Article Text

Download PDFPDF

IHC versus FISH versus NGS to detect ALK gene rearrangement in NSCLC: all questions answered?
  1. Ullas Batra1,
  2. Shrinidhi Nathany2,
  3. Mansi Sharma1,
  4. Sunil Pasricha3,
  5. Abhishek Bansal4,
  6. Parveen Jain1,
  7. Anurag Mehta5
  1. 1 Medical Oncology, Rajiv Gandhi Cancer Institute and Research Centre, New Delhi, India
  2. 2 Molecular Diagnostics, Rajiv Gandhi Cancer Institute and Research Centre, New Delhi, India
  3. 3 Pathology, Rajiv Gandhi Cancer Institute and Research Centre, New Delhi, India
  4. 4 Radiology, Rajiv Gandhi Cancer Institute and Research Centre, New Delhi, India
  5. 5 Laboratory Services, Transfusion Medicine and Research, Rajiv Gandhi Cancer Institute and Research Centre, New Delhi, India
  1. Correspondence to Dr Ullas Batra, Medical Oncology, Rajiv Gandhi Cancer Institute and Research Centre, New Delhi 110085, India; ullasbatra{at}gmail.com

Abstract

Aims Anaplastic lymphoma kinase (ALK) rearranged non-small cell lung carcinoma (NSCLC) is a distinct molecular subtype and rapid approval of ALK tyrosine kinase inhibitors (TKIs) has necessitated rapid and sensitive diagnostic modalities for the detection of this alteration. Gene rearrangements can be identified using many techniques including fluorescence in situ hybridisation (FISH), reverse transcriptase-PCR, next-generation sequencing (NGS) and immunohistochemistry (IHC) for fusion oncoprotein expression. We aimed to determine the concordance between IHC, FISH and NGS for ALK biomarker detection, and determine differences in sensitivity, and survival outcomes.

Methods We analysed the concordance between IHC using D5F3 monoclonal antibody, FISH (break-apart) and NGS using a custom panel containing 71 different ALK variants.

Results Among 71 cases included in this study, FISH was evaluable in 58 cases. The concordance of ALK IHC with FISH was 75.9% and that with NGS was 84.5%. The sensitivities of FISH and NGS were 75.6% and 87.5%, respectively. The median progression-free survival of ALK IHC-positive and FISH-negative group was 5.5 months and that of both positive was 9.97 months.

Conclusion Although NGS offers a better throughput and visualisation, IHC still remains the quintessential screening tool in upfront diagnosis of ALK rearranged NSCLC.

  • genes
  • neoplasm
  • lung neoplasms
  • immunohistochemistry
  • diagnostic techniques and procedures
  • medical oncology

Data availability statement

Data are available on reasonable request. All data relevant to the manuscript have been presented here. Any additional data required will be made available from the corresponding author on reasonable request.

Statistics from Altmetric.com

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.

Introduction

Anaplastic lymphoma kinase (ALK) rearranged non-small cell lung carcinoma (NSCLC) has emerged as a distinct entity and has gained importance since 2007, owing to rapid developments and approvals of various generations of ALK tyrosine kinase inhibitors (TKIs).1 Landmark trials like ALEX, ASCEND and PROFILE 10142 have confirmed the superiority of ALK TKIs over chemotherapy in this subset of patients.3 Owing to the rapid advances in the knowledge and clinicians armamentarium of targeted drugs, robust diagnostic modalities are the need of the hour to accurately and precisely detect ALK rearrangements. To date, three representative methods have been used which include immunohistochemistry (IHC), fluorescence in situ hybridisation (FISH) and reverse transcriptase-PCR (RT-PCR).4

IHC using D5F3 monoclonal antibody on Ventana system has been evaluated in multicentre trials, and is currently approved by the Food and Drug Administration.5 Trials and studies on alectinib have relied solely on IHC-based testing, and confirmed higher response rates.3 However, clinical trials that led to the approval of crizotinib were based on FISH, and hence it still is considered confirmatory. Although discrepancies exist between the two, as assessments using FISH face myriad technical issues like signal instability and user-dependent variability in interpretation. Both these techniques are limited by the premise that they cannot characterise fusion partners/breakpoints.6

