Achilonu Okechinyere, Obaido George, Ogbuokiri Blessing, Aruleba Kehinde, Musenge Eustasius, Fabian June
Division of Epidemiology and Biostatistics, School of Public Health, University of the Witwatersrand, Johannesburg, South Africa.
Center for Human-Compatible Artificial Intelligence (CHAI), Berkeley Institute for Data Science (BIDS), University of California, Berkeley, Berkeley, CA, United States.
Front Digit Health. 2024 Sep 3;6:1427845. doi: 10.3389/fdgth.2024.1427845. eCollection 2024.
In South Africa, between 1966 and 2014, there were three kidney transplant eras defined by evolving access to certain immunosuppressive therapies defined as (before availability of cyclosporine), (when cyclosporine became available), and (availability of tacrolimus and mycophenolic acid). As such, factors influencing kidney graft failure may vary across these eras. Therefore, evaluating the consistency and reproducibility of models developed to study these variations using machine learning (ML) algorithms could enhance our understanding of post-transplant graft survival dynamics across these three eras.
This study explored the effectiveness of nine ML algorithms in predicting 10-year graft survival across the three eras. We developed and internally validated these algorithms using data spanning the specified eras. The predictive performance of these algorithms was assessed using the area under the curve (AUC) of the receiver operating characteristics curve (ROC), supported by other evaluation metrics. We employed local interpretable model-agnostic explanations to provide detailed interpretations of individual model predictions and used permutation importance to assess global feature importance across each era.
Overall, the proportion of graft failure decreased from 41.5% in the era to 15.1% in the era. Our best-performing model across the three eras demonstrated high predictive accuracy. Notably, the ensemble models, particularly the Extra Trees model, emerged as standout performers, consistently achieving high AUC scores of 0.95, 0.95, and 0.97 across the eras. This indicates that the models achieved high consistency and reproducibility in predicting graft survival outcomes. Among the features evaluated, recipient age and donor age were the only features consistently influencing graft failure throughout these eras, while features such as glomerular filtration rate and recipient ethnicity showed high importance in specific eras, resulting in relatively poor historical transportability of the best model.
Our study emphasises the significance of analysing post-kidney transplant outcomes and identifying era-specific factors mitigating graft failure. The proposed framework can serve as a foundation for future research and assist physicians in identifying patients at risk of graft failure.
在南非，1966年至2014年间，根据获得某些免疫抑制疗法的情况，定义了三个肾脏移植时代，分别为（环孢素可用之前）、（环孢素可用时）和（他克莫司和霉酚酸可用时）。因此，影响肾移植失败的因素在这些时代可能有所不同。所以，评估使用机器学习（ML）算法开发的用于研究这些差异的模型的一致性和可重复性，可能会增进我们对这三个时代移植后移植物存活动态的理解。
本研究探讨了九种ML算法在预测三个时代10年移植物存活方面的有效性。我们使用跨越指定时代的数据开发并内部验证了这些算法。这些算法的预测性能通过接受者操作特征曲线（ROC）的曲线下面积（AUC）进行评估，并辅以其他评估指标。我们采用局部可解释模型无关解释来详细解释单个模型预测，并使用排列重要性来评估每个时代的全局特征重要性。
总体而言，移植物失败的比例从时代的41.5%降至时代的15.1%。我们在三个时代表现最佳的模型显示出较高的预测准确性。值得注意的是，集成模型，特别是极端随机树模型，表现突出，在各个时代始终获得0.95、0.95和0.97的高AUC分数。这表明模型在预测移植物存活结果方面具有高度的一致性和可重复性。在评估的特征中，受者年龄和供者年龄是在这些时代始终影响移植物失败的唯一特征，而肾小球滤过率和受者种族等特征在特定时代显示出高度重要性，导致最佳模型的历史可迁移性相对较差。
我们的研究强调了分析肾移植后结果并识别减轻移植物失败的特定时代因素的重要性。所提出的框架可为未来研究奠定基础，并协助医生识别有移植物失败风险的患者。
