Evaluation of optimal cut-offs and dichotomous combinations for two biomarkers to improve patient selection
Fuente:
arXiv
Salvato in:
| Autori principali: | , , |
|---|---|
| Natura: | Preprint |
| Pubblicazione: |
2025
|
| Soggetti: | |
| Accesso online: | |
| Tags: |
Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
|
| _version_ | 1866909573513740288 |
|---|---|
| author | DAngelo, Gina Ran, Di Yu, Binbing |
| author_facet | DAngelo, Gina Ran, Di Yu, Binbing |
| contents | Background Identifying the right cut-off for continuous biomarkers in clinical trials is important to identify subgroups of patients who are at greater risk of disease or more likely to benefit from a drug. The literature in this area tends to focus on finding cut-offs for a single biomarker, whereas clinical trials more often focus on multiple biomarkers. Methods Our first objective was to compare three methods,Youden index, point closest to the (0,1) corner on the receiving operator characteristic curve (ER), and concordance probability, to find the optimal cut-offs for two biomarkers, using empirical and non-empirical approaches. Our second and main objective was to use our proposed logic indicator approach to extend the Youden index and evaluate whether a combination of biomarkers is an improvement over a single biomarker. Results The logic indicator approach created a condition in which either both biomarkers were positive or only one of the biomarkers was positive. A prostate cancer study and a simulated phase 2 lung cancer study were used to illustrate approaches to finding optimal cut-offs and comparing combined biomarkers with single biomarkers. Conclusion Our results can aid in determining whether a single biomarker or a combination of biomarkers is superior in identifying patients who are more likely to respond to treatment. This work can be of great importance in the era of personalized medicine, where many treatments do not provide clinical benefit to average patients. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2504_07159 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | Evaluation of optimal cut-offs and dichotomous combinations for two biomarkers to improve patient selection DAngelo, Gina Ran, Di Yu, Binbing Quantitative Methods Background Identifying the right cut-off for continuous biomarkers in clinical trials is important to identify subgroups of patients who are at greater risk of disease or more likely to benefit from a drug. The literature in this area tends to focus on finding cut-offs for a single biomarker, whereas clinical trials more often focus on multiple biomarkers. Methods Our first objective was to compare three methods,Youden index, point closest to the (0,1) corner on the receiving operator characteristic curve (ER), and concordance probability, to find the optimal cut-offs for two biomarkers, using empirical and non-empirical approaches. Our second and main objective was to use our proposed logic indicator approach to extend the Youden index and evaluate whether a combination of biomarkers is an improvement over a single biomarker. Results The logic indicator approach created a condition in which either both biomarkers were positive or only one of the biomarkers was positive. A prostate cancer study and a simulated phase 2 lung cancer study were used to illustrate approaches to finding optimal cut-offs and comparing combined biomarkers with single biomarkers. Conclusion Our results can aid in determining whether a single biomarker or a combination of biomarkers is superior in identifying patients who are more likely to respond to treatment. This work can be of great importance in the era of personalized medicine, where many treatments do not provide clinical benefit to average patients. |
| title | Evaluation of optimal cut-offs and dichotomous combinations for two biomarkers to improve patient selection |
| topic | Quantitative Methods |
| url | https://arxiv.org/abs/2504.07159 |