Int. J. Cancer: 39, 50-59 (1987) 0 1987 Alan R. Liss, Inc. Publication of the International Union Against Cancer Publication de I'Union lnternationale Contre 18 Cancer QUANTITATIVE AND QUALITATIVE ASPECTS OF RADIOLOCALIZATION IN COLON CANCER PATIENTS OF INTRAVENOUSLY ADMINISTERED MAb B72.3 J.M. ESTEBAN', D. cOLCHER1'5, P. SUGARBAKER', J.A. cARRASQUILL03, G. BRYANT4, A. THOR', J.C. REYNOLDS2, S.M. LARSON3 and J. SCHLOM' 'Laboratory of Tumor Immunology and Biology, National Cancer Institute, National Institutes of Health; 2Surgery Branch, National Cancer Institute. National Institutes of Health: 3Devartment of Nuclear Medicine, Clinical Center, National Institutes of Health; 4Laboratory of Pathology, NationalCancer Institute, Monoclonal antibody (MAb) 672.3 has been previously shown, by in vitro assays, to have a high degree of specificity for carcinomas of the colon, ovary and breast versus normal adult tissues. 672.3 IgG was labelled with '"1 and injected i.v. into 20 patients with known or suspected colorectal cancer. All patients subsequently underwent surgical exploration, with tumor and selected normal tissues removed for staging pur- poses. The selective localization of '"I-MAb 672.3 IgG was demonstrated in biodistribution studies in which the % ID/g of each tumor was compared with that of the normal tissues, thus providing a relative RI for each lesion. Of the tumor lesions, 70% (99/142) had an RI of at least 3 (i.e., 3 times greater uptake per gram than normal tissues), and 3 I % of the tumor lesions had Rls of over 10. Only I2 of 210 (60/0) histolog- ically normal tissues had Rls of > 3; either these tissues were adjacent to or draining tumor masses or, as in the case of 2 patients, the high RI values were apparently due to deposition of immune complexes in the splenic tissues. Several parame- ters were studied to determine factors that might influence MAb localization. Whereas tumors of all histologic types lo- calized the MAb, 31 Yo of the well-differentiated mucinous carcinomas displayed tumor-to-normal ratios greater than 10, while less than 5% of the lesions of other tumor types dem- onstrated similar localization. The expression of the antigen (TAG-72) detected by MAb 672.3 in these tumors, as studied by immunohistochemical techniques using tissue sections, did not always correlate with the outcome of the MAb distribu- tion. No differences in MAb uptake were observed among the carcinoma lesions from numerous anatomic locations, dem- onstrating the ability of i.v. administered 672.3 to reach all the tumor sites. Furthermore, autoradiographic studies of tumors showed good penetration of the MAb into the medial areas of the tumors, regardless of their size. Many of the clinical studies using radiolabelled MAbs for detection of malignancies have emphasized tumor imaging as detected by external gamma scanning (Mach et al., 1981, 1983; Larson et al., 1983; Chatal et al., 1984; Moldofsky et al., 1983, 1984; Armitage et al., 1984; Williams et al., 1984; Epenetos et al., 1984). Although this is one of the end points of these studies, a systematic analysis of the different factors that might influence MAb-tumor binding should prove useful in understanding the mechanisms involved, and consequently could be used to improve the imaging and potential therapeutic results obtained with radiolabelled MAbs. MAb B72.3 is a murine IgGl generated by immunization of mice with a membrane-enriched fraction of a human carci- noma (Colcher et al., 1981). It recognizes a high-molecular- weight glycoprotein (termed tumor-associated glycoprotein [TAG]-72) (Johnson et al., 1986) which is expressed to vary- ing degrees in approximately 90% of colorectal carcinomas, 85% of breast carcinomas and 95% of ovarian carcinomas (Stramignoni et al., 1982; Nuti et al., 1982; Thor et al., 1986a,b). Its relative reactivity with normal adult tissues has been shown by immunohistochemical assays and radioimmu- noassays to be negligible (Thor et al., 1986b). B72.3 is an excellent adjunct when used with immunohistochemical tech- niques for the histopathologic diagnosis of solid tumors and cytologic specimens that traditionally present difficult diag- National Institutes of Health, Bethesda, MD 20892, USA. noses (Johnston et al., 1985; Martin et al., 1986, Nuti et al., 1986). The TAG-72 antigen detected by the B72.3 MAb has been shown to be secreted by the tumor and can be found in the circulation of some patients with colorectal cancer (Pater- son et al., 1986; Mug et al., 1986). It has previously been shown that B72.3 IgG can be conjugated with various radio- nuclides (1251, 13'1 and "'In) with no loss of immunoreactiv- ity, and can be used to selectively localize human colon carcinoma xenografts in athymic mice (Colcher et al., 1984; Keenan et al., 1984). This was shown by external imaging using a gamma camera and, more importantly, by direct anal- ysis of tissue samples. Selective reactivity of radiolabelled MAb with tumors vs. normal tissue was determined by the comparison of cpm/g in the tumor with the cpm/g in normal organs, with tumor:liver ratios, or "radiolocalization in- dices", reaching approximately 18: 1 7d after MAb administration. We report here the results obtained when 1311-B72.3 IgG was administered i.v. to 20 patients with suspected or con- firmed primary, recurrent or metastatic colqrectal carcinomas. The biodistribution of the B72.3 IgG among normal and tumor tissues was correlated with several parameters that could play a role in its distribution, including the histology of the tumors, expression of the TAG-72 antigen and anatomic location of tumor lesions. Patients The 20 patients with confirmed or suspected colorectal car- cinoma were included in previously ongoing NCI-Surgery Branch protocols. With the exception of 2 patients who pre- sented with primary disease, all patients had metastatic carci- noma whose primary lesion had been previously resected. Some had received additional forms of therapy. All the pa- tients were included in the ongoing protocol for staging and/ or debulking of recurrent and/or metastatic disease. The pa- tients' ages ranged from 26 to 70 years (11 males and 9 females), and many of those with recurrent and/or metastatic tumor had advanced disease (see Table I). Approximately 6 to 8 days post injection of the radiolabelled antibody, the patients underwent surgical exploration and tumor resection; selected normal tissues were removed for staging purposes. Monoclonal antibodies B72.3 IgG was purified from ascitic fluid by ammonium sulphate precipitation and ion-exchange chromatography MATERIAL AND METHODS 'To whom reprint requests should be sent. Abbreviations: ABC, avidin-biotin-complex; H and E, hematoxylin and eosin; HPLC, high-performance liquid chromatography; MAb, monoclo- nal antibody; ID/g, injected dose per gram; RI, radiolocalization index; TAG, tumor-associated glycoprotein; Lv., intravenously. Received: July 21, 1986 and in revised form August 27, 1986.
