Journal of Surgical Oncology 51:226-230 (1992)
"'
Gamma Probe Location of Indium-labeled B72.3: An Extension of lmmunoscintigraphy DAVID N. KRAG, MD, MICHAEL K. HASEMAN, MD, PATRICK FORD, MD, LLOYD SMITH, MD, MICHAEL H. TAYLOR, MD, PHILIP SCHNEIDER, PhD, MD, ANU JAMES E. GOODNIGHT, MD, PhD From the Departments of Surgery (L.S., M.H.T., P.S., 1.E.C.Iand Nuclear Medicine (P.F., M.K. H.), University of California at Davis, Sacramento, California; and the Department of Surgery, University of Vermont (D.N.K.), Burlington, Vermont
Eight colorectal and 5 ovarian cancer patients were evaluated with preoperative immunoscintigraphy and intraoperative gamma probe detection of I I I indium-labeled . monoclonal antibody B72.3. Immunoscintigraphy detected the presence of tumor in every patient shown to have tumor at surgery. There was one false-positive scan. A total of 21 pathologically verified lesions were identified at surgery in the 1 1 patients with tumor. Immunoscintigraphy localized 12 (57%) and intraoperative gamma probe detection located 17 (8 I %) of the lesions. lntraoperative probe detection located 6 of 8 lesions smaller than 1 cm and 3 lesions that were not identified on initial surgical exploration. The gamma probe offers information that is complementary to immunoscintigraphy in that ( l ) it aids the surgeon in locating intra- and extra-abdominal lesions previously identified by immunoscintigraphy, (2) it locates lesions too small to be seen by immunoscintigraphy alone, ( 3 ) it locates lesions that otherwise might be missed at surgery, and (4) it provides objective evidence for adequacy of surgical resection of cancer in the abdominal cavity. 0 1992 Wiley-Liss, Inc.
KEYWORDS:colorectal, ovarian cancer, immunoscintigraphy
INTRODUCTION Gamma camera imaging of radiolabeled monoclonal antibodies (immunoscintigraphy) has been undergoing rapid development, particularly over the last decade. The majority of studies involve external gamma camera imaging of cancer patients who have been infused with radiolabeled monoclonal antibodies that bind to tumor associated antigens. Several studies have documented that immunoscintigraphy is comparable or complementary to other radiologic studies for locating solid tumors [ 1-61. The potential benefits of this technology warrant aggressive efforts to surmount a number of limiting factors. These include ( 1 ) inadequate specificity of the monoclonal antibody for the tumor, (2) incomplete penetration of the monoclonal antibody into solid tumors, and (3) development of host antibodies against the infused antibody [7]. Currently, immunoscintigraphy seldom detects tumor deposits less than I cm in diameter. One limitation of 0 1992 Wiley-Liss, Inc.
external scanning is related to rapid attentuation of gamma rays by overlying tissues as the distance between the isotope and the detector increases. The inverse square law dictates that the amount of detectable radiation varies inversely with the square of the distance of the detector from the radiation source. The closer a gamma detector is to the radiation source the better the sensitivity [S].A small hand-held gamma counter is well suited for intraoperative examination of tissues, allowing close approximation of detector to the source. Gamma camera imaging and intraoperative gamma probe detection provide complementary information. Although a hand-held gamma counter will detect smaller lesions than wilI a gamma camera, it will not provide an actual image. It also will not provide optimal information Accepted for publication August 17, 1992. Address reprint requests to Dr. David N. Krag, Surgical Associates Foundation, Inc., 1 South Prospect Street, Burlington, VT 054013444.
Gamma Probe Localization
227
about tissues with which it cannot be in contact. External scanning provides a broad examination of the entire body for local deposits of radiolabeled tissues. The hand-held gamma counter helps localize radiolabeled tissue previously identified with the gamma counter and those deposits within the surgical field that are too small to be detected by the gamma camera. We report here on the use of a commercially available gamma probe (Oncoprobe, Care Wise Medical Products, Morgan Hill, CA) for intraoperative detection of Illindium-labeled monoclonal antibody B72.3 (' I In-CYT103, Cytogen Corporation, Princeton, NJ) in 5 ovarian and 8 colorectal cancer patients.