RT-PCR although has been established as a diagnostic modality owing to its high sensitivity and specificity, and its ability to characterise fusion partners, the yield of good-quality RNA from formalin-fixed paraffin-embedded tumour blocks is the main challenge. This limits its role as first choice diagnostic modality in the clinical setting.7

Targeted RNA sequencing using direct sequencing methods like Sanger’s sequencing and next-generation sequencing (NGS) technologies have emerged and are showing promise. Sanger sequencing is a gold standard platform, which is limited by its sensitivity and requirement of higher quantities of input RNA. NGS overcomes these problems, with higher throughput and massive parallel sequencing.8 However, data on this are limited to a few retrospective studies from the real world and smaller sample size prospective studies. This has not been validated in clinical trials and hence needs further elucidation. This study was a single-centre study evaluating the concordance rates between IHC, FISH and NGS techniques for detection of ALK rearranged NSCLC, and the sensitivity of FISH and NGS against the more widely used IHC method.

Methods

Study design and patient samples

This a prospective single-centre observational study performed between September 2020 and November 2020. Patients aged >18 years of age with biopsy-proven NSCLC and ALK IHC positive between March 2018 and May 2020 with adequate tumour blocks were included in this study. Patients with depleted tissue in FFPE blocks were not included in this study. No extra samples were obtained from the patients. All consecutive ALK IHC-positive cases with adequate tissue blocks were enrolled.

Diagnostic modalities

All patients underwent ALK IHC as a part of routine diagnostic testing using the Ventana (D5F3) CDx Assay (Ventana Medical Systems, Tucson, Arizona, USA). D5F3 is a primary rabbit monoclonal ALK antibody (mAb) clone obtained from Ventana USA. IHC staining was performed using 4 μm thick sections on Ventana Benchmark XT immunostainer (Ventana Medical Systems, Illkirch Graffenstaden, France). The samples were evaluated independently by two pathologists and were scored as IHC positive and IHC negative according to the manufacturer’s recommendations. ALK positive was defined as intense granular cytoplasmic staining. Of these, only ALK IHC-positive cases were included in the study.

Fluorescence in situ hybridisation

All the cases with adequate amounts of tissue in the FFPE tissue blocks were subjected to FISH, using ZytoVision ALK Break Apart FISH probe kit (ZytoVision, Germany). The process and staining were performed as per the manufacturer’s recommendations. Positive cases were defined as those exhibiting split signals with the separation being more than two signal diameters apart. At least 50 tumour cells for each section were counted and interpretation was done independently by two pathologists. Cases with depleted tissue and signal instability/broken signals were classified as unevaluable.

Next-generation sequencing assay

ALK fusions were characterised by NGS in all the cases. A custom RNA-based assay was designed consisting of 71 ALK variants and 12 assay expression controls, using the Ampliseq Designer software. The FFPE tumour block was evaluated on an H&E-stained slide section to ascertain adequate tumour cellularity (at least 30%) and the area was carefully macrodissected to include maximum tumour cell content. After adequate optimisation of the panel on normal samples, RNA was extracted using RNA extraction kit from Promega, and reverse transcribed to complementary DNA using Vilo SuperScript cDNA kit. Template preparation and enrichment were performed using Ion OneTouch Select Template Kit (all from Thermo Fisher Scientific) on Ion OneTouch 2. The prepared libraries were quality checked using Tape Station with high-sensitivity DNA kits, to ensure adequate size and cleanup. The final libraries were sequenced on Ion Personal Genome Machine or the Ion S5 Sequencer using a 318 chip. The raw data were analysed using the Torrent Suite plugin. The analysed reads were checked for quality and presence of >5000 total mapped reads. Samples with <5000 mapped reads were rejected and repeated from the library preparation stage. The variants were called using the Torrent Variant Caller, and Ion Reporter Software and the validity of the called variant was ascertained by visualising the calls on integrative genomic viewer for the presence of at least 30 clean fusion reads. Panel includes two process control target genes, TBP and HMBS. At least one control must have a molecular count of >2 to pass quality checks (as per manufacturer’s protocols).

