Pilot study evaluating the intraoperative localization of radiolabeled monoclonal antibody CC83 in patients with metastatic colorectal carcinoma.

BACKGROUND: CC83, a second-generation monoclonal antibody (MAb) against tumor-associated glycoprotein TAG-72 has been shown to have a higher affinity constant than the anti-TAG MAbs CC49 and B72.3. Clinical studies have shown the effectiveness of both CC49 and B72.3 radiolabeled MAbs in localizing colorectal carcinoma with a hand-held gamma-detecting probe during operation. This current study was designed to assess the safety and tumor-binding ability of radiolabeled CC83 MAb in this setting. METHODS: Seventeen patients with recurrent colorectal cancer underwent intravenous injection with CC83 MAb radiolabeled with iodine 125 (2.0 mCi125I/0.2 mg CC83 MAb). Exploratory laparotomy was carried out 21 to 28 days after injection, consisting of a thorough traditional exploration followed by a survey with a hand-held gamma-detecting probe. All traditionally suspicious and probe-positive tissue was either biopsied or resected and subsequently examined for the presence of carcinoma by using routine histochemical staining techniques. RESULTS: Thirty-two sites were identified as suspicious for cancer by traditional surgical exploration and 39 through intraoperative survey with a hand-held gamma-detecting probe in the seventeen patients completing the study. Biopsy or resection yielded 27 tumor sites when tissue was evaluated by using routine hematoxylin-eosin staining. All 27 tumor sites were localized by the radiolabeled CC83 MAb, whereas 12 additional sites were RIGS positive but hematoxylin-eosin negative, resulting in a sensitivity and positive predictive value of 100% and 69%, respectively. Traditional methods of exploration detected 23 of 27 tumor sites (85% sensitivity), and nine false-positive sites were recorded (72% positive predictive value). Occult tumor was found by using CC83 MAb in four (15%) of 27 sites, altering the surgical plan in three patients. CONCLUSIONS: This initial study indicates that CC83 MAb, when used with RIGS, is safe and sensitive in detecting recurrent intraabdominal colorectal cancer.

The significance of intraoperative periportal lymph node metastasis identification in patients with colorectal carcinoma.

BACKGROUND: Nine patients who underwent Radioimmunoguided Surgery (RIGS) (Neoprobe Corporation, Dublin, OH) procedures for colorectal cancer were found to have disease recurrence in the periportal area. This led to a retrospective study to determine whether periportal lymph node involvement could have been predicted intraoperatively for these patients. METHODS: One hundred twenty-four patients underwent second-look RIGS for recurrent colon and rectal cancer from 1986 to 1992. The monoclonal antibody (MAb) B72.3 was administered as the carrier agent to 87 patients and the CC49 second-generation MAb was administered to 37 patients. Both MAbs were radiolabeled with Iodine-125. RESULTS: Periportal lymph nodes with RIGS-positive tissue were found in 47 (38%) patients, hematoxylin and eosin-positive lymph nodes were found in 13 of 47, and in further immunohistochemical studies performed for 31 of the remaining 34 patients, positive lymph nodes were found in 8, resulting in an incidence of 48% (21/44). A critical review of the nine patients' charts who later presented with a tumor mass in the periportal area demonstrated intraoperative gamma-detecting probe counts in ratios three to five times that of the normal adjacent tissues in the periportal area at the time of first exploration. Probe-directed biopsy was reported to be histologically negative for tumor in these patients, and, thus, the surgeon proceeded assuming the periportal area to be negative. A retrospective study of the periportal lymph nodes of these patients using cytokeratin immunohistochemical analysis identified tumor in five (56%). CONCLUSIONS: These findings suggest that the RIGS system may be a valuable method of intraoperative prediction and detection of periportal lymph node metastasis.

Identification of occult micrometastases in pericolic lymph nodes of Duke's B colorectal cancer patients using monoclonal antibodies against cytokeratin and CC49. Correlation with long-term survival.

