Accuracy of FISH analysis in predicting chromosomal status in patients undergoing preimplantation genetic diagnosis

Size: px

Start display at page:

Download "Accuracy of FISH analysis in predicting chromosomal status in patients undergoing preimplantation genetic diagnosis"

Heather Black
5 years ago
Views:

1 Accuracy of FISH analysis in predicting chromosomal status in patients undergoing preimplantation genetic diagnosis Catherine M. DeUgarte, M.D., a Man Li, M.D., Ph.D., b Mark Surrey, M.D., c Hal Danzer, M.D., c David Hill, Ph.D., c and Alan H. DeCherney, M.D. d a Pacific Fertility Center, Los Angeles, California; b Fertility and Surgical Associates of California, Thousand Oaks, California; c Southern California Reproductive Center, Beverly Hills, California; and d National Institute of Health, Bethesda, Maryland Objective: The purpose of this study was to determine the positive predictive value (PPV) and negative predictive value (NPV) of FISH analysis and to determine which chromosomal abnormalities are most frequently confirmed. Design: Prospective observational. Setting: IVF laboratory. Patient(s): Two hundred forty-one embryos were analyzed from 98 patients. Intervention(s): FISH reanalysis. Main Outcome Measure(s): Embryos that would have been discarded in patients undergoing preimplantation genetic diagnosis (PGD) were fixed and FISH reanalysis was performed. Results of reanalysis were compared with the day 3 diagnosis while PPV and NPV were calculated. Result(s): Among the 241 embryos, 198 embryos were abnormal and 43 were normal by day 3 FISH analysis. The PPV was 83% and the NPV was 81%. PPV was also determined for specific categories of aneuploidy, and certain abnormalities such as monosomies, trisomies, tetrasomies, and polyploidies were frequently confirmed on reanalysis (PPV >80%), whereas Turner syndrome diagnosis was not (PPV ¼ 17%). Conclusion(s): FISH analysis offers a PPV of 83% and NPV of 81% when evaluating a single blastomere in conjunction with PGD. FISH errors and mosaicism are primarily responsible for the errors associated with FISH analysis in PGD. (Fertil Steril Ò 2008;90: Ó2008 by American Society for Reproductive Medicine.) Key Words: PGD, FISH, aneuploidy, confirmation, PPV, NPV Preimplantation genetic diagnosis (PGD) is a procedure that allows testing of embryos for specific genetic disorders before they enter the uterus and before pregnancy has begun (1). The first clinical application of PGD was described by Handyside et al. (2) in PGD was initially used for Mendelian disorders and has been more widely used since 1996, when locus-specific FISH probes became available (3). Preimplantation genetic diagnosis is performed in patients who undergo IVF procedures and is classified by the ESHRE PGD Consortium into the following categories: [1] high-risk PGD, which is done for patients with chromosomal abnormalities and/or single gene defects; [2] low-risk PGD, which has the goal of increasing pregnancy rates in IVF patients, for example, patients with advanced maternal age, repeated miscarriages, or repeated IVF failures (4). Embryo biopsy can be done at three different times during embryo development. The first one is polar body biopsy, in which the first and/or second polar bodies are removed and analyzed (5, 6). The second and most commonly performed time of biopsy is at the cleavage stage, often 3 days after Received August 22, 2005; revised May 17, 2007; accepted July 1, Reprint requests: Catherine Marin DeUgarte, M.D., Pacific Fertility Center, Wilshire Boulevard, Suite 700, Los Angeles, CA (FAX: ; cmdeugarte@hotmail.com). the eggs are harvested. The last stage is the blastocyst stage, and biopsy at this stage is least frequently performed at this time (4). When blastomere biopsy is performed, usually one blastomere is removed, and if removal of two cells is needed, the embryo should be at least at the six-cell stage (7). It is also best to select a blastomere that has one nucleus (8). Misdiagnosis rates for PGD by single-cell blastomere vary with reporting centers. It is therefore recommended that confirmatory diagnosis is performed on embryos diagnosed as abnormal as part of quality control for each lab (4). Mosaicism can contribute to misdiagnosis by single blastomere biopsy and FISH analysis, and has been reported to be prevalent in up to 50% of embryos that are four- to eight-cell stage (9 12). In a previous smaller study by Munne et al. (13), the false negative rate and false positive rates have been reported to be 1.5% and 19%, respectively. Several other studies have reported similar error rates. With this background in mind, the aims of this study were to [1] determine the positive predicted value (PPV) and negative predictive value (NPV) of FISH analysis of day 3 singleblastomere biopsy in low-risk PGD patients, as well as to [2] determine which chromosomal abnormalities are most frequently confirmed /08/$34.00 Fertility and Sterility â Vol. 90, No. 4, October doi: /j.fertnstert Copyright ª2008 American Society for Reproductive Medicine, Published by Elsevier Inc.

