Localization of Malignant Melanoma Using Monoclonal Antibodies Joseph Wasselle, MD; Jeanne Becker, PhD; C. Wayne Cruse, MD; Carmen Espinosa, MD; Charles Cox, MD; Douglas Reintgen, MD \s=b\ Finding a screening test to evaluate patients with cancer for occult metastatic disease, as well as imaging all known disease, is a goal of research efforts. Twenty-nine evaluable patients with deeply invasive (stage I), regional nodal (stage II), or systemic (stage III) melanoma underwent imaging by administration of a preparation of the antimelanoma antibody labeled with technetium 99m. Scan results indicated that 28 of 32 confirmed metastatic sites were imaged with this technique (88% sensitivity). Analysis of the individual positive sites revealed that nodal basins and visceral metastases accounted for the highest percentage of metastatic sites imaged, with 14 (88%) of 16 nodal basin metastases and all four visceral metastases being detected through imaging. Occult nodal disease was detected in the iliac nodal chain in two of the 29 patients. The imaging of benign tumors and nodal basins not containing disease accounted for a confirmed false-positive rate of 21%. Three (10%) of the 29 scan results were confirmed to be false-negative. In vivo tumor localization with monoclonal antibodies showed a sensitivity similar to that of other roentgenographic procedures for identifying metastatic disease and was useful in two of three patients in identifying occult iliac nodal disease, a region that is difficult to evaluate with physical examination and other imaging modalities. (Arch Surg. 1991 ;126:481-484)
Melanoma
has an unpredictable malignant potential. A significant incidence of recurrent metastatic disease is seen within 5 years in stage I and II melanoma. Recurrent disease is predictable for deep primary melanomas (stage I), for patients with regional lymph nodes containing metastatic tumor (stage II), and for patients with isolated visceral mé¬ tastases (stage III) after resection.1 Occult disease present at the time of surgery has been suggested as the basis of these recurrences.2 Metastatic melanoma can involve any organ but has a propensity to occur in the regional lymph nodes, lung, liver, bone, and central nervous system.3 Risk factors have been described in an attempt to predict the incidence of metastatic disease following primary tumor resection. Unfavorable prognostic factors for stage I disease include Breslow tumor thickness greater than 0.76 mm and tumor ulcération.4 Although the probability of metastasis can be assessed, no single test is currently available for confirma¬ tion of staging. The routine workup for clinical staging in¬ cludes chest roentgenography and serum liver function tests. Abnormal test results and patient symptoms are subsequent¬ ly evaluated with computed tomography (CT) of the head, chest, abdomen, and pelvis depending on the abnormality detected and the symptoms. Monoclonal antibodies (MoAbs) to tumor-associated anti¬ gens are becoming an important tool for evaluating tumor
Accepted for publication January 12,1991. From the Departments of Surgery (Drs Wasselle, Becker, Cruse, Cox, and Reintgen) and Pathology (Dr Espinosa), Melanoma Treatment Center, University of South Florida, Tampa. Read before the 43rd Annual Cancer Symposium of the Society of Surgical Oncology, Washington, DC, May 20,1990. Reprint requests to the H. Lee Moffitt Cancer Center, University of South Florida, PO Box 280179, Tampa, FL 33682-0179 (Dr Reintgen).
stage and treatment. Monoclonal antibodies against the 250-
kd cell surface glycoprotein,5 expressed on 90% of all melano¬ recently become available. These antibodies have been tagged with radioactive technetium via a bifunctional chelating agent and developed for detection of malignant melanoma.6 The use of the MoAb Fab fragment rather than whole IgG has been shown to greatly enhance image resolu¬ tion, largely through increased clearance with reduced back¬ ground interference and greater specific localization of the Fab fragment compared with whole IgG in tumors.7 The Melanoma Treatment Center at the University of South Florida, Tampa, conducted this study to evaluate the role of technetium-labeled MoAb scans in the metastatic workup and clinical treatment of patients with advancedstage melanoma and to determine the effect of long-term clinical follow-up on sensitivity results. mas, have
MATERIALS AND METHODS
MoAb· NR-Ml-05, a murine antimelanoma MoAb against the 250-kd highmolecular-weight proteoglycan antigen, was developed and purified for clinical use (NeoRx Corp, Seattle, Wash) using standard hybridoma techniques.8 Fab fragments were also prepared according to
standard methods. The irrelevant murine MoAb NR-2AD (whole IgG), reactive with idiopathic determinants of surface IgG on a B-cell lymphoma and not reactive with normal tissue or other tumors, was also supplied by NeoRx.
