The University of Texas MD Anderson Cancer Center, Houston, Texas.
Cancer Res. 2021 Dec 1;81(23):5810-5812. doi: 10.1158/0008-5472.CAN-21-1929.
Profound advances in computational methods, including artificial intelligence (AI), present the opportunity to use the exponentially growing volume and complexity of available cancer measurements toward data-driven personalized care. While exciting, this opportunity has highlighted the disconnect between the promise of compute and the supply of high-quality data. The current paradigm of ad-hoc aggregation and curation of data needs to be replaced with a "metadata supply chain" that provides robust data in context with known provenance, that is, lineage and comprehensive data governance that will allow the promise of AI technology to be realized to its full potential in clinical practice.
计算方法的重大进展，包括人工智能 (AI)，为利用可用癌症测量数据的指数级增长的数量和复杂性实现数据驱动的个性化护理提供了机会。虽然令人兴奋，但这一机会突显了计算的前景与高质量数据的供应之间的脱节。当前，需要用一个“元数据供应链”取代临时聚合和整理数据的范式，该供应链将以已知的来源（即谱系）和全面的数据治理为背景提供可靠的数据，这将使人工智能技术的承诺在临床实践中充分发挥其潜力成为可能。
