Radioimmunodetection of with Iodine- 131-.3F8: Correlation with Biopsy, Iodine 131-Metaiodobenzylguanidineand Standard Diagnostic Modalities

Samuel D. J. Yeh, Steven M. Larson, Leslie Burch, Brian H. Kushner, Michael Laquaglia, Ronald Finn, and Nai-Kong V. Cheung

Departments ofMedical Imaging, Pediatrics, and Surgery, Memorial Sloan-Kettering Cancer Center and Cornell Medical College, New York, New York

determining the full extent of metastatic disease. A more Iodine-i 31-3F8, a murine IgG3monoclonal antibody specific sensitive and specific imaging modality is [‘@‘I]or [123JJ.. for gangliosideGD2was evaluatedby radioimmunoscintigra metaiodobenzylguanidine (MIBG) (2—4).Iodine- 131- phy in 42 patients with neuroblastoma.Comparisonwas MIBG has also been used in treating patients whose tumors madewith ‘311-metaiodobenzylguanidine(MIBG),99mTcmeth. were refractory to (5—6).Abnormalities ylenediphosphonate(MDP)bonescans,as wellas computed demonstrated by [‘31I]MIBGscintigraphy correlated well axial tomography (CT) or magnetic resonance imaging (MRI). with clinical disease status, as well as assessments by means Iodine-131-3F8detected more abnormal sites (283)than [1311] of other imaging modalities and tumor markers ( 7—8).As MIBG (138) or @‘“Tc-MDP(69), especiallyin patients with the number of patients studied by [‘3I]MIBG increased, extensive disease. In 20 patients with soft-tissuetumors demonstratedby CT/MRI, 131I-3F8detected the disease in tumors were found that had escaped detection (9—10). 18. Uponsurgicalresection,two tumorsinterpretedas neg Moreover, the percentage [‘311]MIBGinjected dose per ative with 131I-3F8imaging revealed ganglioneuroma, one gram (%ID/g) in tumors was low(averaging 0.002%) (11). showingmicroscopicfociof neuroblastoma.In contrast,131l Tumor heterogeneity and inadequate tumor uptake could 3F8 imagingidentifiedtumorsthat were confirmedhistologi limit the clinical efficacy of [‘31I]MIBGin neuroblastoma. cally as . In 26 patients with evidence of Improved methods of tumor localization are needed. Ra marrowdiseaseby antibodyscans, 14/26 had confirmation dioimmunoscintigraphy offers this potential. by iliaccrest marrowaspirate/biopsyexaminations.We con Several monoclonal antibodies against neuroblastoma eludethat 131l-3F8scintigraphyhas clinicalutilityin the man have been tested for diagnosis and therapy (12—14)and agement of patients with neuroblastomaby improvingthe early results appear promising. The antibody we have sensitivityof tumordetection. utilized is 3F8, a murine IgG3specific for disialoganglioside J NucI Med 1991; 32:769—776 GD2 (15). Previous studies have shown that 0D2 is present in high concentrations at about 5—10x 106molecules per cell in human neuroblastomas (16). 3F8 recognizes most neuroblastoma cell lines and reacts intensely with all fresh euroblastoma is the third most common solid tumor human neuroblastoma specimens by immunofluores in childhood accounting for 7%—l4% of the malignancy cence; 3F8 does not react with most adult and fetal tissues in this age group. It affects one in 10,000 children with a (15). In nude mice xenografted with human neuro peak incidence at 2 yr ofage. Most patients have extensive blastoma, ‘@‘Ior ‘25I-labeled3F8 localized in the tumors disease at diagnosis. Dose-intensive chemotherapy and with high uptake (8%—50%ID/g) at 24 hr (17). Sufficient radiotherapy have improved tumor response and patient radiation dose could be delivered by i.v. ‘311-3F8to eradi survival (1). However, drug resistance invariably thwarts cate xenografts completely (18). In preliminary patient attempts to eradicate this disease. Curability of neuro studies (19), ‘3I-3F8localization ranged from 0.04% to blastoma depends on early and accurate detection. Radio 0. 11% ID/g of tumor at 24 hr and was not affected by graphic methods currently available are inadequate in circulating serum GD2. In addition, 3F8 could activate human complement to kill neuroblastoma cells and me diate tumor cytotoxicity in the presence of lymphocytes, Received Jun. 11, 1990; revision accepted Oct. 1, 1990. neutrophils, and activated monocytes (20—22).3F8 also For reprints contact: Dr. Samuel D.J. Yeh, Department of Medical Imaging or Dr.Nai-Kongv. Cheung, DepartmentofPediatncs, MemorialSloan-Kettering has been used for purging tumor cells in bone marrow Cancer Center, 1275 York Ave., New York, NY 10021. prior to cryopreservation and transplantation (23). When

