RNfinity
Research Infinity Logo, Orange eye of horus, white eye of Ra
  • Home
  • Submit
    Research Articles
    Ebooks
  • Articles
    Academic
    Ebooks
  • Info
    Home
    Subject
    Submit
    About
    News
    Submission Guide
    Contact Us
  • Login/sign up
    Login
    Register

Biomedical

Assessment of reproducibility of cancer survival risk predictions across medical centers

rnfinity

info@rnfinity.com

orcid logo

Hung-Chia Chen,

Hung-Chia Chen


James J Chen

James J Chen


  Peer Reviewed

copyright icon

© attribution CC-BY

  • 0

rating
572 Views

Added on

2024-12-03

Doi: https://doi.org/10.1186/1471-2288-13-25

Related Subjects
Anatomy
Biochemistry
Epidemiology
Genetics
Neuroscience
Psychology
Oncology
Medicine
Musculoskeletal science
Pediatrics
Pathology
Pharmacology
Physiology
Psychiatry
Primary care
Women and reproductive health

Abstract

Background

Two key considerations in evaluating survival prediction models are predictability (the ability to accurately predict survival risks) and reproducibility (the ability to generalize predictions across studies or centers). This study explores approaches to assessing the reproducibility of survival risk predictions across medical centers using consistency and transferability metrics.

Methods

Reproducibility was assessed using two public datasets: lung cancer data from four centers and colon cancer data from two centers. Predictability was evaluated using eight survival prediction models, including clinical and gene expression data-based models. Metrics for reproducibility, such as consistency (agreement between centers) and transferability (applicability of models and signatures), were calculated using correlation coefficients.

Results

The analysis demonstrated good consistency and transferability for clinical models, particularly the Cox proportional hazards model. Models combining clinical and gene expression data offered limited improvements. The colon cancer dataset revealed significant variability in gene expression model performance, highlighting challenges in reproducibility across centers.

Conclusions

Models based on clinical variables demonstrated high reproducibility across centers, while gene expression models exhibited limited generalizability. Further studies are needed to enhance reproducibility and address variability in gene expression data.

Key Questions

1. What is the significance of reproducibility in survival prediction models?

Reproducibility ensures that survival prediction models can reliably generalize to data from other centers, enabling broader clinical application.

2. What approaches are used to assess reproducibility?

Consistency and transferability metrics, such as correlation coefficients, are used to evaluate agreement and applicability of predictions across centers.

3. How do clinical and gene expression models compare in reproducibility?

Clinical models, particularly the Cox proportional hazards model, showed better reproducibility than gene expression models, which faced challenges with variability.

4. What are the implications for future research?

Future studies should focus on improving gene expression model reproducibility and addressing variability across datasets to ensure effective clinical integration.

Summary Video Not Available

Review 0

Login

ARTICLE USAGE


Article usage: Dec-2024 to May-2025
Show by month Manuscript Video Summary
2025 May 140 140
2025 April 81 81
2025 March 68 68
2025 February 51 51
2025 January 167 167
2024 December 65 65
Total 572 572
Show by month Manuscript Video Summary
2025 May 140 140
2025 April 81 81
2025 March 68 68
2025 February 51 51
2025 January 167 167
2024 December 65 65
Total 572 572
Related Subjects
Anatomy
Biochemistry
Epidemiology
Genetics
Neuroscience
Psychology
Oncology
Medicine
Musculoskeletal science
Pediatrics
Pathology
Pharmacology
Physiology
Psychiatry
Primary care
Women and reproductive health
copyright icon

© attribution CC-BY

  • 0

rating
572 Views

Added on

2024-12-03

Doi: https://doi.org/10.1186/1471-2288-13-25

Related Subjects
Anatomy
Biochemistry
Epidemiology
Genetics
Neuroscience
Psychology
Oncology
Medicine
Musculoskeletal science
Pediatrics
Pathology
Pharmacology
Physiology
Psychiatry
Primary care
Women and reproductive health

Follow Us

  • Xicon
  • Contact Us
  • Privacy Policy
  • Terms and Conditions

5 Braemore Court, London EN4 0AE, Telephone +442082758777

© Copyright 2025 All Rights Reserved.