- Email: [email protected]
Is the 15-in situ clone protocol necessary to detect amniotic fluid mosaicism?
Is the 15-in situ clone protocol necessary to detect amniotic fluid mosaicism? Edith Y. Cheng, MD, b David A. Luthy, MD, = Deborah F.B. Dunne, MS, a Frederick W. Luthardt, PhD, = and Christine M. Disteche, PhD c
Seattle, Washington OBJECTIVE: Our purpose was to evaluate the 15-clone analysis for detecting amniotic fluid mosaicism by the in situ method. STUDY DESIGN: A 10-year review was performed of all amniotic fluid mosaicism cases at two institutions using the in situ method exclusively, with sequential clonal analysis to determine the first and second clone in which the abnormal cell line occurred. RESULTS: Of the 28,497 amniotic fluid samples, 73 met criteria for amniotic fluid mosaicism by in situ method (0.26%). There were 54 cases (0.19%) with potential clinical significance (23 autosome and 31 sex chromosome mosaicism); 49 of the 54 cases (89%) were detected in the first six clones, including 22 of 23 involving autosomes and 27 of 31 involving sex chromosomes. In one of the six cases detected after clone 6 (46,XX/47,XX, +21) the mosaic cell line was present in 20% of the clones analyzed and was followed by a voluntary termination of the pregnancy, in the other five cases amniotic fluid mosaicism was present in < 20% of the clones; these included one case of 46,XX/47,XX + mar (15% amniotic fluid mosaicism, voluntary termination of pregnancy), two cases of 45,X]46,XY (10% to 12% amniotic fluid mosaicism, both normal at birth), and two cases of 45,W46,XX (8% amniotic fluid mosaicism, lost to follow-up; 12% amniotic fluid mosaicism, voluntary termination of pregnancy). By limiting the analysis to six clones, - 2 0 % of analysis time could be saved per case, but one autosomal amniotic fluid mosaicism case per 10,000 samples could potentially be missed. CONCLUSION: Reducing the number of clones analyzed by in situ method could result in increased efficiency, decreased costs, and minimal loss of sensitivity. (AM J OBSTETGYNECOL1995; 173:1025-30.)
Key words: Amniotic fluid mosaicism, in situ method, 15-clone count
In prenatal cytogenetic diagnosis true mosaicism in amniotic fluid is rare, occurring in approximately 0.1% to 0.3% of samples. "~ However, the efficient diagnosis and interpretation of true amniotic fluid mosaicism remain a challenge. In 1977 Hook ~ determined the n u m b e r of cells to be examined for detection of 10%, 20%, and 30% mosaicism in peripheral blood at 90%, 95%, and 99% confidence levels. This model was extended to the analysis of amniotic fluid and reevaluated by Richkind and Risch 5 and Featherstone et al. ~ to apply to the flask and in situ methods. For the flask method the current practice is to examine a m i n i m u m of 20 cells from two flasks. True amniotic fluid mosaicism is diagnosed when two or more cells with the same
From the Center for Perinatal Studies, Swedish Health Services," and the Departments of Obstetrics and Gynecologyb and Pathology,' University of Washington Medical Center. Partially funded as a special project of the Pacific Northwest Regional Genetics Group by Project No. MCJ411002-11 of the Maternal and Child Health Bureau, Department of Health and Human Se~'oices. Presented in part at the Fifteenth Annual Meeting of the Society of Perinatal Obstetricians, Atlanta, Georgia, January 23-28, 1995. Reprint requests: David A. Luthy, MD, Division of Perinatal Medicine, 747 Broadway, Seattle, WA 98104. Copyright © 1995 by Mosby-Year Book, Inc. 0002-9378/95 $5.00 + 0 6/6/66695
cytogenetic abnormality are found in two or more flasks. For the in situ method one cell each from 15 to 20 clones from multiple culture chambers are analyzed. True mosaicism is defined as the presence of two or more clones with the same abnormality from different culture chambers. These guidelines permit detection of mosaicism at the 20% level or greater with a confidence level of 95% to 99%. 5 Although the flask and in situ methods have been shown to be equally sensitive in detecting true mosaicism, 7' s there continues to be discussion regarding the necessity of additional cytologic evaluations to arrive at an accurate diagnosis and interpretation when true mosaicism is encountered. Hsu et al. ~ specifically addressed this question by reviewing the 50 cases of true mosaicism among 22,000 amniotic fluid samples examined by the Prenatal Diagnosis Laboratory of New York City and concluded that not all cases of true mosaicism required an exhaustive workup for diagnosis and interpretation. Featherstone et al." arrived at similar conclusions that a two-stage approach, dictated by the specific chromosome and its associated phenotype, decreased the amount of unnecessary laboratory effort and was both efficient and cost effective. To further investigate the efficiency of the 15-clone in situ method for detect1025