BACKGROUND: Patients with transmurally invasive, lymph node negative colorectal carcinoma (Dukes' B) have a 5-year survival rate ranging from 53.9% to 84.9%. The authors postulate that patients with Dukes' B colon cancer who die of their disease have occult micrometastases in their pericolic lymph nodes at the time of original diagnosis. In an attempt to identify these occult micrometastases, pericolic lymph nodes from Dukes' B colon cancer resections were stained retrospectively with antibodies against cytokeratin (anti-keratin AE1/AE3, Boehringer Mannheim, Indianapolis, IN) and CC49 (a second-generation monoclonal antibody directed against TAG-72. METHODS: The authors reviewed all Dukes' B (transmurally invasive, lymph node negative) primary colorectal carcinoma resection specimens from the surgical pathology files of the Ohio State University Hospitals between 1984 and 1987. Survival data were obtained from the Tumor Registry of the Arthur G. James Cancer Hospital and Research Institute, Columbus, Ohio. The results were analyzed by univariate and multivariate analysis. RESULTS: Fifty cases with 568 lymph nodes (11.3 per case) were examined with each antibody using standard immunoperoxidase techniques. Positive staining for cytokeratin was seen in 14 patients (33 lymph nodes), 6 of whom died of colon cancer within 66 months (43%). Only 1 of the 36 patients with cytokeratin-negative lymph nodes died of colon cancer over the same time period (3%, P = 0.0009 univariate, P = 0.0013 multivariate). There was no significant difference in survival between the CC49-positive and CC49-negative groups. CONCLUSION: Immunoperoxidase techniques are capable of identifying micrometastatic disease in lymph nodes missed by routine hematoxylin and eosin staining. Further, the presence of cytokeratin-positive cells within lymph nodes correlated with a significantly poorer prognosis. Therefore, cytokeratin staining of pericolic lymph nodes in patients with Dukes' B colorectal cancer is recommended. Larger multicenter studies are needed, however, to confirm these results and to evaluate the appropriateness of adjuvant chemotherapy in patients whose disease is upstaged by immunohistochemical staining.

Intraoperative localization of neuroendocrine tumors with 125I-TYR(3)-octreotide and a hand-held gamma-detecting probe.

BACKGROUND: This study evaluates a novel method of intraoperative localization of endocrine gastroenteropancreatic tumors with a hand-held gamma-detecting probe to detect in situ tumor binding of the radioiodinated somatostatin analog 125I-TYR(3)-octreotide. METHODS: Seven patients with biochemical and radiologic evidence of a specific endocrine tumor, one patient with biochemical evidence of gastrinoma but no tumor localized by conventional imaging techniques, and four patients with equivocal preoperative biochemical or radiologic study results but suspected of harboring a neuroendocrine tumor underwent abdominal exploration with intraoperative injection of 125I-TYR(3)-octreotide. 298 +/- 63 microCi. A hand-held gamma-detecting probe was used during operation to determine whether gross tumor accumulated the radiolabeled analog and occult tumor could be detected. Positive uptake was defined as tumor/background ratios exceeding 2:1. RESULTS: The tumor in all seven patients with gross disease accumulated 125I-TYR(3)-octreotide. Occult tumor beyond that appreciated with preoperative imaging or by routine operative exploration was detected in a patient with carcinoid tumor. In the patient with the occult gastrinoma the probe detected the lesion within the duodenal bulb before duodenotomy and also predicted what proved histologically to be positive peripancreatic adenopathy. There was a single false-positive reading from the stomach in a patient with suspected carcinoid tumor in whom no tumor could be found grossly or histologically. A pancreatic mass that probed negative proved to be an adenocarcinoma of ductal origin. CONCLUSIONS: Tumor-specific peptide-receptor binding can be detected in situ with 125J-TYR(3)-octreotide and a hand-held gamma-detecting probe. This technique may facilitate neuroendocrine tumor localization and operative cytoreduction.

