Guidelines for Imaging and Staging of Neuroblastic Tumors

Note: This copy is for your personal, non-commercial use only. To order presentation-ready copies for distribution to your colleagues or clients, contact us at www.rsna.org/rsnarights.

ORIGINAL RESEARCH n

SPECIAL REPORT

Radiology: Volume 261: Number 1—October 2011 n radiology.rsna.org 243

Guidelines for Imaging and Staging of Neuroblastic Tumors: Consensus Report from the International Neuroblastoma Risk Group Project 1

Hervé J. Brisse , MD , PhD M. Beth McCarville , MD Claudio Granata , MD , PhD K. Barbara Krug , MD Sandra L. Wootton-Gorges , MD Kimio Kanegawa , MD Francesco Giammarile , MD Matthias Schmidt , MD Barry L. Shulkin , MD Katherine K. Matthay , MD Valerie J. Lewington , MD Sabine Sarnacki , MD , PhD Barbara Hero , MD Michio Kaneko , MD, PhD Wendy B. London , PhD Andrew D. J. Pearson , MD Susan L. Cohn , MD Tom Monclair , MD , PhD

Neuroblastoma is an enigmatic disease entity; some tumors disappear spontaneously without any therapy, while others progress with a fatal outcome despite the implementation of maximal modern therapy. However, strong prognostic factors can accurately predict whether children have “good” or “bad” disease at diagnosis, and the clinical stage is cur-rently the most signifi cant and clinically relevant prognostic factor. Therefore, for an individual patient, proper stag-ing is of paramount importance for risk assessment and selection of optimal treatment. In 2009, the International Neuroblastoma Risk Group (INRG) Project proposed a new staging system designed for tumor staging before any treatment, including surgery. Compared with the focus of the International Neuroblastoma Staging System, which is currently the most used, the focus has now shifted from surgicopathologic fi ndings to imaging fi ndings. The new INRG Staging System includes two stages of localized dis-ease, which are dependent on whether image-defi ned risk factors (IDRFs) are or are not present. IDRFs are fea-tures detected with imaging at the time of diagnosis. The present consensus report was written by the INRG Imaging Committee to optimize imaging and staging and reduce interobserver variability. The rationales for using imaging methods (ultrasonography, magnetic resonance imaging, computed tomography, and scintigraphy), as well as tech-nical guidelines, are described. Defi nitions of the terms recommended for assessing IDRFs are provided with ex-amples. It is anticipated that the use of standardized no-menclature will contribute substantially to more uniform staging and thereby facilitate comparisons of clinical trials conducted in different parts of the world.

q RSNA, 2011

Supplemental material: http://radiology.rsna.org/lookup/suppl/doi:10.1148/radiol.11101352/-/DC1

1 From the Imaging Department, Institut Curie, 26 rue d’Ulm, 75005 Paris, France (H.J.B.). The complete list of the author affi liations is at the end of this article. Received August 27, 2010; revision requested October 21; revision received January 28, 2011; accepted March 1; fi nal version accepted March 21. Supported in part by the William Guy Forbeck Research Foundation and Little Heroes Cancer Research Foundation. Address correspondence to H.J.B. (e-mail: [email protected] ).

q RSNA, 2011

244 radiology.rsna.org n Radiology: Volume 261: Number 1—October 2011

SPECIAL REPORT: Guidelines for Imaging and Staging Neuroblastic Tumors Brisse et al

diagnosis ( 9,10 ). Stage M indicates dis-seminated disease (comparable to the INSS stage 4), and stage MS indicates metastatic disease similar to INSS stage 4S disease (with exception of primary tumor size and patient age range). Com-pared with the focus on surgicopatho-logic fi ndings in the INSS, the focus has now shifted to imaging. Since imaging data can be retrospectively reviewed, a system based on preoperative diagnostic image fi ndings will be more robust and reproducible than one based on surgical fi ndings. In addition, because the digital imaging data can be reviewed centrally by expert radiologists, the likelihood of uniform evaluation will be increased. This new staging system is not intended to be a substitute for the INSS, and it is recommended that both systems be used in parallel.

Stage L1 tumors are localized tumors that do not involve vital structures, as defi ned according to the list of IDRFs. The tumor must be confi ned within one body compartment: the neck, chest, ab-domen, or pelvis. The isolated fi nding of intraspinal tumor extension that does not fulfi ll the criteria for an IDRF is con-sistent with stage L1 disease.

Stage L2 tumors are local-regional tumors with one or more IDRFs. The tumor may be ipsilaterally continuous

groups. However, the INSS is not suit-able for the pretreatment risk classifi -cation of patients with localized disease as this staging is based on the extent of tumor removal ( 8 ). Actually, the same tumor can be classifi ed as INSS stage 1 or 3 disease depending on the extent of surgical excision ( 9 ), making direct com-parisons of clinical trials based on the INSS diffi cult. Furthermore, the localized disease of patients whose “treatment” is observation because spontaneous tumor regression is anticipated cannot be prop-erly staged by using the INSS. Assess-ment of lymph node involvement is also diffi cult to apply uniformly ( 8 ).

In 2004, investigators from the ma-jor cooperative groups—the Children’s Oncology Group (COG) from North America and Australia–New Zealand, the German Pediatric Oncology and Hematology Group (GPOH), the Japanese Neuroblastoma Study Group (JNBSG), and the International Society of Pedi-atric Oncology Europe Neuroblastoma (SIOPEN) Group—and China formed the International Neuroblastoma Risk Group (INRG) Task Force, which devel-oped the INRG Staging System (INRGSS) and the INRG Risk Classifi cation System for neuroblastoma ( 2,8 ). The INRGSS, published in 2009, is designed for tu-mor staging before surgery or any other treatment ( Table 2 ) ( 8 ). Localized tu-mors are classifi ed as stage L1 or L2 disease on the basis of whether one or more of 20 IDRFs are present ( Table 3 ) ( 8 ). IDRFs are surgical risk factors, de-tected on images, that make total tumor excision risky or diffi cult at the time of

Neuroblastic tumors (ie, neuroblas-toma, ganglioneuroblastoma, gan-glioneuroma) are the most com-

mon extracranial solid tumors occurring in children ( 1 ). The median age at diag-nosis is about 16 months, and 95% of cases are diagnosed by 7 years of age ( 2,3 ). The most common sites of origin of neuroblastic tumors are the adrenal region (48%), extraadrenal retroperi-toneum (25%), and chest (16%). Less common sites are the neck (3%) and the pelvis (3%) ( 2 ). Forty-eight percent of patients have metastatic disease at diagnosis ( 2 ).

The enigmatic nature of this tumor group has long been recognized: Some tumors undergo spontaneous involution without any therapy, while others prog-ress with a fatal outcome despite the im-plementation of maximal modern ther-apy ( 3 ). Fortunately, strong prognostic factors can be used to accurately predict whether children have “good” or “bad” disease at diagnosis ( 4 ). Clinical stage is currently the most statistically signifi -cant and clinically relevant prognostic factor ( 2 ). Other factors include patient age, serum lactate dehydrogenase level, histologic category and grade of tumor differentiation ( 5 ), status of the MYCN oncogene, DNA ploidy, aberrations of chromosomes 1p and 11q, and genomic profi le ( 5 ). Combinations of biologic fac-tors with age and clinical stage have been used as stratifying criteria to defi ne risk groups ( 2 ), and current treatment pro-tocols are stratifi ed according to risk. Thus, accurate staging at the time of di-agnosis is of paramount importance.

The International Neuroblastoma Staging System (INSS) ( Table 1 ), devel-oped in 1988 (7) and modifi ed in 1993 (6), is still used by most cooperative

Implications for Patient Care

Use of a uniform imaging strategy n

will facilitate the staging of neu-roblastic tumors.

Consistent and uniform reporting n

and staging will greatly facilitate comparisons of risk-based clini-cal trials conducted in different regions of the world.

Patients with neuroblastoma n

worldwide will likely benefi t from the early recognition of effective treatment regimens with improved outcomes.

Advance in Knowledge

This consensus report, written by n

experts representing the major international cooperative pediat-ric cancer study groups, will likely optimize imaging and uni-form reporting for neuroblas-toma staging according to the new International Neuroblastoma Risk Group Staging System.

