Prospective Clinical Studies

Twelve of the 14 published prospective studies of patients with cancer who were treated with chemotherapy that was selected by in vitro DST met the criteria for this review. Two of the studies identified by the search are not included in this review. One study of 53 patients with advanced malignancies was not included because it did not include information on the types of cancer, clinical or pathologic staging, or treatment status before administration of the in vitro–selected chemotherapy. In addition, there were no data on the chemotherapeutic agent or combination of agents selected by the in vitro testing for treating the patients.¹⁸ Another study of 196 patients used the subrenal capsule assay, which is an in vivo assay.¹⁹ With this in vivo assay, it is difficult to distinguish between tumor growth and inflammatory response against the implanted tumor, so this study was not included in our review. A total of 1,545 patients have participated in 12 studies reported from 1981 through 1997 (Table 1). Ten studies were conducted in single institutions,^{11,13,14,20,22,24,26-29} one was conducted in multiple institutions,²¹ and one was conducted in a cooperative group.²⁵

The technologies for performing in vitro DST for large numbers of samples did not become routinely available until the 1980s.^3,8 Different groups of investigators have designed clinical trials to determine the value of DST in selecting chemotherapy treatments. These clinical trials have been reported relatively recently, with nine of the 12 studies published in the 1990s.^{11,13,14,24-29} The studies have included fairly large numbers of patients, in which a median of 115 patients (range, 45 to 470 patients) have been studied. Six of the 12 reports have more than 100 patients participating in the study, and most focus on a specific type of cancer. Four studies reported on patients with lung cancer, two on ovarian cancer, two on gastrointestinal malignancies, one on melanoma, and three on a variety of cancers.

Patient Characteristics

The vast majority of patients had advanced cancers. Five of the studies reported that all of their patients (867) had advanced cancer, but did not specify stage.^{21,22,24,25,27} Three studies provided specific data on the staging for 309 of 316 patients. Forty-one patients had stage I, 30 had stage II, 117 had stage III, and 121 had stage IV disease of multiple tumor types.^13,28,29 Two studies of patients with small-cell lung cancer (SCLC) accrued 54 patients with limited-stage SCLC and 115 patients with extensive-stage SCLC.^11,14 Two of the 12 studies with 193 patients did not provide data on the degree of tumor dissemination.^20,26 These 10 studies showed that approximately 5% had localized disease (41 with stage I and 30 with stage II), 13% had regional disease (117 with stage III and 54 with limited-stage SCLC), and 82% had advanced disease (867 with advanced cancer, 121 with stage IV, and 115 with extensive-stage SCLC).

A minority of patients had received chemotherapy before entering the clinical trials. Six hundred forty-eight of 1,467 patients (44%) enrolled in 11 trials had been treated with chemotherapy. One study of 78 patients did not provide information on previous chemotherapy treatment of patients before entering the clinical trial.²⁶ Two studies of 169 previously untreated patients with SCLC were entered on the trials and then treated with 12 weeks of conventional chemotherapy. The patients were then eligible for in vitro–selected chemotherapy.^11,14 Patients with limited-stage SCLC and those with extensive-stage disease who had no response, a partial response, or relapse after the initial therapy were treated with an in vitro best regiment (IVBR) if DST had been performed.^11,12,14

Tumor Tissue Procurement

One potential limitation on the ability to treat patients with chemotherapy selected by in vitro DST involves the obtainment of tumor tissue. There are only two studies that have evaluated the ability to obtain tumor tissue from consecutive patients with specific cancers.^11,14 These studies evaluated patients with limited- and extensive-stage SCLC. Tumor tissue could be obtained from 16 of 54 patients (30%) with limited-stage SCLC and 83 of 115 patients (72%) with extensive-stage SCLC.^11,12,14 The patients with more accessible cancer (extensive-stage disease) were more than twice as likely to have a tumor specimen obtained for in vitro DST than patients with less accessible cancer (limited-stage disease). The other 10 studies required a tumor specimen to reach the laboratory before entering the patients onto the study, so the information on obtaining tumor tissue was not available.