'
MATERIALS AND METHODS The oncoprobe is a scintillation probe consisting of a sodium iodide crystal coupled to a photomultiplier tube, both of which are contained in a tungsten alloy housing. The threshold and window were adjusted to optimize signal collection for I 'indium. The majority of the measurements were made with a threshold of 150 keV and a window of 50 keV. A collimating head was used to narrow the solid angle of acceptance in an effort to minimize background signals. Although the unit can be gas sterilized, a sterile plastic sleeve designed to cover an ultrasound probe was used to cover the probe. The intensity of radiation detected by the gamma probe was indicated by (1) an analogue needle indicating counts per second, (2) an audio signal that increased the pitch as the radiation level increased, and (3) a print out of cumulative counts over the desired period of time. Cumulative counts were generally made over 10 s. Five ovarian cancer patients and 8 colorectal cancer patients were entered into this prospective, open-label nonrandomized trial at the University of California, Davis. All patients signed an informed consent approved by the UC Davis Human Subject Protection Committee. The mean age of the patients was 65 years. All ovarian cancer patients underwent surgery for possible recurrent cancer, and 3 of the 5 patients had recurrent disease. Seven of the colorectal cancer patients had surgery for primary cancer and one was operated on for hepatic metastases. Monoclonal antibody B72.3 was conjugated with linker-chelator glycyl-tyrosyl-lysyl-diethylenetriaminepentaacetic acid and termed CYT- 103 (Cytogen Corporation, Princeton, New Jersey). One milligram of CYT-I03 was radiolabeled with 5 mCi of '"In chloride (Amersham, Arlington Heights, IL) within 1 h of infusion. Patients were infused with 5 mCi of "'In-CYT-103 over 10 min, and vital signs were monitored for 60 min. All patients had planar and SPECT images of the chest, abdomen, and pelvis obtained at 2-4 days postinfusion. Surgery was performed 4-15 days postinfusion. At surgery, the abdominal cavity was explored with the gamma probe
Fig. 1. Planar scan of colorectal cancer patient showing cervical lesions with increased uptake of radiolabeled antibody.
after the surgeon had made a full inspection and committed to the surgical procedure. This was repeated after the tumor was resected. A lesion was considered positive if the counts of the lesion minus 3 times the standard deviation of the lesion counts (i.e., lesion counts minus 3 times the square root of the lesion counts) were greater than the mean background counts.
RESULTS At surgery, 11 of the 13 patients had histologically proven cancer. Immunoscintigraphy detected at least one lesion in every patient with tumor and there was 1 false positive scan. CT scan was positive in 7 of the 11 patients that had CT scans performed, yielding 6 true positives, 1 false positive, and 2 false negatives. The gamma probe detected at least one lesion in 9 of the 11 patients with tumor. A total of 21 lesions were identified in the 1 1 patients with tumor. Immunoscintigraphy identified 12, CT scan 9, and the gamma probe 7. Eight of the 21 lesions were less than 1 cm in diameter. Immunoscintigraphy detected none of the lesions less than 1 cm, CT scan identified 2, and the gamma probe identified 6. Immmunoscintigraphy and the gamma probe detected increased radioactivity in the cervical region of 2 patients. In one patient, fine-needle aspiration of palpable nodules was guided by the gamma probe. Three of 3 nodules identified as positive by the probe were positive for metastatic adenocarcinoma of the colon (Fig. 1). The second patient had no palpable nodules but had activity localized to the left scalene lymph node region. The
228
Krag et al. TABLE 11. Mean 10 s Counts of Nonmalignant Tissues Obtained With the Gamma Probe Mean counts Tissue Abdominal aorta Lumbar spine Liver right lobe Spleen Anterior abdominal wall Small bowel Colon Pubic symphysis Pelvic sidewall Sacral prominence Diaphragm
Uncollimated 690
Collimated
579
583 240 1,203 611 8 15 18 86 I23 I74
253
90
812 3,185 1,027 30 21 I07
I99 233
Ratio of collimated to uncollimated 1.2 3.4 2.6 1.52 3.8 1.4
5.9 2.3 I .9
3.3 3. I
There were no adverse reactions resulting from the infusion of the radiolabeled antibody. Fig. 2. Planar scan of ovarian cancer patient showing increased uptake of radiolabeled antibody in cervical lymph.
TABLE I. Mean In Vivo and Ex Vivo Counts Accumulated Over 10 s of Tumors > 3 cm and Their Background
In vivo Ex vivo
Tumors >3 cm
Background
793
523
353
79
probe guided the resection of 3 radioactive scalene lymph nodes (Fig. 2), which on histopathology were determined to have a histiocytic reaction but did not contain tumor. Overall the gamma probe identified a total of 5 lymph In-CYT- 103 but nodes that had increased uptake of were histologically negative for tumor. All the radioactive lymph nodes had histiocytic changes. The gamma probe identified 3 tumor deposits that had not been identified by the surgeon. One lesion was a small metastatic nodule of rectal cancer invading the sigmoid colon, one was a small cluster of lymph nodes with ovarian cancer at the iliac artery bifurcation, and the third was a cluster of periaortic lymph nodes positive for ovarian cancer. The mean in vivo and ex vivo counts acquired over 10 s with the gamma probe for all tumors >3 cm and associated background counts are shown in Table I. The normal tissue background counts acquired over 10 s ranged from a high of 3,185 in the liver to a low of 30 in the anterior abdominal wall. The mean counts acquired over 10 s for various nonmalignant areas of the body both collimated and uncollimated are shown in Table 11.
