Highly specific prediction of antineoplastic drug resistance with an in vitro assay using suprapharmacologic drug exposures.

Bayes' theorem has been used to describe the relationship between the accuracy of a predictive test (posttest probability) and the overall incidence of what is being tested (pretest probability). Bayes' theorem indicates that laboratory assays will be accurate in the prediction of clinical drug resistance in tumors with high overall response rates (e.g., previously untreated breast cancer) only when the assays are extremely (greater than 98%) specific for drug resistance. We developed a highly specific drug-resistance assay in which human tumor colonies were cultured in soft agar and drugs were tested at high concentrations for long exposure times. Coefficients for concentration x time exceeded those reported in contemporaneous studies by about 100-fold. We reviewed 450 correlations between assay results and clinical response over an 8-year period. Results were analyzed by subsets, including different tumor histologies, single agents, and drug combinations. Extreme drug resistance (an assay result greater than or equal to SD below the median) was identified with greater than 99% specificity. Only one of 127 patients with tumors showing extreme drug resistance responded to chemotherapy. This negligible posttest probability of response was independent of pretest (expected) probability of response. Once this population of patients with tumors showing extreme drug resistance had been identified, posttest response probabilities for the remaining cohorts of patients varied according to both assay results and pretest response probabilities, precisely according to predictions based on Bayes' theorem. This finding allowed the construction of a nomogram for determining assay-predicted probability of response.

Mediation of cytotoxic immune responses against human tumor-associated antigens by xenogeneic immune RNA.

Xenogeneic immune RNA (I-RNA), extracted from the lymphoid organs of sheep or guinea pigs immunized with human tumor cells, mediated in vitro cytotoxic immune responses that were directed specifically against tumor-associated antigens of human tumor target cells. Normal human peripheral blood lymphocytes from healthy donors became markedly more cytotoxic for human tumor target cells after being incubated with I-RNA extracted from the lymphoid organs of animals that had been immunized with that particular tumor. Gastric carcinoma, malignant melanoma, and carcinoma of the breast were studied. Lymphocytes incubated with RNA from animals immunized with only complete Freund's adjuvant evidenced no increased cytotoxic activity. RNA extracted from the lymphoid organs of animals immunized with normal skin fibroblasts that were autologous to the immunizing tumor, when incubated with normal allogeneic lymphocytes, also mediated cytotoxic immune reactions against tumor target cells. These immune responses probably were directed principally against normal transplantation antigens. However, when lymphocytes that were autologous to the immunizing tumor and/or the tumor target cells were incubated with RNA from animals immunized with autologous normal fibroblasts, cytotoxicity did not increase. Only I-RNA extracted from donor animals specifically immunized with tumor cells mediated cytotoxic antitumor immune responses when incubated with autologous lymphocytes.

Specificity of antitumor immune reactions mediated by xenogeneic immune RNA.

Evidence that xenogeneic immune RNA (I-RNA) mediated specific cytotoxic immune responses against human tumor-associated antigens was obtained from in vitro studies in two autologous melanoma systems. In these systems, malignant melanoma target cells, matching normal fibroblast target cells, lymphocyte effector cells, and melanoma and normal skin tissue used to immunize RNA donor animals were derived from the same autochthonous hosts. When incubated with autologous lymphocytes, I-RNA extracted from the lymphoid organs of donor animals immunized with melanoma tissue mediated immune reactions against autologous melanoma target cells in vitro. I-RNA from animals immunized with normal skin tissue from autochthonous hosts did not increase the cytotoxicity of autologous lymphocytes for autologous melanoma cells. Using autologous fibroblasts as target cells, we detected no increase in cytotoxicity when autologous lymphocytes were incubated with RNA from animals immunized either with melanoma tissue or normal skin tissue from the autochthonous host. By contrast, when allogeneic lymphocytes were used as effector cells, RNA extracted from animals immunized either with melanoma tissue or normal skin mediated cytotoxic immune reactions against melanoma target cells and normal fibroblast target cells derived from the same patient.