Preparation of Technetium 99m MoAb Conjugate An imaging kit was developed that uses a diamide dimercaptide N2S2 formed by the esterification of 4,5-bisthioacetamide pentanoate with 2,3,5,6-tetraflourophenol using carbodiimide (NeoRx Corp). The ""Tc-NgSj, complex was produced by reconstituting stannous gluconate with 1 mL of sterile water and then mixing 0.75 mL of r"Tc]pertechnetate with this solution.9
Glacial acetic acid-0.2N hydrochloric acid solution and buffers used to adjust the pH of the stannous gluconate solution to form a^Tc ligand ester. NR-Ml-05 was added to this active ester complex and incubated at room temperature for 20 minutes. The resulting were
99mTc-N2S2 antibody conjugate was then purified using ion exchange columns and micropore filters.10
Patients and Study Design
Twenty-nine patients seen at the Melanoma Treatment Center at University of South Florida, Tampa, with invasive stage I, regional nodal stage II, or systemic stage III disease were evaluated. Four of the patients had undergone no previous therapy, 25 had undergone resections of the primary tumor, and none had undergone previous chemotherapy. One patient had undergone immunotherapy using an irradiated tumor cell vaccine. All participants gave written consent to receive the radioconjugate. The histologie diagnosis of melanoma was confirmed in all patients by the clinic dermatopathologist(C.E.). The patients received three antibody preparations intravenously: 40 mg of unlabeled irrelevant murine antibody (NR-2AD), followed the
25 minutes later with 7.5 mg of an unlabeled whole antimelanoma antibody (NR-MI-05), followed 5 minutes later with 10 mg of 1110Mbq Tc-labeled Fab antimelanoma antibody (NR-Ml-05). Unla¬ beled NR-Ml-05 IgG and excess unlabeled irrelevant NR-2AD-IgG were administered before the radiolabeled MoAb Fab fragment be-
Downloaded From: http://archsurg.jamanetwork.com/ by a Michigan State University User on 06/13/2015
earlier studies11 demonstrated blocking of the reticuloendothelial system by the combination of the irrelevant MoAb and the intact immunoglobulin. This resulted in an enhanced uptake of the tracer in the tumor. All patients were observed for signs and symptoms of toxicity following MoAb injection. A laxative was administered 3 hours after injection to clear the gastrointestinal tract of radioisotope. Wholebody images were taken beginning 5 hours after injection in the anterior and posterior projections using a gamma camera. A dedi¬ cated computer was used to measure regional uptake of the radiolabeled antibody. Left and right lateral views of the head were also obtained. Occasionally, counts were repeated 24 hours later if images were questionable. All scans were reviewed and images of positive sites were then correlated with lesions found with conventional roent¬ genography and physical examinations. All patients underwent chest roentgenography, and serum sam¬ ples were obtained for liver function tests and lactic dehydrogenase determinations. Computed tomographic scans of the chest, head, abdomen, and pelvis were obtained when clinically warranted. All positive sites, determined by either MoAb scans or conventional roentgenographic studies, underwent surgical biopsies, and the spec¬ imens were evaluated by a pathologist. Specimens from clinically positive sites and lymph nodes from elective dissections were resect¬ ed and sent for routine pathologic determination. All patients were followed up after surgery. cause
Assay for Human Antimurine Antibody (HAMA) In 16 of 29 patients studied, serum samples were drawn 5,8, and 16 weeks after surgery and analyzed for HAMA response using a previ¬ ously published enzyme-linked immunosorbent assay technique.9 Titers of antibody against NR-Ml-05 were considered significantly elevated if they were more than 2 SDs above the mean normal level of serum units for the controls.
RESULTS
Optimal images were acquired at a mean of 5 hours after infusion of ""Tc-NR-Ml-05 Fab fragment. Nonspecific up¬ take was seen in the gallbladder, lung, bone marrow, kidneys, testicles, ovaries, and bladder. A total of 32 sites were identified to contain tumor from pathologic specimens, 28 of which were MoAb-scan positive (88% sensitivity). False-negative imaging occurred at a rate of 12%. The Table summarizes the overall sensitivity of confirmed metastatic lesions in various organ sites from all 29 patients. Analysis of individual sites determined to be posi¬ tive by scanning revealed that nodal basins and visceral mé¬ tastases accounted for the highest percentage of metastatic sites imaged, with 14 (88%) of 16 known lymph node métasta¬ ses and all four known visceral métastases being detected. In soft tissues, 10 (83%) of 12 sites known to contain disease were imaged. Occult disease was detected in two (7%) of 29 pa¬ tients. Six patients eventually died of their disease and dis¬ ease recurred in 10 patients. Assuming the recurrences were secondary to undetected occult disease at the time of scan¬ ning, the sensitivity falls to 66%. The false-positive rate was 21%. These scans included three nodal basins not containing tumor and three benign tumors: a lipoma, a breast adenoma, and a meningioma. Sev¬ en patients had areas of poorly defined uptake that corre¬ sponded to areas of scar formation, infection, and, in one case, a previous tumor cell vaccine and bacille Calmette-Guérin injection site. An attempt was made to determine the effect of MoAb scanning on clinical decision making. In three patients with lower-extremity melanomas, the scan showed positive tumor uptake in the iliac nodal chain. Computed tomographic scans were then performed, the results of which were negative. Dissections of these areas revealed occult micrometastatic disease in two of the three patients. Eleven patients with gross disease underwent MoAb scan imaging. Nine patients with deeply invasive tumor or micrometa-
Sensitivity of Monoclonal Antibodies Labeled With Technetium 99m Fab-NR-M1-05 for Known Melanoma Métastases
NO. (%) Of Lesions
NO. Of Known Lesions
Lesion Site
Imaged 9 (82) 1 (100) 1 (100) 1 (100) 2 (100) 12(86) 2(100) 28 (88)
11
Skin/subcutaneous Mediastinum
1
Liver
1
Brain
1
Lung Regional lymph nodes Distant lymph nodes
2 14
Total
32
2
static disease in their regional node dissections underwent MoAb scanning. None of the nine had evidence of disease either clinically or roentgenographically at the time of MoAb scanning. All nine patients had concurrent negative scan results. Although the follow-up was relatively short (mean, 11.4 months; range, 3 to 21 months), only one of the patients with previously negative scan results developed metastatic disease, making the predictive value of a completely negative scan 89%. These results need to be confirmed with a longer
follow-up.
