Developing robust and effective artificial intelligence (AI) models in medicine requires access to large amounts of patient data. The use of AI models solely trained on large multi-institutional datasets can help with this, yet the imperative to ensure data privacy remains, particularly as membership inference risks breaching patient confidentiality. As a proposed remedy, we advocate for the integration of differential privacy (DP). We specifically investigate the performance of models trained with DP as compared to models trained without DP on data from institutions that the model had not seen during its training (i.e., external validation) - the situation that is reflective of the clinical use of AI models. By leveraging more than 590,000 chest radiographs from five institutions, we evaluated the efficacy of DP-enhanced domain transfer (DP-DT) in diagnosing cardiomegaly, pleural effusion, pneumonia, atelectasis, and in identifying healthy subjects. We juxtaposed DP-DT with non-DP-DT and examined diagnostic accuracy and demographic fairness using the area under the receiver operating characteristic curve (AUC) as the main metric, as well as accuracy, sensitivity, and specificity. Our results show that DP-DT, even with exceptionally high privacy levels (epsilon around 1), performs comparably to non-DP-DT (P>0.119 across all domains). Furthermore, DP-DT led to marginal AUC differences - less than 1% - for nearly all subgroups, relative to non-DP-DT. Despite consistent evidence suggesting that DP models induce significant performance degradation for on-domain applications, we show that off-domain performance is almost not affected. Therefore, we ardently advocate for the adoption of DP in training diagnostic medical AI models, given its minimal impact on performance.

翻译：开发稳健高效的医学人工智能（AI）模型需要获取大量患者数据。仅使用大型多机构数据集训练的AI模型有助于解决这一问题，但确保数据隐私的迫切性依然存在，尤其是成员推断攻击可能泄露患者机密。作为应对方案，我们主张整合差分隐私（DP）。我们重点研究了使用DP训练的模型与未使用DP训练的模型在模型训练期间未接触过的机构数据（即外部验证）上的性能表现——这一场景反映了AI模型在临床实际应用中的情况。通过利用来自五个机构的超过59万张胸部X光片，我们评估了差分隐私增强型领域迁移（DP-DT）在诊断心脏肥大、胸腔积液、肺炎、肺不张以及识别健康受试者方面的有效性。我们将DP-DT与非DP-DT进行对比，以受试者工作特征曲线下面积（AUC）作为主要指标，同时辅以准确率、灵敏度和特异度，检验诊断准确性和人口公平性。结果表明，即使在隐私保护水平极高（epsilon约为1）的情况下，DP-DT的性能与非DP-DT相当（所有领域的P>0.119）。此外，与非DP-DT相比，DP-DT在几乎所有亚组中的AUC差异均小于1%。尽管现有证据一致表明DP模型在域内应用中会导致性能显著下降，但我们发现其域外性能几乎不受影响。因此，鉴于DP对性能的影响微乎其微，我们强烈主张在训练诊断性医疗AI模型时采用差分隐私。