Skip to main content

Frozen section evaluation of breast carcinoma sentinel lymph nodes: a retrospective review of 1,940 cases.

Poling JS, Tsangaris TN, Argani P, Cimino-Mathews A

Publication: Breast Cancer Res Treat (2014) • Vol 148 • Page 355-61PMID: 25318925View on PubMed

Purpose

This paper retrospectively compared the final breast cancer diagnosis based on formalin-fixed paraffin-embedded (FFPE) sections and follow-up visits with that based on frozen sections of sentinel lymph node biopsies (SLNBs) from the same patient.

Conclusion

Frozen sections had a sensitivity of 76.3%, a specificity of 99.9%, a positive predictive value (PPV) of 99.7% and a negative predictive value (NPV) of 94% for the final diagnosis. False negatives in the frozen section occurred in 4.9% of cases, mostly due to tissue or block sampling issues, but 20% were due to a need for IHC detection to diagnose carcinoma and 6% were due to misinterpretation of the frozen section. The majority of false-negatives occurred in cases with metastases that were less than 2.0 mm (73%).  In one case, isolated tumor cells were seen in the frozen section but not the FFPE section and after IHC analysis, the authors believe this to be a false positive. Interestingly, additional axillary dissection (AD) was performed for a higher percentage of false negative cases obtained between 2003 and 2008 than between 2009 and 2012 (53% versus 27%, p<0.05). 

Studies (1)

Purpose

This study investigated the diagnostic accuracy of frozen sections of SLNB in the diagnosis of breast tumors. Data from 1,940 frozen sentinel lymph node biopsies with matched FFPE sections were evaluated. Cases with a discrepancy in the data from the frozen section to the final diagnosis based on FFPE sections and clinical follow-up were broken down by reason and, when necessary, determined by examination of frozen and FFPE slides.

Summary of Findings

Frozen sections had a sensitivity of 76.3%, a specificity of 99.9%, a positive predictive value (PPV) of 99.7% and a negative predictive value (NPV) of 94% for the final diagnosis using FFPE sections. Of the 1,940 cases, 95 specimens (4.9%) were false negatives for carcinoma on the frozen section and were diagnosed using the FFPE specimen as being metastatic ductal carcinoma (64 cases), lobular carcinoma (20 cases), and mammary carcinomas with features of both ductal and lobular carcinoma (11 cases). The majority of the 95 false negatives were found to be due to tissue-(37 cases, 39%) or block-sampling issues (33 cases, 35%), but 20% (19 cases) were due to a need for IHC detection to diagnose carcinoma and 6% (6 cases) were due to misinterpretation of the frozen section. Importantly, the majority of false-negatives occurred in cases with metastases that were less than 0.2 mm or contained only isolated tumor cells (27%) or in metastasis of 0.2-2.0 mm (46%).  In one case, isolated tumor cells were seen in the frozen section but not by FFPE and after IHC analysis, the authors believe this to be a false positive. Additional axillary dissection (AD) was performed for a higher percentage of false negative cases obtained between 2003 and 2008 than between 2009 and 2012 (53% versus 27%, p<0.05). 

  • Biospecimens

    TissueBreast
    TissueLymph Node
  • Preservation Types

    FormalinFrozen
  • Diagnoses

    NeoplasticCarcinoma
  • Platforms

    ProteinImmunohistochemistry
    MorphologyLight microscopy
    MorphologyH-and-E microscopy
  • Pre-analytical Factors

    ClassificationPre-Analytical FactorValue(s)
    PreaquisitionDiagnosis/ patient condition
    Lobular carcinoma
    Mammary carcinoma with features of both ductal and lobular carcinoma
    Metastatic ductal carcinoma
    PreaquisitionPrognostic factor
    Metastases 0.2-2.0 mm
    Metastases >2.0 mm
    Metastasis <0.2 mm or isolated tumor cells
    Biospecimen PreservationType of fixation/preservation
    Formalin (buffered)
    Frozen