Two cases illustrate the clinical use of MoAb scanning. A 55-year-old white man presented with palpable adenopathy in his right axilla 2 years after undergoing a wide local excision for a Clark level II melanoma on his right shoulder. During the workup, a MoAb scan with lateral (Fig 1, top) and anterior (Fig 1, center) views of the head revealed an isolated, totally asymptomatic brain mass. A CT scan (Fig 1, bottom) con¬ firmed the finding of an intracranial mass. The MoAb scan of the right axilla also showed the palpable adenopathy (Fig 2). The patient underwent a right axillary dissection and craniotomy. On pathologic evaluation, he had metastatic disease in his right axilla; however, the craniotomy revealed a meningioma.
In the second case, the MoAb scan revealed occult nodal disease. A 44-year-old white man, with a history of a Clark level IV melanoma removed from his left calf, presented with a palpable intransit lesion on the medial aspect of his left thigh and palpable left inguinal adenopathy. Results of a CT scan of the pelvis and abdomen were negative. However, a MoAb scan revealed uptake in the intransit thigh lesion (Fig 3, top), uptake in the left inguinal node region (Fig 3, bottom), and uptake in the iliac node region (Fig 3, bottom). The patient underwent a wide excision of his intransit lesion, therapeutic left groin dissection, and elective iliac nodal dissection. The pathologic examination showed metastatic disease in all three sites, with micrometastatic disease noted in the iliac nodes that were less than 1.0 cm in diameter. In this instance, the MoAb scan led to more extensive surgery that revealed occult métastases. Three (10%) of the 29
scan
results
were
confirmed to be
false-negative on pathologic examination. Two of the falsenegative scan results occurred with known abdominal-wall
métastases, and the other scan did not reveal micrometastatic
disease in lymph node basins found to be pathologically posi¬ tive after an elective lymph node resection. Four (14%) of the 29 scans had mixed results, revealing true-positive, falsepositive, and false-negative combinations at different sites. No toxicity to the MoAb infusion was observed in any patient, and no significant increase in HAMA specific for NRMl-05 was seen in the 16 patients tested.
Downloaded From: http://archsurg.jamanetwork.com/ by a Michigan State University User on 06/13/2015
2.—Anterior image of the right shoulder obtained 5 hours after intravenous administration of technetium 99m-Fab-NR-M1-05 mono¬ clonal antibodies. Note the positive tracer uptake in the axillary lymph nodes (arrow), with normal blood pool distribution seen in the lung and axillary vasculature.
Fig
Fig 1.—Top, Lateral view of the head after monoclonal antibody injection shows enhancing left meningioma. Center, Anterior view of the head shows the same lesion. Bottom, Computed tomographic scan of the head reveals a benign meningioma at the left. COMMENT The information provided by in vivo tumor localization with MoAbs suggests that "Tc-labeled MoAb (Fab fragment) is a sensitive and safe method of detecting metastatic melanoma. These scans imaged 88% of detectable melanomas and visual-
Fig 3.—Top, Scan of a medial thigh lesion (arrow) in a patient with a previously resected primary melanoma of the calf. Bottom, Scan of occult iliac nodal (wide arrow) and palpable inguinal nodal disease (thin arrow). Nonspecific uptake is seen in the testes, and excreted complex is seen in the urinary bladder.
Downloaded From: http://archsurg.jamanetwork.com/ by a Michigan State University User on 06/13/2015
ized all gross disease in 11 (50%) of 22 patients, confirming the absence of gross disease in nine (41%) of 22 patients, and detecting occult disease in two (9%) of 22 patients. In the patients with occult disease detected with MoAb scanning, further roentgenographic testing and more extensive sur¬ gery were performed based on the MoAb results. Negative scan results in a patient with deeply invasive stage I or micrometastatic stage II disease had a predictive value of 89% for no further disease development. If this holds true with longer follow-up, then MoAb scanning may be used as a screening test to initially evaluate patients with an earlystage melanoma for metastatic disease and perhaps replace the battery of screening tests used now, including liver func¬ tion tests, chest roentgenography, and CT scans when appli¬ cable. Although large prospective trials are needed to com¬ pare NR-Ml-05 scanning with more standard imaging modalities, the results may also be used as a prognostic indicator for patients and their families to confirm the absence of systemic micrometastatic disease. The MoAb scan was able to detect occult disease in two patients that was otherwise undetectable on physical exami¬ nation or routine roentgenographic examinations. Both of these occult nodal métastases developed in the iliac nodal basins from lower-extremity melanomas. It is very difficult, if not impossible, to clinically examine the iliac nodal chain. The roentgenographic procedure of choice, a CT scan of the pelvis, will distinguish between enlarged, pathologic-appearing nodes (ie, larger than 2.0 cm) and normal-sized nodes, but cannot answer the question ofwhether the nodes are involved with metastatic disease or reactive in nature. Numerous studies have shown that micrometastases usual¬ ly develop in nodes that are less than 1.0 cm in diameter12 and would not be identified as abnormal with the CT scan. In these two cases, iliac dissection was added to the standard superfi¬ cial lymph node dissection based on the results of the MoAb scan combined with the prognostic variables for the particular patient's primary melanoma. Micrometastatic disease was pathologically confirmed in two of three iliac nodal basins. The MoAb scan could help to identify the need for a more exten¬ sive ilioinguinal node dissection in patients with aggressive, high-risk primary melanomas of the lower extremity or in patients with grossly involved inguinal lymph nodes. Technetium 99m-N2S2 has previously been shown to be excreted in bile and by the kidneys, which explains the local¬ ization in the gallbladder, kidneys, and urinary bladder.6,11,13 The 21% false-positive rate seen in nodal basins and benign tumors and the nonspecific, poorly defined uptake seen in areas of inflammation may be related to increased vascular permeability or increased blood flow." Another possible ex¬ planation involves the presence of submicroscopic disease overlooked at the time of pathologic examination.15 Because of the 21% false-positive rate in this patient population, clinical judgment based on prognostic factors for melanoma must be included in the analysis of data generated by the scan to avoid unnecessary surgery. The recurrence of disease in 10 patients is most likely due to micrometastatic disease not found on initial evaluations.2 Since only one patient who had completely negative MoAb scan results had a recurrence and all other patients who had recurrences had gross disease at the time of MoAb scanning, one may postulate that clinically apparent metastatic deposits containing high tumor volumes may act as a sink for the MoAb and not allow radiolocalization to micrometastatic deposits. However, it is possible that the MoAb scanning is not sensi¬ tive enough to detect all micrometastatic foci. Further stud¬ ies must be performed on patients with no discernible disease to answer this question. The inability to reliably reveal the presence of micrometastatic disease in patients with concur¬ rent gross disease resulted in an increase in the number of
false-negative scan results (34%).