Mediation of cytotoxic immune responses against human tumor-associated antigens by allogeneic immune RNA.

Allogeneic immune RNA (I-RNA), extracted from the peripheral blood lymphocytes of patients putatively cured of cancer, mediated cytotoxic immune reactions that apparently were directed specifically against human tumor-associated antigens. I-RNA was extracted from the peripheral blood lymphocytes of patients with various types of cancer. Patients selected had not been previously sensitized to HL-A or other normal transplantation antigens or to blood group antigens. Normal human peripheral blood lymphocytes were incubated with these allogeneic I-RNA preparations and tested for cytotoxicity against human target cells in vitro. Allogeneic I-RNA mediated cytotoxic immune reactions only against tumor target cells of the same histologic type as the I-RNA donor. I-RNA's extracted from peripheral blood lymphocytes of melanoma patients mediated cytotoxic immune reactions only against melanoma cells. Similarly, only I-RNA's extracted from the lymphocytes of patients with colon cancer mediated cytotoxic immune reactions against colon carcinoma cells, and only I-RNA's from the lymphocytes of breast cancer patients mediated immune reactions against breast cancer target cells. Allogeneic I-RNA extracted from peripheral blood lymphocytes of cancer patients possibly mediated specific cytotoxic immune reactions that were directed against common tumor-associated antigens shared by human tumors of similar histologic type.

Purification and immunologic evaluation of human melnoma-associated antigens.

Melanoma-associated antigens (MAA) were isolated and their functional immunologic properties were evaluated. Spent fetal calf serum-free culture media and 3-m KCI extracts of cultured human melanoma cells grown in this medium were used as antigen sources. Ultracentrifugal flotation on KBr was used to separate MAA and HLA antigens present in the extracts or spent culture media; thus interference by histocompatibility antigens was prevented in subsequent tests of tumor antigenic activity. MAA purified in this manner retained their immunologic functions as evidenced by their ability to produce delayed cutaneous hypersensitivity reactions in patients with melanoma, specifically combine with antimelanoma xenoantibody, and elicit production of functionally specific xenoantibody. Possible structural differences between HLA antigens and MAA were considered in evaluation of the data.

In vitro pharmacodynamics of 1-beta-D-arabinofuranosylcytosine: synergy of antitumor activity with cis-diamminedichloroplatinum(II).

1-beta-D-Arabinofuranosylcytosine (ara-C) was tested at a concentration of 10 micrograms/ml in the human tumor colony-forming assay against 55 human tumors of various histological types. Using the criterion for sensitivity of at least 70% inhibition of colony formation, 12 tumors (22%) were sensitive to ara-C. ara-C was most active against lung tumors (3 of 8 tumors were sensitive), and melanomas (6 of 8 sensitive). However, ara-C was not active against breast cancer (0 of 7) or colon cancer (0 of 3), and only 1 of 13 ovarian cancers was sensitive to ara-C. The activity of ara-C against melanoma and other solid tumors was confirmed using a thymidine incorporation assay. The time (t) and concentration (C) dependency of the cytotoxicity of ara-C and other chemotherapeutic agents was determined. Most agents such as Adriamycin, cis-diamminedichloroplatinum(II) (cis-platinum), and bleomycin were found to follow the C x t rule. That is, as the drug concentration was doubled, an equivalent amount of cell kill was achieved in half the time. However, the activity of ara-C was more concentration dependent than time dependent. ara-C was more effective when cells were exposed to high concentrations for short time periods. Synergy of activity between ara-C and cis-platinum was demonstrated in the breast 231 and melanoma M19 cell lines. No synergy of interaction between these two drugs was observed in the colon HT29 and lung P3 cell lines. When fresh biopsy specimens were tested with the combination, there was evidence of a synergistic interaction in 9 of 36 (25%). Maximum cytotoxicity was obtained when cells were exposed to ara-C 2 h before the addition of cis-platinum. The addition of cis-platinum before ara-C decreased the synergism.

