Treatment of aggressive lymphoma in relapse is difficult. Patients who initially present with these diseases often know they have a malignancy considered curable in many cases, and diagnosis of relapse can be devastating. For this reason, it is useful to know the individual patient's risk of relapse prior to starting initial therapy, since it may be appropriate to treat patients with poor prognoses with intensive programs or investigational studies. In the private practice setting, most patients with these diseases receive CHOP or similar cyclophosphamide and doxorubicin-containing regimens at the time of initial diagnosis. However, there are certain disease-related features which determine whether these patients have a high or low risk of relapse, and investigators are now using combinations of these features to determine which patients may be safely treated with CHOP and which may benefit from more intensive chemotherapy management. For example, the International Prognostic Factor Index system, now in common usage, delineates four different groups of patients with differing complete remission, freedom from progression, and overall survival rates. The Tumor Score System, developed at MDACC, delineates only two groups with very different survival rates, and may be a better scoring system for patients with diffuse large cell lymphomas, primarily because of its inclusion of the serum beta(2)-microglobulin level prior to treatment, an important predictor of relapse. In addition to pretreatment features, certain treatment-related factors are also important in determining the risk of relapse, including the dose of chemotherapy administered and the rapidity of response. Results of a gallium scan with SPECT imaging may be an important method of confirming complete response, and should be incorporated into treatment programs, whether the treatment is standard CHOP or an investigational program. For the patient with relapse or progressive disease following induction with CHOP or a similar regimen, the type of response to initial therapy plays an important role in determining potential response to salvage therapy, including high-dose therapy followed by stem cell rescue. Patients for whom initial treatment fails to achieve any response have a very poor chance of responding to any currently used standard-dose program for relapse. Those with partial responses have a better chance of responding to relapse therapy, but a high risk of disease progression or early relapse, and those with a prior complete response to initial therapy have a good chance of responding to relapse therapy, especially those in whom the complete response lasted more than a year. For these reasons, stem cell transplant (SCT) protocols routinely require complete response with initial therapy as a requirement for entry, although "good partial remission" may be acceptable at certain centers. Other limitations for SCT protocols include age greater than 60 or 65 years, significant chronic obstructive pulmonary, renal, or cardiac disease, a poor performance status, and central nervous system or marrow involvement. For these reasons, there is a continued need for newer treatment programs which offer the potential for higher response rates and better survival rates, not only for those for whom SCT is not an option, but also for those who must have an adequate response to "standard dose therapy" prior to selecting SCT as a treatment option. Three broad groups of relapse therapy for aggressive lymphoma have been described, based upon the drugs contained within these regimens. These include platinum-based, mitoxantrone-based, and ifosfamide-based chemotherapy regimens. Results with these programs vary widely and are likely different because of tumor-related features prior to relapse therapy, including size of mass, beta(2)-microglobulin level, LDH level, and type of response to initial therapy. Other features, such as dose of therapy, specific drugs utilized, and number of prior treatments also play important roles in determining results with relapse therapy. In a study of DHAP followed by transplant or more DHAP, DHAP induced a response in 56% of patients, and at 5 years, significantly more of the responders to DHAP who were subsequently treated with high-dose therapy and bone marrow transplant were free of disease compared to those who continued to receive DHAP after response to this regimen. Therefore, high-dose therapy is clearly better for DHAP responders than is continued DHAP. However, results for the overall population are still not good when non-responders are included in the analysis, and DHAP, a first-generation platinum regimen, may not be the optimal regimen to use prior to high-dose therapy followed by peripheral stem cell rescue. At MDACC, we have extensively investigated various combinations containing ifosfamide and etoposide. The most recently reported regimen, MINE-ESHAP, contains mesna, ifosfamide, mitoxantrone, and etoposide, followed after adequate response with etoposide, methylprednisolone, high-dose cytarabine, and continuous infusion as cis-platinum, a second-generation platinum regimen. This strategy resulted in a complete response in 47% of the patients treated, with a 44% complete response in patients with intermediate-grade lymphoma, 56% in those with low-grade lymphoma and 36% in those with transformed lymphoma. Results varied according to type of response achieved with initial therapy, and serum LDH and beta(2)-microglobulin levels prior to treatment with MINE-ESHAP. Using more intensive doses of ifosfamide and etoposide, we have described therapy for 36 patients with relapsed aggressive lymphomas, prior to pheresis and SCT. Results of this study are encouraging: 42% entered complete response with ifosfamide-etoposide and the overall survival was 52%, with a progression-free survival of 32%. Therapy with a similar regimen, combining ifosfamide, carboplatin, and etoposide in standard doses (ICE) has also been described. This regimen has been extensively studied in patients with relapsed aggressive lymphomas and Hodgkin's disease, followed by SCT. In patients with relapsed lymphomas, ICE has achieved a 66% complete response rate, with 89% undergoing transplant. Overall survival in these studies is affected by the quality of the response to ICE. The same program was used to treat 65 patients with Hodgkin's disease. The response rate to ICE was 88%, and the 5-year event-free survival for those transplanted was 68%. These factors predicted outcome: B symptoms, extranodal disease, and complete response less than 1 year. Finally, we have recently studied paclitaxel in combination with topotecan for relapsed and refractory aggressive lymphomas. These and newer combinations should be further developed to treat patients in relapse of aggressive lymphomas.