Previous reports described a substantial incidence of HAMA formation against MoAbs.9,10 No significant increase was seen in HAMA reaction against specific MoAb NR-Ml05. This could be explained by an antiglobulin response against the irrelevant antibody (NR-2AD) as opposed to the specific antibody (NR-Ml-05), as previously described by Schroff et al,16 or the short duration and one-time exposure of the patient to the foreign protein. The use of Tc has been shown to be advantageous over other radiolabeled isotopes because of its relatively short half-life (6.2 hours), low radiation exposure, and 140 keV gamma emission for which modern nuclear medicine gamma cameras have high-resolution and detection capabilities.10 An¬ other advantage over iodine 131 or indium 111 is that tumor uptake is optimal within 5 to 7 hours after MoAb injection of ""Tc as opposed to 24 to 72 hours with 131I or mIn.17 The problem of liver uptake is not seen with "Te as it is with mInMoAb conjugates,11 making ""Tc-labeled Fab fraction superi¬ or in detecting liver métastases. Detection of malignant melanoma with 99mTc-labeled Fab, NR-Ml-05, is a sensitive, safe, and simple scanning proce¬ dure. It provides reliable information on the presence and absence of gross disease and, in some instances, the presence of occult disease. On preliminary investigation, this informa¬ tion is helpful in predicting prognosis and, in a small number of patients, making clinical treatment decisions. References 1. Balch CM, Soong S-J, Shaw HM, Milton GW. Cutaneous Melanoma: Clinical Management and Treatment Results Worldwide. Philadelphia, Pa: JB Lippincott; 1985. 2. Trojani M, de Mascarel I, Bonichon F, CoindreJM, Delsol G. Micrometastases to axillary lymph nodes from carcinoma of breast: detection by immunohistochemistry and prognostic significance. Br J Cancer. 1987;55:303-306. 3. Akslen LA, Hove LM, Hartveit F. Metastatic distribution in malignant melanoma. Invasion Metastasis. 1987;7:253-263. 4. Balch CM, Soong S-J, Milton GW. A comparison of prognostic factors and surgical results with localized (stage 1) melanoma treated in Alabama, USA, and New South Wales, Australia. Ann Surg. 1982;196:677-684. 5. Burmol TF, Reisfeld RA. Unique glycoprotein-proteoglycan complex defined by monoclonal antibodies on human melanoma cells. Proc Natl Acad Sci USA. 1982;79:1245-1249. 6. Fritzberg AR. Advances in Tc-99m labeling of antibodies. Nucl Med.
1987;26:7-12.
7. Larson SM. Radiolabeled monoclonal anti-tumor antibodies in diagnosis and therapy. J Nucl Med. 1985;26:538-545. 8. Morgan AC, Galloway DR, Reisfeld RA. Production and characterization of monoclonal antibody to a specific glycoprotein. Hybridoma. 1990;1:27-36. 9. Eary JF, Schroff RW, Abrams PG, et al. Successful imaging of malignant melanoma with technetium-99m-labeled monoclonal antibodies. J Nucl Med.
1989;30:25-32. 10. Lamki LM, Zukiwski AA, Shanken LJ, et al. Radioimaging of malignant
melanoma using 99m-Tc-labeled Fab fragment reactive with a high molecular weight melanoma antigen. Cancer Res. 1990;50:904-908. 11. Lamki LM, Murray JL, Rosenblum MG, Patt YZ, Babian R, Unger MW. Effect of unlabeled monoclonal antibody (MoAb) on biodistribution and pharmokinetics of 111 Indium labeled MoAb. Nucl Med Commun. 1988;9:553-564. 12. Wade DS, Herrera L, Castillo NB, Petrelli NJ. Metastases to the lymph nodes in epidermoid carcinoma ofthe anal canal studied by a clearing technique. Surg Gynecol Obstet. 1989;169:238-242. 13. Siccardi AG, Buraggi GL, Callegaro L, et al. Multicenter study of immunoscintigraphy with radiolabeled monoclonal antibodies in patients with melanoma. Cancer Res. 1986;46:4817-4822. 14. Sands H, Jones PL, Shah SA, Palme D, Vesselle RL, Gallagher BM. Correlation of vascular permiability and blood flow with monoclonal antibody uptake by human Clouser and renal cell renographs. Cancer Res. 1988;48:188\x=req-\ 193. 15. Becker J, Wasselle JA, Reintgen DS. The immunobiology of the regional lymph node in malignant melanoma. In: Program and Abstracts of the 13th Annual Seminar of Cancer Research in Florida ofthe American Cancer Society; March 3,1990; Orlando, Fla. 16. Schroff RW, Beatty PG, Abrams PG, et al. Use of irrelevant murine monoclonal antibody (MAb) to reduce anti-globulin responses to relevant MAb. JNucl Med. 1987;28:615. 17. Carasquillo JA, Abrams PG, Schroff RW, et al. Effects of In-111 9.2.27 monoclonal antibody dose on the imaging and biodistribution of indium-111 9.2.27 anti-melanoma monoclonal antibody. JNucl Med. 1988;29:39-47.