Published online before print 10.1148/radiol.11101352 Content code:

Radiology 2011; 261:243–257

Abbreviations: FDG = fl uorine 18 fl uorodeoxyglucose IDRF = image-defi ned risk factor INRG = International Neuroblastoma Risk Group INRGSS = INRG Staging System INSS = International Neuroblastoma Staging System

Author contributions: Guarantors of integrity of entire study, H.J.B., F.G.; study concepts/study design or data acquisition or data analysis/interpretation, all authors; manuscript drafting or manu-script revision for important intellectual content, all authors; manuscript fi nal version approval, all authors; literature research, H.J.B., M.B.M., S.L.W., B.L.S., K.K.M., S.S., B.H., A.D.J.P., T.M.; clinical studies, C.G., K.K., F.G., M.S., B.L.S., K.K.M., S.S., M.K., S.L.C., T.M.; statistical analysis, W.B.L.; and manuscript editing, H.J.B., M.B.M., C.G., K.B.K., S.L.W., B.L.S., K.K.M., V.J.L., S.S., B.H., W.B.L., A.D.J.P., S.L.C., T.M.

Potential confl icts of interest are listed at the end of this article.



report. Although ganglioneuroma is a lo-calized tumor and tends to be relatively homogeneous, the imaging characteris-tics of neuroblastoma, ganglioneuroblas-toma, and ganglioneuroma are similar ( 13 ). The histologic type cannot be dis-criminated with imaging and is based on pathologic analysis fi ndings only.

The diagnosis of neuroblastic tumor is usually suspected with a high degree of confi dence on the basis of the patient’s age and imaging pattern. Before obtain-ing pathologic specimens or specific markers (urinary catecholamines, MIBG uptake), the most usual challenge is the differential diagnosis of renal versus ex-trarenal retroperitoneal tumor—that is, neuroblastoma versus Wilms tumor—since both of these neoplasms occur dur-ing childhood ( 18 ). Multiplanar imaging

the chairs of their respective cooperative groups (INRG, SIOPEN, COG, JNBSG, GPOH). The present recommendations from this group were developed in con-sensus on the basis of professional expe-riences and review of the literature. This consensus report is provided to optimize imaging and uniform reporting for stag-ing of neuroblastic tumors and to facili-tate comparisons of risk-based clinical trials conducted in different regions of the world.

Imaging of Primary Tumors

Imaging of Neuroblastic Tumors The imaging patterns of neuroblastic tumors have been thoroughly described ( 10–17 ) and are beyond the scope of this

within body compartments—for exam-ple, a left-sided abdominal tumor with left-sided chest involvement should be considered stage L2 disease. However, a clearly left-sided abdominal tumor with right-sided chest (or vice versa) involve-ment is defi ned as metastatic disease.

Stage M tumor refers to distant met-astatic disease (ie, not contiguous with the primary tumor), except as defi ned for stage MS. Nonregional (distant) lymph node involvement is metastatic disease. However, an upper abdominal tumor with enlarged lower mediastinal nodes or a pelvic tumor with inguinal lymph node involvement is considered local-regional disease. Ascites and pleu-ral effusion, even with malignant cells, do not constitute metastatic disease unless they are remote from the body compartment of the primary tumor.

Stage MS tumor refers to metastatic disease in patients younger than 18 months (547 days), with the metasta-sis confi ned to the skin, liver, and/or bone marrow. Bone marrow involve-ment should be limited to less than 10% of all nucleated cells in the culture smears or biopsy samples. Iobenguane I 123 (MIBG) scintigraphy fi ndings must be negative in bone and bone marrow. Provided there is MIBG uptake in the primary tumor, bone scans are not re-quired. The primary tumor can be with or without IDRFs (stage L2 or L1), and there is no restriction regarding crossing or infi ltration of the midline.

The INRGSS has profound implica-tions for imaging, and the radiologist’s infl uence is clearly increased. Further discussion regarding the defi nitions and importance of the IDRFs was perceived to be necessary. Therefore, an interna-tional group of neuroblastoma experts was formed to prepare this special report to disseminate recommendations for imaging and staging. After the initiative from the INRG Executive Committee, six pediatric radiologists (H.J.B. [chair], M.B.M., C.G., K.K., K.B.K., S.L.W.), four nuclear physicians (F.G., V.J.L., M.S., B.L.S.), three pediatric surgeons (M.K., T.M., S.S.), four pediatric oncol-ogists (S.L.C., B.H., K.K.M., A.D.J.P.), and a statistician (W.B.L.) were invited to participate after having been chosen by

Table 1

Descriptions of Original INSS Tumor Stages

Tumor Stage Description

1 Localized tumor with complete gross excision, with or without microscopic residual disease; representative ipsilateral lymph nodes negative for tumor microscopically. Nodes attached to

and removed with the primary tumor may be positive.2A Localized tumor with incomplete gross excision; representative ipsilateral nonadherent lymph

nodes negative for tumor microscopically2B Localized tumor with or without complete gross excision, with ipsilateral nonadherent lymph

nodes positive for tumor; enlarged contralateral lymph nodes negative microscopically3 Unresectable unilateral tumor infi ltrating across the midline (beyond the opposite side of the

vertebral column) with or without regional lymph node involvement, or midline tumor with bilateral extension via infi ltration (unresectable) or lymph node involvement

4 Any primary tumor with dissemination to distant lymph nodes, bone, bone marrow, liver, skin, and/or other organs (except as defi ned for stage 4S disease)

4S Localized primary tumor (as defi ned for stage 1, 2A, or 2B disease) with dissemination limited to skin, liver, and/or bone marrow (limited to infants younger than 1 year, marrow involvement

of less 10% of total nucleated cells, and MIBG scan fi ndings negative in the marrow)

Source.—Reference 6.

Table 2

Descriptions of New INRG Tumor Stages

Tumor Stage Description

L1 Localized tumor not involving vital structures, as defi ned by the list of IDRFs, and confi ned to one body compartment

L2 Local-regional tumor with presence of one or more IDRFsM Distant metastatic disease (except stage MS tumor)MS Metastatic disease in children younger than 18 months, with metastases confi ned to skin,

liver, and/or bone marrow

Source.—Reference 8 . Complete defi nitions of these stages are cited in the text. IDRFs = image-defi ned risk factors.



we recommend that magnetic resonance (MR) or computed tomographic (CT) images always be obtained at the time of diagnosis for accurate staging.

There currently is no consensus about the optimal imaging modality for assessing local disease. Both MR imaging and CT are routinely used, depending on local availability and the radiologist’s preference. In one multiin-stitutional study ( 20 ), CT and MR imag-ing had statistically similar but relatively poor performance in assessing the fea-tures of local disease. However, the aim of that study was to assess the relevance of imaging for assessment of INSS stage but not IDRFs. In the absence of evidence-based criteria, we describe the advan-tages and limitations of both modalities in the following text.

CT is widely available in imaging departments, and current multidetector CT machines allow very fast acquisitions without motion artifacts, reducing the need for sedation. Intraspinal extension is also well depicted, and postprocess-ing software allows multiplanar recon-structions for accurate coronal and sag-ittal views. However, with CT, the use of intravenous iodine-based contrast mate-rial (with its potential risks) is required to increase the soft-tissue contrast and assess the relationships between the tumor and adjacent vessels. Moreover, CT is associated with substantial radia-tion exposure, and children are known to have a higher inherent sensitivity to the negative effects of ionizing radiation ( 21 ). Nevertheless, for patients requir-ing radiation therapy, CT is still the reference-standard technique for defi n-ing the radiation fi elds and dose planning ( 22 ). A preoperative CT study is there-fore usually required in trials where the postoperative radiation fi eld is deter-mined on the basis of the preoperative (postinduction) residual tumor exten-sion. However, although MR imaging is not yet widely available or used for pa-tients with neuroblastoma, there is cur-rently increasing use of this examination as a treatment-planning method ( 23 ).

MR imaging has long been recog-nized as an effective method for imaging neuroblastoma ( 10,14,16,17,20 ); how-ever, to our knowledge, its superiority

Rationale for Use of Imaging Modalities When an abdominal or pelvic tumor is suspected in a child, ultrasonography (US) is usually the fi rst imaging examina-tion performed because of its wide avail-ability and noninvasiveness. Gray-scale and duplex Doppler US allows accurate localization of neuroblastoma and can help defi ne the relationship of this tumor with adjacent organs and vessels. US is also quite effective for imaging the liver in children. In newborns, US may even depict intraspinal involvement. However, US is associated with important limita-tions: low interobserver reproducibility, limited assessment of highly calcifi ed tu-mors because of acoustic shadowing, and marked restrictions in the retrospective data review that is mandatory for clini-cal trials. Surgeons and radiation oncol-ogists also require more extensive imag-ing for treatment planning. Therefore,

analysis usually allows demonstration of the renal or extrarenal origin of the mass. However, a large neuroblastoma invading the kidney can be mistaken for an exophytic Wilms tumor, and vice versa.