The 12 studies of in vitro DST have typically not included a surgical procedure under general anesthesia solely for obtaining tumor tissue for DST. Patients from eight studies had tumor samples collected from malignant effusions, bone marrow containing tumor cells, or solid tumor specimens from procedures performed as part of a diagnostic evaluation or as a therapeutic procedure.^{11,12,20,21,24-26,28,29} In one study of 165 patients with non–small-cell lung cancer (NSCLC), tumor tissue for chemosensitivity testing was obtained from primary resections or staging procedures.¹³ In one study of 51 patients with melanoma, biopsies from skin or subcutaneous sites were the source for DST.²² Forty-five patients with NSCLC had samples obtained for DST without major surgery or incidentally during an otherwise indicated major surgery.²⁷ Fifty-four patients with limited-stage SCLC had 18 surgical biopsies performed under general anesthesia solely for obtaining tumor tissue for DST.¹⁴

DST

Four studies evaluated drug sensitivity by human tumor colony-forming assay,^20-22,25 one by capillary cloning system,²⁴ four by differential staining cytotoxicity assay,^11,13,14,27 and three by the formazan salt 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT) assay^26,28,29 (Table 2). A median of 64% (range, 24% to 95%) of the tumor samples had adequate cell numbers for DST. Three studies that used the MTT assay^26,28,29 had higher percentages of specimens that were able to be analyzed for drug sensitivity (87% to 92%).

The median number of chemotherapeutic agents tested against the tumor specimens was 10 drugs (range, four to 31) in the 12 different trials. In five of the trials, patients were treated with single chemotherapy agents selected by DST.^20-22,24,25 Seven and 12 different single agents were identified in two trials by in vitro DST.^20,22 The other three trials that used single chemotherapy agents for treatment did not specify the single agents selected by DST.^21,24,25 In the other seven trials, IVBR treatment consisted of a median of nine different chemotherapy combinations (range, five to 20).^{11,13,14,26-29} Thus DST selected different single agents and drug combinations for the treatment of the patients.

Efficacy of Chemotherapy Treatment

A mean of only 33% (range, 11% to 88%) of the patients in these 12 studies were able to be treated with IVBR compared with a mean of 47% (range, 0% to 81%) who were treated with empiric or clinician's choice regimen (eight studies). One study selected the two treatment arms based on randomization.²⁴ Four studies did not treat patients with empiric chemotherapy.^20,22,26,28 Patients in the other eight trials who had no assessable lesions or for whom DST was unsuccessful were treated on the empiric chemotherapy arm.^{11,13,14,21,25,27,29}

Seven of the 12 studies compared the outcome of patients treated with chemotherapy selected by in vitro DST with the outcome of those treated with empiric chemotherapy. Nine of the 12 studies allocated the patients in two different groups (IVBR and empiric chemotherapy). Seven studies determined response rates for each group. The other two studies reported response rates only for the IVBR group. The remaining three studies had a single treatment group, and only two of them detailed the response rate. Five of the seven studies that determined the response rate for both groups showed that the overall response rate to IVBR was 11% to 36% higher than the overall response rate to empiric chemotherapy.^{11,21,24,25,27} The response rate to both IVBR and empiric chemotherapy in a study of limited-stage SCLC was 100%.¹⁴ The difference between IVBR and empiric chemotherapy response rate was less than 5% in one of the trials.¹³ Seven studies showed that the mean response rate to in vitro–selected therapy was 26% (range, 10% to 100%) compared with 18% for empiric treatment (range, 0% to 100%).

Five of the 12 studies compared the survival rate of patients treated with chemotherapy that was selected by in vitro DST with the survival rate of those treated with empiric chemotherapy.^{11,13,14,24,25} Three studies determined the survival rate for only patients treated with IVBR.^27-29 The other four studies do not present any data on patient survival. The IVBR treatment group in two of the five studies had a longer median survival than the group of patients treated with empiric chemotherapy. In one study, 19 patients with a variety of advanced cancers treated with in vitro–selected single-agent chemotherapy lived a median of 8.6 months compared with 4.2 months for the 36 patients randomized to empiric treatment.²⁴ A study of patients with limited-stage SCLC showed that eight patients treated with in vitro–selected chemotherapy for their five to eight cycles lived a median of 38.5 months compared with 19 months for the 44 patients treated with empiric chemotherapy.¹⁴ The patients in the other three trials who were treated with in vitro–selected chemotherapy had similar survival compared to those treated with empiric chemotherapy. Two trials of 23 patients with ovarian cancer and 21 patients with extensive-stage SCLC who were treated with in vitro–selected chemotherapy survived a median of 6.2 and 9.1 months, respectively, compared with 101 ovarian cancer and 68 extensive-stage SCLC patients treated with empiric chemotherapy who survived 7.3 and 10.6 months, respectively.^11,25 The other trial showed that 21 patients with advanced NSCLC who were treated with in vitro–selected chemotherapy had a survival duration that was 4 months shorter (7.2 months) than that of the 69 patients treated with empiric chemotherapy (11.2 months).¹³ Only one trial that was not randomized showed a significant difference in survival between the group of patients treated with empiric chemotherapy compared with the group of patients treated with in vitro–selected chemotherapy.¹⁴