Transcript
Int. J . Cancer: 39, 50-59 (1987) 0 1987 Alan R. Liss, Inc.
Publication of the International Union Against Cancer Publication de I'Union lnternationale Contre 18 Cancer
QUANTITATIVE AND QUALITATIVE ASPECTS OF RADIOLOCALIZATION IN COLON CANCER PATIENTS OF INTRAVENOUSLY ADMINISTERED MAb B72.3 J.M. ESTEBAN', D. cOLCHER1'5, P. SUGARBAKER', J.A. cARRASQUILL03, G. BRYANT4, A. THOR', J.C. REYNOLDS2, S.M. LARSON3 and J. SCHLOM' 'Laboratory of Tumor Immunology and Biology, National Cancer Institute, National Institutes of Health; 2Surgery Branch, National Cancer Institute. National Institutes of Health: 3Devartment of Nuclear Medicine, Clinical Center, National Institutes of Health; 4Laboratory of Pathology, NationalCancer Institute,
Monoclonal antibody (MAb) 672.3 has been previously shown, by in vitro assays, t o have a high degree of specificity for carcinomas of the colon, ovary and breast versus normal adult tissues. 672.3 IgG was labelled with '"1 and injected i.v. into 20 patients with known or suspected colorectal cancer. All patients subsequently underwent surgical exploration, with tumor and selected normal tissues removed for staging pur- poses. The selective localization of '"I-MAb 672.3 IgG was demonstrated in biodistribution studies in which the % ID/g of each tumor was compared with that of the normal tissues, thus providing a relative RI for each lesion. Of the tumor lesions, 70% (99/142) had an RI of at least 3 (i.e., 3 times greater uptake per gram than normal tissues), and 3 I % of the tumor lesions had Rls of over 10. Only I2 of 210 (60/0) histolog- ically normal tissues had Rls of > 3; either these tissues were adjacent t o or draining tumor masses or, as in the case of 2 patients, the high RI values were apparently due to deposition of immune complexes in the splenic tissues. Several parame- ters were studied to determine factors that might influence MAb localization. Whereas tumors of all histologic types lo- calized the MAb, 31 Yo of the well-differentiated mucinous carcinomas displayed tumor-to-normal ratios greater than 10, while less than 5% of the lesions of other tumor types dem- onstrated similar localization. The expression of the antigen (TAG-72) detected by MAb 672.3 in these tumors, as studied by immunohistochemical techniques using tissue sections, did not always correlate with the outcome of the MAb distribu- tion. No differences in MAb uptake were observed among the carcinoma lesions from numerous anatomic locations, dem- onstrating the ability of i.v. administered 672.3 to reach all the tumor sites. Furthermore, autoradiographic studies of tumors showed good penetration of the MAb into the medial areas of the tumors, regardless of their size.
Many of the clinical studies using radiolabelled MAbs for detection of malignancies have emphasized tumor imaging as detected by external gamma scanning (Mach et al . , 1981, 1983; Larson et al . , 1983; Chatal et al . , 1984; Moldofsky et al . , 1983, 1984; Armitage et al . , 1984; Williams et al., 1984; Epenetos et al., 1984). Although this is one of the end points of these studies, a systematic analysis of the different factors that might influence MAb-tumor binding should prove useful in understanding the mechanisms involved, and consequently could be used to improve the imaging and potential therapeutic results obtained with radiolabelled MAbs.
MAb B72.3 is a murine IgGl generated by immunization of mice with a membrane-enriched fraction of a human carci- noma (Colcher et al., 1981). It recognizes a high-molecular- weight glycoprotein (termed tumor-associated glycoprotein [TAG]-72) (Johnson et al . , 1986) which is expressed to vary- ing degrees in approximately 90% of colorectal carcinomas, 85% of breast carcinomas and 95% of ovarian carcinomas (Stramignoni et al . , 1982; Nuti et al . , 1982; Thor et al., 1986a,b). Its relative reactivity with normal adult tissues has been shown by immunohistochemical assays and radioimmu- noassays to be negligible (Thor et al., 1986b). B72.3 is an excellent adjunct when used with immunohistochemical tech- niques for the histopathologic diagnosis of solid tumors and cytologic specimens that traditionally present difficult diag-
National Institutes of Health, Bethesda, MD 20892, USA.
noses (Johnston et al., 1985; Martin et al . , 1986, Nuti et al., 1986). The TAG-72 antigen detected by the B72.3 MAb has been shown to be secreted by the tumor and can be found in the circulation of some patients with colorectal cancer (Pater- son et al . , 1986; Mug et al . , 1986). It has previously been shown that B72.3 IgG can be conjugated with various radio- nuclides (1251, 13'1 and "'In) with no loss of immunoreactiv- ity, and can be used to selectively localize human colon carcinoma xenografts in athymic mice (Colcher et a l . , 1984; Keenan et al., 1984). This was shown by external imaging using a gamma camera and, more importantly, by direct anal- ysis of tissue samples. Selective reactivity of radiolabelled MAb with tumors vs. normal tissue was determined by the comparison of cpm/g in the tumor with the cpm/g in normal organs, with tumor:liver ratios, or "radiolocalization in- dices", reaching approximately 18: 1 7d after MAb administration.
We report here the results obtained when 1311-B72.3 IgG was administered i.v. to 20 patients with suspected or con- firmed primary, recurrent or metastatic colqrectal carcinomas. The biodistribution of the B72.3 IgG among normal and tumor tissues was correlated with several parameters that could play a role in its distribution, including the histology of the tumors, expression of the TAG-72 antigen and anatomic location of tumor lesions.
Patients The 20 patients with confirmed or suspected colorectal car-
cinoma were included in previously ongoing NCI-Surgery Branch protocols. With the exception of 2 patients who pre- sented with primary disease, all patients had metastatic carci- noma whose primary lesion had been previously resected. Some had received additional forms of therapy. All the pa- tients were included in the ongoing protocol for staging and/ or debulking of recurrent and/or metastatic disease. The pa- tients' ages ranged from 26 to 70 years (11 males and 9 females), and many of those with recurrent and/or metastatic tumor had advanced disease (see Table I). Approximately 6 to 8 days post injection of the radiolabelled antibody, the patients underwent surgical exploration and tumor resection; selected normal tissues were removed for staging purposes. Monoclonal antibodies
B72.3 IgG was purified from ascitic fluid by ammonium sulphate precipitation and ion-exchange chromatography
MATERIAL AND METHODS
'To whom reprint requests should be sent.