'
DISCUSSION Intraoperative probes for detecting localized deposits of radioisotopes in tissue have been intermittently used for at least SO years. Until the last decade, detection was limited to radioactive material delivered to tumors in a relatively nonspecific manner. Intraoperative probes were used to detect brain tumors radiolabeled with "phosphorus [9], thyroid tumors radiolabeled with I3'iodine [lo], submucosal lung cancers radiolabeled with 57Co-bleomycin [ 1 I], and ostoid osteoma labeled with "'?echnetium [ 12,131. The identification of carcinoembryonic antigen (CEA) in 1965 [ 141, the development of hybridoma technology in 1975 [IS] and the application of this technology for delivery of radioisotopes in humans in 1978 [ 161 greatly increased the ability to deliver selectively a variety of radioisotopes to tumor tissue [4-61. Clinical trials of intraoperative detection of radiolabeled antibodies began in 1984 with a case report describing intraoperative detection of radiolabeled colorectal cancer [ 171. Subsequent experience by this group at Ohio State has led to the development of a commercially available gamma probe (Neoprobe Corporation, Columbus, OH). They reported that at least 77% of primary and recurrent colorectal cancers were detected, occult lesions were detected, and patient management was improved [ 18-20], The probe is designed to detect tissues radiolabeled with "'iodine in the absence of significant background radiation. Improved tumor to background ratios are expected when background radiation is very low. The downside of using 125. iodine is that ( I ) it is not suitable for preoperative imaging, and (2) it takes a mean of 22 days to achieve a suitable clearance of background activity prior to intraoperative detection.
Gamma Probe Localization
Our approach has been to view intraoperative gamma detection as an extension of preoperative immunoscintigraphy. Our experience has been with the Oncoprobe, a commercially available gamma counter designed for intraoperative use. We found that the probe detected tumor in 9 of the 1 1 patients with cancer. One of the patients with a false negative result had a delay in surgery until 15 days after infusion of "'In-CYT-103. By this time the antibody had cleared from the tumor and the indium had decayed to levels too low for effective gamma counting. The other patient that had a false negative evaluation with the probe had a large lesion in the liver. As is frequently the case with external imaging, high background counts from 'indium in the liver resulted in decreased ability of the gamma probe to discriminate hepatic tumors from background hepatic tissue. Of the 21 lesions identified in the 1 1 patients evaluated, 8 of the lesions were less than 1 cm in diameter. External scanning detected none of the lesions less than 1 cm whereas the probe identified 6. The ability to detect small lesions appears to be a unique feature of intraoperative gamma detection and is related to the ability to place the detector directly on the tissues in question. The gamma probe identified 5 lymph nodes that had increased uptake of '"In-CYT-103 but were histologically negative for tumor. The basis for increased I 'indium presence in occasional tumor-negative lymph nodes has not been determined. This phenomenon has been observed with I 'indium-labeled antibodies other than B72.3 (M Haseman: personal communication). We can speculate that the lymph nodes are processing circulating antigen shed by the tumor which is then targeted in the lymph node by "'In-CYT-103. Immunohistochemistry, however, was unable to detect TAG-72 antigen in the radiolabeled lymph nodes. Alternatively, reactive lymph nodes undergoing a histiocytic reaction may actively bind and accumulate mouse protein in a nonspecific manner. The gamma probe identified 3 tumor deposits that had not been identified by the surgeon. The probe provided evidence for the inadequacy of resection in these cases. More patients will need to be evaluated to better delineate the ability of the gamma probe to identify surgically occult lesions. High background counts remain a limiting factor for identifying small deposits of tumor with the gamma probe. This was particularly evident for lesions in the liver or in the pelvis. Table I1 indicates the range of background counts for nonmalignant areas of the body. It is clear that the liver, large blood vessels, and the bone marrow have high background counts. If the probe is directed toward a high background organ, falsely elevated counts will occur even at a distance. Use of the collimator and care in directing the probe minimizes artificial elevation of counts from a distant source. Cervical metastasis were detected with external gamma camera imaging in one of the eight patients with
''
'
229
colorectal cancer. The radiolabeled cervical lesions were readily localized for biopsy with the gamma probe. A previous study with 'In-CYT- 103 immunoscintigraphy reported a detection rate of extraabdominal colorectal cancer metastasis of 9% [2]. In this subset of patients, the probe may be very useful for guiding the surgeon to extraabdominal lesions detected by gamma camera imaging [2 I]. Confirmation of extraabdominal metastasis would be an important addition to staging of patients with colorectal or ovarian cancer. In conclusion we report on the use of a commercially available gamma probe for intraoperative detection of I I 'indium-labeled monoclonal antibody B72.3 in colorectal and ovarian cancer patients. The gamma probe appears to offer information that is complementary to immunoscintigraphy in that ( I ) it aids the surgeon in localizing intra- and extra-abdominal lesions previously identified by immunoscintigraphy, (2) it localizes lesions too small to be identified by immunoscintigraphy alone, (3) it localizes lesions that might otherwise be missed at surgery, and (4) it provides objective evidence for adequacy of surgical resection of cancer in the abdominal cavity.
