Detection of ALK and ROS1 rearrangements using Next Generation Sequencing in lung cancer patients: comparison between FISH, IHC and NGS

Sergi Clavé1,2, Natalia Rodon3, Lara Pijuan1, Alba Dalmases1,2, Olga Díaz3, Álvaro Taus4, Marta Lorenzo1,2, Pedro Rocha4, Ana M. Muñoz-Mármol5, Glòria Oliveras6, Joaquim Bosch6, Blanca Espinet1,2, Beatriz Bellosillo1,2, Xavier Puig3, Edurne Arriola2,4, Marta Salido1,2.

1 Servei de Patologia, Hospital del Mar, Barcelona. 2 Programa de Recerca en Càncer, Institut Hospital del Mar d’Investigacions Mèdiques, Barcelona. 3 BIOPAT. Biopatologia Molecular SL, Grup Assistència, Barcelona. 4 Servei d’Oncologia, Hospital del Mar, Barcelona. 5 Servei d’Anatomia Patològica, Hospital Germans Trias i Pujol, Badalona. 6 Departament d’Oncologia Mèdica, Hospital Universitari Dr. Josep Trueta, Girona.

Detection of ALK and ROS1 rearrangements in non-small cell lung cancer (NSCLC) is required for directing patient care. While fluorescence in situ hybridization (FISH) and immunohistochemistry (IHC) have been established as gold standard methods, next generation sequencing (NGS) platforms are called to be at least equally successful, but also more compatible with multiplexing and diagnostic workflows. Our aim was to investigate the performance of NGS in the detection of rearranged cases.


Thirty NSCLC samples were selected retrospectively from our database (n=2.399) based on previous ALK (n=25) and ROS1 (n=5) FISH results (positive or inconclusive) and material availability. Cases were tested by both FISH (Abbott Molecular) and IHC (Cell Signaling) in paraffin blocks, and were reviewed centrally to determine the tumor area and content percent. Both DNA and RNA were manually extracted from 4-6 unstained sections of 10-μm. Ion Torrent sequencing technology with Oncomine™ Focus Assay (ThermoFisher Scientific) were applied using 10 ng of DNA and RNA from each sample.


Patients characteristics were: median age 62 years, 17 were males and adenocarcinoma (ADC) was the most common histology (72%). Regarding FISH results, 14 cases had split signals, 14 had isolated 3’ signals, and two had negative FISH patterns with isolated 5’ signals. Testing with IHC, four out of the 30 cases had negative results: as expected, two cases with FISH isolated 5’ signals were negative and also two cases with FISH isolated 3’ signals (discordance FISH vs. IHC). Tumor material came from small biopsies in 18 cases, from cytology specimens in six cases and from surgically resected patients in six cases. Ten cases (33%) were non-evaluable by NGS due to insufficient sequencing coverage (seven of these 10 cases were small biopsies with low DNA and RNA input). NGS technology detected positive ALK and ROS1 fusions in 75% of the samples (15 out of 20 assessable samples), being EML4-ALK (E13;A20) and EZR-ROS1 (E10;R34) the most prevalent. Regarding the five negative NGS result, two showed isolated ALK 5’ signals considered negative by FISH; two had isolated ALK 3’ signals and were also IHC negative (histologically both cases were non-ADC); and one case which was positive for ALK rearrangement by FISH and IHC was reviewed by both techniques in a newly unstained section and re-classified as non-rearranged.


NGS technology for detecting ALK and ROS1 rearrangements in NSCLC could be considered as a screening test although the success rate is closely related to the correct evaluation of the initial amount of tumor tissue, particularly in small biopsies. NGS technology could be used as an additional molecular technique for cases with unexpectedly inconclusive or discordant FISH/IHC results.

Texto completo

Abstract exhibited at the: