In order to improve survival in a disease where the majority of deaths occur from metastases, the integration of systemic chemotherapy is crucial. Research efforts must continue to focus on refining case selection criteria, improving complete response proportions, and overcoming drug resistance. The blanket recommendation of a single therapeutic strategy such as radical surgery, chemotherapy, or radiation therapy to all patients is quickly becoming an outdated approach. Refinements in the understanding of the clinical, pathologic, and molecular features of urothelial tumors will ultimately improve case selection. Evaluation of NM23 RNA levels, or DNA ploidy and T138 surface antigen expression, which have been shown to correlate with metastatic potential, may hold important therapeutic implications. The use of hematopoietic growth factors has the potential to improve both the tolerance of chemotherapy and complete response proportions, a prerequisite for cure. A recent report from Japan of granulocyte colony-stimulating factor with MVAC and other chemotherapy regimens for urothelial tumors corroborated an initial report in reducing the duration of neutropenia. However, the dose response curves for most of the known active agents are not well defined and, ultimately, new agents and strategies will be required. Gallium nitrate, when administered by continuous intravenous infusion, has significant single agent activity in cisplatin-refractory patients with 9/31 responses (29%), including 6 CRs (19%) and further studies are warranted. Drug resistance remains a major obstacle, and as the mechanisms are unravelled, more rational therapies can be designed. For example, resistance to Adriamycin (doxorubicin; Adria Laboratories, Columbus, OH) and vinblastine, two components in the MVAC regimen, are mediated in part by the MDR1 gene. Attempts are ongoing to identify prospectively those tumors with high levels of expression that may be more amenable to treatment with drugs that are not affected by this mechanism. The neoadjuvant approach allows an in vivo assessment of response to chemotherapy as well as the potential for bladder preservation. In most cases additional therapy directed at the primary is required as clinical understaging is a significant problem and pCR proportions are less than 30%. For some patients, initial surgery followed by treatment based on pathologic criteria may represent a better strategy. In these cases the recommendation for adjuvant treatment potentially limits therapy to a population of patients for whom therapy is essential. Based on available data, this would include patients with positive lymph nodes at the time of surgery. Ideally, patients with invasive bladder cancer should be entered into clinical trials designed to assess the impact of these strategies on survival.(ABSTRACT TRUNCATED AT 400 WORDS)