''
ACKNOWLEDGMENTS We thank Patrick Newberry for serving as clinical coordinator. This work was supported in part by the Cytogen Corporation. REFERENCES I . Riva P, Moscatelli G , Paganelli G, Benini S, Siccardi A: Antibody guided diagnosis: An Italian experience on CEA-expressing tumours. Int. J. Cancer Suppl. 2.1 14-120, 1988. 2. Doerr Rj, Hani AN, Krdg DN, Mitchell E: Radiolabeled antibody imaging in the management of colorectal cancer: Results of a multicenter clinical study. Ann Surg 214:l 18-124, 1991. 3. Patt YZ, Lamki LM, Shanken J, et al.: Imaging with Indium"' labeled anti-carcinoembryonicantigen monoclonal antibody ZCE025 of recurrent colorectal or carcinoembryonic antigen-producing cancer occult metastases. J Clin Oncol 8:124&1254, 1990. 4. Haller DG: Monoclonal antibody imaging in the management of patients with colorectal cancer. J CIin Oncol6:1213-121.5, 1988. 5 . Larson SM: Lymphoma, melanoma, colon cancer: Diagnosis and treatment with radiolabeled monoclonal antibodies. Radiology 16.5 :297-304, 1987. 6. Byers VS, Baldwin RW: Therapeutic strategies with monoclonal antibodies and immunoconjugates. Immunology 65:329-335, 1988. 7. Epstein AL, Khawli LA: Tumor biology and monoclonal antibodies: Overview of basic principles and clinical considerations. Antibiot lmmunoconjug Radiopharm 4:373-384, I99 I . 8. Barber HB, Barrett HH, Woolfenden JM, Myers KJ, Hickernell TS: Comparison of in vivo scintillation probes and gamma cameras for detection of small, deep tumours. Phys Med Biol 34:727739, 1989. 9. Selverstone B, Sweet WH, Robinson CV: The clinical use of radioactive phosphorus in the surgery of brain tumors. Ann Surg 130~643-651, 1949. 10. Harris CC, Bigelow RR, Francis HE, Kelley GG, Bell RP: A CsI(T1)-crystal surgical scintillation probe. Nucleonics 14:102108, 1956. 1 1 . Woolfdenden JM, Nevin WS, Barber HB, Donahue DJ: Lung
230
12.
13. 14. 15.
16.
17.
Krag et al. cancer detection using a miniature sodium iodide detector cobalt-57 bleomycin. Chest 85:84-88, 1984. Colton CL, Orth ED, Hardy G: J Bone Joint Surg 65A:10191022, 1983. Szypryt EP, Hardy JG, Colton CL: J Bone Joint Surg 685(4):643646. 1986. Gold P, Freedinan SO: Demonstration of tumor-specific antigens in human colonic carcinomas by immunological tolerance and absorption techniques. J Exp Med 121:439, 1965. Kohler G , Milstein C: Continuous cultures of fused cells secreting antibody of predefined specificity. Nature 256:495-497, 1975. Goldenberg DM, Deland FH, Kim E, et al.: Use of radiolabeled antibodies to carcino-embryonic antigen for the detection and localization of diverse cancers by external photoscanning. N Engl J Med 298: 1384-1388, 1978. Aitkin DR, Hinkle GH, Thurston MO: A gamma-detecting probe for radioimmune detection of CEA-producing tumors: Successful experimental use and clinical case report. Dis Colon Rectum 27:279, 1984.
18. NierodaCA, Mojzisik C, Sardi A, FerraraPJ, Hinkle G , Thurston MO, Martin EW Jr: Radioimmunoguided surgery in primary eoIon cancer. Cancer Prevention and Detection 14:651-656, 1990. 19. Martin EW Jr, Mojzisik CM, Hinkle GH Jr, Sampsel J , Siddl MA, Tuttle SE, Sickle-Santanello 5 , Colcher D, Thurston MO, Bell JG, Farrar WB, Schlom J: Radioimmunoguided surgery using monoclonal antibody. Am J Surg 156:38&392, 1988. 20. Hinkle GH, Mojzisik CM, Loesch JA, eta].: The evolution of the Radioimmunoguided Surgery System@):An innovative technique for the intraoperative detection of tumors. Antibody Inimunol Radiopharm 4:339-358, 1991. 21. Kuhn JA, Corbisiero RM, Buras R R , Carroll RG, 5eatty JD: Intraoperative gamma detection probe combined with preoperative Iindium labeled antibody imaging for localization of recurrent carcinoma. Presented at the Society of Surgical Oncology, 1991.
