Accurate detection of altered anaplastic lymphoma kinase (ALK) expression is critical for the selection of lung cancer patients eligible for ALK-targeted therapies. To overcome intrinsic limitations and discrepancies of currently available companion diagnostics for ALK, we developed a simple, affordable and objective PCR-based predictive model for the quantitative measurement of any ALK fusion as well as wild-type ALK upregulation. This method, optimized for low-quantity/-quality RNA from FFPE samples, combines cDNA pre-amplification with ad hoc generated calibration curves. All the models we derived yielded concordant predictions when applied to a cohort of 51 lung tumors, and correctly identified all 17 ALK FISH-positive and 33 of the 34 ALK FISH-negative samples. The one discrepant case was confirmed as positive by IHC, thus raising the accuracy of our test to 100%. Importantly, our method was accurate when using low amounts of input RNA (10 ng), also in FFPE samples with limited tumor cellularity (5-10%) and in FFPE cytology specimens. Thus, our test is an easily implementable diagnostic tool for the rapid, efficacious and cost-effective screening of ALK status in patients with lung cancer.
Keywords: ALK; FFPE NSCLC; RT-qPCR; cytology specimens; inverse prediction.