Local extension of neuroblastoma mainly consists of perivascular involve-ment with arterial encasement, infi ltra-tion of adjacent soft tissues and organs (mainly the kidneys and liver), and infi ltration of the foramina and epidu-ral space of the spinal canal when the primary tumor arises from a paraspi-nal sympathetic chain. Lymph node involvement occurs in about 30% of cases. Lymph nodes are usually located close to the primary tumor, although distant lymph nodes (previously classi-fi ed as stage 4N tumors if there were no other metastatic sites) may be observed occasionally ( 19 ).

Table 3

Descriptions of IDRFs

Anatomic Region Description

Multiple body compartments

Ipsilateral tumor extension within two body compartments (ie, neck and chest, chest and abdomen, or abdomen and pelvis)

Neck Tumor encasing carotid artery, vertebral artery, and/or internal jugular veinTumor extending to skull baseTumor compressing trachea

Cervicothoracic junction Tumor encasing brachial plexus rootsTumor encasing subclavian vessels, vertebral artery, and/or carotid arteryTumor compressing trachea

Thorax Tumor encasing aorta and/or major branchesTumor compressing trachea and/or principal bronchiLower mediastinal tumor infi ltrating costovertebral junction between T9 and T12 vertebral levels

Thoracoabdominal junction Tumor encasing aorta and/or vena cavaAbdomen and pelvis Tumor infi ltrating porta hepatis and/or hepatoduodenal ligament

Tumor encasing branches of superior mesenteric artery at mesenteric rootTumor encasing origin of celiac axis and/or origin of superior mesenteric arteryTumor invading one or both renal pediclesTumor encasing aorta and/or vena cavaTumor encasing iliac vesselsPelvic tumor crossing sciatic notch

Intraspinal tumor extension

Intraspinal tumor extension (whatever the location) provided that more than one-third of spinal canal in axial plane is invaded, the perimedullary

leptomeningeal spaces are not visible, or the spinal cord signal intensity is abnormal

Infi ltration of adjacent organs and structures

Pericardium, diaphragm, kidney, liver, duodenopancreatic block, and mesentery

Source.—Reference 8 . Conditions that should be recorded but are not considered IDRFs are multifocal primary tumors, pleural effusion with or without malignant cells, and ascites with or without malignant cells.



tumor and the neighboring structure (Fig E3 [online]). When a tumor is sep-arated from a vital structure, an IDRF is not present.

2. Contact means that no visible layer is present between the tumor and the neighboring structure. For an artery, contact means that less than 50% of the vessel’s circumference is in contact with the tumor ( Fig 3 , Fig E4 [online]). In addition, the term f lat-tened is used to describe veins with a reduced diameter that still have a par-tially visible lumen ( Figs 2, 3 , Fig E5 [online]). When a tumor is in contact with a vital structure or is fl attening a vein without encasement, an IDRF is not present, except in the case of renal vessels (see following text).

3. Encasement means that the neigh-boring structure is surrounded by the tumor. When a tumor is encasing a vital structure, according to previous defi ni-tions, an IDRF is present. Encasement of a vessel means that more than 50% of the circumference is in contact with the tumor ( Figs 2–5 , Fig E6 [online]). Total encasement means that a vital structure (organ or vessel) is completely surrounded by the tumor. A fl attened vein with no visible lumen is also con-sidered to be encased.

4. Compression is used only when referring to the airways. When a tumor is in contact with the airways and causes the short axis of the air-ways to be reduced ( Figs 5, 6 , Fig E5 [online]), this pattern is considered an IDRF. For other vital structures (ie, neighboring organs), a contact may be associated with displacement—that is, an abnormal anatomic loca-tion—or distortion—that is, an abnor-mal anatomic shape—of the structure (Figs E7, E8 [online]). However, these situations are not considered IDRFs unless there is infi ltration or total encasement.

5. Infi ltration refers to involvement of vital structures other than vessels. Infi ltration of the vessel wall cannot be demonstrated at imaging. An infi ltrating tumor has extension into a neighboring organ, causing the margins between the tumor and the infi ltrated struc-ture to be lacking or ill defi ned (Fig E8

tumor should always be assessed to aid surgical decisions.

For patients who are either observed only or treated with primary chemo-therapy, imaging has to be repeated, either to document tumor regression or as a preoperative assessment. Since an IDRF may disappear with tumor re-gression ( 26 ), reporting should include reassessment of the patient’s IDRF sta-tus. A checklist of IDRFs is provided as supplemental material (Fig E13 [on-line]) to facilitate and ensure uniform and correct staging at diagnosis; how-ever, this checklist is not intended to be a substitute for the local radiologist’s report.

Anatomic Compartments and Multifocality Most tumors occur in a single anatomic compartment (ie, neck, chest, abdomen, pelvic cavity). However, some tumors may extend into an adjacent compart-ment. This extension often increases the risk of injury to vital structures dur-ing surgery, and such lesions should be classifi ed as stage L2 tumors, even if no other IDRF is present in each compart-ment. This usually occurs with cervico-thoracic lesions arising from the sym-pathetic chain ( 25 ), lower mediastinal tumors extending into the retroperito-neum along the aorta ( Fig 1 ), upper ret-roperitoneal tumors extending into the inferior mediastinum ( Fig 2 ) or pelvis, or pelvic presacral tumors extending ei-ther upward in the abdomen around the aorta or inferior vena cava or laterally through the greater sciatic foramen.

Multifocal primary neuroblastomas are rare and may be familial ( 27,28 ). They can manifest as synchronous or metachronous noncontiguous tumors. When multifocal tumors occur, each neoplasm should be staged according to the greatest extent of disease. This spe-cial feature should be recorded but by itself is not considered an IDRF.

Defi nitions of Terms to Describe Relationships between Primary Tumor and Vital Structures We recommend that the following terms be used to describe the primary tumor:

1. Separation means that a visible layer, usually fat, is present between the

over CT for local-regional staging has not been demonstrated. Compared with CT, MR imaging offers the advantages of higher contrast resolution and the lack of ionizing radiation. There is an ongoing unresolved discussion regard-ing the need for gadolinium-based con-trast material injection to assess neuro-blastoma extension. Gadolinium-based contrast material may improve the as-sessment of infi ltration into adjacent tissues and tumor vascularity. However, T1- and T2-weighted sequences provide excellent contrast resolution, and ves-sels can be displayed suffi ciently with-out the use of contrast material ( 16 ). In one study, the accuracy of T2-weighted and gadolinium-enhanced T1-weighted sequences in defi ning the local-regional tumor extent were compared, and no difference was found ( 24 ).

The main limitations of MR imag-ing are the local availability in some countries and, as compared with CT, the need for sedation in young children because of the longer acquisition time. In patients with intraspinal extension of primary paraspinal tumors, MR imaging is the recommended imaging modality ( 14,25 ) because of the excellent visual-ization of the spinal cord, nerve roots, and subarachnoid spaces. Specifi c tech-nical recommendations for imaging neu-roblastic tumors in children, including CT scanning settings, MR sequences, three-dimensional measurement meth-ods, and data storage, are provided in the Appendix (online).

Recommended Terms to Describe IDRFs To ensure uniform tumor staging, it is recommended that radiologists use spe-cifi c terms to describe the relationships between tumors and the neighboring vital structures—those entities that cannot be sacrifi ced without impairing function. As a part of treatment planning, these relationships should be systematically discussed by a multidisciplinary team that includes ideally radiologists, nuclear medicine physicians, surgeons, and pedi-atric oncologists.

Although metastatic disease (stage M or MS) is staged regardless of the local-regional tumor extent, the absence or presence of an IDRF for the primary



[online]). When a tumor infi ltrates a vital structure, an IDRF is present.

6. The term invasion was also originally included in the IDRF list to describe relationships with the renal pedicle ( 8 ); however, this is a less well-defi ned term. Because surgical dissec-tion of the renal pedicle is particularly risky in patients with neuroblastoma, it was decided that an IDRF is present even if the strict criteria for encase-ment are not fulfi lled—that is, even if the tumor is in contact with the renal vessels only. (This specifi c location is addressed in the following section.) The term invasion was also used to describe spinal canal involvement and is equiva-lent to the term infi ltration for this spe-cifi c location. (This specifi c location is addressed in the following section.)

