Are Real-World Survival Outcomes in Metastatic Breast Cancer Transportable Between the US and Austria?
Author(s)
Harlan Pittell, PhD1, Uwe Siebert, MD, MPH, MSc, ScD2, Xiaolong Jiao, MS, MD3, Stella Mokiou, BSc, MSc, PhD4, Per-Olof Thuresson, MSc5, Lu Chen, PhD6, Carol Hawkes, MSc, PhD7, Danalyn Byng, MSc, PhD8, Gabriel Rinnerthaler, MD9, Simon Peter Gampenrieder, MD9, Richard Greil, MD9, Mohamed S. Ali, PharmD, SM10, Philani Mpofu, PhD10, Elsie Horne, PhD11, Qianyi Zhang, MS12, Amit Samani, PhD, MD13, Blythe Adamson, MPH, PhD10.
1Senior Research Scientist, Flatiron Health, New York, NY, USA, 2UMIT TIROL – University for Health Sciences and Technology, Hall in Tirol, Austria, TIROL, Austria, 3Pfizer, Wynnewood, PA, USA, 4Pfizer, REDHILL, United Kingdom, 5Roche Products Ltd, Basel, Switzerland, 6Genentech, San Francisco, CA, USA, 7Daiichi Sankyo Europe GmbH, Slough, United Kingdom, 8Daiichi Sankyo Europe GmbH, Munich, Germany, 9Austrian Group Medical Tumor Therapy (AGMT), Salzburg, Austria, 10Flatiron Health, New York, NY, USA, 11Flatiron Health UK, London, United Kingdom, 12Flatiron Health Inc, Jersey City, NJ, USA, 13Flatiron Health, London, United Kingdom.
1Senior Research Scientist, Flatiron Health, New York, NY, USA, 2UMIT TIROL – University for Health Sciences and Technology, Hall in Tirol, Austria, TIROL, Austria, 3Pfizer, Wynnewood, PA, USA, 4Pfizer, REDHILL, United Kingdom, 5Roche Products Ltd, Basel, Switzerland, 6Genentech, San Francisco, CA, USA, 7Daiichi Sankyo Europe GmbH, Slough, United Kingdom, 8Daiichi Sankyo Europe GmbH, Munich, Germany, 9Austrian Group Medical Tumor Therapy (AGMT), Salzburg, Austria, 10Flatiron Health, New York, NY, USA, 11Flatiron Health UK, London, United Kingdom, 12Flatiron Health Inc, Jersey City, NJ, USA, 13Flatiron Health, London, United Kingdom.
OBJECTIVES: Real-world evidence (RWE) can inform global treatment decisions, but its transportability across settings is uncertain. This study assessed whether survival outcomes in metastatic breast cancer (mBC) are transportable between US and Austrian real-world datasets.
METHODS: We analyzed patient-level data from the US Flatiron Health Research Database and Austria AGMT registry; the latter may reflect a more selective population due to its consent-based enrollment. Both cohorts included adults diagnosed with mBC in 2015-2024. Baseline characteristics were compared. A pooled logistic regression model, fit to the US cohort and standardized for shared clinical factors, was applied to Austria to predict monthly overall survival (OS) from treatment start. Predicted survival curves were compared to observed (Kaplan-Meier) survival curves for the overall Austria cohort and by tumour subtype.
RESULTS: The analysis included 1,292 Austrian and 21,215 US patients, with similar median ages (65 vs 64 years). Obesity (38% vs 19%) and recurrent mBC (67% vs 60%) were more common in the US. The distribution of HR/HER2 subtypes was similar between the US and Austrian cohorts. Median unadjusted OS was longer in Austria than the US (42.4 vs 33.3 months). The US-trained model underestimated Austrian OS (predicted vs observed, 31.0 vs 42.4 months). By subtype, predicted versus observed median OS in the Austria cohort was 42.2 vs 58.5 months for HR+/HER2+ (n: US=3458, Austria=191), 35.7 vs 45.3 months for HR+/HER2− (n: US=14,799, Austria=898), 37.2 vs 56.0 months for HR−/HER2+ (n: US=976, Austria=61), and 10.6 vs 18.4 months for HR−/HER2− (n: US=1982, Austria=142).
CONCLUSIONS: Differences in survival estimates between these mBC datasets highlight uncertainty around the transportability of RWE across settings. However, the generalizability of the AGMT registry to the broader Austrian mBC population is unclear. Further research is needed to explore how dataset composition and healthcare system differences may contribute to observed survival variation.
METHODS: We analyzed patient-level data from the US Flatiron Health Research Database and Austria AGMT registry; the latter may reflect a more selective population due to its consent-based enrollment. Both cohorts included adults diagnosed with mBC in 2015-2024. Baseline characteristics were compared. A pooled logistic regression model, fit to the US cohort and standardized for shared clinical factors, was applied to Austria to predict monthly overall survival (OS) from treatment start. Predicted survival curves were compared to observed (Kaplan-Meier) survival curves for the overall Austria cohort and by tumour subtype.
RESULTS: The analysis included 1,292 Austrian and 21,215 US patients, with similar median ages (65 vs 64 years). Obesity (38% vs 19%) and recurrent mBC (67% vs 60%) were more common in the US. The distribution of HR/HER2 subtypes was similar between the US and Austrian cohorts. Median unadjusted OS was longer in Austria than the US (42.4 vs 33.3 months). The US-trained model underestimated Austrian OS (predicted vs observed, 31.0 vs 42.4 months). By subtype, predicted versus observed median OS in the Austria cohort was 42.2 vs 58.5 months for HR+/HER2+ (n: US=3458, Austria=191), 35.7 vs 45.3 months for HR+/HER2− (n: US=14,799, Austria=898), 37.2 vs 56.0 months for HR−/HER2+ (n: US=976, Austria=61), and 10.6 vs 18.4 months for HR−/HER2− (n: US=1982, Austria=142).
CONCLUSIONS: Differences in survival estimates between these mBC datasets highlight uncertainty around the transportability of RWE across settings. However, the generalizability of the AGMT registry to the broader Austrian mBC population is unclear. Further research is needed to explore how dataset composition and healthcare system differences may contribute to observed survival variation.
Conference/Value in Health Info
2025-11, ISPOR Europe 2025, Glasgow, Scotland
Value in Health, Volume 28, Issue S2
Code
HTA38
Topic
Health Technology Assessment
Topic Subcategory
Decision & Deliberative Processes
Disease
No Additional Disease & Conditions/Specialized Treatment Areas, Oncology