ELCC | Poster | Real-World Biomarker Testing and Treatment Patterns in Belgian Lung Cancer Patients: Insights from the AIBED Study
At the European Lung Cancer Congress 2025, Dr. Lore Decoster (UZ Brussel) presented the poster: "432P - Real-World Biomarker Testing and Treatment Patterns in Belgian Lung Cancer Patients: Insights from the AIBED Study"
This first-of-its-kind Belgian study harnesses AI and NLP to provide real-world insights into lung cancer patient trajectories, bridging the gap between clinical trial results and routine clinical practice.
Background
Understanding real-world implementation of lung cancer treatments is critical as the landscape of routine clinical practice does not fully reflect clinical trials. This study describes demographics, biomarker testing, and treatment patterns in Belgian patients with lung cancer.
Methods
This multicenter secondary data study analyzed routinely collected electronic health record (EHR) data from five Belgian hospitals (2019–2021). Structured data were combined with Natural Language Processing (NLP)-processed unstructured data to build a standardized OMOP-CDM warehouse per hospital for federated analyses. Patients were grouped into stage IB-IIIB resected non-small cell lung cancer (NSCLC), unresected stage III NSCLC, stage IV NSCLC, and extensive-stage small cell lung cancer (ES-SCLC) cohorts.
Results
Data from 1952 patients could be captured for further analysis. Of these, 1699 (87.0%) had NSCLC and 253 (13.0%) had SCLC, of which 86.6% classified as ES-SCLC. Stage IB–IIIB resected NSCLC comprised 17.3%, unresected stage III NSCLC 12.3%, and stage IV NSCLC 45.0% of the total population. Brain metastases around time of diagnosis were reported in 16.4% of stage IV NSCLC (, 22.6% of EGFR-mutated stage IV NSCLC,) and 15.1% of ES-SCLC patients. In NSCLC cohorts, reported PDL1 positivity ranged from 45.6–56.1%, and EGFR mutation positivity from 4.8–11.0%. Among 294 stage IB–IIIB resected NSCLC patients, resection status was retrieved for 83.0% of patients, of which 93.0% with complete resection. Of 327 stage III NSCLC patients, 63.9% were unresected, 36.7% had PD-L1 positivity reported, 18.3% completed chemoradiotherapy, and 11.3% initiated durvalumab. In stage IV NSCLC (n = 764), 11.0% had EGFR mutations, and 4.4% received osimertinib. For ES-SCLC (n = 219), 80.9% received systemic therapy, and 45.7% received chemo-immunotherapy.
Conclusions
This first-of-its-kind Belgian study uses AI and NLP to deliver comprehensive real-world insights into lung cancer patient trajectories, identifying gaps in biomarker testing and treatment uptake to inform improvements in care delivery.
Authors: Lore Decoster (UZ Brussel), PaulGermonpré (AZ Maria Middelares), Daphne Verleyen (AstraZeneca Belux), ImkeMasuy (LynxCare), Bart Verheyden (AstraZeneca Belux), Clara L. Oeste(LynxCare), Emma Callewaert (AZ Groeninge), Sofie Derijcke (AZ Groeninge), ElsWauters (UZ Leuven), Annelies Janssens (UZA), Katoo M. Muylle (AstraZenecaBelux), Annelies T. Verbiest (UZA).
Available on ELCC 2025 - On-demand Poster Display - ELCC 2025 - Onsite and Online Congress
Background
Understanding real-world implementation of lung cancer treatments is critical as the landscape of routine clinical practice does not fully reflect clinical trials. This study describes demographics, biomarker testing, and treatment patterns in Belgian patients with lung cancer.
Methods
This multicenter secondary data study analyzed routinely collected electronic health record (EHR) data from five Belgian hospitals (2019–2021). Structured data were combined with Natural Language Processing (NLP)-processed unstructured data to build a standardized OMOP-CDM warehouse per hospital for federated analyses. Patients were grouped into stage IB-IIIB resected non-small cell lung cancer (NSCLC), unresected stage III NSCLC, stage IV NSCLC, and extensive-stage small cell lung cancer (ES-SCLC) cohorts.
Results
Data from 1952 patients could be captured for further analysis. Of these, 1699 (87.0%) had NSCLC and 253 (13.0%) had SCLC, of which 86.6% classified as ES-SCLC. Stage IB–IIIB resected NSCLC comprised 17.3%, unresected stage III NSCLC 12.3%, and stage IV NSCLC 45.0% of the total population. Brain metastases around time of diagnosis were reported in 16.4% of stage IV NSCLC (, 22.6% of EGFR-mutated stage IV NSCLC,) and 15.1% of ES-SCLC patients. In NSCLC cohorts, reported PDL1 positivity ranged from 45.6–56.1%, and EGFR mutation positivity from 4.8–11.0%. Among 294 stage IB–IIIB resected NSCLC patients, resection status was retrieved for 83.0% of patients, of which 93.0% with complete resection. Of 327 stage III NSCLC patients, 63.9% were unresected, 36.7% had PD-L1 positivity reported, 18.3% completed chemoradiotherapy, and 11.3% initiated durvalumab. In stage IV NSCLC (n = 764), 11.0% had EGFR mutations, and 4.4% received osimertinib. For ES-SCLC (n = 219), 80.9% received systemic therapy, and 45.7% received chemo-immunotherapy.
Conclusions
This first-of-its-kind Belgian study uses AI and NLP to deliver comprehensive real-world insights into lung cancer patient trajectories, identifying gaps in biomarker testing and treatment uptake to inform improvements in care delivery.
Authors: Lore Decoster (UZ Brussel), PaulGermonpré (AZ Maria Middelares), Daphne Verleyen (AstraZeneca Belux), ImkeMasuy (LynxCare), Bart Verheyden (AstraZeneca Belux), Clara L. Oeste(LynxCare), Emma Callewaert (AZ Groeninge), Sofie Derijcke (AZ Groeninge), ElsWauters (UZ Leuven), Annelies Janssens (UZA), Katoo M. Muylle (AstraZenecaBelux), Annelies T. Verbiest (UZA).
Available on ELCC 2025 - On-demand Poster Display - ELCC 2025 - Onsite and Online Congress
