Prognostic factors for survival in patients treated in phase I clinical trials

Background. Patients with advanced or metastatic cancer treated in Phase I clinical trials are considered to have a poor prognosis. Survival from first treatment in a Phase I trial was determined for 349 patients. Univariate and multivariate survival analyses were performed to determine whether potential prognostic factors and distinct risk groups could be identified. Methods. Patients were identified retrospectively from a large data base of patients with advanced or metastatic cancer treated in Phase I clinical trials at the University of Chicago between February 1987 and October 1991. Results. With a median follow-up of 29 months, 10% of patients were alive at the time of analysis. Twentynine percent were alive 1 year after the initiation of Phase I chemotherapy, and the median survival from first treatment in a Phase I study was 6.5 months. Multivariate analysis indicated that a better pretreatment performance status, a higher pretreatment serum albumin concentration, a lower pretreatment platelet count, no prior cisplatin chemotherapy, and a genitourinary or gynecologic cancer diagnosis were predictive of better survival. Three risk groups incorporating these five variables were identified, with median survivals of 12.7, 7.4, and 3.5 months for the good, intermediate, and poor risk groups, respectively. Conclusion. Patients treated in Phase I clinical trials have a median survival of 6.5 months, even though some patients have been treated at early (subtoxic and potentially subtherapeutic) dose levels. The results of this study may allow the identification of patients who are likely to survive long enough to contribute information on acute and cumulative drug-related toxicities and, possibly, tumor response.