Downloaded From: http://archsurg.jamanetwork.com/ by a Michigan State University User on 06/13/2015
Melanoma
has an unpredictable malignant potential. A significant incidence of recurrent metastatic disease is seen within 5 years in stage I and II melanoma. Recurrent disease is predictable for deep primary melanomas (stage I), for patients with regional lymph nodes containing metastatic tumor (stage II), and for patients with isolated visceral mé¬ tastases (stage III) after resection.1 Occult disease present at the time of surgery has been suggested as the basis of these recurrences.2 Metastatic melanoma can involve any organ but has a propensity to occur in the regional lymph nodes, lung, liver, bone, and central nervous system.3 Risk factors have been described in an attempt to predict the incidence of metastatic disease following primary tumor resection. Unfavorable prognostic factors for stage I disease include Breslow tumor thickness greater than 0.76 mm and tumor ulcération.4 Although the probability of metastasis can be assessed, no single test is currently available for confirma¬ tion of staging. The routine workup for clinical staging in¬ cludes chest roentgenography and serum liver function tests. Abnormal test results and patient symptoms are subsequent¬ ly evaluated with computed tomography (CT) of the head, chest, abdomen, and pelvis depending on the abnormality detected and the symptoms. Monoclonal antibodies (MoAbs) to tumor-associated anti¬ gens are becoming an important tool for evaluating tumor
Accepted for publication January 12,1991. From the Departments of Surgery (Drs Wasselle, Becker, Cruse, Cox, and Reintgen) and Pathology (Dr Espinosa), Melanoma Treatment Center, University of South Florida, Tampa. Read before the 43rd Annual Cancer Symposium of the Society of Surgical Oncology, Washington, DC, May 20,1990. Reprint requests to the H. Lee Moffitt Cancer Center, University of South Florida, PO Box 280179, Tampa, FL 33682-0179 (Dr Reintgen).
stage and treatment. Monoclonal antibodies against the 250-
kd cell surface glycoprotein,5 expressed on 90% of all melano¬ recently become available. These antibodies have been tagged with radioactive technetium via a bifunctional chelating agent and developed for detection of malignant melanoma.6 The use of the MoAb Fab fragment rather than whole IgG has been shown to greatly enhance image resolu¬ tion, largely through increased clearance with reduced back¬ ground interference and greater specific localization of the Fab fragment compared with whole IgG in tumors.7 The Melanoma Treatment Center at the University of South Florida, Tampa, conducted this study to evaluate the role of technetium-labeled MoAb scans in the metastatic workup and clinical treatment of patients with advancedstage melanoma and to determine the effect of long-term clinical follow-up on sensitivity results. mas, have
MATERIALS AND METHODS
MoAb· NR-Ml-05, a murine antimelanoma MoAb against the 250-kd highmolecular-weight proteoglycan antigen, was developed and purified for clinical use (NeoRx Corp, Seattle, Wash) using standard hybridoma techniques.8 Fab fragments were also prepared according to
standard methods. The irrelevant murine MoAb NR-2AD (whole IgG), reactive with idiopathic determinants of surface IgG on a B-cell lymphoma and not reactive with normal tissue or other tumors, was also supplied by NeoRx.
Preparation of Technetium 99m MoAb Conjugate An imaging kit was developed that uses a diamide dimercaptide N2S2 formed by the esterification of 4,5-bisthioacetamide pentanoate with 2,3,5,6-tetraflourophenol using carbodiimide (NeoRx Corp). The ""Tc-NgSj, complex was produced by reconstituting stannous gluconate with 1 mL of sterile water and then mixing 0.75 mL of r"Tc]pertechnetate with this solution.9
Glacial acetic acid-0.2N hydrochloric acid solution and buffers used to adjust the pH of the stannous gluconate solution to form a^Tc ligand ester. NR-Ml-05 was added to this active ester complex and incubated at room temperature for 20 minutes. The resulting were
99mTc-N2S2 antibody conjugate was then purified using ion exchange columns and micropore filters.10
Patients and Study Design
Twenty-nine patients seen at the Melanoma Treatment Center at University of South Florida, Tampa, with invasive stage I, regional nodal stage II, or systemic stage III disease were evaluated. Four of the patients had undergone no previous therapy, 25 had undergone resections of the primary tumor, and none had undergone previous chemotherapy. One patient had undergone immunotherapy using an irradiated tumor cell vaccine. All participants gave written consent to receive the radioconjugate. The histologie diagnosis of melanoma was confirmed in all patients by the clinic dermatopathologist(C.E.). The patients received three antibody preparations intravenously: 40 mg of unlabeled irrelevant murine antibody (NR-2AD), followed the
25 minutes later with 7.5 mg of an unlabeled whole antimelanoma antibody (NR-MI-05), followed 5 minutes later with 10 mg of 1110Mbq Tc-labeled Fab antimelanoma antibody (NR-Ml-05). Unla¬ beled NR-Ml-05 IgG and excess unlabeled irrelevant NR-2AD-IgG were administered before the radiolabeled MoAb Fab fragment be-
Downloaded From: http://archsurg.jamanetwork.com/ by a Michigan State University User on 06/13/2015