Production of antisera with specificity for malignant melanoma and human fetal skin.

Complement-dependent cytotoxic antibodies to common cell surface antigens of cultured melanoma cells were produced in guinea pigs. At appropriate dilution, melanoma antisera were cytotoxic only to melanoma target cells. Following absorption with pooled lymphoid cells, additional absorption with melanoma cells but not absorption with fibroblasts or carcinoma cells was found to remove all cytotoxic activity from melanoma antisera. Absorption with human fetal skin cells but not with autologous fetal visceral cells was found to remove cytotoxicity from melanoma antisera. Tissue type-specific antigens may be shared by human malignant melanomas and fetal skin of black racial origin (at 16 to 18 weeks of gestation). The methods may be useful in the production of xenogeneic antisera with "operational monospecificity" for common melanoma-specific antigens. Sera from 47 patients with malignant melanoma failed to evidence specific cytotoxicity for melanoma target cells.

Structure-activity relationships defining the cytotoxicity of catechol analogues against human malignant melanoma.

The cytotoxic activities of three new synthetic catechol analogues, beta-[(p-hydroxyphenyl)amino]alanine (Compound 1), N delta-(p-hydroxyphenyl)ornithine (Compound 2), and N delta-(m-hydroxyphenyl)ornithine (Compound 3), were determined against 10 human melanoma and 5 nonmelanoma cell lines. Activities of L-DOPA and 3,4-dihydroxybenzylamine were also measured. Dose-response curves were obtained and concentrations in micrograms/ml required to give 90% inhibition of colony formation (IC90) were calculated. Using a cutoff IC90 of less than 2.5 as a definition of activity. Compound 2 was active in 6 of 10 melanoma and 0 of 5 nonmelanoma cell lines while both Compound 1 and L-DOPA methyl ester were active in 1 of 10 melanomas and 0 of 5 nonmelanomas. Compound 3 was inactive in all cell lines and all IC90 values exceeded 100. 3,4-Dihydroxybenzylamine was active in 3 of 8 melanomas and 1 of 5 nonmelanomas. Regression analysis of IC90 values for Compound 2 and tyrosinase levels in the 15 cell lines yielded a correlation coefficient of 0.93 (P less than 0.001). By contrast, a similar analysis for 3,4-dihydroxybenzylamine gave a correlation coefficient of 0.17 (P greater than 0.05). Spectrophotometric data indicated that Compounds 1 and 2 were oxidized by tyrosinase to quinones. Cytotoxic activity was blocked by the tyrosinase inhibitor phenylthiocarbamide. Since the rates of activation of Compounds 1 and 2 were identical, the higher activity of Compound 2 was probably due to its higher lipophilicity and greater intracellular accumulation. Compounds 1 and 2 exhibited greater potency and selectivity against malignant melanoma than did the natural product L-DOPA methyl ester.

In vitro and in vivo cytotoxicity of rhodamine 123 combined with hyperthermia.

Because both Rhodamine 123 (R123) and hyperthermia have been shown to be cytotoxic, we examined their effect, independently and in combination, on five different human malignant cell lines in vitro and on cultured melanoma cells grown intradermally in nude mice. The cell lines examined include two human melanomas, UCLA-SO-M14 and UCLA-SO-M21, the colon cancer cell line HT29, the human lung cancer cell line P3, and the human breast cancer cell line B231. R123 and hyperthermia, when used in combination, were found to be cytotoxic for these five different human malignant cell lines in vitro. The two agents together appear to enhance the cytotoxic effect of each alone, as documented by synergistic ratios ranging from 2.31 to 45 for the different cell lines. In the "nude" mouse model, animals were treated with a combination of R123 and hyperthermia (43 degrees C for 90 min). A statistically significant (P = 0.04) decrease in tumor growth rate was observed when compared with the rate of tumor growth in untreated animals. The results suggest a potential role for R123 in combination with hyperthermia in the treatment of malignant cells.

