Pilot Studies for Proficiency Testing Using Fluorescence In Situ Hybridization With Chromosome-specific DNA Probes

Abstract

Fluorescence in situ hybridization using chromosome-specific DNA probes is rapidly becoming part of clinical laboratory practice for certain congenital and neoplastic disorders. Current legislation requires proficiency testing for clinical laboratory studies. To evaluate the efficacy of fluorescence in situ hybridization proficiency testing, we invited 19 representative institutions to participate in three pilot studies. One study used probes for the X and Y chromosomes to evaluate metaphase spreads and interphase nuclei. Another study used probes for bcr and abl to detect bcr/abl fusion in interphase nuclei in chronic myelogenous leukemia. The third study used a D22S75 probe to detect microdeletions in metaphase spreads from a patient with velocardiofacial syndrome. The results of these studies demonstrate that proficiency testing with fluorescence in situ hybridization is attainable using either metaphase or interphase preparations, and that either microscope slides or fixed cell pellets are suitable.

Fluorescence in situ hybridization (FISH) with chromosome-specific DNA probes is rapidly becoming part of accepted clinical practice. Fluorescence in situ hybridization can be used to study cells from various tissues in either metaphase or interphase and can be done in 1 to 2 days. With locus-specific probes, FISH can be performed to detect microdeletions in congenital disorders, such as Prader-Willi, Angelman, Williams, velocardiofacial, and other microdeletion syndromes.^1–4 Fluorescence in situ hybridization can be performed with whole-chromosome DNA libraries to study structurally abnormal chromosomes, and probes for the centromeric regions of chromosomes can detect aneuploidy in interphase and metaphase cells.^5–7 Certain DNA probes can be used with FISH to detect rearranged oncogenes in neoplastic disorders, such as chronic myelogenous leukemia and acute promyelocytic leukemia.^8,9

Because the application of FISH in clinical practice has advanced rapidly, the International Standing Committee on Human Cytogenetic Nomenclature recently developed and published standard nomenclature for reporting results of this technology.¹⁰ In addition, the American College of Medical Genetics (ACMG) has produced guidelines for validation of FISH probes and procedures in clinical practice.¹¹ Although many publications demonstrate the accuracy of FISH, the Food and Drug Administration has approved only three FISH tests for clinical practice.

No accrediting agency has formally begun proficiency testing for FISH procedures. In 1997, the Cytogenetics Resource Committee of the College of American Pathologists (CAP) and ACMG plans to provide the first nationally available proficiency test for clinical laboratories performing FISH procedures. The CAP/ACMG committee anticipates that approximately 150 laboratories will participate in the initial proficiency test. To simulate clinical practice, the proficiency test will be designed to permit participants to obtain their own DNA probes and to use their laboratories’ own standard procedures to process and score test specimens. Objective methods of recording and interpreting the results of the proficiency test will be used for ease of computer analysis and interlaboratory comparisons of data.

To determine whether proficiency testing is possible for FISH, we designed three pilot studies and invited representative laboratory directors (piloteers) to participate in each. In the first pilot study, DNA probes for the X and Y chromosomes were used to study both interphase nuclei and metaphase spreads in peripheral blood specimens. In the second pilot study, probes for bcr and abl sequences were used to detect malignant interphase nuclei in chronic myelogenous leukemia bone marrow. In the third pilot study, a locus-specific DNA probe for a chromosome 22 sequence was used to detect a microdeletion associated with velocardiofacial syndrome in metaphase spreads from peripheral blood. This report describes data from these pilot studies and evaluates the efficacy of proficiency testing for FISH procedures with chromosome-specific DNA probes.

MATERIALS AND METHODS

Proficiency test pilot materials consisted of fixed cells or microscope slides with fixed cells that were prepared by using standard cytogenetic methods. Each specimen for proficiency test pilot materials was processed by a central laboratory using standard culture and harvesting procedures for subsequent chromosome or FISH analysis.^3,7,8 To aid in the preparation of slides, a CDS-5 Cytogenetic Drying Chamber (Thermotron, Holland, Mich) was used.¹² Slides or cell pellets were coded for each pilot study so that piloteers did not know the condition or sex of the individuals from whom the specimens were collected. Piloteers obtained DNA probes and used their routine hybridization and scoring methods for FISH. After each study, piloteers were given a summary of all results to permit comparisons of their findings with those of others.

Pilot I

This study evaluated FISH with probes for the X and Y chromosomes and was designed to determine whether proficiency testing was best done using fixed cells or microscope slide preparations. This study involved 12 piloteers from different institutions and was divided into two parts with six piloteers in each. Each piloteer received either an XX or an XY preparation and a mixture of XX and XY cells. The XX or XY preparations and the XX/XY cell mixtures were prepared from fixed cells from phytohemagglutinin-stimulated cultures of peripheral blood from a normal female and a normal male.

