THE IMPACT OF USING ARTIFICIAL INTELLIGENCE IN REPORTING HIGH RESOLUTION NGS RESULTS UPON PRIORITIZATION OF EMBRYOS SELECTED FOR TRANSFER

Abstract

There are many factors that influence the interpretation of the plots of high resolution NGS (hr-NGS) in PGT A examination of the embryos, such as the Genetist experience, cut-off points, and the chromosomal number. The PGT A results differentiate between euploidy embryos that are given first priority for embryo transfer (ET), second priority for low mosaic, and aneuploidy or high level mosaic those who are not transferred . This study compares between manual calling of the results versus using artificial intelligence platform. This is a prospective ongoing study. A total of 720 biopsied sample from 280 patients processed according to Miseq Illumina protocol. Results are manually called subjective methodology (SM) using Bluefuse software and also uploaded up to PGTaiSM platform Coopersurgical. The mean age of the female partners is 34 years. There was a difference in 138 results out of 720 (19%) comparing between the “SM” and PGTaiSM platform interpretation. The following table shows the differences between four categories in both SM and PGTai results:Tabled 1categoriesSMPGTaiSMNo. of embryosEuploid3114 low level mosaic11 high level mosaic5 aneuploid1 QA failAneuploid271 low level mosaic15 high level mosaic1 euploid10 QA failHigh level mosaic292 low level mosaic14 euploid12 aneuploid1 QA failLow level mosaic4612 high level mosaic28 euploid5 aneuploid1 QA failThe PGTai platform changes some of the results in all reported categories. That would change the prioritization of embryos for transfer. There were 16 embryos that were reported euploidy by SM, while they are deemed “untransferable” by PGTai, being 5 aneuploidy and 11 being of high level mosaic. Again, in the same group 14 embryos were of low mosaic level which would have been given a second priority. In the aneuploidy group only 2 embryos (1 euploid and 1 low level mosaic) would have been possibly transferred according to PGTai.In the high level mosaic group 16 embryos would have been given priority for embryo transfer as 14 were found euploidy while 2 were of low mosaic level. In the low level mosaic group 28 embryos would have been given first priority as found to be euploid, while 17 embryos would have must been considered to be as 12 of high level mosaic and 5 were aneuploidy. Open table in a new tab The PGTai platform changes some of the results in all reported categories. That would change the prioritization of embryos for transfer. There were 16 embryos that were reported euploidy by SM, while they are deemed “untransferable” by PGTai, being 5 aneuploidy and 11 being of high level mosaic. Again, in the same group 14 embryos were of low mosaic level which would have been given a second priority. In the aneuploidy group only 2 embryos (1 euploid and 1 low level mosaic) would have been possibly transferred according to PGTai. In the high level mosaic group 16 embryos would have been given priority for embryo transfer as 14 were found euploidy while 2 were of low mosaic level. In the low level mosaic group 28 embryos would have been given first priority as found to be euploid, while 17 embryos would have must been considered to be as 12 of high level mosaic and 5 were aneuploidy. PGTai is thought to be more valid than manual subjective calling of results as based upon true data set of live births and sustained pregnancy outcomes. This study shows a strong difference of 19% in results that changes how embryos are prioritized for ET. Moreover, same embryos would have been not transferred as being aneuploidy as of high mosaic level. Further studies are needed to evaluate if using PGTai would improve live birth rate and decrease miscarriage by better selection of embryos.