Intra-abdominal patterns of disease dissemination in colorectal cancer identified using radioimmunoguided surgery

Survival Advantage Following TAG-72 Antigen-Directed Cancer Surgery in Patients With Colorectal Carcinoma: Proposed Mechanisms of Action

Patients with colorectal carcinoma (CRC) continue to have variable clinical outcomes despite undergoing the same surgical procedure with curative intent and having the same pathologic and clinical stage. This problem suggests the need for better techniques to assess the extent of disease during surgery. We began to address this problem 35 years ago by injecting patients with either primary or recurrent CRC with 125I-labeled murine monoclonal antibodies against the tumor-associated glycoprotein-72 (TAG-72) and using a handheld gamma-detecting probe (HGDP) for intraoperative detection and removal of radioactive, i.e., TAG-72-positive, tissue. Data from these studies demonstrated a significant difference in overall survival data (p < 0.005 or better) when no TAG-72-positive tissue remained compared to when TAG-72-positive tissue remained at the completion of surgery. Recent publications indicate that aberrant glycosylation of mucins and their critical role in suppressing tumor-associated immune response help to explain the cellular mechanisms underlying our results. We propose that monoclonal antibodies to TAG-72 recognize and bind to antigenic epitopes on mucins that suppress the tumor-associated immune response in both the tumor and tumor-draining lymph nodes. Complete surgical removal of all TAG-72-positive tissue serves to reverse the escape phase of immunoediting, allowing a resetting of this response that leads to improved overall survival of the patients with either primary or recurrent CRC. Thus, the status of TAG-72 positivity after resection has a significant impact on patient survival.

Comparison of two threshold detection criteria methodologies for determination of probe positivity for intraoperative in situ identification of presumed abnormal 18F-FDG-avid tissue sites during radioguided oncologic surgery

Background

Intraoperative in situ identification of ¹⁸F-FDG-avid tissue sites during radioguided oncologic surgery remains a significant challenge for surgeons. The purpose of our study was to evaluate the 1.5-to-1 ratiometric threshold criteria method versus the three-sigma statistical threshold criteria method for determination of gamma detection probe positivity for intraoperative in situ identification of presumed abnormal ¹⁸F-FDG-avid tissue sites in a manner that was independent of the specific type of gamma detection probe used.

Methods

From among 52 patients undergoing appropriate in situ evaluation of presumed abnormal ¹⁸F-FDG-avid tissue sites during ¹⁸F-FDG-directed surgery using 6 available gamma detection probe systems, a total of 401 intraoperative gamma detection probe measurement sets of in situ counts per second measurements were cumulatively taken.

Results

For the 401 intraoperative gamma detection probe measurement sets, probe positivity was successfully met by the 1.5-to-1 ratiometric threshold criteria method in 150/401 instances (37.4%) and by the three-sigma statistical threshold criteria method in 259/401 instances (64.6%) (P < 0.001). Likewise, the three-sigma statistical threshold criteria method detected true positive results at target-to-background ratios much lower than the 1.5-to-1 target-to-background ratio of the 1.5-to-1 ratiometric threshold criteria method.

Conclusions

The three-sigma statistical threshold criteria method was significantly better than the 1.5-to-1 ratiometric threshold criteria method for determination of gamma detection probe positivity for intraoperative in situ detection of presumed abnormal ¹⁸F-FDG-avid tissue sites during radioguided oncologic surgery. This finding may be extremely important for reshaping the ongoing and future research and development of gamma detection probe systems that are necessary for optimizing the in situ detection of radioisotopes of higher-energy gamma photon emissions used during radioguided oncologic surgery.

Intraoperative detection of ¹⁸F-FDG-avid tissue sites using the increased probe counting efficiency of the K-alpha probe design and variance-based statistical analysis with the three-sigma criteria

BACKGROUND

Intraoperative detection of (18)F-FDG-avid tissue sites during 18F-FDG-directed surgery can be very challenging when utilizing gamma detection probes that rely on a fixed target-to-background (T/B) ratio (ratiometric threshold) for determination of probe positivity. The purpose of our study was to evaluate the counting efficiency and the success rate of in situ intraoperative detection of (18)F-FDG-avid tissue sites (using the three-sigma statistical threshold criteria method and the ratiometric threshold criteria method) for three different gamma detection probe systems.