1026 Cheng et al.
October 1995 Am J Obstet Gynecol
Table I. Distribution of amniotic fluid samples and true amniotic fluid mosaic cases at University of Washington and Swedish Medical Center between Jan. 1, 1984, and April 14, 1994
) Total AF samples Total AFM cases % of AFM cases No. of AFM cases excluded Final No. of AFM cases included Distribution of AFM cases Autosome (%) Sex chromosome (%)
SwedishMedicalCenter
I
UniversityofWashington ]
Total
19,710 60 0.30 12 48
8787 28 0.32 4 24
24,497 88 0.31 16 72
21 (44%, 21/48) 27 (56%, 27/48)
14 (58%, 14/24) 10 (42%, 10/24)
35 (49%, 35/72) 37 (51%, 37/72)
AF, Amniotic fluid; AFM, amniotic fluid mosaic. Table II. Classification and distribution of amniotic fluid mosaic cases (N = 72)
] Autosomal (total) Tri-21 Tri-20 Tri-18 Tri-14 Tri-8 Markers Isochromosomes* Sex chromosomes (total) X/XX X/XY XY/XYY XY/XXY XX/XXX Total No. of clinically significant AFMt % total AFM % total AF samples
SwedishMedicalCenter ]University of Washington 21 6 3 0 1 0 7 4 27 13 9 2 2 1 37 77% (37/48) 0.19% (37/19,710)
14 3 5 1 0 1 3 1 10 5 3 1 0 1 17 71 (17/24) 0.19% (17/8,787)
]
Total 35 9 8 1 1 1 10 5 37 18 12 3 2 2 54 75% (54/72) 0.19% (54/28,497)
AFM, Amniotic fluid mosaicism; AF, amniotic fluid. *Isochromosomes: 3 = i(20q), 1 = i(12p), 1 = idic(15). *Clinically significant mosaics include Tri-21, Tri-18, Tri-14, Tri-8, de novo markers, i(12p), idic(15), X/XX, X/XY.
ing true amniotic fluid mosaicism, we undertook a review of the prenatal cytogenetic experience of two large prenatal diagnostic centers in Washington over the past 10 years.
Material and methods T h e cytogenetic databases of the prenatal diagnosis programs at the University of Washington and Swedish Health Services, Seattle, Washington, were examined for all cases of true amniotic fluid mosaicism detected between Jan. 1, 1984, and April 14, 1994. All amniotic fluid samples were analyzed by the in situ method at both institutions. Backup flasks were used only when additional cells from another culture were needed to establish or exclude amniotic fluid mosaicism. True amniotic fluid mosaicism was defined as the presence of the same chromosomal abnormality in two or more different culture dishes. Cases were included in the study when they met the definition for true amniotic fluid mosaicism. Cases that were excluded from the study included (1) cases in
which the cytogenetics file was not available for review (n = 2), (2) cases in which the mosaicism was not the primary defect (n = 12), and (3) cases in which a parent had the same karyotype (n = 2). The remaining cases were reviewed for demographic, obstetric, and cytogenetic information, including the method of amniotic fluid analysis (flask, in situ method, or both), the chromosomal abnormality detected, the total n u m b e r of clones or cells examined, the proportion of abnormal clones present, and the clonal n u m b e r at which the abnormality was first detected and the clonal n u m b e r at which the second abnorfiaal clone was detected. Follow-up information was also reviewed whenever possible for confirmation of mosaicism and abnormal clinical outcome. Cases were analyzed for the proportion of aneuploidy involving sex chromosomes, autosomes, markers, and structural rearrangements and for the proportion representing potentially clinically significant and nonsignificant abnormalities. The classification of amniotic fluid mosaicism with potential clinical significance re-
Cheng et al.
Volume 173, Number 4 Am J Obstet Gynecol
(1)
100
n..
80
o
60
o
40
cO
..........Detected Rate - Significant AFM -- Detected Rate - Total AFM
e-"
o
1027
20 0 0
5
10
15
20
N u m b e r of Clones Analyzed Fig. 1. Cumulative amniotic fluid mosaicism detection rate based on number of clones analyzed. quired previously r e p o r t e d documentation of phenotypic abnormalities associated with a specific chromosome mosaicism, All cases involving autosomes (except for chromosome 20), markers, structural rearrangements, and sex chromosome aneuploidy involving 45,X/46,XX or 45,X/46,XY were considered to be potentially clinically significant. T h e r e were eight amniotic fluid mosaicism cases involving trisomy 20 that were not considered to be potentially clinically significant. Hsu et al. '° reviewed the outcome of 103 cases of trisomy 20 mosaicism identified at amniocentesis and found that 90% (90/103) of the cases were associated with a grossly normal phenotype. Of the 10 cases demonstrating an abnormal phenotype, one was also 45,X; one child was subsequently diagnosed with William syndrome; and one fetus died in utero because of Rh disease. Thus, although the clinical interpretation of trisomy 20 mosaicism continues to cause some anxiety in prenatal diagnosis counseling, the majority of cases involving trisomy 20 mosaicism in amniotic fluid appears to result in a normal outcome. Cases involving sex chromosome mosaicism 46,XY/47,XYY, 46,XY/47,XXY, and 46,XX/47,XXX were interpreted as not being clinically significant. The detection rates at 5, 10, 15, and 20 clones were calculated for all cases and for cases that were interpreted as being potentially clinically significant. Results
Between Jan. 1, 1984, and April 14, 1994, 28,497 amniotic fluid samples were examined at the University of Washington and Swedish Health Services. Table I summarizes the distribution of cases between the two centers and the distribution of chromosome abnormalities. Eighty-eight cases (0.31%) of true mosaicism were identified; 16 cases were excluded, resulting in 72 final cases (0.25%) for analysis. Forty-nine percent (35/72) of cases were related to mosaicism involving autosomes
(including markers) and 51% (37/72) to sex chromosome aneuploidy. Table II details the classification and distribution of amniotic fluid mosaicism cases. Fifty-four of 28,497 cases, representing an incidence of 0.19% for true amniotic fluid mosaicism, were interpreted as having potential clinical significance. This represented 75% (54/72) of cases of amniotic fluid mosaicism in this study. The number and distribution of cases (total and clinically significant) identified on the basis of clonal number at which the first abnormal clone was detected is illustrated in Fig. 1. The cumulative detection rates based on the identification of the first abnormal clone within the first 5, 10, 15, and 20 clones for all 72 cases of amniotic fluid mosaicism and for the 54 clinically significant cases are detailed in Table III and demonstrated graphically in Fig. 2. For both groups approximately 80% of cases of amniotic fluid mosaicism were identified by the fifth clone, 89% by the sixth clone, and 94% by the tenth clone. Fig. 3 summarizes the relationship between the degree of amniotic fluid mosaicism and the clonal number at which the first abnormal clones was detected. There were nine cases with potential clinical significance in which the first abnormal clone was not identified until after the fifth clone. These cases, outlined in Table IV, represent those with the lowest level of mosaicism. Five underwent termination of pregnancy, two were lost to follow-up, and two were delivered of phenotypically normal infants. Comment
T h e pressure to maximize the effect of each health care dollar has forced a reevaluation of many traditional tests and procedures. In prenatal cytogenetic diagnosis a balance must be maintained between the cost of ensuring a high rate of detected amniotic fluid mosaicism and the financial, medical, and emotional burden of missed abnormalities, procedural risks and
1028
Cheng et al.