earlier studies11 demonstrated blocking of the reticuloendothelial system by the combination of the irrelevant MoAb and the intact immunoglobulin. This resulted in an enhanced uptake of the tracer in the tumor. All patients were observed for signs and symptoms of toxicity following MoAb injection. A laxative was administered 3 hours after injection to clear the gastrointestinal tract of radioisotope. Wholebody images were taken beginning 5 hours after injection in the anterior and posterior projections using a gamma camera. A dedi¬ cated computer was used to measure regional uptake of the radiolabeled antibody. Left and right lateral views of the head were also obtained. Occasionally, counts were repeated 24 hours later if images were questionable. All scans were reviewed and images of positive sites were then correlated with lesions found with conventional roent¬ genography and physical examinations. All patients underwent chest roentgenography, and serum sam¬ ples were obtained for liver function tests and lactic dehydrogenase determinations. Computed tomographic scans of the chest, head, abdomen, and pelvis were obtained when clinically warranted. All positive sites, determined by either MoAb scans or conventional roentgenographic studies, underwent surgical biopsies, and the spec¬ imens were evaluated by a pathologist. Specimens from clinically positive sites and lymph nodes from elective dissections were resect¬ ed and sent for routine pathologic determination. All patients were followed up after surgery. cause
Assay for Human Antimurine Antibody (HAMA) In 16 of 29 patients studied, serum samples were drawn 5,8, and 16 weeks after surgery and analyzed for HAMA response using a previ¬ ously published enzyme-linked immunosorbent assay technique.9 Titers of antibody against NR-Ml-05 were considered significantly elevated if they were more than 2 SDs above the mean normal level of serum units for the controls.
RESULTS
Optimal images were acquired at a mean of 5 hours after infusion of ""Tc-NR-Ml-05 Fab fragment. Nonspecific up¬ take was seen in the gallbladder, lung, bone marrow, kidneys, testicles, ovaries, and bladder. A total of 32 sites were identified to contain tumor from pathologic specimens, 28 of which were MoAb-scan positive (88% sensitivity). False-negative imaging occurred at a rate of 12%. The Table summarizes the overall sensitivity of confirmed metastatic lesions in various organ sites from all 29 patients. Analysis of individual sites determined to be posi¬ tive by scanning revealed that nodal basins and visceral mé¬ tastases accounted for the highest percentage of metastatic sites imaged, with 14 (88%) of 16 known lymph node métasta¬ ses and all four known visceral métastases being detected. In soft tissues, 10 (83%) of 12 sites known to contain disease were imaged. Occult disease was detected in two (7%) of 29 pa¬ tients. Six patients eventually died of their disease and dis¬ ease recurred in 10 patients. Assuming the recurrences were secondary to undetected occult disease at the time of scan¬ ning, the sensitivity falls to 66%. The false-positive rate was 21%. These scans included three nodal basins not containing tumor and three benign tumors: a lipoma, a breast adenoma, and a meningioma. Sev¬ en patients had areas of poorly defined uptake that corre¬ sponded to areas of scar formation, infection, and, in one case, a previous tumor cell vaccine and bacille Calmette-Guérin injection site. An attempt was made to determine the effect of MoAb scanning on clinical decision making. In three patients with lower-extremity melanomas, the scan showed positive tumor uptake in the iliac nodal chain. Computed tomographic scans were then performed, the results of which were negative. Dissections of these areas revealed occult micrometastatic disease in two of the three patients. Eleven patients with gross disease underwent MoAb scan imaging. Nine patients with deeply invasive tumor or micrometa-
Sensitivity of Monoclonal Antibodies Labeled With Technetium 99m Fab-NR-M1-05 for Known Melanoma Métastases
NO. (%) Of Lesions
NO. Of Known Lesions
Lesion Site
Imaged 9 (82) 1 (100) 1 (100) 1 (100) 2 (100) 12(86) 2(100) 28 (88)
11
Skin/subcutaneous Mediastinum
1
Liver
1
Brain
1
Lung Regional lymph nodes Distant lymph nodes
2 14
Total
32
2
static disease in their regional node dissections underwent MoAb scanning. None of the nine had evidence of disease either clinically or roentgenographically at the time of MoAb scanning. All nine patients had concurrent negative scan results. Although the follow-up was relatively short (mean, 11.4 months; range, 3 to 21 months), only one of the patients with previously negative scan results developed metastatic disease, making the predictive value of a completely negative scan 89%. These results need to be confirmed with a longer
follow-up.
Two cases illustrate the clinical use of MoAb scanning. A 55-year-old white man presented with palpable adenopathy in his right axilla 2 years after undergoing a wide local excision for a Clark level II melanoma on his right shoulder. During the workup, a MoAb scan with lateral (Fig 1, top) and anterior (Fig 1, center) views of the head revealed an isolated, totally asymptomatic brain mass. A CT scan (Fig 1, bottom) con¬ firmed the finding of an intracranial mass. The MoAb scan of the right axilla also showed the palpable adenopathy (Fig 2). The patient underwent a right axillary dissection and craniotomy. On pathologic evaluation, he had metastatic disease in his right axilla; however, the craniotomy revealed a meningioma.
