Mixture Model to Predict the Cumulative Incidence of Relapses in Follicular Lymphoma : Need for Longer Follow-up or Alternative Outcomes
Bibliographic record
Abstract
Background: The effectiveness of treatments in follicular lymphoma is evaluated by the reduction in the rate of recurrence. One of the methodological difficulties is related to the presence of competitive events such as death or second primary malignancies. The related cumulative incidence functions (CIF) are always estimated by using the non-parametric Aalen-Johansen estimator and the associated predictors by using the Fine and Gray approach or by considering a cause-specific Cox model. In the present study, we investigated if the use of a parametric mixture model would provide additional information on the interpretation of the impact of treatments in the management of follicular lymphoma. Methods: We used the RELEVANCE study database (Morschhauser et al.NEJM 2018, JCO 2022) because of the prolonged follow-up of patients treated for follicular lymphoma. RELEVANCE study is a multicenter, phase 3 randomized clinical trial that evaluates rituximab lenalidomide (R 2) as compared to rituximab plus chemotherapy (R-chemo), in patients with previously untreated follicular lymphoma. All 1030 patients received a rituximab maintenance. Results: With a median follow up of 72 months, median progression-free survival (PFS) was not reached in both groups (Kaplan-Meier estimator): the 6-years PFS was 60% and 59% for R 2 and R-chemo, respectively (Cox model, HR = 1.03 [IC95% CI, 0.84 to 1.27]). The CIF of relapse at 72 months was 41% and 39% in the R 2 and R-chemo, respectively (Aalen-Johansen estimator, Figure 1). We did not highlight any significant difference between the treatments in terms of the CIF of relapse (Fine and Gray model, HR = 0.938 [IC95% CI, 0.754 to 1.17], for the same prognostic time, the CIF of death was 2,5% and 2,9% in the R 2 and R-chemo, respectively). We then performed the test with the mixture model, and confirmed there was no significant differences across groups at 72 months. Indeed, 64% and 96% of patients have not relapsed and died respectively. However the mixture model approach allows for extrapolation, and thus we predicted the long-term CIF of relapse to be 67% if patients were followed-up 20 years across groups. The mixture model extrapolation calculated that 60% of relapse would occurr between 72 months and 240 months (20 years). Conclusion: There was no difference at 72 months for survival across tests; but the mixture model, that allow extrapolation, suggested that a difference could be revealed with a greater follow-up. We concluded that future studies of survival in FL will require longer follow-up. This comment appears of particular importance in the context of the forthcoming immunotherapies such as bispecific and CAR-T cells. We can also propose for hematologic malignancies with a chronic evolution, FL, alternative primary outcomes not based on survival, allowing a shorter read out, such as the disability-adjusted life years (DALYs) or quality-adjusted life years (QALYs).
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 machine prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. The Gemma side is a direct model label for every work in the frame, read from the title-only record. The Codex side is a classifier learned from the 10,348 direct Codex labels and calibrated to design-weighted sample rates; fields without enough sample support carry no Codex call. Candidate is the union of the two sides; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels.
Distilled classifier scores by category (both heads)
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.012 | 0.024 |
| Meta-epidemiology (narrow) | 0.001 | 0.001 |
| Meta-epidemiology (broad) | 0.002 | 0.004 |
| Bibliometrics | 0.003 | 0.001 |
| Science and technology studies | 0.001 | 0.001 |
| Scholarly communication | 0.002 | 0.001 |
| Open science | 0.002 | 0.001 |
| Research integrity | 0.002 | 0.002 |
| Insufficient payload (model declined to judge) | 0.004 | 0.001 |
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 source (direct Gemma or distilled Codex), 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".