Diagnostic performance of an artificial intelligence-driven cardiac-structured reporting system for myocardial perfusion SPECT imaging.
Diagnostic performance of an artificial intelligence-driven cardiac-structured reporting system for myocardial perfusion SPECT imaging.
复制标题
DOI：
10.1007/s12350-018-1432-3
复制
发表时间：
2020-10
期刊：
影响因子：
--
通讯作者：
Esteves F
中科院分区：
文献类型：
--
作者：
Garcia EV;Klein JL;Moncayo V;Cooke CD;Del'Aune C;Folks R;Moreiras LV;Esteves F
通过将自动生成的报告与核心脏病学专家生成的实际临床报告的结果进行直接比较来描述和验证人工智能驱动的结构化报告系统。我们的 AI 报告系统使用从 MPI 研究中提取的定量参数来自动生成符合指南的结构化报告 (sR)。一种新的非参数方法为 17 个 LV 节段中的每一个节段生成休息和压力、灌注和增厚的分布函数，然后将其转换为节段灌注不足、缺血的确定性因子 (CF)。然后，这些 CF 将被输入到我们的一组启发式规则中，用于得出诊断结果和印象，并传播到结构化报告中，称为人工智能驱动的结构化报告 (AIsR)。 AIsR 检测 CAD 和缺血的诊断准确性在 1,000 名接受过休息/应激 SPECT MPI 的患者中进行了测试。在高特异性水平上，在 100 名患者的子集中，AIsR 与 9 名专家对 CAD (p = .33) 或缺血 (p = .37) 的印象之间的一致性没有统计学差异。这种高特异性水平还为 1000 名患者提供了全球和区域结果的最高准确度。在所有比较中，这些准确度在统计上显着优于其他两个水平（SN/SP 权衡、高灵敏度）。该人工智能报告系统自动生成结构化自然语言报告，其诊断性能可与专家相媲美。
To describe and validate an AI driven structured reporting system by direct comparison of automatically generated reports to results from actual clinical reports generated by nuclear cardiology experts. Quantitative parameters extracted from MPI studies are used by our AI reporting system to generate automatically a guideline compliant structured report (sR). A new non-parametric approach generates distribution functions of rest and stress, perfusion and thickening, for each of 17 LV segments that are then transformed to certainty factors (CF) that a segment is hypoperfused, ischemic. These CFs are then input to our set of heuristic rules used to reach diagnostic findings and impressions propagated into a structured report referred as an AI driven structured Report (AIsR). The diagnostic accuracy of the AIsR for detecting CAD and ischemia was tested in 1,000 patients who had undergone rest /stress SPECT MPI. At the high-specificity level, in a subset of 100 patients, there were no statistical differences in the agreements between the AIsR and nine experts' impressions of CAD (p = .33) or ischemia (p = .37). This high-specificity level also yielded the highest accuracy across global and regional results in the 1000 patients. These accuracies were statistically significantly better than the other two levels (SN/SP tradeoff, high sensitivity) across all comparisons. This AI reporting system automatically generates a structured natural language report with a diagnostic performance comparable to those of experts.
登录
查看更多内容
影响因子：
37.8
作者：
Cerqueira, MD;Weissman, NJ;Verani, MS
通讯作者：
Verani, MS
影响因子：
2.4
作者：
Esteves, Fabio P.;Galt, James R.;Garcia, Ernest V.
通讯作者：
Garcia, Ernest V.
影响因子：
24
作者：
Rozanski, Alan;Gransar, Heidi;Berman, Daniel S.
通讯作者：
Berman, Daniel S.
影响因子：
8.5
作者：
EZQUERRA, N;MULLICK, R;GARCIA, EV
通讯作者：
GARCIA, EV
影响因子：
7.5
作者：
Haddad, M;Adlassnig, KP;Porenta, G
通讯作者：
Porenta, G