Piloteers in part I received two microscope slide preparations and two fixed cell pellets. Slide A and pellet A were mixtures of approximately 50% XX and 50% XY cells. Slide B and pellet B were XY cells only. Piloteers in part II received two slides and two pellets. Slide A and pellet A were XX cells only; slide B and pellet B were mixtures of approximately 80% XX and 20% XY cells. For each specimen (slide and pellet), piloteers were asked to score 100 interphase nuclei and 20 metaphase spreads using probes for the X and Y chromosomes.

Materials were mailed on October 2, 1995, and the piloteers were asked to return their data on supplied forms by October 20, 1995.

Pilot II

This study evaluated FISH in interphase nuclei using probes for bcr and abl. Initially, 16 piloteers volunteered for this project, but one was unable to complete the study. Only microscope slide preparations were sent to piloteers in pilot II.

A central laboratory collected cells from apheresed specimens from two patients. According to clinical workup, routine cytogenetic evaluation, and FISH analysis with probes for bcr and abl, one of the patients had Philadelphia chromosome (Ph)-positive chronic myelogenotis leukemia, and the other had acute leukemia that was Ph-negative. For each piloteer, two slides (one was a backup slide) were prepared for each of three specimens; the slides were labeled A, B, and C, respectively. Slide A had approximately 20% cells from the Ph-positive specimen and 80% cells from the Ph-negative specimen. Slide B had only cells from the Ph-negative specimen. Slide C had only cells from the Ph-positive specimen.

Each piloteer was asked to analyze 200 consecutive, scorable, interphase nuclei from each specimen using FISH and probes for bcr and abl. Piloteers were asked to provide the percentage of cells with apparent bcr/abl fusion for each specimen and to interpret the possible clinical significance of their findings. Piloteers were also asked to complete a questionnaire regarding their experience with FISH, whether they had established normal and abnormal ranges for bcr/abl, the source of their probes, and whether they had a quality control program for bcr/abl analysis.

Slides were mailed on January 16, 1996, and the piloteers were asked to return their data on supplied forms by February 2, 1996.

Pilot III

This study evaluated FISH on metaphase spreads using probes for the detection of the D22S75 locus. This probe can be used to detect microdeletions in chromosome 22 of patients with Di-George sequence, velocardiofacial syndrome, and related diseases.^3,4,13 Fifteen piloteers participated in this study.

A central laboratory collected peripheral blood from a normal individual (specimen 1) and from a patient with a microdeletion of chromosome 22 that included the D22S75 locus (specimen 2). For each specimen, two slides were prepared for each piloteer. The slides for specimen 1 were labeled 1A and 1B; the slides for specimen 2 were labeled 2A and 2B.

Each piloteer was asked to use FISH with a probe for the D22S75 locus to analyze 20 consecutive, scorable, metaphase spreads from slide A of each specimen. Piloteers were instructed to use slide B of either specimen only if slide A was unsatisfactory.

Each piloteer was asked to report the number of metaphase spreads they scored and the number of cells with a D22S75 signal pattern consistent with normal or microdeletion. Based on these data, the percentage of cells for a normal or microdeletion D22S75 signal pattern was calculated. Piloteers were asked to complete a questionnaire regarding their experience with FISH, the source of their probes, whether they have established normal and abnormal ranges for D22S75 probes, and whether they had a quality control program for D22S75 analysis.

Slides for this pilot were mailed on May 6, 1996, and the piloteers were asked to return their data on supplied forms by May 24, 1996.

RESULTS

Pilot I

Part I

Six piloteers participated in part I; five used dual-colored probes (the X probe was red, and the Y probe was green) to score both X and Y chromosomes in each cell. Each of these five piloteers met the goal to report on 100 interphase nuclei and 20 metaphase spreads on slide A, pellet A, slide B, and pellet B (Table 1).

Table 1.