IDRF Assessment Based on Anatomic Location Cervical neuroblastomas arise mainly from the superior cervical sympathetic chain behind the internal carotid artery. These tumors usually extend anteriorly and laterally, displacing the carotid ar-tery and internal jugular vein. They may also extend medially, compressing the airway (Fig E5 [online]), and upward to the skull base along the carotid artery. Intraspinal extension is usually not ob-served at this anatomic level. IDRFs are present when the tumor encases the carotid artery ( Fig 4 ), vertebral artery, or internal jugular vein; extends to the

Figure 1

Figure 2

Figure 2: (a) Sagittal T1-weighted and (b, c) axial T2-weighted MR images in 2½-year-old girl with medial retroperitoneal neuroblastoma. Tumor extends into inferior mediastinum (*) via infiltration of diaphragma (arrowhead in b ). Tumor completely encases aorta, celiac artery (arrow in a and b ), superior mesenteric artery (not shown), and both renal arteries (arrow in c , left renal artery not shown). All conditions are IDRFs. Inferior vena cava is displaced and fl attened against the mass (arrowhead in c ).

Figure 1: Contrast material–enhanced (iomeprol, Iomeron 350; Bracco Imaging, Evry, France) (a) axial and (b) coronal CT images in 8-month-old girl with left mediastinal calcifi ed infi ltrating neuroblastoma, with contiguous extension into upper retroperi-toneum (*). Multiple compartment extension is classifi ed as L2 disease, regardless of other IDRFs. Aorta (long arrow) is completely encased. Left main bronchus is compressed by the tumor (arrowhead in b ). About two-thirds of spinal canal in axial plane is invaded (arrowheads in a ), and spinal cord is anteriorly displaced. Associated bilateral pleural infi ltration (short arrows) also is seen. All of these conditions are IDRFs.



complications is considerably increased. Associated foraminal and intraspinal extensions are also possible. IDRFs are present when the tumor encases the brachial plexus roots, subclavian vessels, or vertebral or carotid artery or when it compresses the trachea ( Fig 5 ).

Most thoracic neuroblastomas arise from the paraspinal sympathetic chains in the posterior mediastinum. Foram-inal and intraspinal extensions (ie, dumbbell tumors) are often present at this anatomic level. Surgical removal of the intraspinal component is usually not recommended at diagnosis, but it may be necessary as an emergency proce-dure in the presence of acute neurologic symptoms caused by spinal cord com-pression or ischemia. Cord compression may be diagnosed because of neurologic signs or on the basis of imaging fi ndings. An IDRF is present when more than one-third of the spinal canal in the axial plane is invaded (ie, infi ltrated) ( Figs 1, 6, Fig E9 [online]), the leptomeningeal fl uid spaces are no longer visible, or the spinal cord MR signal intensity is abnor-mal (ie, imaging features that are usu-ally observed in patients with neurologic symptoms).

A lower mediastinal tumor, infi ltrat-ing the costovertebral junction between the T9 through T12 vertebral levels, is associated with a theoretical risk of spinal cord ischemia caused by surgi-cal injury of the anterior spinal (Ad-amkiewicz) artery ( 29 ) and by itself is considered an IDRF (Fig E4 [online]). However, the risk of injury to the spi-nal cord is very low in children, prob-ably because of the abundant collateral blood fl ow that develops in response to the rather slow growth of most tumors in this location. Thus, angiography is not mandatory, even if it is performed with noninvasive methods such as CT or MR imaging.

The other major surgical diffi cul-ties are related to relationships with the descending aorta and the pulmo -nary pedicles. Encasement of the aorta ( Fig 1 , Fig E9 [online]) or major branches, compression of the trachea or principal bronchi ( Fig 1 ), and infi l-tration of adjacent structures such as the pericardium, pleura ( Fig 1 ), and

diffi culties ( 26 ). The stellate ganglion is located above the subclavian artery at the level of the origin of the vertebral artery. When dissection of neurovas-cular elements in this region is needed for tumor excision, the risk of surgical

skull base ( Fig 4 ); or compresses the trachea.

Cervicothoracic neuroblastomas aris-ing from the region of the stellate ganglion are rare but associated with particular imaging patterns and surgical

Figure 3

Figure 3: Diagrams illustrate relationship between tumor (NB) and vessels on two views in planes orthogo-nal and parallel to vessel axis, respectively. (a) Contact means no visible layer is present between tumor and neighboring structure. For an artery, contact means that less than 50% of vessel’s circumference is in con-tact with the tumor. A vein that has a reduced diameter but still a partially visible lumen is called a fl attened vein, which is considered equivalent to contact. In such situations, an IDRF is not present (except for renal vessels). (b) Encasement means that vessel is surrounded by the tumor and is classifi ed as an IDRF. Encase-ment of a vessel means that more than 50% of vessel’s circumference is in contact with the tumor; 100% of the circumference in contact with the tumor would indicate total encasement. In addition, a fl attened vein with no visible lumen is classifi ed as equivalent to encasement and must also be considered an IDRF.

Figure 4

Figure 4: (a) Axial fat-saturated T2-weighted and (b) sagittal contrast-enhanced (gadoterate meglumine, Dotarem; Guerbet, Roissy, France) T1-weighted MR images in 4½-year-old girl with right cervical ganglioneuro-blastoma. Internal carotid artery (arrow) is partially encased by tumor, with about 80% of artery circumference in contact with tumor, and tumor (*) extends to skull base (arrowhead). Both conditions are IDRFs.



diaphragm are considered IDRFs. In our experience, the diffi culties in sur-gically dissecting infi ltrating mediasti-nal tumors obviously arising from the periaortic sympathetic plexuses are usually the same as those encountered with infi ltrating periaortic abdominal tumors. At imaging, these tumors are ill-defi ned masses infi ltrating adjacent structures, and they differ from well-circumscribed tumors (Fig E4 [online]) arising from paravertebral sympathetic chains.

Abdominal neuroblastomas arise from the adrenal gland, sympathetic gan-glia (celiac, superior, and inferior mesen-teric ganglia), or sympathetic fi bers and plexuses along the aorta and its main branches. The intimate relationships be-tween the primary tumor and abdominal vessels usually represent the main chal-lenge for the surgeon. Therefore, accurate analyses and descriptions of all arteries and veins are critical parts of the imag-ing reports. The main abdominal vessels (aorta, celiac axis, superior and inferior mesenteric arteries, renal arteries and veins, inferior vena cava, iliac arteries and veins, portal vein) should be sepa-rately assessed and classifi ed according to the terms just defi ned. Tumors par-tially or totally encasing the origin of the celiac axis ( Figs 2, 7 ), the superior mesenteric artery, the aorta ( Fig 1, Fig E6 [online]), one or both renal pedicles ( Figs 2, 7, Fig E6 [online]), the inferior vena cava, or the iliac vessels (Fig E10 [online]) are IDRF positive. Since injury to the inferior mesenteric artery almost never causes complications, encasement of this vessel is not by itself an IDRF.

Renal pedicles are frequently in-volved in retroperitoneal neuroblasto-mas. In the fi rst study of the Localized Neuroblastoma European Study Group ( 30 ) of the International Society of Pe-diatric Oncology, 51% of patients with localized abdominal neuroblastoma had IDRFs, and 45% of these had involve-ment of the renal pedicles. In the Sham-berger et al series ( 31 ), 15% of children treated for abdominal neuroblastoma required nephrectomies or had a renal infarction during surgery. Furthermore, the risk for nephrectomy in children who underwent tumor excision up front

Figure 5

Figure 5: Nonenhanced (a) coronal and (b) sagittal T1-weighted MR images in 16-month-old girl with right cervicothoracic ganglioneuroblastoma. (a) Trachea (arrowheads) is displaced and reduced in diameter by the tumor. (b) Tumor encases C8 and T1 brachial plexus roots, T2 root (arrowheads), subclavian artery (long arrow), and origin of vertebral artery (short arrow). All of these conditions are IDRFs.

Figure 6

Figure 6: Diagram illustrates relationship of tumor with airways and spinal canal on axial section of medi-astinum. (a) Trachea (Tr) diameter is normal (no IDRF). Although spinal foramina and epidural fat are invaded by the tumor, less than one-third of spinal canal in axial plane is invaded. Leptomeningeal fl uid spaces and the spinal cord (SC) MR signal intensity are normal (no IDRF). (b) Trachea diameter is compressed, and this condition is an IDRF. More than one-third of spinal canal in axial plane is invaded, and this also is an IDRF, although leptomeningeal fl uid spaces are still visible and spinal cord MR signal intensity is still normal. Ao = aortic arch, Az = azygos vein, Es = esophagus, NB = neuroblastoma, St = sternum, SVC = superior vena cava, T4 = body of fourth thoracic vertebra.



was twice that of those who underwent resection after chemotherapy. Therefore, even isolated contact with renal vessels is considered an IDRF-positive condi-tion ( Fig 8 ).