Prospective Clinical Studies

We have identified and reviewed 12 prospective trials performed throughout the world. These trials have been limited by several factors that have made it difficult to determine if there is a potential benefit for administering in vitro–selected chemotherapy. These include the small percentage of patients entered onto a trial who actually receive the chemotherapy; most of these patients have already received chemotherapy and/or have cancers that are relatively resistant to chemotherapy, and most of the patients have advanced-stage solid tumors. A small percentage of patients (approximately 33%) entered onto these studies have been able to be treated with an in vitro–selected regimen. This was due to several factors related to the in vitro testing techniques as well as to clinical issues. Approximately 28% of the patients did not have enough tumor cells obtained for testing. Technical reasons for inadequate in vitro drug testing assay include difficulty obtaining sufficient viable tumor cells (when samples are derived from metastatic sites), crushed samples, tumor necrosis, lack of colony formation, and insufficient dye conversion. Fewer than one half of the patients who had adequate tumor cells (33% of the total) were actually treated with a regimen selected by in vitro DST. Therefore, fewer than one third of the patients entering the trials were actually treated with in vitro–selected chemotherapy agent(s).

Tumor Tissue Procurement

There is relatively little data on the ability to obtain tumor tissue from consecutive patients entered onto a study. The only studies we could identify were the studies performed at our institution. Tumor tissue was obtained from 99 of 169 (59%) consecutive patients with SCLC.^11,12,14 All of the other 10 studies required a tumor specimen to reach the laboratory; therefore, information about the ability to harvest tumor specimens was not provided. More than 80% of patients treated on the 12 trials had disseminated advanced-stage solid tumors. Nine of these trials obtained tumor tissue from patients during routine staging procedures. Our analysis of patients with extensive-stage SCLC demonstrated that patients with cancer that is disseminated enough to allow biopsy under local anesthesia survive a shorter period of time than patients who do not.¹¹ Therefore, patients who have tumor that is relatively easy to obtain with procedures performed under local anesthesia may be those who are destined to have relatively short survival. Thus these trials tend to select patients with a poor prognosis who are unlikely to survive adequate lengths of time to assess the potential benefit of chemotherapy selected by in vitro DST.

DST

Most of these trials focused on solid tumors that were relatively resistant to chemotherapy. Two studies evaluated patients with NSCLC, two studies evaluated patients with gastrointestinal malignancies, and one study evaluated patients with melanoma.^13,22,27-29 Three studies evaluated solid tumors that were more sensitive to chemotherapy (SCLC and ovarian cancer).^11,14,20 Both SCLC trials administered in vitro–selected agents after administration of four cycles of empiric therapy, so the cancer had already been treated.^11,14 Three fourths of patients with ovarian cancer had been previously treated with chemotherapy.²⁰ More than one half of the 771 patients who were entered onto three studies that evaluated different tumor types were previously treated with chemotherapy.^21,24,25 Therefore, nearly all these studies were of patients who had previously treated tumors, cancers that were highly resistant to chemotherapy, or both. Forty percent of the patients were treated with a combination of chemotherapeutic agents, and 60% were treated with a single agent. The five trials that treated patients' cancers with single agents that were selected by in vitro DST were less likely to have partial and complete responses (23%)^20-22,24,25 than those treated with combination chemotherapy (36%).^11-14,26-29 All of the reported studies were performed before the new generation of chemotherapeutic agents was available for testing or administration (taxanes, camptothecins, gemcitabine, and vinorelbine).

In vitro DST of patients' tumors may be more likely to provide useful clinical information if drugs are used that do not have a high degree of cross-resistance and if the tumors are heterogeneous in their sensitivity to the agents. The in vitro DST selected seven to 12 different single agents and five to 20 different chemotherapy combinations. This means that in vitro DST thus far has identified many different chemotherapeutic agents for different patients. The introduction of new classes of agents in the last 10 years (taxanes, camptothecins, gemcitabine, and vinorelbine) has provided agents that are not cross-resistant with the other agents tested in the past. Therefore, the introduction of these newer chemotherapeutic agents provides even more potential combinations for treatment.