Prospective correlative chemosensitivity testing in high-dose intraarterial chemotherapy for liver metastases.

Clinical response of liver metastases treated by high-dose intraarterial chemotherapy (HDIAC) delivered via the hepatic artery was predicted by a modification of the human tumor colony-forming assay (HTCFA) originally described by Hamburger and Salmon [Science (Wash. DC), 197:461-463, 1977. In a first set of experiments, the immediate clinical response to HDIAC was determined in 12 patients with colorectal liver metastases. Biopsies were taken immediately before and after HDIAC, and cells were plated in the HTCFA. Three patients received intraoperative 4-epidoxorubicin and another 9 received mitomycin C by 15-min intraarterial infusions. Sensitivity in the HTCFA was defined as 50% inhibition of colony formation in tumors exposed to the chemotherapeutic agent, compared to the untreated controls. Clinical response was accurately predicted by the HTCFA in 11 of 12 cases. Eight patients had a regression of disease following HDIAC treatment with mitomycin C, as evidenced by either greater than 50% reduction in carcinoembryonic antigen serum level (7 patients) or regression of tumor by computed tomography scan (1 patient). Three patients had no evidence of clinical response to epidoxorubicin, and their tumors were resistant to epidoxorubicin in the HTCFA. One tumor was sensitive to mitomycin C in the HTCFA, but serum carcinoembryonic antigen in the patient continued to increase following HDIAC. The HTCFA was also performed on untreated biopsies following incubation in vitro with the drug used for HDIAC. Results correlated with clinical response in all 12 cases. In a second set of experiments, the HTCFA was used to predict the long-term clinical response to HDIAC of 30 patients with liver metastases. One patient had breast cancer metastases, one patient had carcinoid liver metastases, 4 had liver metastases of malignant melanoma, and 24 patients had colorectal liver metastases. All 21 of the patients whose tumors were sensitive in vitro had clinical response, while 6 of 9 patients predicted by the HTCFA to be resistant had no clinical response. Our results demonstrate a high correlation between the HTCFA and clinical response.

Development of a miniaturized, improved nucleic acid precursor incorporation assay for chemosensitivity testing of human solid tumors.

Two technological problems limit the usefulness of chemosensitivity assays: low success rates (generally 30-60%); and the requirement for large numbers of tumor cells (5 X 10(5)/dish). To solve these problems, we developed a miniaturized, improved, nucleic acid precursor incorporation assay (MINI-assay). In this new assay, 0.3-1.5 X 10(5) tumor cells were plated in double-layer agarose in 16-mm wells of a Costar (No. 3524) 24-well cluster dish. After 72 h of incubation, 5 microCi [3H]thymidine were added to each well. After an additional 24 h of incubation, the trichloroacetic acid-precipitable material was collected and counted by liquid scintillation. We found that 280 of 351 (80%) human solid tumors gave evaluable chemosensitivity results. Labeling efficiency was optimum when the plating density was between 1.5 and 3 X 10(4) cells/well. Radioisotope uptake was less efficient in 35-mm Petri dishes and in the 7-mm wells. The MINI-assay was particularly suitable for small specimens (less than 1 g) and for tumor types that usually yield small numbers of viable tumor cells (19 of 30 breast cancers and 56 of 71 sarcomas were evaluable). The artifacts of colony counting (cell clumps, debris, clots) were also eliminated with this assay. With high evaluability rates, the requirement of fewer cells, a short duration (5 days), and ease of quantitation, the MINI-assay is widely applicable to chemosensitivity testing in human tumors.

Clinical correlations with chemosensitivities measured in a rapid thymidine incorporation assay.