The adjacent organs and structures should also be precisely assessed and classifi ed according to the earlier defi ned terms. Tumors infi ltrating the porta he-patis (liver hilum) ( Fig 7 ), diaphragm ( Fig 2, Fig E11 [online]), kidneys (Figs E6, E8 [online]), liver, duodenopancre-atic block ( Fig 7, Fig E6 [online]), or mesentery (Fig E12 [online]) are IDRF positive. Neuroblastomas arising from the inferior mesenteric region (organ of Zuckerkandl) more frequently infi ltrate the mesentery ( 32 ) than do neuroblas-tomas in other locations. Regarding the kidneys, infi ltration may occur directly through the cortex (usually from adre-nal neuroblastomas) (Fig E8 [online]) or through the hilum along renal ves-sels (Fig E6 [online]) (usually from ret-roperitoneal neuroblastomas).

Lumbar paraspinal neuroblastomas are much less common. Frequently clas-sifi ed as having abdominal locations be-cause of their anterior extension, these tumors are different from the above de-scribed retroperitoneal ones. They may be associated with foraminal and in-traspinal extension (dumbbell tumors) (Fig E11 [online]), sharing the same pattern and neurologic issues as medi-astinal primary tumors. As in the medi-astinum, an IDRF is present when more than one-third of the spinal canal in the axial plane is invaded (Fig E11 [online]), when the leptomeningeal spaces are not visible, or when the spinal cord MR sig-nal intensity is abnormal. Tumors involv-ing the lumbar plexus (Fig E10 [online]) are associated with the inherent risk of surgical injury and should be classifi ed as IDRF positive.

Pelvic neuroblastomas arise mainly from the upper hypogastric sympathetic plexus or presacral sympathetic ganglia. Pelvic organs are usually anteriorly dis-placed but not infi ltrated. Major exten-sion occurs along the iliac vessels and into the lumbosacral foramina. Tumors encasing the aorta, inferior vena cava, or iliac vessels (Fig E10 [online]) are con-sidered IDRF positive. Posterior lumbar

Figure 7

Figure 7: Contrast-enhanced (iomeprol) (a, b) axial and (c) coronal CT images in 14-month-old boy with medial retroperitoneal neuroblastoma. Tumor completely encases origin of celiac axis (arrowhead in a ), hepatic artery (downward-pointing long arrow), portal vein (upward-pointing long arrow), and left renal artery (arrowhead in b ). Tumor infi ltrates porta hepatis (ie, liver hilum) (short arrows) and duodenopancreatic block (*). All conditions are IDRFs.

Figure 8

Figure 8: Contrast -enhanced (iomeprol) (a) coronal and (b) sagittal CT images in 4-year-old girl with left retroperitoneal neuroblastoma. Upper pole of tumor is in contact with left renal artery (arrow), and left renal vein (arrowhead) is fl attened. Although renal vessels are not strictly encased, these conditions are IRDFs since surgical dissection of renal pedicle is particularly risky.



using both imaging and bilateral bone marrow aspirate and biopsy. Distant metastases must be assessed by using MIBG scintigraphy, and the study must be performed before tumor excision ( Fig 10 ) ( 8,34 ). Guidelines have been published for MIBG scanning in chil-dren ( 34–36 ).

Since there is no physiologic up-take of MIBG in bone or bone marrow, MIBG scanning is an accurate method for detecting osteomedullary metastases ( Fig 10 ), with sensitivity and specifi city estimated at 90% and 100%, respectively ( 37,38 ). MIBG single photon emission computed tomography (SPECT) enables better depiction of the small focal up-take that is diffi cult to visualize on planar MIBG scans, especially in areas close to intense physiologic uptake such as the

traindication to primary excision of the extraspinal tumor component. More rarely, tumors extend laterally into the gluteal region through the greater sciatic foramen; this condition is also considered to be IDRF positive. On axial CT and MR images, the anatomic level of the greater sciatic foramen is de-fi ned by a sagittal-oblique line between the spine of the ischium and the lat-eral margin of the sacrum ( Fig 9 ).

Imaging of Metastatic Disease

Bone Marrow and Bone Metastases Distant metastases of neuroblastomas are located primarily in bone marrow (70%) or bone (55%) ( 33 ). Bone mar-row involvement must be assessed by

or sacral foraminal extensions are IDRF negative, but intraspinal extension is IDRF positive when more than one-third of the spinal canal in the axial plane is invaded.

Intraspinal tumor extension be-low the level of the L2 vertebra leads to radicular involvement ( Fig 9 ) rather than spinal cord compression. This condition sel dom leads to emergency neurosurgery and is usually not a con-

Figure 9

Figure 9: (a) Axial fat-saturated T2-weighted and (b) sagittal nonenhanced T1-weighted MR images in 14-month-old boy with pelvic neuroblastoma infi ltrat-ing sacral holes and dural cul de sac (not an IDRF) (arrows). This low-situated spinal extension does not lead to spinal cord compression; rather, it leads to radicular involvement only. Tumor extends into right gluteal region (arrowhead) through greater sciatic foramen (an IDRF). Dotted lines between the spine of the ischium and the lateral margin of the sacrum represent the greater sciatic foramen location.

Figure 10

Figure 10: Anterior (left) and posterior (right) planar MIBG scans of 5-year-old girl with left retroperitoneal neuroblastoma. MIBG uptake in primary tumor (arrow), as well as diffuse osteomedullary metastases in cranial vault, skull base, sternum, spine, humerus, ribs, pelvis, femurs, and tibias, is visible.



If required in a specifi c study, dedicated conventional radiography or CT may be helpful in patients with MIBG-positive osteomedullary metastases for distinguish-ing bone marrow from cortical bone involvement.

The role of fl uorine 18 fl uorodeox-yglucose (FDG) positron emission to-mography (PET) in the setting of neu-roblastoma is still under investigation. The feasibility of PET in young chil-dren, the associated radiation exposure, and the higher cost, as compared with those associated with MIBG scanning, may explain the hitherto limited use of this technique. Another limitation is the high physiologic uptake of FDG in the brain, which may hide skull loca-tions. However, FDG PET may help defi ne the distribution of neuroblasto-mas that fail to take up MIBG ( 42,45 ) ( Fig 12 ). FDG PET is superior for

before examination) ( 8 ). Bone scan has a reported sensitivity of 70%–78% and a specifi city of 51% for the detection of bone metastases ( 38,42 ). The high physiologic uptake of the growing meta-physes in children, especially infants, may be misinterpreted as metastases or hide small metastases ( 43 ).

Skeletal surveys have previously been recommended for infants with negative bone scan fi ndings ( 7,43 ). However, in our experience, these rarely add impor-tant information and accordingly are not considered helpful. CT examina-tions performed to assess local disease also commonly depict bone metastases. Although CT was used in some previ-ous protocols as a prognostic part of the staging ( 44 ), it is known to be in-accurate for detecting bone metastases because of its low sensitivity compared with that of MIBG or bone scan ( 20 ).

liver and the bladder. Postprocessing tools also enable fusion of SPECT and CT images ( Fig 11 ), improving anatomic localization ( 39–41 ).

One unequivocal MIBG-positive le-sion at a distant site is suffi cient to de-fi ne metastatic disease. However, a single equivocal lesion on MIBG scans requires confirmation with another imaging modality—namely, conventional radiogra-phy, and MR imaging or CT if the radio-graphic fi ndings are negative—or with biopsy. To better assess the extent of disease and the response to therapy, the use of a semiquantitative scoring system is strongly recommended ( 34 ).

Technetium Tc 99m medronate bone scintigraphy is usually not required, except in cases in which the primary tumor is not MIBG avid or MIBG positivity cannot be confi rmed (ie, if the primary tumor has been removed

Figure 11

Figure 11: MIBG SPECT (top tow) and CT (middle and bottom rows) scans in 12-year-old boy with right retroperitoneal neuroblastoma. Top far left: MIBG uptake of primary tumor (arrow) is well depicted on anterior maximum intensity projection SPECT image. Multiplanar SPECT reconstructions (top row: second from left, third from left, and far right images) can be coregistered and fused with CT images (middle panel) to depict anatomic location and MIBG activity on the same (fused) image (bottom row) in each of the three reference cutting planes: axial (right), sagittal (middle), and coronal (right).



diffusion restriction because of high cellu-lar density ( 55 ). Therefore, MR imaging could become a major imaging tool in children, allowing staging of both local-regional and distant disease. However, further studies are needed to assess the accuracy of this imaging strategy.

Other Metastatic Sites The anatomic location of involved lymph nodes should be precisely assessed. An upper abdominal tumor with contigu-ous or noncontiguous enlarged lower mediastinal nodes or a pelvic tumor with contiguous or noncontiguous inguinal lymph nodes is considered local-regional disease. However, distant (nonregional) lymph nodes depicted by CT, MR im-aging, or MIBG scanning (eg, supra-clavicular lymph nodes associated with retroperitoneal primary tumor) are con-sidered metastatic (stage M) disease.