Preclinical data has shown that the new generation of agents (taxanes, camptothecins, and gemcitabine) has cytotoxicity against lung cancer cell lines that are resistant to the older agents (etoposide and cisplatin). Two SCLC cell lines (NCI-H69 and OC-NYH) were established from previously treated patients.³² Three cell lines derived from the parental NCI-H69 were selected in vitro for resistance to doxorubicin, etoposide, or carmustine. Four other cell lines derived from the parental OC-NYH were selected in vitro for resistance to VM-26, topotecan, camptothecin, and cisplatin. These studies showed lack of cross-resistance between cisplatin and taxane or a topoisomerase I and topoisomerase II inhibitor. Tsai et al³³ studied the in vitro sensitivity of 12 NSCLC cell lines to etoposide, cisplatin, and gemcitabine. He demonstrated that gemcitabine has little cross-resistance to etoposide or cisplatin.

Several clinical studies have demonstrated the lack of cross-resistance between taxanes and anthracyclines, taxanes and platinum compounds, and gemcitabine and platinum compounds. Phase II trials of patients with previously treated ovarian cancer, breast cancer, and NSCLC have shown response rates of approximately 20% or more with the new generation of chemotherapeutic agents (paclitaxel, docetaxel, and gemcitabine). Phase II studies in patients with ovarian cancer previously treated with cisplatin (50 patients) or carboplatin (44 patients) had response rates of 19% to gemcitabine and 48% to paclitaxel.^34,35 Phase II trials for 76 patients who had previously received chemotherapy regimens for breast cancer had a 33% response rate to paclitaxel.^36,37 Phase II trials of 69 women with metastatic breast cancer refractory to anthracycline-based chemotherapy had a 55% objective response rates to docetaxel.^38,39 Docetaxel has antitumor activity against platinum-resistant NSCLC as well. Two phase II studies of 88 patients with NSCLC previously treated with cisplatin or carboplatin had response rates to docetaxel of 21% and 14%.^40,41 The lack of clinical cross-resistance observed thus far with the new generation of chemotherapy agents is encouraging for the development of new clinical studies based on in vitro drug sensitivity assays. The testing has the potential to identify agents that may cause tumor regression in patients who are refractory to other agents.

Efficacy of Chemotherapy Treatment

The response rates have been compared in seven trials. In all of the trials, the patients' response rates to treatment with in vitro–selected chemotherapy were at least as good as those achieved with empiric treatment. Five of the trials had response rates that were more than 10% higher in patients treated with in vitro–selected chemotherapy compared with patients treated with empiric therapy.^{11,21,24,25,27} One trial had a response rate that was within 5% for patients treated with in vitro–selected chemotherapy and empiric chemotherapy.¹³ In one study of limited-stage SCLC, there was no difference in the overall response rate between IVBR and empiric chemotherapy treatment groups.¹⁴ The survival of patients treated with in vitro–selected chemotherapy and empiric chemotherapy has been evaluated in five trials. Two trials had survival durations that were similar for both groups, within 2 months of each other.^11,25 One trial showed that survival was 4 months shorter in the IVBR arm compared with that of the empiric arm,¹³ and two trials showed that survival was 4 and 19 months longer in the IVBR arm compared with the empiric arm.^14,24 Only one of these trials was randomized, and all are plagued with small numbers. The median number of patients treated on the IVBR arms was 23 (range, eight to 246) and the median number treated with empiric therapy was 68 (range, four to 358). Therefore, the survival information is difficult to interpret.

Our evaluation of this data suggests to us that although this technique has been evaluated for more than 20 years, the potential clinical benefit has not yet been adequately examined by the trial design. We believe that the data accumulated thus far show that the in vitro–selected chemotherapy seems to be at least as active as empiric regimens. The technologies for selecting chemotherapeutic regimens are well established.⁸ Future studies should be designed so investigators can define the potential clinical benefit of administering chemotherapy regimens that have been determined by in vitro DST to patients with cancer. The following are recommendations that we believe should be incorporated into future clinical trial designs.

We believe that the appropriate design for testing a tumor would be one that is resected routinely where the tumor can be cultured consistently. This would avoid the problem of studying patients whose cancer is advanced enough so that tumor tissue can be biopsied under local anesthesia. Information should be gathered about the number of tumors that can be collected from consecutive patients to determine the ability to get tumors to the laboratory for in vitro DST. Patients with previously untreated stage III ovarian cancer, stage II and III NSCLC, and locally advanced breast cancer fit this definition. All of these tumors respond to a variety of chemotherapeutic agents, including the recently introduced drugs.

We believe patients could be randomized to receive either empiric or in vitro–selected chemotherapy when the IVBR testing has been completed. This would allow investigators to determine the potential benefit of the in vitro–selected treatment. The important end point should be survival. Many studies have shown an improved response rate,^11,21,24,25 but information about potential benefit for survival needs to be established.