Abbreviations: ABC, avidin-biotin-complex; H and E, hematoxylin and eosin; HPLC, high-performance liquid chromatography; MAb, monoclo- nal antibody; ID/g, injected dose per gram; RI, radiolocalization index; TAG, tumor-associated glycoprotein; Lv., intravenously.
Received: July 21, 1986 and in revised form August 27, 1986.
RADIOLOCALIZATION IN COLON CANCER
TABLE I - SUMMARY OF PATIENTS INJECTED WITH "'1-872.3 I&
51
Injected dose
mg mCi activitvl
Scan Specific resultz Name Age Sex Carcinomatous lesions
E.L. M.S.
J.R. P.H.
H.F. C.C.
T.R.
J.D.
C.M.
M.P.
E.S.
H.K.
D.F.
B.B. E.P.
R.C.
H.L.
H.B.
J.P.
M.L.
70 37
44 38
56 38
52
27
65
57
59
49
66
57 56
45
18
54
63
64
F M
M M
F F
M
M
F
F
M
F
M
F M
M
F
M
M
F
3.70 3.84
20.00 19.72
19.24 19.24
17.80
0.39
0.45
0.40
0.34
0.22
0.27
1.35 1.32
1.18
1 .00
1.06
0.86
0.17
1.1 1.7
10.0 10.0
10.0 10.0
10.0
0.8
1.5
1.9
2.0
1.6
2 .o 10.0 10.0
9.4
8.3
10.0
10.0
2.0
0.30 0.44
0.50 0.51
0.52 0.52
0.56
2.08
3.33
4.75
5.88
7.18
7.41
7.41 7.58
7.99
8.30
9.43
11.63
11.76
Metastases to liver < 5 cm diameter. Metastases to upper and lower lobes of left lung and lower
lobe of right lung, < 1 cm diameter. Metastasis in pelvis, 1.5 cm diameter. Metastases to right lung, right and left lobes of liver. Im-
plants in pericardium, soft tissues of chest and abdominal cavities, 1 cm to 14x8 cm.
Carcinoma of right colon, 3 ~ 2 . 5 cm. Peritoneal implants on serosa of sigmoid colon, omentum,
vagina and various soft tissues. Sizes ranged from 1.5 cm to 3 ~ 2 . 5 cm.
Recurrent tumor in previous colonic anastomotic site; metas- tases and implants in serosa of ileum mesentery, left ureter and vas deferens. Size ranged from < 1 cm to 6x4 cm.
Metastases to right lobe of liver, omentum and retroperito- neum measuring up to 4X2.5 cm.
Recurrent tumor to previous colon anastomotic site, metas- tases to liver and abdominal wall. Sizes ranged from <0.5 cm to 2x2.4 cm.
Extensive peritoneal disease in serosa of small and large in- testine, spleen and liver capsules, omentum, abdominal wall and pelvis. Metastases to lymph nodes of the small and large intestines. Sizes ranged from <0.5 to 3 ~ 2 . 8 cm .
Four metastases to right lobe of liver ranging from < 1 cm to 9 . 5 ~ 6 cm.
Residual tumor in rectum (3.7 x 3.2) and metastases to re- gional lymph nodes.
Metastases to right lung and right lobe of liver measuring 1.5 and 2.5 cm, respectively.
Metastasis to liver measuring 3 ~ 2 . 5 cm. Carcinoma of colon, rectum and ileum; metastases to re-
gional lymph nodes; implants and free mucin in the perito- neal cavity. Location of primary tumor unknown. Sizes ranged from < 1 cm to 10x4 cm.
Metastases to a retroperitoneal lymph node of < 1 cm and liver (not biopsied).
Recurrent tumor in rectum and small bowel with extensive peritoneal implants in serosa of colon and ileum; appendix, omentum, ovaries, fallopian tubes and soft tissues. Sizes ranged from < 1 cm to 18x5 cm.
Carcinoma of colon and appendix. Extensive implants to ser- osa of bowel, omentum, diaphragm and abdominal soft tissues. Sizes ranged from microscopic foci to 5 cm.
Extensive peritoneal implants in the peritoneal serosa and perihepatic tissues. Sizes up to 2.2 cm in maximal dimension.
- Recurrent rectal tumor extending into the pelvis, reaching the pouch of Douglas. 4 x 3 cm.
'mCi/rng.-*Confirmed at surgery.
(Colcher et al., 1984). The purified IgG was filtered and all end lots were tested for freedom from Mycoplasma, from 12 adventitious viruses, and from pyrogenicity, as well as for sterility and general safety. Iodination of hh4b Bn.3
B72.3 was labelled with Na1311 by the Iodogen method (Colcher et al., 1984). Purified IgG (500gg) was labelled with approximately ~ m ~ i of N ~ ~ ~ ~ I in a glass vial coated with 250pg of Iodogen. After 10 min incubation, the free 1311 was removed by gel filtration through a Sephadex G-10 column. The labelled MAb was then diluted in saline containing 1% human serum albumin. Labelling efficiency was approxi- mately 60%, and the specific activity ranged from approxi- mately 6 to 12.5mCi per mg of antibody.
The immunoreactivity of all radiolabelled antibody prepa- rations was tested in a solid-phase radioimmunoassay with colon carcinoma extracts (LS-174T) known to react with MAb B72.3 (Colcher et al., 1984). More than 80% of the antibody remained immunoreactive after radiolabelling based on se- quential saturation solid-phase RIAs. Prior to administration to patients, unlabelled B72.3 IgG was added to the radiola- belled MAb to obtain the desired specific activity; the sample was then packaged into sterile ampules, shown to be apyro- genic by the Limulus amebocyte lysate assay, and then tested for sterility. Antibody administration and sample collection
The patients received from 0.17 to 20 mg of MAb, with activities ranging from 0.8 to 10 mCi, by slow infusion (over
52 ESTEBAN ET AL.
60 min) via a peripheral vein. No adverse reactions were observed in any of the patients during or after the infusion. Following administration, blood samples were collected at 5, 30, 60, 120, and 240 min and at daily intervals thereafter until surgery. Patient imaging
The patients were imaged within 2 hr of MAb administration and daily up to surgery. A large field of view gamma camera with a high energy collimator was utilized. A 20% window over the 364 KeV gamma ray of I3’I was used to obtain anterior and posterior whole-body images as well as multiple spot views (5 to 10 min each). In addition to analogue images, digital images were recorded with a Hewlett Packard Scinti- graphic computer.
Biodistribution studies All the tissues removed at surgery were labelled according
to organ and anatomic location and were immediately weighed on an analytical balance and counted in a gamma counter in order to establish the percentage of the injected activity per gram of tissue (% ID/@. The percentage ID/g of histologically confirmed normal liver or colon, or the average of several normal tissues, was arbitrarily assigned a value of 1. The percentage of each biopsy specimen was divided by the per- centage ID/g of the normal tissue(s) used as standard to deter- mine the radiolocalization indices (RI). Soft tissues such as adipose tissue or muscle, bone or skin were not included among the normal tissues used for the calculations so as not to artificially increase the RIs by their typically low uptake. Biopsies containing less than 20% tumor tissue were excluded from the tabulation.