Statistical analysis

Statistical calculations and analysis were done using MedCalC Softare (Ostend, Belgium).Descriptive statistics were used to summarise demographics and clinical characteristics.Categorical variables were reported as frequency and percentage. For the calculation of concordance estimates, only cases recorded as evaluable were included. The FISH unevaluable group was excluded from this analysis. To calculate the concordance and sensitivity, standard 2×2 tables were constructed as shown in table 1 and the following formulae were used. Specificity could not be calculated as the inclusion criteria of the study included ALK IHC-positive cases only.

Table 1

Calculation and interpretation for sensitivity and concordance rates

The study was frozen on 31 May 2020 for the last follow-up for interim analysis. The progression-free survival (PFS) was defined as time between date of initiation of ALK-directed TKI up to date of progression (evaluated radiologically)/death. The overall survival (OS) was defined as the time between the date of diagnosis and date of death/last follow-up. The patients who were lost to follow-up or alive up to last follow-up date were censored. While computing Kaplan-Meier curves, log rank tests could not be performed owing to limited numbers in the various subgroups.

Concordance rate=100%×(a+d)/(a+b+c+d).

Sensitivity=100%×d/(b+d).

Results

Demographics and clinical features

The median age of the cohort was 53 years (range: 23–71 years) with a slight male preponderance (n=37, 52.1%; females: 34, 47.9%). The clinicopathological profiles of these patients are depicted in table 2. A total of 60 patients received TKI treatment as first line: crizotinib, 7 ceritinib and 10 alectinib.

Table 2

Baseline characteristics of the study population

ALK evaluable cases

A total of 71 cases were included which underwent ALK testing using all the three modalities, that is, IHC, FISH and NGS. Among these, there were 13 (18.3%) unevaluable cases due to signal instability and tumour depletion, and hence 58 cases were available to assess the sensitivity, agreement and concordance rates. An overview of the concordance and discordance between all the three techniques is depicted in figure 1. A Venn diagram depicting overlap of the three modalities is depicted in figure 2

Figure 1

Consort diagram depicting an overview of the 71 cases included in this study. ALK, anaplastic lymphoma kinase; FISH, fluorescence in situ hybridizsation; IHC, immunohistochemistry; NSCLC, non-small cell lung carcinoma; NGS, next-generation sequencing.

Figure 2

Venn diagram showing overlap of the three modalities in the study cohort. FISH, fluorescence in situ hybridisation; IHC, immunohistochemistry; NGS, next-generation sequencing.

Figure 4

Kaplan-Meier curve showing progression-free survival among FISH-negative (0), FISH-positive (1) and FISH-unevaluable (2) groups. FISH, fluorescence in situ hybridisation; TKI, tyrosine kinase inhibitor.

Comparison of ALK status assessed by FISH versus IHC

Among the 58 evaluable cases of FISH, there was concordance for a positive result in 44 (75.9%) cases. The rest of the 14 (24.1%) cases were negative on FISH for ALK rearrangement, however were positive for ALK IHC. The sensitivity was 75.6%.

Comparison of ALK status assessed by NGS versus IHC

The comparison was made in the cohort which was tested for all the three modalities, hence 58 cases were evaluated of which 49 cases were concordant between IHC and NGS. The concordance rate was 84.5%. The most commonly encountered fusion partner was EML4 in all the 49 cases, among which variant 1 (EML4exon13-ALKexon20) was seen in 22 cases, followed by variant 3 (EML4exon6a/b-ALKexon20) in 13 cases and variant 2 (EML4exon20-ALKexon20) in 9 cases. One case had variant 5a (EML4exon2-ALKexon20) and three cases showed novel breakpoints in EML4. The characterisation of variants is important as different fusion transcripts may have different protein stabilities and hence differential responses to various ALK TKIs. Discordance was seen in nine (15.5%) cases. The sensitivity was calculated as 87.5% for NGS. Among the nine discordant cases, three cases were also negative by FISH. The other six cases however were both IHC and FISH positive.