'
"'
Gamma Probe Location of Indium-labeled B72.3: An Extension of lmmunoscintigraphy DAVID N. KRAG, MD, MICHAEL K. HASEMAN, MD, PATRICK FORD, MD, LLOYD SMITH, MD, MICHAEL H. TAYLOR, MD, PHILIP SCHNEIDER, PhD, MD, ANU JAMES E. GOODNIGHT, MD, PhD From the Departments of Surgery (L.S., M.H.T., P.S., 1.E.C.Iand Nuclear Medicine (P.F., M.K. H.), University of California at Davis, Sacramento, California; and the Department of Surgery, University of Vermont (D.N.K.), Burlington, Vermont
Eight colorectal and 5 ovarian cancer patients were evaluated with preoperative immunoscintigraphy and intraoperative gamma probe detection of I I I indium-labeled . monoclonal antibody B72.3. Immunoscintigraphy detected the presence of tumor in every patient shown to have tumor at surgery. There was one false-positive scan. A total of 21 pathologically verified lesions were identified at surgery in the 1 1 patients with tumor. Immunoscintigraphy localized 12 (57%) and intraoperative gamma probe detection located 17 (8 I %) of the lesions. lntraoperative probe detection located 6 of 8 lesions smaller than 1 cm and 3 lesions that were not identified on initial surgical exploration. The gamma probe offers information that is complementary to immunoscintigraphy in that ( l ) it aids the surgeon in locating intra- and extra-abdominal lesions previously identified by immunoscintigraphy, (2) it locates lesions too small to be seen by immunoscintigraphy alone, ( 3 ) it locates lesions that otherwise might be missed at surgery, and (4) it provides objective evidence for adequacy of surgical resection of cancer in the abdominal cavity. 0 1992 Wiley-Liss, Inc.
KEYWORDS:colorectal, ovarian cancer, immunoscintigraphy
INTRODUCTION Gamma camera imaging of radiolabeled monoclonal antibodies (immunoscintigraphy) has been undergoing rapid development, particularly over the last decade. The majority of studies involve external gamma camera imaging of cancer patients who have been infused with radiolabeled monoclonal antibodies that bind to tumor associated antigens. Several studies have documented that immunoscintigraphy is comparable or complementary to other radiologic studies for locating solid tumors [ 1-61. The potential benefits of this technology warrant aggressive efforts to surmount a number of limiting factors. These include ( 1 ) inadequate specificity of the monoclonal antibody for the tumor, (2) incomplete penetration of the monoclonal antibody into solid tumors, and (3) development of host antibodies against the infused antibody [7]. Currently, immunoscintigraphy seldom detects tumor deposits less than I cm in diameter. One limitation of 0 1992 Wiley-Liss, Inc.
external scanning is related to rapid attentuation of gamma rays by overlying tissues as the distance between the isotope and the detector increases. The inverse square law dictates that the amount of detectable radiation varies inversely with the square of the distance of the detector from the radiation source. The closer a gamma detector is to the radiation source the better the sensitivity [S].A small hand-held gamma counter is well suited for intraoperative examination of tissues, allowing close approximation of detector to the source. Gamma camera imaging and intraoperative gamma probe detection provide complementary information. Although a hand-held gamma counter will detect smaller lesions than wilI a gamma camera, it will not provide an actual image. It also will not provide optimal information Accepted for publication August 17, 1992. Address reprint requests to Dr. David N. Krag, Surgical Associates Foundation, Inc., 1 South Prospect Street, Burlington, VT 054013444.
Gamma Probe Localization
227
about tissues with which it cannot be in contact. External scanning provides a broad examination of the entire body for local deposits of radiolabeled tissues. The hand-held gamma counter helps localize radiolabeled tissue previously identified with the gamma counter and those deposits within the surgical field that are too small to be detected by the gamma camera. We report here on the use of a commercially available gamma probe (Oncoprobe, Care Wise Medical Products, Morgan Hill, CA) for intraoperative detection of Illindium-labeled monoclonal antibody B72.3 (' I In-CYT103, Cytogen Corporation, Princeton, NJ) in 5 ovarian and 8 colorectal cancer patients.