Role of carcinoembryonic antigen in predicting resectability of recurrent colorectal cancer.

The reported low resectability rate for patients with recurrent colorectal cancer who have carcinoembryonic antigen (CEA) levels > 11 has led us to perform this study. One hundred twenty-four patients who underwent Radioimmunoguided Surgery (RIGS) procedures for recurrent colorectal cancer from 1986 to the present were studied. In surgery, all patients underwent a traditional exploration followed by survey with a hand-held, gamma-detecting probe to detect preinjected radiolabeled monoclonal antibodies attached to cancer cells. Sites of metastases included: 72 liver (58.1 percent), 23 pelvis (18.5 percent), 15 distant lymph nodes (12.1 percent), 2 anastomotic (1.6 percent), and 12 other sites (9.7 percent). The resectability rate was 43.5 percent (54 patients). The mean preoperative CEA level for patients with resectable disease was significantly lower than for patients with unresectable disease (P = 0.017): unresectable--mean, 87.1; SD, 141.0; minimum, 0.3; maximum, 501; resectable--mean, 36.6; SD, 59.3; minimum, 0.3; maximum, 329. The CEA level for patients with liver metastasis did not vary significantly from those patients without metastasis: 70 vs. 58.2 (P = 0.58). Those patients with resectable liver tumors had lower mean CEA levels than those with unresectable liver, approaching significance: 41.6 vs. 91.9 (P = 0.065). Other metastatic sites had a mean CEA level of: pelvic, 72.6; distant lymph nodes, 47.8; anastomotic, 2.7; and other sites, 53.8. These data suggest that there is a significant difference between the preoperative CEA level of the resectable and unresectable recurrent colorectal cancer patients, but the large standard deviation does not justify abandonment of exploration for any CEA level.

Radioimmunoguided radical prostatectomy and lymphadenectomy.

BACKGROUND: The feasibility of using radioimmunoguided surgery (RIGS) (Neoprobe Corp., Columbus, OH) for intraoperative detection of prostate cancer was examined in a pre-Phase I clinical study involving 10 patients having radical prostatectomy and lymphadenectomy. METHODS: Patients were injected with iodine 125-radiolabeled B72.3 monoclonal antibody, which has been shown previously to bind to TAG-72, a pancarcinoma and oncofetal antigen. At a mean of 26 days after injection, RIGS was performed with a specially designed intraoperative gamma-detecting probe. RESULTS: By comparing probe counts with counts of appropriate background tissues, the RIGS system successfully localized tumor to the prostate of all 10 patients. Clinically occult and histologically confirmed bilateral intraprostatic tumor was identified in three patients. One additional patient had bilateral positive intraprostatic probe count ratios with the RIGS technique; on histologic examination, tumor was identified unilaterally, and extensive high-grade prostatic intraepithelial neoplasia was found on the contralateral side. Probe count ratios were positive in the lymph nodes of three patients; two had tumor confirmed histologically. CONCLUSIONS: The current investigation supports the feasibility of the RIGS technique and the need for additional studies.

Intraoperative detection of colorectal cancer with radioimmunoguided surgery and CC49, a second-generation monoclonal antibody.