Results of Pilot Study I Using Probes for X and Y Chromosomes, Part I^*

Piloteer	FISH Strategy	Interphase			Metaphase			Interphase			Metaphase
		% XX	% XY	% Other	% XX	% XY	% Other	% XX	% XY	%Other	% XX	% XY	%Other
		Slide A, XX and XY Mixture						Pellet A, XX and XY Mixture
1	Dual	49	49	2	25	75	0	42	57	1	35	65	0
3	Dual	48	52	0	40	60	0	53	45	2	25	75	0
4	Dual	46	50	4	15	80	5	48	51	1	20	80	0
5	Dual	38	62	0	40	60	0	35	65	0	30	70	0
6	Dual	36	54	10	15	85	0	47	45	8	35	60	5
2	Single	19	20	61	ND	50	NA	13	39	48	9	70	21
		Slide B, XY Cells Only						Pellet B, XY Cells Only
1	Dual	1	99	0	0	100	0	0	100	0	0	100	0
3	Dual	0	100	0	0	100	0	0	100	0	0	100	0
4	Dual	0	99	1	0	100	0	0	99	1	0	100	0
5	Dual	0	100	0	0	100	0	0	100	0	0	100	0
6	Dual	4	89	7	0	100	0	0	98	2	0	95	5
2	Single	ND	23	NA	ND	10	NA	30	70	0	0	80	20

^*FISH indicates fluorescence in situ hybridization; ND, not done; and NA, not applicable.

Among the piloteers using dual-colored probes, the range in percentage of XX interphase nuclei on slide A and pellet A was 36% to 49% and 35% to 53%, respectively. By comparison, the range in percentage of XX metaphase spreads on slide A and pellet A was 15% to 40% and 20% to 35%, respectively. The range in percentage of XY interphase nuclei for slide B and pellet B was 89% to 100% and 98% to 100%, respectively. By comparison, the percentage of XY metaphase spreads for slide B was 100% for each piloteer, but ranged from 95% to 100% for pellet B.

For slide A, piloteers 1, 4, and 6 reported 21% to 31% more XX cells than XY cells in interphase nuclei than in metaphase spreads. Piloteer 3 reported 8% more XX cells than XY cells in interphase than in metaphase. Piloteer 5 reported 2% fewer XX cells than XY cells in interphase than metaphase.

In part I, variation among piloteers was mostly due to cells with signal patterns other than XX or XY. The percentage of non-XX/XY cells for slide A and pellet A interphase nuclei was 0% to 10% and 0% to 8%, respectively, and for metaphase spreads was 0% to 5% and 0% to 5%, respectively. The percentage of non-XX/XY cells for slide B and pellet B interphase nuclei ranged from 0% to 7% and 0% to 2%, respectively, and for metaphase spreads was 0% for slide B and ranged from 0% to 5% in pellet B.

Piloteer 2 used X and Y probes separately and thus could not report the number of X and Y signals in the same cells. The results of this piloteer were incomplete. For the X chromosome probe, no metaphase spreads were reported for either the A or B slides, and no interphase nuclei were reported for slide B.

Part II

Three piloteers used different colored X and Y probes; the other three piloteers used X and Y probes separately. Each piloteer who used dual-colored X and Y probes achieved the goal to report on 100 interphase nuclei and 20 metaphase spreads on slide A, pellet A, slide B, and pellet B (Table 2).

Table 2.

Results of Pilot Study I Using Probes for X and Y Chromosomes, Part II^*

Piloteer	FISH Strategy	Interphase			Metaphase			Interphase			Metaphase
		% XX	% XY	% Other	% XX	% XY	% Other	% XX	% XY	% Other	% XX	% XY	% Other
		Slide A, XX Cell Only						Pellet A, XX Cells Only
7	Dual	91	0	9	95	0	5	81	0	19	95	0	5
8	Dual	98	0	2	100	0	0	65	0	35	90	0	10
9	Dual	95	0	5	90	5	5	97	0	3	95	0	5
10	Single	93	0	NA	100	0	NA	NR	NR	NR	NR	NR	NR
11	Single	92	0	NA	100	0	NA	80	0	NA	95	0	NA
12	Single	93	0	NA	100	0	NA	96	0	NA	100	0	NA
		Slide B, XX and XY Mixture						Pellet B, XX and XY Mixture
7	Dual	74	13	13	70	20	10	75	15	10	70	30	0
8	Dual	73	22	5	65	30	5	70	25	5	75	25	0
9	Dual	74	23	3	65	35	0	85	15	0	75	25	0
10	Single	93	0	NA	100	0	NA	NR	NR	NR	NR	NR	NR
11	Single	0	45	NA	0	20	NA	85	53	NA	75	25	NA
12	Single	80	19	NA	60	40	NA	80	20	NA	80	50	NA

^*FISH indicates fluorescence in situ hybridization; NA, not applicable; and NR, no results.

Among piloteers using dual-colored probes, the percentage of XX interphase nuclei for slide A and pellet A ranged from 91% to 98% and 65% to 97%, respectively. By comparison, the percentage of XX metaphases for slide A and pellet A ranged from 90% to 100% and 90% to 95%, respectively. The percentage of XX interphase nuclei on slide B and pellet B ranged from 73% to 74% and 70% to 85%, respectively. The percentage of XX metaphases on slide B and pellet B ranged from 65% to 70% and 70% to 75%, respectively.

