Evaluation of UroVysion™ FISH and Cytology Testing – Concordance and Cost Effectiveness Comparison between Cotesting vs. Non-Cotesting Samples.
Niti Manglik, Patricia Wasserman, Silvat Sheikh-Fayyaz, Nora Morgenstern. North Shore Long Island Jewish Health System, New Hyde Park, NY
Background: FISH is widely used in urothelial cancer patients however, there are no strict guidelines for ordering this expensive test. Some clinicians order it only in cases with equivocal cytology whereas others use it more liberally. Combined FISH and cytology may guide clinicians well in straight forward cases but in fair number of cases they will give conflicting results making it difficult to plan treatment. This problem is further complicated by the fact that the amount/quality of cells might be different in each urine specimen due to variable shedding of urothelium. We hypothesize that if these tests are done on the same urine sample (cotesting); they will have a better concordance and hence will be easier to interpret for clinicians, as opposed to when they are done on different urine specimens (non-cotesting).
Design: We studied 203 patients (410 samples) who had urine cytology and FISH tests at our institution during 2006-2009 and divided their samples in two groups: Cotesting and Non-Cotesting. Cotesting category had urine cytology and FISH on same sample. Non Co-testing group had these tests on different urine samples. We compared concordance of cytology with corresponding FISH results in both groups to evaluate if it was different in two groups. By concordance we mean that when cytology was abnormal corresponding FISH was positive and vice a versa.
Results: Concordance between cytology and FISH was 84% (174) in the Cotesting group (total 208) whereas it was lower in the Non-Cotesting group 74% [(75), (from the total of 101)]. The FISH and cytology results in the Cotesting group were concordant in 10% more cases than in the non-cotesting group. We had 399 FISH studies in our institution during 2009 out of which 100 (25%) were uninformative (insufficient number of cells). We had surgical follow up in 84 cases in which sensitivity and specificity for cytology was 42%, 65%; and for was FISH 29%, 79% respectively.
Conclusions: Our study shows that when FISH and cytology were ordered on different urine sample they showed 10% lower concordance. Based on our results we propose an algorithmic approach for ordering FISH testing. Cytology should always be ordered when ordering FISH. If cytology does not show sufficient number of cells then a repeat sample for FISH analysis should be requested. This practice will not only give a better concordance between cytology and FISH, but will also eliminate FISH testing on sparsely cellular specimens with uninformative results.
Category: Genitourinary (including renal tumors)
Tuesday, March 1, 2011 9:30 AM
Poster Session III # 189, Tuesday Morning