College of American Pathologists
Printable Version

  New school of FISH:
  solid tumor testing


cap today

January 2004
Cover Story

William Check, PhD

If it looks like a duck and quacks like a duck, it must be a duck. Unless it’s a FISH.

Like a typical cytologic or histologic technique, FISH-fluorescence in situ hybridization-is performed on cells fixed on a slide and stained. However, the stain is not a cytochemical or histochemical stain such as the Papanicolaou or H&E stain, but a molecular genetic stain that consists of fluorescently labeled DNA probes, and the abnormalities being sought are genetic, not morphologic. We can only conclude that FISH is a hybrid, not so much duck as platypus. It is the hybrid character of FISH that defines its advantages and disadvantages relative to cytologic and histologic techniques and more conventional molecular techniques such as polymerase chain reaction, or PCR.

“FISH combines the morphology of tissues or cells with molecular assessment,” says Marek Skacel, MD, associate staff in the Department of Anatomic Pathology at the Cleveland Clinic Foundation. “It is not as good for morphology as light microscopy, but it allows co-localization of molecular changes in cells.” Co-localization of genetic or chromosomal changes relies on a nuclear counterstain (DAPI) and can be further enhanced by looking at conventionally stained preparations in parallel to FISH. “This is an advantage compared to amplification techniques like PCR where you grind up cells and have no idea where the genetic changes are coming from,” Dr. Skacel says.

Kevin Halling, MD, PhD, uses FISH primarily to detect malignant cells in cytology specimens. “FISH is a very powerful tool,” says Dr. Halling, assistant professor of pathology and laboratory medicine, Mayo Medical School, and co-director of the clinical molecular genetics laboratory at the Mayo Clinic. “It has the advantage of being able microscopically to detect rare tumor cells that typical molecular genetic techniques might not pick up because they are diluted with normal cells.”

Similar considerations led Bonnie King, PhD, associate research scientist in the Department of Genetics at Yale University School of Medicine, to choose FISH for her work on breast ductal lavage specimens. “If we had extracted DNA from our whole samples,” she says, “we would have gotten nucleic acids from noncancerous infiltrating immune cells as well as normal breast epithelial cells, which could have confused our results. So we either had to use cell separation strategies or an in situ approach like FISH.”

These remarks emphasize FISH's molecular genetic character. Emphasizing the histologic side, Gregory Tsongalis, PhD, associate professor of pathology at the University of Connecticut and director of molecular pathology in the Department of Laboratory Medicine and Pathology at Hartford (Conn.) Hospital, refers to FISH as “a new special stain.”

“I don't want to trivialize the technology,” he says, “but it has changed enough in the past few years that it has become as routine as a special stain for a pathologist to use. A lot of laboratories can now bring this in-house.”

As Dr. Halling puts it, “Instead of using a Pap stain, we are using a FISH stain to identify malignant cells.”

Much of FISH's current value in molecular oncology comes from recent research on genetic alterations in solid tumors, says John Coon, MD, PhD, professor of pathology and otolaryngology at Rush Medical College and director of molecular diagnostics at Rush Medical Laboratories, Chicago. “We are in a discovery phase of unraveling genomic abnormalities in solid tumors because of some very important array-based techniques that have been brought to bear just in the last few years,” Dr. Coon says. Particularly useful, in Dr. Coon's view, has been comparative genomic hybridization, or CGH, which can survey for copy number abnormalities for a large number of genes in solid tumor samples.

“This technique is producing a lot of useful data about which gene amplifications and deletions might be linked to clinical aspects of solid tumors, especially prognosis and response to certain therapies,” Dr. Coon says. Results of CGH studies often translate into FISH probes, which have the advantage of being able to pinpoint specific genes involved in amplified or deleted areas. “It is often more practical in diagnostic laboratories to do FISH studies than relatively technically demanding array-based studies,” Dr. Coon says.

New insights into the significance of copy number abnormalities in specific genes, he says, are likely to translate soon into diagnostic tools. “FISH is quite an attractive and practical way to assess those abnormalities in clinical tumor samples,” he adds.

Already these new clinical applications of FISH are having an effect. Even now, “Requests for FISH testing are soaring through the roof,” Dr. Tsongalis says. “We can't keep up with them, to the point where we have cross-trained three of our molecular technologists to do FISH and we are thinking of cross-training two more.” One reason for what Dr. Tsongalis calls the “exploding” volume in FISH testing is that his laboratory brought several tests in-house, shortening turnaround time and improving interactions with physicians.

The second reason for the volume increase is the expanding number of applications for FISH, not only for BCR/ABL and other translocations but also in solid tumors. “That was where solid tumor oncology was really falling behind in molecular diagnostics,” Dr. Tsongalis says. “We didn't have a lot of good markers. Now with more applications and better clinical utility, we need to be able to isolate targets and identify which cells are abnormal.” Solid tumors tend to have many chromosomal and genetic abnormalities, leading to multiple probes, such as the four probes in the UroVysion FISH assay for bladder cancer surveillance. “Panels make some pathologists nervous,” Dr. Tsongalis says. “We need to become comfortable looking at several abnormalities at the same time.”

Says Dr. Halling, “The use of multiprobe cocktails increases both the sensitivity and specificity of the assays over single- or dual-probe assays.”

Speakers at a symposium at the recent Association for Molecular Pathology meeting presented data on the value of FISH in the diagnosis or prognosis of several solid tumors, including carcinoma of the bladder, lung, larynx, and breast.

Dr. Skacel discussed Vysis UroVysion (Vysis/Abbott), an FDA-cleared assay for detecting recurrent bladder cancer. Urothelial carcinoma, or UC, has a high recurrence rate, with about 70 percent of UC cases recurring as noninvasive tumors and up to 20 percent of them progressing eventually to invasive cancer. Current surveillance tools include scheduled cystoscopy, which detects low-grade papillary tumors but sometimes misses high-grade carcinoma in situ, and cytology at the time of cystoscopy, which detects high-grade CIS but frequently misses low-grade papillary tumors.

Using FISH probes to detect recurrent UC in the urine relies on the fact that urothelial cells, especially UC cells, readily exfoliate and that UC cells, even from low-grade tumors, have a high frequency of chromosomal abnormalities. UroVysion’s four-probe set—CEP7, CEP3, CEP17, and LSI (locus-specific identifier) 9p21, all bearing different fluorophores—targets the common abnormalities in UC: It is directed to the centromeres of chromosomes 3, 7, and 17 and the P16 gene on chromosome 9. In 95 percent or more of high-grade UC cases, there is a gain of at least two of these chromosomes in the neoplastic cells. The gain of two or more chromosomes in a cell is called polysomy. Polysomy isn’t present in low-grade papillary UC. Instead, there is a deletion of the 9p21 locus, which with UroVysion is observed as homozygous 9p21 deletion, that is, loss of both copies of the gold 9p21 probe in the neoplastic cells. A small fraction of UC have trisomy (three copies) of one chromosome, such as CEP7 trisomy, with disomy (two copies) of the other chromosomes. Since FDA clearance was based on detecting either polysomy or homozygous 9p21 deletion, a laboratory using an isolated trisomy as a criterion for the presence of UC must validate it in-house.