2 MATERIALS AND METHODS In between June 2004 and December 2004 at the Assisted Reproductive Technology Reproductive Center in Beverly Hills, CA, patients who were scheduled for an IVF-PGD procedure consented to the study. The study was approved by the Western institutional review board, and consent was obtained before the day of the oocyte retrieval. The patients who underwent retrieval had received gonadotropins in conjunction with a GnRH agonist or antagonist protocol, and ovulation was triggered with hcg approximately 36 hours before oocyte retrieval. Those patients desiring PGD had their embryos analyzed 3 days after oocyte retrieval when one blastomere was removed and FISH analysis was performed (Fig. 1A). The embryos read as normal were either transferred or cryopreserved. The embryos that were diagnosed as abnormal on day 3 and would have therefore been discarded were included in the study. Those normal embryos that were not deemed suitable for cryopreservation were also included in the study. Those embryos that were diagnosed as inconclusive or those that had no signal after confirmation with FISH was performed were excluded from the study. The embryos that were included in the study were then fixed entirely to slides on days 4 7 using Tween-20 solution and Carnoy s fixative (3:1 methanol to acetic acid solution), a modification of previously described techniques (14, 15), after which the slides were then treated in pepsin for cytoplasm removal. Following fixation, the slides were then placed in methanol, were allowed to dry, and a map of the FIGURE 1 (A) Diagram of methods. (B) Embryos evaluated for the study and the number of embryos in each category. blastomere nuclei was then established for future FISH analysis. A five-chromosome probe (Vysis) for chromosome 13 (red), 18 (aqua), 21 (green), X (blue), and Y (yellow) was then applied, after which DNA codenaturation was performed at 73 C for 5 minutes, followed by hybridization at for 4 hours at 37 C. In a few cases, when three chromosomes were checked on day 3, a three-chromosome probe was applied on confirmation day as well (chromosomes 18, X, and Y). The slides were then washed in 0.4 Sodium chloride/ Sodium citrate with 0.3% NP-40 at 73 C for approximately 2 minutes and then in 2 Sodium chloride/sodium citrate with 0.1% NP-40 at room temperature for 1 minute. Antifade solution was then applied for counterstaining, and FISH analysis was done using a fluorescence microscope (Leica, St. Gallen, Switzerland) with filter sets for observation of spectrum red/aqua/green/blue and yellow (Applied Imaging Corp., San Jose, CA). All nuclei were analyzed, counted, and the results were then entered in a computerized data base. Statistical analysis involved the calculation of PPVs and NPVs, along with their corresponding standard errors for the abnormal and normal embryos as well as for specific chromosomal abnormalities. Computed specificity and sensitivity were projected based on the observed PPVs and NPVs in this study and prevalence for patients screened by the same lab. RESULTS Overall, a total of 4,125 nuclei from 241 embryos were analyzed from 98 patients (average age 38.5) who underwent IVF-PGD, after 10 embryos were excluded secondary to lack of signal on FISH analysis. Of these, 198 embryos were abnormal and 43 were normal by day 3 FISH analysis (Fig. 1B). Confirmation with FISH was performed and the PPV, NPV, and the standard error were calculated (Table 1). Embryos were confirmed as abnormal when >50% of the nuclei were deemed abnormal on confirmation and normal if >50% of the nuclei were determined normal on confirmation with FISH. This was derived from previous studies using this cutoff value such as that by Ziebe et al. (16). Among the embryos diagnosed as abnormal on day 3, 164 of 198 were confirmed as abnormal on confirmation (83% ¼ PPV). Thirty-four embryos diagnosed as abnormal on day 3 were, in fact, normal, and the reasons for misdiagnosis were as follows: 10 were mosaic, 6 had signal overlap, 4 had split signals, 3 were tetraploidies, 2 were conservative diagnosis, 2 from Y failing hybridization, and the others were in the following categories: chromosomal self-correction, fragmented DNA, and inaccurate diagnosis. Among the embryos diagnosed as normal on day 3, 35 of 43 were confirmed as normal on confirmation, giving us an NPV of 81%. The projected sensitivity and specificity were calculated to be 94% and 60%, respectively, and were derived based on a presumed 39% prevalence of abnormal embryos in the same lab. All eight embryos that were abnormal in this category were misdiagnosed secondary to mosaicism DeUgarte et al. Accuracy of FISH analysis in PGD patients Vol. 90, No. 4, October 2008