Histologic studies After the tissues were weighed and counted, they were fixed
in 10% buffered formalin, embedded in paraffin and sec- tioned. The sections, 5p in thickness, were stained with H and E for light microscopy studies. The percentage of tumor per total cells, the histologic type of the tumor, and other micro- scopic characteristics were then tabulated. Immunohistochemical studies
Sections, mounted on gelatinized slides, were used in a modified avidin-biotin-peroxidase complex (ABC) immuno- histochemical assay to semi-quantitate the TAG-72 antigenic expression in all the tumors and normal tissues. The sections were incubated with MAb B72.3 (40 pglml) in PBS containing 0.1 % bovine serum albumin overnight at 4°C. Slides were then processed by the standard ABC method (Hsu et al . , 1981). The slides were scored on the basis of the percentage of cells stained and the location of the stain (i .e. , associated with the membrane, cytoplasm, and intra or extracellular mu- cin). The intensity of that stain was quantitated from very strong ( 3 + ) , strong (2+), weak positive (1+) to negative (-1. Autoradiographic studies
Sections on slides were deparaffinized in xylene and rehy- drated through steps of alcohol into phosphate-buffered saline. The slides were then immersed in photographic emulsion (11- ford K5) for 5 sec, then dried and dehydrated overnight in light-tight containers. After incubations of 3-6 weeks in boxes containing Drierite at -7O”C, the slides were developed with Kodak photographic developer (D-19) for 6 min. The process was then stopped by immersing the slides in 1 % acetic acid, and then fixed with a photographic fixer. After rehydration, the tissues were counterstained with Mayer’s hematoxylin for 15 min, after which the presence and distribution of the silver grains was determined.
This technique had some technical limitations due to the small number of counts per slide and spatial resolution ob- tained with the 13’1 radionuclide used in this protocol. Since the patients underwent exploratory laparotomy approximately 7 days post-inoculation of the labelled MAb, at which time only 0.04-10.37 X ID/g was present in the tumor, the corresponding counts present in a 5 p section of the tumor after being fixed, embedded and sectioned were further re- duced. By the time that the slides were actually covered with the emulsion, the radionuclide had gone through approxi- mately 2 half-lives. Therefore, only those tumors with the highest counts contained enough activity for autoradiographic visualization.
RESULTS
Biodistribution of l3’I-Bn. 3 IgG in patients Twenty patients who received I3’I-B72.3 IgG underwent
exploratory laparotomy approximately 7 days after injection of the radiolabelled MAb. The patients’ age and sex, together with specific activities of the l3lI-B72.3 IgG administer- ed, location and size of the tumors, are given in Table I. Two of the 20 patients (E.P., H.F.) had primary disease, while the rest had recurrent and/or metastatic tumors. Eight patients had extensive disease spreading into the peritoneal cavity; the rest had more localized disease centered in the organs to which colorectal carcinomas usually metastasize.
All the tissues removed at surgery were weighed and counted to establish the radiolocalization indices (RIs) which were defined as the ratio between the uptake of 131I-B72.3% ID per gram of tumor vs. that of normal tissues (see “Material and Methods”). Tumor and normal tissues were classified as such only after histologic examination. From the pool of normal tissues, skin and soft tissues were purposely excluded, since the uptake of radiolabel by these tissues is very low. The resection margins of intestines, liver and/or spleen were the tissues selected for the RI calculations when available. An RI of > 3 was arbitrarily considered to be “positive” for MAb localization. Seventy percent (99/142) of tumors had RIs > 3 ranging from 3.4 to 30, whereas only 6% (12/210) normal tissues had RIs > 3 (to be discussed below). Tables I1 through V show 4 representative patterns of MAb biodistributions encountered. An analysis of the % ID/g in the various tumor biopsies revealed a range of 0.04-10.37 x for those biop- sied on days 6-13 post-inoculation of the radiolabelled B72.3 IgG, with an average of 1 . 7 6 ~ 1 0 - . Patient MP (Table 11) had a large number of biopsies taken; all tumor biopsies removed from various sites demonstrated RIs of >3, with some as high as 30. RIs of < 3 were observed for all 5 normal tissue biopsies. Biopsies from patient M.L. (Table 111) follow the same pattern, showing a large number of normal tissues with no preferential MAb uptake (RIs <3), with all tumor biopsies showing enhanced MAb localization with RIs > 3. In patient H.L. (Table IV) preferential MAb localization was observed in some, but not all, tumor lesions. Patient J.P. (Table V) exemplifies one of two patients in whom high RIs were observed in histologically normal spleen (to be discussed below).
On studying the factors that could correlate between the specific localization of the label and the characteristics of the tissues, it became obvious that the first parameter involved was the percentage of tumor present in each tissue studied. This relationship is shown in patient M.P. (Fig. 1) in whom the progressive increase in the percentage of tumor in each biopsy generally corresponded to increasing RI ratios. How- ever, this was not a strict relationship; direct examination of tissues shows that antigenic heterogeneity between different biopsies are clearly a factor (see below). However, differences
RADIOLOCALIZATION IN COLON CANCER 53
TABLE I1 - MAb B72.3 LOCALIZATION IN BIOPSY SPECIMENS OF PATIENT M.P.
Biopsy Immunoperoxidase
reactivity Percent RI2
% IDIG' ( X 1n-3)
Tumor Mucin Cells Mucin \ .- I
Carcinoma site Liver 6.90 30.5 953 764 15' + + + 6 Serosa, small bowel (n:4) 4.95 21.8 100 80 20 +++ Omentum, lesser 3.68 16.2 90 72 10 ++ Site not specified 3.37 14.9 90 72 20 +++ Spleen 3.07 13.6 70 56 25 ++ Colon, distal site 3.11 13.1 80 64 25 ++ Lymph nodes, colon 2.65 11.7 80 68 15 +++ Serosa, small bowel 2.51 11.1 60 48 25 +++ Serosa, small bowel 1.56 6.9 90 72 25 +++ Lymph node, small bowel (n:2) 1.49 6.6 60 51 25 ++ Small bowel (necrotic) 1.20 5.3 10 8 15 + Retroperitoneum (n:2) 0.74 3.3 20 15 20 + Nonnal tissue Adipose tissue 0.07 0.3 0 Colon, distal margin 0.28 1.2 307 Muscle, abdominal waIl 0.53 2.4 0 Peritoneum, scar tissue abdomen 0.30 1.3 0 Small bowel (edematous) 0.28 1.3 0 'Percentage of injected dose of I3'I MAb B72.3 per gram of tissue; surgery was performed 8 days post-administration of the antibcdy.-'RI determined by dividing the %
ID/g of tumors by the averaged % ID/g of all normal tis~ues.-~Percenta e of tissue that is carcin~rna.-~Percentage of tissue that is mucin.-'Percentage of carcinoma cells reactive with B72.3.-'Scaled from very strong ( + + + ) t o negative (-).%orma1 colon mucosa adjacent to tumor.
