The Working Formulation tested the clinical utility of a consensus diagnosis based on morphology of patients with lymphoma. The variable to be tested was histology. Unfortunately, other factors affecting outcome were not kept constant such that the results are confounded by many variables including, most notably, stage and treatment.
Patients having stages I and II were included, thereby allowing a 200% variation in survival within some histologic subtypes (Table 1). Additionally, few patients received treatment of curative potential. These shortfalls of the Working Formulation suggest that a reappraisal of what constitutes an intermediate-grade histology is in order.
The question of whether any subset of follicular lymphoma is potentially curable has fueled controversy for at least 15 years.[4-6] The answer is elusive for at least four reasons:
First, criteria used to determine what constitutes a follicular architecture are variably applied (prominent or vague nodularity, complete or partial nodularity). Within the Southwest Oncology Group (SWOG), we classify a lymphoma as follicular regardless of the degree or extent of the follicular pattern. That is, any vestige of a follicular pattern constitutes a follicular lymphoma.
Second, the subtypes of follicular lymphoma cannot be reproducibly determined.
Third, follicular large-cell is an uncommon histology. In order to have adequate numbers of patients for analysis, many investigators have incorporated the complete spectrum of stages and treatments into studies, thereby confounding interpretation of outcome.[8-10]
Fourth, most reported series have a relatively short follow-up period (less than the known median survival [approximately 8 years] of follicular lymphoma), thereby creating the appearance of a tail or plateau in the survival curve.[9-10]
We have recently analyzed outcome of 53 patients with follicular large-cell lymphoma (expert review by multiple hematopathologists), having stages III and IV disease, treated with CHOP (cyclophosphamide [Cytoxan, Neosar], doxorubicin(Drug information on doxorubicin) [Adriamycin], vincristine [Oncovin], and prednisone(Drug information on prednisone)), and followed for up to 25 years. We compared the outcome of these patients with follicular large-cell to 336 follicular small-cleaved-cell and follicular mixed small- and large-cell patients similarly matched for stage and treatment. There is no evidence of a plateau in the survival curve, and nearly every patient eventually relapsed with lymphoma (Figure 1). Furthermore, the survival curves for the other histologic subtypes of follicular lymphoma are not significantly different. Thus, our study suggests that follicular large-cell lymphoma fails the test for intermediate-grade classification in the sense that it is not curable. The Nebraska Lymphoma Study Group has reported similar findings for all stages of follicular large-cell lymphoma.
This lymphoma is relatively uncommon in the United States and seems to be disappearing as a distinct entity. The original entity was a heterogeneous group of diseases, and at least two distinct clinical and biologic subtypes have been split off from the original category, namely, lymphoblastic lymphoma and mantle-cell lymphoma.[13,14] These histologic subtypes require special treatment considerations; neither do well with standard cyclic combination chemotherapy (see articles in this ONCOLOGY supplement by Magrath and Armitage . In addition, some small-cleaved-cell lymphomas have intact follicular dendritic cells detectable on immunohistochemical staining and indicative of persistent follicular center cell architecture that is not evident histologically. The remaining small number of patients seem to have an outcome similar to large-cell lymphoma.
This category of lymphoma is highly heterogeneous and is not easily separated from diffuse large-cell lymphoma. Ideally, it is a disease of B cells with admixed small and large malignant cells. On occasion, the admixed small cells are reactive and the condition is referred to as T-cell-rich B-cell lymphoma. Many peripheral T-cell lymphomas are included in this group, and these lymphomas have a different outcome than the B-cell counterpart.
There are no recent data concerning this histologic subtype to suggest that outcome, of at least the B-cell lymphomas (approximately 70%), differs significantly from that of diffuse large-cell lymphoma. Recent unpublished updates of the Intergroup Lymphoma Study that compared CHOP to other second- and third-generation regimens demonstrate that survival curves for diffuse mixed small- and large-cell lymphoma and for diffuse large-cell lymphoma are virtually superimposable. For that reason, the new Revised European-American Lymphoma (REAL) Classification does not distinguish diffuse mixed as a separate entity.
Diffuse large-cell lymphoma is by far and away the most common intermediate-grade histologic subtype. It is the prototype of potentially curable lymphomas demonstrating a relatively flat survival curve after 6 or 7 years of disease-free follow-up. The relevant question for clinicians is whether diffuse large-cell, immunoblastic subtype (Group H), constitutes a significantly different disease. In that regard there is some controversy, but most investigators believe that: 1) the two entities are difficult to distinguish reproducibly, and 2) survival in prospective studies is virtually identical.[1,17] For these reasons, the name diffuse large-cell lymphoma is frequently used to include Working Groups G and H.