Radioimmunoguided surgery (RIGS) has been employed intraoperatively in cases of colorectal cancer to assess the extent of local tumor spread and metastatic disease. This technique uses radiolabeled monoclonal antibodies (MAbs) directed against tumor-associated antigens, and a hand-held gamma-detection probe to detect the radiolabel fixed to tumor tissue. Recently introduced is an MAb directed against tumor-associated glycoprotein (anti-TAG), CC49. Sixty patients were entered into the initial study. Eighteen of 21 (86%) primary tumors were localized by the CC49 MAb and the gamma-detecting probe. Twenty-nine of 30 (97%) recurrent tumors were localized. Antibody dose did not affect localization. Specimens were divided into tissue types I through IV, based on antibody localization and hematoxylin and eosin (H&E) staining: type I, RIGS (-) and histologically (-); type II, RIGS (-) and histologically (+); type III, RIGS (+) and histologically (-); type IV, RIGS (+) and histologically (+). Type IV tissue were further classified by whether they were grossly apparent, IVa, or grossly inapparent, IVb (occult). Occult tumor found by RIGS and confirmed by H&E staining (type IV) had localization ratios similar to RIGS-positive, histology-negative tissue (type III). Traditionally found cancer (type IV) had significantly higher ratios. In 12 of 24 patients (50%) with primary tumors and 14 of 30 patients (47%) with recurrent tumors, RIGS with CC49 altered the planned operative procedure. Radioimmunoguided surgery with CC49 provides useful, immediate intraoperative information not available by other techniques.

Radioimmunoguided surgery challenges traditional decision making in patients with primary colorectal cancer.

BACKGROUND: Initial experience with the radioimmunoguided surgery system (RIGS) has been found to impact on decision making in patients with recurrent colorectal cancers. Reported here is experience with RIGS-influenced therapeutic decisions in patients with primary colorectal cancer. METHODS: Thirty-six evaluable patients with primary cancers were injected with the second-generation anti-tumor-associated glycoprotein antibody CC49 labeled with 1 to 2 mCi iodine 125. Pharmacokinetic determination and precordial counts were obtained after injection and weekly until levels were less than 20 counts/2 sec. At surgery abdominal and pelvic explorations were performed, first traditionally by inspection and palpation and then with the hand-held, gamma-detecting probe. RIGS-positive tissue was considered cancerous and removed if possible. RESULTS: Thirty patients (83%) had positive antibody localization at surgery. Of those patients with localization, in 24 (80%) additional information was obtained at the time of surgery. In 11 patients (34%) staging changes were made as a result of RIGS exploration. New findings resulted in operative changes in nine patients (25%). Eleven (30%) of the original 36 patients became eligible for adjuvant chemotherapy based on current recommendations because of RIGS findings. CONCLUSIONS: In conclusion, the RIGS system provides immediate staging information that impacts on therapeutic interventions, challenging the adequacy of traditional procedures alone for primary colorectal cancer exploration.

Radioimmunoguided surgery using iodine 125 B72.3 in patients with colorectal cancer.

Preliminary data using B72.3 murine monoclonal antibody labeled with iodine 125 suggested that both clinically apparent as well as occult sites of colorectal cancer could be identified intraoperatively using a hand-held gamma detecting probe. We report the preliminary data of a multicenter trial of this approach in patients with primary or recurrent colorectal cancer. One hundred four patients with primary, suspected, or known recurrent colorectal cancer received an intravenous infusion of 1 mg of B72.3 monoclonal antibody radiolabeled with 7.4 x 10 Bq of iodine 125. Twenty-six patients with primary colorectal cancer and 72 patients with recurrent colorectal cancer were examined. Using the gamma detecting probe, 78% of the patients had localization of the antibody in their tumor; this included 75% of primary tumor sites and 63% of all recurrent tumor sites; 9.2% of all tumor sites identified represented occult sites detected only with the gamma detecting probe. The overall sensitivity was 77% and a predictive value of a positive detection was 78%. A total of 30 occult sites in 26 patients were identified. In patients with recurrent cancer, the antibody study provided unique data that precluded resection in 10 patients, and in another eight patients it extended the potentially curative procedure.

Radioimmunoguided surgery (RIGS) in recurrent colorectal cancer.