TABLE I11 - MAb B72.3 LOCALIZATION IN BIOPSY SPECIMENS OF PATIENT M.L.
'Percentage of injected dose of "'I-MAb B72.3 per gram of tissue; surgery was performed 13 days post-administration of the antibody.-'RI determined by dividing the X ID/g of tumors by the averaged % IDlg of normal ~olon.-~Percenta e of tissue that is carcinoma-4Percentage of tissue that is mu~in.-~Percentage of carcinoma cells reactive with B72.3-'Scaled from very strong (+ + +) to negative (-).%orma1 endocervical epithelium reactive with B72.3.
in the percentage of tumor present could not explain all the differences in MAb uptake in the different biopsy specimens.
Anatomic sites of the tissues and Rls
The percentage of non-tumor tissues from different sites with RIs > 3 was also determined in order to study the differ- ences in uptake by the normal organs. The vast majority of the histologically normal tissues examined showed no preferential
Tumor biopsies were grouped according to their anatomic uptake of the radiolabelled B72.3 IgG, regardless Of their site site or organ of origin, including 9 of the anatomic sites from of origin (Table VII); a few tissue biopsies, however, did have which biopsies were excised more frequently. As shown in increased deposition of radiolabel. Table VI, the percent of tumor lesions with €Us > 3 among Histologically normal spleen in 4 of 5 biopsies from 3 the different sites was similar, ranging from 50 to loo%, with patients had RIs > 3. The 4 spleen biopsies with elevated RIs, the exception of the spleen in which only 1 of 4 tumors was however, were obtained from the 2 patients who had high positive; however, this latter observation could very well be levels of circulating immune complexes, as determined by due to the small number of biopsy samples available. HPLC analyses of plasma; these antibody-antigen complexes
54 ESTEBAN ET AL.
TABLE IV - MAb B72.3 LOCALIZATION IN BIOPSY SPECIMENS OF PATIENT H.L
Biopsy Immunoperoxidase
reactivity Percent RI2
Tumor Mucin Cells Mucin
Site of carcinoma Fallopian tube (1x2) 0.22 6.0 753 604 705 + + + 6 Fallopian tube, hydrosalpim 0.17 4.7 40 32 45 +++ Rectum, distal margin (n:2) 0.17 4.7 80 64 75 +++ Ovary 0.15 4.2 70 56 40 +++ Appendix 0.15 4.0 85 68 75 +++ Serosa, ileocecum 0.13 3.6 80 64 75 +++ Colon, lymph node 0.10 2.9 80 64 60 +++ Serosa rectum, implant 0.08 2.2 60 48 70 +++ Lymph node, spleen hilum 0.07 1.9 75 60 60 +++ Lymph node, colon 0.05 I .4 30 24 60 +++ Falciform ligament 0.05 1.3 30 24 60 +++ Ileum 0.05 1.3 20 16 80 +++ Omentum (n:3) 0.05 1.3 80 64 75 +++ Colon 0.04 1.3 25 20 60 +++ Normal tissue Adipose tissue, abdominal wall 0.02 0.5 0 Skin, parastoma 0.09 1.4 0 Ureter 0.05 1.1 0 'Percentage of injected dose of I3'I-MAb B72.3 per gram of tissue. surgery was performed 8 days post-administration of the antibody.-2RI determined by dividing the I
IDlg of the tumors by the averaged % IDlg of all normal t i s s u e ~ . - ~ ~ ~ ~ ' % e e Table 11.
TABLE V - MAb B72.3 LOCALIZATTON IN BIOPSY SPECIMENS OF PATIENT J.P.
'Percentage of injected dose of '3iI-MAb B72 3 per gram of tissue; surgery was performed 7 days post-administration of the antIbody.-'RI determrned by dividing the 96 ID/g of the tumors by the averaged % IDlg of normal tissues (excluding spleen samples) -3 %ee Table 11.
may thus have been deposited in the spleen. The spleen biopsy from a third patient did not have an elevated RI, nor did this patient have any circulating immune complexes. Slightly ele- vated RIs were also observed in 8 additional tissues. These findings however, can most probably be explained by one of the following reasons: (i) tissue was immediately adjacent to tumor; (ii) tissue was draining from a carcinoma (Le., gall bladder in patient with extensive hepatic metastasis); (iii) tis- sue was richly vascularized due to abundant granulation tissue; or (iv) occurrence of nonspecific uptake of the radiolabelled MAb by macrophages, as in 2 biopsies from lymph nodes with extensive lipogranulomas. An analysis of the 210 normal biop- sies indicated that no normal tissue was consistentIy positive for preferential MAb uptake (Table VII). Tumor histology and 81s
All tumor biopsies were studied by light macroscopy on H and E-stained slides to determine their type, grade, presence of necrosis, and other histologic characteristics. The tumors represented approximately equivalent numbers of well-differ-
entiated adenocarcinomas, well-differentiated adenocarcino- mas of the mucinous type, and moderately differentiated adenocarcinomas (Table VIII). One tumor was classified as a signet-ring-cell variant of adenocarcinoma. The percentage of lesions with RIs > 3 was very similar for each histologic group with values of 56, 75, 84 and 47%, respectively (Table VIII). However, there was a striking difference in the histologic characteristics of the tumor when lesions with RIs of more than 10 were examined. Thirty-one percent (25/81) tumors of the mucinous type had €Us of > 10 while only 0, 5, and 0% of the well-moderately-differentiated and signet-ring-cell adeno- carcinomas, respectively, had RIs of > 10 (Table VIII). Mu- cinous adenocarcinomas are therefore the histological type of tumor with the overall highest uptake of the I3lI-B72.3 IgG. TAG-72 expression and RIs
Immunohistochemical studies were performed, using the ABC immunoperoxidase method, on all the tissues excised at surgery to determine possible correlations between the TAG- 72 content of the tumor cells and/or the mucin and their RIs.
RADIOLOCALIZATION IN COLON CANCER 55
100
90
60 70
60
50
40
30
20-
0 0 - 0 .
