[MOL] Diagnostic assays and chemobiotherapy.... [02656]

American Society of Clinical Oncology 36th Annual Meeting
Day 4 - May 23, 2000

Diagnostic Assays and Chemobiotherapy

Jeffrey Weber, MD

The prevention, assessment, and treatment of melanoma were important topics at the 36th Annual Meeting of the American Society of Clinical Oncology in New Orleans, Louisiana. Vaccine approaches and new prognostic markers were discussed alongside several new strategies for single- and multiple-agent treatment. Unfortunately, none of these chemotherapeutic strategies broke the long-standing threshold of the 20% response rate. However, diagnostic refinements using PCR and ultrasound and a novel approach to postchemobiotherapy treatment for metastatic disease generated some excitement.

Toward a Vaccine
Dr. Shurin and colleagues,^[1] from the University of Pittsburgh, Pennsylvania, presented data from a study that suggested that tumor supernatants from murine melanomas induced downregulation of important dendritic cell markers and apoptosis of dendritic cells (DC), with increases of proapoptotic Bax protein and decrease of antiapoptotic Bcl-2. These data raise concerns about the use of DC in cancer patients, since tumor suppression of DC function would compromise the efficacy of adoptive transfer -- as discussed in a number of posters previously presented at this conference.
Over the past decade, Dr. Bystryn and colleagues,^[2] from Kaplan Comprehensive Cancer Center in New York, NY, have tested a shed polyvalent melanoma vaccine in a large number of melanoma patients. They summarized results from 31 patients in which extensive immunologic analyses of immune reactivity in fresh blood samples were performed. They found that 17 of 20 strongly HLA-binding epitope peptides were recognized by peripheral blood melanoma-specific T cells derived from immunized patients. The results seem to indicate that the shed vaccine induced a broad immune response, and that a larger variety of antigen peptides than previously thought are capable of eliciting an immune response in patients. The assay used was an ELISPOT test, which enumerates antigen-specific cytotoxic T lymphocytes (CTLs) from fresh blood without expansion. Of note, there was profound patient to patient variation in ELISPOT response to different peptides, with no evidence seen of any HLA-A3, -A11, or -B7 restricted immunodominant peptide.
Dr. Xie and colleagues,^[3] from the New York University School of Medicine, New York, NY, analyzed a subset of 24 patients with resected stage III disease who received the same shed vaccine to determine whether the development of IgG and IgM antimelanoma antibodies correlated with clinical outcome. Using Western blots of whole melanoma extracts, they found that 54% of patients had IgM responses and 20% had IgG responses. Overall and disease-free survival correlated with elevated IgM but not IgG levels, indicating that IgM may serve as a surrogate marker for clinical benefit.
An update to a DNP-modified autologous melanoma vaccine trial for resected stage III melanoma was presented by Dr. Berd and colleagues,^[4] from the Thomas Jefferson University in Philadelphia, Pennsylvania, who reported their experience with 214 patients with a median follow-up of 3.6 years. The 5-year overall survival was 47%, and survival correlated with the number of positive nodes. DTH reaction to the unmodified cells was associated with significantly longer survival (P < .001). Based on these data, the authors are currently engaged in a randomized study of a DNP-hapten modified autologous tumor cell vaccine compared with high-dose IFN-alpha.

Assays for Prediction of Clinical Course
Dr. Hsueh and colleagues,^[5] from the Johns Hopkins Cancer Center in Baltimore, Maryland, discussed a serum assay that has been shown to detect a complex in melanoma patients consisting of an antibody and a 90 kD glycoprotein of unknown function called TA90. Over a 3-year period, 571 patients had TA90 assays performed. The levels in patients with disease were slightly higher, on average, than those in resected patients without disease (P = .042). TA90 levels increased with stage and tumor burden. The sensitivity of the assay for prediction of nodal status was 83%, and for prediction of recurrence, an encouraging 74%. However, 21% of TA90-negative patients still had a relapse compared with 35% of positive patients. The range of values for the assay overlapped the mean. The authors felt that this validated the use of TA90 as a predictive assay.
Dr. Benez and colleagues,^[6] from the University of Tübingen, Germany, used immunomagnetic bead sorting with an antibody known as 9.2.27 to identify melanoma cells circulating in peripheral blood. In spiking experiments, 1 cell per mL of blood was detected. No detectable cells were found in blood from controls or from patients with primary cutaneous melanoma, but they were found in 3 of 29 (10%) patients with resected stage III disease and 13 of 85 (15%) patients with metastatic stage IV disease. The results suggest that while the assay would not be practical for early detection of metastases, viable melanoma cells circulate.
Another attempt to define a prognostic factor in melanoma was made by Dr. Salti and colleagues,^[7] from the University of Illinois at Chicago, who wanted to determine whether there was a correlation between expression of microphthalmia transcription factor (Mitf) and survival in melanoma. After immunohistochemical staining for nuclear Mitf of tissue from primary melanoma resections, they found that disease-free and overall survival were better in patients with Mitf-positive samples than in those who were negative (P = .008 for disease-free survival; P = .005 for overall survival). In this cohort of 63 patients with intermediate thickness melanomas, Mitf staining was considered to be a prognostic marker.

