Reinforcement Learning for Precision Oncology.
Reinforcement Learning for Precision Oncology.
复制标题
精准肿瘤学的强化学习。
DOI：
10.3390/cancers13184624
复制
发表时间：
2021-09-15
期刊：
影响因子：
5.2
通讯作者：
Middeke JM
中科院分区：
文献类型：
--
作者：
Eckardt JN;Wendt K;Bornhäuser M;Middeke JM
信息技术与癌症研究的加速融合预示着肿瘤学临床决策的新方法和新模型的出现。强化学习作为机器学习的主要分支之一，具有开发高性能决策支持工具的潜力。然而，最近许多肿瘤学强化学习的研究都存在共同的缺点和缺陷，需要解决这些缺点和缺陷，以便为未来的临床实践开发安全、可解释和可靠的算法。精确肿瘤学的基础是对恶性疾病的遗传和分子机制的不断了解，并为个体患者提供不同的治疗途径。医疗数据的日益复杂导致了机器学习技术的实施，这些技术广泛应用于使用监督或无监督学习进行风险评估和结果预测。强化学习（RL）仍然在很大程度上被忽视，它通过探索环境的潜在动态来解决顺序任务，并通过采取行动来塑造环境，以便随着时间的推移最大化累积奖励，从而实现最佳的长期结果。RL最近的突破在游戏玩法和自动驾驶方面取得了显著的成果，通常可以实现类似人类甚至超人的性能。虽然这种类型的机器学习有可能成为一种有用的决策支持工具，但它也带来了一系列独特的挑战，需要解决这些挑战，以确保其适用性、有效性和安全性。在这篇综述中，我们重点介绍了RL在肿瘤学研究方面的最新进展，并指出了未来研究中需要考虑的挑战和陷阱，以便成功开发基于RL的精确肿瘤学决策支持系统。
The accelerating merger of information technology and cancer research heralds the advent of novel methods and models for clinical decision making in oncology. Reinforcement learning—as one of the major subspecialties in machine learning—holds the potential for the development of high-performance decision support tools. However, many recent studies of reinforcement learning in oncology suffer from common shortcomings and pitfalls that need to be addressed for the development of safe, interpretable and reliable algorithms for future clinical practice. Precision oncology is grounded in the increasing understanding of genetic and molecular mechanisms that underly malignant disease and offer different treatment pathways for the individual patient. The growing complexity of medical data has led to the implementation of machine learning techniques that are vastly applied for risk assessment and outcome prediction using either supervised or unsupervised learning. Still largely overlooked is reinforcement learning (RL) that addresses sequential tasks by exploring the underlying dynamics of an environment and shaping it by taking actions in order to maximize cumulative rewards over time, thereby achieving optimal long-term outcomes. Recent breakthroughs in RL demonstrated remarkable results in gameplay and autonomous driving, often achieving human-like or even superhuman performance. While this type of machine learning holds the potential to become a helpful decision support tool, it comes with a set of distinctive challenges that need to be addressed to ensure applicability, validity and safety. In this review, we highlight recent advances of RL focusing on studies in oncology and point out current challenges and pitfalls that need to be accounted for in future studies in order to successfully develop RL-based decision support systems for precision oncology.
登录
查看更多内容
DOI：
10.1146/annurev-statistics-022513-115553
发表时间：
2014
影响因子：
7.9
作者：
Chakraborty B;Murphy SA
通讯作者：
Murphy SA
影响因子：
4.1
作者：
Grote, Thomas;Berens, Philipp
通讯作者：
Berens, Philipp
DOI：
10.1109/jbhi.2020.3004663
发表时间：
2020-11-01
影响因子：
7.7
作者：
Daoud, Salma;Mdhaffar, Afef;Freisleben, Bernd
通讯作者：
Freisleben, Bernd
影响因子：
7.7
作者：
Charron, Odelin;Lallement, Alex;Meyer, Philippe
通讯作者：
Meyer, Philippe
影响因子：
5.2
作者：
Comes MC;La Forgia D;Didonna V;Fanizzi A;Giotta F;Latorre A;Martinelli E;Mencattini A;Paradiso AV;Tamborra P;Terenzio A;Zito A;Lorusso V;Massafra R
通讯作者：
Massafra R