'
MATERIALS AND METHODS The oncoprobe is a scintillation probe consisting of a sodium iodide crystal coupled to a photomultiplier tube, both of which are contained in a tungsten alloy housing. The threshold and window were adjusted to optimize signal collection for I 'indium. The majority of the measurements were made with a threshold of 150 keV and a window of 50 keV. A collimating head was used to narrow the solid angle of acceptance in an effort to minimize background signals. Although the unit can be gas sterilized, a sterile plastic sleeve designed to cover an ultrasound probe was used to cover the probe. The intensity of radiation detected by the gamma probe was indicated by (1) an analogue needle indicating counts per second, (2) an audio signal that increased the pitch as the radiation level increased, and (3) a print out of cumulative counts over the desired period of time. Cumulative counts were generally made over 10 s. Five ovarian cancer patients and 8 colorectal cancer patients were entered into this prospective, open-label nonrandomized trial at the University of California, Davis. All patients signed an informed consent approved by the UC Davis Human Subject Protection Committee. The mean age of the patients was 65 years. All ovarian cancer patients underwent surgery for possible recurrent cancer, and 3 of the 5 patients had recurrent disease. Seven of the colorectal cancer patients had surgery for primary cancer and one was operated on for hepatic metastases. Monoclonal antibody B72.3 was conjugated with linker-chelator glycyl-tyrosyl-lysyl-diethylenetriaminepentaacetic acid and termed CYT- 103 (Cytogen Corporation, Princeton, New Jersey). One milligram of CYT-I03 was radiolabeled with 5 mCi of '"In chloride (Amersham, Arlington Heights, IL) within 1 h of infusion. Patients were infused with 5 mCi of "'In-CYT-103 over 10 min, and vital signs were monitored for 60 min. All patients had planar and SPECT images of the chest, abdomen, and pelvis obtained at 2-4 days postinfusion. Surgery was performed 4-15 days postinfusion. At surgery, the abdominal cavity was explored with the gamma probe
Fig. 1. Planar scan of colorectal cancer patient showing cervical lesions with increased uptake of radiolabeled antibody.
after the surgeon had made a full inspection and committed to the surgical procedure. This was repeated after the tumor was resected. A lesion was considered positive if the counts of the lesion minus 3 times the standard deviation of the lesion counts (i.e., lesion counts minus 3 times the square root of the lesion counts) were greater than the mean background counts.
RESULTS At surgery, 11 of the 13 patients had histologically proven cancer. Immunoscintigraphy detected at least one lesion in every patient with tumor and there was 1 false positive scan. CT scan was positive in 7 of the 11 patients that had CT scans performed, yielding 6 true positives, 1 false positive, and 2 false negatives. The gamma probe detected at least one lesion in 9 of the 11 patients with tumor. A total of 21 lesions were identified in the 1 1 patients with tumor. Immunoscintigraphy identified 12, CT scan 9, and the gamma probe 7. Eight of the 21 lesions were less than 1 cm in diameter. Immunoscintigraphy detected none of the lesions less than 1 cm, CT scan identified 2, and the gamma probe identified 6. Immmunoscintigraphy and the gamma probe detected increased radioactivity in the cervical region of 2 patients. In one patient, fine-needle aspiration of palpable nodules was guided by the gamma probe. Three of 3 nodules identified as positive by the probe were positive for metastatic adenocarcinoma of the colon (Fig. 1). The second patient had no palpable nodules but had activity localized to the left scalene lymph node region. The
228
Krag et al. TABLE 11. Mean 10 s Counts of Nonmalignant Tissues Obtained With the Gamma Probe Mean counts Tissue Abdominal aorta Lumbar spine Liver right lobe Spleen Anterior abdominal wall Small bowel Colon Pubic symphysis Pelvic sidewall Sacral prominence Diaphragm
Uncollimated 690
Collimated
579
583 240 1,203 611 8 15 18 86 I23 I74
253
90
812 3,185 1,027 30 21 I07
I99 233
Ratio of collimated to uncollimated 1.2 3.4 2.6 1.52 3.8 1.4
5.9 2.3 I .9
3.3 3. I
There were no adverse reactions resulting from the infusion of the radiolabeled antibody. Fig. 2. Planar scan of ovarian cancer patient showing increased uptake of radiolabeled antibody in cervical lymph.
TABLE I. Mean In Vivo and Ex Vivo Counts Accumulated Over 10 s of Tumors > 3 cm and Their Background
In vivo Ex vivo
Tumors >3 cm
Background
793
523
353
79
probe guided the resection of 3 radioactive scalene lymph nodes (Fig. 2), which on histopathology were determined to have a histiocytic reaction but did not contain tumor. Overall the gamma probe identified a total of 5 lymph In-CYT- 103 but nodes that had increased uptake of were histologically negative for tumor. All the radioactive lymph nodes had histiocytic changes. The gamma probe identified 3 tumor deposits that had not been identified by the surgeon. One lesion was a small metastatic nodule of rectal cancer invading the sigmoid colon, one was a small cluster of lymph nodes with ovarian cancer at the iliac artery bifurcation, and the third was a cluster of periaortic lymph nodes positive for ovarian cancer. The mean in vivo and ex vivo counts acquired over 10 s with the gamma probe for all tumors >3 cm and associated background counts are shown in Table I. The normal tissue background counts acquired over 10 s ranged from a high of 3,185 in the liver to a low of 30 in the anterior abdominal wall. The mean counts acquired over 10 s for various nonmalignant areas of the body both collimated and uncollimated are shown in Table 11.