In the second case, the MoAb scan revealed occult nodal disease. A 44-year-old white man, with a history of a Clark level IV melanoma removed from his left calf, presented with a palpable intransit lesion on the medial aspect of his left thigh and palpable left inguinal adenopathy. Results of a CT scan of the pelvis and abdomen were negative. However, a MoAb scan revealed uptake in the intransit thigh lesion (Fig 3, top), uptake in the left inguinal node region (Fig 3, bottom), and uptake in the iliac node region (Fig 3, bottom). The patient underwent a wide excision of his intransit lesion, therapeutic left groin dissection, and elective iliac nodal dissection. The pathologic examination showed metastatic disease in all three sites, with micrometastatic disease noted in the iliac nodes that were less than 1.0 cm in diameter. In this instance, the MoAb scan led to more extensive surgery that revealed occult métastases. Three (10%) of the 29
scan
results
were
confirmed to be
false-negative on pathologic examination. Two of the falsenegative scan results occurred with known abdominal-wall
métastases, and the other scan did not reveal micrometastatic
disease in lymph node basins found to be pathologically posi¬ tive after an elective lymph node resection. Four (14%) of the 29 scans had mixed results, revealing true-positive, falsepositive, and false-negative combinations at different sites. No toxicity to the MoAb infusion was observed in any patient, and no significant increase in HAMA specific for NRMl-05 was seen in the 16 patients tested.
Downloaded From: http://archsurg.jamanetwork.com/ by a Michigan State University User on 06/13/2015
2.—Anterior image of the right shoulder obtained 5 hours after intravenous administration of technetium 99m-Fab-NR-M1-05 mono¬ clonal antibodies. Note the positive tracer uptake in the axillary lymph nodes (arrow), with normal blood pool distribution seen in the lung and axillary vasculature.
Fig
Fig 1.—Top, Lateral view of the head after monoclonal antibody injection shows enhancing left meningioma. Center, Anterior view of the head shows the same lesion. Bottom, Computed tomographic scan of the head reveals a benign meningioma at the left. COMMENT The information provided by in vivo tumor localization with MoAbs suggests that "Tc-labeled MoAb (Fab fragment) is a sensitive and safe method of detecting metastatic melanoma. These scans imaged 88% of detectable melanomas and visual-
Fig 3.—Top, Scan of a medial thigh lesion (arrow) in a patient with a previously resected primary melanoma of the calf. Bottom, Scan of occult iliac nodal (wide arrow) and palpable inguinal nodal disease (thin arrow). Nonspecific uptake is seen in the testes, and excreted complex is seen in the urinary bladder.
Downloaded From: http://archsurg.jamanetwork.com/ by a Michigan State University User on 06/13/2015
ized all gross disease in 11 (50%) of 22 patients, confirming the absence of gross disease in nine (41%) of 22 patients, and detecting occult disease in two (9%) of 22 patients. In the patients with occult disease detected with MoAb scanning, further roentgenographic testing and more extensive sur¬ gery were performed based on the MoAb results. Negative scan results in a patient with deeply invasive stage I or micrometastatic stage II disease had a predictive value of 89% for no further disease development. If this holds true with longer follow-up, then MoAb scanning may be used as a screening test to initially evaluate patients with an earlystage melanoma for metastatic disease and perhaps replace the battery of screening tests used now, including liver func¬ tion tests, chest roentgenography, and CT scans when appli¬ cable. Although large prospective trials are needed to com¬ pare NR-Ml-05 scanning with more standard imaging modalities, the results may also be used as a prognostic indicator for patients and their families to confirm the absence of systemic micrometastatic disease. The MoAb scan was able to detect occult disease in two patients that was otherwise undetectable on physical exami¬ nation or routine roentgenographic examinations. Both of these occult nodal métastases developed in the iliac nodal basins from lower-extremity melanomas. It is very difficult, if not impossible, to clinically examine the iliac nodal chain. The roentgenographic procedure of choice, a CT scan of the pelvis, will distinguish between enlarged, pathologic-appearing nodes (ie, larger than 2.0 cm) and normal-sized nodes, but cannot answer the question ofwhether the nodes are involved with metastatic disease or reactive in nature. Numerous studies have shown that micrometastases usual¬ ly develop in nodes that are less than 1.0 cm in diameter12 and would not be identified as abnormal with the CT scan. In these two cases, iliac dissection was added to the standard superfi¬ cial lymph node dissection based on the results of the MoAb scan combined with the prognostic variables for the particular patient's primary melanoma. Micrometastatic disease was pathologically confirmed in two of three iliac nodal basins. The MoAb scan could help to identify the need for a more exten¬ sive ilioinguinal node dissection in patients with aggressive, high-risk primary melanomas of the lower extremity or in patients with grossly involved inguinal lymph nodes. Technetium 99m-N2S2 has previously been shown to be excreted in bile and by the kidneys, which explains the local¬ ization in the gallbladder, kidneys, and urinary bladder.6,11,13 The 21% false-positive rate seen in nodal basins and benign tumors and the nonspecific, poorly defined uptake seen in areas of inflammation may be related to increased vascular permeability or increased blood flow." Another possible ex¬ planation involves the presence of submicroscopic disease overlooked at the time of pathologic examination.15 Because of the 21% false-positive rate in this patient population, clinical judgment based on prognostic factors for melanoma must be included in the analysis of data generated by the scan to avoid unnecessary surgery. The recurrence of disease in 10 patients is most likely due to micrometastatic disease not found on initial evaluations.2 Since only one patient who had completely negative MoAb scan results had a recurrence and all other patients who had recurrences had gross disease at the time of MoAb scanning, one may postulate that clinically apparent metastatic deposits containing high tumor volumes may act as a sink for the MoAb and not allow radiolocalization to micrometastatic deposits. However, it is possible that the MoAb scanning is not sensi¬ tive enough to detect all micrometastatic foci. Further stud¬ ies must be performed on patients with no discernible disease to answer this question. The inability to reliably reveal the presence of micrometastatic disease in patients with concur¬ rent gross disease resulted in an increase in the number of
false-negative scan results (34%).
