Evaluation of ALK gene rearrangement in central nervous system metastases of non-small-cell lung cancer using two-step RT-PCR technique

Lack of high quality of mRNA in FFPE samples, as well as too high sensitivity of RT-PCR assays are the main reasons of RT-PCR failure in screening of ALK gene rearrangement [13, 18]. Therefore, guidelines primarily recommend IHC and FISH techniques in selection of patients for treatment with ALK-TKIs [8]. On the other hand, comprehensive RT-PCR allows to identify the fusion partners for ALK gene and may have an informative value in prediction of effectiveness of treatment and patients’ survival [8, 9, 11, 13]. It was indicated that diversity of ALK fusion partners may influence on therapeutic efficacy of ALK inhibitors. In the future, RT-PCR assays could be used for monitoring of the effectiveness of treatment in mRNA isolated from liquid biopsy [3, 13, 19]. To date, utility of RT-PCR based on originally designed probes (with re-sequencing) or commercially available AmoyDx EML4-ALK Fusion Gene Detection Kit (Amoy Diagnostics, China) was described in screening for ALK rearrangement in NSCLC patients [11‐13]. Moreover, only variants 1 and 3a/b which represent 80% of all ALK fusion partners were broadly described in literature and they were revealed in clinical utility [1, 4, 12].

In the following study, we used EML4-ALK Fusion Gene Detection Kit (Entrogen, USA) that allowed to find out the mRNA containing abnormal ALK sequences in 21% of CNS metastases from NSCLC (30/145). All positive results were observed during reaction with first set of probes covering seven fusion variants (1, 2, 3a/b, 5a/b and 6). The amplification was not observed in second set of probes covering next two variants (4 and 7) Our study was the first RT-PCR analysis of expression of mRNA containing ALK abnormal sequences in CNS metastases of NSCLC worldwide. However, frequency of abnormal mRNA expression remained in compliance with data obtained by Wallander et al., Zhang et al., and Teixido et al. These authors observed RT-PCR positive results in 24% (11/4), 43% (9/21) and 12.5% (25/200) of NSCLC primary tumors [7, 11, 12]. In those four studies, incidence of expression of mRNA with ALK abnormal sequences is higher than ALK rearrangement frequency examined by FISH method and incidence of abnormal ALK protein expression examined with IHC techniques [8, 9]. Finally, in our study all truly positive patients had diagnosed lung adenocarcinoma, and all patients with squamous cell carcinoma who had positive RT-PCR results showed IHC and FISH negative results. However, there are single data reporting the incidence of ALK rearrangement in lung squamous cell carcinoma (1.3%) and not-otherwise specified NSCLC (4.5%) [9].

Several studies postulated that in ALK abnormalities diagnosis, RT-PCR shows the higher sensitivity in comparison to IHC and FISH [8, 9, 11‐13]. Moreover, authors indicated that sensitivity and specificity of RT-PCR compared to IHC and FISH is in the range from 95 to 100% [3, 20, 21]. We confirmed that the RT-PCR assay had 100% sensitivity, but low specificity in comparison to FISH technique. Whereas, IHC assay is characterized by higher specificity than RT-PCR when compared to FISH method. However, in our previous study we reported that tissues obtained from CNS metastases of NSCLC contained high number of macrophages, as well as nerve and glial cells, which imitate a false positive background in IHC staining. Moreover, FISH analysis gives relatively often non-diagnostic results [15].

Teixido et al. examined ALK abnormalities in 137 adenocarcinoma patients using three different methods: FISH, IHC and RT-PCR. Authors obtained positive results in 5.8, 10 and 17.5% of patients; respectively. All ALK rearranged patients diagnosed with FISH method were also positive in IHC and RT-PCR assays. However, six IHC-positive patients (43% of all IHC-positive patients) and 16 patients with RT-PCR positive results (67% of all RT-PCR-positive patients) had confirmed diagnosis of ALK rearrangement in FISH method [11]. Wallander et al. reported high concordance between RT-PCR and FISH methods (80% of consistent result), as well as between RT-PCR and IHC methods (70% of consistent result). However, FISH and IHC showed agreement in only 67% of analyzed samples [12]. On the other hand, Zhang et al. observed overall compliance results of IHC and FISH, but RT-PCR assay showed lower sensitivity than IHC method and gave large percentage of negative or equivocal results [7]. Also French multicenter study indicated 39% of false negative RT-PCR results compared to positive IHC/FISH results [22]. Shun et al. reported that the sensitivity of RT-PCR was lower than IHC assay. Furthermore, authors declared that two negative RT-PCR results showed late amplification of PCR products and they also had weak IHC expression [13]. Authors indicated different reasons for discordance between RT-PCR, FISH and IHC results. Wang et al. reported that subtle splitting of red and green FISH signals could be interpreted as false negative FISH results. They also suggested that the cross-contamination of RT-PCR samples may lead to non-specific amplification causing false positive RT-PCR results [11]. Wallander et al. suggested that discrepancies between results may be caused by various detectability of different variants of fusion genes [12].

In our study, discordant results between detectability of ALK disorders using different molecular methods might be caused by the degradation of mRNA in FFPE tissue blocks. However, more likely is the existence of a few nuclei with ALK gene rearrangement in examined specimens, which entitled to RT-PCR positive results and simultaneously to FISH negative results. Negative FISH results could be issued even if the existence of single nuclei with ALK gene rearrangement in the examined slides are visualized (Fig. 1). The slides were obtained from the patient, who was previously classified as true negative using IHC [15]. However, in following study the sample showed presence of ALK abnormal sequence in RT-PCR, therefore, we also performed FISH analysis, which presented the presence of single rearranged cells (Fig. 1d). In FISH method, when the rearrangement-positive cells rate is less than 15%, the specimen is interpreted as negative for ALK gene rearrangement. Indeed, RT-PCR has an extremely high sensitivity that was enough to detect single rearranged cells in analyzed material. Taking into account that only patients harboring at least 15% of rearranged cells in FISH analysis show benefit from ALK-TKIs treatment [8, 9], the utility of RT-PCR assays in qualification of NSCLC patients for targeted therapy appears to be insufficient [11].

However, cell lines studies confirmed associations between fusion variants and sensitivity to ALK inhibitors in patients with ALK gene rearrangement [18, 23, 24]. Yoishida et al. observed longer PFS and higher percentage of disease control after crizotinib administration in patients harboring ALK fusion variant 1 than in patients with other variants of ALK gene rearrangement [18]. Necessity of ALK fusion variants identification may increase the clinical significance of comprehensive RT-PCR analysis [12].

Summarizing, RT-PCR performed with the EML4-ALK Fusion Gene Deletion Kit (Entrogen, USA) allowed us to preselect a huge cohort of patients for diagnosis of ALK gene rearrangement using FISH assay. In FISH reaction, we confirmed the presence of ALK gene rearrangement only in samples with expression of abnormal ALK protein visualized in IHC assay and only in adenocarcinomas. RT-PCR technique showed lower specificity in comparison to IHC and FISH methods with 100% of sensitivity. Therefore, utility of RT-PCR technique both in screening patients for detection of ALK abnormalities and in qualifying patients for molecularly targeted therapies is disputable. However, our analysis allowed to detect ALK gene abnormalities in the CNS metastases of NSCLC using RT-PCR that was first observed in such unique material worldwide.