External quality assessment (EQA) schemes assess the performance of predictive biomarker testing in lung and colorectal cancer and have previously demonstrated variable error rates. No information is currently available on the underlying causes of incorrect EQA results in the laboratories. Participants in EQA schemes by the European Society of Pathology between 2014 and 2018 for lung and colorectal cancer were contacted to complete a survey if they had at least one analysis error or test failure in the provided cases. Of the 791 surveys that were sent, 325 were completed including data from 185 unique laboratories on 514 incorrectly analyzed or failed cases. For the digital cases and immunohistochemistry, the majority of errors were interpretation-related. For fluorescence in situ hybridization, problems with the EQA materials were reported frequently. For variant analysis, the causes were mainly methodological for lung cancer but variable for colorectal cancer. Post-analytical (clerical and interpretation) errors were more likely detected after release of the EQA results compared to pre-analytical and analytical issues. Accredited laboratories encountered fewer reagent problems and more often responded to the survey. A recent change in test methodology resulted in method-related problems. Testing more samples annually introduced personnel errors and lead to a lower performance in future schemes. Participation to quality improvement projects is important to reduce deviating test results in laboratories, as the different error causes differently affect the test performance. EQA providers could benefit from requesting root cause analyses behind errors to offer even more tailored feedback, subschemes, and cases.

外部质量评估（EQA）方案评估了肺癌和大肠癌中预测性生物标志物测试的性能，并且先前已证明了可变错误率。目前尚无有关实验室中EQA结果不正确的根本原因的信息。如果2014年至2018年间欧洲病理学会针对肺癌和大肠癌的EQA计划参与者至少有一个分析错误或测试失败，请联系他们完成调查。在发送的791个调查中，有325个已完成，包括来自185个独立实验室的514个错误分析或失败案例的数据。对于数字病例和免疫组织化学，大多数错误与解释相关。对于荧光原位杂交，EQA材料的问题经常被报道。对于变体分析，原因主要是肺癌的方法论原因，而结直肠癌的原因则各不相同。与分析前和分析问题相比，在发布EQA结果后更有可能检测到分析后（文书和解释）错误。经过认证的实验室遇到的试剂问题更少，对调查的回复更多。测试方法的最新变化导致方法相关的问题。每年测试更多样本会引入人为错误，并导致未来方案的性能降低。参与质量改进项目对于减少实验室中的测试结果很重要，因为不同的错误原因会不同地影响测试性能。