Vaccine Assays
Dr. Keilholz and colleagues,^[8] from the Free University in Berlin, Germany, described 2 trials of 4 melanoma peptides derived from tyrosinase and administered with GM-CSF. Patients who either had metastatic disease or who were rendered free of disease by resection (NED) were treated with multiple injections of peptides. Analysis of peripheral blood samples was by the ELISPOT test. In 7 of 11 NED patients, reactive CTL were generated, compared with a positive assay in only 2 of 13 patients with metastatic disease. Of note, the 2 immunologic responders were also the only patients with evidence of clinical benefit, having stable disease and a mixed response, respectively. Modest but reproducible levels of positive CTL were observed. The authors concluded that immune suppression by the metastatic disease burden may compromise the effectiveness of vaccination.
Dr. Eisnebeis and colleagues,^[9] from the University of Illinois at Chicago, used a nested PCR technique for multiple melanoma antigens to analyze 82 specimens of peripheral blood or sentinel nodes. As has been shown previously, PCR positivity increased with stage, but only 68% of stage IV patients were positive for at least 1 marker. In stage III resected patients, 5 of 9 PCR-positive patients have relapsed, compared with 1 of 10 PCR-negative patients, suggesting that PCR positivity may have prognostic significance. Dr. Palmieri and colleagues,^[10] from the Institute of Molecular Genetics in Alghero, Italy, tested peripheral blood and sentinel lymph node samples using multimarker PCR and found that an increased number of positive markers correlated with increased chance of relapse.
Dr. Reynolds and colleagues,^[11] from the New York University School of Medicine, New York, NY, used the presence or absence of circulating melanoma cells that were marker-positive as a surrogate for melanoma progression in a vaccine trial. Prior to treatment with a polyvalent shed cell vaccine, 12 of 41 patients were PCR positive; after 3 months 9 were positive; this figure dropped to 3 at 5 months. Of 62 patients with negative cells, the relapse-free survival was 17.2 months, and for the group with positive cells, the survival was 13.1 months, a statistically significant difference (P = .003). Surprisingly, the median survival in patients who converted from positive to negative was 12.9 months, compared with 16.8 months for patients who remained negative.

Using Ultrasound as an Assay
Dr. Lassau and colleagues,^[12] from the Institute Gustave-Roussy in Villejuif, France, used ultrasound in 65 patients with primary cutaneous melanoma to estimate the thickness of the tumor noninvasively, and color Doppler ultrasound to determine vessel density as a surrogate marker for angiogenesis. The ultrasound determination strongly correlated with micrometer Breslow thickness, and of 65 patients who underwent a color Doppler measurement, none of the 41 without vessels relapsed, none of 8 with single vessels relapsed, and 6 of 13 patients with more than 2 vessels developed metastases. The results suggest that Doppler ultrasound may have value as a noninvasive prognostic predictor.
Dr. Voit and colleagues,^[13] from Humboldt University, Berlin, Germany, also used ultrasound for diagnosis, but as a tool to prospectively analyze in-transit and draining lymph node basins in resected stage I/II patients. Metastases were detected by physical exam in 61 of 242 cases, but 240 of 242 were recognized by ultrasound, with an excellent predictive value of 83%. These data strongly support the further testing and refinement of ultrasound for the detection of recurrence of melanoma. By contrast, Dr. Wagner and colleagues,^[14] from the Indiana University Cancer Center Melanoma Program in Indianapolis, Indiana, found FDG-PET imaging to be a poor screening tool, since sensitivity for detection of regional nodal metastases was 0.24 and the overall sensitivity for the detection of stage IV disease was 0.08.