METHODS

Of 58 patients undergoing (18)F-FDG-directed surgery for known or suspected malignancy using gamma detection probes, we identified nine (18)F-FDG-avid tissue sites (from amongst seven patients) that were seen on same-day preoperative diagnostic PET/CT imaging, and for which each (18)F-FDG-avid tissue site underwent attempted in situ intraoperative detection concurrently using three gamma detection probe systems (K-alpha probe, and two commercially-available PET-probe systems), and then were subsequently surgical excised.

RESULTS

The mean relative probe counting efficiency ratio was 6.9 (± 4.4, range 2.2-15.4) for the K-alpha probe, as compared to 1.5 (± 0.3, range 1.0-2.1) and 1.0 (± 0, range 1.0-1.0), respectively, for two commercially-available PET-probe systems (P < 0.001). Successful in situ intraoperative detection of 18F-FDG-avid tissue sites was more frequently accomplished with each of the three gamma detection probes tested by using the three-sigma statistical threshold criteria method than by using the ratiometric threshold criteria method, specifically with the three-sigma statistical threshold criteria method being significantly better than the ratiometric threshold criteria method for determining probe positivity for the K-alpha probe (P = 0.05).

CONCLUSIONS

Our results suggest that the improved probe counting efficiency of the K-alpha probe design used in conjunction with the three-sigma statistical threshold criteria method can allow for improved detection of 18F-FDG-avid tissue sites when a low in situ T/B ratio is encountered.

124I-HuCC49deltaCH2 for TAG-72 antigen-directed positron emission tomography (PET) imaging of LS174T colon adenocarcinoma tumor implants in xenograft mice: preliminary results

Background

¹⁸F-fluorodeoxyglucose positron emission tomography (¹⁸F-FDG-PET) is widely used in diagnostic cancer imaging. However, the use of ¹⁸F-FDG in PET-based imaging is limited by its specificity and sensitivity. In contrast, anti-TAG (tumor associated glycoprotein)-72 monoclonal antibodies are highly specific for binding to a variety of adenocarcinomas, including colorectal cancer. The aim of this preliminary study was to evaluate a complimentary determining region (CDR)-grafted humanized C_H2-domain-deleted anti-TAG-72 monoclonal antibody (HuCC49deltaC_H2), radiolabeled with iodine-124 (¹²⁴I), as an antigen-directed and cancer-specific targeting agent for PET-based imaging.

Methods

HuCC49deltaC_H2 was radiolabeled with ¹²⁴I. Subcutaneous tumor implants of LS174T colon adenocarcinoma cells, which express TAG-72 antigen, were grown on athymic Nu/Nu nude mice as the xenograft model. Intravascular (i.v.) and intraperitoneal (i.p.) administration of ¹²⁴I-HuCC49deltaC_H2 was then evaluated in this xenograft mouse model at various time points from approximately 1 hour to 24 hours after injection using microPET imaging. This was compared to i.v. injection of ¹⁸F-FDG in the same xenograft mouse model using microPET imaging at 50 minutes after injection.

Results

At approximately 1 hour after i.v. injection, ¹²⁴I-HuCC49deltaC_H2 was distributed within the systemic circulation, while at approximately 1 hour after i.p. injection, ¹²⁴I-HuCC49deltaC_H2 was distributed within the peritoneal cavity. At time points from 18 hours to 24 hours after i.v. and i.p. injection, ¹²⁴I-HuCC49deltaC_H2 demonstrated a significantly increased level of specific localization to LS174T tumor implants (p = 0.001) when compared to the 1 hour images. In contrast, approximately 50 minutes after i.v. injection, ¹⁸F-FDG failed to demonstrate any increased level of specific localization to a LS174T tumor implant, but showed the propensity toward more nonspecific uptake within the heart, Harderian glands of the bony orbits of the eyes, brown fat of the posterior neck, kidneys, and bladder.

Conclusions

On microPET imaging, ¹²⁴I-HuCC49deltaC_H2 demonstrates an increased level of specific localization to tumor implants of LS174T colon adenocarcinoma cells in the xenograft mouse model on delayed imaging, while ¹⁸F-FDG failed to demonstrate this. The antigen-directed and cancer-specific ¹²⁴I-radiolabled anti-TAG-72 monoclonal antibody conjugate, ¹²⁴I-HuCC49deltaC_H2, holds future potential for use in human clinical trials for preoperative, intraoperative, and postoperative PET-based imaging strategies, including fused-modality PET-based imaging platforms.

