Send to

Choose Destination
J Thorac Oncol. 2015 Apr;10(4):595-602. doi: 10.1097/JTO.0000000000000444.

ALK testing in lung adenocarcinoma: technical aspects to improve FISH evaluation in daily practice.

Author information

*Institute of Pathology, Locarno, Switzerland; Departments of †Surgical and Morphological Sciences and §Department of Pathology, Ospedale di Circolo, Varese, Italy; and ‡Department of Molecular and Translational Medicine, Section of Pathology, School of Medicine, University of Brescia, Brescia, Italy.



Anaplastic lymphoma kinase (ALK) gene rearrangement characterizes a subgroup of patients with lung adenocarcinoma who may benefit from ALK inhibitors. Fluorescence in situ hybridization (FISH) with a break-apart/split-signal strategy is the gold standard to investigate ALK. The cutoff to define ALK positivity has been settled at 15% or greater. A subset of patients has ALK borderline status, showing 15% ± 5% positive cells. Several aspects, both biological and technical, might influence signals evaluation, making FISH interpretation a challenging task. To improve ALK evaluation, we classified the different FISH patterns on the basis of the type of the split signals, namely short, long, far away, and deleted.


We investigated ALK gene status by FISH in 244 lung adenocarcinomas and in a series of ALK negative cell lines samples, collected in three Institutions.


ALK positive profile was found in 12% of patients; long, deleted, and far-away splits were the primary patterns observed. ALK borderline profile characterized 10% of samples; long and deleted splits were significantly more frequent in those borderline finally classified as ALK positive, whereas short split were mostly detected in those borderline patients finally classified as ALK negative (p = 3.4 × 10). In the ALK negative control series, short split was the predominant pattern. Concordance was observed among different operators and probes for both samples and controls.


Difficulties in ALK FISH signal interpretation might be bypassed using this detailed scoring system, which is highly reproducible, helps clarify borderline samples (according to split type), and provides experimental evidence that 15% is a reasonable cutoff to overcome the assay-dependent background noise.

[Indexed for MEDLINE]
Free full text

Supplemental Content

Full text links

Icon for Elsevier Science
Loading ...
Support Center