3 TABLE 1 Positive and negative predictive values of confirmation FISH using >50 nuclei criteria, stratified by age, day of confirmation, and morphology. PPV PPV SE NPV NPV SE Age % (n ¼ 41) 7.3% 81.8% (n ¼ 22) 8.2% % (n ¼ 26) 3.8% 60.0% (n ¼ 5) 21.9% % (n ¼ 32) 6.9% 100.0% (n ¼ 7) n/a % (n ¼ 57) 4.8% 100.0% (n ¼ 2) n/a % (n ¼ 42) 4.5% 71.4% (n ¼ 7) 17.1% All ages (19 46) 83.3% 2.6% 81.4% 5.9% Day of confirmation % (n ¼ 89) 4.0% 100.0% (n ¼ 13) n/a % (n ¼ 61) 4.1% 88.9% (n ¼ 9) 10.5% 6 or % (n ¼ 48) 6.1% 66.7% (n ¼ 21) 10.3% Morphology Mor/Blast 72.9% (n ¼ 85) 4.8% 91.3% (n ¼ 23) 5.9% Cell/deg 91.2% (n ¼ 113) 2.7% 70.0% (n ¼ 20) 10.2% Note: PPV, positive predictive value; PPV SE, standard error of PPV; NPV, negative predictive value; NPV SE, standard error of NPV; day of confirmation, the day when the reanalysis was performed. Morphology is the embryo grade on the day the reanalysis was performed. Mor/Blast is when the embryo is either a morula or blastocyst, cell/deg is when the embryo is at all other cell stages or degenerated. Further statistical analysis was performed by age stratification (Table 1). Additionally, PPV and NPV were calculated by evaluating the specific day of confirmation, the day in which the reanalysis was done for all the ages combined. When reanalysis was performed on day 4, PPV was 83%, whereas NPV was 100%. On day 5, PPV and NPV were 88% and 89%, respectively, and on day 6 or 7 they were 77% and 67%, respectively. Additional analysis was performed looking specifically at the embryo morphology on the day of reanalysis. If on the day of reanalysis the embryo was at the morula or blastocyst stage, the PPV was 73% and the NPV was 91%, whereas at the cleavage stage or when the embryo was degenerated, the PPV and NPV were 91% and 70%, respectively. An additional calculation was performed by defining embryos as abnormal or normal if >80% of the nuclei were concordant on confirmation (Table 2). In other words, additional calculations were performed defining an embryo as normal if >80% of the nuclei were normal. The PPV was 88%, NPV was 56%, and the projected sensitivity and specificity were 71% and 79%, respectively. An additional aim of the study was to evaluate the PPVs of specific day 3 abnormalities (Table 3). In the embryos that had an abnormal diagnosis on day 3 (164 of 198), the exact abnormal diagnosis was confirmed in 56% (91 of 164) of the embryos, whereas 44% (73 of 164) were confirmed as abnormal, but had a different abnormality. The following abnormalities were confirmed in >80% of the cases (PPV >80): monosomies, trisomies, tetrasomies, polyploidies, abnormal nuclear fragments, embryos with two abnormalities, or those with more than two abnormalities. However, in those embryos that were diagnosed as Turner syndrome on day 3, only 17% were confirmed as abnormal upon confirmation with FISH. Within the Turner categories, mis-diagnosis was usually the case, secondary to poor hybridization or mosaicism (Fig. 2). DISCUSSION This study shows that PGD and FISH analysis of a single blastomere is a good technique but it is not 100% accurate, TABLE 2 Positive and negative predictive values of confirmation FISH using >80 nuclei criteria, stratified by age. Age PPV PPV SE NPV NPV SE % 6.9% 59.1% 10.5% % 0.0% 60.0% 21.9% % 5.8% 57.1% 18.7% % 4.3% 50.0% 35.4% % 3.3% 42.9% 18.7% All Ages (19-46) 87.9% 2.3% 55.8% 7.6% Fertility and Sterility â 1051