In summary, intermediate-grade lymphomas, from a clinicians perspective, constitute a different subset of diseases than the Working Formulation originally defined. Our expectations are that the disease has aggressive growth characteristics and a potential for cure. In that regard, Working Groups E, F, G, and H probably best fit the definition.
All patients with intermediate-grade non-Hodgkins lymphoma are initially treated with CHOP chemotherapy. Patients with localized disease (stages I, IE, and non-bulky II and IIE) generally receive three cycles of CHOP followed by involved-field radiotherapy (CHOP-3 plus RT) based on a randomized trial of 401 patients treated with either CHOP-3 plus RT or eight cycles of CHOP (CHOP-8) alone.
CHOP-3 plus RT has been shown to result in better overall survival, better time to treatment failure, and less toxicity than CHOP-8. Estimates of overall survival at 5 years were 86% for patients with stages I and IE disease treated with CHOP-3 plus RT, and 74% for patients with stages II and IIE disease (non-bulky). Furthermore, there was less overall toxicity, especially left ventricular failure, for patients treated with CHOP-3 plus RT.
Patients with advanced disease (stages bulky-II, III, and IV) received eight cycles of CHOP based on a randomized Intergroup Study of 899 patients treated with either CHOP or one of three second- or third-generation regimens including m-BACOD (methotrexate [Rheumatrex], leucovorin, bleomycin [Blenoxane], doxorubicin, cyclophosphamide(Drug information on cyclophosphamide), vincristine, dexamethasone(Drug information on dexamethasone)), ProMACE-CytaBOM (prednisone, doxorubicin, cyclophosphamide, etoposide(Drug information on etoposide) [VePesid], cytarabine [Cytosar], bleomycin(Drug information on bleomycin), vincristine, mitoxantrone(Drug information on mitoxantrone), leucovorin), or MACOP-B (methotrexate, leucovorin, doxorubicin, cyclophosphamide, vincristine, bleomycin, prednisone). In that study, overall survival and time to treatment failure rates were similar for all treatment arms, but patients treated with CHOP had less life-threatening and fatal toxic events. Estimates of overall survival at 5 years for patients treated on any of the four arms were 49% for stage bulky-II and 46% for stages III and IV (combined).
These two studies define standard treatment for intermediate-grade non-Hodgkins lymphoma and establish a benchmark for future comparisons (Table 1). They don't, however, conclude the search for effective and safe therapy for patients with intermediate-grade non-Hodgkins lymphoma. At least 25% of patients with localized disease, and 50% of patients with advanced disease, will relapse and die of lymphoma within 5 years following treatment with CHOP-based regimens.
Following relapse patients with recurrent intermediate-grade non-Hodgkins lymphoma are treated with a variety of salvage regimens and simultaneously evaluated for high-dose chemotherapy and autologous transplant. Patients with drug-sensitive lymphoma who are free of any complicating medical problems generally receive high-dose chemotherapy and autotransplant based on a randomized trial by Philip et al. In that study, for the transplantation group, the event-free survival at 5 years was 46% and overall survival was 53%; for the standard salvage chemotherapy group, the event-free survival at 5 years was 12%, and overall survival was 32%.
Attempts to broaden the role of high-dose chemotherapy and autotransplant by treating patients with chemotherapy-resistant disease, or by treating patients in complete remission, have not led to convincing proof of benefit. In randomized trials of patients with chemotherapy-resistant disease, high-dose chemotherapy has not been shown to be better than conventional chemotherapy.[20-21] Likewise, Haioun et al, in a trial of patients in complete remission, have shown no difference in disease-free survival or overall survival among 541 patients randomized to receive consolidation by either sequential chemotherapy or autotransplant. In that study, a subgroup analysis of 236 randomized high-risk patients (defined as high-intermediate and high-risk according to the International Prognostic Index, see Table 2) revealed a significant difference in disease-free survival, but not in overall survival.
There are a large number of standard-dose chemotherapy regimens in use for treating recurrent intermediate-grade lymphoma, and there are little data to distinguish them from each other. Two platinum-based regimens developed by the M. D. Anderson group have gained wide acceptance and have a singular advantage as preinduction treatment prior to autotransplant by limiting exposure to anthracyclines.