Iodine-i31-3F8 Scintigraphyof Neuroblastoma•Yehet al 769 injected intravenously, it has induced prolonged remis (19). Specific activity of ‘@‘Iwas 6—12mCi/ag iodide. Radiola sions in some patients who failed conventional chemo beledantibody3F8 must have>50% bindingby in vitroantigen therapy (24). The purpose of this study is to test whether binding assay, >95% TCA precipitable and <3% free iodine by ‘3I-3F8radioimmunoscintigraphy can provide superior radio-thinlayerchromatography.Periodictestingof radiolabeled antibodieswasperformedto ensuresterilityaswellasthe absence sensitivity and specificity in the detection of metastatic of pyrogen. neuroblastoma, when compared to [‘311]MIBG,techne Patients were given a saturated solution of potassium iodide tium-labeled methylene diphosphonate bone scan (99mTc (SSK1)orallyfor 7—10daysto block ‘@‘Iuptake by their thyroid MDP), computed axial tomography (D'), magnetic reso glands.Followinga negativeskin test with intradermalinjection nance imaging (MR!), tumor biopsies, as well as standard of 1 @igofunlabeled 3F8, 311-labeled3F8 (0.5 mg ofprotein with bone marrow examinations. 2 mCi of ‘@‘I)was given over a period of 20 mm. Immediately upon the completion ofthe infusion, anterior and posterior total bodycounts weremeasuredusingan uncollimatedgamma cam MATERIALS AND METhODS era with the window set for ‘@‘Iand positioned at a distance of Fifty-four radioimmunoscintigraphic studies using ‘31I-3F8 10feet;thiswassimilarto our routinedosimetricprocedurefor wereperformedin 42 consecutivepatients with either StageIII patients with thyroid cancer. Daily measurements were made. unresectableor StageIV neuroblastoma.Allofthese patientshad Serial blood samples were obtained 10 mm, 24, 48, and 72 hr previous surgery and chemotherapy. They were referred to Me after injection to measure blood clearance. Anterior and posterior morial Sloan-Kettering Cancer Center for either bone marrow total-body images were obtained at 24, 48, and sometimes 120 transplantation, continued chemotherapy, or other treatment hr on a Gemini 700 (General Electric, Milwaukee, WI) gamma modalities. Pathologic confirmation ofthe diagnosis was obtained camera with a high-energycollimator.Whenevernecessary,lat in each patient. Clinical staging was carried out as part of this cml views as well as multiple spot views of special interest were study. It includedbilateralposterior iliac crest marrow biopsies also obtained. Total-body images were obtained at a scanning and four aspirates (anterior and posterior left and right iliac speedof 15—20cpm forMDPand 6—8cpm forantibodyimaging, crests), CT or MRI ofprevious or suspected sites ofdisease, bone depending on the count rates at the sternum (bone scan) and scans, as well as 24-hr urine collections for catecholamines, liver (antibody and MIBG). Spot views were acquired at 10 mm vanillylmandeicacid(VMA),and serumLDH.Completeclinical per view. Digital data were stored at 256 x 256 matrix size. remission (CR) is defined as no evidence ofdisease radiologically, Iodine-l3l-MIBG was purchasedfrom the Nuclear Pharmacy biochemically, as well as by biopsy. In this group ofCR patients, of the University of Michigan, Ann Arbor, MI. The thyroids of bonescanswerenormalexceptforfourpatientswhohad isolated the patients imaged with [‘311]MIBGwere protected by a saturated areas of faint uptake that improved or normalized on follow-up solution ofpotassium iodide administered orally(five drops daily) bone scans.They also had repeatedlynegativeradiographicand for 5 days. The [‘3I]MIBGdose was 1mCi/l.7 m2.The first four marrow examinations.This study was approvedby the Institu patients, however, received a lower dose of 0.5 mCi/l .7 m2 tional Review Board and informed consents were signed by the becauseof Internal ReviewBoardprotocolrequirements.Total patients or their parents. There were 21 males and 21 females body or multiple spot imagesof the entire body were obtained with an age range from 1 to 24 yr [a majority of patients were 24 and 48 hr after the administrationof the tracer. Iodine-l3l- between2—8yr (seeTable 1)].There weretwelvepatients older MIBG images were obtained within 1 wk prior to the antibody than 6, and two over 20 yr of age.Clinically,their diseasestatus study in 19 patients and within 1 mo in 22 patients (mean 26 could be divided into four groups: days,median 21 days). Technetium-99m-MDP was used for bone scanning. The dose A. Progressivedisease on chemotherapy with positive bone was 25 mCi/70 kg. Images were obtained on the Omega 500 marrowexaminations. (General Electric), Gemini 700 (General Electric), or Genesis B. Progressivediseaseon chemotherapywith negativebone (ADAC)2—3hr after injection.Spot viewswereobtainedwhen marrowexaminations. ever necessary. Seventeen patients had bone scans performed less C. Activebut stablediseaseon chemotherapy. than 10 days priorto the antibody study (mean 4.6 days, median D. Completeclinicalremission. 5 days), and 25 patients had bone scans within 30 days (mean 10.4 days, median 8 days) of the study. The number of abnormal sites of uptake was The monoclonalantibody 3F8 was manufactured under BB estimated. Quantitative interpretation was sometimes difficult IND 2323 and radiolabeled under BB IND 2299 at Memorial becauseof the contiguousnature of the lesions.The number of Sloan-Kettering Cancer Center according to guidelines of the lesions should be considered as best estimates for comparison Officeof BiologicsResearchand ReviewCenter for Drugs and purposes. Biologics,Foodand DrugAdministration.For qualityassurance, hybridoma 3F8 was found to be negative for adventitious agents by MAP, S@L, and XC plaque assays, as well as negative for RESULTS reversetranscriptase.MAPtestingincludedscreeningfor murine leukovirus, LCM virus isolation by intracerebral inoculation, Validation of 131I-3F8Scintigraphy by Histology murine saliva gland virus, mouse thymic virus, EDIM, and LDH Surgical resection ofthe soft-tissue tumor(s) was carried virus isolations.Purified antibody 3F8 had to pass MAP and out in nine patients after imaging with ‘311-3F8.Seven sterilitytesting(bacteria,mycoplasma,and fungalcultures),rab were antibody scan-positive and all were confirmed by the bit pyrogen testing, as well as safety testing in mice and guinea Department of Pathology as neuroblastomas. Two were pigs.3F8wasthen conjugatedto ‘@‘Ibythechloramine-Tmethod ‘311-3F8negative and were diagnosed as ganglioneuromas;

770 The Journalof NuclearMedicine•Vol. 32 •No. 5 •May 1991 one of these tumors had microscopic foci of neuro TABLE 1 blastoma. Iodine- 13 l-3F8 imaging revealed hot spots Summary of Clinical Status and Comparison of Imaging Results suggestive of marrow involvement (corresponding sites were negative on bone scan or x-rays) in 26 patients. Iliac Numberabnormalradionuclideof uptakePt. crest bone marrow aspirates and biopsies confirmed mar row disease in 14. Many of the abnormal marrow sites marrow detected by 3F8 imaging were scattered and did not involve no. AgesitesSex 3F8 MIBG MDPBone exam the iliac crest. Thus, random iliac crest marrow samplings GroupA: Progressivediseaseon chemotherapywith positive might have missed these tumors. marrow 1 9 M 30 10 2 + Correlation between 1311-3F8Imaging and Other 2 24 F 28 7 12 + Imaging Modalities 3 4 M 20 8 6 + The clinical status and imaging findings in the 42 pa 4 15 F 19 10 5 + tients were summarized in Table 1. Nineteen patients 5 6 F 15 21 0 + 6 14 M 15 10 2 + (Groups A and B) had progressive disease. Thirteen of 7 5 M 13 10 10 + them had bone marrow aspirates and/or biopsies showing 8 4 M 11 12 4 + involvement by neuroblastoma (Patients 1—13).Iodine 9 4 M 11 3 2 + 13 1-3F8 detected 177 sites of abnormal radionuclide up 10 2 F 6 1 0 + take, [‘311]MIBG92, and 99mTc..MDP 44. One patient (No. 11 2 M 5 0 1 + 12 4 M 4 0 0 + 13) with microscopic recurrence of marrow disease (a few 13 7 M 0 0 0 + scattered clumps in a high power field of the marrow smear) was antibody imaging-negative. Six patients had GroupB: Progressivediseaseon chemotherapywith negativemar progressive disease on chemotherapy but without histo row144F30124—154F9notdone4—165M833—173F720—187F523—197M410— logic evidence of marrow involvement (Patients 14—19). Five of these patients were imaged with all three modali ties. The number ofabnormal sites when compared totaled 54 (‘3I-3F8),20 ([‘3I]MIBG)and 10 (99mTc@MDP), spectively. Seven patients had active but stable disease (Patients 20—26,Group C); one patient had positive bone GroupC: Activebut stablediseaseon chemotherapy marrow biopsy. Thirty-six abnormal sites were detected 20 4 F 11 8 3 + by ‘31I-3F8,22 by [‘311]MIBG,and 8 by 99mTc@MDP. 21 5 M 8 6 0 — Patient 26 had an isolated retroperitoneal ganglioneuroma 22 24 F 6 1 2 — with scattered foci of neuroblastoma. In 16 patients (Nos. 23 5 F 4 2 1 — 27—42,Group D) who were in clinical remission, five had 24 5 F 4 2 0 — 25 3 F 3 3 2 — abnormal radionuclide uptake by ‘3I-3F8scintigraphy, 26 8 M 0 0 0 — although fewer abnormal sites per patient were found, when compared to patients in Groups A—C.Even fewer GroupD:Completeclinicalremission hot spots were noted in [‘311]MIBGstudies. Four patients 27 2 M 6 1 0 in Group D had slightly abnormal bone scans. The iliac 28 2 F 7 2 0 29 6 M 1 1 0 crest uptake of 99mTcMDp in Patients 30 and 34 were 30 4 M 1 0 1 attributable to postoperative changes related to bone mar 31 5 F 1 0 0 row biopsies and harvests. Patient 33 had minimal increase 32 15 M 0 0 3 in the left acetabulum without interval changes over a 33 3 M 0 0 2 period of 1 yr and repeatedly had negative x-rays studies. 34 3 M 0 0 1 35 5 M 0 0 0 Patient 32 had three faint areas of uptake in the thoracic 36 3 F 0 0 0 spine and negative radiographic studies. After therapy, 37 6 M 0 0 0 these patients (Nos. 30, 32, 33, 34) showed continuous 38 2 F 0 0 0 improvement until all bone scans became normal after 2 39 6 F 0 0 0 yr. 40 3 F 0 0 0 41 3 F 0 0 0 Comparison ofthese imaging modalities among patients 42 2 M 0 0 0 grouped by disease status is summarized in Table 2. Most ofthe abnormalities were seen in the bone or bone marrow. Although ‘311-3F8detected more sites than [‘311]MIBGin individual subjects, the proportion of patients who were patients had obvious disease at the time of the study and scan-positive (or -negative) showed excellent overall agree both died within 6 mo. The other two patients developed ment (Table 3). Abnormalities of ‘311-3F8images were recurrent marrow disease at 8 and 16 mo follow-up, re seen in four patients who had negative MIBG studies. Two spectively.

