RET Fusion Testing in Patients With NSCLC: The RETING Study

Abstract

Introduction: RET inhibitors with impressive overall response rates are now available for patients with NSCLC, yet the identification of RET fusions remains a difficult challenge. Most guidelines encourage the upfront use of next-generation sequencing (NGS), or alternatively, fluorescence in situ hybridization (FISH) or reverse transcriptase-polymerase chain reaction (RT-PCR) when NGS is not possible or available. Taken together, the suboptimal performance of single-analyte assays to detect RET fusions, although consistent with the notion of encouraging universal NGS, is currently widening some of the clinical practice gaps in the implementation of predictive biomarkers in patients with advanced NSCLC.

Methods: This situation prompted us to evaluate several RET assays in a large multicenter cohort of RET fusion-positive NSCLC (n = 38) to obtain real-world data. In addition to RNA-based NGS (the criterion standard method), all positive specimens underwent break-apart RET FISH with two different assays and were also tested by an RT-PCR assay.

Results: The most common RET partners were KIF5B (78.9%), followed by CCDC6 (15.8%). The two RET NGS-positive but FISH-negative samples contained a KIF5B(15)-RET(12) fusion. The three RET fusions not identified with RT-PCR were AKAP13(35)-RET(12), KIF5B(24)-RET(9) and KIF5B(24)-RET(11). All three false-negative RT-PCR cases were FISH-positive, exhibited a typical break-apart pattern, and contained a very high number of positive tumor cells with both FISH assays. Signet ring cells, psammoma bodies, and pleomorphic features were frequently observed (in 34.2%, 39.5%, and 39.5% of tumors, respectively).

Conclusions: In-depth knowledge of the advantages and disadvantages of the different RET testing methodologies could help clinical and molecular tumor boards implement and maintain sensible algorithms for the rapid and effective detection of RET fusions in patients with NSCLC. The likelihood of RET false-negative results with both FISH and RT-PCR reinforces the need for upfront NGS in patients with NSCLC.

Keywords: FISH; Lung carcinoma; Next-generation sequencing; RET fusions; RT-PCR.

Figure 1

Flowchart of samples in the RETING study. FISH, fluorescence in situ hybridization; NGS, next-generation sequencing; RT-PCR, reverse transcriptase–polymerase chain reaction.

Figure 2

Representative examples of RET FISH-positive NSCLCs using the Vysis RET Probe (A,B) and the ZytoVision RET Probe (C,D). (A,C) A typical break-apart pattern is shown with one fused signal and one break-apart signal per nucleus (arrows). (B,D) An isolated 3’ signal pattern is depicted (red signals with the Vysis probe and green signals with the ZytoVision probe) (arrows). All four cases were scored using the BioView Duet scoring system and were RET NGS-positive. See text for details. Original magnification: x1000. FISH, fluorescence in situ hybridization; NGS, next-generation sequencing.

Figure 3

Representative examples of RET FISH patterns: borderline break-apart positive (A), false-negative (B), and (C) typical negative fusion signal pattern (C). (A) A tumor with a CCDC6-RET fusion showing a borderline break-apart positive pattern. (B) A tumor with a KIF5B-RET fusion showing insufficient separation between the red and the green signals (i.e., FISH false-negative). (C) A typical example of a tumor without RET fusions exhibits two fused signals. All images correspond to the Vysis RET probe and were interpreted using the BioView Duet scoring system. The fusion status was confirmed by NGS. See text for details. Original magnification: x1000. FISH, fluorescence in situ hybridization.

Figure 4

Typical features of NSCLC with RET fusions. (A) signet ring cells, (B) psammoma bodies, and (C) pleomorphic nuclei (hematoxylin-eosin, original magnification X200 [A-C]).