The DHAP (dexamethasone, cytarabine(Drug information on cytarabine), and cisplatin(Drug information on cisplatin) [Platinol]) regimen is most frequently used, yielding a 44% to 59% overall response rate in recurrent intermediate- and high-grade lymphoma.[19,23] (The response rates to DHAP vary widely, 21% to 64%, as they do in all salvage studies, depending on whether patients had primary chemotherapy-resistance or relapsed after obtaining an initial response to a doxorubicin-containing regimen.) However, a recent update including long-term follow-up, compares DHAP to ESHAP (etoposide, solumedrol [methylprednisolone], cytarabine, and cisplatin), and demonstrates an advantage for patients treated with ESHAP in regard to time to treatment failure and overall survival.
There are many clinical features associated with survival duration of patients with intermediate-grade non-Hodgkins lymphoma, and many more publications attempting to determine the optimum combination of risk factors (see Moore and Cabanillas ).
Shipp et al have recently completed an analysis of the most commonly reported clinical prognostic factors by assessing outcome in 2,031 patients treated with potentially curative therapy. This International Prognostic Index has been largely successful by allowing a standardized application of clinical variables to trial design and interpretation of result reporting. Two examples of its use are detailed to demonstrate the potential for use in this group of intermediate-grade lymphomas.
First, a consistent and reproducible system of determining prognosis can be helpful in designing future trials by honing the patient group to be studied. The recently published Southwest Oncology Group study of localized lymphoma provides an example. Inasmuch as a large number of patients were cured with relatively well-tolerated treatment (CHOP-3 plus RT), it makes little sense to include those cured patients in future trials of more intensive or costly treatment designed for the subset of patients who failed treatment. In our planned study of localized lymphoma, we have applied the International Prognostic Index modified for localized disease by specifying non-bulky II as a poor risk factor (rather than stage III and IV). Analysis reveals that for patients without any risk factors (stage I, less than 60 years, normal LDH, good performance status, and less than two extranodal sites of disease) 5-year survival is 95%, whereas patients having any adverse risk factor(s) have a 5-year survival of 69%. By eliminating patients without any risk factors, we will spare about 30% of patients more intensive therapy, and we will target patients having a reasonable possibility of relapse with current standard therapy.
Second, a system using clinically available data allows comparisons between studies to help explain apparent inconsistencies and avoid unnecessary time and resource allocation to large randomized trials. The Intergroup Lymphoma Study comparing CHOP to second- and third-generation regimens was controversial from the time of conception, and the results remained controversial long after completion. Pilot data suggested that ProMACE-CytaBOM was superior to CHOP (the justification for the randomized comparison). The randomized comparison result showed no difference in outcome and was presumed by some investigators to reflect lack of experience with the regimen, and thus, an inadequate test of the regimen. Early application of the International Prognostic Index might have changed the design of that study. The International Prognostic Index Comparison of Risk Group Composition allows for comparison of risk group composition between different studies and may help explain observed differences in outcome. As Table 2 summarizes, there was a remarkable disparity in patient selection between institutions participating in the Intergroup Lymphoma Study and at the NCI.[15, 25] The proportion of patients in the lowest risk group (one adverse prognostic factor) was 44% at the NCI, twice that of the Intergroup Study. The cause of the disparate results between the pilot study and the randomized trial is no longer mysterious.
Approximately 12% of patients with intermediate-grade lymphoma will have a T-cell malignancy, and 88% will have a B-cell disease.[26,27] The importance of determining cell lineage was historically controversial, but the vast majority of recent reports, as summarized by Gascoyne, document a shortened disease-free survival, shortened survival, or both.  The issue is consequential, as getting the facts straight is the first step toward improving therapy.
Vose et al have changed therapy for patients with T-cell lineage by increasing the dose-intensity with autotransplants and report an early experience of possible improved outcome. Increasing the dose-intensity might be a correct solution if one extrapolates from work by Klimecki et al who showed that normal circulating blood cells displayed a quantifiable hierarchy with regard to drug-resistant protein expression. T-cell subsets consistently expressed more p-glycoprotein than B-cells, and therefore, might benefit from increased dose-intensity. Regardless, clinicians should exercise caution when treating a younger patient with a T-cell intermediate-grade lymphoma.
Clinicians look forward to the opportunity to treat patients with intermediate-grade non-Hodgkins lymphoma. Most patients respond to treatment and there is that opportunity to effect a cure. However, there is only one small subset of patients for whom standard therapy is acceptable. Those patients have stage I disease, are less than 60 years old, have a normal LDH, and are generally asymptomatic. We can offer such patients a 95% probability of cure, and thats not bad in medicine. However, for all other patients, the outcome afforded by standard therapy is not acceptable, and until more of these patients are enrolled in clinical trials, the longer we have to look forward to offering unacceptable treatment.