- H
- 0 0 0
- 0
- 0 0
- - - 0
0 0
10-0 0
0 - 4 I I I I I I I 1
FIGURE 1 - Relationship between the radiolocalization index of I3'I- B72.3 IgG and percentage of tumor cells. All tissue biopsies, non-tu- mor (0) and tumor (0) removed from patient M.P. were studied mi- croscopically. The percentage of tumor present in each biopsy was plotted against its radiolocalization index.
In many patients the percentage of cells that scored positive varied within each tumor and from tumor to tumor, demon- strating once again the great antigenic heterogeneity present in most of the neoplasms. The pattern of staining also varied in different tumors of the same patient. MAb reactivity varied from the luminal border of the cell, intracytoplasmic or mucin- associated patterns were present, or frequently a combination of all or some of the patterns (Fig. 2). The intensity of staining of the mucin was scored from 0 to 3 + (3+ being the most intense). As shown in Table VIII, all the mucinous adenocar- cinomas had the strongest immunoperoxidase staining, with all muck encountered also having strong staining. In the other histologic groups, the intensity of immunoperoxidase staining ranged from negative to moderately positive, most of the tumors showing weakly positive patterns. Of interest was the strong positivity found in 80% of tumor cells in the only case of the signet-ring-cell type.
When the antigen content of the tumor cells was to be correlated with the radiolocalization, no clear trend was ob- served. Although, in a few patients, the tumors with the highest percentage of reactivity gave the best RIs (patients M.L., M.P.), there were also cases in which such correlation could not be established. An example of the latter can be seen in patient H.L. (Table IV) with strong MAb immunoperox- idase reactivity in up to 80% of the cells, most of which was
mucin-associated; this patient, however, had only 9/19 (47 %) tumor biopsies with RIs > 3, none of these being > 10. The inverse was also true: one patient, B.B., with minimal and weak antigenic expression (< 5 % of the cells) and with weak and scant rnucin reactivity had RIs > 3 in all 5 tumors studied. Thus, while there was a trend toward high antigenic content and increased MAb localization, there remain at this time unknown factors that account for differences in MAb localiza- tion among different tissues. Autoradiographic studies
Autoradiography was employed to observe the penetration and distribution of the radiolabelled MAb through the tumor mass (Fig. 3). The scatter of the radionuclide did not permit determination of the exact location within the cell (i. e . , mem- brane-bound, intracytoplasmic, etc.). The distribution of the silver grains was heterogeneous, foci of intense activity alter- nating with areas lacking grains, but radioactivity was equally distributed in the medial and the peripheral regions of the tumor, indicating a good penetration of the label throughout the tumor mass (Fig. 3a). MAb localization was predomi- nantly seen within the mucinous pools of tumors (Fig. 3b) although concentration in the cells was also present.
DISCUSSION
Radiolabelled MAbs have recently been used in clinical trials for the detection of many types of malignancies including colorectal, breast and ovarian cancer, as well as melanomas (Mach et al., 1981, 1983; Larson et al., 1983; Chatal et al., 1984; Moldofsky et al., 1983, 1984; Armitage et al., 1984; Williams et al., 1984; Epenetos et al., 1984). Many of the clinical trials undertaken thus far have not attempted to corre- late scan results with parameters such as quantitative aspects of MAb localization and the histopathology of the tissues that were localized. If radiolabelled MAbs are to be considered for therapeutic trials, the quantitative aspects of MAb localization for carcinoma versus normal tissues will be of even greater importance.
We had the opportunity to evaluate a large number of tissues (tumor and normal) resected from 20 patients with colorectal carcinomas that had previously been injected i.v. with 13'1- B72.3 IgG. Seventy percent (99/142) of the tumors had RIs > 3 , and of these, 17% (24042) had RIs of > 10. These ratios represent the highest values reported in this type of clinical trial (Mach et al., 1981, 1983; Larson et al., 1983; Chatal et al., 1984; Moldofsky et al., 1984). Furthermore, soft tissues such as adipose tissue or muscle, bone or skin were not included in the normal tissues used for calculation of the ratios; therefore, the RIs are not artificially increased by the
TABLE VI - LOCALIZATION OF '3'1 872.3 Igc TO CARCINOMA LESIONS
RI' Positive RI/ 3-10 > 10 total* (%) Site of carcinoma
< 3
Liver 4 24 1 25/29 (86) Colon 5 10 2 12/17 (71) Small intestine 3 1 3 417 (57) Lymph node 4 7 2 9/13 (69) Lung 3 6 2 811 1 (73) Peritoneum 18 18 8 26/44 (59) Pelvis 1 1 0 112 (50) Soft tissue 2 3 3 618 (75) Spleen 3 0 1 114 (25) Othe? 0 5 2 717 (100) Total 43 (30%) 75 (53%) 24 (17%) 991142 (70) 'The $6 IDlg of every biopsy specimen was divided by the % IDlg of the normal tissue(s) used as standard to determine the
radiolocalization indices (RI).-'RIs > 3 are considered positive for MAb localization.-'Ovanes (n:3), Fallopian tubes (n:2), pouch of Douglas (n:l), and vena cava @ I ) .
56 ESTEBAN E r AL.
TABLE VII - UPTAKE OF "'1.B72.3 IgG BY NORMAL ORGANS
121210 (6) 'The % IDlg of every biopsy specimen was divided by the % IDlg of the normal tissue(s) used as standard to determine the
radiolocalization indices (RI).-'RIs > 3 are considered positive for MAh Iocalizati~n.-~Tissues adjacent to t~mor.-~Abundant lipogmnulomas positive for TAG-72 by immunoperoxidase ~taining.-~Granulation tissue.-6Patients with circulating immune compIexe~.-~Gall-bladder (n:5), vagina (n:2), cervix (n:2), uterus (n:2), fallopian tube (n:2), hone (n:2), skin (n:2), 1 each of: epicardium, ureter, vas deferens, seminiferous duct, and portal vein .-8Gall-bladder from patient with liver metastasis.
TABLE VlII - PATIENTS GROUPED BY TUMOR HISTOLOGY AND IMMUNOPEROXIDASE REACTIVITIES
lmmunoperoxidase Radiolocalization reactivity index (RI)' RI > 3/Total RI > IO/Total
'The % ID/g of every biopsy specimen was divided by the % IDlg of the normal tissue(s) used as standard to determine the radiolocalization indices (RI).-'Average percentage of carcinoma cells reactive with B72.3.-'Average mucin reactivity with B72.3: very strong (++ +) to negative (-).-4Average percentage of tumor that is mucin.
typical low uptake encountered in those tissues. If they had been included, many tumor to tissue ratios would have been >50:1.
