Predicting Prognostic Factors in Kidney Transplantation Using A Machine Learning

Predicting Prognostic Factors in Kidney Transplantation: A Machine Learning Approach to Enhance Outcome Prediction

Kidney transplantation (KT) is the most effective treatment for end-stage renal disease, offering improved quality of life and long-term survival. However, predicting transplant survival and assessing prognostic factors is complex due to the multifaceted nature of patient variables and individualized treatments. Traditional methods have fallen short in their predictive accuracy. This study aims to develop machine learning algorithms capable of parsing extensive clinical data to identify key prognostic indicators that can potentially forecast survival rates for KT recipients. By incorporating baseline characteristics of donors and recipients, the model strives to unearth patterns linking donor and recipient profiles, thereby offering insights into modifiable factors that could influence postoperative outcomes. The goal is to provide a tool that aids clinicians in improving the prognosis and quality of life for KT recipients.

Study Overview

• Status: Completed
• Conditions: Kidney Transplant Failure and Rejection
• Intervention / Treatment: Other: Prognostic factors affecting graft survival

Detailed Description

Kidney transplantation (KT) is the most effective treatment modality for end-stage renal disease (ESRD), offering patients the opportunity to ahieve improved quality of life and long-term survival. Advances in surgical techniques and immunosuppressive regimens have substantially decreased immediate postoperative complications and acute rejection episodes.

Considering that KT is the most frequently performed organ transplantation, improving the longevity of transplant survival could benefit many individuals. The efficacy of KT is often gauged by graft function, which is a critical determinant of the graft's long-term survival and a key metric in evaluating transplant success. While post-transplant graft function is influenced by a spectrum of variables-from the characteristics of donors and recipients to immunosuppressive strategies-this complexity presents challenges in forecasting outcomes, particularly over the long term. Traditional methods, such as the kidney donor risk index (KDRI) and Cox regression analyses, have fallen short in their predictive accuracy.

The prediction of transplant survival and the assessment of prognostic factors are complex due to the multifaceted nature of patient variables and the individualization of perioperative treatments. Yet, with the rise of machine learning and advanced computational analytics, researchers are now poised to decode the intricacies of data with clinical significance, potentially transforming patient care post-transplantation. The integration of deep learning algorithms into clinical practice in the field of transplantation is a relatively nascent area but is rapidly gaining traction.

This study aims to develop machine learning algorithms capable of parsing extensive clinical data to pinpoint key prognostic indicators which can potentially forecast survival rates for KT recipients. By incorporating baseline characteristics of both donors and recipients, the present model strives to unearth patterns linking donor and recipient profiles, thereby offering insights into modifiable factors that could influence postoperative outcomes. Through this, we seek to provide a tool that aids clinicians in improving the prognosis and quality of life for KT recipients.

• Study Type: Observational
• Enrollment (Actual): 4077

Contacts and Locations

Study Locations

• Korea, Republic of
  • Seoul, Korea, Republic of, 05505
    • Asan Medical Center

Participation Criteria

Eligibility Criteria

• Ages Eligible for Study: Child; Adult; Older Adult
• Accepts Healthy Volunteers: No

• Sampling Method: Non-Probability Sample

Study Population

4077 patients who underwent kidney transplantation at Asan Medical Center from June 1990 to May 2015

Description

Inclusion Criteria:

• Patients who have received kidney transplantation (including multiple times of transplantation) at this hospital.
• Patients who have listened to and understood a detailed explanation of this study, and have voluntarily decided to participate and provided written consent.

Exclusion Criteria:

• Patients who are receiving a multi-organ transplantation (e.g. simultaneous pancreas and kidney transplantation, simultaneous heart and kidney transplantation)

Study Plan

How is the study designed?

Number of groups / cohorts

1

Cohorts and Interventions

Group / Cohort	Intervention / Treatment
Kidney transplant patients; Patients who underwent kidney transplantation at a single center	Other: Prognostic factors affecting graft survival; The primary outcome measured was a 5-year graft survival, defined as the absence of any need for dialysis or re-transplantation five years following the initial transplantation

What is the study measuring?

Primary Outcome Measures

Outcome Measure	Measure Description	Time Frame
5-year graft survival	The primary outcome measured was a 5-year graft survival, defined as the absence of any need for dialysis or re-transplantation five years following the initial transplantation	5 years

Collaborators and Investigators

• Sponsor: Sung Shin
• Collaborators: Korea Health Industry Development Institute; Asan Institute for Life Sciences

Study record dates

Study Major Dates

• Study Start (Actual): January 1, 2023
• Primary Completion (Actual): January 1, 2024
• Study Completion (Actual): February 1, 2024

Study Registration Dates

• First Submitted: April 28, 2024
• First Submitted That Met QC Criteria: April 28, 2024
• First Posted (Actual): May 1, 2024

Study Record Updates

• Last Update Posted (Actual): May 1, 2024
• Last Update Submitted That Met QC Criteria: April 28, 2024
• Last Verified: April 1, 2024

More Information

Terms related to this study

• Keywords: prognosis; survival; kidney transplant; deep learning
• Other Study ID Numbers: 2022-1276

Plan for Individual participant data (IPD)

• Plan to Share Individual Participant Data (IPD)?: UNDECIDED
• IPD Plan Description: The authors will provide the raw data supporting the findings of this article upon request, without any unwarranted delays or restrictions.

Drug and device information, study documents

• Studies a U.S. FDA-regulated drug product: No
• Studies a U.S. FDA-regulated device product: No