Since 1986, 191 patients with recurrent colorectal cancer have undergone surgical exploration 2 to 43 days after injection of 1.0 to 0.25 mg of monoclonal antibody (MAb) (B72.3 or 17-1A) radiolabeled with 5.0 to 1.0 mCi of 125I. The intraoperative use of a hand-held gamma detector (Neoprobe 1000) demonstrated that MAb identified tumor in 73% of cases. Clearer intraoperative definition of tumor margins and identification of occult tumor assisted the surgeon in the resection of liver metastases as well as nodal and pelvic disease. Unsuspected nodal disease was identified. The external use of the Neoprobe to scan the sacral region and intrarectal and intravaginal use led to the avoidance of operative procedures by defining inoperable disease. In approximately 25% of cases, the surgical procedure was modified based on Neoprobe findings. RIGS system provides a method of immediate intraoperative staging which may prevent additional recurrences, lead to earlier institution of adjuvant therapy, and result in improved survival.

Intraoperative radioimmunodetection of ovarian cancer using monoclonal antibody B72.3 and a portable gamma-detecting probe.

To assess the value of radioimmunoguided surgery in the intraoperative detection of ovarian cancer, we used monoclonal antibody B72.3, radiolabeled with 125I, and a hand-held gamma-detecting probe in 13 women with ovarian cancer undergoing exploratory laparotomy. B72.3, which recognizes a tumor-associated glycoprotein, TAG 72, was injected 12-29 days preoperatively (intraperitoneally in four cases, intravenously in nine, and by both routes in one). Intraoperatively, the abdomen was surveyed with the probe and probe counts were correlated with biopsies and excised specimens studied by routine histologic stains. Probe counts were positive in four of seven evaluable patients with histologically confirmed disease. In three of these four patients, the probe detected cancer in specimens interpreted as normal on frozen histologic sections. The probe also identified microscopic cancer in the one patient who had no gross disease. The specificity of the probe was 70%. Preoperative computed tomography was normal in all patients, including those with tumors as large as 3 cm. This pilot study shows the ability of radioimmunoguided surgery to detect occult ovarian cancer.

Radioimmunoguided surgery in primary colon cancer.

Radioimmunoguided surgery (RIGS), the intraoperative use of a hand-held gamma detecting probe (GDP) to identify tissue containing radiolabeled monoclonal antibody (MAb), was performed upon 30 patients with primary colon carcinoma. Each patient received an intravenous injection of MAb B72.3 (1.0 to 0.25 mg) radiolabeled with 125I (5.0 to 1.0 mCi) 8 to 34 days before exploration. The GDP was used to measure radioactivity in colon tissue, tumor bed, nodal drainage areas, and areas of suspected metastases. Antibody localized to histologically documented tumor in 23 of 30 patients (77%). Tumor margins were more clearly defined in 20 of 30 patients (67%). GDP counts led to major alterations in surgical resection in five patients (17%) and changes in adjuvant therapy in four (14%). GDP counts identified occult liver metastases in two patients (7%) and correctly indicated the benign nature of liver masses in three (10%). In four patients (13%), occult nodal metastases were identified. RIGS can precisely delineate tumor margins, define the extent of nodal involvement, and localize occult tumor, providing a method of immediate intraoperative staging that may lessen recurrences and produce higher survival rates.

The impact of radioimmunoguided surgery (RIGS) on surgical decision-making in colorectal cancer.

Radioimmunoguided surgery (RIGS system) was performed in ten patients with rectal or low sigmoid colon carcinoma with the use of a hand-held gamma detector (Neoprobe 1000) intraoperatively and externally after injection of radiolabeled (125I) monoclonal antibody to detect pelvic and metastatic tumor. Fifteen procedures, including six exploratory laparotomies, four transperineal explorations, two transsacral explorations, one transvaginal biopsy, one brachytherapy, and one transanal polypectomy, were performed. Two patients had previous low anterior resection, seven abdominoperineal resection, and one a rectal polypectomy. Five patients had previous pelvic radiation therapy. Reoperation was indicated by elevated CEA levels in seven patients (70 percent), persistent pelvic pain in six (60 percent), and a suspicious radiologic study in seven (70 percent). RIGS system localized tumors verified by histopatholoy in all ten patients (100 percent); one patient with a positive CT scan and probe findings lacked histopathologic confirmation on frozen section, but had a tumor confirmed on permanent histology. Five major abdominal operations were avoided; in five patients major modifications were made in the surgical procedure based on probe findings. Six received chemotherapy or radiation therapy based on findings of the RIGS system. In six patients with negative or equivocal CT scans, the RIGS system localized histopathologically confirmed tumor. Major abdominal procedures can be avoided, the surgical approach modified, and other modes of therapy instituted earlier with the use of the RIGS system.