October 1995 Am J Obstet Gynecol
30
25 O9
O9 20 O
+6
15
.Q
E
10
Z
5
0 0
5
10
15
20
Clone Number of Abnormality First Detected Fig. 2. Number and distribution of amniotic fluid mosaicism cases based on detection of first abnormal clone.
100 • Autosomal Mosaics - Significant
E
80
[] A u t o s o m a l M o s a i c s - Non-significant •
O3 O tO
o S e x M o s a i c s - Non-significant
"i
40
"
20
• S e x M o s a i c s - Significant
•
i,
60
,o
13.
•
I
"
....... |++....+|....+.=~ ............ ............ =...+.++.... •
+
e
O
+
[]
0 ..................................................................................................................... ~
[]
-
o
•
0
0
5
10
15
20
Clone Number of Abnormality First Detected Fig. 3. Percent amniotic fluid mosaicism versus detection of first abnormal clone.
false-positive diagnosis, and financial costs that society must bear. T h e results of the current study are similar to the results of previous studies of amniotic fluid mosaicism.l-3. 5.9 We have demonstrated that the detection of true amniotic fluid mosaicism could be accomplished by reducing the number of clones analyzed by the in situ m e t h o d with minimal loss of test sensitivity. If analysis were limited to the first five or six clones, the sensitivity for amniotic fluid mosaicism would have decreased from 95% to 80% or 89%, respectively, and would have resulted in the failure to detect approxi-
mately one case with potential clinical significance per 3 0 0 0 amniotic fluid samples (nine of 28,497) overall, or one of 9000 or 10,000 samples, respectively, for autosomes. If a 10-clone analysis had been done, t h e sensitivity would have increased to 93% to 94% (Table III). Of the nine cases in this study that would have been missed with the five-done protocol, only three cases involving autosomes (tri-8, tri-21, and de novo marker) would potentially have h a d the most clinical i m p a c t if continued to term. O f the remaining six cases, all were related to sex chromosome mosaicism and represented
Volume 173, Number 4 Am J Obstet Gynecol
Cheng et al. 1029
Table I l L Detection rates based on diagnosis made within first 5, 10, 15, and 20 clones for all amniotic fluid mosaic cases (N = 72) and for the 54 clinically significant cases
Clone No. All AFM (n = 72)
Clinically significant AFM (n = 54)
I
5 10 15 20 5 10 15 20
c~,s detected (No.) 57 67 70 72 45 51 52 54
]
Detectionrate (%) 79 93 97 100 83 94 96 100
(57/72) (67/72) (70/72) (45/54) (51/54) (52/54)
AFM, Amniotic fluid mosaicism.
Table IV. Cases with potential clinical significance in which first abnormal clone was identified after clone No. 5
J Swedish Medical Center (n = 5)
University of Washington (n = 4)
Karyotype
First abnormal clone Mosaicism (%)
X/XY X/XX X/XX X/XY X/XY X/XY XY/XY, + 8 XX/XX, + 21 XX/XX, + mark
6 9 15 16 18 6 6 8 9
34 8 11 7 12 20 21 19 15
Outcome VTOP Lost to follow-up (no VTOP) VTOP Normal newborn Normal newborn NA VTOP VTOP VTOP
VTOP, Voluntary termination of pregnancy; NA, not available.
very low levels of mosaicism. Two of these fetuses were phenotypically normal newborns. Hsu 11 reviewed the outcome of 601 cases of mosaicism obtained from a North American survey and found that only 10% of sex chromosome mosaics were associated with any phenotypic abnormalities. Ninety-five percent of 45,X/46,XY fetuses had normal male genitalia at birth and only eight of 98 cases of 45,X/46,XX females had some features of T u r n e r syndrome. Therefore it is most likely that the four remaining cases of sex chromosome mosaicism for which we did not have follow-up would have been clinically normal as well. If a 10-clone protocol had been used, five cases of amniotic fluid mosaicism would have been missed overall. Of these five cases, only three may have been clinically significant, but none involved autosomes (Tables III and IV). O n the basis of time-motion and efficiency studies done in our laboratory, we estimate that 15% to 20% of technical analysis time could be saved per case if the n u m b e r of clones analyzed were reduced to five to six clones. Although there would be no savings on supplies, equipment, or other fixed costs, cytogeneticist time, n u m b e r of subcultures, or time in culture, substantial technical (analysis) time could be saved, resulting in increased efficiency, decreased costs and turnaround time, and minimal loss of sensitivity. The decision to adopt a five-, six-, or 10-clone protocol is arbitrary and reflects the balance between the desired sensitivity and
laboratory costs. As the pressure to decrease costs and increase productivity continues, this approach may have advantages over other less expensive cytogenetic alternatives, such as single-cell fluorescent in situ hybridization, because the majority of cases of amniotic fluid mosaicism would still be detected. Large collaborative studies may be necessary to determine the optimal trade-off between sensitivity and costs. REFERENCES
1. Bui TH, Iselius L, Lindsten J. European collaborative study on prenatal diagnosis: mosaicism, pseudomosaicism and single abnormal cells in amniotic fluid cell cultures. Prenat Diagn 1984;4:145-62. 2. Hsu LYF. United States survey on chromosome mosaicism and pseudomosaicism in prenatal diagnosis. Prenat Diagn 1984;4:97-130. 3. Worton RG, Stern R. A Canadian collaborative study of mosaicism in amniotic fluid cell cultures. Prenat Diagn 1984;4:131-44. 4. Hook EB. Exclusion of chromosomal mosaicism: tables of 90%, 95%, and 99% confidence limits and comments on use. Am J Hum Genet 1977;29:94-7. 5. Richkind KE, Risch NJ. Sensitivity of chromosomal mosaicism detection by different tissue culture methods. Prenat Diagn 1990;10:519-27. 6. Featherstone T, Cheung SW, Spitznagel E, Peakman D. Exclusion of chromosomal mosaicism in amniotic fluid cultures: determination of number of colonies needed for accurate analysis. Prenat Diagn 1994;14:1009-17. 7. Claussen U, Schafer HH, Trampisch JJ. Exclusion of chromosomal mosaicism in prenatal diagnosis. Hum Genet 1984;67:23-8. 8. Cheung SW, Spitznagel E, Featherstone T, Crane JP.
