We are entering an extraordinary era in the treatment of patients with indolent non-Hodgkin’s lymphomas (NHL). Several decades were devoted to expending resources on comparing combinations and permutations of conventional agents, but with no beneficial impact on survival.
Recently, however, new, exciting, and more active agents with unique mechanisms of action have entered the clinical arena. The result has been a dramatic alteration of our therapeutic paradigm. Two of the most important of the new agents that are commercially available are the purine analog fludarabine (Fludara) and rituximab(Drug information on rituximab) (Rituxan), a chimeric anti-CD20 monoclonal antibody.
Fludarabine is now considered a reasonable alternative to alkylating agents as initial therapy of low-grade NHL because it induces higher response rates with less cumulative myelosuppression and a lower risk of secondary malignancies. Indeed, in a recent randomized study of previously untreated patients, fludarabine induced a higher response rate with more complete remissions than a standard CVP (cyclophosphamide, vincristine, prednisone(Drug information on prednisone)) regimen.[1]
Nevertheless, patients inevitably relapse and eventually require salvage therapy. Rituximab has rapidly assumed a role as a treatment of choice for patients with relapsed or refractory disease, avoiding exposure to repeated courses of more toxic chemotherapy programs.
Despite the impressive response rates with these newer drugs, there are still no cures. As a result, a major focus of current clinical research has been to develop more effective combinations such as fludarabine, mitoxantrone(Drug information on mitoxantrone) (Novantrone), and dexamethasone(Drug information on dexamethasone) (FND), developed by McLaughlin and coworkers at M.D. Anderson Cancer Center,[2] and the cyclophosphamide(Drug information on cyclophosphamide)-fludarabine regimen piloted by the Eastern Cooperative Oncology Group (ECOG).[3]
To determine whether these new regimens will improve outcomes requires randomized studies. In one ongoing ECOG trial, previously untreated patients with low-grade NHL are initially randomized to fludarabine plus cyclophosphamide or CVP. Responders then receive consolidation with rituximab or observation alone. M.D. Anderson investigators are comparing their FND regimen with combined or sequenced rituximab.
While these combination trials are accruing patients, an impressive series of interesting new agents are in earlier stages of clinical development. Two of the most exciting agents in clinical trials for NHL are the radioimmunoconjugates Bexxar[4] and IDEC-Y2B8.[5] Both are monoclonal antibodies directed against CD20; the former is conjugated to iodine-131, the latter to yttrium-90.
Cytotoxic drugs under evaluation include compound 506U78, a new purine analog[6]; flavopiridol, a cyclin inhibitor[7]; and bryostatin [8] and UCN-01, inhibitors of protein kinase C.
As more of these newer agents become available, how will the clinician determine which is the best option to offer patients? We have traditionally relied on the literature and based the treatment choice on the published response rates and durability of those responses.
However, differences in results among treatments may not reflect merely the activity of the regimen but, rather, differences in the characteristics of the patients accrued and the manner in which the study outcomes are measured. The relative numbers of relapsed vs refractory patients can also have a large impact on study results.
Differences in Response Criteria
In general, each major cooperative cancer treatment group with active clinical investigations in NHL has developed its own means of assessing and describing response rates and duration. Unfortunately, the differences among these criteria are sufficient to result in substantial confusion.
For example, to be characterized as a complete remission, a protocol generally requires that all lymph nodes involved with NHL return to normal size. However, the definition of “normal” varies among studies. Prior to treatment, a “normal” node is 1 cm in diameter. Nevertheless, following treatment, nodes rarely shrink below 1 cm, not necessarily because they are involved with tumor, but as the result of the presence of necrosis or fibrosis.
This problem is emphasized in a recent study by Grillo-López et al.[9] Using the database generated from the 166-patient rituximab pivotal trial, these investigators calculated what the complete remission rate would be if the bidimensional requirement for “normal” node size was considered to be 2 cm × 2 cm, 1.5 cm × 1.5 cm, or 1 cm × 1 cm.
Whereas the overall response rate (approximately 48%) did not change appreciably with the use of different criteria, the complete remission rate underwent a significant decrease from 28% to 18% to 6%, respectively, when the different criteria were used.
In addition, the frequency with which patients are evaluated for disease progression and when that measurement begins during the course of treatment can also result in artificial differences in time to progression, response duration, and other endpoints.
Uniform response criteria facilitate the interpretation of data and the ability to compare results among various studies. Therefore, standardization would help us identify new agents with promising activity. In addition, criteria provide a framework upon which to test scientific correlative studies of new biologic and immunologic insights into the diseases being studied.
The April 1999, issue of the Journal of Clinical Oncology contains the recommendations of an international workshop that developed standardized criteria for assessing response and other endpoints for clinical research trials in NHL.[10]
This project began several years ago as a collaborative effort by several investigators from the National Cancer Institute-sponsored cooperative groups, the NCI, and the pharmaceutical industry. A workshop was held at the National Cancer Institute on February 25-26, 1998, with a subsequent meeting on May 16, 1998, that included US, European, and Canadian lymphoma experts.
The product of these discussions was a set of guidelines for response assessment in adult patients with indolent and aggressive NHL (Table).
In phase II trials, particularly in the setting of relapsed and refractory disease, where the activity of a new agent may be the most important objective, response rates are meaningful and may provide support for approval by regulatory agencies. However, the primary goal of clinical NHL research is to identify therapies that will prolong the event-free survival, if not the overall survival, of our patients. Those endpoints should represent the goals of most phase III trials.
Response rates will continue to be difficult to interpret as long as the criteria are based on consensus opinion of clinical data rather than more precise and reliable measures of minimal residual disease. Over time, additional clinically relevant laboratory and imaging studies may be identified that may mandate modification of these recommendations. Until that time, the International Guidelines will be required for all NCI-sponsored clinical trials in NHL. We hope that they will improve communication among international investigators and comparisons among clinical trials so that more effective treatments can be identified.