Radioimmunoguided surgery in recurrent colorectal cancer: the role of carcinoembryonic antigen, computerized tomography, and physical examination.

From January 1986 to December 1987, 32 patients with recurrent colorectal cancer had second-look radioimmunoguided surgery (RIGS system). All patients had pathologic confirmation of recurrence. The RIGS system identified 81% of recurrences, and in six patients recurrent tumor was identified only by RIGS. All patients had physical examination, carcinoembryonic antigen (CEA) assay, and computerized tomography of the abdomen and pelvis. Detection of recurrence was based on symptoms in six, elevated CEA value in 25, and physical examination in one. The CEA was elevated preoperatively in 30 patients; two false-negative results occurred in symptomatic patients who had pelvic recurrence. The median CEA value in those with liver recurrence was 30 ng/ml (range 5.2 to 298) and for pelvic recurrence 13 ng/ml (range 1.9 to 31) (P less than .05). The overall sensitivity of CT was 41% (abdomen other than liver 37%, liver 56%, and pelvis 22%). The combination of elevated CEA, symptoms, and physical findings identified 100% of recurrences. We conclude that a rising CEA remains the most accurate indicator of recurrence. CT should not be done routinely to detect recurrent colorectal cancer unless CEA is elevated or the patient is symptomatic. In our study the intraoperative use of the RIGS system aided the surgeon in identifying occult tumors.

The use of a hand-held gamma detector improves the safety of isolated limb perfusion.

We used two hand-held gamma-detecting probes (GDP) (Neoprobe 1000 system) capable of detecting small gamma emissions to monitor leakage in patients undergoing hyperthermic isolated limb perfusion (HILP) who received 800 microCi Technetium 99m pentetate through the perfusate. The percentage of gamma-ray leakage was calculated by a simultaneous reading of two probes at 1-minute intervals (one over the precordial area and one over the thigh) and this was compared to results of simultaneous blood sampling from the perfusate and systemic circulation at 15-minute intervals for gamma well counting (GWC). The percentage of leakage recorded by the GDPs was essentially identical to that detected by the GWC (7.3% and 8.2%, respectively at the conclusion of the perfusion). The GDP gives an immediate and accurate indication of the percentage of leakage during HILP, making it a safer procedure.

Staging of carcinoma of the breast using a hand-held gamma detecting probe and monoclonal antibody B72.3.

Radioimmunoguided Surgery (RIGS) uses a hand-held gamma detecting probe to identify radiolabeled monoclonal antibodies (Mab). Fourteen patients with carcinoma of the breast proved at biopsy received Mab B72.3 (5 millicuries of 125I per 1 milligram, Iodo-Gen method) intravenously six to 26 days before exploration. Probe counts were measured intraoperatively in mammary tissue and axillary lymph nodes. In the mammary tissue, the RIGS system identified tumor that was histologically confirmed in seven of eight patients and confirmed the absence in four of six patients. Probe counts were suspicious for tumor that was not proved histologically in two of 14 patients. Unsuspected tumor was identified in three of 14 patients. In axillary tissue, probe counts identified one of two tumors that were confirmed histologically and verified the absence of tumor in eight of 12 patients. Probe counts in axillary tissue were suspicious for tumor that could not be documented histologically in four of 14 patients. RIGS appears to be able to identify residual, subclinical and multicentric carcinoma of the breast and accurately delineate the pattern of antigenic drainage of tumor into adjacent lymph nodes.

