Dataset Information

Predicting Long-term Survival After Allogeneic Hematopoietic Cell Transplantation in Patients With Hematologic Malignancies: Machine Learning-Based Model Development and Validation.

ABSTRACT:

Background

Scoring systems developed for predicting survival after allogeneic hematopoietic cell transplantation (HCT) show suboptimal prediction power, and various factors affect posttransplantation outcomes.

Objective

A prediction model using a machine learning-based algorithm can be an alternative for concurrently applying multiple variables and can reduce potential biases. In this regard, the aim of this study is to establish and validate a machine learning-based predictive model for survival after allogeneic HCT in patients with hematologic malignancies.

Methods

Data from 1470 patients with hematologic malignancies who underwent allogeneic HCT between December 1993 and June 2020 at Asan Medical Center, Seoul, South Korea, were retrospectively analyzed. Using the gradient boosting machine algorithm, we evaluated a model predicting the 5-year posttransplantation survival through 10-fold cross-validation.

Results

The prediction model showed good performance with a mean area under the receiver operating characteristic curve of 0.788 (SD 0.03). Furthermore, we developed a risk score predicting probabilities of posttransplantation survival in 294 randomly selected patients, and an agreement between the estimated predicted and observed risks of overall death, nonrelapse mortality, and relapse incidence was observed according to the risk score. Additionally, the calculated score demonstrated the possibility of predicting survival according to the different transplantation-related factors, with the visualization of the importance of each variable.

Conclusions

We developed a machine learning-based model for predicting long-term survival after allogeneic HCT in patients with hematologic malignancies. Our model provides a method for making decisions regarding patient and donor candidates or selecting transplantation-related resources, such as conditioning regimens.

SUBMITTER: Choi EJ

PROVIDER: S-EPMC8938832 | biostudies-literature | 2022 Mar

REPOSITORIES: biostudies-literature

ACCESS DATA

Publications

Predicting Long-term Survival After Allogeneic Hematopoietic Cell Transplantation in Patients With Hematologic Malignancies: Machine Learning-Based Model Development and Validation.

Choi Eun-Ji EJ Jun Tae Joon TJ Park Han-Seung HS Lee Jung-Hee JH Lee Kyoo-Hyung KH Kim Young-Hak YH Lee Young-Shin YS Kang Young-Ah YA Jeon Mijin M Kang Hyeran H Woo Jimin J Lee Je-Hwan JH

JMIR medical informatics 20220307 3

<h4>Background</h4>Scoring systems developed for predicting survival after allogeneic hematopoietic cell transplantation (HCT) show suboptimal prediction power, and various factors affect posttransplantation outcomes.<h4>Objective</h4>A prediction model using a machine learning-based algorithm can be an alternative for concurrently applying multiple variables and can reduce potential biases. In this regard, the aim of this study is to establish and validate a machine learning-based predictive mo ...[more]

PMID: 35254275

Similar Datasets

Project description:A minimally toxic nonmyeloablative regimen was developed for allogeneic hematopoietic cell transplantation (HCT) to treat patients with advanced hematologic malignancies who are older or have comorbid conditions.To describe outcomes of patients 60 years or older after receiving minimally toxic nonmyeloablative allogeneic HCT.From 1998 to 2008, 372 patients aged 60 to 75 years were enrolled in prospective clinical HCT trials at 18 collaborating institutions using conditioning with low-dose total body irradiation alone or combined with fludarabine, 90 mg/m(2), before related (n = 184) or unrelated (n = 188) donor transplants. Postgrafting immunosuppression included mycophenolate mofetil and a calcineurin inhibitor.Overall and progression-free survival were estimated by Kaplan-Meier method. Cumulative incidence estimates were calculated for acute and chronic graft-vs-host disease, toxicities, achievement of full donor chimerism, complete remission, relapse, and nonrelapse mortality. Hazard ratios (HRs) were estimated from Cox regression models.Overall, 5-year cumulative incidences of nonrelapse mortality and relapse were 27% (95% CI, 22%-32%) and 41% (95% CI, 36%-46%), respectively, leading to 5-year overall and progression-free survival of 35% (95% CI, 30%-40%) and 32% (95% CI, 27%-37%), respectively. These outcomes were not statistically significantly different when stratified by age groups. Furthermore, increasing age was not associated with increases in acute or chronic graft-vs-host disease or organ toxicities. In multivariate models, HCT-specific comorbidity index scores of 1 to 2 (HR, 1.58 [95% CI, 1.08-2.31]) and 3 or greater (HR, 1.97 [95% CI, 1.38-2.80]) were associated with worse survival compared with an HCT-specific comorbidity index score of 0 (P = .003 overall). Similarly, standard relapse risk (HR, 1.67 [95% CI, 1.10-2.54]) and high relapse risk (HR, 2.22 [95% CI, 1.43-3.43]) were associated with worse survival compared with low relapse risk (P < .001 overall).Among patients aged 60 to 75 years treated with nonmyeloablative allogeneic HCT, 5-year overall and progression-free survivals were 35% and 32%, respectively.

Project description:Allogeneic hematopoietic cell transplantation is a curative procedure for hematologic malignancies but is associated with a significant risk of non-relapse mortality (NRM). The Hematopoietic Cell Transplantation-Comorbidity Index (HCT-CI) is a prognostic tool that discriminates this risk in all age groups. A recent survey of transplant physicians demonstrated that 79% of pediatric providers used the HCT-CI infrequently, and most reported concerns about its applicability in the younger population. We conducted a retrospective study using the Center for International Blood and Marrow Transplant Research database to examine the impact of expanded HCT-CI definitions on NRM in pediatric and young adult patients with hematologic malignancies. We included 5790 patients <40 years old receiving allogeneic transplants between 2008 and 2017 to examine broader definitions of comorbidities in the HCT-CI, including history of mechanical ventilation and fungal infection, estimated glomerular filtration rate, and body mass index (BMI) percentiles. Multivariable Fine-Gray models were created to determine the effect of each HCT-CI defining comorbidity and its modification on NRM and were used to develop 2 novel risk scores. We next developed the expanded HCT-CI for children and young adults (youth with malignancies; expanded ymHCT-CI), where 23% patients had an increased comorbidity score, compared to the HCT-CI. Comorbidities with hazard ratio < 1.2 were then removed to create the simplified HCT-CI for children and young adults (youth with malignancies; simplified ymHCT-CI), which demonstrated higher scores corresponded to a greater risk of NRM (P < .001). These novel comorbidity indexes with broader definitions are more relevant to pediatric and young adult patients, and prospective studies are needed to validate these in the younger patient population. It remains to be seen whether the development of these pediatric-specific and practical risk indexes increases their use by the pediatric transplant community.

Dataset Information

Predicting Long-term Survival After Allogeneic Hematopoietic Cell Transplantation in Patients With Hematologic Malignancies: Machine Learning-Based Model Development and Validation.

Background

Objective

Methods

Results

Conclusions

Publications

Predicting Long-term Survival After Allogeneic Hematopoietic Cell Transplantation in Patients With Hematologic Malignancies: Machine Learning-Based Model Development and Validation.

Similar Datasets

OmicsDI is part of the ELIXIR infrastructure

Tweets