'
DISCUSSION Intraoperative probes for detecting localized deposits of radioisotopes in tissue have been intermittently used for at least SO years. Until the last decade, detection was limited to radioactive material delivered to tumors in a relatively nonspecific manner. Intraoperative probes were used to detect brain tumors radiolabeled with "phosphorus [9], thyroid tumors radiolabeled with I3'iodine [lo], submucosal lung cancers radiolabeled with 57Co-bleomycin [ 1 I], and ostoid osteoma labeled with "'?echnetium [ 12,131. The identification of carcinoembryonic antigen (CEA) in 1965 [ 141, the development of hybridoma technology in 1975 [IS] and the application of this technology for delivery of radioisotopes in humans in 1978 [ 161 greatly increased the ability to deliver selectively a variety of radioisotopes to tumor tissue [4-61. Clinical trials of intraoperative detection of radiolabeled antibodies began in 1984 with a case report describing intraoperative detection of radiolabeled colorectal cancer [ 171. Subsequent experience by this group at Ohio State has led to the development of a commercially available gamma probe (Neoprobe Corporation, Columbus, OH). They reported that at least 77% of primary and recurrent colorectal cancers were detected, occult lesions were detected, and patient management was improved [ 18-20], The probe is designed to detect tissues radiolabeled with "'iodine in the absence of significant background radiation. Improved tumor to background ratios are expected when background radiation is very low. The downside of using 125. iodine is that ( I ) it is not suitable for preoperative imaging, and (2) it takes a mean of 22 days to achieve a suitable clearance of background activity prior to intraoperative detection.
Gamma Probe Localization
Our approach has been to view intraoperative gamma detection as an extension of preoperative immunoscintigraphy. Our experience has been with the Oncoprobe, a commercially available gamma counter designed for intraoperative use. We found that the probe detected tumor in 9 of the 1 1 patients with cancer. One of the patients with a false negative result had a delay in surgery until 15 days after infusion of "'In-CYT-103. By this time the antibody had cleared from the tumor and the indium had decayed to levels too low for effective gamma counting. The other patient that had a false negative evaluation with the probe had a large lesion in the liver. As is frequently the case with external imaging, high background counts from 'indium in the liver resulted in decreased ability of the gamma probe to discriminate hepatic tumors from background hepatic tissue. Of the 21 lesions identified in the 1 1 patients evaluated, 8 of the lesions were less than 1 cm in diameter. External scanning detected none of the lesions less than 1 cm whereas the probe identified 6. The ability to detect small lesions appears to be a unique feature of intraoperative gamma detection and is related to the ability to place the detector directly on the tissues in question. The gamma probe identified 5 lymph nodes that had increased uptake of '"In-CYT-103 but were histologically negative for tumor. The basis for increased I 'indium presence in occasional tumor-negative lymph nodes has not been determined. This phenomenon has been observed with I 'indium-labeled antibodies other than B72.3 (M Haseman: personal communication). We can speculate that the lymph nodes are processing circulating antigen shed by the tumor which is then targeted in the lymph node by "'In-CYT-103. Immunohistochemistry, however, was unable to detect TAG-72 antigen in the radiolabeled lymph nodes. Alternatively, reactive lymph nodes undergoing a histiocytic reaction may actively bind and accumulate mouse protein in a nonspecific manner. The gamma probe identified 3 tumor deposits that had not been identified by the surgeon. The probe provided evidence for the inadequacy of resection in these cases. More patients will need to be evaluated to better delineate the ability of the gamma probe to identify surgically occult lesions. High background counts remain a limiting factor for identifying small deposits of tumor with the gamma probe. This was particularly evident for lesions in the liver or in the pelvis. Table I1 indicates the range of background counts for nonmalignant areas of the body. It is clear that the liver, large blood vessels, and the bone marrow have high background counts. If the probe is directed toward a high background organ, falsely elevated counts will occur even at a distance. Use of the collimator and care in directing the probe minimizes artificial elevation of counts from a distant source. Cervical metastasis were detected with external gamma camera imaging in one of the eight patients with
''
'
229
colorectal cancer. The radiolabeled cervical lesions were readily localized for biopsy with the gamma probe. A previous study with 'In-CYT- 103 immunoscintigraphy reported a detection rate of extraabdominal colorectal cancer metastasis of 9% [2]. In this subset of patients, the probe may be very useful for guiding the surgeon to extraabdominal lesions detected by gamma camera imaging [2 I]. Confirmation of extraabdominal metastasis would be an important addition to staging of patients with colorectal or ovarian cancer. In conclusion we report on the use of a commercially available gamma probe for intraoperative detection of I I 'indium-labeled monoclonal antibody B72.3 in colorectal and ovarian cancer patients. The gamma probe appears to offer information that is complementary to immunoscintigraphy in that ( I ) it aids the surgeon in localizing intra- and extra-abdominal lesions previously identified by immunoscintigraphy, (2) it localizes lesions too small to be identified by immunoscintigraphy alone, (3) it localizes lesions that might otherwise be missed at surgery, and (4) it provides objective evidence for adequacy of surgical resection of cancer in the abdominal cavity.
