Prognostic factors for adult patients with Burkitt lymphoma treated with dose-adjusted EPOCH-R

TO THE EDITOR:

Burkitt lymphoma (BL) is a highly aggressive B-cell lymphoma characterized by high tumor proliferation and frequent involvement of the bone marrow (BM) and/or central nervous system (CNS).¹ It commonly affects children and adolescents, and >90% of pediatric patients are cured with highly dose-intensive chemotherapy, including patients with high-risk features such as leukemic disease or CNS involvement.^2-5 Highly dose-intensive chemotherapy is also effective in adults, but the clinical outcomes are worse overall, in part, because of poor treatment tolerance and/or potentially worse tumor biology than that in younger patients.^6,7 The pharmacodynamically adjusted regimen dose-adjusted etoposide, prednisone, vincristine, cyclophosphamide, doxorubicin, and rituximab (DA-EPOCH-R) was developed to overcome the adverse effects of high tumor proliferation through continuous low-dose administration.^8,9 Prospective studies have confirmed high efficacy and tolerance in adults with BL.^10,11 Recent genomic studies identified distinct genetic and epigenetic BL subtypes associated with disparate clinical outcomes, and both pediatric and adult BL share common pathobiology.¹² Thus, it remains unclear whether the prognostic effect of age is related to poor treatment tolerance, high-risk biology, or their combination.¹³ 

Disease-related factors including CNS involvement and tumor burden are also associated with poor outcomes.^13,14 Recently, a multicenter retrospective study reported clinical results from 633 adult patients who were treated with 3 common BL regimens.¹⁵ From this data set, the BL International Prognostic Index (BL-IPI) was developed, which stratified patients into risk categories based on 4 factors: age ≥ 40 years, an Eastern Cooperative Oncology Group (ECOG) performance status ≥2, serum lactate dehydrogenase (LDH) levels > 3 times the upper limit of normal (ULN), and CNS involvement.¹⁶ Patients with low-risk (no risk factor), intermediate-risk (1 risk factor), and high-risk (≥2 risk factors) disease had differences in 3-year progression-free survival of 92%, 72%, and 53%, respectively, in the derivation cohort and 96%, 82%, and 63%, respectively, in the validation cohort.

Herein, we analyze the prognostic utility of the BL-IPI in adult patients with untreated BL in a multicenter prospective trial (NCI 9177) of risk-adapted DA-EPOCH-R, in which treatment tolerance was not a major issue (the trial was clinically registered at www.clinicaltrials.gov as #NCT01092182).

NCI 9177 was a multicenter study of DA-EPOCH-R for patients with untreated BL aged ≥18 years with any HIV status.¹¹ Treatment was risk adapted based on baseline prognostic factors independent of age. Patients with low-risk disease received 3 cycles without CNS prophylaxis. Patients with high-risk disease received 6 cycles along with either prophylactic intrathecal therapy or an intensified intrathecal therapy schedule for active CNS disease.

Event-free survival (EFS) was calculated from study entry until progression, documented active disease, death, or last follow-up. Overall survival was calculated from study entry to death from any cause or last follow-up. Kaplan-Meier estimation and log-rank tests were used to determine the prognostic utility of the BL-IPI components.

The study was approved by local institutional review boards of participating institutions, and all patients signed informed consent forms.

In total, 113 patients were enrolled, including 15 (13%) considered at low-risk based on protocol-specified criteria. Median patient age was 49 years (range, 18-86 years) and the BM and/or cerebrospinal fluid was involved in 29 (26%) patients. At a median follow-up of 58.7 months, the 4-year EFS and overall survival for all patients were 84.5% and 87.0%, respectively.¹¹ Applying the BL-IPI model, 31 (27%) patients were at low risk, 55 (49%) were at intermediate risk, and 27 (24%) were at high risk (Figure 1A). Based on BL-IPI, no difference was noted in the 5-year EFS between patients with low-risk and those with intermediate-risk disease: 83.6% (95% confidence interval [CI], 65-93) vs 94.2% (95% CI, 83-98; P = .13), but high-risk BL-IPI predicted a worse 5-year EFS of 66.7% (95% CI, 46-81) compared with 90.3% (95% CI, 82-93; P = .003) in combined groups of patients with low/intermediate-risk disease (Figure 1B).

Individual components of the BL-IPI were analyzed for prognostic impact by univariate analysis (Figure 1C). Notably, the 5-year EFS was similar for patients aged ≥40 years (n = 70) vs <40 years (n = 43) at 86.7% (95% CI, 76-93) vs 81.1% (95% CI, 66-90; hazard ratio [HR], 0.7; P = .50). Interestingly, of the 15 patients who were considered at low risk based on protocol-specified criteria and received only 3 cycles of chemotherapy; 12 (80%) would have been considered to be at intermediate risk based on BL-IPI because of age >40 years. The 5-year EFS for these patients was 100%. Patients with serum LDH levels ≥ 3 × ULN (n = 26) had an inferior 5-year EFS compared with those with serum LDH levels < 3 × ULN (n = 87): 65.4% (95% CI, 44-80) vs 90.4% (95% CI, 82-95; HR, 4.4; P = .001). Similarly, patients with an ECOG performance status ≥ 2 (n = 21) had an inferior 5-year EFS compared with those with an ECOG performance status of 0 or 1 (n = 92): 61.9% (95% CI, 38-79) vs 89.8% (95% CI, 81-95; HR, 4.8; P = .0004), and patients with CNS involvement (n = 11) had a markedly inferior 5-year EFS compared with those without CNS involvement (n = 102): 45.5% (95% CI, 17-71) compared with 88.8% (95% CI, 81-94; HR, 7.0; P = .0001).