Iodine-131-3F8Scintigraphy of Neuroblastoma •Yeh et al 771 TABLE 2 TABLE 4 Comparison of Imaging Modalities Evaluation of Chest/Abdomen/Pelvic Soft-Tissue Masses: Correlation Between 131l-3F8Imaging and CT/MRI in 42 Averagenumberof lesionsdetected Patients perpatient CT/MRI ClinicalStatus 131l-3F8 [131IJMIBG @“Tc-MDP + @ 131I3F8 18 1* A. Progressivedeease 13.6±9.0* 7.1 ±6.2 3.4 ±3.9 - 2t 21 withpositiveBM(1- 1W 92 44 13)t B. Progressivedeease 10.8 ±10.8 4.0 ±4.5 2.0 ±1.9 Likelihoodof agreement= 93%±8%(95%confidencelimits).

withnegativeBM 54 20 10 * This patient had a “hot spot― on [131l]MIBG in the presacral soft (14,16—19) tissueonrepeatedstudies.Both131l-3F8,bonescan,andlaparatomy C.Activebutstable 5.1±1.23622813.63.1 ±2.91.1± failed to reveal any evidenceof tumor. disease(20—26) t False-negative: first patient had abdominal ganglioneuroma and D. Complete remission .0±0.89k16472.20.25 ±0.580.44± scatteredfociof neuroblastomaSecondpatienthadganglioneuroma. (27—42)