A rapid assay for in vitro chemosensitivity testing measuring [3H]thymidine incorporation has been developed. Results of this assay correlate highly with chemosensitivities determined by the soft-agar clonogenic assay. A correlative study was carried out on 219 solid tumor specimens to assess the ability of the rapid assay to predict clinical response to antineoplastic therapy. One hundred forty-two of 219 tumors (65%) yielded chemosensitivity data. Of these, 33 were evaluable for in vitro-in vivo correlations. In vitro sensitivity (greater than or equal to 80% inhibition of thymidine uptake) was associated with clinical response in 6 of 13 patients. In vitro resistance was associated with progressive disease in 20 of 20 patients. The rapid assay offers several advantages over the soft-agar clonogenic assay, including higher success rate, avoidance of clumping artifact, shorter time course (5 days), and very low false-negative rate. Further refinement may be necessary, but the rapid assay appears to have potential for individualizing solid tumor chemotherapy.

Comparison of drug sensitivity among tumor cells within a tumor, between primary tumor and metastases, and between different metastases in the human tumor colony-forming assay.

The human tumor colony-forming assay was used to compare chemosensitivity among tumor cells within a primary tumor, between primary tumor and metastases, and between different metastases. No significant differences in cloning efficiency were found in any of the three comparison studies. However, considerable differences in chemosensitivities were observed between different parts of the same tumor and between the primary tumor and metastases. Two different parts of the same tumor were comparably assayed for nine primary tumors. In nine paired samples which allowed in vitro drug sensitivity testing, there was no satisfactory correlation of sensitivity to cytostatic drugs. Cell suspensions were prepared from 28 primary tumors and from metastases taken from the same patient. In 14 paired samples which formed sufficient colonies for determination of drug effect, the data showed no satisfactory correlation of chemosensitivity between a primary tumor and its metastases. Both tumor samples from different metastatic sites of the same patient formed sufficient colonies in seven of eight instances. In the seven paired samples, there was strong association of chemosensitivity (p less than 0.005). The results indicate that the reported discrepancies of in vitro and in vivo results in clinical trials using the tumor colony-forming assay for predicting resistance or sensitivity to cytostatic drugs may be due to therapeutic heterogeneity among tumor colony-forming units within a primary tumor and between a primary tumor and its metastases.

Application of a human tumor colony-forming assay to new drug screening.

The applicability of a human tumor colony-forming assay to drug screening was investigated in terms of feasibility, validity, and potential for discovering new antitumor drugs. Feasibility was addressed in a pilot study during which basic methods, appropriate assay quality controls, and a standardized protocol for screening were developed. Considerable variability was noted in the availability and colony growth of different tumor types. The majority of the evaluable experiments utilized breast, colorectal, kidney, lung, melanoma, or ovarian tumors. For many tumor types, little evidence of growth was observed, or only rare specimens formed colonies. Colony-forming efficiencies ranged from 0.05 to 0.11% for the six most useful tumors listed above. A set of quality control measures was developed to address technical problems inherent in the assay. Testing of standard agents in the pilot study established that most of these agents could be detected as active. However, it also identified three assay limitations: compounds requiring systemic metabolic activation are inactive; medium constituents may block the activity of certain antimetabolites; and compounds without therapeutic efficacy may be positive in the assay. The assay categorized nontoxic clinically ineffective agents as true negatives with 97% accuracy. Of 79 compounds which were negative in the current National Cancer Institute prescreen (leukemia P388), 14 were active in the assay. Several demonstrated outstanding in vitro activity and are structurally unrelated to compounds already in development or in clinical trials. A subset of these active compounds were found to lack activity in a P388 in vitro colony-forming assay. This indication of differential cytotoxicity to human tumor cells makes this subset of compounds particularly interesting as antitumor drug leads. The demonstrated sensitivity to most standard agents, discrimination of nontoxic compounds, reproducibility of survival values within assays and between laboratories, and evidence of ability to identify active compounds which were negative in the in vivo prescreen suggest that the human tumor colony-forming assay may be a valuable tool for antitumor drug screening. However, because of technical limitations inherent in the current assay methodology, this must be confined to selected tumor types and limited to screening on a moderate scale.

A new micromethod for the detection of HL-A antigens on cultured human tumor cells.

