Calibration of the Becton Dickinson FocalPoint GS slide profiler is essential for optimal performance: Avoiding the algorithm supersaturation cascade effect pitfall

摘要

BACKGROUNDThe MAVARIC study supported the use of the FocalPoint GS (FPGS) imaging system "No Further Review" (NFR) technology for cervical screening and recommended further investigation. A validation study (Nuttall et al.) was performed by Cervical Screening Wales before implementing the NFR slide reporting technology within the cervical screening program in Wales, United Kingdom. METHODA total of 45,317 SurePath liquid-based cytology cervical screening samples were submitted for FPGS scanning within four Welsh cytology laboratories between 2006 and 2011. The study, Computer Assisted Evaluation, Screening and Reporting, involved scanning the slides using the FPGS and comparing the results with manual screening performed under established Cervical Screening Wales protocols. RESULTSAn increased number of abnormal cases presented in the NFR reporting category, significantly greater than that previously encountered. This anomaly resulted in higher false-negative rates with potentially life-changing consequences for the screening participant. Subsequent investigation determined that this increase in cases created an algorithm cascade or "sump" effect, which resulted in an unprecedented increased number of samples categorized as NFR. This exceeded the calibration parameters set for the FPGS and was thought to be caused by an increase in the number of younger women attending for screening following the death of a young reality television celebrity from cervical cancer. CONCLUSIONAdequate and timely calibration of FPGS technology is vital for quality assurance of the results produced, particularly following events that may impact on cervical precancer incidence rates. Failure to do so can result in potentially catastrophic screening incidents that are avoidable.