CLINICAL UTILITY AND DIAGNOSTIC CONCORDANCE OF ALK IMMUNOHISTOCHEMISTRY AND FISH IN NON SMALL CELL LUNG CANCER: A SYSTEMATIC REVIEW AND META-ANALYSIS

Abstract

Background: Detection of anaplastic lymphoma kinase (ALK) rearrangements guides targeted treatment in non small cell lung cancer (NSCLC). Immunohistochemistry (IHC) and fluorescence in situ hybridization (FISH) interrogate different biological features, and discordant results may influence treatment eligibility. Objective: To evaluate concordance and diagnostic accuracy of ALK IHC relative to FISH and assess the clinical significance of discordant results. Methods: The PRISMA 2020 study-selection diagram includes 728 database records, 574 title-and-abstract records screened, 43 full-text reports assessed, and four studies included. Original diagnostic studies reporting paired ALK IHC and FISH results were evaluated through January 2026. Sensitivity and specificity relative to FISH were synthesized using separate random-effects inverse-variance models with logit transformation. Additional studies describing discordant findings were reviewed qualitatively. Results: Four cohorts contributed 3,241 paired evaluable specimens, with 3,184 concordant results (98.24%). Pooled sensitivity of IHC relative to FISH was 86.3% (95% CI 81.8–89.9%) and pooled specificity was 99.3% (95% CI 98.9–99.5%). Individual sensitivities ranged from 85.1% to 100%, while specificities ranged from 99.3% to 100%. In the largest cohort, 33 specimens were FISH-positive/IHC-negative and 17 were FISH-negative/IHC positive. RNA-based reassessment of discrepant cases demonstrated that FISH positivity without protein expression did not always indicate a functional ALK fusion. Conclusion: Validated ALK IHC exhibits excellent overall agreement and specificity relative to FISH. Discordance is infrequent but clinically relevant, especially near FISH positivity thresholds or with atypical signals. Orthogonal molecular testing should be considered for equivocal or discordant cases. Generalization of pooled estimates is limited by study selection, assay differences and the imperfect reference-standard properties of FISH.