soft tissues. 18 F-FDG-PET helps to identify tumors based on their high 18 F-FDG avidity. This feature resulted in the incorporation of 18 F-FDG-PET/MRI in the staging of lymphoma and can replace bone marrow biopsies in evaluating bone marrow metastases ( ; ).
18F-FDG-PET/MRI
While 18 F-FDG-PET could be combined with both CT and MRI for attenuation correction and anatomical coregistration of radiotracer data, utilizing MRI as an ionizing radiation-free modality can reduce the radiation exposure ( ). In both HL and NHL 18 F-FDG-PET/MRI ( Fig. 6.1 ) has proven to be a reliable and lower radiation dose alternative to 18 F-FDG-PET/CT ( ).
Diffusion-weighted imaging (DWI) has been utilized for the WB staging of lymphoma in adults and children ( ; ). WB DWI could be easily integrated into tumor staging protocols for pediatric lymphoma patients ( ). The addition of DWI improved staging agreement with FDG-PET/CT ( ). Furthermore, in staging pediatric patients with lymphoma, Littooji et al. introduced WB DWI as a radiation-free alternative to FDG-PET/CT ( ). DWI results in higher sensitivity and accuracy in diagnosing specific subtypes of lymphomas, for example, MALT lymphomas, which have variable or no FDG uptake ( ).
DWI could also be utilized to predict clinical response at the end of treatment. Concordant changes of DWI and 18 F-FDG uptake tend to correlate with the response at the end of treatment in pediatric lymphoma patients. Moreover, several patients may demonstrate partial or complete responses also on interim scans ( ). Interim-apparent diffusion coefficient (ADC) can also help recognize nonresponder lesions in HL patients ( ).
Ewing sarcoma
Clinical background
Ewing sarcoma, the second most common bone malignancy in the pediatric population ( ), has an incidence of 1–3 cases per 10 6 children and accounts for approximately 3% of pediatric malignancies. Ewing sarcoma family of tumors comprises classic osseous Ewing sarcoma of the bone, primitive neuroectodermal tumor, and Askin tumor. They all share a common translocation between chromosomes 11 and 22 and contain similar small round blue cells ( ). At diagnosis, 15%–46% of patients present with distant metastases ( ). Despite recent advancements in treatment, the overall prognosis remains poor (5-year survival rate of approximately 25%–30% in metastatic patients).
Diagnosis
Classic osseous Ewing sarcoma can affect any bone but predilects the femur, ilium, tibia, humerus, fibula, ribs, and sacrum. Ewing sarcomas of the long bones are often diaphyseal and demonstrate aggressive permeative destruction with associated soft tissue components. Before cross-sectional imaging, a radiograph of the primary tumor should be obtained. On radiographs, the typical pattern would be that of a nongeographic, permeative, and moth-eaten appearance with a wide zone of transition, usually accompanied by cortical destruction, soft tissue components, and characteristic onion skin or hair on end periosteal reaction ( ). The absence of tumor osteoid helps to differentiate it from osteosarcoma.
18F-FDG-PET/MRI
MRI imaging features closely mirror the radiographic findings: a large permeative soft tissue component extending from the bone marrow through the cortex. The soft tissue component usually demonstrates homogeneous and intermediate T1w signal and homogeneous low to intermediate T2w signal, secondary to the high cellularity of this tumor ( ). Larger lesions can be heterogeneous with central hemorrhage and necrosis. Enhancement may be homogeneous or peripheral and nodular. Spiculated subperiosteal low signal corresponds to the radiographic hair on end periosteal reaction.
On 18 F-FDG-PET, the tumor may show areas of up-regulated glucose metabolism which correlate with its biological aggressiveness. 18 F-FDG-PET can help to decide the areas to biopsy.
Areas of metabolic activity may or may not overlap with areas of diffusion restriction. However, a high degree of overlap is indicative of highly aggressive sarcomas. Besides local staging, 18 F-FDG-PET/MRI also provides excellent WB staging ( Fig. 6.2 ). Metastases are hematogenous and predominantly hit the lungs and bone marrow ( ). FDG-PET combined with MRI is sensitive even for subtle bone marrow metastases. Since osseous metastases in Ewing sarcoma are lytic, PET has better sensitivity in detecting such hypermetabolic lesions than bone scans ( ). One of the limitations of the PET/MRI is identifying subcentimeter pulmonary metastases which are not metabolically active ( ). Mean SUV values of the primary tumor range from 5.3 for patients with no metastases at presentation to 11.3 for patients with metastases at presentation ( ).
Assessment of treatment response using MRI alone is less accurate since only minimal morphological changes may be evident despite significant tumor viability reduction ( ). FDG-PET allows better detection of progression or response since the biochemical changes in the tumor occur before morphological changes ( ; ).