4 TABLE 3 Results of FISH reanalysis and PPV of embryos in specific diagnostic categories. Diagnostic category Normal Abnormal-confirmed Diff. Abnormality PPV Monosomy (N ¼ 43) % Trisomy (N ¼ 35) % Tetrasomy (N ¼ 6) % Polyploidy (N ¼ 16) % Turner (N ¼ 12) % Two abnormalities (N ¼ 28) % More than two abnormalities (N ¼ 21) % Nuclear frag-abnormal (N ¼ 27) % Other (N ¼ 10) % Total (N ¼ 198) % Note: Diagnostic category is the diagnosis on day 3 (monosomy includes monosomy 13, 18, or 21; trisomy includes trisomy 13, 18, or 21, tetrasomy includes tetrasomy 13, 18, or 21; polyploidy includes triploid, tetraploid, and polyploid; Turner is monosomy X. Two abnormalities includes any two abnormalies, that is, trisomy 18 and monosomy 21; more than two abnormalities includes three or more abnormalities as well as haploid; nuclear frag-abnormal includes two or more nuclear fragments, each with an abnormal diagnosis; other includes possible split, overlapped or unclear signals, nuclear fragments that are normal and nuclear fragments that have normal and abnormal fragments). Reanalysis results are in separate columns and represent the following: normal 50% or more of the nuclei are normal on reanalysis; abnormalconfirmed 50% or more of the nuclei are the same abnormality as that assigned on day 3, for example, monosomy 13 that is confirmed as monosomy 13; diff abnormality 50% or more of the nuclei are abnormal, but the abnormality is different than that diagnosed on day 3, for example, monosomy 13 that is now diagnosed as monosomy 21. PPV is positive predictive value. and therefore, the results should be interpreted with caution. In addition, several abnormalities are more likely to be confirmed than others. In this study we noted that out of the 198 abnormal embryos, 164 were confirmed when the whole embryo was analyzed by FISH. The PPV was 83%, which signifies that 17% of embryos are misdiagnosed as abnormal on day 3, when they are, in fact, normal. These embryos are therefore not transferred, perhaps reducing the chance of pregnancy in some patients. Of note, 56% of the embryos with a specific abnormal diagnosis on day 3 were confirmed, whereas 44% of embryos were confirmed as abnormal, but had a different abnormality upon reanalysis. We also analyzed embryos in this study that were considered normal on day 3. Of the 43 available normal embryos, 35 were confirmed as normal on FISH reanalysis, giving us an NPV of 81. However, this part of the study should be interpreted with caution, because there were fewer embryos in this category available for research. It is a lot more difficult to obtain normal embryos for reanalysis because most normal embryos are either transferred or cryopreserved. Those embryos that did not meet the minimum morphologic criteria for cryopreservation were also included in this study; therefore, the results may have been biased by a higher proportion of abnormal embryos in the euploid group. The reason for the high NPV might also be attributed to the fact that many errors occur at the postzygotic stage, and that embryos that are deemed to be abnormal may have more errors in the regulation of the cell cycle (16). Furthermore, other studies have shown that growth retardation in embryos could be an indication of abnormal chromosomal makeup (17, 18). We have also shown that PPV and NPV are best when confirmation is performed on day 4 or 5, which makes sense because degeneration after day 5 is greater. Our data shows that when reanalysis is performed on embryos that are at the morula or blastocyst stage, the PPV was 73% and the NPV was 91%, whereas at the cleavage stage or when the embryo was degenerated, the PPV and NPV were 91% and 70%, respectively. An additional aim of the study was to evaluate the positive predictive values of specific day 3 abnormalities to help physicians when they are looking at a PGD report in conjunction with morphology of the embryos. A previous study had revealed that when looking at specific abnormalities, concordance with day 3 diagnosis was present in 97% of those diagnosed with trisomy, 65% for monosomy, and 100% in those diagnosed with complex abnormal aneuploidies (19). The following abnormalities were confirmed in our study in >80% of the cases: monosomies, trisomies, tetrasomies, polyploidies, abnormal nuclear fragments, embryos with two abnormalities, or those with more than two abnormalities. However, in the embryos that were diagnosed as Turner syndrome on day 3, only 17% were confirmed as abnormal upon confirmation with FISH. This means that when a Turner syndrome diagnosis is encountered on day 3, several measures 1052 DeUgarte et al. Accuracy of FISH analysis in PGD patients Vol. 90, No. 4, October 2008