Liver and subcutaneous metastases are mainly observed in infants (Pepper syndrome, stage 4S or MS) ( 33 ). Liver metastases can be assessed with US, CT, or MR imaging. Liver metastases may manifest in one of two forms: focal masses or diffuse infi ltration. The latter

results in children showed whole-body MR imaging to have higher sensitivity than bone scan for the detection of bone marrow metastases (50 ,51 ), and the short-tau inversion-recovery sequence was described as an effi cient sequence in children with small cell tumors (47 ,48 ). One small series (51) involving patients with neuroblastoma provided encour-aging results: Whole-body MR imag-ing had higher sensitivity than MIBG scintigraphy. However, compared with MIBG scintigraphy, whole-body MR im-aging has several limitations. The fi rst limitation is the low specifi city of MR signal intensity abnormalities. Whole-body MR imaging also has low pre-dictive value in the detection of ex-traskeletal metastases (51). Finally, the capability of MR imaging for assessing tumor response has not been reported in patients with neuroblastoma, whereas MIBG scoring is possible and aids in as-sessing therapeutic response.

Diffusion-weighted MR imaging has been described as a promising tool for assessing metastatic response to ther-apy for various neoplasms ( 52–54 ), and neuroblastoma is known to demonstrate

demonstrating soft-tissue lesions and INSS stage 1 and 2 tumors, but MIBG scanning is more sensitive for the detec-tion of bone metastases ( 45,46 ).

The role of whole-body MR imag-ing in bone marrow assessment is still not defi ned in patients with neuroblas-toma. In the fi rst publications related to the MR assessment of bone and bone marrow metastases from neuroblas-toma, this method demonstrated high sensitivity (47 ,48 ). The combination of nonenhanced and gadolinium-enhanced T1-weighted MR sequences has been considered the optimal method ( 49 ). In one large study ( 20 ) comparing the accuracies of CT, MR imaging, and bone scan in staging neuroblastoma, MR imaging alone and bone scan alone yielded similar results for the depic-tion of skeletal metastases. However, in that study, the whole skeleton was not assessed with MR imaging, and MIBG scintigraphy was not included for comparison.

During the late 1990s, whole-body MR imaging was reported as an effec-tive method for the depiction of bone and bone marrow metastases. Early

Figure 12

Figure 12: FDG PET and CT images in 22-month-old boy. Left: axial nonenhanced CT image (top) and fused CT and attenuation-corrected PET image (bottom). Middle: corresponding axial PET images with (top) and without (bottom) attenuation correction. Right: anterior maximum intensity projection FDG PET image. Right-sided abdominal tumor (arrows, top left and middle; short arrow, right) demonstrates peripheral FDG-avid areas and central low activity related to tumor necrosis. Invaded regional retroperitoneal lymph nodes (arrowhead) are well depicted. Osteomedullary metastases (long arrows, right) in spine and limbs are visible on maximum intensity projection.



Although not a staging issue, when skull base involvement is clinically sus-pected or depicted on MIBG or bone scans, brain MR imaging or CT should be performed to detect optic nerve com-pression ( 63 ), which may require emer-gency treatment. Figure 13 is a fl owchart of the imaging strategy at diagnosis for patients with neuroblastoma—both local-regional disease and distant metastases—including mandatory and optional examinations.

Conclusion

Patients with neuroblastoma should be stratifi ed at the time of diagnosis on the basis of reproducible imaging criteria. Increased focus on imaging enhances the role of radiologists in diagnosis and treatment planning.

Treatment strategies must be decided by the individual cooperative groups. Nevertheless, an international group has for many years used the absence of IDRFs to select patients for primary surgical treatment ( 30 ). It is therefore anticipated that the knowledge accu-mulated on the basis of the concept of IDRFs will become an increasingly larger integrated part of therapy plan-ning, both for treatment protocols and for individual patients.

These guidelines can be consid-ered a platform for further studies. As with all projects involving international collaboration, compromises had to be made to reach consensus guidelines and recommendations. However, with con-tinuous efforts, these can be improved and updated as needed. To do so, new and accurate information has to be ac-quired. If incorporated into protocols and used systematically, the guidelines will facilitate the collection of consistent and standardized data that will be use-ful for evaluating the effect of IDRFs on patient outcomes.

Author Affi liations: From the Imaging Depart-ment, Institut Curie, 26 rue d’Ulm, 75005 Paris, France (H.J.B.); Department of Radiological Sciences (M.B.M.) and Division of Nuclear Medicine (B.L.S.), St Jude Children’s Research Hospital, Memphis, Tenn; Department of Radi-ology, Ospedale Gaslini, Genova, Italy (C.G.); Departments of Radiology (K.B.K.), Nuclear

Figure 13

Figure 13: Flowchart of neuroblastoma imaging workup at diagnosis. i.v. = intravenous. XR = radiograph. ∗ = recommended examinations.

pattern is primarily observed in infants and is sometimes missed at CT because it may uniformly increase the parenchy-mal attenuation ( 13 ), although the liver is usually enlarged.

Lung and pleural metastases are more common among patients with MYCN -amplifi ed tumors, but they are rarely observed (3.6% of patients with INSS stage 4 disease) ( 56 ) and there-fore are not routinely screened for with CT. The CT pattern of lung metastases is nonspecifi c ( 57 ). CT is the best method for assessing these metastases since both MIBG scintigraphy and FDG PET have low sensitivity in depicting small lung nodules. However, hybrid MIBG SPECT and combination PET/CT ma-chines will benefi t from the high sensi-tivity of CT for better lesion depiction.

Although pleural disease is associated with reduced survival rates in patients

with metastatic disease ( 57 ), isolated pleural effusion (or ascites) is rare in pa-tients with local-regional disease, and its effect on outcome is still unclear. Therefore, effusions and ascites should be recorded, but these fi ndings alone are not considered IDRFs, and neither condition is considered metastatic disease.

Central nervous system parenchy-mal or meningeal metastatic spread is also very rare and is usually observed during recurrences ( 58,59 ). Therefore, it is usually not systematically screened for at diagnosis unless specifi c symp-toms (other than those related to spinal cord compression) are present. The ra-diologic features of central nervous sys-tem metastases vary from single large parenchymal lesions with possible cys-tic or hemorrhagic features or calcifi ca-tions to diffuse meningeal involvement ( 58,60–62 ).



15 . Brisse H , Edeline V , Michon J , Couanet D , Zucker J , Neuenschwander S . Current strat-egy for the imaging of neuroblastoma [in French] . J Radiol 2001 ; 82 ( 4 ): 447 – 454 .

16 . Sofka CM , Semelka RC , Kelekis NL , et al . Magnetic resonance imaging of neuroblas-toma using current techniques . Magn Reson Imaging 1999 ; 17 ( 2 ): 193 – 198 .

17 . Berdon WE , Ruzal-Shapiro C , Abramson SJ , Garvin J . The diagnosis of abdominal neu-roblastoma: relative roles of ultrasonogra-phy, CT, and MRI . Urol Radiol 1992 ; 14 ( 4 ): 252 – 262 .

18 . Wu YH , Song B , Xu J , et al . Retroperito-neal neoplasms within the perirenal space in infants and children: differentiation of renal and non-renal origin in enhanced CT images . Eur J Radiol 2010 ; 75 ( 3 ): 279 – 286 .

19 . Abramson SJ , Berdon WE , Stolar C , Ruzal-Shapiro C , Garvin J . Stage IVN neuroblas-toma: MRI diagnosis of left supraclavicular “Virchow’s” nodal spread . Pediatr Radiol 1996 ; 26 ( 10 ): 717 – 719 .

20 . Siegel MJ , Ishwaran H , Fletcher BD , et al . Staging of neuroblastoma at imaging: report of the radiology diagnostic oncology group . Radiology 2002 ; 223 ( 1 ): 168 – 175 .

21 . Brody AS , Frush DP , Huda W , Brent RL ; American Academy of Pediatrics Section on Radiology . Radiation risk to children from computed tomography . Pediatrics 2007 ; 120 ( 3 ): 677 – 682 .

22 . Gatcombe HG , Marcus RB Jr , Katzenstein HM , Tighiouart M , Esiashvili N . Excellent local control from radiation therapy for high-risk neuroblastoma . Int J Radiat Oncol Biol Phys 2009 ; 74 ( 5 ): 1549 – 1554 .

23 . Jonsson JH , Karlsson MG , Karlsson M , Nyholm T . Treatment planning using MRI data: an analysis of the dose calculation ac-curacy for different treatment regions . Ra-diat Oncol 2010 ; 5 : 62 .