Osteosarcoma
Clinical background
Osteosarcomas, the most common primary malignant bone tumors in the pediatric population, arise from mesenchymal stem cells or from committed osteoblast precursors ( ). They are characterized by the production of osteoid tumor matrix. According to WHO, they are classified into eight different subtypes: conventional, parosteal, periosteal, telangiectatic, small cell, low-grade central, secondary, and high-grade surface ( ). Other uncommon subtypes include gnathic osteosarcoma and osteosarcomatosis, not included in the WHO classification. This discussion is focused predominantly on conventional osteosarcoma, the most common subtype.
Diagnosis
Common sites of involvement include the metaphysis of the long bones, predominantly of the distal femur, proximal tibia, proximal humerus, and proximal femur. Other bones, such as the pelvis and mandible, can also be involved. Primary osteosarcoma of the wrist and ankle are extremely rare. On the radiographs, it usually presents as a sclerotic, nongeographic aggressive and permeative mass with a sunburnst periosteal reaction ( ).
18F-FDG-PET MRI
Conventional osteosarcoma
Hybrid imaging technique provides excellent sensitivity in assessing local and metastatic disease. Conventional osteosarcomas originate in the center of the medullary cavity. They exhibit a characteristic zoning phenomenon: as the tumor propagates from the medullary cavity to the periphery, the older and mature tumor cells in the medullary cavity ossifies first, followed by the newer peripheral nonossified components of the tumor, which reflect the most actively proliferating part. The latter demonstrates hyperenhancement, diffusion restriction, and FDG uptake; meanwhile, the more mature ossifying central part shows decreased or even absent abnormal metabolic activity and enhancement. This also helps guiding potential sites for biopsy. Occasionally, a characteristic peripheral circumferential “Donut” type of uptake is visualized, which, based on our institutional experience, is associated with a highly aggressive behavior.
Telangiectatic osteosarcoma
This tumor occurs mostly in the metaphysis of the long bones, with the femur being the most common site. As the name suggests, the tumor is composed of telangiectatic elements resulting in characteristic MRI findings, that are an aggressive expansile permeative lesion with multiple blood fluid levels, internal necrosis, and mostly peripheral enhancing soft tissue components. Differentiating this lesion from primary aneurysmal bone cyst (ABC) or secondary ABC is challenging; however, the presence of tumor osteoid in the nonnecrotic peripheral aspect of the tumor on CT or radiographs can help narrow the differential ( ).
Juxtacortical osteosarcoma
Unlike the conventional type, these tumors arise within the periosteum and only secondarily involve the bone marrow ( ). There are three types based on the WHO classification: parosteal, periosteal, and high-grade surface OS.
Parosteal osteosarcoma
This sarcoma originates from the outer fibrous layer of the periosteum and presents as a sessile mushroom-like lesion in the metaphysis of the long bones, characteristically from the posterior aspect of the distal femur ( ; ). On imaging, a lobulated exophytic mass with a central osteoid matrix is seen adjacent to the bone. The lesion has a characteristic “stuck on” appearance and a cleavage plane, separating the mass from the bone marrow (“string sign”) ( ).
Periosteal osteosarcoma
It is predominantly made up of cartilaginous elements, since it arises from the inner germinative layer of the periosteum, thus resulting in high T2w signal ( ). On radiographs, the periosteal reaction perpendicular to the cortex results in a “hair on end” appearance. Medullary involvement is rare.
High-grade surface osteosarcoma
These are usually large at presentation, ranging from 4.5 to 22 cm, and are more aggressive than other surface osteosarcomas, with greater circumferential bone involvement and medullary invasion. Otherwise, imaging features overlap those of periosteal, parosteal, and conventional OS ( ).
Initial staging
Evaluation for potential involvement of adjacent joints, muscles, and neurovascular structures plays a major role for planning local tumor resection. These features are depicted with exquisite details by MRI. It is important to evaluate the adjacent vasculature for tumor thrombosis by assessing flow voids and enhancement patterns. Occasionally, linear FDG uptake along the vessels, corresponding to the tumor thrombus, can be visualized ( Fig. 6.3 ).
On MRI, skip lesions can be diagnosed as areas of abnormal marrow signal, marrow replacement, or enhancement. Diagnosing skip lesions on MRI becomes challenging in patients with physiologic red marrow changes which show high T2w, low T1w signal, and contrast enhancement. In such instances, FDG-PET can diagnose metabolically avid skip lesions.
Osteosarcomas metastasize early and hematogenously, particularly to bones and lungs. Since PET/MRI is limited in assessing lung metastases, chest CT is included for staging ( ).
Treatment monitoring
Since sarcomas do not demonstrate significant reduction in size posttreatment, PET can provide important prognostic information. Percentage necrosis on histology may used for quantifying treatment response ( ).
You're Reading a Preview
Become a Clinical Tree membership for Full access and enjoy Unlimited articles
If you are a member. Log in here