When studying the factors that could correlate the specific localization of the label with the characteristics of the tissues, it became obvious that the major parameter involved was the percentage of tumor present in the biopsy specimen. This relationship is exemplified in patient M.P. (Fig. 1) in whose biopsy the progressive increase in percentage of tumor was proportional to increasing RIs ratios. A great deal of antigenic heterogeneity was seen within a tumor mass and between the different metastases in the same patient, as demonstrated by immunohistochemical techniques. However, there was no strict
correlation between the antigenic content of the tumor and the radiolocalization of the MAb. The explanation for this finding may be related to 2 factors: (i) specific localization is a multi- factorial phenomenon, of which absolute antigenic content might not be the most important factor , and (ii) the immuno- peroxidase technique may not be the most valid method of determining the true antigenic content of a tumor. In this regard one must remember that the tissue has been fixed and sectioned and intracytoplasmic antigen exposed.
The tumors were analyzed and classified histologically into 3 groups containing: well-differentiated adenocarcinomas, moderately-differentiated adenocarcinomas , and mucinous ad- enocarcinomas. Of the mucinous adenocarcinomas, 3 1 % ( 2 9
58 ESTEBAN ET AL.
FIGURE 3 - Autoradiographs of carcinomas from patients injected with I3'I-B72.3 showing the distribution of the radiolabelled MAb. (a) Low- power view of a section taken from the medial portion of the carcinoma showing activity evenly distributed throughout the tumor mass, indi- cating the ability of the radionuclide-MAb complex to penetrate deep into the tumor. Note the silver grains associated with both mucin and cells (arrows). (b) A different section from the same patient (H.L.), showing an almost exclusive localization of the activity with the cells. Only a few tumor cells in the field contain silver grains, probably due to the antigenic heterogeneity of the tumor, as previously seen by immunohis- tochemical studies. (Scale bar = 1 0 0 p ~ , a; 2 5 p ~ , b)
~ ~~ ~
FIGURE 2 - TAG-72 content of carcinomas as determined by ABC immunoperoxidase method using MAb B72.3. Four examples of the different reactivities encountered are depicted. (a) Moderately differ- entiated adenocarcinoma from patient M.L. showing strong intracy- toplasmic staining with B72.3, but minimal (<5%) and weak mucin reactivities. Note the lack of stain in approximately 40% of the cells of the gland located in the center, demonstrating the antigenic heter- ogeneity found in most carcinomas. (b) Well-differentiated mucinous adenocarcinoma from patient E.P. is a rare example of poor radiolo- calization encountered in tumors of this type. Abundant pools of mu- cin with stain intensity ranging from weak to very strong were seen in all the tumor biopsies examined. In spite of the abundant antigen con- tent of the mucin, only 2 of 15 carcinoma lesions had RIs > 3. (c) Well- differentiated adenocarcinoma (H.F.) with stain located exclusively on the apical areas of the cells. This pattern was frequently associated with much andlor cytoplasmic reactivity. (d) Signet-ring-cell adenocarci- noma from patient H.L. This tumor had the strongest reactivity, in both percentage of cells and intensity of stain. As shown, the stain is pri- marily located in the intracytoplasmic mucin. We hypothesize that the MAb could not reach the antigen in vivo, which explains the less than optimal radiolocalization achieved in this patient. (Scale bars = 2 5 p ~ , a d ; 1 0 0 ~ ~ , b,r)
81) had RIs > 10 vs. 0 and 5% of the well- and moderately- differentiated groups respectively; these results indicate the capacity of MAb B72.3 to localize preferentially in this type of tumor rather than in the others. This is most likely due to the fact that the TAG-72 antigen is abundantly excreted into the numerous much pools. This has been shown by a high peritoneal localization in colorectal cancer patients with pseu- domyxoma peritonei (large amount of much in the peritoneal cavity) (Colcher et al., 1986) and by autoradiographic studies (see Fig. 2). Once MAb B72.3 localizes within the extracel- lular mucinous pools, it is likely to behave as a sequestered compartment. Furthermore, most of these mucinous tumors had scant cellularity, in many cases representing less than 10% of the total tumor volume, indicating that the mucin could be the determining factor for good MAb localization in some tumors.
The anatomic site of the tumor lesions studied was represen- tative of the most common sites to which colorectal carcino- mas metastasize. Localization of the MAb in tumors from various organs and anatomic sites was seen. Ofthe 210 normal
RADIOLOCALIZATION IN COLON CANCER 59
tissues studied, only the spleens of 2 patients had high RIs; these 2 patients had circulating immune complexes, and there- fore it is assumed that these complexes were deposited in the spleens. Eight additional tissues immediately adjacent to tumor also had slightly elevated RIs. No normal tissues however, were consistently positive, thus demonstrating the selective reactivity of the B72.3 IgG for tumor tissue in vivo. Moreover, no toxicity was observed in any of the patients receiving various doses of MAb.
In summary, radiolabelled MAb B72.3 was able to reach the majority of tumors located in the most frequent metastatic sites. Our preliminary results show the ability of MAb B72.3 to penetrate the medial areas of tumors, rather than remain in the perivascular peripheral areas; however, further autora- diography studies need to be conducted to better define this
point. These and other studies using suitable radionuclides at appropriate doses will determine the potential application of MAb B72.3 in diagnosis and perhaps therapy of various car- cinomas. We believe that the systematic approach used in this study is necessary to understand and maximize the potential of MAbs as radiopharmaceutical agents.
ACKNOWLEDGEMENTS
We thank Ms. D. Simpson, Ms. P. Perentesis, Ms. M. Lora and Mr. P. Maloney for their expert technical assistance. We acknowledge the help of the radiopharmacists: Mr. M. Rot- man, Mr. R. Fejka, Ms. J. Englert, and Mr. R. Farkas. We also thank Dr. R. Mariani-Costantini for his help in reading some of the histological sections.