For slide B, piloteers 7, 8, and 9 reported 4% to 9% more XX cells than XY cells in interphase than in metaphase.

In part II, variation among piloteers was mostly due to cells with signal patterns other than XX or XY. The range in percentage of non-XX/XY cells for slide A and pellet A interphase nuclei was 2% to 9% and 3% to 35%, respectively, and for metaphases was 0% to 5% and 5% to 10%, respectively. The range in percentage of non-XX/XY cells for slide B and pellet B interphase nuclei was 3% to 13% and 0% to 10%, respectively, and for metaphases was 0% for each piloteer for pellet B.

Three piloteers in part II used X and Y probes singly. The results were generally problematic. Only piloteer 12 reported interpretable results using probes for the X and Y chromosomes separately.

Pilot II

Fifteen piloteers participated in this study. Their results are summarized in Table 3.

Table 3.

Results of Pilot Study II Using Probes for bcr and abl^*

Piloteer	Specimen A†				Specimen B†				Specimen C†
	Total Nuclei	Fusion	Percent Fusion	Results‡	Total Nuclei	Fusion	Percent Fusion	Results	Total Nuclei	Fusion	Percent Fusion	Result†
1	200	75	38	3	200	20	10	1	200	188	94	2
2	200	52	26	3	200	7	4	1	200	140	70	2
3	200	47	24	3	245	6	3	1	208	169	81	2
4	300	67	22	3	200	2	1	1	200	180	90	2
5	204	68	34	3§	202	10	5	1	200	191	96	2
6	200	54	27	3	200	4	2	1	200	167	84	2
7	200	33	17	3	200	3	2	6	200	151	76	2
8	200	37	19	3	200	7	4	4||	200	199	100	2
9	200	47	24	3	200	3	2	1	200	188	94	2
10	200	82	41	3	200	29	15	1	200	200	100	2
11	301	86	29	3	395	14	4	1	395	324	82	2
12	176	34	19	3	187	7	4	1	192	183	94	2
13	200	33	17	3	200	8	4	1	200	190	95	2
14	200	25	13	3	200	2	1	1	200	195	98	2
15	300	82	27	3	500	57	11	7	200	198	99	2
Total	3281	822	NA	…	3529	179	NA	…	3195	2665	NA	…
Median	200	52	24	…	200	7	4	…	200	188	94	…
SD	42.7	20.5	8	…	89.3	14.5	4	…	50.4	42.3	9	…
Maximum	301	86	41	…	500	57	15	…	395	324	100	…
Minimum	176	25	13	…	187	2	1	…	192	140	70	…

^*NA, not applicable.

^†Specimen A had a mixture of normal and Philadelphia chromosome–positive cells. Specimen B had only normal cells. Specimen C had cells from an untreated patient with chronic myelogenous leukemia.

^‡1 indicates within normal limits; 2, similar to untreated chronic myelogenous leukemia; 3, possible treated patient with chronic myelogenous leukemia; 4, ambiguous result; 5, technically unsatisfactory; 6, initially reported 53% of nuclei had multiple bcr signals; and 7, three abl signals were observed in many nuclei.

^§Piloteer indicated “result could also represent an emerging population.”

^||Piloteer indicated they did not know how to interpret 4% of cells with bcr/abl fusion.

Specimen A

For all piloteers, the mean percentage of cells detected with bcr/abl fusion was 25; the standard deviation (SD) was 8, and the range was 13 to 41.

Each piloteer indicated their results might occur in a patient with chronic myelogenous leukemia who had been treated. Piloteer 5 commented “this result could represent an emerging population.” Indeed, about 10% of untreated patients with chronic myelogenous leukemia are mosaic for normal and Ph-positive cells,⁸ and these patients would be expected to have a lower than expected percentage of cells with bcr/abl fusion.

Specimen B

The mean percentage of cells with bcr/abl fusion for all piloteers was 5%; the SD was 4%, and the range was 1% to 15%. Twelve piloteers indicated their results were “within normal limits.”

Piloteer 15 reported 11% of cells had bcr/abl fusion in this specimen and also indicated they saw three abl signals in a significant number of nuclei. No additional information from this piloteer was given to explain this unexpected result. Piloteer 7 initially reported multiple bcr signals in 53% of nuclei in this specimen, but only 2% of cells had bcr/abl fusion. Subsequently, this piloteer suggested they had a problem with excess background fluorescent signals due to inappropriate washing conditions after the hybridization step. This piloteer repeated the FISH studies on the second slide and reported that 3.5% cells had bcr/abl fusion; no cells had extra signals of any kind. Piloteer 8 reported ambiguous results but did not say why. This piloteer subsequently reported that they did not know how to interpret the meaning of 4% cells with bcr/abl fusion.