Comparison of ALK status assessed by NGS and FISH

Among the 58 cases, 38 (65.5%) cases showed concordance for presence of an ALK rearrangement by both NGS and FISH and 3 cases showed concordance for a negative result. Among the 17 (29.3%) discordant cases, 11 cases were NGS positive, but FISH negative. However, only six cases were FISH positive, and NGS negative.

Survival outcomes

Only 28 patients were included for survival analysis. Of the remaining 30 cases, 11 cases did not receive treatment and 7 cases died/discontinued the drug before first response evaluation.

The other 12 cases were newly recruited and hence have not had first response evaluation at the time of study end date. The median follow-up was 12.5 months. The median OS for IHC-positive, FISH-positive and IHC-positive and FISH-negative groups were not reached. With respect to the PFS, the median PFS of IHC-positive, FISH-positive group was 9.97 months (95% CI 7.01 to 14.3 months), whereas in the IHC-positive, FISH-negative group was 5.5 months (95% CI 2.8 to 7.6 months) (figure 3). Among the IHC-positive and FISH-unevaluable group, out of 13 cases, 11 patients received ALK TKI (8: crizotinib, 1: ceritinib and 2: alectinib), of which 6 patients on crizotinib progressed. The median PFS in the FISH-unevaluable group on crizotinib was 12.13 months. However, the values did not reach statistical significance as the numbers in the response evaluation group were small and this is an interim report.

Figure 3

Kaplan-Meier curve showing progression-free survival among next-generation sequencing (NGS)-negative (0) and NGS (1) groups. NS, next generation sequencing; TKI, tyrosine kinase inhibitor.

Among the IHC-positive and NGS-negative groups, all nine cases received ALK TKI (seven: crizotinib, two: ceritinib), of which three cases progressed (two on crizotinib and one on ceritinib) of which one was FISH negative, while the other two were FISH positive. Due to small sample size in NGS and FISH discordant groups, PFS could not be derived. The Kaplan-Meier curves for PFS are depicted in figures 2 and 4.

Discussion

In this early exploratory analysis of patients with advanced ALK rearranged NSCLC, 81.6% (58/71) had a valid FISH result. Among these, 75.9% (44/58) cases were both IHC and FISH positive, whereas, 24.1% (14/58) were IHC positive and FISH negative. A valid ALK FISH result could not be obtained in 13 (19.4%) cases owing to inconclusive results, signal intensity problems or depleted tumour blocks. The concordance for positive results for FISH was 75.9% with IHC and 65.5% with NGS. Since this is an interim analysis and the study has not reached its end point, the survival outcomes did not reach statistical significance.

In the ALEX study in 303 patients of ALK rearranged advanced NSCLC, FISH was not evaluable in 20.1% cases owing to inconclusive results/sample unavailability.9 The concordance between IHC and FISH for positive results observed in ALEX was 83.9%, whereas in PROFILE 10142 it was 85%. This is higher than that observed in our early experience (75.9%) which may be attributed to pre-analytics including tissue availability as well as observer-dependent variability. In a large meta-analysis of 42 studies, pooled data of >11 000 ALK rearranged NSCLC cases the concordance rate was 80.6% (95% CI 73.3 to 86.1).10 Hence, it is clearly evident that the percentage variation is not constant and hence may be attributed to variabilities in pre-analytical factors and interpretative dilemmas. Several other studies have evaluated the concordance between these techniques, for example, Pekar-Zlotin et al 11 found the most accurate of the three modalities discussed herein, however, the numbers of ALK-positive cases were limited (n=5); Xu et al 12 evaluated concordance of Ventana IHC for ALK with RTPCR in 1720 patients with NSCLC; Lin et al 13 used IHC, FISH and NGS in 55 patients and they did not find any difference in overall response rate, disease control rates among the various groups. In another study evaluating IHC and NGS,14 they found that there was no statistically significant difference between the two subgroups.

