Bulky disease and/or very high LDH define a high-risk LBCL subgroup within IPI 1–2, but only in younger patients

Abstract

Abstract Introduction The International Prognostic Index (IPI) is a widely used tool for risk stratification of newly diagnosed large B-cell lymphoma (LBCL) patients. However, some patients experience refractory disease or relapse even within the low or low-intermediate risk groups (IPI 0–2). To improve risk discrimination, several modifications to the IPI have been proposed. Recent data suggest that patients within IPI 1–2 with bulky disease and/or very high lactate dehydrogenase (VH-LDH) levels have outcomes similar to those with IPI 3 (Maurer, ASH 2023), and could therefore be grouped with the intermediate-high and high-risk (H-Risk) IPI categories. This approach is already being incorporated into clinical trials (NCT06356129). However, external validation outside the US cohort is limited. We aimed to independently validate whether LBCL patients with IPI 1–2 and bulky disease and/or VH-LDH have survival outcomes comparable to those with IPI 3. Methods Consecutive patients with newly diagnosed systemic LBCL registered in the NiHiL project (NCT03199066) between 2010–2020 were identified. Inclusion criteria were clinical stage II–IV, age 18–80 years, IPI 1–5, complete baseline data, first-line R-CHOP-like therapy (n=2,303). Bulky disease was defined as ≥7.5 cm, and VH-LDH levels as >1.3× upper limit normal per prior report. IPI 1–2 H-Risk was defined by the presence of bulky and/or VH-LDH within the IPI 1–2 group. The primary endpoint was event-free survival (EFS) comparison between IPI 1–2 H-Risk versus IPI 1–2 low-risk (L-Risk) and IPI 3 groups. Hazard ratios (HRs) were estimated using Cox proportional hazards models. Age-stratified subgroup analysis was performed in patients aged ≤60 years and >60 years. Results Among 2,303 included patients, the median age was 67 years (IQR 59–72), with 71% being >60 years old, 48% were female, 78% had stage III–IV disease, and 34% had ECOG performance status 2–4. The IPI distribution was 36% IPI 1–2 (n=828), 29% IPI 3 (n=664), and 35% IPI 4–5 (n=811). Among IPI 1–2 patients, 440 (53%) were classified as L-Risk and 388 (47%) as H-Risk. The comparison of H-Risk versus L-Risk IPI 1-2 groups revealed the following differences: H-Risk patients were younger (median age 59 vs 66 years, P<0.01), had more frequently elevated LDH (70% vs 21%, P <0.01) and bulky disease (77% vs 0%, P <0.01, by definition), as well as a shorter diagnosis-to-treatment interval (median 24 vs 35 days, P <0.01). With a median follow-up of 8.1 years, 2-year EFS was 86% in the IPI 1–2 L-Risk group, 75% in IPI 1–2 H-Risk, and 65% in IPI 3; corresponding 5-year EFS rates were 73%, 69%, and 54%; with no significant differences between H-Risk vs L-Risk group (HR=1.08, 95% CI 0.87–1.35, P=0.47), while the EFS of IPI 3 patients was inferior when compared to IPI 1–2 H-Risk (HR=1.58, 95% CI 1.31–1.91, P<0.01). Due to the observed age imbalance between L-Risk and H-Risk groups, we performed an age-stratified analysis. Among patients aged ≤60 years (n=659), 361 had IPI 1–2, including 150 (42%) L-Risk and 211 (58%) H-Risk, while 203 (31%) had IPI 3. The 2-year EFS was 89% in L-Risk, 74% in H-Risk, and 65% in IPI 3; 5-year EFS was 81%, 69%, and 58%, respectively. Patients in the H-Risk group had significantly inferior EFS compared to L-Risk (HR=1.53, 95% CI 1.04–2.24, P=0.03), whereas the difference between IPI 1–2 H-Risk and IPI 3 was of borderline significance (HR=1.35, 95% CI 1.00–1.84, P=0.05). Among patients between 60 and 80 years (n=1,644), 467 (28%) had IPI 1–2, including 290 (62%) L-Risk and 177 (38%) H-Risk, and 461 (28%) had IPI 3. The 2-year EFS was 84% in L-Risk, 76% in H-Risk, and 64% in IPI 3; 5-year EFS was 70%, 68%, and 52%, respectively; with no difference between L-Risk and H-Risk groups (HR=1.00, 95% CI 0.75–1.38 P=0.98), while IPI 3 was associated with worse outcomes compared to H-Risk (HR=1.59, 95% CI 1.24–2.05, P<0.01). Conclusion The external validation in the NiHiL project demonstrated that IPI 1–2 H-Risk LBCL patients had significantly better EFS than those with IPI 3, while the EFS between IPI 1–2 H-Risk and L-Risk did not statistically differ. Thus, we were unable to confirm the findings reported in the US cohort (Maurer, ASH 2023), raising questions about the clinical utility of this stratification. However, in patients aged ≤60 years, the H-Risk group had significantly worse outcomes than L-Risk, supporting the potential prognostic relevance of this stratification in younger patients. Funding: NU21-03-00411