Skip to main content

Systematic evaluation of PAXgene® tissue fixation for the histopathological and molecular study of lung cancer.

Southwood M, Krenz T, Cant N, Maurya M, Gazdova J, Maxwell P, McGready C, Moseley E, Hughes S, Stewart P, Salto-Tellez M, Groelz D, Rassl D

Publication: J Pathol Clin Res (2020) • Vol 6 • Page 40-54PMID: 31571426View on PubMed

Purpose

This paper compared hematoxylin and eosin (H&E) and immunohistochemical (IHC) staining, DNA and RNA yield, and next-generation sequencing (NGS) results in matched formalin-fixed paraffin embedded (FFPE) and PAXgene fixed paraffin-embedded (PFPE) lung parenchyma and tumor resection specimens, Temno biopsy specimens, and fine-needle aspiration (FNA) specimens. Two different DNA extraction kits were used for FFPE specimens and two NGS workflows were used for both FFPE and PFPE specimens.

Conclusion

Generally, H&E and IHC staining were comparable between FFPE and PFPE specimens, regardless of specimen acquisition method, but some differences were observed. The DNA yield was higher from PFPE than FFPE specimens using either extraction method, but the differences in yield were smaller for the MAXwell Kit when measured by Qubit than NanoDrop. The RNA yield was higher for PFPE compared to FFPE specimens as indicated by lower Beta-actin cycle threshold (CT) values. NGS quality metrics for both GeneReader and QIAseq indicated superior performance in PFPE than FFPE specimens, but significance was dependent on metric and specimen type. More variants were found in PFPE than FFPE specimens using GeneReader but this was not further broken down by specimen type. More variants were found in FFPE than PFPE tumor blocks and Temno biopsies using QIAseq and fewer variants were found in FFPE than PFPE FNA specimens. However, overall concordance in genomic findings using QIAseq was good with the same actionable mutations identified in matched PFPE and FFPE specimens. Similarly, mutation identification using the Therascreen EGFR PCR assay was concordant in matched FFPE and PFPE resection and biopsy specimens. When specimen types were compared, the majority of variants were identified in tumor block specimens, regardless of fixation method, but PFPE FNA was superior to FFPE FNA.

Studies (1)

Purpose

This study compared H&E and IHC staining, DNA and RNA yield, and NGS results in matched FFPE and PFPE lung parenchyma and tumor resection specimens, Temno biopsy specimens, and FNA specimens. Two different DNA extraction and NGS workflows were used. Paired lung tumor and background tissue from lung cancer resections from eight patients were fixed in formalin or PAXgene for 24-72 h. Additionally, twenty-three 16G Temno needle core tumor biopsies were fixed in formalin or PAXgene for 24-72 h. Twenty-three FNA biopsies were obtained from unfixed specimens using a wide-bore needle were placed in the methanol-based ThinPrep Cytolyt fixative or PAXgene-FNA fixative for 1 h or more. After clotting with thrombin, ThinPrep FNA specimens were fixed in formalin for 24-72 h. After fixation, PAXgene specimens were placed in PAXgene stabilizer solution and stored at -20°C (storage of formalin-fixed specimens was not described) All specimens were paraffin embedded at 60°C and blocks were stored at -20°C. H&E staining was performed on fresh 4 µm sections and nuclear, cytoplasmic, and cell membrane staining was scored 0-4 by two blinded observers. For IHC staining, sections were subjected to antigen retrieval with high pH antigen retrieval solution (PT module, DakoCytomation) and then stained using antibodies to Ki67, MNF116, p63, cytokeratin (CK) 7, CK5/6, and thyroid transcription factor-1. IHC staining was scored 1 or 2 (unacceptable), 3 (borderline), or 4-5 (acceptable). DNA and RNA were extracted and analyzed using two different workflows. For the first workflow, DNA was extracted from eight paired FFPE and PFPE resection specimens and twenty-three Temno Biopsies using the QIAamp DNA FFPE Kit and the PAXgene Tissue DNA Kit, respectively, quantified by spectrophotometer and the QUBIT dsDNA Broad Range Fluorimetry, and the integrity evaluated by Tapestation 2200. Genotyping was conducted on matched PAXgene and FFPE resection and Temno specimens from seven patients using the Therascreen EGFR RGQ PCR Kit and by sequencing using the QIAGEN GeneReader NGS workflow with the QIAact Lung DNA Panel. RNA was extracted from the same eight paired FFPE and PFPE resection and biopsy specimens using the RNeasy FFPE Kit and the PAXgene Tissue RNA Kit; respectively and evaluated by real-time PCR amplification of Beta-actin. For the second workflow, DNA/RNA was extracted from FFPE specimens using the Maxwell 16 DNA/RNA Purification Kit and from PFPE specimens using the PAXgene Tissue DNA/RNA Kit and yield and purity were determined using a Qubit 2.0 Fluorometer and a Nanodrop 2000 Spectrophotometer. DNA fragmentation was evaluated using the Fragment Analyzer Automated system for High Sensitivity Genomic DNA. Libraries were prepared using the QIAseq Targeted DNA Panel assay and sequenced on a MiSeq or Illimina NextSeq 500.