Dr. Skacel cited Dr. Halling's pivotal study on UroVysion, which found the assay to be 81 percent sensitive for UC and 96 percent specific (cytology, 58 percent and 98 percent). (P value for sensitivity difference between FISH and cytology was 0.001 and for specificity difference was 0.564.) UroVysion was more sensitive at every stage and for grades two and three but not grade one (J Urol. 2000; 165: 1768-1775). In five cited studies, UroVysion approached 100 percent specificity in all but one, a prospective study that is still in progress at the Cleveland Clinic, in which UroVysion showed specificity approaching 90 percent at nine-month followup (Yoder B, et al. Acta Cytol. October 2003, abstract). Dr. Halling also published a comparison of four approved assays for detecting UC: BTA stat, hemoglobin dipstick, telomerase, and UroVysion. That study revealed that UroVysion had the highest accuracy (that is, the highest combined sensitivity and specificity) of the markers evaluated. UroVysion plus cystoscopy was 98 percent sensitive and more than 95 percent specific. (Dr. Halling and others have a patent on the UroVysion bladder cancer detection probe set and receive royalties from sales of the product.)

In the FDA-cleared mode, manual slide preparations of urine samples are made at three dilutions; the assay is then performed on the dilution with the optimal cellularity (approximately 100 to 200 cells as determined with a phase contrast microscope). After the FISH hybridization, 25 cytologically abnormal cells, as observed with the DAPI nuclear counterstain, are scored. (Dr. Halling notes that many labs don't like the slide preparation aspect of the test, and that instruments to automate the slide preparation are being developed.) It can also be performed on ThinPrep slides, as the Cleveland Clinic laboratory does, an approach that is not FDA-cleared or approved and requires validation. (The validation of the Cleveland Clinic technique can be found in Anal Quant Cytol Histol. 2001;23:381-387.)

With ThinPrep slides, cytologic examination is first done on the same slide, and UroVysion is not done if cytology is positive. “Most people don't do cytology when preparing the slides for FISH manually,” Dr. Skacel says, “unless they split the sample, which can cause sampling discrepancies.” Using ThinPrep slides is currently more expensive because it requires a higher volume of UroVysion probe (10 µL instead of 3 µL); however, in the future this could be dealt with by reducing the size of the target area on the ThinPrep slide, Dr. Skacel says. Hybridization with UroVysion is an overnight procedure, but true turnaround time is up to seven working days because the test is batched and not performed every day in most laboratories.

Several studies have shown that UroVysion often resolves suspicious or atypical cytology, detecting carcinomas in many cases. Dr. Skacel has demonstrated further that UroVysion has an overall 85 percent sensitivity for carcinomas in cases with atypical or negative cytology (J Urol. 2003; 169: 2101-2105). “We looked at cases that had either atypical or normal cytology from patients with available subsequent biopsy-proven diagnosis,” he says. “We found that the majority of cancers that presented with atypical or even normal cytology can be detected by FISH. So now we routinely apply FISH to samples with negative cytology in addition to samples with atypical cells.”

UroVysion can also give “anticipatory positive” results. A high percentage of patients with negative cystoscopy and normal cytology but positive FISH will develop recurrent tumor over the next six to 12 months. Urologists at the Cleveland Clinic follow such patients at shorter intervals.

Dr. Skacel concludes that UroVysion is an “effective molecular genetic assay for bladder cancer surveillance.” It can be used as a primary test or a reflex test for atypical or negative cytology. UroVysion is seen as an ancillary test to cytology at the Cleveland Clinic, and its laboratory offers “cytology and FISH for bladder cancer.” Although Dr. Skacel estimates that about 90 percent of the cases they receive go to FISH, performing cytology serves as an additional quality control step and helps eliminate testing of technically unsatisfactory as well as cytologically positive cases. “As pathologists, we prefer to know exactly what we’re working with when performing a specialized test such as FISH,” he says. He predicts that FISH will soon become the standard of care for UC surveillance and will probably extend into screening of clinically high-risk patients—smokers and those with hematuria.

In Dr. Halling's experience, clinicians are just getting used to FISH being used as an adjunct to urine cytology. “My personal feeling,” he says, “is that over the next decade it may supplant urine cytology-it is more sensitive and just as specific.”