This discrepancy poses a diagnostic dilemma for the molecular pathologist, and a therapeutic dilemma to the treating clinician in order to optimally strategise the institution of ALK-directed therapy, especially in treatment-naïve cases. These discrepancies may result from a multitude of factors including both technical and ALK biology-related factors. Technical factors may result from poor tissue processing and fixation causing fracture of FISH signals, and non-specific staining.15 With respect to NGS, the lability of RNA, the laborious library preparation steps may cause technical discrepancies. Biological factors leading to discrepant results may be due to ALK amplification giving rise to false IHC positivity, although this is reported to be of rare occurrence.4 ALK IHC therefore currently outscores both NGS and FISH in terms of initial screening, however, the need for confirmatory testing is still unmet and is a matter of debate.

In terms of survival, the ALEX study reported a median PFS of 22.4 months for alectinib in the IHC-positive and FISH-uninformative group and 1.33 months in the IHC-positive/FISH-negative group.3,9 In our study however, the PFS was evaluated with respect to crizotinib and showed 12.13 and 5.5 months, respectively. This may be attributed to the smaller proportion of patients who took ALK TKI therapy, which prevents us from making a conclusive remark on the survival outcomes at this point. Nevertheless, similar trends of relatively lower rates of response and short PFS have been observed in other real-world experiences and controlled trials as well.

Interestingly, a subset of this population has not progressed indicating that they are true IHC positive and false FISH negative. Conversely, a smaller subset in both positive groups have also not shown durable response to ALK TKI, which may be due to either sample size limitations, imbalance in confounding factors like smoking, age, performance status or maybe related to copy number changes in ALK gene. However, NGS may resolve these discrepancies by accurately characterising the fusion partners. McLeer-Florin et al 16 demonstrated high sensitivity and specificity of NGS-based testing in 76 cases of IHC and FISH discordant cases. However, in our cohort among the 14 IHC and FISH discordant cases, NGS was positive in 11 cases, 3 cases were also negative on NGS. Whether the breakpoints involved affect patterns of FISH signals or interpretation needs further validation in larger cohorts. IHC-negative cases were not included in this study as this was the initial screening tool used. This may limit the calculation of negative predictive value, which will require larger cohorts.

NGS-based testing is fast emerging as a one stop solution in lung cancer diagnostics. However, from our experience, it is evident that IHC, especially in ALK, has its own merits. Availability, affordability and sensitivity of IHC makes it irreplaceable. NGS offers higher throughput and additional information owing to the multiplexing, however, cannot be considered as a true surrogate for IHC.

Take home messages

  • Conventional diagnostic modalities like immunohistochemistry (IHC), fluorescence in situ hybridisation (FISH) and reverse transcriptase-PCR for anaplastic lymphoma kinase rearrangement detection have been used since its discovery.

  • Although FISH can detect chromosome rearrangements, it suffers from inherent problems of signal intensity variation and interobserver variability.

  • Although next-generation sequencing (NGS) offers partner characterisation, higher throughput and broader panel-based testing, it is limited by sensitivity and laborious workflow processes.

  • As is evident from ALEX study data, IHC still remains the modality of choice for screening, which can further be confirmed using NGS for variant characterisation, which may aid in therapy decision making.

Data availability statement

Data are available on reasonable request. All data relevant to the manuscript have been presented here. Any additional data required will be made available from the corresponding author on reasonable request.

Ethics statements

Patient consent for publication

Ethics approval

The study protocol was approved by the institutional review board and scientific committee and was in accordance with the Declaration of Helsinki.

References

Footnotes

  • Handling editor Runjan Chetty.

  • Contributors All authors have contributed to the conductance of the study and preparation of this manuscript, as well all authors have reviewed the final manuscript.

  • Funding This work was supported by an educational grant funded by Pfizer Products India Private Limited (Number: 63631449).

  • Competing interests None declared.

  • Provenance and peer review Not commissioned; externally peer reviewed.