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.

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.

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.

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.

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.

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.

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.

Reoperation directed by carcinoembryonic antigen level: the importance of a thorough preoperative evaluation

Many asymptomatic patients suspected to have recurrent colorectal cancer based on an elevated carcinoembryonic antigen level will be spared unnecessary operation if strict attention is paid to their preoperative evaluation. Liver and renal function should be assessed. Unresectable extraabdominal and intraabdominal recurrence or metastases should be excluded. Patients being evaluated for recurrence after curative resection of a rectosigmoid cancer should undergo a bone scan. Having satisfactorily ensured normal results for these investigations, the surgeon should then proceed to search for an intraabdominal source of tumor recurrence.

Radioimmunoguided surgery using the monoclonal antibody B72.3 in colorectal tumors

The authors have developed a hand-held gamma-detecting probe (GDP) for intraoperative use that improves the sensitivity of external radioimmunodetection. Radiolabeled monoclonal antibody (MAb) B72.3 was injected in six patients with primary colorectal cancer and 31 patients with recurrent colorectal cancer an average of 16 days preoperatively. The GDP localized the MAb B72.3 in 83 percent of sites. The technique, known as a radioimmunoguided surgery (RIGS) system did not alter the surgical procedure in patients with primary colorectal cancer but did alter the approach in 26 percent (8/31) of patients with recurrent colorectal cancer. Two patients avoided unnecessary liver resections and two underwent extraabdominal approaches to document their disease. The RIGS system may influence the short-term morbidity and mortality of surgery for colorectal cancer. Larger series and longer follow-up are needed to determine whether the RIGS system confers a survival advantage to the patient with colorectal cancer.

Variation in recognition of recurrent colonic cancer by different CEA assays

Forty-six patients were followed with serial CEA determinations by two CEA assays after curative resection of a primary or recurrent colonic cancer. Thirteen have had a histologically proven recurrence, of which only six (46 percent) were predicted by both assays, while seven (54 percent) were predicted by one assay only. The assays appear complementary in indicating tumor recurrence, and preliminary findings suggest that it may be beneficial to follow patients with more than one CEA assay after resection of a primary or recurrent colonic cancer.

Neurologic complications of gastric partitioning

Six patients who had gastric reduction for morbid obesity suffered severe complex neurologic disturbances that included confusion and inappropriate behavior. All were profoundly weak or paraplegic, and recovery was delayed and incomplete. Encephalopathy was documented by EEG and neurologic examination. The cause is uncertain. Acute catabolism of lipid may predispose to damage of the nervous system, but relative vitamin deficiency is a more obvious and treatable explanation.

Complications of gastric restrictive operations in morbidly obese patients

After gastric restrictive procedures in morbidly obese persons, the reported morbidity is 15 per cent, and mortality is 1 per cent. Pulmonary embolism continues to be the most serious complication. Gastric perforations must be recognized early and treated promptly, as must the common postoperative complications: pneumonia, atelectasis, and wound infection.

Prevention of early failure of stapled gastric partitions in treatment of morbid obesity

Clinical observations and data obtained in animal experiments indicate that the gastric partition is most susceptible to the stresses of solid food boluses between the second and sixth postoperative weeks, predisposing the operation to early failure. Tensiometer studies indicate that by the sixth to the eighth postoperative week, the staple line is as strong as uninjured stomach. Liquid diet restriction for eight weeks after gastric partitioning reduces the incidence of early operative failure. Until safe and effective methods of increasing staple line strength are established, all patients having gastric partitioning should be restricted to a liquid diet for eight weeks after operation.