许多出版物已经证明了深度学习（DL）算法在基于图像的癌症诊断方面与临床医生相当甚至表现更好，但这些算法经常被视为对手而不是合作伙伴。尽管临床医生在DL辅助下的方法具有很大的潜力，但没有研究系统地量化了临床医生在基于图像的癌症识别中在有或没有DL辅助下的诊断准确性。我们对临床医生在基于图像的癌症识别中有无DL辅助的诊断准确性进行了系统量化，对于任何类型的研究设计都允许，重点比较未经辅助的临床医生和DL辅助的临床医生在使用医学影像学中的癌症识别方面的表现。排除使用医学波形数据图形材料和研究图像分割而非分类的研究。筛选纳入二元诊断准确性数据和列联表以进行进一步荟萃分析的研究。分别分析癌症类型和影像学方式的两个亚组。我们搜索了PubMed、Embase、IEEEXplore和Cochrane图书馆，涵盖了2012年1月1日至2021年12月7日发表的研究，共发现9796篇研究，其中48篇研究符合系统评价的标准。其中25项研究比较了未经辅助的临床医生和DL辅助的临床医生，并提供了足够的数据供统计综合分析。我们发现未经辅助的临床医生的汇总敏感度为83％（95％CI 80％-86％），DL辅助的临床医生汇总敏感度为88％（95％CI 86％-90%）。未经辅助的临床医生的汇总特异度为86％（95％CI 83％-88％），DL辅助的临床医生的汇总特异度为88％（95％CI 85％-90％）。DL辅助的临床医生的汇总敏感度和特异度值均高于未经辅助的临床医生，分别为1.07（95％CI 1.05-1.09）和1.03（95％CI 1.02-1.05）。预定义的亚组中也观察到DL辅助临床医生的类似诊断性能。总之，在基于图像的癌症识别中，DL辅助临床医生的诊断性能似乎比未经辅助临床医生更好。但是，需要谨慎行事，因为审查的研究提供的证据并未涵盖现实临床实践中涉及的所有微观细节。将临床实践的定性见解与数据科学方法相结合可能会改进DL辅助实践，但需要进一步的研究。PROSPERO CRD42021281372；https://www.crd.york.ac.uk/prospero/display_record.php？RecordID=281372。©彭学、司明宇、秦东旭、魏炳睿、塞缪尔·西瑞、叶子辰、陈明阳、王素梦、宋成、张博、丁明、张文玲、白安英、闫慧娇、荡乐、赵玉倩、雷米拉·热扎克、张少凯、乔友林、屈一民、姜宇。 原始出版物发表于 Journal of Medical Internet Research（https://www.jmir.org），2023年3月2日。

A number of publications have demonstrated that deep learning (DL) algorithms matched or outperformed clinicians in image-based cancer diagnostics, but these algorithms are frequently considered as opponents rather than partners. Despite the clinicians-in-the-loop DL approach having great potential, no study has systematically quantified the diagnostic accuracy of clinicians with and without the assistance of DL in image-based cancer identification.We systematically quantified the diagnostic accuracy of clinicians with and without the assistance of DL in image-based cancer identification.PubMed, Embase, IEEEXplore, and the Cochrane Library were searched for studies published between January 1, 2012, and December 7, 2021. Any type of study design was permitted that focused on comparing unassisted clinicians and DL-assisted clinicians in cancer identification using medical imaging. Studies using medical waveform-data graphics material and those investigating image segmentation rather than classification were excluded. Studies providing binary diagnostic accuracy data and contingency tables were included for further meta-analysis. Two subgroups were defined and analyzed, including cancer type and imaging modality.In total, 9796 studies were identified, of which 48 were deemed eligible for systematic review. Twenty-five of these studies made comparisons between unassisted clinicians and DL-assisted clinicians and provided sufficient data for statistical synthesis. We found a pooled sensitivity of 83% (95% CI 80%-86%) for unassisted clinicians and 88% (95% CI 86%-90%) for DL-assisted clinicians. Pooled specificity was 86% (95% CI 83%-88%) for unassisted clinicians and 88% (95% CI 85%-90%) for DL-assisted clinicians. The pooled sensitivity and specificity values for DL-assisted clinicians were higher than for unassisted clinicians, at ratios of 1.07 (95% CI 1.05-1.09) and 1.03 (95% CI 1.02-1.05), respectively. Similar diagnostic performance by DL-assisted clinicians was also observed across the predefined subgroups.The diagnostic performance of DL-assisted clinicians appears better than unassisted clinicians in image-based cancer identification. However, caution should be exercised, because the evidence provided in the reviewed studies does not cover all the minutiae involved in real-world clinical practice. Combining qualitative insights from clinical practice with data-science approaches may improve DL-assisted practice, although further research is required.PROSPERO CRD42021281372; https://www.crd.york.ac.uk/prospero/display_record.php?RecordID=281372.©Peng Xue, Mingyu Si, Dongxu Qin, Bingrui Wei, Samuel Seery, Zichen Ye, Mingyang Chen, Sumeng Wang, Cheng Song, Bo Zhang, Ming Ding, Wenling Zhang, Anying Bai, Huijiao Yan, Le Dang, Yuqian Zhao, Remila Rezhake, Shaokai Zhang, Youlin Qiao, Yimin Qu, Yu Jiang. Originally published in the Journal of Medical Internet Research (https://www.jmir.org), 02.03.2023.