Marileila Varella-Garcia, PhD, director of the Cytogenetics Core at the University of Colorado Cancer Center, is exploring FISH for detecting premalignant cells in people at risk for lung cancer. “Chromosomal abnormalities are frequent events in lung cancer and probably early events in tumorigenesis,” Dr. Varella-Garcia says. “We set out to see whether these abnormalities might be a useful biomarker for assessment of lung cancer risk.”

Initially, she and her colleagues assessed karyotype-wide chromosomal abnormalities in bronchial epithelium of heavy smokers by spectral imaging using a 24-probe set called Human SkyPaint (available from ASI). Chromosomal abnormalities were detected at high frequency-67 percent-in metaphases of cells from primary cultures of bronchial epithelium. They included mono-, tri-, and tetrasomy (+7 +8, and +18), polyploidy, and structural changes (deletions, inversions, translocations, and so on). “All of those changes are unusual in normal cells, even when short-term cultured,” Dr. Varella-Garcia says. Chromosomal abnormalities seen in dividing bronchial epithelium cells were confirmed in interphase cells using centromeric probes (CEPs 3, 7, 8, 18, X, Y) and region-specific probes.

In their next step, these investigators assessed chromosomal abnormalities in interphase cells of premalignant bronchial epithelium and sputum of heavy smokers using multi-target FISH-the LAVysion Multi-color Assay (Vysis/Abbott), which contains four probes: LSI 5p15, CEP6, LSI 7p12 (EGFR gene), and LSI 8q24 (c-myc gene). Two biopsies were inspected from each patient: one with normal histology and one with mild dysplasia or worse. Their analysis showed, first, that chromosomal abnormalities can be detected in sectioned bronchial epithelium biopsies by multicolor FISH analysis. This was not a foregone conclusion. “When you section tissue, you lose part of the nuclear contents,” Dr. Varella-Garcia explains. “How much have you lost for each cell? What is the expected number of chromosomes in a normal cell in these sections?” Investigating samples with normal histology sectioned to the same thickness as the test samples allowed them to define a threshold for abnormalities.

FISH-detected chromosomal abnormalities increased in frequency by 20 to 30 percent from normal to dysplastic epithelium. However, concordance between FISH and histology abnormalities happened in only 53 percent of specimens. “We were surprised at that low concordance rate,” Dr. Varella-Garcia says. “We are not sure what it means.” One practical possibility is that FISH and histology score will be complementary in predicting clinical outcome.

In an intriguing finding, some sections with normal histology had one abnormal FISH target. Dr. Varella-Garcia cites the gains of the two chromosomes-abnormality criterion (that is, polysomy) for a UroVysion-positive. “We are not looking for tumor cells, but for abnormal cells at the premalignant level,” she notes, “so we may need a different criterion.” Will one target suffice to identify a premalignant sample? To address this question, they are following patients clinically.

A separate study showed that preserved sputum is a feasible biological sample for multi-target FISH analysis and that chromosomal abnormalities can be detected in epithelial nuclei in sputum of people at risk for lung cancer. However, only 86 percent of specimens yielded a result. Better strategies for sputum preservation are needed. A promising finding was that levels of chromosomal abnormalities by FISH in sputum collected within 12 months of diagnosis were significantly associated with the development of lung cancer. Unfortunately, sensitivity was poor, 41 percent. Sputum cytology had 56 percent sensitivity. Since the two tests were largely independent, with only 43 percent concordance, combining them (either test positive) increased sensitivity to 83 percent while retaining specificity, 80 percent. “FISH nicely complements cytology as a biomarker for lung cancer risk,” Dr. Varella-Garcia concludes.

Vysis plans to do more preclinical trials with LAVysion. The goal is to have it qualify as an in vitro diagnostic for lung cancer.

Dr. Varella-Garcia is submitting a grant application to use FISH as an intermediate endpoint in chemoprevention studies. “If you know that aneusomy is related to cancer,” she says, “you can use aneusomy as an intermediate endpoint that will show any benefit of chemoprevention more quickly than using a hard endpoint.”