A rapid microcytotoxicity assay for the detection of HL-A antigens on tissue culture cells derived from human solid tumors is described. Tumor cells were prelabeled with 125Iododeoxyuridine. Isotopically labeled tumor cells were reacted with up to 37 highly selected HL-A antisera and diluted rabbit complement. Results of the HL-A typing of nine human tumor cell lines are reported. Three melanoma cell lines showed individually distinct HL-A profiles at the first HL-A locus which agreed with the antigenic pattern of the tumor donor's autologous lymphocytes. Less reactivity was noted with HL-A antisera defining second locus specificities on the three melanoma cell lines, whereas some other cell lines showed more HL-A reactions than required to present a "full house". This method obviates the necessity for visually enumerating residual tumor target cells.

In vitro assessment of antineoplastic therapy. New indication for thoracotomy?

Selection and determination of the efficacy of antineoplastic agents has been dependent upon the trial and error method of observing measurable disease. Such methods subject the patient not only to loss of precious time but to needless toxicity if the drug is ineffective. The clonogenic assay, an in vitro assessment of tumor cell sensitivity to antineoplastic agents, has the potential for individualizing therapy. In this assay, tumor cells exposed to various drugs are cloned in soft agar. In the 16 primary and 24 metastatic pulmonary tumors tested with this technique, a growth rate of 80% was achieved. Fifty-five percent of the primary tumors and 60% of the metastatic lesions responded in vitro to one or more of the test drugs. There were twelve possible correlations between in vitro and in vivo results. In four of 12 assays, in vivo sensitivity was predicted and three of four patients demonstrated a clinical response. No drug that was inactive in vitro had activity in vivo. Prior knowledge of in vitro sensitivity may dictate a more aggressive surgical approach to pulmonary metastatic disease, whereas in vitro resistance would call for more conservative treatment. Just as with estrogen receptor status in breast cancer, data derived from the clonogenic assay may ultimately be of such import that thoracotomy would be warranted solely for the purpose of obtaining tissue for the assay.

Mediation of immune responses to tumor antigens in vitro by immune RNA.

It was shown that normal nonimmune C3H mouse spleen cells became specifically cytotoxic to chemically-induced syngeneic C3H tumor cells by incubation with xenogeneic I-RNA extracted from the lymphoid organs of specifically immunized guinea pigs. This response was specific for the tumor used to immunize the I-RNA donor. In a totally syngeneic system, we showed that syngeneic I-RNA extracted from the spleens of tumor-bearing rats mediated cytotoxic immune reactions which were directed specifically against the tumor-associated antigens of syngeneic rat tumor target cells. Active antitumor I-RNA synthesis in the lymphoid organs of I-RNA donor animals reached a maximum between days 14 and 21, depending on the route of administration and the nature of the immunizing tumor. Active I-RNA preparations were insensitive to treatment with deoxyribonuclease or pronase, but were inactivated by ribonuclease treatment; thereby indicating that the active moiety was one or more species of RNA. The active fractions of the I-RNA preparations had sedimentation values in sucrose density gradients of 12-16S, and comprised only a small fraction of the total RNA present in the lymphoid cells. Active antitumor I-RNA appeared to be localized in the cytoplasm of sensitized lymphoid cells, rather than in the nucleus. Lymphocytes from normal human donors as well as from cancer patients, when incubated with xenogeneic or allogeneic I-RNA, became specifically cytotoxic for human tumor cells in vitro. Crossreactivity among tumors of the same histologic type was observed, but not crossreactivity with tumors of other histologic types. Xenogeneic I-RNA extracted from the lymphoid organs of donor animals immunized either iwth tumor cells or normal tissues, following incubation with normal allogeneic lymphocytes, mediated cytotoxic immune reactions which were directed both against tumor-associated antigens and normal transplantation antigens. However, when autologous lymphocytes were used as effector cells, only immune reactions directed against tumor-associated antigens were observed. Allogeneic I-RNA extracted from peripheral blood lymphocytes of human cancer patients mediated specific cytotoxic immune reactions that were directed against common tumor-associated antigens shared by human tumors of similar histologic type. I-RNA's directed against "self" normal cell surface antigens appear to be recognized as self by lymphocytes, and immune responses against these self antigens are not elicited. On the other hand, I-RNA's directed against "nonself" tumor-associated antigens induce lymphocytes to effect specific antitumor immune responses. Our data are consistent with the hypothesis that I-RNA is an information-containing ribonucleic acid molecule capable of mediating immune reactions in vitro which are specific for the tumor-associated antigens of the tumor used to immunize the I-RNA donor.