Near-infrared fluorescence labeled anti-TAG-72 monoclonal antibodies for tumor imaging in colorectal cancer xenograft mice

Anti-TAG-72 monoclonal antibodies target the tumor-associated glycoprotein (TAG)-72 in various solid tumors. This study evaluated the use of anti-TAG-72 monoclonal antibodies, both murine CC49 and humanized CC49 (HuCC49deltaCH2), for near-infrared fluorescent (NIR) tumor imaging in colorectal cancer xenograft models. The murine CC49 and HuCC49deltaCH2 were conjugated with Cy7 monofunctional N-hydroxysuccinimide ester (Cy7-NHS). Both in vitro and in vivo anti-TAG-72 antibody binding studies were performed. The in vitro study utilized the human colon adenocarcinoma cell line LS174T that was incubated with Cy7, antibody-Cy7 conjugates, or excessive murine CC49 followed by the antibody-Cy7 conjugates and was imaged by fluorescence microscopy. The in vivo study utilized xenograft mice, bearing LS174T subcutaneous tumor implants, that received tail vein injections of Cy7, murine CC49-Cy7, HuCC49deltaCH2-Cy7, or nonspecific IgG-Cy7 and were imaged by the Xenogen IVIS 100 system from 15 min to 288 h. The biodistribution of the fluorescence labeled antibodies was determined by imaging the dissected tissues. The in vitro study revealed that the antibody-Cy7 conjugates bound to LS174T cells and were blocked by excessive murine CC49. The in vivo study demonstrated that murine CC49 achieved a tumor/blood ratio of 15 at 96 h postinjection. In comparison, HuCC49deltaCH2-Cy7 cleared much faster than murine CC49-Cy7 from the xenograft mice, and HuCC49deltaCH2-Cy7 achieved a tumor/blood ratio of 12 at 18 h postinjection. In contrast, Cy7 and Cy7 labeled nonspecific IgG resulted in no demonstrable tumor accumulation. When mice were injected with excessive unlabeled murine CC49 at 6 h before the injection of murine CC49-Cy7 or HuCC49deltaCH2-Cy7, both the intensity and retention time of the fluorescence from the tumor were reduced. In summary, the Cy7 labeled murine CC49 and HuCC49deltaCH2 demonstrate tumor-targeting capabilities in living colorectal cancer xenograft mice and provide an alternative modality for tumor imaging.

Gamma probes and their use in tumor detection in colorectal cancer

The purpose of this article is to summarize the role of gamma probes in intraoperative tumor detection in patients with colorectal cancer (CRC), as well as provide basic information about the physical and practical characteristics of the gamma probes, and the radiopharmaceuticals used in gamma probe tumor detection. In a significant portion of these studies, radiolabeled monoclonal antibodies (Mabs), particularly 125I labeled B72.3 Mab that binds to the TAG-72 antigen, have been used to target tumor. Studies have reported that intraoperative gamma probe radioimmunodetection helps surgeons to localize primary tumor, clearly delineate its resection margins and provide immediate intraoperative staging. Studies also have emphasized the value of intraoperative gamma probe radioimmunodetection in defining the extent of tumor recurrence and finding sub-clinical occult tumors which would assure the surgeons that they have completely removed the tumor burden. However, intraoperative gamma probe radioimmunodetection has not been widely adapted among surgeons because of some constraints associated with this technique. The main difficulty with this technique is the long period of waiting time between Mab injection and surgery. The technique is also laborious and costly. In recent years, Fluorine-18-2-fluoro-2-deoxy-D-glucose (18F-FDG) use in gamma probe tumor detection surgery has renewed interest among surgeons. Preliminary studies during surgery have demonstrated that use of FDG in gamma probe tumor detection during surgery is feasible and useful.

Radioimmunoguided surgery (RIGS), PET/CT image-guided surgery, and fluorescence image-guided surgery: Past, present, and future

(125)I-labeled anti-TAG-72 antibodies were applied in radioimmunoguided surgery (RIGS) to remove gross and occult tumors. It is challenging to handle (125)I-labeled materials. PET/CT image-guided surgery utilizes (18)FDG to monitor the biochemical activity of the tumor and to integrate pre- and postoperative imaging for complete tumor removal. PET/CT image-guided surgery only detects later stage disease. Fluorescence image-guided surgery using anti-TAG-72 antibodies may provide opportunities for intraoperative cancer detection of both gross and occult tumors.