FIGURE 2 (A) Embryo on day 3, blastomere analysis done for 18, X, and Y. Reading was 1818X, Turner syndrome (two aqua signals and one red or yellow signal that appears orange after picture is taken).

5 FIGURE 2 (A) Embryo on day 3, blastomere analysis done for 18, X, and Y. Reading was 1818X, Turner syndrome (two aqua signals and one red or yellow signal that appears orange after picture is taken). (B) Same embryo on day 6, a blastocyst, after fixation and FISH analysis. Of the 60 nuclei, 55 were normal female (each with two red or yellow signals and two aqua signals) and 5 were Turner syndrome. Looking back on the day 3 image, the X signals appear close together. may be taken to confirm the diagnosis. Rehybridization can be performed, because the Y signal may occasionally be weak, misleading the results and subsequently having one X as part of the reading. Rebiopsy of that embryo can be performed on day 3 or 4 to evaluate if, in fact, Turner diagnosis exists. One could speculate the reasons why the PPVand NVP are not 100%: clearly, the main problem is the mosaicism that interferes with accurate PGD results. As Munne et al. (20) have described, mosaicism that arises from the first embryonic division can be detected by single-cell biopsy on day 3, whereas mosaicism that occurs at later stages cannot. The chance of detecting mosaicism in an eight-cell embryo that is diploid has been shown to be 25% to 50%, depending on when mosaicism occurs. Mosaicism could also be caused by culture conditions, and therefore, this may have also affected results. This study has a number of strengths, including the large number of embryos used and the categories of abnormalities that were evaluated. However, potential limitations are recognized. Foremost among these is the mosaicism of the embryo. When one blastomere is removed from a mosaic embryo, false negative or false positive results are possible. FISH techniques are also not without error, and split, overlapped, or loss of signals can result in false diagnoses. The cutoff to determine if an embryo was normal was 50%, a number lower than previous studies, which could also have contributed to the increased number of false positive and false negative rates. In conclusion, FISH analysis of a single blastomere can be useful in conjunction with PGD for selected patients. However, the technique is not without errors, and the results should be individualized. FISH errors and mosaicism are prevalent, and should be considered when PGD reports are reviewed. Acknowledgments: The authors acknowledge Dr. Jeff Gornbein from UCLA for his assistance with the statistical analysis in for this study. REFERENCES 1. Thornhill AR, Snow K. Molecular diagnostics in preimplantation genetic diagnosis. J Mol Diagn 2002;4: Handyside AH, Pattinson JK, Penketh RJ, Delhanty JD, Winston RM, Tuddenham EG. Biopsy of human preimplantation embryos and sexing by DNA amplification. Lancet 1989;18: Verlinsky Y, Cohen J, Munne S, Gianaroli L, Simpso JL. Over a decade of experience with preimplantation genetic diagnosis: a multicenter report. Fertil Steril 2004;82: Thornhill AR, dedie-smulders CE, Geraedts JP, Harper JC, Harton GL, Lavery SA, et al. ESHRE PGD Consortium Best practice guidelines for clinical preimplantation diagnosis (PGD) and preimplantation genetic screening (PGS). Hum Reprod 2005;20: Munne S, Dailey T, Sultan KM, Grifo J, Cohen J. The use of first polar bodies for preimplantation diagnosis of aneuploidy. Hum Reprod 1995;10: Verlinsky Y, Cieslak J, Freidine M, Ivakhnenko V, Wolf G, Kovalinskaya L, et al. Polar body diagnosis of common aneuploidies by FISH. J Assist Reprod Genet 1996;13: Van de Velde H, De Vos A, Sermon K, Staessen C, De Rycke M, Van Assche E. Embryo implantation after biopsy of one or two cells from Fertility and Sterility â 1053