Specimen C

The mean percentage of cells with bcr/abl fusion for all piloteers was 90%; the SD was 9%, and the range was 70% to 100%. Each piloteer reported the results were “similar to untreated chronic myelogenous leukemia.”

Established Ranges

The range for cells with bcr/abl in untreated patients with chronic myelogenous leukemia varied among piloteers. Seven piloteers reported that they had not established either a normal or an abnormal range for bcr/abl. Seven piloteers indicated they had established the normal and abnormal ranges for bcr and abl, and one other piloteer reported having a normal range, but not for untreated patients with chronic myelogenous leukemia. The seven piloteers who had established normal ranges reported ranges from less than 5% to less than or equal to 15% cells with bcr/abl fusion signals.

Experience

In 1995, seven piloteers processed 10 or fewer cases, three studied 20 to 40 cases, two studied 60 to 80 cases, and three studied 100 or more cases.

Source of Probes

Nine piloteers used bcr/abl probes from Vysis Inc (Downers Grove, III), five used Oncor, Inc (Gaithersburg, Md) probes, and one used a home brew detection system but with Oncor probes.

Quality Control Programs

Twelve piloteers reported they had established a quality control program for FISH with bcr/abl; three piloteers reported they did not. We did not inquire about the specifics of these quality control programs.

Comments

Each piloteer provided comments; however, no consistent pattern was apparent. No comments were sufficiently critical to warrant a change in the procedure if a national proficiency test were modeled after this pilot.

Pilot III

Fifteen piloteers participated in this study. For slide 1, each piloteer correctly reported a normal pattern for D22S75 signals (Table 4). Only one piloteer used the B slide. A normal pattern for D22S75 signals was reported by 12 piloteers in 20 of 20 metaphase spreads, by one piloteer in 25 of 25 metaphase spreads, and by two piloteers in 19 of 20 metaphase spreads. Collectively, the piloteers examined 305 metaphase spreads, and 303 were reported to have a normal pattern of D22S75 signals. Thus, the collective piloteer metaphase hybridization efficiency (number of cells with a normal pattern of D22S75 signals divided by the total number of cells scored) for slide 1 was 99.34%.

Table 4.

Results of Slide 1 (Normal Individual) From Pilot Study III Using the D22S75 Probe

Piloteer	No. of Metaphases Analyzed	No. of Metaphases With Normal Pattern of D22S75 Signals	Percentage of Metaphases Normal for D22S75	No. of Metaphases With Microdeletion Pattern of D22S75 Signals	Percentage of Metaphases With Microdeletion of D22S75	Status of D22S75 Hybridization Site
1	20	19	95	1	5	Normal
2	20	20	100	0	0	Normal
3	20	20	100	0	0	Normal
4	20	20	100	0	0	Normal
5	20	20	100	0	0	Normal
6	20	20	100	0	0	Normal
7	20	20	100	0	0	Normal
8	20	20	100	0	0	Normal
9	20	20	100	0	0	Normal
10	25	25	100	0	0	Normal
11	20	20	100	0	0	Normal
12	20	19	95	1	5	Normal
13	20	20	100	0	0	Normal
14	20	20	100	0	0	Normal
15	20	20	100	0	0	Normal
Total	305	303	1490	2	10	…
Mean	20.3	20.2	99	0.1	1	…
SD	1.3	1.4	2	0.4	2	…
Maximum	25	25	100	1	5	…
Minimum	20	19	95	0	0	…

For slide 2, each piloteer correctly reported a microdeletion pattern of D22S75 signals (Table 5). No piloteer used the B slide. Of the 15 piloteers, 14 reported a microdeletion D22S75 signal pattern in 20 of 20 metaphase spreads; the other piloteer reported a microdeletion DD22S75 signal pattern in 20 of 20 metaphase spreads; the other piloteer reported a microdeletion D22S75 signal pattern in 25 of 25 metaphase spreads. Collectively, the piloteers reported a microdeletion D22S75 signal pattern in 305 of 305 metaphase spreads. Thus, the collective piloteer metaphase hybridization efficiency for slide 2 was 100%.

Table 5.