Dr. Halling is investigating use of the LAVysion kit for detecting lung cancer in sputum specimens and bronchial secretions and brushings. “Preliminary studies suggest it is more sensitive than cytology and just as specific,” he reports. “It is quite promising for picking up lung cancer earlier.” (A patent, on which Dr. Halling is listed as a co-inventor, is pending for LAVysion.)

“The nice thing about the FISH tests for bladder cancer, lung cancer, and other tumors,” he says, “is that they not only diagnose tumors, but will probably add prognostic and therapeutic information.” One LAVysion probe is for the epidermal growth factor receptor gene, or EGFR, whose protein product is the target of the drug Iressa (gefitinib), which is approved for non-small cell lung cancer. “A story analogous to the HER2/Herceptin story could develop here,” Dr. Halling says.

In laryngeal cancer, too, a FISH probe for EGFR is being explored for its prognostic value. Dr. Coon explains that EGFR is a member of a family of growth factor receptors that also includes HER2. When these receptors are deregulated in cancer cells, they can drive cell proliferation and block apoptosis. Laryngeal carcinoma is an important head and neck cancer with a highly variable clinical course. A prognostic marker such as EGFR that stratifies patients into risk categories might have practical importance in planning therapy. “We were fortunate in being able to acquire a patient sample set where patients were all treated pretty much the same way by a single surgeon and long-term followup was available,” Dr. Coon says.

He assessed the prognostic significance of EGFR amplification or deletion and gene dosage (chromosome 7 polysomy) by FISH probes in 59 patients with advanced squamous cell carcinoma of the larynx. FISH results were compared with EGFR expression via immunohistochemistry. Overexpression by IHC was defined as greater than 70 percent of tumor cells with strong surface staining.

During 12 years of followup, the difference in survival between patients whose tumors had low or high IHC expression was only marginally significant. For abnormal FISH results (EGFR amplification or deletion), the survival difference compared with normal EGFR copy number was clearly significant. Comparing chromosome 7 disomy versus polysomy produced an interesting observation: None of the patients with disomy died of their disease, although many with polysomy did not die of the disease either.

Most important, a striking survival difference was seen when patients were divided into three groups: those with disomy (0/10 dead); those with polysomy and normal EGFR gene copy number (6/33 dead); and patients with polysomy plus abnormal EGFR copy number (10/16 dead). Dr. Coon combined EGFR copy number, chromosome 7 polysomy/disomy, and EGFR protein expression by IHC to create three categories. These categories discriminated survival very well (P < 0.0001). Dr. Coon posits a possible interpretation for the independent contribution of chromosome 7 polysomy. “There is a lot of genetic information on chromosome 7,” he notes, “so duplication of something other than the EGFR gene might be contributing.”

He concludes that EGFR amplification or deletion and chromosome 7 polysomy may be of prognostic significance in surgically treated advanced squamous cell carcinoma of the larynx. “Adding EGFR-related markers may increase prognostic discrimination,” he suggests.

Dr. Tsongalis is working with FISH probes to evaluate several aspects of breast carcinoma. Along with surgical pathologist Andrew Ricci Jr., MD, he is exploring the significance of HER2 positivity in ductal carcinoma in situ. “Apparently what we are seeing here, and what other people have reported in a few cases, is HER2 amplification in a noninvasive lesion,” Dr. Tsongalis says. “Right now Herceptin is approved only for invasive breast cancer,” he notes, “so people who do FISH need to know what they are looking at.” It is important not to treat with Herceptin on the basis of HER2 amplification in noninvasive cells. It is also important to know if this finding puts a woman at increased risk of invasive breast cancer. “These questions require clinical followup,” Dr. Tsongalis says.

He is also working with a four-probe breast aneusomy panel to evaluate multicentric breast disease. “In a majority of women with breast cancer, there is more than one lesion, typically within the same breast,” Dr. Tsongalis says. “These tumors are thought to have the same origin. We are using FISH to get a handle on these different foci of tumor cells-are they genetically identical? If they are different, do they behave differently?” He is using the same FISH panel to study genetic relatedness in cases of primary versus metastatic disease and synchronous versus metachronous disease.

