The combination of FISH and IHC for ALK assessment in NSCLC is more powerful than either technique alone

Aim Examination of the proposition that detection of ALK overexpression by IHC in NSCLC may act as triage or replacement for FISH. Method Results on 221 NSCLCs referred for ALK FISH testing (Abbott-Vysis LSI ALK Breakapart FISH Probe) were reviewed, FISH hybridisation pattern classified and compared to ALK IHC (Clone D5F3). Results 10 cases (4.5%) had insufficient material for testing. Thirty-four cases were FISH +ve, four equivocal (1.8%) and 173 negative. Seventeen positive cases had classic ‘split’ signal; 14 had isolated 3’ red signals, two had combined pattern and two cases (both IHC +ve) had variant patterns with additional 5’ (green) signals (previously regarded as FISH–ve). Twenty-nine of these 34 cases were IHC +ve (85% IHC sensitivity), three were IHC equivocal, 1 was untested, and 1 IHC–ve. Two FISH equivocal cases (minimal split) were strongly IHC +ve. One of 173 FISH–ve cases was strongly IHC +ve. IHC on 196 cases scored 32 as positive, 14 equivocal (7%) and 150 negative. 25 IHC +ve cases were straightforward FISH +ve, six had variant/equivocal FISH patterns, one was FISH–ve (97% IHC specificity). Three of 14 IHC equivocal cases were FISH +ve, as was one IHC–ve case. Discussion Combination of IHC and FISH for detection of ALK mutation is more powerful than either test in isolation. Without availability of IHC, six cases with equivocal or variant FISH patterns would likely have been missed, however four times more cases are equivocal by IHC than by FISH. Occasional cases (2 in this cohort: