Validation of a Prognostic Model to Assess the Risk of CNS Disease in Patients with Aggressive B-Cell Lymphoma
Bibliographic record
Abstract
Abstract Introduction: Despite the improvement of outcome of aggressive B-cell lymphomas in the rituximab treatment era, central nervous system (CNS) relapse continues to pose a significant management problem. It remains a challenge to select patients (pts) in whom specific diagnostic procedures to identify CNS disease at diagnosis should be performed and to target a high risk group in whom a CNS prophylaxis strategy is warranted. The German High-Grade Non-Hodgkin Lymphoma Study Group (DSHNHL) recently proposed a new prognostic model incorporating the 5 IPI factors (age > 60 y, LDH > N, stage 3 or 4, extranodal (EN) sites > 1) in addition to kidney/adrenal gland involvement to predict the risk of secondary CNS disease in pts with aggressive B-cell lymphoma. This model effectively stratified patients into 3 risk groups: low risk (0-1 factors, 2 y CNS relapse risk .6% (95% CI 0.0-1.2); intermediate risk (2-3 factors, 2 y CNS relapse risk 3.4 %( 95% CI 2.2-4.6)) and high risk (4-6 factors, 2 y CNS relapse risk of 10.2% (95% CI 6.3-14.1)) (Schmitz et al. Hematol. Oncol. 2013: 31, 047a). Herein, we sought to validate this model in an independent cohort of DLBCL treated with R-CHOP chemotherapy at the British Columbia Cancer Agency (BCCA). Methods: The DSHNHL dataset on which the model was initially developed was comprised of 2164 patients with aggressive B-cell lymphomas (n= 1735, 80.2% diffuse large B-cell lymphoma (DLBCL)), 18-80 years of age who were treated with rituximab with (CHO(E) P-like chemotherapy on prospective studies. The BCCA Lymphoid Cancer Database was screened to identify all patients with DLBCL treated with curative intent R-CHOP chemotherapy. Results: In total, 1597 patients were diagnosed with DLBCL at the BCCA and received at least one cycle of curative intent R-CHOP chemotherapy. The median follow-up for living patients was 4.2 years. Pts in the BCCA population-based DLBCL cohort were more likely to have poor risk features including PS >1, advanced age and a high IPI score (Table 1). Applying the 6 factor model, very similar risk groups were identified: low risk (0-1 factors 2 year CNS relapse risk .8% (95% CI 0.0-1.6%).; intermediate risk (2-3 factors 2 year CNS relapse risk 3.9% (95% CI 2.3-5.5%); and high risk (4-6 factors 2 y CNS relapse risk 12% (95% CI 7.9-16.1%) (Figure 1). The median time to CNS relapse was 6.7 months from the time of diagnosis in the BCCA group and was 7.2 months in the DSHNHL group highlighting that this event typically occurs early in the disease course. In both datasets kidney/adrenal involvement was highly associated with CNS relapse (2 year CNS risk BCCA 33%; 14% DSHNHL), the difference likely reflecting the higher risk pts in the BCCA population-based setting. Conclusions: We have validated the proposed DSHNHL prognostic model for CNS relapse in an independent dataset. The model identifies a high risk group in which diagnostic procedures to rule out CNS disease are highly recommended at diagnosis including MRI head, and cerebrospinal fluid analysis by cytology and flow cytometry and consideration of CNS-directed therapies. Kidney/adrenal involvement is consistently associated with a high risk of CNS relapse in the rituximab treatment era for which CNS prophylaxis should be incorporated into front-line therapy. Table 1 Clinical factor BCCA N=1597 DSHNHL N=2164 Age > 60 years * Median age 1035 (65%) 65 years (16-94) 974 (45%) 58 years (18-80) Median follow-up 4.6 years 2.9 years Male sex 915 (57%) 1244 (57.5%) PS > 1* 584 (37%) 247 (11%) Elevated LDH 1147 (53.0%) 737 (49.0%) EN > 1 396 (25%) 479 (22%) Stage 3 or 4 916 (57%) 1148 (53%) IPI * 0,1 2 3 4,5 463 (31%) 359 (24%) 350 (23%) 329 (22%) 1009 (47%) 523 (24%) 398 (18%) 231 (11%) Bulky disease > 7cm 636 (41%) 1027 (47.5%) * P< 0.05 Figure 1 Figure 1. Disclosures Savage: F Hoffmann-La Roche: Other. Villa:F Hoffmann-La Roche: Other. Sehn:Roche: Research Funding. Pfreundschuh:Roche: Advisory Board Other, Research Funding. Gascoyne:Hoffman La-Roche: Research Funding. Connors:Seattle Genetics, Inc.: Research Funding; Roche: Research Funding.
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
How this classification was reachedexpand
Full frame distilled prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. Learned from the 10,348 direct Codex labels and 10,348 direct Gemma labels. Candidate is the union of thresholded teacher heads; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels or direct frontier model labels.
Codex and Gemma teacher scores by category
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.000 | 0.000 |
| Meta-epidemiology (narrow) | 0.000 | 0.000 |
| Meta-epidemiology (broad) | 0.000 | 0.000 |
| Bibliometrics | 0.000 | 0.000 |
| Science and technology studies | 0.000 | 0.000 |
| Scholarly communication | 0.000 | 0.000 |
| Open science | 0.000 | 0.000 |
| Research integrity | 0.000 | 0.000 |
| Insufficient payload (model declined to judge) | 0.000 | 0.000 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from itClassification
machine, unvalidatedMachine predicted; a candidate call from one teacher head, not a consensus.
How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".