a Mean ± s.d.; BM = bone marrow test. 3 (Fig. 1), illustrating the superiority of antibody over t Numbers in parentheses correspond to the patient numbers in other imaging modalities. Figure 2 illustrates widespread Table1. bone or bone marrow disease, as well as right supraclavi * Total number of abnormal sites of uptake. cular, infraclavicular, and axillary lymph node involve . Abnormal @“9c-MDPuptake disappearedin two patients without treatment.Onepatienthad increased @“Tc-MDPuptakeafter iliac ments in Patient 1. Intense uptake of ‘31I-3F8and [‘@‘I] crest biopsy and anotherone at the surgicalsite after marrow MIBG in the liver of Patient 7 corresponded to the region harvesting. of the tumor defined by the abdominal CT (Fig. 3). In volvement of the lower portion of the liver and multiple areas of spine and right femur was found in Patient 10 Correlation between 1311-3F8and CT/MRI (Fig. 4). The [‘31I]MIBGlocalization in the liver tumor Because ofthe extensive abnormalities noted in ‘311-3F8 was poor. A large retrothoracic and retroperitoneal tumor and [‘31IJMIBGscintigraphies, it was not feasible to obtain was demonstrated in Patient 23 (Fig. 5). independent confirmation of all lesions. However, it was The body retention, plasma radioactivity, and plasma possible to correlate imaging information at soft-tissue sites half-lives of ‘311-3F8in patients with either extensive dis in chest, abdomen and pelvis with conventional CT or ease (8 patients from Group A) or patients with minimal MM. As shown in Table 4, the likelihood of agreement abnormalities (16 patients from Group D) were compared was 93% ±8%. Two patients had negative ‘311-3F8scans in Figure 6. Patients with more widespread abnormalities despite the presence of tumor. When analyzed histologi by antibody imaging had both faster body and blood cally, one showed ganglioneuroma and the other scattered clearance. No adverse side effects were observed. foci of neuroblastoma. Iodine-13 1-3F8 imaging was repeated in 12 patients, Abnormal radionucide uptake in the ‘311-3F8images ranging from 2 wk to 1yr after their first ‘311-3F8exposure. was clearly evident at 24—48hr following injection of the None of them had neutralizing human antimouse anti labeled antibody. Since the target-to-non-target ratio was bodies (HAMA) or anaphylactoid antibodies. One patient sufficiently high, no additional background subtraction developed a transitory episode of hypotension and tran was necessary to provide clear delineation of lesions. sient loss of consciousness that responded to hydration Rarely were liver and spleen imaged and the occurrence without neurologic sequelae. Nine patients showed interim was only during the immediate postinjection period when disease progression with excellent correlation of ‘31I-3F8 blood radioactivity was high. Iodine-13l-3F8 scans showed imaging with known sites ofdisease. One patient had rapid widespread bone and bone marrow involvement in Patient clearance of ‘311-3F8and no images could be obtained. Two patients continued to be free of metastatic disease by TABLE 3 ‘311-3F8scintigraphy and remained alive without evidence Correlation Between 1311-3F8and [131l]MIBG* oftumor relapse. Imaging in 41 Patients [‘31l]MIBG DISCUSSION + @ 13113F8 24 4 Our study demonstrated that ‘311-3F8could detect pri 0 13 mary and metastatic neuroblastomas with excellent sensi tivity and specificity. Histologic examination of ‘31I-3F8- Ukelihood of agreement= 90% ±9% (95% confidence limits). positive sites confirmed the presence of neuroblastoma. * Iodine-131-MIBG images were obtained within 1 wk prior to the Iodine-13l-3F8 was found to detect more abnormal pri antibodystudyin 19 patientsandwithin1 mo in 22 patients(mean mary and metastatic sites than either [‘311]MIBGor 99mTc@ 26 days,median21 days). MDP bone scans. The correlation between ‘311-3F8and

772 The Journal of Nudear Medicine •Vol. 32 •No. 5 •May1991 @ @‘ . - ‘

MIBG

L R

MOP 3F8 FIGURE 1. Posteriorviews of Patient 3: 1311-3F8imaging showed widespread abnormal uptake in the calvarium, spine, ribs, pelvis and all four extremities. Free iodine was trapped by the stomach and was also excreted in the urine (fillingthe bladder).MIBGlocalizedin the samegen @ R L eral areas. Parotid, cardiac, and hepatic uptakeof [1311]MIBGwere seen.However, owing to the adjacentnormalliveractivity, spine abnormalitieswere not well deline A i@ ated. The @‘Tc-MDPbone scan showed increased uptake in the right hip, left shoul der,andrightposterior8th rib.Radioactive markers in the images (marked by arrows) were placed on the right side (R) of the patient.L = left sideof the patient.

CT/MRI for soft-tissue abnormalities was excellent. lo Two other monoclonal antibodies against neuro dine-l31-3F8 may provide a more sensitive diagnostic test blastoma have been used to define tumor deposits. Iodine for the detection of metastatic marrow disease than ran 13 l-UJ 13A was found to localize in 15 sites of primary dom marrow aspirates or biopsies. and metastatic lesions from 9 patients with neuroblastoma.

3F8 MDP

AP .,i. ‘.

R R@ - R

I 4 h@ ø@4 ,. , 0 *4 4 o 4 , p , FIGURE 2. Patient 1: 131I-3F8uptake MIBG was evidentin the calvanum,right axillary, both shoulders,spine,pelvis,and extrem ities.MIBGscan(posteriorviews)showed increased uptake in the right humerus, right clavicle, right axilla, spine and ex L .:ç. , S R tremities, with less tumor versus normal tissuecontrast.The @“Tc-MDPbonescan showed increased uptake in the shafts of both femurs. A = anteriorview, P = pos tenor view, A = right side of patient, and L = left sideof patient.

