24
IASLC ATLAS OF ALK TESTING IN LUNG CANCER
To minimize technical bias, we 1st Reader-50 tumor cells recommend a two-step assessment strategy with two independent <10% Positive >50% Positive 10%-50% Positive scorers (Figure 11). The first scorer scores 50 tumor cells. A rate of Equivocal rearrangement-positive cells less than 10% (i.e., rearrangement in 2nd Reader-50 tumor cells fewer than five of the 50 cells) is 1st + 2nd Readers-100 tumor cells considered negative; a rate greater than 50% (i.e., more than 25 of 50 <15% Positive ?15% Positive cells) is considered positive; and a Specimen is Specimen is rate of 10% to 50% (i.e., 5 to 25 of positive for ALK negative for ALK 50 cells) is considered equivocal rearrangement rearrangement and additional scoring should be done. In that case, a second inde- Figure 11. Recommended scoring algorithm for ALK FISH. pendent scorer scores an additional 50 tumor cells, and a final rate of rearrangement-positive cells is calculated on the sum of the first and second scores. If the final rate is 15% or more, the specimen is interpreted as positive for ALK gene rearrangement; if the rate is less than 15%, the specimen is interpreted as negative for ALK gene rearrangement. Note: The 15% cutoff is primarily based on testing with the Vysis LSI ALK Break Apart FISH Probe Kit (Abbott Molecular) and should be validated when a different reagent is used.
Laboratory Validation
The ALK FISH assay should be properly validated in the laboratory before testing is offered in a clinical setting (Halling 2012, Saxe 2012). The accuracy of the results—that is, the degree to which the assay discriminates between normal (ALK negative) and abnormal (ALK positive) should be — compared with the accuracy at another laboratory where the validated assay is being performed properly and/or compared with the accuracy for a previously validated method in the same laboratory. The precision or reproducibility of results should be verified according to the degree of agreement between measurements conducted on the same specimen by different technologists and/or at different times, and the entire analytic process should be verified. Verification of accuracy and precision should be repeated periodically. Moreover, the analytic sensitivity and specificity of the assay should be verified in specimens with known genotype. ALK wild-type NSCLC and benign tissue must be evaluated to assess the distance in a true-positive split signal (a distance of at least two times the signal diameter). These tasks are simpler when commercial reagents are used and require higher level of details when laboratory-developed reagents are used.
Challenges
ALK break-apart FISH has been associated with four primary challenges: false-negative and falsepositive results, atypical FISH signal profiles, borderline rates of rearrangement-positive cells, and the need for repeat testing.