Combined use of preoperative 18F FDG-PET imaging and intraoperative gamma probe detection for accurate assessment of tumor recurrence in patients with colorectal cancer

Background

The purpose of this study was to combine intraoperative gamma probe (GP) detection with preoperative fluorine 18-fluoro-2-deoxy-glucose positron emission tomography (¹⁸F FDG-PET) imaging in order to improve detection of tumor recurrence in colorectal cancer (CRC) patients.

Methods

Twenty-one patients (12 females, 9 males) with a mean age of 54 years (range 31–78) were enrolled. Patients were suspected to have recurrent CRC by elevated CEA (n = 11), suspicious CT findings (n = 1), and clinically suspicious findings (n = 9). Preoperative FDG-PET scan and intraoperative GP study were performed in all patients. Mean time interval between preoperative FDG-PET scan and surgery was 16 days (range 1–41 days) in 19 patients. For intraoperative GP studies, 19 patients were injected with a dose of 10–15 mCi ¹⁸F FDG at approximately 30 minutes before the planned surgery time. In two patients, the intraoperative GP study was performed immediately after preoperative FDG-PET scan.

Results

Preoperative FDG-PET and intraoperative GP detected 48 and 45 lesions, respectively. A total of 50 presumed site of recurrent disease from 20 patients were resected. Thirty-seven of 50 presumed sites of recurrent disease were histological-proven tumor positive and 13 of 50 presumed sites of recurrent disease were histological-proven tumor negative. When correlated with final histopathology, the number of true positive lesions and false positive lesions by preoperative FDG-PET and intraoperative GP were 31/9 and 35/8, respectively. Both preoperative FDG-PET and intraoperative GP were true positive in 29 lesions. Intraoperative GP detected additional small lesions in the omentum and pelvis which were not seen on preoperative FDG-PET scan. FDG-PET scan demonstrated additional liver metastases which were not detected by intraoperative GP. Preoperative FDG-PET detected distant metastasis in the lung in one patient. The estimated radiation dose received by a surgeon during a single 18F FDG GP surgery was below the occupational limit.

Conclusion

The combined use of preoperative FDG-PET and intraoperative GP is potentially helpful to the surgeon as a roadmap for accurately locating and determining the extent of tumor recurrence in patients with CRC. While intraoperative GP appears to be more sensitive in detecting the extent of abdominal and pelvic recurrence, preoperative FDG-PET appears to be more sensitive in detecting liver metastases. FDG-PET is also a valuable method in detecting distant metastases.

Pilot study using a humanized CC49 monoclonal antibody (HuCC49DeltaCH2) to localize recurrent colorectal carcinoma

BACKGROUND

CC49 is a monoclonal antibody directed against a pancarcinoma antigen (TAG-72) expressed by colorectal cancers. The use of murine CC49 in radioimmunoguided surgery (RIGS) was problematic because of the human anti-mouse antibodies (HAMA) generated. This study was designed to assess the clearance, safety, and effectiveness of localization of a complimentarity determining region (CDR)-grafted humanized domain-deleted antitumor CC49 antibody (HuCC49DeltaCH2).

METHODS

After thyroid blockade, 1 mg of HuCC49DeltaCH2 radiolabeled with 2 mCi of iodine-125 was administered. All patients subsequently underwent traditional exploration followed by a survey with the gamma-detecting probe. In five patients, exploration was performed 10 to 24 days after injection, when precordial counts were sufficiently low (<30 counts per 2 seconds [cp2s]). Traditionally suggestive and probe-positive tissue was biopsied or excised and examined for the presence of carcinoma, when considered appropriate by the operating surgeon. Serum was assessed for HAMA.

RESULTS

Seventeen sites were identified as suggestive of carcinoma on traditional exploration and 21 by RIGS. Of these, pathologic correlation was obtained in 15. The sensitivity of RIGS was 92%, and the positive predictive value was 100%. None of the patients expressed significant HAMA.

CONCLUSIONS

This initial study indicates that the HuCC49DeltaCH2 monoclonal antibody, when used with RIGS, is safe and sensitive in detecting recurrent intra-abdominal colon cancer.

Immunoscintigraphy and intra-operative radioimmunodetection in the treatment of colorectal carcinoma

PURPOSE

Intra-operative radioimmunodetection of malignant involved lymph nodes follows the pre-operative immunoscintigraphy in the treatment of patients with colorectal carcinoma. The aims of this clinical study were to determine the sensitivity of the method, to compare the results in study when using Oncoscint and CEA-Scan and to evaluate the importance of the method of surgery and postoperative adjuvant therapy.