''
ACKNOWLEDGMENTS We thank Patrick Newberry for serving as clinical coordinator. This work was supported in part by the Cytogen Corporation. REFERENCES I . Riva P, Moscatelli G , Paganelli G, Benini S, Siccardi A: Antibody guided diagnosis: An Italian experience on CEA-expressing tumours. Int. J. Cancer Suppl. 2.1 14-120, 1988. 2. Doerr Rj, Hani AN, Krdg DN, Mitchell E: Radiolabeled antibody imaging in the management of colorectal cancer: Results of a multicenter clinical study. Ann Surg 214:l 18-124, 1991. 3. Patt YZ, Lamki LM, Shanken J, et al.: Imaging with Indium"' labeled anti-carcinoembryonicantigen monoclonal antibody ZCE025 of recurrent colorectal or carcinoembryonic antigen-producing cancer occult metastases. J Clin Oncol 8:124&1254, 1990. 4. Haller DG: Monoclonal antibody imaging in the management of patients with colorectal cancer. J CIin Oncol6:1213-121.5, 1988. 5 . Larson SM: Lymphoma, melanoma, colon cancer: Diagnosis and treatment with radiolabeled monoclonal antibodies. Radiology 16.5 :297-304, 1987. 6. Byers VS, Baldwin RW: Therapeutic strategies with monoclonal antibodies and immunoconjugates. Immunology 65:329-335, 1988. 7. Epstein AL, Khawli LA: Tumor biology and monoclonal antibodies: Overview of basic principles and clinical considerations. Antibiot lmmunoconjug Radiopharm 4:373-384, I99 I . 8. Barber HB, Barrett HH, Woolfenden JM, Myers KJ, Hickernell TS: Comparison of in vivo scintillation probes and gamma cameras for detection of small, deep tumours. Phys Med Biol 34:727739, 1989. 9. Selverstone B, Sweet WH, Robinson CV: The clinical use of radioactive phosphorus in the surgery of brain tumors. Ann Surg 130~643-651, 1949. 10. Harris CC, Bigelow RR, Francis HE, Kelley GG, Bell RP: A CsI(T1)-crystal surgical scintillation probe. Nucleonics 14:102108, 1956. 1 1 . Woolfdenden JM, Nevin WS, Barber HB, Donahue DJ: Lung
230
12.
13. 14. 15.
16.
17.
Krag et al. cancer detection using a miniature sodium iodide detector cobalt-57 bleomycin. Chest 85:84-88, 1984. Colton CL, Orth ED, Hardy G: J Bone Joint Surg 65A:10191022, 1983. Szypryt EP, Hardy JG, Colton CL: J Bone Joint Surg 685(4):643646. 1986. Gold P, Freedinan SO: Demonstration of tumor-specific antigens in human colonic carcinomas by immunological tolerance and absorption techniques. J Exp Med 121:439, 1965. Kohler G , Milstein C: Continuous cultures of fused cells secreting antibody of predefined specificity. Nature 256:495-497, 1975. Goldenberg DM, Deland FH, Kim E, et al.: Use of radiolabeled antibodies to carcino-embryonic antigen for the detection and localization of diverse cancers by external photoscanning. N Engl J Med 298: 1384-1388, 1978. Aitkin DR, Hinkle GH, Thurston MO: A gamma-detecting probe for radioimmune detection of CEA-producing tumors: Successful experimental use and clinical case report. Dis Colon Rectum 27:279, 1984.
18. NierodaCA, Mojzisik C, Sardi A, FerraraPJ, Hinkle G , Thurston MO, Martin EW Jr: Radioimmunoguided surgery in primary eoIon cancer. Cancer Prevention and Detection 14:651-656, 1990. 19. Martin EW Jr, Mojzisik CM, Hinkle GH Jr, Sampsel J , Siddl MA, Tuttle SE, Sickle-Santanello 5 , Colcher D, Thurston MO, Bell JG, Farrar WB, Schlom J: Radioimmunoguided surgery using monoclonal antibody. Am J Surg 156:38&392, 1988. 20. Hinkle GH, Mojzisik CM, Loesch JA, eta].: The evolution of the Radioimmunoguided Surgery System@):An innovative technique for the intraoperative detection of tumors. Antibody Inimunol Radiopharm 4:339-358, 1991. 21. Kuhn JA, Corbisiero RM, Buras R R , Carroll RG, 5eatty JD: Intraoperative gamma detection probe combined with preoperative Iindium labeled antibody imaging for localization of recurrent carcinoma. Presented at the Society of Surgical Oncology, 1991.
'