Nyberg et al.
Exclusion of chromosomal mosaicism in amniotic fluid cultures: efficacy of in situ versus flask techniques. Prenat Diagn 1990;10:41-57. 9. Hsu LY, Kaffe S, Jenkins EC, et al. Proposed guidelines for diagnosis of chromosome mosaicism in amniocytes based on data derived from chromosome mosaicism and pseudomosaicism studies. Prenat Diagn 1992;12:555-73.
October 1995 Am J Obstet Gynecol
10. Hsu LYF, Kaffe S, Perlis TE. A revisit of trisomy 20 mosaicism in prenatal diagnosis-an overview of 103 cases. Prenat Diagn 1991;11:7-15. 11. Hsu LYF. Chromosomal disorders. In: Reece EA, Hobbins JC, Mahoney MJ, Petrie RH, eds. Principles and practice of fetal-maternal medicine. Philadelphia: JB Lippincott, 1992:433-40.
Role of prenatal ultrasonography in women with positive screen for Down syndrome on the basis of maternal serum markers David A. Nyberg, MD,= David A. Luthy, MD,= Edith Y. Cheng, MD,b Robert C. Sheley, MD, = Robert G. Resta, MS," and Michelle A. Williams, ScD"
Seattle, Washington OBJECTIVE: Our purpose was to evaluate the usefulness of prenatal ultrasonography among women with a positive screen for fetal Down syndrome on the basis of three biochemical markers-maternal serum ~-fetoprotein, human chorionic gonadotropin, and unconjugated estriol. STUDY DESIGN: A total of 395 women underwent prenatal ultrasonography at a single institution after being identified as screen positive (midtrimester risk -> 1 : 195) on the basis of triple-marker screening between 15 and 18 weeks. Ultrasonographic findings were compared with the biochemical markers and the eventual fetal outcome for these patients. Ultrasonographic abnormalities that were evaluated included structural defects, nuchal thickening or cystic hygroma, echogenic bowel, cerebral ventricular dilatation, pylectasis, and shortened femur. RESULTS: Among 395 patients, 374 (94.7%) had normal karyotype by genetic amniocentesis (n = 232) or postnatal follow-up (n = 142), 18 (4.5%) proved to have Down syndrome, and three had other karyotypic abnormalities. One or more ultrasonographic abnormalities were found in nine of 18 (50%) with Down syndrome compared to 27 of 377 (7.2%) other fetuses (p < 0.001). Fetuses with abnormal ultrasonography results included three with other chromosome abnormalities and five with nonchromosomal anomalies. An abnormal ultrasonography result increased the risk of Down syndrome by 5.6-fold (25% from 4.5%) and a negative result reduced the risk by 45% (2.5% from 4.5%). The value of uitrasonography is further enhanced when all chromosome abnormalities and nonchromosomal anomalies are considered. CONCLUSION: Abnormal ultrasonographic findings increase the risk for Down syndrome, whereas normal findings are less predictive of normalcy. After correction for inaccurate menstrual dates, genetic amniocentesis should be offered in spite of a normal ultrasonography result among women with positive triple screen. (AM J OBSTETGYNECOL1995;173:1030-5.)
Key words: Fetus, chromosome abnormalities, prenatal ultrasonography, prenatal diagnosis, Down syndrome, ct-fetoprotein, triple screen
Biochemical markers obtained from maternal serum during the second trimester can help detect fetal Down From the Center for Perinatal Studies, Swedish Hospital Medical Center,a and the Department of Obstetrics and Gynecology, University of Washington.b Presented in part at the FifteenthAnnual Meeting of the Society of Perinatal Obstetricians, Atlanta, Georgia, January 23-28, 1995. Reprint requests: David A. Nyberg, MD, Seattle Ultrasound Associates, 1229 Madison St., Suite 1150, Seattle, WA 98104. Copyright © 1995 by Mosby-Year Book, Inc. 0002-9378/95 $5.00 + 0 6/6/68589 1030
syndrome. An association with Down syndrome has been established for low et-fetoprotein (AFP)," 2 elevated h u m a n chorionic gonadotropin (hCG), 3' 4 and low estriol levels. ~ The combination of all three biochemical markers together with maternal age can be used to calculate a specific, individual risk for Down syndrome. Prenatal testing of women who are screen positive has proved to be more efficacious than screening based on maternal age alone. 69 With this approach various studies have reported sensitivities of 60% to 65% (range
Seattle, Washington OBJECTIVE: Our purpose was to evaluate the 15-clone analysis for detecting amniotic fluid mosaicism by the in situ method. STUDY DESIGN: A 10-year review was performed of all amniotic fluid mosaicism cases at two institutions using the in situ method exclusively, with sequential clonal analysis to determine the first and second clone in which the abnormal cell line occurred. RESULTS: Of the 28,497 amniotic fluid samples, 73 met criteria for amniotic fluid mosaicism by in situ method (0.26%). There were 54 cases (0.19%) with potential clinical significance (23 autosome and 31 sex chromosome mosaicism); 49 of the 54 cases (89%) were detected in the first six clones, including 22 of 23 involving autosomes and 27 of 31 involving sex chromosomes. In one of the six cases detected after clone 6 (46,XX/47,XX, +21) the mosaic cell line was present in 20% of the clones analyzed and was followed by a voluntary termination of the pregnancy, in the other five cases amniotic fluid mosaicism was present in < 20% of the clones; these included one case of 46,XX/47,XX + mar (15% amniotic fluid mosaicism, voluntary termination of pregnancy), two cases of 45,X]46,XY (10% to 12% amniotic fluid mosaicism, both normal at birth), and two cases of 45,W46,XX (8% amniotic fluid mosaicism, lost to follow-up; 12% amniotic fluid mosaicism, voluntary termination of pregnancy). By limiting the analysis to six clones, - 2 0 % of analysis time could be saved per case, but one autosomal amniotic fluid mosaicism case per 10,000 samples could potentially be missed. CONCLUSION: Reducing the number of clones analyzed by in situ method could result in increased efficiency, decreased costs, and minimal loss of sensitivity. (AM J OBSTETGYNECOL1995; 173:1025-30.)
Key words: Amniotic fluid mosaicism, in situ method, 15-clone count
In prenatal cytogenetic diagnosis true mosaicism in amniotic fluid is rare, occurring in approximately 0.1% to 0.3% of samples. "~ However, the efficient diagnosis and interpretation of true amniotic fluid mosaicism remain a challenge. In 1977 Hook ~ determined the n u m b e r of cells to be examined for detection of 10%, 20%, and 30% mosaicism in peripheral blood at 90%, 95%, and 99% confidence levels. This model was extended to the analysis of amniotic fluid and reevaluated by Richkind and Risch 5 and Featherstone et al. ~ to apply to the flask and in situ methods. For the flask method the current practice is to examine a m i n i m u m of 20 cells from two flasks. True amniotic fluid mosaicism is diagnosed when two or more cells with the same
From the Center for Perinatal Studies, Swedish Health Services," and the Departments of Obstetrics and Gynecologyb and Pathology,' University of Washington Medical Center. Partially funded as a special project of the Pacific Northwest Regional Genetics Group by Project No. MCJ411002-11 of the Maternal and Child Health Bureau, Department of Health and Human Se~'oices. Presented in part at the Fifteenth Annual Meeting of the Society of Perinatal Obstetricians, Atlanta, Georgia, January 23-28, 1995. Reprint requests: David A. Luthy, MD, Division of Perinatal Medicine, 747 Broadway, Seattle, WA 98104. Copyright © 1995 by Mosby-Year Book, Inc. 0002-9378/95 $5.00 + 0 6/6/66695
cytogenetic abnormality are found in two or more flasks. For the in situ method one cell each from 15 to 20 clones from multiple culture chambers are analyzed. True mosaicism is defined as the presence of two or more clones with the same abnormality from different culture chambers. These guidelines permit detection of mosaicism at the 20% level or greater with a confidence level of 95% to 99%. 5 Although the flask and in situ methods have been shown to be equally sensitive in detecting true mosaicism, 7' s there continues to be discussion regarding the necessity of additional cytologic evaluations to arrive at an accurate diagnosis and interpretation when true mosaicism is encountered. Hsu et al. ~ specifically addressed this question by reviewing the 50 cases of true mosaicism among 22,000 amniotic fluid samples examined by the Prenatal Diagnosis Laboratory of New York City and concluded that not all cases of true mosaicism required an exhaustive workup for diagnosis and interpretation. Featherstone et al." arrived at similar conclusions that a two-stage approach, dictated by the specific chromosome and its associated phenotype, decreased the amount of unnecessary laboratory effort and was both efficient and cost effective. To further investigate the efficiency of the 15-clone in situ method for detect1025
1026 Cheng et al.
October 1995 Am J Obstet Gynecol
Table I. Distribution of amniotic fluid samples and true amniotic fluid mosaic cases at University of Washington and Swedish Medical Center between Jan. 1, 1984, and April 14, 1994
) Total AF samples Total AFM cases % of AFM cases No. of AFM cases excluded Final No. of AFM cases included Distribution of AFM cases Autosome (%) Sex chromosome (%)
SwedishMedicalCenter
I
UniversityofWashington ]
Total
19,710 60 0.30 12 48
8787 28 0.32 4 24
24,497 88 0.31 16 72
21 (44%, 21/48) 27 (56%, 27/48)
14 (58%, 14/24) 10 (42%, 10/24)
35 (49%, 35/72) 37 (51%, 37/72)
AF, Amniotic fluid; AFM, amniotic fluid mosaic. Table II. Classification and distribution of amniotic fluid mosaic cases (N = 72)
] Autosomal (total) Tri-21 Tri-20 Tri-18 Tri-14 Tri-8 Markers Isochromosomes* Sex chromosomes (total) X/XX X/XY XY/XYY XY/XXY XX/XXX Total No. of clinically significant AFMt % total AFM % total AF samples
SwedishMedicalCenter ]University of Washington 21 6 3 0 1 0 7 4 27 13 9 2 2 1 37 77% (37/48) 0.19% (37/19,710)
14 3 5 1 0 1 3 1 10 5 3 1 0 1 17 71 (17/24) 0.19% (17/8,787)
]
Total 35 9 8 1 1 1 10 5 37 18 12 3 2 2 54 75% (54/72) 0.19% (54/28,497)
AFM, Amniotic fluid mosaicism; AF, amniotic fluid. *Isochromosomes: 3 = i(20q), 1 = i(12p), 1 = idic(15). *Clinically significant mosaics include Tri-21, Tri-18, Tri-14, Tri-8, de novo markers, i(12p), idic(15), X/XX, X/XY.
ing true amniotic fluid mosaicism, we undertook a review of the prenatal cytogenetic experience of two large prenatal diagnostic centers in Washington over the past 10 years.
Material and methods T h e cytogenetic databases of the prenatal diagnosis programs at the University of Washington and Swedish Health Services, Seattle, Washington, were examined for all cases of true amniotic fluid mosaicism detected between Jan. 1, 1984, and April 14, 1994. All amniotic fluid samples were analyzed by the in situ method at both institutions. Backup flasks were used only when additional cells from another culture were needed to establish or exclude amniotic fluid mosaicism. True amniotic fluid mosaicism was defined as the presence of the same chromosomal abnormality in two or more different culture dishes. Cases were included in the study when they met the definition for true amniotic fluid mosaicism. Cases that were excluded from the study included (1) cases in
which the cytogenetics file was not available for review (n = 2), (2) cases in which the mosaicism was not the primary defect (n = 12), and (3) cases in which a parent had the same karyotype (n = 2). The remaining cases were reviewed for demographic, obstetric, and cytogenetic information, including the method of amniotic fluid analysis (flask, in situ method, or both), the chromosomal abnormality detected, the total n u m b e r of clones or cells examined, the proportion of abnormal clones present, and the clonal n u m b e r at which the abnormality was first detected and the clonal n u m b e r at which the second abnorfiaal clone was detected. Follow-up information was also reviewed whenever possible for confirmation of mosaicism and abnormal clinical outcome. Cases were analyzed for the proportion of aneuploidy involving sex chromosomes, autosomes, markers, and structural rearrangements and for the proportion representing potentially clinically significant and nonsignificant abnormalities. The classification of amniotic fluid mosaicism with potential clinical significance re-
Cheng et al.
Volume 173, Number 4 Am J Obstet Gynecol
(1)
100
n..
80
o
60
o
40
cO
..........Detected Rate - Significant AFM -- Detected Rate - Total AFM
e-"
o
1027
20 0 0
5
10
15
20
N u m b e r of Clones Analyzed Fig. 1. Cumulative amniotic fluid mosaicism detection rate based on number of clones analyzed. quired previously r e p o r t e d documentation of phenotypic abnormalities associated with a specific chromosome mosaicism, All cases involving autosomes (except for chromosome 20), markers, structural rearrangements, and sex chromosome aneuploidy involving 45,X/46,XX or 45,X/46,XY were considered to be potentially clinically significant. T h e r e were eight amniotic fluid mosaicism cases involving trisomy 20 that were not considered to be potentially clinically significant. Hsu et al. '° reviewed the outcome of 103 cases of trisomy 20 mosaicism identified at amniocentesis and found that 90% (90/103) of the cases were associated with a grossly normal phenotype. Of the 10 cases demonstrating an abnormal phenotype, one was also 45,X; one child was subsequently diagnosed with William syndrome; and one fetus died in utero because of Rh disease. Thus, although the clinical interpretation of trisomy 20 mosaicism continues to cause some anxiety in prenatal diagnosis counseling, the majority of cases involving trisomy 20 mosaicism in amniotic fluid appears to result in a normal outcome. Cases involving sex chromosome mosaicism 46,XY/47,XYY, 46,XY/47,XXY, and 46,XX/47,XXX were interpreted as not being clinically significant. The detection rates at 5, 10, 15, and 20 clones were calculated for all cases and for cases that were interpreted as being potentially clinically significant. Results
Between Jan. 1, 1984, and April 14, 1994, 28,497 amniotic fluid samples were examined at the University of Washington and Swedish Health Services. Table I summarizes the distribution of cases between the two centers and the distribution of chromosome abnormalities. Eighty-eight cases (0.31%) of true mosaicism were identified; 16 cases were excluded, resulting in 72 final cases (0.25%) for analysis. Forty-nine percent (35/72) of cases were related to mosaicism involving autosomes
(including markers) and 51% (37/72) to sex chromosome aneuploidy. Table II details the classification and distribution of amniotic fluid mosaicism cases. Fifty-four of 28,497 cases, representing an incidence of 0.19% for true amniotic fluid mosaicism, were interpreted as having potential clinical significance. This represented 75% (54/72) of cases of amniotic fluid mosaicism in this study. The number and distribution of cases (total and clinically significant) identified on the basis of clonal number at which the first abnormal clone was detected is illustrated in Fig. 1. The cumulative detection rates based on the identification of the first abnormal clone within the first 5, 10, 15, and 20 clones for all 72 cases of amniotic fluid mosaicism and for the 54 clinically significant cases are detailed in Table III and demonstrated graphically in Fig. 2. For both groups approximately 80% of cases of amniotic fluid mosaicism were identified by the fifth clone, 89% by the sixth clone, and 94% by the tenth clone. Fig. 3 summarizes the relationship between the degree of amniotic fluid mosaicism and the clonal number at which the first abnormal clones was detected. There were nine cases with potential clinical significance in which the first abnormal clone was not identified until after the fifth clone. These cases, outlined in Table IV, represent those with the lowest level of mosaicism. Five underwent termination of pregnancy, two were lost to follow-up, and two were delivered of phenotypically normal infants. Comment
T h e pressure to maximize the effect of each health care dollar has forced a reevaluation of many traditional tests and procedures. In prenatal cytogenetic diagnosis a balance must be maintained between the cost of ensuring a high rate of detected amniotic fluid mosaicism and the financial, medical, and emotional burden of missed abnormalities, procedural risks and
1028
Cheng et al.
October 1995 Am J Obstet Gynecol
30
25 O9
O9 20 O
+6
15
.Q
E
10
Z
5
0 0
5
10
15
20
Clone Number of Abnormality First Detected Fig. 2. Number and distribution of amniotic fluid mosaicism cases based on detection of first abnormal clone.
100 • Autosomal Mosaics - Significant
E
80
[] A u t o s o m a l M o s a i c s - Non-significant •
O3 O tO
o S e x M o s a i c s - Non-significant
"i
40
"
20
• S e x M o s a i c s - Significant
•
i,
60
,o
13.
•
I
"
....... |++....+|....+.=~ ............ ............ =...+.++.... •
+
e
O
+
[]
0 ..................................................................................................................... ~
[]
-
o
•
0
0
5
10
15
20
Clone Number of Abnormality First Detected Fig. 3. Percent amniotic fluid mosaicism versus detection of first abnormal clone.
false-positive diagnosis, and financial costs that society must bear. T h e results of the current study are similar to the results of previous studies of amniotic fluid mosaicism.l-3. 5.9 We have demonstrated that the detection of true amniotic fluid mosaicism could be accomplished by reducing the number of clones analyzed by the in situ m e t h o d with minimal loss of test sensitivity. If analysis were limited to the first five or six clones, the sensitivity for amniotic fluid mosaicism would have decreased from 95% to 80% or 89%, respectively, and would have resulted in the failure to detect approxi-
mately one case with potential clinical significance per 3 0 0 0 amniotic fluid samples (nine of 28,497) overall, or one of 9000 or 10,000 samples, respectively, for autosomes. If a 10-clone analysis had been done, t h e sensitivity would have increased to 93% to 94% (Table III). Of the nine cases in this study that would have been missed with the five-done protocol, only three cases involving autosomes (tri-8, tri-21, and de novo marker) would potentially have h a d the most clinical i m p a c t if continued to term. O f the remaining six cases, all were related to sex chromosome mosaicism and represented
Volume 173, Number 4 Am J Obstet Gynecol
Cheng et al. 1029
Table I l L Detection rates based on diagnosis made within first 5, 10, 15, and 20 clones for all amniotic fluid mosaic cases (N = 72) and for the 54 clinically significant cases
Clone No. All AFM (n = 72)
Clinically significant AFM (n = 54)
I
5 10 15 20 5 10 15 20
c~,s detected (No.) 57 67 70 72 45 51 52 54
]
Detectionrate (%) 79 93 97 100 83 94 96 100
(57/72) (67/72) (70/72) (45/54) (51/54) (52/54)
AFM, Amniotic fluid mosaicism.
Table IV. Cases with potential clinical significance in which first abnormal clone was identified after clone No. 5
J Swedish Medical Center (n = 5)
University of Washington (n = 4)
Karyotype
First abnormal clone Mosaicism (%)
X/XY X/XX X/XX X/XY X/XY X/XY XY/XY, + 8 XX/XX, + 21 XX/XX, + mark
6 9 15 16 18 6 6 8 9
34 8 11 7 12 20 21 19 15
Outcome VTOP Lost to follow-up (no VTOP) VTOP Normal newborn Normal newborn NA VTOP VTOP VTOP
VTOP, Voluntary termination of pregnancy; NA, not available.
very low levels of mosaicism. Two of these fetuses were phenotypically normal newborns. Hsu 11 reviewed the outcome of 601 cases of mosaicism obtained from a North American survey and found that only 10% of sex chromosome mosaics were associated with any phenotypic abnormalities. Ninety-five percent of 45,X/46,XY fetuses had normal male genitalia at birth and only eight of 98 cases of 45,X/46,XX females had some features of T u r n e r syndrome. Therefore it is most likely that the four remaining cases of sex chromosome mosaicism for which we did not have follow-up would have been clinically normal as well. If a 10-clone protocol had been used, five cases of amniotic fluid mosaicism would have been missed overall. Of these five cases, only three may have been clinically significant, but none involved autosomes (Tables III and IV). O n the basis of time-motion and efficiency studies done in our laboratory, we estimate that 15% to 20% of technical analysis time could be saved per case if the n u m b e r of clones analyzed were reduced to five to six clones. Although there would be no savings on supplies, equipment, or other fixed costs, cytogeneticist time, n u m b e r of subcultures, or time in culture, substantial technical (analysis) time could be saved, resulting in increased efficiency, decreased costs and turnaround time, and minimal loss of sensitivity. The decision to adopt a five-, six-, or 10-clone protocol is arbitrary and reflects the balance between the desired sensitivity and
laboratory costs. As the pressure to decrease costs and increase productivity continues, this approach may have advantages over other less expensive cytogenetic alternatives, such as single-cell fluorescent in situ hybridization, because the majority of cases of amniotic fluid mosaicism would still be detected. Large collaborative studies may be necessary to determine the optimal trade-off between sensitivity and costs. REFERENCES
1. Bui TH, Iselius L, Lindsten J. European collaborative study on prenatal diagnosis: mosaicism, pseudomosaicism and single abnormal cells in amniotic fluid cell cultures. Prenat Diagn 1984;4:145-62. 2. Hsu LYF. United States survey on chromosome mosaicism and pseudomosaicism in prenatal diagnosis. Prenat Diagn 1984;4:97-130. 3. Worton RG, Stern R. A Canadian collaborative study of mosaicism in amniotic fluid cell cultures. Prenat Diagn 1984;4:131-44. 4. Hook EB. Exclusion of chromosomal mosaicism: tables of 90%, 95%, and 99% confidence limits and comments on use. Am J Hum Genet 1977;29:94-7. 5. Richkind KE, Risch NJ. Sensitivity of chromosomal mosaicism detection by different tissue culture methods. Prenat Diagn 1990;10:519-27. 6. Featherstone T, Cheung SW, Spitznagel E, Peakman D. Exclusion of chromosomal mosaicism in amniotic fluid cultures: determination of number of colonies needed for accurate analysis. Prenat Diagn 1994;14:1009-17. 7. Claussen U, Schafer HH, Trampisch JJ. Exclusion of chromosomal mosaicism in prenatal diagnosis. Hum Genet 1984;67:23-8. 8. Cheung SW, Spitznagel E, Featherstone T, Crane JP.
Nyberg et al.
Exclusion of chromosomal mosaicism in amniotic fluid cultures: efficacy of in situ versus flask techniques. Prenat Diagn 1990;10:41-57. 9. Hsu LY, Kaffe S, Jenkins EC, et al. Proposed guidelines for diagnosis of chromosome mosaicism in amniocytes based on data derived from chromosome mosaicism and pseudomosaicism studies. Prenat Diagn 1992;12:555-73.
October 1995 Am J Obstet Gynecol
10. Hsu LYF, Kaffe S, Perlis TE. A revisit of trisomy 20 mosaicism in prenatal diagnosis-an overview of 103 cases. Prenat Diagn 1991;11:7-15. 11. Hsu LYF. Chromosomal disorders. In: Reece EA, Hobbins JC, Mahoney MJ, Petrie RH, eds. Principles and practice of fetal-maternal medicine. Philadelphia: JB Lippincott, 1992:433-40.
Role of prenatal ultrasonography in women with positive screen for Down syndrome on the basis of maternal serum markers David A. Nyberg, MD,= David A. Luthy, MD,= Edith Y. Cheng, MD,b Robert C. Sheley, MD, = Robert G. Resta, MS," and Michelle A. Williams, ScD"
Seattle, Washington OBJECTIVE: Our purpose was to evaluate the usefulness of prenatal ultrasonography among women with a positive screen for fetal Down syndrome on the basis of three biochemical markers-maternal serum ~-fetoprotein, human chorionic gonadotropin, and unconjugated estriol. STUDY DESIGN: A total of 395 women underwent prenatal ultrasonography at a single institution after being identified as screen positive (midtrimester risk -> 1 : 195) on the basis of triple-marker screening between 15 and 18 weeks. Ultrasonographic findings were compared with the biochemical markers and the eventual fetal outcome for these patients. Ultrasonographic abnormalities that were evaluated included structural defects, nuchal thickening or cystic hygroma, echogenic bowel, cerebral ventricular dilatation, pylectasis, and shortened femur. RESULTS: Among 395 patients, 374 (94.7%) had normal karyotype by genetic amniocentesis (n = 232) or postnatal follow-up (n = 142), 18 (4.5%) proved to have Down syndrome, and three had other karyotypic abnormalities. One or more ultrasonographic abnormalities were found in nine of 18 (50%) with Down syndrome compared to 27 of 377 (7.2%) other fetuses (p < 0.001). Fetuses with abnormal ultrasonography results included three with other chromosome abnormalities and five with nonchromosomal anomalies. An abnormal ultrasonography result increased the risk of Down syndrome by 5.6-fold (25% from 4.5%) and a negative result reduced the risk by 45% (2.5% from 4.5%). The value of uitrasonography is further enhanced when all chromosome abnormalities and nonchromosomal anomalies are considered. CONCLUSION: Abnormal ultrasonographic findings increase the risk for Down syndrome, whereas normal findings are less predictive of normalcy. After correction for inaccurate menstrual dates, genetic amniocentesis should be offered in spite of a normal ultrasonography result among women with positive triple screen. (AM J OBSTETGYNECOL1995;173:1030-5.)
Key words: Fetus, chromosome abnormalities, prenatal ultrasonography, prenatal diagnosis, Down syndrome, ct-fetoprotein, triple screen
Biochemical markers obtained from maternal serum during the second trimester can help detect fetal Down From the Center for Perinatal Studies, Swedish Hospital Medical Center,a and the Department of Obstetrics and Gynecology, University of Washington.b Presented in part at the FifteenthAnnual Meeting of the Society of Perinatal Obstetricians, Atlanta, Georgia, January 23-28, 1995. Reprint requests: David A. Nyberg, MD, Seattle Ultrasound Associates, 1229 Madison St., Suite 1150, Seattle, WA 98104. Copyright © 1995 by Mosby-Year Book, Inc. 0002-9378/95 $5.00 + 0 6/6/68589 1030
syndrome. An association with Down syndrome has been established for low et-fetoprotein (AFP)," 2 elevated h u m a n chorionic gonadotropin (hCG), 3' 4 and low estriol levels. ~ The combination of all three biochemical markers together with maternal age can be used to calculate a specific, individual risk for Down syndrome. Prenatal testing of women who are screen positive has proved to be more efficacious than screening based on maternal age alone. 69 With this approach various studies have reported sensitivities of 60% to 65% (range