Iodine-i 31-3F8 Scintigraphy of Neuroblastoma •Yeh et al 773 @ ;@•@:.

therapy dose exceeded 2.8 mCi/kg. Objective responses 48h were obtained in one of the four patients. The use of another IgG1monoclonal antibody BW 575/9 was recently reported. Good localization was seen in two out of three patients (14). Comparison of 99mTc@BW575/9 and [123!] MIBG was made in seven children with Stage IV neuro blastoma. The majority of tumor sites was detected by 4 both modalities. However, 5 of 26 lesions were undetect able by [‘231]MIBGand 8 were not detected by immuno scintigraphy (27). It was concluded that there was a good correlation between these two modalities, and that their 3F8 combined use could improve the detectability of metastatic disease (26,27). Similar to our previous report of ‘311-3F8imaging (19), FIGURE3. Posteriorviewsof1311-3F8and[1311]MIBGscansthis study was carried out in a group of patients who had of Patient7 showedintenseuptakein the superiorposteriorarea of the liver. A large tumor mass was evident in the CT scan. been previously treated with chemotherapy, radiotherapy, Spine uptake in the vertebral column was seen in the 131l-3F8 and/or surgery. A great deal more bone marrow (or bone) imagesbut was less intensein the [1311]MIBGscan. disease was seen in this population than expected on clinical grounds. Since the bone marrow abnormalities were numerous, confirmation of the presence of tumor in One false-positive and two false-negative localizations were all areas of increased uptake was not practical. Unless reported (13). This IgG1antibody was also used for target marrow disease was diffuse, negative iliac crest marrow ing ‘@‘Itherapy in four patients who failed other conven examinations could be due to sampling error. Conversely, tional treatments (25). Liver uptake was seen in every ifmarrow disease was scant, positive 1311-3F8images would patient, and bone marrow toxicity was noted when the be unlikely. Detection of minimal disease in bone marrow

3F8 MIBG

•1 @ R@ ‘. *T@ S PO$T. A p

@;, A p .‘j'@ MDP

FIGURE 4. Patient10: 131l-3F8imaging showed increased uptake in the lower per

@ tion of the right lobe of the liver, multiple 0@ • • • areas in the mid- and lower thoracic spine, and right proximal and distal femurs. The large hepatic lesion was also found in CT, R S@ R MRI, and @‘Tc-sulfurcolloid liver scan ;@,. (data not shown).The [1311]MIBGshowed increaseduptake in the right femur, left . . * iliacbone,andright renalfossabut not the . . . 4 large tumor behind the liver. The MDP bonescanappearednormal.Anteriorview . , ¶. , (A)is on the leftand posteriorview(P) is on the right. A P

774 The Journal of Nuclear Medicine •Vol. 32 •No. 5 •May 1991 3F8 MDP

@ A P 4 A P

R

*

FIGURE 5. Patient23: 1311-3F8imaging showed a tumor mass in the posterior thoraxextendingintothe retroperitoneum. Irregularuptake in the spine and femurs was also noted. MIBG imaging showed left superiormediastinal,retrothoracicand retropentoneal abnormalities. The @“Tc MDPbonescanappearednormalin these areas.A = anteriorview and P = posterior view.

that was histologically disease-free (by Wright-Giemsa and of measurement was the individual patient (Table 3), a Hematoxylin-Eosin stains) has been previously reported detailed comparison of individual lesions detected by using GD2 antibodies (i.e., 3F8) (28), as well as UJ 13A either modality suggested that ‘311-3F8detected signifi (29) and BW 575/9 (30). Immunostaining of marrow cantly more abnormalities than [1311]MIBG, especially for specimens in patients with positive ‘31I-3F8scans have metastatic sites in marrow (or bone) (Table 2). More confirmed marrow disease in selected cases (unpublished lesions were evident because there was either more tumor data). heterogeneity with respect to the uptake of [‘31I]MIBG,or In order to compare the sensitivity of radiolabeled 3F8 that the tumor-to-non-tumor ratio was superior for ‘@‘I and MIBG in the detection of metastatic neuroblastoma, 3F8. For soft-tissue abnormalities, chest and abdominal we chose ‘@‘Iconjugates because they were the common CT or MRI showed strong correlation with antibody im conditions of usage. Although the likelihood of agreement aging (Table 4) and histologic confirmation of hot spots between 13I-3F8 and [‘311]MIBGwas 90% when the unit was obtained in seven patients. As we have previously reported (19), 3F8 did not localize to ganglioneuroma. In addition, microscopic neuroblastoma foci also escaped Total Body Retention Blood Radioactivity detection (19). The nature of increased uptake in a few

Percent Radioactivity patients who had apparent clinical remissions may become 1oo@- clear with longer follow-up. Nine patients who had nega

8OI@ @@dmid tive or focally positive initial antibody studies, showed multiple sites of abnormalities (either new or with in 6O@ creased intensity) during the subsequent imaging studies