An assessment of prolonged reactivity of seven monoclonal antibodies against CX-1 tumor xenografts using a hand-held gamma-detecting probe.

The biodistribution and kinetics of 7 monoclonal antibodies (MAb) with known reactivity against CX-1 tumor were examined over 21 days using a hand-held gamma-detecting probe (Neoprobe system). Twenty-eight immuno-deprived (athymic) nude mice implanted with human colon adenocarcinoma CX-1 xenografts were injected intraperitoneally with 50 microCi of 125I-labeled antibodies (4 mice/antibody). Of the 7 monoclonal antibodies, 4 were anti-CEA (MA, MB, MC, and MD), 2 were anti-TAG 72 (B72.3 NCI and B72.3 fermented) and one was anti-colorectal cancer (17-1A). Daily probe counts were recorded in duplicate over the tumor site and the contralateral nontumor site (background), and tumor-to-background (Tu/Bkg) ratios were calculated. Animals were sacrificed on day 21, and blood, heart, liver, spleen, lungs, kidneys, intestine, muscle, and the tumor were removed for gamma well counting. All antibodies identified the tumor as early as 24 h postinjection and specific tumor localization improved over time. Patterns of prolonged tumor binding varied considerably from one antibody to another, although all but one (MB) showed continuously increasing Tu/Bkg ratios. These data indicate progressive clearance of the antibodies from the background tissue and a persistence of labeled MAb activity in tumor resulting in improved tumor localization with increasing postinjection time.

Intra-abdominal recurrence of colorectal cancer detected by radioimmunoguided surgery (RIGS system).

Since 1986, 32 patients with metastatic colorectal cancer have undergone second-look radioimmunoguided surgery (RIGS system). The primary tumor was located in the right and transverse colon in 11 patients, left and sigmoid colon in 16, and rectum in five. The carcinoembryonic antigen level was elevated in 30 patients (94%); all patients underwent a computed tomographic scan of the abdomen and pelvis. The overall sensitivity of the computed tomographic scan was 41% (abdomen other than liver, 27%; liver, 58%; and pelvis, 22%). The RIGS system identified recurrent tumor in 81% of the patients. The most common site of metastasis was the liver (41%), independent of the primary location. Local/regional recurrences alone accounted for 40% of all recurrences. In six patients (18%), recurrent tumor was found only with the RIGS system. The RIGS system is more dependable in localizing clinically obscure metastases than other methods, and carcinoembryonic antigen testing remains the most accurate preoperative method to indicate suspected recurrences.

Radioimmunoguided surgery using monoclonal antibody.

The potential proficiency of radioimmunoguided surgery in the intraoperative detection of tumors was assessed using labeled monoclonal antibody B72.3 in 66 patients with tissue-proved tumor. Monoclonal antibody B72.3 was injected 5 to 42 days preoperatively, and the hand-held gamma-detecting probe was used intraoperatively to detect the presence of tumor. Intraoperative probe counts of less than 20 every 2 seconds, or tumor-to-adjacent normal tissue ratios less than 2:1 were considered negative (system failure). Positive probe counts were detected in 5 of 6 patients with primary colon cancer (83 percent), in 31 of 39 patients with recurrent colon cancer (79 percent), in 4 of 5 patients with gastric cancer (80 percent), in 3 of 8 patients with breast cancer (37.5 percent), and in 4 of 8 patients with ovarian cancer (50 percent) undergoing second-look procedures. Additional patients in each group were scored as borderline positive. Overall, radioimmunoguided surgery using B72.3 identified tumors in 47 patients (71.2 percent), bordered on positive in 6 patients (9.1 percent), and failed to identify tumor in 13 patients (19.7 percent). Improved selection of patients for antigen-positive tumors, the use of higher affinity second-generation antibodies, alternate routes of antibody administration, alternate radionuclides, and more sophisticatedly bioengineered antibodies and antibody combinations should all lead to improvements in radioimmunoguided surgery.