24 . Kornreich L , Horev G , Kaplinsky C , Ziv N , Grünebaum M . Neuroblastoma: evaluation with contrast enhanced MR imaging . Pe-diatr Radiol 1991 ; 21 ( 8 ): 566 – 569 .

25 . Sauvat F , Brisse H , Magdeleinat P , et al . The transmanubrial approach: a new operative approach to cervicothoracic neuroblastoma in children . Surgery 2006 ; 139 ( 1 ): 109 – 114 .

26 . Squire R , Sarnacki S , Haider N , Gerrard M , Rubie H , Michon J . The outcome of surgical procedures at diagnosis in localised infant neuroblastoma and the effect of chemo-therapy on resectability: European Infant Neuroblastoma Study [abstr] . Pediatr Blood Cancer 2007 ; 49 : 402 .

27 . Maris JM , Kyemba SM , Rebbeck TR , et al . Molecular genetic analysis of familial

References 1 . Pearson ADJ , Pinkerton R . Neuroblastoma .

In: Pinkerton R , Plowman PN , Pieters R , eds. Paediatric oncology . London, England : Arnold , 2004 ; 386 – 414 .

2 . Cohn SL , Pearson AD , London WB , et al . The International Neuroblastoma Risk Group (INRG) classifi cation system: an INRG Task Force report . J Clin Oncol 2009 ; 27 ( 2 ): 289 – 297 .

3 . Castleberry RP , Pritchard J , Ambros P , et al . The International Neuroblastoma Risk Groups (INRG): a preliminary report . Eur J Cancer 1997 ; 33 ( 12 ): 2113 – 2116 .

4 . Woods WG , LaQuaglia MP . The drama of the gifted disease . J Clin Oncol 2009 ; 27 ( 2 ): 172 – 173 .

5 . Shimada H , Umehara S , Monobe Y , et al . International neuroblastoma pathology clas-sifi cation for prognostic evaluation of pa-tients with peripheral neuroblastic tumors: a report from the Children’s Cancer Group . Cancer 2001 ; 92 ( 9 ): 2451 – 2461 .

6 . Brodeur GM , Pritchard J , Berthold F , et al . Revisions of the international criteria for neuroblastoma diagnosis, staging, and re-sponse to treatment . J Clin Oncol 1993 ; 11 ( 8 ): 1466 – 1477 .

7 . Brodeur GM , Seeger RC , Barrett A , et al . International criteria for diagnosis, staging, and response to treatment in patients with neuroblastoma . J Clin Oncol 1988 ; 6 ( 12 ): 1874 – 1881 .

8 . Monclair T , Brodeur GM , Ambros PF , et al . The International Neuroblastoma Risk Group (INRG) staging system: an INRG Task Force report . J Clin Oncol 2009 ; 27 ( 2 ): 298 – 303 .

9 . Kushner BH , LaQuaglia MP , Kramer K , Cheung NK . Radically different treatment recommendations for newly diagnosed neu-roblastoma: pitfalls in assessment of risk . J Pediatr Hematol Oncol 2004 ; 26 ( 1 ): 35 – 39 .

10 . Siegel MJ , Jaju A . MR imaging of neuroblas-tic masses . Magn Reson Imaging Clin N Am 2008 ; 16 ( 3 ): 499 – 513, vi .

11 . McHugh K . Renal and adrenal tumours in children . Cancer Imaging 2007 ; 7 : 41 – 51 .

12 . Mehta K , Haller JO , Legasto AC . Imaging neuroblastoma in children . Crit Rev Com-puted Tomogr 2003 ; 44 ( 1 ): 47 – 61 .

13 . Lonergan GJ , Schwab CM , Suarez ES , Carlson CL . Neuroblastoma, ganglioneuro-blastoma, and ganglioneuroma: radiologic-pathologic correlation . RadioGraphics 2002 ; 22 ( 4 ): 911 – 934 .

14 . Hiorns MP , Owens CM . Radiology of neu-roblastoma in children . Eur Radiol 2001 ; 11 ( 10 ): 2071 – 2081 .

Medicine (M.S.), and Pediatric Oncology and Hematology, Children’s Hospital (B.H.), Univer-sity of Cologne, Cologne, Germany; Department of Radiology, University of California, Davis Medical Center and UC Davis Children’s Hospi-tal, Sacramento, Calif (S.L.W.); Department of Radiology, Aichi Children’s Health and Medical Center, Aichi, Japan (K.K.); Nuclear Medicine Department, CHLS, Hospices Civils de Lyon and Faculté de Médecine, Lyon, France (F.G.); Children’s Oncology Group Statistics and Data Center, University of Florida, Gainesville, Fla, and Department of Pediatrics, University of California San Francisco, San Francisco, Calif (K.K.M.); Nuclear Medicine Department, Royal Marsden NHS Foundation Trust, Sutton, United Kingdom (V.J.L.); Department of Pediatric Surgery, Hôpital Necker-Enfants-Malades, Paris, France (S.S.); Department of Pediatric Surgery, Institute of Clinical Medicine, University of Tsukuba, Ibaraki, Japan (M.K.); Department of Biostatistics, Division of Hematology/Oncology, Children’s Hospital Boston/Dana-Farber Cancer Institute, Harvard Medical School, Boston, Mass (W.B.L.); Section of Paediatrics, Institute of Cancer Research and Royal Marsden Hospital, Sutton, United Kingdom (A.D.J.P.); Department of Pediatrics, University of Chicago, Chicago, Ill (S.L.C.); and Department of Paediatric Surgery, Oslo University Hospital, Rikshospitalet, Oslo, Norway (T.M.).

Cooperative Groups: International Neuroblas-toma Risk Group (INRG): Hervé J. Brisse, Francesco Giammarile, Katherine K. Matthay, Wendy B. London, Andrew D. J. Pearson, Susan L. Cohn, and Tom Monclair. International So-ciety of Pediatric Oncology Europe Neuroblas-toma Group (SIOPEN): Hervé J. Brisse, Clau-dio Granata, Valerie J. Lewington, and Sabine Sarnacki. Children’s Oncology Group (COG): M. Beth McCarville, Sandra L. Wootton-Gorges, and Barry L. Shulkin. Japanese Neuroblastoma Study Group (JNBSG): Kimio Kanegawa and Michio Kaneko. German Pediatric Oncology and Hematology Group (GPOH): K. Barbara Krug, Matthias Schmidt, and Barbara Hero.

Disclosures of Potential Confl icts of Interest: H.J.B. No potential confl icts of interest to dis-close. M.B.M. No potential confl icts of interest to disclose. C.G. No potential confl icts of inter-est to disclose. K.B.K. No potential confl icts of interest to disclose. S.L.W. No potential confl icts of interest to disclose. K.K. No potential con-fl icts of interest to disclose. F.G. Financial ac-tivities related to the present article: none to disclose. Financial activities not related to the present article: receives royalties from Elsevier Masson. Other relationships: none to disclose. M.S. No potential confl icts of interest to dis-close. B.L.S. No potential confl icts of interest to disclose. K.K.M. No potential confl icts of inter-est to disclose. V.J.L. No potential confl icts of interest to disclose. S.S. No potential confl icts of interest to disclose. B.H. No potential confl icts of interest to disclose. M.K. No potential con-fl icts of interest to disclose. W.B.L. No potential confl icts of interest to disclose. A.D.J.P. No po-tential confl icts of interest to disclose. S.L.C. No potential confl icts of interest to disclose. T.M. No potential confl icts of interest to disclose.



neuroblastoma . Eur J Cancer 1997 ; 33 ( 12 ): 1923 – 1928 .

28 . Hiyama E , Yokoyama T , Hiyama K , et al . Multifocal neuroblastoma: biologic behavior and surgical aspects . Cancer 2000 ; 88 ( 8 ): 1955 – 1963 .

29 . Boglino C , Martins AG , Ciprandi G , Sousinha M , Inserra A . Spinal cord vascular injuries following surgery of advanced thoracic neu-roblastoma: an unusual catastrophic com-plication . Med Pediatr Oncol 1999 ; 32 ( 5 ): 349 – 352 .

30 . Cecchetto G , Mosseri V , De Bernardi B , et al . Surgical risk factors in primary surgery for localized neuroblastoma: the LNESG1 study of the European International Society of Pediatric Oncology Neuroblastoma Group . J Clin Oncol 2005 ; 23 ( 33 ): 8483 – 8489 .

31 . Shamberger RC , Smith EI , Joshi VV , et al . The risk of nephrectomy during local con-trol in abdominal neuroblastoma . J Pediatr Surg 1998 ; 33 ( 2 ): 161 – 164 .

