Treatment patterns, survival, and early mortality prediction in registry-coded adult malignant glioma: a population-based SEER study, 2000–2021

Population registries support large-scale studies of glioma outcomes, but historical histology codes and incomplete molecular data preclude direct alignment with modern WHO-defined diffuse glioma entities. We examined treatment patterns, survival, and early-mortality prediction among adults with registry-coded malignant glioma in a SEER export covering 2000–2021. We included adults with positive histologic confirmation, malignant behavior, ICD-O-3 9380/3, 9381/3, or 9382/3, and non-missing survival time. Zero-month survival records were retained and assigned 0.5 months for time-to-event analysis. Treatment was classified into eight patterns defined by combinations of surgery, radiotherapy, and chemotherapy. Analyses included Kaplan-Meier estimation, multivariable Cox models, inverse probability of treatment weighting among surgically treated patients, a 3-month landmark analysis, a 2018–2021 grade-adjusted analysis, and machine-learning models for 12-month mortality among patients with ascertainable 12-month status. The analytic cohort included 4,555 patients, including 217 with zero-month survival. Overall, 2,522 deaths occurred. Reverse Kaplan-Meier median follow-up was 127.0 months, whereas the raw median time to death or last contact was 37.0 months. The most frequent treatment patterns were surgery only (1,254 patients) and triple therapy (1,219 patients). With surgery only as the reference, adjusted HRs were 1.48 (95% CI 1.32–1.67) for triple therapy and 1.97 (95% CI 1.69–2.30) for radiotherapy plus chemotherapy; these estimates were interpreted as observational associations because of substantial case-mix differences. IPTW reduced measured covariate imbalance, although residual confounding remained likely. Treatment estimates differed in the 2018–2021 grade-adjusted subset, consistent with confounding by disease severity, calendar era, and subset composition. Gradient boosting yielded an AUROC of 0.839 (95% CI 0.814–0.860) and a Brier score of 0.134 for baseline-available 12-month mortality prediction. A further 324 patients censored alive before 12 months were excluded because their endpoint status was indeterminate. This SEER export supports observational analyses of registry-coded malignant glioma, but not causal treatment comparisons or inference for modern WHO-defined molecular subgroups. The findings define the interpretive limits of historical registry-coded glioma data and indicate that baseline registry variables can support internally validated early-mortality risk stratification.

Authors

Institutions

Publication Details

Journal
Cancer Cell International
Published
2026-11-14
DOI
https://doi.org/10.1186/s12935-026-04502-0
Primary Topic
Glioma Diagnosis and Treatment
Type
article
Field-Weighted Citation Impact
0.00
Controls
|||
ALL TIME
JAN
FEB
MAR
APR
MAY
JUN
JUL
AUG
SEP
OCT
NOV
article

Treatment patterns, survival, and early mortality prediction in registry-coded adult malignant glioma: a population-based SEER study, 2000–2021

Chuangnan Yan, Ziye Xu, Qiuling Tang, Hui Chen et al.
Cancer Cell International
Glioma Diagnosis and Treatment
article

Treatment patterns, survival, and early mortality prediction in registry-coded adult malignant glioma: a population-based SEER study, 2000–2021

Chuangnan Yan, Ziye Xu, Qiuling Tang, Hui Chen, Peihong Xu
article en

Abstract

Population registries support large-scale studies of glioma outcomes, but historical histology codes and incomplete molecular data preclude direct alignment with modern WHO-defined diffuse glioma entities. We examined treatment patterns, survival, and early-mortality prediction among adults with registry-coded malignant glioma in a SEER export covering 2000–2021. We included adults with positive histologic confirmation, malignant behavior, ICD-O-3 9380/3, 9381/3, or 9382/3, and non-missing survival time. Zero-month survival records were retained and assigned 0.5 months for time-to-event analysis. Treatment was classified into eight patterns defined by combinations of surgery, radiotherapy, and chemotherapy. Analyses included Kaplan-Meier estimation, multivariable Cox models, inverse probability of treatment weighting among surgically treated patients, a 3-month landmark analysis, a 2018–2021 grade-adjusted analysis, and machine-learning models for 12-month mortality among patients with ascertainable 12-month status. The analytic cohort included 4,555 patients, including 217 with zero-month survival. Overall, 2,522 deaths occurred. Reverse Kaplan-Meier median follow-up was 127.0 months, whereas the raw median time to death or last contact was 37.0 months. The most frequent treatment patterns were surgery only (1,254 patients) and triple therapy (1,219 patients). With surgery only as the reference, adjusted HRs were 1.48 (95% CI 1.32–1.67) for triple therapy and 1.97 (95% CI 1.69–2.30) for radiotherapy plus chemotherapy; these estimates were interpreted as observational associations because of substantial case-mix differences. IPTW reduced measured covariate imbalance, although residual confounding remained likely. Treatment estimates differed in the 2018–2021 grade-adjusted subset, consistent with confounding by disease severity, calendar era, and subset composition. Gradient boosting yielded an AUROC of 0.839 (95% CI 0.814–0.860) and a Brier score of 0.134 for baseline-available 12-month mortality prediction. A further 324 patients censored alive before 12 months were excluded because their endpoint status was indeterminate. This SEER export supports observational analyses of registry-coded malignant glioma, but not causal treatment comparisons or inference for modern WHO-defined molecular subgroups. The findings define the interpretive limits of historical registry-coded glioma data and indicate that baseline registry variables can support internally validated early-mortality risk stratification.

Cancer Cell International
Shantou University (CN), Cancer Hospital of Shantou University Medical College (CN), Second Affiliated Hospital of Shantou University Medical College (CN)
Glioma Diagnosis and Treatment
AI Navigator

Ask Laika to Summarize, Analyze, and Connect papers live on the map.

Summarize Papers & Methodologies

Extract key findings, datasets, and comparative methods across publications.

Benchmark Rankings & Visual Analytics

Rank top research institutions, authors, funders, topics, and journals by Field-Weighted Citation Impact (FWCI) and paper volume with instant charts.

Connect Distant Disciplines

Bridge topological clusters on the map to find hidden collaborative intersections.

Laika access coming soon