Echefu Gift, Shah Rushabh, Sanchez Zanele, Rickards John, Brown Sherry-Ann
Division of Cardiovascular Medicine, University of Tennessee, Memphis, TN, USA.
Medical College of Wisconsin, Milwaukee, WI, USA.
Am Heart J Plus. 2024 Oct 31;48:100479. doi: 10.1016/j.ahjo.2024.100479. eCollection 2024 Dec.
Numerous cancer therapies have detrimental cardiovascular effects on cancer survivors. Cardiovascular toxicity can span the course of cancer treatment and is influenced by several factors. To mitigate these risks, cardio-oncology has evolved, with an emphasis on prevention and treatment of cardiovascular complications resulting from the presence of cancer and cancer therapy. Artificial intelligence (AI) holds multifaceted potential to enhance cardio-oncologic outcomes. AI algorithms are currently utilizing clinical data input to identify patients at risk for cardiac complications. Additional application opportunities for AI in cardio-oncology involve multimodal cardiovascular imaging, where algorithms can also utilize imaging input to generate predictive risk profiles for cancer patients. The impact of AI extends to digital health tools, playing a pivotal role in the development of digital platforms and wearable technologies. Multidisciplinary teams have been formed to implement and evaluate the efficacy of these technologies, assessing AI-driven clinical decision support tools. Other avenues similarly support practical application of AI in clinical practice, such as incorporation into electronic health records (EHRs) to detect patients at risk for cardiovascular diseases. While these AI applications may help improve preventive measures and facilitate tailored treatment to patients, they are also capable of perpetuating and exacerbating healthcare disparities, if trained on limited, homogenous datasets. However, if trained and operated appropriately, AI holds substantial promise in positively influencing clinical practice in cardio-oncology. In this review, we explore the impact of AI on cardio-oncology care, particularly regarding predicting cardiotoxicity from cancer treatments, while addressing racial and ethnic biases in algorithmic implementation.
许多癌症治疗方法会对癌症幸存者产生有害的心血管影响。心血管毒性可能贯穿癌症治疗过程，并受多种因素影响。为降低这些风险，心脏肿瘤学应运而生，其重点在于预防和治疗因癌症及癌症治疗导致的心血管并发症。人工智能（AI）在改善心脏肿瘤学治疗效果方面具有多方面的潜力。目前，AI算法利用临床数据输入来识别有心脏并发症风险的患者。AI在心脏肿瘤学中的其他应用机会涉及多模态心血管成像，在这方面算法还可利用成像输入为癌症患者生成预测风险概况。AI的影响还延伸至数字健康工具，在数字平台和可穿戴技术的开发中发挥着关键作用。已经组建了多学科团队来实施和评估这些技术的疗效，评估由AI驱动的临床决策支持工具。其他途径也同样支持AI在临床实践中的实际应用，例如将其纳入电子健康记录（EHR）以检测有心血管疾病风险的患者。虽然这些AI应用可能有助于改进预防措施并为患者提供量身定制的治疗，但如果在有限的、同质化的数据集中进行训练，它们也可能会延续并加剧医疗保健差距。然而，如果经过适当的训练和操作，AI在积极影响心脏肿瘤学的临床实践方面具有巨大潜力。在本综述中，我们探讨了AI对心脏肿瘤学护理的影响，特别是在预测癌症治疗引起的心脏毒性方面，同时解决算法实施中的种族和民族偏见问题。