40 when their clinical status deteriorated. In the past few years, a large number of patients with @ 2: :@ neuroblastoma have been imaged with 1311..or [‘23I]MIBG. Although excellent sensitivity and specificity of MIBG in 0 hr 24 hr 48 hr 120 hr 0 hr 24 hr 48 hr 120 hr Time after t-131-3F8 tnjection Time after t-131-3F8 tnjection detecting neuroblastoma were reported in the early studies FIGURE 6. Clearanceof 131I-3F8.Both total-bodyretention (2—4),subsequentexperiencehasdemonstratedsome fail (left)and plasmaradioactivity(right)were expressedas percent ures in patients with active disease (9). Recently, it was agesof the initialcounts at time 0 hour after antibodyinjection. reported that [‘23I]MIBGunderestimated the skeletal in

Iodine-i 31-3F8 Scintigraphy of Neuroblastoma •Yeh et al 775 volvement in neuroblastoma. Four patients who had neg relation with tumor histopathology. J NuclMed 1989;30:474—480. 12. Cheung NKV, Saarinen U, Neely J, et al. Development of neuroblastoma ative [‘231]MIBGstudies later died ofdisseminated disease antibodies for potential utilization in diagnosis and therapy. In: Evans AE, (31). It is likely that MIBG may not be able to detect all D'Angrio GJ, SeegerRC, eds.Advancesin neuroblastomaresearch.New tumor sites in neuroblastoma, even when [‘231]MIBGor a York: Alan R. Liss; 1985:501—505. 13. Goldman A, Vivian G, Gordon I, et al. Immunolocalization of neuro larger dose of [‘31IJMIBGis used. Iodine-l31-MIBG ap blastoma using radiolabeled UJ 13A. J Ped pears to be effective in the treatment of malignant pheo 1984;105:252—256. chromocytoma (32). The results of [‘311]MIBGtherapy in 14. Baum RP, Maul FD, Schwarz A, et al. Diagnosis and treatment of stage IV neuroblastoma with Tc-99m- and 1-131-labeled monoclonal antibody advanced neuroblastoma has been less encouraging (5-6). BW 575/9. JNuclMed 1989;30:904. Since ‘31I-3F8has a higher tumor uptake (0.08% ID/g) 15. Cheung NKV, Saarinen UM, Neely JE, et a!. Monoclonal antibodies to a than [‘311]MIBG(0.002%) (11) and appears to have less glycolipid antigen on human neuroblastoma cells. Cancer Res 1985; 45:2642—2649. heterogeneity sites, this antibody is currently used as our 16. Wu ZL, Schwartz E, Seeger R, et al. Expression of GD2 ganglioside by standard for staging patients with advanced neuro untreated primary human neuroblastomas. Cancer Res 1986;46:440—443. blastoma. In addition, ‘311-3F8with its relative low back 17. Cheung NKV, Neely JE, Landmeier B, et al. Targeting ofganglioside GD2 monoclonal antibody to neuroblastoma. J NuclMed 1987;28:1577—1583. ground activity in normal tissues has been utilized in 18.CheungNKV,LandmeierB,NeelyJ, etal.Completetumorablationwith targeted radioimmunotherapy of neuroblastoma (33). iodine-131-radiolabeled disialoganglioside GD2 specific monoclonal anti body against human neuroblastoma xenografted in nude mice. J Nail Cancer Inst 1986;77:739—745. ACKNOWLEDGMENTS 19. Miraldi FD, Nelson AD, Kraly C, et al. Diagnostic imaging of human The authors want to thank Dr. Irene Cheung for her critical neuroblastoma with radiolabeled antibody. Radiology 1986;l61:413—418. 20. Munn DH, Garnick M, Cheung NKV. Effects of parenteral recombinant review and editing ofthe manuscript; Mr. Joon Kim, Ms. Heather human macrophage colony-stimulating factor on monocyte number, phe James, and Ms. Elsa Seguritan for their technical assistance in notype, and antitumor cytotoxicity in nonhuman primates. Blood antibodyiodinationand assays;Mr. Finn Augensen,Mr. Upen 1990;75:2042—2048. dra Bhatt, Mr. BensonEgbuta,Mr. Kevin Kavanagh,Mr. Mat 21. KushnerBH, CheungNKV. GM-CSFenhances 3F8 monoclonal antibody dependent cellular cytotoxicity against human and neuro thew Senatore, Ms. Molikuity Thomas, and Ms. Virrel Williams blastoma.Blood1989;73:l936—1941. for their technical assistance in patient scanning; and Ms. Avery 22. Munn DH, Cheung NKV. Phagocytosis of tumor cells by human mono Whitefortypingthe manuscript. cytes cultured in recombinant macrophage colony stimulating factor. J Supported in part by grant CA33049 from the National Insti Exp Med 1990;172:231—237. tutes of Health. 