Results of Slide 2 (Specimens With Microdeletion) From Pilot Study III Using the D22S75 Probe

Piloteer	No. of Metaphases Analyzed	No. of Metaphases With Normal Pattern of D22S75 Signals	Percentage of Metaphases Normal for D22S75	No. of Metaphases With Microdeletion Pattern of D22S75 Signals	Percentage of Metaphases With Microdeletion of D22S75	Status of D22S75 Hybridization Site
1	20	0	0	20	100	Microdeletion
2	20	0	0	20	100	Microdeletion
3	20	0	0	20	100	Microdeletion
4	20	0	0	20	100	Microdeletion
5	20	0	0	20	100	Microdeletion
6	20	0	0	20	100	Microdeletion
7	20	0	0	20	100	Microdeletion
8	20	0	0	20	100	Microdeletion
9	20	0	0	20	100	Microdeletion
10	25	0	0	25	100	Microdeletion
11	20	0	0	20	100	Microdeletion
12	20	0	0	20	100	Microdeletion
13	20	0	0	20	100	Microdeletion
14	20	0	0	20	100	Microdeletion
15	20	0	0	20	100	Microdeletion
Total	305	0	0	305	1500	…
Mean	20.3	0.0	0	20.3	100	…
SD	1.3	0.0	0	1.3	0	…
Maximum	25	0	0	25	100	…
Minimum	20	0	0	20	100	…

Established Range

Eleven piloteers reported to have established the normal range for the percentage of metaphase spreads with D22S75 signals in normal individuals: four piloteers indicated the range was less than 95%, five indicated the range was higher than 98%, and two indicated the range was 100%.

Eight piloteers reported to have established the range for metaphase spreads with a microdeletion pattern for D22S75 signals in patients with DiGeorge sequence, velocardiofacial syndrome, and related diseases. Except for one piloteer, these numbers ranged from less than 95% to 100% of cells with a microdeletion of a D22S75 hybridization site. One piloteer reported a range of less than 90%, but this figure may have been a clerical error

Experience

Each of the piloteers used the D22S75 probe in their clinical practice in 1995. The number of times the piloteers used D22S75 ranged from 13 to 99; the mean was 44.7, and the SD was 28.3. Collectively, the 15 piloteers performed 670 studies using D22S75. Each piloteer encountered cases with microdeletions; collectively, the piloteers found microdeletions in 120 (17.9%) patients. The number of patients reported by individual piloteers to have had deletions of a D22S75 locus ranged from 3 to 18; the mean was 8.0, and the SD was 4.6.

Source of Probes

Although each piloteer indicated they had obtained probes from Oncor, Inc, one piloteer reported they had both Oncor, Inc and Vysis, Inc probes for D22S75. This piloteer did not state which probe they actually used in this survey.

Control Probe

Fourteen piloteers scored only metaphase spreads that had two control signals (ie, a DNA probe for another chromosome 22 sequence). One piloteer reported they did not use a control probe in the scoring process.

Fluorescence In Situ Hybridization in Clinical Practice

All piloteers indicated they used the D22S75 probe in clinical practice as an adjunct to conventional cytogenetics. This was interpreted to mean that the piloteers used the D22S75 FISH studies only as an adjunct to conventional chromosome study.

The number of metaphase cells examined by piloteers ranged from 10 to 50: four piloteers reported 10 cells, three reported 10 to 20 cells, three reported 15 cells, three reported 20 cells, one reported 25 cells, and one reported 50 cells.

Scoring Criteria

Seven piloteers scored D22S75 signals at the chromosome level (ie, one or more FISH signals at the appropriate site per chromosome equals a single hybridization site per chromosome). Eight piloteers scored signals at the chromatid level (ie, two FISH signals at the appropriate site per chromosome equals a single hybridization site per chromosome).

Quality Control Programs

Fourteen piloteers had established a quality control program for FISH using the D22S75 probe; one piloteer had not. Nine piloteers reported having a written quality control program. We did not inquire about the nature of these quality control programs.

Comments

Most piloteers provided comments, but no consistent observation was reported. No comments were sufficiently critical to warrant a change in the procedure if a national proficiency test were modeled after this pilot.

COMMENT

In January 1996, 243 laboratories enrolled in the ACMG/CAP Cytogenetic Survey were asked about using FISH in their practice. Data were returned by 221 participants; 153 laboratories were sufficiently interested to volunteer for pilot studies involving FISH and expressed support for proficiency testing.

Each laboratory was queried about its use of six probes for congenital disorders (Table 6). Probes for the centromeres of different chromosomes were used by 166 laboratories on more than 3109 specimens. These probes were frequently used to study marker chromosomes and aneuploidy. Locus-specific probes for SNRPN, D15S10, and GABRB3 were used by 172 laboratories to study more than 2756 specimens. These probes were generally used to detect microdeletions in the Prader-Willi and Angelman syndromes.¹ D22S75 was used by 156 laboratories on more than 2501 specimens. Other common FISH probes used in the diagnosis of congenital disorders included X and Y probes (more than 2188 utilizations by 167 laboratories), whole-chromosome libraries (more than 1361 utilizations by 162 laboratories), and a probe to detect deletions in the elastin gene for Williams syndrome (more than 638 utilizations by 128 laboratories).