We previously reported that patients in NCI 9177 with BM, peripheral blood, and/or CNS (BM/CNS) involvement had inferior outcomes: the 4-year EFS for patients with no involvement of the cerebrospinal fluid or BM (n = 69) vs that for those with involvement of either (n = 29) was 92.4% (95% CI, 83-97) vs 58.6% (95% CI, 39-74; P = 001), respectively.¹¹ Herein, we analyzed these risk factors within the BL-IPI risk groups. Twenty-nine (26%) patients had BM/CNS involvement, including 19 (70%) who were at high risk and 10 (30%) who were at low or intermediate risk based on BL-IPI (Figure 2A). Specifically, BM or peripheral blood involvement was observed in 3 (10%), 7 (13%), and 16 (59%) patients in the low-, intermediate-, and high-risk BL-IPI groups, respectively. All 11 patients with CNS involvement were categorized as being at high risk by the BL-IPI. Patients with BM/CNS involvement who were at high risk based on the BL-IPI had a markedly worse 5-year EFS than those without involvement: 52.6% (95% CI, 29-72) compared with 100% (P = .03; Figure 2B). Notably, among low-/intermediate-risk disease, BM/CNS involvement was an adverse prognostic factor, with a 5-year EFS of 70.0% (95% CI, 33-89) compared with 93.0% (95% CI, 84-97; P = .03) in patients without BM/CNS involved.

Dose-adjusted EPOCH-R is included in National Comprehensive Cancer Network guidelines for untreated BL and has become a preferred regimen at many institutions. We have demonstrated that the BL-IPI is prognostic in a prospective clinical trial of DA-EPOCH-R and could be used for risk stratification in future studies. Patients with low- or intermediate-risk disease based on the BL-IPI had an excellent 5-year EFS of 90.3% when treated with DA-EPOCH-R, whereas those with high-risk disease had a 5-year EFS of only 66.7%. Notably, patients at high risk per the BL-IPI comprised only 24% of our cohort, compared with nearly 50% of the patients included in the original BL-IPI analysis. These data suggest that the patient population in this clinical trial does not represent the full spectrum of the disease in adults. Although the BL-IPI model was prognostic overall as well as for the individual factors of LDH, ECOG PS of ≥2, and CNS involvement, the lack of prognostic value of age ≥40 with the DA-EPOCH-R regimen suggests that when treatment tolerance is moderated, age is not associated with high-risk disease biology. In this way, age may only carry prognostic information for patients who are treated with other highly dose-intensive regimens but not DA-EPOCH-R.

Our data also suggest that the other components of the BL-IPI such as LDH level, ECOG performance status, and, especially, CNS involvement are highly prognostic with the use of DA-EPOCH-R. Indeed, with the composite risk factor of BM and/or CNS involvement, we were able to further risk stratify patients within BL-IPI categories. These data are important when considering which adult patients with BL should be prioritized for novel treatment approaches. In NCI 9177, patients without BM/CNS involvement made up 70% of the protocol-specified high-risk cohort and had a 5-year EFS of 92%. It will be difficult to improve upon these outcomes. Conversely, patients with BM/CNS involvement had a 5-year EFS of only 59%, with many of the events related to early toxic death often on the first cycle despite the relatively good treatment tolerance of DA-EPOCH-R. Taken together, these data suggest that further intensification of standard chemotherapy regimens in those with BM/CNS involvement may not improve outcomes because it will also further increase the risk of treatment-related toxicity, particularly in older patients with an impaired performance status. In summary, adult patients with BL that should be considered high-risk for treatment failure should be extended beyond the BL-IPI and include all patients with CNS involvement. These patients should be prioritized for novel treatment approaches including PI3K pathway inhibitors, chimeric antigen receptor T-cell therapy, and other forms of immunotherapy.

Acknowledgments: The Cancer Therapy Evaluation Program and Lymphoid Malignancies Branch, National Cancer Institute, National Institutes of Health, provided funding for this study. Research reported in this publication was supported by the National Cancer Institute of the National Institutes of Health under award numbers U10CA180888 and UG1CA233230, as well as the AIDS Malignancy Consortium grant UMI CA121947. The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health.

Contribution: M.R. contributed in conception and design of the study; M.R., R.L., and S.M.S. analyzed and assembled the data; R.L. wrote the first draft that was edited by M.R.; all authors collected and assembled data, reviewed the analysis, and provided comments, approval, and input on the final version; and M.R., primarily, along with all authors are accountable for all the aspects of work.

Conflict-of-interest disclosure: K.D. reports research support from Genentech. J.S.A. is a consultant for AbbVie, AstraZeneca, BeiGene, bluebird bio, Bristol Myers Squibb (BMS), Celgene, Epizyme, Incyte, Kymera, Genmab, Genentech, Ono Pharma, Mustang Bio, MorphoSys, Regeneron, Century, Kite Pharma, Lilly, Janssen, Takeda, Caribou Biosciences, Inrerius, and Cellectar. B.K.L. reports support from MEI Pharma Inc. B.S.K. recieved research support from AbbVie, Genentech, Celgene, AstraZeneca, and ADCT; and is a consultant for AbbVie, BMS, BeiGene, ADCT, Genentech, MEI, Janssen, Kite, and TG therapeutics. A.N. receives research support from Pharmacyclics/AbbVie and Cornerstone, and is a consultant for Janssen, Pharmacyclics/AbbVie, Epizyme, ADC, Cornerstone, and Kite/Gilead. The remaining authors declare no competing financial interests.

Correspondence: Mark Roschewski, Lymphoid Malignancies Branch, CCR, NCI, NIH, Bldg 10, Room 12c442, 10 Center Dr, Bethesda, MD 20892; e-mail: mark.roschewski@nih.gov.