Summary of Findings

Generally, H&E and IHC staining were comparable between FFPE and PFPE specimens, regardless of specimen acquisition method, but FFPE Temno biopsies and FNA specimens had significantly lower nuclear staining scores than matched PFPE specimens. H&E scores were modestly correlated between matched FFPE and PFPE Temno biopsies (r2=0.3455, P=0.0269) and matched PFPE and FNA biopsies (r2=0.8919, P<0.0001). The DNA yield was higher from PFPE lung parenchyma, tumor blocks, and Temno biopsies extracted using the PAXgene Tissue Kit than from matched FFPE specimens extracted using QIAamp (P=0.0011, P=0.008, and P=0.0053; respectively) or the MAXwell 16 DNA Kit (P=0.042, P=0.0015, and P=0.0002; respectively), but the differences in yield were smaller for the MAXwell Kit when measured by Qubit than NanoDrop. The RNA yield was higher for PFPE compared to FFPE specimens as indicated by lower Beta-actin CT values for lung parenchyma (P=0.0002), tumor blocks (P<0.0001), and Temno biopsies (P=0.0002). Using the Therascreen EGFR PCR assay, one patient was found to have a pG719X EGFR mutation in PFPE and FFPE tumor blocks and Temno biopsies and another to have both a p.L858R and a p.S768I EGFR mutation in PFPE and FFPE tumor blocks and Temno biopsies, indicating that either fixative is suitable for analysis by Therascreen. GeneReader NGS found a comparable percentage of specimens with a read quality score >25 and percentage of reads in targeted region between matched PFPE and FFPE specimens of parenchymal, tumor block, and Tenmo biopsies, but PFPE specimens had a higher percentage of on-target reads in FNA specimens (P=0.0354), a higher percentage coverage >100X and >60X in tumor blocks (P=0.0102 and P=0.002, respectively), a higher percentage unique molecular identifier (UMI) reads in tumor blocks (P=0.0202) and Tenmo biopsies (P=0.0045), and UMI read mean quality score in Tenmo biopsies (P=0.0006). Importantly, the number of variants called by GeneReader was higher in PFPE than FFPE specimens (P=0.0084) and, when broken down by variant type, this was found to be due to a significantly higher number of variants of uncertain significance in PAXgene specimens (P=0.0093). In the second workflow, QIAseq libraries had larger mean fragment size and higher depth of coverage for PFPE versus formalin-fixed lung parenchyma specimens (P=0.00357 and P=0.0365, respectively) and tumor FNA specimens (P=0.0001 and P=0.0102, respectively) and higher depth of coverage for PFPE than FFPE Temno biopsies (P=0.0017). While concordance in genomic findings using QIAseq was generally good, there were more variants found in tumor blocks and Temno biopsies when formalin-fixed than when PAXgene-fixed and more variants found in PFPE FNA specimens than in FFPE FNA specimens. Importantly, PFPE and FFPE specimens performed equally well in identifying actionable EGFR mutations with five identified in three patients, regardless of fixation method. When specimen types were compared, the majority of variants were identified in tumor block specimens, regardless of fixation method, but PFPE FNA was superior to FFPE FNA.

  • Biospecimens

    TissueLung
  • Preservation Types

    FormalinPAXgene
  • Diagnoses

    NeoplasticCarcinoma
    NeoplasticNormal Adjacent
  • Platforms

    MorphologyH-and-E microscopy
    ProteinImmunohistochemistry
    DNASpectrophotometry
    DNAFluorometry
    DNANext generation sequencing
  • Pre-analytical Factors

    ClassificationPre-Analytical FactorValue(s)
    Biospecimen AcquisitionMethod of tissue acquisition
    Biopsy
    Fine needle aspiration
    Surgical resection
    Biospecimen PreservationType of fixation/preservation
    Formalin (buffered)
    PAXgene
    Next generation sequencing SpecificTechnology platform
    GeneReader
    QIAseq