6 cleavage-stage embryos with a view to preimplantation genetic diagnosis. Prenat Diagn 2000;20: Munne S, Cohen J. Unsuitability of multinucleated human blastomeres for preimplantation genetic diagnosis. Hum Reprod 1993;8: Munne S, Magli C, Adler A, Wright G, de Boer K, Mortimer D, et al. Treatment-related chromosome abnormalities in human embryos. Hum Reprod 1997;12: Katagiri Y, Katayama S. Influence of mosaicism on sexing of human preembryos detected by the polymerase chain reaction. J Assist Reprod Genet 1996;13: Handyside AH, Delhanty JD. Preimplantation genetic diagnosis: strategies and surprises. Trends Genet 1997;13: Baart EB, Van Opstal D, Los FJ, Fauser BC, Martini E. Fluorescence in situ hybridization analysis of two blastomeres from day 3 frozen-thawed embryos followed by analysis of the remaining embryo on day 5. Hum Reprod 2004;19: Munne S, Marquez C, Magli C, Morton P, Morrison L. Scoring criteria for preimplantation genetic diagnosis of numerical abnormalities for chromosomes X, Y, 13, 16, 18 and 21. Mol Hum Reprod 1998;4: Dozortsev DI, McGinnis KT. An improved fixation technique for fluorescence in situ hybridization for preimplantation genetic diagnosis. Fertil Steril 2001;76: Velilla E, Escudero T, Munne S. Blastomere fixation techniques and risk of misdiagnosis for preimplantation genetic diagnosis of aneuploidy. Reprod Biomed Online 2002;4: Ziebe S, Lundin K, Loft A, Bergh C, Nyboe Andersen A, Selleskog U, et al. CEMAS II and Study Group. FISH analysis for chromosomes 13, 16, 18, 21, 22, X and Y in all blastomeres of IVF pre-embryos from 144 randomly selected donated human oocytes and impact on pre-embryo morphology. Hum Reprod 2003;18: Bielanska M, Tan S, Ao A. Chromosomal mosaicim throughout human preimplantation development in vitro: incidence, type and relevance to embryo outcome. Hum Reprod 2002;17: Magli MC, Gianaroli L, Munne S, Ferraretti A. Incidence of chromosomal abnormalities from a morphologically normal cohort of embryos in poor-prognosis patients. J Assist Reprod Genet 1998;15: Magli MC, Jones GM, Gras L, Gianaroli L, Korman L, Trouson AO. Chromosome mosaicism in day 3 aneuploid embryos that develop to morphologically normal blastocysts in vitro. Hum Reprod 2000;15: Munne S, Weier HU, Grifo J, Cohen J. Chromosome mosaicism in human embryos. Biol Reprod 1994;51: DeUgarte et al. Accuracy of FISH analysis in PGD patients Vol. 90, No. 4, October 2008

Incidence of Chromosomal Abnormalities from a Morphologically Normal Cohort of Embryos in Poor- Prognosis Patients

Incidence of Chromosomal Abnormalities from a Morphologically Normal Cohort of Embryos in Poor- Prognosis Patients M. C. MAGLI,1 L. GIANAROLI,1,3 S. MUNNE,2 and A. P. FERRARETTI1 Submitted: December 29,