Dr. Tsongalis is conducting what he calls “probably the most exciting FISH application for us right now,” assessing sentinel lymph nodes. What do small clusters of tumor cells in a sentinel node mean clinically? Could these potentially metastasize? Or do they represent only a few cells that the person's immune system can destroy? By using FISH probes to determine whether the genetic pattern of tumor cells in sentinel lymph node matches the primary tumor, and then monitoring clinical outcomes, Dr. Tsongalis and his coworkers hope to answer these questions. “We are able to detect these tumor cells very readily in situ,” Dr. Tsongalis says, “so we can do multiple staining on them, such as for cytokeratin.”

To optimize FISH testing, Dr. Tsongalis has set up a cooperative effort between himself, the molecular laboratory, the director of the IHC laboratory, and surgical pathology. “Everybody is on the same page to make sure we are looking at what we are supposed to be looking at,” he says. “We have a pretty comprehensive reporting scheme that should be easy to replicate in other places.” He believes that most pathology departments now have the core competencies; it is a matter of how well people work together. “FISH testing can't be done in a molecular laboratory with no input from tissue morphology,” Dr. Tsongalis says.

An intraductal approach to detecting early breast cancer is being undertaken by Dr. King, who uses FISH to analyze breast ductal lavage fluid. The ductal lavage technique was introduced in 1999 and uses a device called a microcatheter to cannulate an individual milk duct opening. The duct is flushed with saline, and exfoliated breast epithelial cells are harvested.

“The initial approach to evaluating ductal lavage cells for cancer-associated changes was to use cytology,” Dr. King says. “Our goal was to use a genetic method to confirm and extend cytologic findings.” Based on the chromosomal abnormalities most frequently reported during breast cancer progression, Dr. King chose centromeric probes to chromosomes 1, 8, 11, and 17.

She set up a collaboration with breast surgeon Rogsbert Phillips, MD, of Decatur, Ga., who did ductal lavage on women before they underwent biopsy or mastectomy and sent cells to Dr. King for FISH analysis. Collaborating cytopathologist David Rimm, MD, PhD, evaluated the cells for cytologic changes. Cytology identified abnormalities in seven of 15 evaluable lavages collected from cancer patients and in four of 19 lavages harvested from benign cases. “So cytology had 47 percent sensitivity and 79 percent specificity on evaluable specimens,” Dr. King says. “FISH did a little better”: Numeric chromosome changes were seen in 10 of 14 malignant cases and only two of 18 benign cases, for sensitivity of 71 percent and specificity of 89 percent (Clin Cancer Res. 2003; 9 :1509-1516).

For this initial study, Dr. King selected centromeric probes, which detect changes in whole chromosome copy number but not region-specific deletions. Centromeric probes give “a loud robust signal,” Dr. King says, which increases the chance of finding chromosomal abnormalities. “Our future plan,” she says, ““is to also look at chromosomal changes that involve specific deletions at discrete regions. In this first study we just wanted to prove that we could detect cancer-associated alterations in lavage specimens.”

Analyzing ductal lavage fluids has revealed remarkable cell heterogeneity (Cancer Cytopathology. 2002;9: 244- 249). “Ductal lavage populations are not as simple as we thought,” Dr. King says. “Not only are we collecting exfoliated ductal epithelial cells, we are also getting infiltrating cells in huge numbers. The intraductal compartment often has thousands of macrophages.” It is this heterogeneity that makes FISH's ability to localize chromosomal abnormalities in target cells preferable to total nucleic acid extraction, so-called grind-and-find techniques.

