Evaluation of a Prognostic 23-Gene Expression Panel in a Set of 390 Multiethnic Hodgkin Lymphoma Patients
Bibliographic record
Abstract
Classic Hodgkin lymphoma (CHL) is a curable lymphoma with unique epidemiological features, including rare neoplastic cells, and an adolescent/young adult age (AYA) incidence peak higher in females, correlated with socioeconomic status. We previously reported that RNA expression measured in formalin-fixed paraffin-embedded (FFPE) tissues using a panel of 23 immune response genes was associated with 5-year overall survival (OS) in predominantly White/European advanced staged CHL patients treated with ABVD chemotherapy (Scott et al., 2013). Here, we conducted a study to evaluate if the 23-gene expression panel predicted 5-year OS in a real-world set of FFPE tumor tissue from racially/ethnically diverse populations in the U.S. and Mexico from the Multi- Ethnic Study of Hodgkin Lymphoma (MESH). Methods: We collected tumor blocks from 891 patients from 9 hospitals and 2 Residual Tissue Repositories in the U.S. and Mexico. Demographic and clinical data, including survival by month, age at diagnosis, birth year, race/ethnicity, sex, histology were abstracted. Patients were diagnosed from 1978 to 2018. CHL diagnoses and histological subtypes were validated by histopathology review. RNA was extracted and NanoString gene expression profiling was performed using 250-gene codeset including the published 23-gene expression (GE) panel (Scott et. al., 2013). 390 cases passed QC after applying a modified normalization protocol (Chan et.al., 2017). The multiethnic risk index was derived from a linear equation of the log 2 GE in MESH multiplied by the established regression coefficients (Scott et al., 2013). Next, we examined a dichotomized score using a cut-off value at the 75 th percentile. Both were tested for association with 5-year OS using a Cox proportional hazards model using R software 4.3.0 version adjusting for age as a continuous and categorical variable (15-39 and 40+ years), type of hospital that provided the samples, sex, EBV tumor status (assigned using EBER2), race/ethnicity, and histology. The 5-year OS was computed by right-censoring any patient that survived longer than 5 years. Proportionality assumptions were tested (using an alpha level of 0.05). Stratified models by categorical age and race/ethnicity were adjusted for patient demographical variables and effect heterogeneity was examined using a t-test. Result: The analytic set comprised 47% females and 44% Hispanic, 24% White, 20% Black, 9% Asian, and 3% Pacific Islander/Native Hawaiians. 62% were 15-39 years and 32% 40+ years at diagnosis. The multivariate analysis, excluding age, indicated an association between the risk index and 5-year OS (HR =1.41 (1.08, 1.93), p=0.01). However, including age attenuated the association (HR=1.18 (0.89, 1.57), p=0.25). When stratified by age categories, the continuous risk index was associated with 5-year OS only in the older age group: 15-39 years (HR = 0.89 (0.54, 1.47), p=0.65), and >40 years (HR = 1.47 (1.08, 2.01), p=0.02), indicating effect heterogeneity by age groups (p=0.09). We did not observe significant effect heterogeneity by EBV status (p=0.6), race/ethnicity (p=0.72), nor sex (p=0.33). When the dichotomized score was examined, the model was attenuated after adjusting for age, but stratification analysis showed an association in the older adults (HR= 2.8 (1.42, 5.51), p<0.01) with significant effect heterogeneity across age (p=0.02) (Figure 1). Evaluating the association between individual genes and OS, we found effect heterogeneity by age; two genes from the 23-gene panel indicated protective effects in the AYA but hazardous in older adults: B2M (p effect heterogeneity' =0.048) and APOL6 (p'=0.059). Individual genes with heterogeneity of association across race/ethnicity included RNF144B (p' =0.0021) with increased hazards in Whites compared to Blacks, whereas WDR83 (p'=0.018), PRF1 (p'=0.0018), LYZ (p'=0.0028) had higher hazards in Blacks compared to Whites, and RAPGEF2 (p'=0.036) had higher hazards in Hispanics compared to Whites. Conclusion: The risk index and dichotomized score derived from the previously developed 23-gene panel was significantly associated with OS among >40 year age group, with evidence of age effect heterogeneity. We detected 2 genes with an association that varied by age, and five across race/ethnicity, suggesting the importance of including diverse racial/ethnic and age groups in CHL patient populations when developing predictive models.
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 | 0.000 | 0.001 |
| 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) | 0.001 | 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 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".