Natural thermal resistance of human tumor cells and the role of prostaglandin.

We found that human malignant melanoma cells had varying thermal sensitivity and that some exhibited natural thermal resistance, a heretofore unrecognized phenomenon. Samples of 73 melanoma cell suspensions were heated at 42 degrees C for 1 hour before plating in the soft agar clonogenic assay or the thymidine assay for proliferating cells. We observed greater than 75% cell kill after hyperthermia in 39 (53%) tumors. Native thermal resistance was apparent in 17 (23%) tumors and growth enhancement in 17 (23%) at this temperature and exposure time. We postulated that prostaglandin, known to protect stomach mucosa against thermal injury, has a role in stabilizing the tumor cell membrane exposed to heat. Three melanoma cell lines known to be thermosensitive were heated to 42 degrees C without and with exogenous prostaglandin E2 (PGE2). The survival of colony-forming cells was increased in all three lines in the presence of 30 microM PGE2. A naturally thermoresistant cell line was exposed to 1 microgram/ml indomethacin for 24 hours before hyperthermic treatment. The survival of colony-forming cells was significantly decreased compared to cells not treated with indomethacin. The addition of 30 microM of exogenous PGE2 to indomethacin-treated cells reestablished thermal resistance. These preliminary data suggest that some tumor cells synthesize prostaglandins to render the cell thermoresistant. Treatment with indomethacin blocks prostaglandin synthesis and induces thermosensitivity. These discoveries may have important clinical applications for hyperthermia treatment of human cancers.

Heterogeneity of human metastatic clones by in vitro chemosensitivity testing. Implications for the clinical application of the clonogenic assay.

To determine whether in vitro chemosensitivities of clones from metastases of human tumors varied, biopsy specimens of two separate metastatic lesions were obtained from 75 patients. Significant tumor growth (greater than 30 colonies per plate) occurred in both specimens in 49 of the 75 patients. Biopsies were performed simultaneously in 22 patients (synchronous) and sequentially in 27 patients (metachronous). Tumor samples were scored as being resistant (less than 50% inhibition of colony formation) or sensitive (greater than 50% inhibition) to each drug. The two tumor samples from each patient were compared for differences in sensitivity after exposure with standard panels of chemotherapeutic agents. A total of 272 individual drug comparisons were possible. Variations were evident in 49 (40%) of 121 metachronous drug comparisons and in 42 (28%) of 151 synchronous drug comparisons. We concluded that multiple metastases of human tumors can vary in chemosensitivity in vitro; this finding seemed to reflect the heterogeneity of metastatic clones.

Biopsy of thoracic neoplasms for assay of chemosensitivity. New indication for thoracotomy.

Eighty-six thoracic neoplasms, both primary and metastatic, were removed at thoracotomy from 86 patients and were tested for chemosensitivity in the clonogenic assay. Substantial tumor growth was achieved in 79% (67/86). Fifty-two percent (16/31) of the primary lung tumors and 45% (15/33) of the metastatic tumors were sensitive to at least one tested drug. Clinical correlations between in vitro chemosensitivity and in vivo response were possible in 20 patients. The assay was 83% accurate for predicting in vivo sensitivity and 86% accurate for predicting in vivo resistance. The value of the assay as it pertains to lung cancer has been demonstrated. On the basis of the results, thoracotomy is indicated in selected patients as a diagnostic procedure to obtain tissue for chemosensitivity testing.