PATIENTS AND METHODS

121 patients with colorectal tumours (106 primary and 15 recurrent) were operated on using radioimmunoguided surgery (RIGS). The study compared results of pre-operative immunoscintigraphy, intra-operative radioimmunodetection and postoperative histological examination. Histological investigation used classical H&E staining. In histologically negative and RIGS positive cases the immunohistochemical investigation was supplemented. Two radiopharmaceuticals were used Oncoscint CR 103 (MAb B72.3, Satumomab Pendetide), labelled with 111In in 56 patients and CEA-Scan (IMMU 4-Fab' fragments MAb against CEA, Arcitumomab), labelled with 99mTc in 65 patients.

RESULTS

The relationship between RIGS positive results and histological examination was statistically assessed after 38 operations and the most acceptable RIGS evaluating index was determined. All subsequent results were evaluated by this index. Immunoscintigraphy of tumour was positive in 112 cases (92.6%). Fifty-five RIGS positive cases of malignant infiltrated lymph nodes were confirmed by 43 histologically positive examinations (78%). In this group 9 cases were discovered only by immunohistochemistry. Sixty-six remaining RIGS negative results were confirmed in 62 (94%) cases by negative histology.

CONCLUSIONS

Both immunoscintigraphy and RIGS enable one to make a more accurate diagnosis. While treating the primary disease the use of RIGS may help in assessment of necessary extent of operation performance and in staging of the disease by revealing occult lymph nodes involved. Pre-operative immunoscintigraphy seems to be a useful diagnostic method for detection of tumour recurrence. When comparing two radiopharmaceuticals used, CEA-Scan seems to be more suitable for diagnostic studies, but using the Oncoscint for tumour recurrence detection had some specific benefit, too.

Benefits of radioimmunoguided surgery for pelvic recurrence

AIM

Surgery for recurrent rectal cancer is usually traumatic and of questionable curative value. The use of radioimmunoguided surgery (RIGS) in enhancing the surgeon's assessment of the extent of disease in these patients was investigated.

METHODS

Twenty-one patients diagnosed with recurrent pelvic cancer were operated using the RIGS(O)system. Preoperative assessment included CTs of chest, abdomen and pelvis as well as colonoscopy. Patients were injected with CC49, a monoclonal antibody (MoAb) labelled with 125I. Surgical exploration was followed by survey with the gamma-detecting probe.

RESULTS

Surgical exploration identified eight intra-colorectal recurrences, nine extra-colonic pelvic recurrences and five extra-pelvic lymph node metastases. RIGS exploration confirmed all intra-colonic recurrences except for one (patient with no MoAb localization), identified 13 pelvic recurrences and 10 lymph node metastases. There were seven patients with occult findings (33%), resulting in a modified surgical procedure. Surgery included five abdomino-perineal resections, six low anterior resections, seven excisions of presacral tumour, eight total abdominal hysterectomy and bilateral salpingo-oophorectomy, one pelvic exenteration and one post-exenteration. There were no operative deaths. Eight patients had minor complications, and one patient had a major complication with reoperation due to urinary leak. The mean follow-up was 18 months. Ten patients died of disease.

CONCLUSION

Although not curative, RIGS can help the surgeon in the decision-making process through better disease staging.

Identification of lymph node metastases in recurrent colorectal cancer

Lymph node metastases are an important prognostic prediction factor in patients with recurrent colorectal cancer, particularly those with liver metastasis. Fifty-six patients with recurrent colorectal cancer were operated by us using the RIGS (radioimmunoguided surgery) technology. Patients were injected with 1 mg monoclonal antibody (MoAb) CC49 labeled with 2 mCi 125I. In surgery, traditional exploration was followed by survey with a gamma-detecting probe. Sixty of 151 patients enrolled in the Neo2-14 Phase III study for recurrent colorectal cancer were diagnosed with liver metastases based on preoperative CT. In 17/56 patients (30%), RIGS identified at least one tumor site confirmed by pathology (H&E). This resulted in 16 major changes in surgical plan. RIGS performance varied between lymphatic and non-lymphatic tissue, with positive predictive value (PPV) of 100% and negative predictive value (NPV) of 94% for non-lymphoid tissue, compared to PPV of 46.5% and NPV of 100% for the lymphoid tissue. Thirty-five out of 60 patients were considered resectable after traditional evaluation. RIGS identified occult tumor in 10 of these patients (28.5%). 7/10 occult patients expired (70%), while only 7/25 of the non-occult patients expired (28%) (P = 0.046). In localizing patients, no RIGS activity in lymph nodes signifies no tumor, while H&E confirmation is needed for decisions based on RIGS activity in the lymph nodes. RIGS provides important staging information, identifying patients for whom surgery may be done with curative intent.