Chemotherapeutic Strategies
Weekly paclitaxel (150 mg/m²) was tested as second-line treatment for metastatic melanoma by Dr. Zonder and colleagues,^[15] from the University Hospital in Zaragoza, Spain. Of 15 eligible patients, there were 2 partial responses and 1 mixed response, for a total response rate of 13%, suggesting that the regimen had minimal activity and was probably not worth pursuing in its present form.
Temozolomide has been used extensively in phase II/III clinical trials for melanoma, and Bafaloukos and colleagues from the Hellenic Cooperative Oncology Group^[16] in Athens, Greece, tested it with doxetaxel for patients with stage IV disease. Fifty-one patients received monthly cycles of oral Temozolomide at 150 mg/m² for 5 days plus an injection of doxetaxel at 80 mg/m². For the 40 evaluable patients, there were 2 complete responses and 8 partial responses, including responses in 2 of 5 patients with CNS disease, for a total response rate of 25%. These encouraging data merit a repeat multicenter phase II trial. Temozolomide was also evaluated by Dr. Arance and colleagues,^[17] from the Christie Hospital in Manchester, England, who compared response rates and survival in metastatic melanoma patients receiving either 5 doses spaced 8 hours apart of temozolomide, temozolomide for 5 days with IFN-alpha, or temozolomide for 5 days with thalidomide. Median survival was only 6.5 months, and there was no clear difference in survival in any of the arms, with an overall response rate of 16%.
Paclitaxel and DTIC were evaluated in a phase II trial by Feun and colleagues,^[18] from the Sylvester Comprehensive Cancer Center in Miami, Florida. They administered DTIC 800 mg/m² with paclitaxel 135 mg/m² every 3 weeks to patients with stage IV disease. paclitaxel was escalated to 250 mg/m² and DTIC to 1000 mg/m². The response rate in 25 patients was 12%, not different from the response rate of either agent alone, suggesting that no additive effect was observed.

Chemobiotherapy Regimens
A number of refinements of chemobiotherapy regimens were discussed. Dr. O'Day and colleagues,^[19] from the John Wayne Cancer Institute at Saint John's Health Center in Santa Monica, California, used a regimen that has been published with decrescendo IL-2, IFN-alpha, DTIC, cisplatin, and vinblastine with GM-CSF to treat patients with stable disease or a partial response after four 28-day cycles of IL-2 and GM-CSF. Four of 20 partial responders became complete responders, and 1 SD became a partial responder, for a response rate of 25%. Treatment was well tolerated, with no IL-2 dose reductions over 12 months.
Dr. Kamanabrou and colleagues^[20] treated 109 patients with a regimen of sequential IL-2, IFN-alpha, cisplatin, DTIC, and carmustine. The overall response rate was 38%, with 11% complete responses. More than 31 months after treatment, 8 of 12 complete response patients are still in remission more than 31 months from treatment. An outpatient regimen of DTIC, cisplatin, IL-2, IFN-alpha and GM-CSF in 14-day cycles every 28 days was used by Gajewski and colleagues^[21] from the University of Chicago. Among 9 patients treated, 2 responses have been seen so far with good tolerance.^[21] Tres and colleagues,^[22] from the University Hospital in Zaragoza, Spain, put together an outpatient regimen of the same reagents given sequentially. What made their regimen unique was that steroids were allowed on day 1 prior to the immunotherapy as an antiemetic. Three patients were admitted to the hospital because of toxicity, but there were 7 responders of the first 14 evaluable patients. Median survival of the group so far is 21 months, which is impressive.
Dr. Gibbs and colleagues,^[23] from the University of Colorado in Denver, treated 60 patients with an inpatient concurrent chemobiotherapy regimen that included a decrescendo IL-2 regimen to reduce toxicity, GM-CSF after chemotherapy for support, and substitution of temozolomide for DTIC. The response rate was a disappointing 20% with only 1 complete response, but patients with treated CNS disease were not excluded. Median survival overall was 11 months, not different than other published chemobiotherapy regimens.
Dr. Kersten and colleagues,^[24] from the Netherlands Cancer Institute, Amsterdam, The Netherlands, treated 40 patients with IL-2, IFN-alpha, GM-CSF, and escalating cycles of temozolomide. Starting at a dose of 150 mg/m² for 5 days, it was increased to 250 mg/m². Fifteen of 19 patients had no significant toxicity at 200 mg/m², but at 250 mg/m², 7 of 16 patients had grade III or IV thrombocytopenia. Responses were 17% at 150 mg/m², 33% at 200 mg/m², and 50% at 250 mg/m². Because there was some evidence of a dose-response effect, the authors recommended starting at a dose of 200 mg/m² and escalating to 250 mg/m² if thrombocytopenia was not an issue.
Dr. Becker and colleagues,^[25] from the University of Wuerzburg in Germany, chose to use the alkylator fotemustine combined with IL-2 and IFN-alpha for patients with metastatic uveal melanoma. In an unusual twist, patients with disease limited to the liver (50%) received intrahepatic arterially infused fotemustine, whereas those patients with extrahepatic disease were given the chemotherapy intravenously. Of 33 evaluable patients, 7 showed a partial response. Median survival was 10 months. The novel cytokine IL-12 and interferon-alpha were tested in a phase II trial by Carson and colleagues from Ohio State University in Columbus.^[26] No objective responses were seen, but a biologic effect of the IL-12 was observed at the highest dose of IL-12 (300 ng/kg given intravenously).

Clinical Implications

IgM may serve as a surrogate marker for clinical benefit in melanoma vaccine studies.
TA90 and Mitf may be effective markers for disease progression.
A number of chemobiotherapy regimens have shown promise for the treatment of melanoma patients.