23. Kushner BH, Gulati SC, Cheung NKV. Effective purging of neuroblastoma from bone marrow (BM) using IgG3 monoclonal antibody (MoAb) that mediate human complement-dependent as well as cell-mediated cytotox REFERENCES icities [Abstract]. Blood 1988;72(suppl 1):394a. 24. Cheung NKV, Burch L, Kushner BH, et al. Monoclonal antibody 3F8 can 1. Kushner BH, Cheung NKV. Treatment of neuroblastoma. In: DeVita V, effect durable remissions in neuroblastoma (NB) patients (pta) refmctory Rosenberg S, Heilman S, eds. Cancer principles and practice of oncology: to chemotherapy: a Phase II trial. Proc Am Assoc Can Res l990;3l:288. updates,volume2. Philadelphia:Lippincott, 1988:1—11. 25. Kemshead iT, Goldman A, Jones D, et al. Therapeutic application of 2. MunknerT. I-131-meta-iodobenzylguanidinescintigraphyof neuroblas radiolabelled monoclonal antibody UJ13A in children with disseminated tomas. SeminNuclMed 1985;15:154—160. neuroblastoma—a Phase I study. In: Evans AE, D'Angrio GJ, Seeger RC, 3. Hadley GP, Rabe E. Scanning with iodine-131-MIBG in children with eds. Advances in neuroblasioma research. New York: Alan R. Liss; solid tumors. An initial appraisal. JNuclMed 1986;27:620—626. 1985:533—544. 4. Feine U. Metaiodobenzylguanidine in the diagnosis ofneuroblastomas. In: 26. Berthold F, Waters W, Sieverts H, et al. Immunoscintigraphic imaging of Winkler C, ed. Nuclearmedicinein clinicaloncology.Berlin,Heidelberg: MIBG-negative metastases in neuroblastoma. Am J Ped Hemat Oncol Springer-Verlag, 1986:122—128. l990;12:61—62. 5. VoutePA, HoefnagelCA, de KrakerJ. I-131-meta-iodobenzylguanidine27. Smolarz K, Waters W, Sieverts H, et al. Immunoscintigraphy with Tc in diagnosis and treatment of neuroblastoma. Bull Cancer 1988;75:107— 99m-labeled monoclonal antibody BW575 compared with I-123-MIBG 111. scintigraphy in neuroblastoma [Abstract]. Radiology 1989;173: 152a. 6. Beierwaltes WH. Application of I-131-meta-iodobenzylguanidine. NucI 28. Cheung NKV, Von Hoff DD, Strandjord S, et a!. Detection of neuro Med Biol1987;14:183—l89. blastoma cells in bone marrow using GD2specific monoclonal antibodies. 7. Geatti 0, Shapiro B, SissonJ, et a!. Iodine-131-metaiodobenzylguanidine J Clin Oncol l986;4:363—369. scintigraphy for the localization of neuroblastoma: preliminary experience 29. Allan PM, Garson JA, Harper E, et al. Biological characterization and in ten cases. JNuclMed 1985;26:736—742. clinical applications of a monoclonal antibody recognizing an antigen 8. Yeh 5DJ, Helson L, Benua RS. Correlation between iodine-131-MIBG restricted to neuroectodermal tissues. mi J Cancer l983;3 1:59 1—598. imaging and biological markers in advanced neuroblastoma. Clin Nucl 30. Berthold F, Schneider A, Schumacher R, et al. Detection of minimal Med 1988;13:46—52. disease in bone marrow ofneuroblastoma patients by immunofluorescence. 9. Bouvier JF, Philip T, Chauvot P, et al. Pitfalls and solutions in neuro PediairHemaiol Oncol 1989;6:73—83. blastoma diagnosis using radioiodine MIBG. Our experience about 50 31. Gordon I, Peters AM, Gutman A, et al. Skeletal assessment in neuro cases. In: Evans AE, D'Angrio GJ, Knudson AG, Seeger RC, eds. Advances blastoma—the pitfalls of iodine-123-MIBG scans. J NucI Med 1990; in neuroblastoma research. New York: Alan R. Liss; 1988:707—720. 31: 129—134. 10. Schmiegelow K, Simes MA, Agertoft L, et al. Radio-iodobenzylguanidine 32. Sisson JC, Shapiro B, Beierwaltes WH, et al. Radiopharmaceutical treat scintigraphy of neuroblastoma: conflicting results, when compared with ment ofmalignant pheochromocytoma. J NuclMed l984;25:197—206. standard investigation. MedPed Oncol 1989;17:126—130. 33. Cheung NKV, Yeh SDJ, Gulati 5, et al. I-i 3l-3F8 targeted radiotherapy 11. Moyes J5, Babich JW, Caner R, et al. Quantitative study of radioiodinated of neuroblastoma (NB): a phase I clinical trial. Proc Am Assoc Can Res metaiodobenzylguanidine uptake in children with neuroblastoma: cor 1990;31:284.

776 The Journal of Nuclear Medicine •Vol. 32 •No. 5 •May 1991