Intraoperative radioimmunodetection

The radioimmunoguided surgery (RIGS) system employs a monoclonal antibody (CC49), a radionuclide (Iodine-125), and a hand-held gamma-detecting probe (the Neoprobe model 1000). The prototype cancer studied has been colorectal cancer. The antibody identifies a type of mucin, the by-product of the adenocarcinoma cancer cell. The RIGS system localizes up to 90% of colorectal cancers and finds additional RIGS-positive tissues in >50% of the patients. More than 90% of the RIGS-positive visceral tumors are identified by routine hematoxylin-and-eosin (H&E) light microscopy, but the RIGS-positive lymph nodes are H&E occult tissues in >70% of the cases. Enhanced, more time-consuming methods have been developed to confirm hidden cancer cells in these lymph node tissues. Survival data confirm the importance of RIGS-positive tissues. RIGS-positive tissues remaining at the completion of the surgical procedure portend a much poorer outcome than if the patient is deemed RIGS-negative at the completion of the surgical procedure (i.e., all RIGS-positive tissue was removed at surgery).

First results for resetting the antitumor immune response by immune corrective surgery in colon cancer

BACKGROUND

A critical step for cancer recurrence is the failure of the cellular immune response. It is suspected that chronic humoral immune responses against some tumor-associated antigens (TAA) can contribute to that failure.

METHODS

In this study, we tested the ability of an immune corrective surgical procedure to prevent recurrences of colon cancer in stages I, II, and III. Radiolabeled anti-TAG antibodies injected intravenously become concentrated on TAG-72 immune complexes presented by follicular dendritic cells, which are responsible for the persistent humoral response against TAG-72 TAA. Using a hand-held gamma probe, we can intraoperatively detect and remove lymph nodes involved in TAG-72 presentation. By removing these lymph nodes, together with the tumor tissue, presentation and source of TAG-72 are drastically reduced.

RESULTS

The impact of this TAA suppression on the tumor recurrence process is analyzed in a sample of 24 patients. The immune corrective surgical procedure did not increase morbidity. Five years after surgery the following were disease free: 5 of 5 stage I, 6 of 6 stage II, and 10 of 13 stage III. The global survival of this group was 87.5%. Compared with the standard surgical treatment of colon cancer (58% survival for the same stages), this surgical immune corrective procedure introduces a statistically significant improvement of 29% (P <0.001).

CONCLUSIONS

The surgical removal of lymph nodes involved in the persistent humoral immune response against TAA has an important beneficial impact on colon cancer treatment.

Prognostic relevance of occult tumor cells in lymph nodes of colorectal carcinomas: an immunohistochemical study

PURPOSE

Whereas lymph node metastases in colorectal carcinoma are an important prognostic factor, the prognostic relevance of occult tumor cells in lymph nodes is not elucidated at present. Therefore, our study intended to assess the rate of patients with occult tumor cells in histopathologically negative lymph nodes. Furthermore, we tried to evaluate an eventual influence of these occult tumor cells on patients' prognoses.

METHODS

For examination, we used paraffin blocks of lymph nodes, tumor-negative by conventional histopathology, from 49 patients with colorectal carcinoma (Stage I-III) after a curative (R0) tumor resection in 1987. After preparation of tissue blocks using the serial sectioning technique, the specimens were stained with the alkaline phosphatase, antialkaline phosphatase method and two monoclonal antibodies (AE1/AE3 and Ber-EP4).

RESULTS

In 13 of 49 patients (26.5 percent), we disclosed tumor cells, mostly located in subcapsular sinuses as single cells or in groups. There was a good correlation between the detection rate and N category, tumor stage, and grading. Moreover, 33 percent of patients in Stage I/II with occult tumor cells (N0+) developed a local relapse and/or distant metastases in contrast to 12 percent of patients without tumor cells (N0-). With a median follow-up of 84 months, we found no difference in disease-free survival between the tumor cell negative and positive groups in Stage I/II patients.

CONCLUSION

The results show that occult tumor cells might increase the risk for development of a local tumor relapse and/or distant metastases but do not influence patients' prognoses at all.