REFERENCES
ARMITAGE, N.C., PERKINS, A.C., PIMM, M.V., FARRANDS, P.A., BALD- WIN, R.W., and HARDCASTLE, J.D., The localization of an anti-tumor monoclonal antibody (791T/36) in gastrointestinal tumours. J. Surg., 71,
CHATAL, J-F., SACCAVINI, J-C., FUMOLEAU, P., DOUILLARD, J-Y., CUR- TET, C., KREMER, M., LE MEVEL, B., and KOPROWSKI, H., Immunoscin- tigraphy of colon carcinoma. J. nucl. Med.. 25, 307-314 (1984). COLCHER, D., ESTEBAN, J.M., CARRASQUILLO, J.A., SUGARBAKER, P., REYNOLDS, J.C., BRYANT, G., LARSON, S.M., and SCHLOM, J., Compari- son of route of administration of radiolabelled MAb in patients with colorectal cancer. J. nucl. Med., 27, 902 (Abstract), (1986). COLCHER, D., HAND, P., NUTRI, M., and SCHLOM, J., A spectrum of monoclonal antibodies reactive with human mammary tumor cells. Proc. natl. Acad. Sci. (Wash.), 78, 3199-3203 (1981). COLCHER, D., KEENAN, A.M., LARSON, S.M., and SCHLOM, J., Prolonged binding of a radiolabelled monoclonal antibody (B72.3) used for the in situ radioimmunodetection of human colon carcinoma xenografts. Cancer Res., 44, 57445751 (1984). EPENETOS, A.A., SHEPHERD, J., BRITTON, K.E., HAWKINS, L., NIMMON, C.C., TAYLOR-PAPADIMITRIOU, J., DURBIN, H., MALPAS, J.S., MATHER, S., GRANOWSKA, M., DUKE, D., and BODMER, W.F., Radioimmunodi- agnosis of ovarian cancer using IWlabelled, tumor-associated monoclo- nal antibodies. Cancer Detect. Prevent., 7,45-49 (1984). Hsu, S.M., RAINE, L., and FAUGER, H., Use of avidin-biotin-peroxidase complex (ABC) in immunoperoxidase technique: a comparison between ABC and unlabeled antibody (PAP) procedures. J. Hisrochem. Cytochem.,
JOHNSON, V.G., SCHLOM, J., PATERSON, A.J., BENNETT, J., MAGNANI, J.L., and COLCHER, D., Analysis of a human tumor-associated glycopro- tein (TAG-72) identified by monoclonal antibody B72.3. Cancer Res., 46,
JOHNSTON, W. W., SZPAK, C.A., ~ T T I C H , S.C., THOR, A,, and SCHLOM, J., Use of a monoclonal antibody (B72.3) as an immunocytochemical adjunct to diagnosis of adenocarcinoma in human effusions. Cancer Rex,
KEENAN, A.M., COLCHER, D., LARSON, S.M., and SCHLOM, J., Radioim- munoscintigraphy of human colon cancer xenografts in mice with ra- dioiodinated monoclonal antibody B72.3. J . nucl. Med., 25, 1197-1203 (1984). KLUG, T.L., SATTLER, M.A., COLCHER, D., and SCHLOM, J., Monoclonal antibody immunoradiometric assay for an antigenic determinant (CA72) on a novel pancarcinoma antigen (TAG-72). Znt. J. Cancer, 38, 661-669 (1986). LARSON, S.M., BROWN, J.P., WRIGHT, P.W., CARRASQUILLO, J.A., HELLSTROM, I., and HELLSTROM, K.E., Imaging of melanoma with 1-131 labelled monoclonal antibodies. J. nucl. Med., 24, 123-129 (1983). MACH, J-P., BUCHEGGER, F., FORNI, M., RITSCHARD, J., BERCHE, C.,
407-412 (1984).
29,577-580 (1981).
850-857 (1986).
45, 1894-1900 (1985).
LUMBROSO, J-D., SCHREYER, M., GIRARDET, C., ACCOLLA, R.S., and CARREL, S., Use of radiolabelled monoclonal anti-CEA antibodies for the detection of human carcinomas by external photoscanning and tomoscin- tigraphy. Immunol. Today, 2,239-249 (1981). MACH, J.P., CHATAL, J-F., LUMBROSO, J-D., BUCHEGGER, F., FORNI, M., RITSCHARD, J., BERCHE, C., DOUILLARD, J-Y., CARREL, S., HERYLN, M., STEPLEWSKI, Z., and KOPROWSKI, H. Tumor localization in patients by radiolabelled monoclonal antibodies against colon carcinoma. Cancer Res. ,
MARTIN, S.E., MOSHIRI, S., THOR, A., VILASI, V., CHU, E.W., and SCHLOM, J., Identification of adenocarcinoma in cytospin preparations of effusions using monoclonal antibody B72.3. Amer. J. d in . Path., 86, 10- 18 (1986). MOLDOFSKY, P.J., POWE, J., MULHERN, C.B., JR., HAMMOND, N., SEARS, H.F., GATENBY, R.A., STEPLEWSKI, Z. , and KOPROWSKI, H., Metastatic colon carcinoma detected with radiolabelled F(ab’)z monoclonal antibody fragments. Radiology, 149, 549-555 (1983). MOLDOFSKY, P.J., SEARS, H.F., MULHERN, C.B., JR., HAMMOND, N.D., POWE, J., GATENBY, R.A., STEPLEWSKI, Z., and KOPROWSKI, H., Detec- tion of metastatic tumor in normal-sized retroperitoneal lymph nodes by monoclonal antibody imaging. N . Engt. J . Med., 311, 106-107 (1984). NUTI, M., MOTTOLESE, M., VIORA, M., DONNORSO, R.P., SCHLOM, J., and NATALI, P.G., Use of monoclonal antibodies to human breast tumor- associated antigens in fine-needle aspirate cytology. Int. J . Cancer, 37,
NUTI, M., TERAMOTO, Y.A., MARIANI-COSTANTINI, R., HORAN HAND, P., COLCHER, D., and SCHLOM, J., A monoclonal antibody (B72.3) defines patterns of distribution of a novel tumor-associated antigen in human mammary carcinoma cell populations. Int. J . Cancer, 29,539-545 (1982). PATERSON, A.J., SCHLOM, J., SEARS, H.F., BENNETT, J., and COLCHER, D., A radioimmunoassay for the detection of human tumor-associated glycoprotein (TAG-72) using monoclonal antibody B72.3. Znt. J. Cancer,
STRAMIGNONI, D., BOWEN, R., ATKINSON, B.F., and SCHLOM, J., Differ- ential reactivity of monoclonal antibodies with human colon adenocarci- nomas and adenomas. Znt. J . Cancer, 31,543-552 (1982). THOR, A., GORSTEIN, F., OHUCHI, N., SZPAK, C.A., JOHNSTON, W. W., and SCHLOM, J., Tumor-associated glycoprotein (TAG-72) in ovarian car- cinomas defined by monoclonal antibody B72.3. J. nut. Cancer Znst., 76,
THOR, A., OHUCHI, N., SZPAK, C.A., JOHNSTON, W.W., and SCHLOM, J., The distribution of oncofetal antigen TAG-72 recognized by monoclonal antibody B72.3. Cancer Res., 46, 3118-3124 (1986b). WILLIAMS, M.R., PERKINS, A.C., CAMPBELL, F.C., PIMM, M .V., HARDY, J.G., WASTIE, M.L., BLAMEY, R.W., and BALDWIN, R. W., The use of monoclonal antibody 791T136 in the immunoscintigraphy of primary and metastatic carcinoma of the breast. Clinical Oncology, 10,375-381 (1984).