Previous reports described a substantial incidence of HAMA formation against MoAbs.9,10 No significant increase was seen in HAMA reaction against specific MoAb NR-Ml05. This could be explained by an antiglobulin response against the irrelevant antibody (NR-2AD) as opposed to the specific antibody (NR-Ml-05), as previously described by Schroff et al,16 or the short duration and one-time exposure of the patient to the foreign protein. The use of Tc has been shown to be advantageous over other radiolabeled isotopes because of its relatively short half-life (6.2 hours), low radiation exposure, and 140 keV gamma emission for which modern nuclear medicine gamma cameras have high-resolution and detection capabilities.10 An¬ other advantage over iodine 131 or indium 111 is that tumor uptake is optimal within 5 to 7 hours after MoAb injection of ""Tc as opposed to 24 to 72 hours with 131I or mIn.17 The problem of liver uptake is not seen with "Te as it is with mInMoAb conjugates,11 making ""Tc-labeled Fab fraction superi¬ or in detecting liver métastases. Detection of malignant melanoma with 99mTc-labeled Fab, NR-Ml-05, is a sensitive, safe, and simple scanning proce¬ dure. It provides reliable information on the presence and absence of gross disease and, in some instances, the presence of occult disease. On preliminary investigation, this informa¬ tion is helpful in predicting prognosis and, in a small number of patients, making clinical treatment decisions. References 1. Balch CM, Soong S-J, Shaw HM, Milton GW. Cutaneous Melanoma: Clinical Management and Treatment Results Worldwide. Philadelphia, Pa: JB Lippincott; 1985. 2. Trojani M, de Mascarel I, Bonichon F, CoindreJM, Delsol G. Micrometastases to axillary lymph nodes from carcinoma of breast: detection by immunohistochemistry and prognostic significance. Br J Cancer. 1987;55:303-306. 3. Akslen LA, Hove LM, Hartveit F. Metastatic distribution in malignant melanoma. Invasion Metastasis. 1987;7:253-263. 4. Balch CM, Soong S-J, Milton GW. A comparison of prognostic factors and surgical results with localized (stage 1) melanoma treated in Alabama, USA, and New South Wales, Australia. Ann Surg. 1982;196:677-684. 5. Burmol TF, Reisfeld RA. Unique glycoprotein-proteoglycan complex defined by monoclonal antibodies on human melanoma cells. Proc Natl Acad Sci USA. 1982;79:1245-1249. 6. Fritzberg AR. Advances in Tc-99m labeling of antibodies. Nucl Med.
1987;26:7-12.
7. Larson SM. Radiolabeled monoclonal anti-tumor antibodies in diagnosis and therapy. J Nucl Med. 1985;26:538-545. 8. Morgan AC, Galloway DR, Reisfeld RA. Production and characterization of monoclonal antibody to a specific glycoprotein. Hybridoma. 1990;1:27-36. 9. Eary JF, Schroff RW, Abrams PG, et al. Successful imaging of malignant melanoma with technetium-99m-labeled monoclonal antibodies. J Nucl Med.
1989;30:25-32. 10. Lamki LM, Zukiwski AA, Shanken LJ, et al. Radioimaging of malignant
melanoma using 99m-Tc-labeled Fab fragment reactive with a high molecular weight melanoma antigen. Cancer Res. 1990;50:904-908. 11. Lamki LM, Murray JL, Rosenblum MG, Patt YZ, Babian R, Unger MW. Effect of unlabeled monoclonal antibody (MoAb) on biodistribution and pharmokinetics of 111 Indium labeled MoAb. Nucl Med Commun. 1988;9:553-564. 12. Wade DS, Herrera L, Castillo NB, Petrelli NJ. Metastases to the lymph nodes in epidermoid carcinoma ofthe anal canal studied by a clearing technique. Surg Gynecol Obstet. 1989;169:238-242. 13. Siccardi AG, Buraggi GL, Callegaro L, et al. Multicenter study of immunoscintigraphy with radiolabeled monoclonal antibodies in patients with melanoma. Cancer Res. 1986;46:4817-4822. 14. Sands H, Jones PL, Shah SA, Palme D, Vesselle RL, Gallagher BM. Correlation of vascular permiability and blood flow with monoclonal antibody uptake by human Clouser and renal cell renographs. Cancer Res. 1988;48:188\x=req-\ 193. 15. Becker J, Wasselle JA, Reintgen DS. The immunobiology of the regional lymph node in malignant melanoma. In: Program and Abstracts of the 13th Annual Seminar of Cancer Research in Florida ofthe American Cancer Society; March 3,1990; Orlando, Fla. 16. Schroff RW, Beatty PG, Abrams PG, et al. Use of irrelevant murine monoclonal antibody (MAb) to reduce anti-globulin responses to relevant MAb. JNucl Med. 1987;28:615. 17. Carasquillo JA, Abrams PG, Schroff RW, et al. Effects of In-111 9.2.27 monoclonal antibody dose on the imaging and biodistribution of indium-111 9.2.27 anti-melanoma monoclonal antibody. JNucl Med. 1988;29:39-47.
Downloaded From: http://archsurg.jamanetwork.com/ by a Michigan State University User on 06/13/2015