Table 6.

Summary of Probe Utilization for Congenital Disorders in 1995^*

No. of Cases	Centromeric Probes	SNRPN/D15S10/GABRB3	D22S75	X and Y	Whole, Chromosome Libraries	Elastin
None	54	49	64	53	58	89
1–10	79	57	60	99	93	95
11–30	53	74	56	40	53	27
31–50	6	25	23	11	9	4
51–100	11	10	12	8	6	3
>100	17	6	5	9	1	0
Total No, of laboratories	220	221	220	220	220	218
No. of FISH laboratories	166	172	156	167	162	129
No. of specimens studied	>3109	>2756	>2501	>2188	>1361	669

^*Data provided by American College of Medical Genetics/College of American Pathologists cytogenetics survey participants. FISH indicates fluorescence in situ hybridization.

The survey also inquired about five probes often used for malignant hematologic disorders (Table 7). Probes for bcr and abl were used for chronic myelogenous leukemia by 111 laboratories on more than 1424 specimens. Probes for X and Y chromosomes were used by 81 laboratories to study more than 1353 specimens. Other common probes for neoplastic disorders included centromeric probes (more than 1255 specimens studied by 101 laboratories), whole-chromosome libraries (more than 451 specimens studied by 92 laboratories), and PML/RARA (more than 391 specimens studied by 62 laboratories).

Table 7.

Summary of Survey of Probe Utilization for Neoplastic Disorders in 1995^*

No. of Cases	bcr/abl	X and Y	Centromeric Probes	Whole Chromosome Libraries	PML/RARA
None	105	133	114	123	152
1–10	68	55	61	77	52
11–30	27	8	24	10	4
31–50	5	5	9	2	3
51–100	4	5	1	2	2
>100	7	8	6	1	1
Total No. of laboratories	216	214	215	215	214
No. of FISH laboratories	111	81	101	92	62
No. of specimens studied	>1424	>1353	>1255	>451	>391

^*Data provided by American College of Medical Genetics/College of American Pathologists cytogenetics survey participants. FISH indicates fluorescence in situ hybridization.

Nineteen cytogenetic laboratories were involved in these pilot studies, but only nine participated in all three investigations. Some piloteers had more experience than others, but in general the participants in each of these pilot studies were representative of US laboratories performing FISH. We did not inquire about previous experience in the first pilot study. The experience reported with bcr and abl in the second pilot study varied considerably. One piloteer had not processed any cases with bcr and abl in 1995, and six others had analyzed fewer than 10 cases; the remaining piloteers had evaluated 20 or more cases. In the third pilot, each piloteer had used the D22S75 probe in their practices in 1995; the number of cases studied ranged from 13 to 99. In the third pilot, we inquired about piloteer experience with both normal and abnormal cases using the D22S75 probe in 1995. We learned that 17.9% of the collective specimens studied by the piloteers had a deletion of a D22S75 locus. This percentage is similar to published series at institutions that see large numbers of patients with features of DiGeorge sequence, velocardiofacial syndrome, or related diseases. For example, Larson and Butler⁵ reported deletions of the D22S75 locus in 3 of 20 patients presenting with DiGeorge sequence, velocardiofacial syndrome, or related diseases. In another study, Crifasi et al³ reported deletions of the D22S75 locus in 9 of 42 patients with features of DiGeorge sequence, velocardiofacial syndrome, or related diseases.

These pilot studies were not the first to demonstrate that multiple laboratories can work together to evaluate probes using FISH. In one study, the Great Lakes Regional Genetics Group enlisted the aid of 23 laboratories with varying experience (some with no FISH experience) to evaluate the SNRPN probe to study metaphase spreads.¹⁴ In a separate study, another group of 26 Great Lakes Regional Genetics Group laboratories successfully completed a study using FISH probes for the X and Y chromosomes to study interphase nuclei.¹⁵ In still another study, three different cytogenetic laboratories successfully collaborated to evaluate probes for the X and Y chromosomes to detect XX and XY cells in bone marrow from patients who had undergone opposite-sex bone marrow transplantation.^16,17 Although a national proficiency test would probably involve 120 or more laboratories, the results of these pilot studies and others^14–17 suggest it would be possible to coordinate multiple institutions for FISH studies.

