|
A predictive model based on five risk factors may help guide researchers and clinicians in the planning of chemotherapy for women with advanced or recurrent cervical cancer, according to researchers with the Gynecologic Oncology Group.
Using a pooled multivariate analysis of three published phase III Gynecologic Oncology Group (GOG) studies involving 816 patients evaluable for response, the investigators identified the following prognostic factors for decreased patient response to cisplatin-based chemotherapy, Dr. David H. Moore said in an interview:
With the exception of race, these factors also were independently associated with increased risk of death, while patient age under 50 years (which was dropped from the list) was predictive of poor response only with cisplatin monotherapy, said Dr. Moore of St. Francis Hospitals & Health Centers in Indianapolis.
Patients with none of the five factors would be considered at low risk for poor response to cisplatin or cisplatin-containing therapy, while having two to three factors would imply medium risk and four or five would place the patient at high risk of poor response.
The investigators used data from the GOG 110, 169, and 179 trials to develop the model. They validated it by analyzing the GOG 149 database of 428 patients in a published phase III trial of cisplatin plus ifosfamide vs. cisplatin/ifosfamide plus bleomycin.
"That test upheld our model, so what we're proposing within the GOG is that in future trials, this model should be used as a defined eligibility criterion such that, if you have a patient deemed at high risk, she should be moved into nonplatinum therapies or biologics," Dr. Moore said in an interview.
There's also a message for clinicians, he added: "If a patient with cervical cancer has a pelvic recurrence, has never had complete response with chemoradiation therapy, is African American, and has a performance status of 2, I think this person would be unlikely to derive any benefit from a platinum-based therapy and perhaps should be steered into an investigational trial."
Dr. Moore noted that response was chosen as the primary end point of the study because data from GOG protocols 110 and 169 showed that adding ifosfamide to cisplatin and paclitaxel to platinum significantly improved response rates but had no effect on survival.
While topotecan improved survival in protocol 179, Dr. Moore cautioned that increasing numbers of patients entering such trials have received prior chemotherapy in addition to radiation. "Even though there's an impact on survival of the addition of an agent to platinum versus platinum alone, the question is, is it really better treatment, or have we pushed the response rates to single-agent cisplatin in our control group to such a low level that it's not so much that our experimental arm is doing a lot better? Rather, it may be that our control arm is now doing so poorly that we see a survival advantage with the combination therapy," he said.
Discussant Dr. Robert L. Coleman said predictive modeling is important for hypothesis generation and for developing strata for future trials. "In addition, this type of analysis gives us valuable insight into anticipated outcomes," said Dr. Coleman of University of Texas M.D. Anderson Cancer Center, Houston.
Limitations of the study, he added, include artificially allocated equal weighting and use of factors not equally represented in the individual trials from which they were derived. "Lastly, it's not entirely clear that such modeling will be applicable to cohorts not treated with platinum or in whom alternative strategies may be used, such as biologics," he said.
Meanwhile, the GOG cervix and vulva committee is planning a chemotherapy protocol for patients with advanced, recurrent/metastatic cervical cancer, said Dr. Moore. "They do not intend to use this model in the next study for defining eligibility. However, they intend to apply the model to see if its predictive value extends beyond platinum-containing regimens," he said.
The next study will be a four-arm trial with platinum being absent from two of those arms, Dr. Moore added.
Moore D.H. et al. Factors predictive of response to cisplatin-based chemotherapy in stage IVB persistent or recurrent cervical carcinoma: A multivariate analysis of three Gynecologic Oncology Group trials. Abstract 5534.
Commentary |
In 1999, the National Cancer Institute issued a clinical alert that declared cisplatin-based concurrent chemoradiation as the standard of care of patients with stages IB-IVA carcinoma of the cervix. The alert was based on the results of five major randomized trials that showed a 24%-51% reduction in mortality with this approach, as opposed to radiation alone or in combination with other radiosensitizers. Prior to that clinical alert, most patients with recurrent or disseminated disease were chemotherapy naive and responded best to cisplatin-based chemotherapy. Gynecologic Oncology Group (GOG) studies since 1999 have shown a marked reduction in the response rate to cisplatin-based regimens; and the GOG phase II effort has identified a number of active nonplatinum agents. These facts prompted Moore et al. to evaluate the GOG database for patient-related characteristics that predicted a lower response to standard platinum-based chemotherapy. The study identifies five risk factors that predict poor response to platinum-based chemotherapy for advanced or recurrent cervix carcinoma: performance status 1 or 2 vs. 0, African American ethnicity, recurrence in the pelvis as opposed to elsewhere, prior concurrent chemoradiation, and recurrence within 1 year of diagnosis. The investigators suggest that presence of four or five of these factors identifies a high-risk patient who should be considered for nonplatinum-based chemotherapy or biologics rather than standard platinum-based regimens. However, while the study certainly suggests that patients with several risk factors will not do as well with standard therapy, insufficient data for the superior efficacy of nonplatinum regimens in these patients argues for clinical trials rather than for the use of nonplatinum regimens as routine practice in these patients. The GOG has planned a study of nonplatinum regimens in this setting to begin the process of assessing whether these regimens represent a realistic and effective alternative approach for these high-risk patients. Until these data are available, the standard of care for advanced or recurrent disease should remain platinum-based combination chemotherapy. — J. Tate Thigpen, M.D.
|