32 . Berdon WE , Stylianos S , Ruzal-Shapiro C , Hoffer F , Cohen M . Neuroblastoma arising from the organ of Zuckerkandl: an unusual site with a favorable biologic outcome . Pe-diatr Radiol 1999 ; 29 ( 7 ): 497 – 502 .

33 . DuBois SG , Kalika Y , Lukens JN , et al . Metastatic sites in stage IV and IVS neu-roblastoma correlate with age, tumor biol-ogy, and survival . J Pediatr Hematol Oncol 1999 ; 21 ( 3 ): 181 – 189 .

34 . Matthay KK , Shulkin B , Ladenstein R , et al . Criteria for evaluation of disease extent by (123)I-metaiodobenzylguanidine scans in neu-roblastoma: a report for the International Neuroblastoma Risk Group (INRG) Task Force . Br J Cancer 2010 ; 102 ( 9 ): 1319 – 1326 .

35 . Olivier P , Colarinha P , Fettich J , et al . Guide-lines for radioiodinated MIBG scintigraphy in children . Eur J Nucl Med Mol Imaging 2003 ; 30 ( 5 ): B45 – B50 .

36 . Franzius C , Schmidt M , Hero B , Pfl uger T , Hahn K . Procedure guidelines for MIBG-scintigraphy in children [in German] . Nuklearmedizin 2008 ; 47 ( 3 ): 132 – 138 .

37 . Lumbroso J , Guermazi F , Hartmann O , et al . Sensitivity and specifi city of meta-iodobenzylguanidine (mIBG) scintigraphy in the evaluation of neuroblastoma: analy-sis of 115 cases [in French] . Bull Cancer 1988 ; 75 ( 1 ): 97 – 106 .

38 . Jacobs A , Delree M , Desprechins B , et al . Consolidating the role of *I-MIBG-scintigraphy in childhood neuroblastoma: fi ve years of clinical experience . Pediatr Radiol 1990 ; 20 ( 3 ): 157 – 159 .

39 . Gelfand MJ , Elgazzar AH , Kriss VM , Mas-ters PR , Golsch GJ . Iodine-123-MIBG SPECT versus planar imaging in children with neu-

ral crest tumors . J Nucl Med 1994 ; 35 ( 11 ): 1753 – 1757 .

40 . Rozovsky K , Koplewitz BZ , Krausz Y , et al . Added value of SPECT/CT for correlation of MIBG scintigraphy and diagnostic CT in neu-roblastoma and pheochromocytoma . AJR Am J Roentgenol 2008 ; 190 ( 4 ): 1085 – 1090 .

41 . Rufi ni V , Shulkin B . The evolution in the use of MIBG in more than 25 years of experi-mental and clinical applications . Q J Nucl Med Mol Imaging 2008 ; 52 ( 4 ): 341 – 350 .

42 . Shulkin BL , Hutchinson RJ , Castle VP , Yanik GA , Shapiro B , Sisson JC . Neuroblas-toma: positron emission tomography with 2-[fl uorine-18]-fl uoro-2-deoxy-D-glucose com-pared with metaiodobenzylguanidine scintig-raphy . Radiology 1996 ; 199 ( 3 ): 743 – 750 .

43 . Kaufman RA , Thrall JH , Keyes JW Jr , Brown ML , Zakem JF . False negative bone scans in neuroblastoma metastatic to the ends of long bones . AJR Am J Roentgenol 1978 ; 130 ( 1 ): 131 – 135 .

44 . De Bernardi B , Gerrard M , Boni L , et al . Excellent outcome with reduced treatment for infants with disseminated neuroblastoma without MYCN gene amplifi cation . J Clin Oncol 2009 ; 27 ( 7 ): 1034 – 1040 .

45 . Sharp SE , Shulkin BL , Gelfand MJ , Salis-bury S , Furman WL . 123I-MIBG scintigra-phy and 18F-FDG PET in neuroblastoma . J Nucl Med 2009 ; 50 ( 8 ): 1237 – 1243 .

46 . Taggart DR , Han MM , Quach A , et al . Comparison of iodine-123 metaiodobenzyl-guanidine (MIBG) scan and [18F]fl uorode-oxyglucose positron emission tomography to evaluate response after iodine-131 MIBG therapy for relapsed neuroblastoma . J Clin Oncol 2009 ; 27 ( 32 ): 5343 – 5349 .

47 . Mazumdar A , Siegel MJ , Narra V , Luchtman-Jones L . Whole-body fast inversion recov-ery MR imaging of small cell neoplasms in pediatric patients: a pilot study . AJR Am J Roentgenol 2002 ; 179 ( 5 ): 1261 – 1266 .

48 . Kellenberger CJ , Epelman M , Miller SF , Babyn PS . Fast STIR whole-body MR imag-ing in children . RadioGraphics 2004 ; 24 ( 5 ): 1317 – 1330 .

49 . Meyer JS , Siegel MJ , Farooqui SO , Jaramillo D , Fletcher BD , Hoffer FA . Which MRI sequence of the spine best reveals bone-marrow metas-tases of neuroblastoma? Pediatr Radiol 2005 ; 35 ( 8 ): 778 – 785 .

50 . Daldrup-Link HE , Franzius C , Link TM , et al . Whole-body MR imaging for detection of bone metastases in children and young adults: comparison with skeletal scintigra-phy and FDG PET . AJR Am J Roentgenol 2001 ; 177 ( 1 ): 229 – 236 .

51 . Goo HW , Choi SH , Ghim T , Moon HN , Seo JJ . Whole-body MRI of paediatric malignant

tumours: comparison with conventional on-cological imaging methods . Pediatr Radiol 2005 ; 35 ( 8 ): 766 – 773 .

52 . Schmidt GP , Reiser MF , Baur-Melnyk A . Whole-body MRI for the staging and follow-up of patients with metastasis . Eur J Radiol 2009 ; 70 ( 3 ): 393 – 400 .

53 . Darge K , Jaramillo D , Siegel MJ . Whole-body MRI in children: current status and future applications . Eur J Radiol 2008 ; 68 ( 2 ): 289 – 298 .

54 . Jacobs MA , Pan L , Macura KJ . Whole-body diffusion-weighted and proton imaging: a re-view of this emerging technology for moni-toring metastatic cancer . Semin Roentgenol 2009 ; 44 ( 2 ): 111 – 122 .

55 . Uhl M , Altehoefer C , Kontny U , Il’yasov K , Büchert M , Langer M . MRI-diffusion im-aging of neuroblastomas: fi rst results and correlation to histology . Eur Radiol 2002 ; 12 ( 9 ): 2335 – 2338 .

56 . Dubois SG , London WB , Zhang Y , et al . Lung metastases in neuroblastoma at ini-tial diagnosis: a report from the Interna-tional Neuroblastoma Risk Group (INRG) project . Pediatr Blood Cancer 2008 ; 51 ( 5 ): 589 – 592 .

57 . Kammen BF , Matthay KK , Pacharn P , Gerb-ing R , Brasch RC , Gooding CA . Pulmonary metastases at diagnosis of neuroblastoma in pediatric patients: CT fi ndings and prog-nosis . AJR Am J Roentgenol 2001 ; 176 ( 3 ): 755 – 759 .

58 . Matthay KK , Brisse H , Couanet D , et al . Central nervous system metastases in neu-roblastoma: radiologic, clinical, and biologic features in 23 patients . Cancer 2003 ; 98 ( 1 ): 155 – 165 .

59 . D’Ambrosio N , Lyo JK , Young RJ , Haque SS , Karimi S . Imaging of metastatic CNS neu-roblastoma . AJR Am J Roentgenol 2010 ; 194 ( 5 ): 1223 – 1229 .

60 . Aronson MR , Smoker WR , Oetting GM . Hem-orrhagic intracranial parenchymal metasta-ses from primary retroperitoneal neuroblas-toma . Pediatr Radiol 1995 ; 25 ( 4 ): 284 – 285 .

61 . Kellie SJ , Hayes FA , Bowman L , et al . Primary extracranial neuroblastoma with central ner-vous system metastases characterization by clinicopathologic fi ndings and neuroimaging . Cancer 1991 ; 68 ( 9 ): 1999 – 2006 .

62 . Quackenbush KE , Luna-Fineman S , Magee JF , et al . Neuroblastoma involvement of the falx cerebri . Pediatr Blood Cancer 2009 ; 53 ( 7 ): 1337 – 1339 .

63 . Lau JJ , Trobe JD , Ruiz RE , et al . Metastatic neuroblastoma presenting with binocular blindness from intracranial compression of the optic nerves . J Neuroophthalmol 2004 ; 24 ( 2 ): 119 – 124 .

Documents

Guidelines for Imaging and Staging of Neuroblastic Tumors