Each of these pilot studies was expected to be completed within 2 to 3 weeks. Except for one piloteer in pilot II, each volunteer successfully completed the work within the time allotted. Thus, we believe that a FISH proficiency test can be completed by a large group of laboratories within a relatively short time period. Similar conclusions were reported by the Great Lakes Regional Genetics Group.^14,15

In each pilot study, nearly every piloteer used commercial probes. The source of the probes did not seem to influence the success of the study. Data from these pilot studies and from previous reports^14–17 suggest that results are similar whether probes are provided by the organizing group or provided by the laboratories themselves.

In the first pilot study, we investigated possible differences between prepared slide preparations and cell pellets from which the pilot laboratories would make slides. We found that the percentage of XX and XY cells in interphase nuclei and metaphase spreads were generally comparable in slides provided versus fixed cell pellets. Thus, the possibility of success and accuracy of proficiency testing seems equivalent for either type of specimen. In proficiency testing, the use of fixed cell pellets would allow each laboratory to make its own slides and thereby extend the depth of testing to include slide preparation. However, it would be necessary for the central laboratory to culture many more cells to provide adequate fixed cell pellets for all participants. Using prepared slides for proficiency testing would circumvent this problem and would be less expensive. The Great Lakes Regional Genetics Group FISH investigations demonstrated similar results.^14–15

In the first pilot study, several piloteers reported a different ratio of XX to XY cells in interphase than in metaphase. We suspect the mitotic index from the culture of this male blood specimen was higher than for the female. It is common knowledge among cytogenetic laboratories that blood specimens from different individuals often have different mitotic indices.

Piloteers used a variety of scoring procedures in pilot I. Some of these differences stemmed from the fact that some piloteers used dual-colored probes and others used X and Y probes separately. We had considerable difficulty collecting consistent data for piloteers using X and Y probes separately. We concluded that for X and Y chromosome studies, it may be necessary to restrict proficiency testing to dual-colored probes for the X and Y chromosomes for evaluable results.

No problems with scoring were evident in the second and third pilot studies. In pilot II, except for three problematic reports for specimen B, the results were concordant among the various piloteers. Each of the three problems did not appear to be related to specimen quality, but rather to technical problems with the FISH procedure or inexperience with using this method. The results for pilot III were also highly concordant among the piloteers. The results for pilots II and III bode well for doing proficiency testing with bcr/abl on interphase nuclei and D22S75 on metaphase spreads.

We observed one difference in the scoring procedure among the piloteers in pilot III, but this did not interfere with the accuracy of the test. Seven piloteers scored D22S75 signals at the chromosome level (ie, the observation of at least one D22S75 signal on a chromosome is necessary to classify the D22S75 locus as normal). Eight piloteers scored signals at the chromatid level (ie, the observation of one D22S75 signal on each chromatid is required to classify the D22S75 locus as normal).

In pilot I, several piloteers reported that the metaphase distribution on the slides they received was not ideal. These comments were made mainly by piloteers using X and Y probes alone, and these participants had the most technical difficulties with FISH testing. Nevertheless, because of this experience with pilot I, we anticipated similar problems with satisfactory metaphase preparations in pilot III. However, piloteers apparently had no problem finding 20 metaphase spreads as instructed on each slide. Only one laboratory used the backup slide to obtain its results, and that decision was not related to metaphase quality. Furthermore, the collective piloteer hybridization efficiency was 99.34% for specimen 1 and 100% for specimen 2, indicating consistent and evaluable metaphase quality in pilot III. Although three piloteers provided favorable comments about metaphase quality, two piloteers provided somewhat negative comments about metaphase quality. Ten other piloteers had no comment about metaphase quality, and we interpreted this to mean that metaphase quality was acceptable.

The results of these pilot studies suggest that proficiency testing for FISH with chromosome-specific DNA probes is attainable. Furthermore, this can be done at relatively low cost and will provide objective data with high concordance among laboratories. Although these pilot studies focused on probes for the X and Y chromosome and locus-specific bcr/abl and D22S75 probes, FISH procedures for these DNA probes are similar to those for many other probes. Thus, the CAP/ACMG Cytogenetics Resource Committee believes that large-scale proficiency testing with other probes is also feasible. The results of these pilot studies are particularly important for the incorporation of FISH into clinical practice since the Clinical Laboratory Improvement Amendments of 1988 require proficiency testing for all clinical laboratory procedures.¹⁸

Figure 1.

A, Representative interphase nuclei showing X (red) and Y (green) chromosome hybridization signals from pilot I. B, Representative interphase nuclei showing bcr (red), abl (green) and bcr/abl fusion (yellow) signals from pilot II. C, Representative metaphase spread showing a microdeletion of a D22S75 hybridization site from pilot III. (Normal chromosome 22 showing two hybridization signals and the chromosome 22 with a microdeletion showing only one hybridization signal.)