Author of

• 11866

  +

  P2.18 - Poster Session 2 - Pathology (ID 176)

  • Event: WCLC 2013
  • Type: Poster Session
  • Track: Pathology
  • Presentations: 1
  • Moderators:
  • Coordinates: 10/29/2013, 09:30 - 16:30, Exhibit Hall, Ground Level

  • 11871

    +

    P2.18-005 - Morphological and genetic classification of lung cancer: variation in practice and implications for tailored treatment (ID 956)

    09:30 - 09:30  |  Author(s): A. Nicholson
    • Abstract

    Background
    The rational use of tailored therapy for lung cancer depends crucially on high quality pathology. Not only must subtyping of the tumour be achieved consistently and with accuracy, but as much material as possible from increasingly small specimens must be preserved for the genetic analysis upon which such treatment is increasingly predicated. Although there is a general presumption that pathologists have risen to this challenge, reliable data is sparse and the nature and degree of variation in practice and quality is unknown.

    Methods
    We collected and scrutinised anonymised information, including pathology reports, on all consecutive, newly-diagnosed patients with lung cancer referred to 19 lung cancer units across the UK for a period of 6 months commencing late 2011, a total of 1507 cases. Centres surveyed ranged from district general hospitals to specialist regional cardiothoracic units.

    Results
    Achievement of a positive tissue diagnosis of malignancy ranged from 53 to 88%, figures accompanied by ‘suspicious but non-diagnostic’ rates of 10 and 2% respectively. Despite apparent adherence to the diagnostic criteria and terminology of the WHO classification of tumours of the lung, variation in proportions of tumour subtypes was wide, the prevalence of squamous carcinoma, for example, varying from less than 10 to greater than 50%. The proportion of tumours unclassified beyond ‘non-small cell lung cancer not otherwise specified’ varied from 3 to 20% despite the almost universal use of immunochemistry (most often TTF-1, class 7 cytokeratins and p63) to aid in this differential diagnosis. Testing for EGFR gene mutations was directly instigated by the pathologist at diagnosis in only 4 of the 19 centres and the proportion of tumours tested ranged from 12 to 92%.

    Conclusion
    Variations in practice amongst pathologists and arguably in the quality of pathology ranged widely across the centres surveyed, raising important questions about variable expertise of pathologists, adherence to guidelines, applicability and rigour of external quality assessment and, ultimately, the reliability of the pathology that crucially underpins the management of lung cancer.

• 12441

  +

  P3.06 - Poster Session 3 - Prognostic and Predictive Biomarkers (ID 178)

  • Event: WCLC 2013
  • Type: Poster Session
  • Track: Biology
  • Presentations: 1
  • Moderators:
  • Coordinates: 10/30/2013, 09:30 - 16:30, Exhibit Hall, Ground Level

  • 12467

    +

    P3.06-027 - A comparison of FISH and immunohistochemistry in the detection of ALK rearrangement in lung adenocarcinoma (ID 2392)

    09:30 - 09:30  |  Author(s): A. Nicholson
    • Abstract

    Background
    Personalised treatments for lung cancer are being increasingly employed. Around 4% of pulmonary adenocarcinomas have rearrangements of the ALK gene resulting in oncogenic fusion proteins. Such cases are sensitive to the small molecule tyrosine kinase inhibitor Crizotinib, and detection of the ALK rearrangement is crucial to the treatment of these patients. Currently the only approved test for the detection of ALK-positive NSCLC is FISH. Various immunohistochemical assays have been proposed as alternatives or screening tools, which are cheaper, quicker and easier to perform and interpret than FISH. Validation of these immunohistochemical tests is therefore important.

    Methods
    15 FISH positive cases were obtained from local archives at the Royal Marsden and Royal Brompton Hospitals, with accompanying data on crizotinib therapy and clinical response. A further 14 FISH negative and FISH uncertain cases were also retrieved. 3 antibodies were optimised for immunohistochemical detection of ALK: D5F3, marketed by Ventana and using their proprietary automated system, ALK1 (Dako), and 5A4 (Abcam). All three antibodies were applied to sections from the archival cases. Antibodies were semiquantitatively scored on their intensity (0-3) and proportion of malignant cells stained (0-100%). Cutoffs for positivity/negativity were set by ROC analysis to optimise correct classification.

    Results
    Positivity according to the Vysis FISH assay (i.e. appropriately abnormal FISH signal in at least 15% of cells) was taken as the gold standard. The Ventana assay (D5F3) was markedly more intense but showed higher background than the other two antibodies. ROC analysis of staining intensity showed that the Ventana and Dako antibodies should be considered positive when staining is of 'moderate' intensity or above. The Abcam assay was discriminatory when any positivity was seen. Scoring of proportion did not improve specificity or sensitivity with any of the antibody tests. Subsequent analyses were of intensity scoring. Taking all cases together, all 3 antibodies were highly specific (100%) and sensitive (Ventana 86%, Abcam 79% and Dako 71%). In excision specimens, all 3 assays showed sensitivity of 83%. In cytology and small biopsy specimens, Ventana performed the best (specificity of 88% compared to 75% for Dako and 63% for Abcam) although the numbers are small. No ‘false positive’ (-ve FISH, +ve IHC) cases were seen; the only disparity between FISH and IHC were 'false negative' cases. Two cases were FISH-positive but universally IHC-negative (with all 3 assays). Of these, one was treated with crizotinib, and failed to respond.

    Conclusion
    IHC is a highly specific (100%) and sensitive (71%-86%) test for ALK rearrangement in archival FFPE tumour tissue. In this study of 15 ALK-rearranged tumours, the Ventana D5F3 assay performed the best, despite having relatively high background staining. Occasional cases are FISH-positive but IHC-negative. One such case was not responsive to Crizotinib, further supporting IHC as a test predictive of drug response. IHC has the potential to replace FISH for the detection of candidates for Crizotinib therapy. However, further work is required to direct the treatment of tumours with discordant FISH and immunohistochemical tests.