“Our main goal in this work was to show that we could find breast cancer-associated genetic alterations in ductal lavage cells harvested from women with in situ and invasive cancer,” Dr. King says, “and we did that unequivocally.” She says that the chromosomal abnormalities visualized with FISH were so clear that “a five-year-old child could score them.” Sensitivity of FISH, 71 percent, was low, however, perhaps because many advanced lesions no longer shed into the ductal system. Dr. Phillips also made touch preparations from the tumors she excised, and in some cases the preparations from very advanced lesions gave a high degree of chromosomal changes, while the ductal lavage cells from the same woman were normal. “While this limits the application of ductal lavage for detecting the later stages of breast cancer progression,” Dr. King says, “advanced lesions should be picked up by other modalities.”

Based on this study, Dr. King doesn’t think it probable that ductal lavage has acceptable sensitivity for detecting breast cancer with present methods, “although that has not been exhaustively defined,” she cautions. Dr. King believes her work, along with several other studies, indicates that ductal lavage may be most useful for risk assessment. “What we have done with this study is to validate ductal lavage cells for biomarker development,” she says. “We have shown that we can detect chromosomal alterations.” This makes lavage cells valid for risk assessment, in which one tries to catch lesions early and predict which will progress to invasive cancer. “Numeric changes in several chromosomes of our FISH panel have been seen in the earliest stages of breast cancer progression,” Dr. King says.

A critical question is how many genetic or chromosomal changes need to be seen to predict that an early lesion will progress to cancer. In a multicenter trial evaluating the microcatheter device on high-risk but asymptomatic women, almost one-fourth of evaluable specimens contained atypical cells on cytology (J Natl Cancer Inst. 2001; 93: 1624–1632). “This is one of the big challenges facing the intraductal approach,” Dr. King says. The technique reveals a lot of cytologic atypia in the breast, a finding associated with an increased relative risk for invasive breast cancer. However, studies show that only about 10 to 25 percent of women diagnosed with atypia go on to get invasive breast cancer. “We need to be able to discriminate or stratify these atypical cases,” Dr. King says. “That is where a method like FISH may help.”

As FISH becomes more widely used, personnel issues will arise. Dr. Tsongalis says the main obstacle to FISH testing in his laboratory now is training technologists to read slides on the fluorescence microscope. “We can process hundreds of slides per day, but who will read them?” he asks. One solution is automated spot readers. A MetaSystems instrument, he says, “seems to be one of the nicest systems on the market for doing that right now.”

Dr. Halling is collaborating with Vysis and MetaSystems to develop automated reading of UroVysion slides. “I think these machine-versus-man comparisons need to be done,” he says. “We just have to wait and see whether instruments will save time and reduce costs.” Dr. Halling is somewhat concerned that the instrument won't be able to pick up rare cells as well as humans do. “But I think it is important that they work,” he says. Dr. Halling agrees with Dr. Skacel that UroVysion might eventually be used for screening for bladder cancer. If that happens, test volumes will be high. “It will be important to find ways to reduce technologist time,” he says.

Dr. Skacel finds that the volume of requests for UroVysion assays is already increasing to the point that the Cleveland Clinic laboratory is looking into purchasing an automated reader soon. “Such an instrument will become absolutely necessary once test volume exceeds a certain level, as we expect. We currently do 25 or so UroVysion assays per week,” he says. “If that volume doubles, it will be difficult to maintain our turnaround time.” Automated counters will also make it easier to scan all cells and to archive images at the time of scanning, which is currently cumbersome.

In addition to the analytical and technical issues surrounding FISH testing of solid tumors, there are what Dr. Halling calls political issues, such as who is going to perform FISH testing. Because FISH assays can be applied to just about any type of cytologic specimen, Dr. Halling believes that cytologists and cytopathologists will be increasingly involved in performing and interpreting these tests, rather than the molecular diagnostics laboratory.

“I do think there will be a field of molecular cytology that forms over the next decade,” he says. “Cytologists won't just be doing conventional stains anymore to look for cells.”

William Check is a medical writer in Wilmette, Ill.