HPV Detection in Head and Neck Squamous Carcinoma: A Comparison of Methods
M Brandwein-Gensler, M Prystowsky, M McLemore, G Nuovo, RV Smith, N Schlecht. Montefiore/Einstein, Bronx, NY; Ohio State University College of Medicine, Columbus, OH; Einstein, Bronx, NY
Background: Substantial evidence confirms a causal relationship between HPV and some HNSCC, most commonly oropharyngeal SCC and HPV16. HPV+HNSCC may have a better response to chemo-radiation and longer survival than HPV-HNSCC. Future treatments will be tailored based on HPV status and pathologists will be requested to make this determination. IHC for p16 and ISH with HPV DNA probes are touted as surrogates to PCR analyses. IHC is readily available, easily performed, and easily interpreted, although few reports address optimal cut-off values. In contrast, HPV ISH is not routinely available and interpretation is more difficult. It is important and timely to compare the sensitivity and specificity of commercial ISH/IHC tests to the gold-standard of HPV RNA PCR on frozen samples.
Design: Frozen tumors were tested by RT-PCR for HPV16 E6/E7 RNA. Tissue microarrays (TMA) were constructed from corresponding FFPE tumors and tested for p16 by IHC (BD Pharmingen antibody). Nuclear and cytoplasmic staining was quantified for intensity and % distribution; these values were incrementally correlated with HPV16 E6/E7 RNA status. The optimal combination of cut-off values was selected based on maximum area under the ROC curve. HPV ISH was assessed on these TMAs with two different probes: the INFORMR HPV-III Fam16(B) probe (Ventana Medical Systems, Tucson, AZ) and HPV 16/18 probe (DakoCytomation, Carpinteria, CA).
Results: Sixty-eight tumors were studied; HPV16 E6/E7 was detected in 39% by RT-PCR. The optimum cut-off values for p16+, as determined by ROC values, was nuclear plus cytoplasmic staining ≥+2, ≥75%. The sensitivity and specificity for p16 by IHC is 50% (95% CI 30%, 70%) and 93% (95% CI 81,96%), respectively, compared to RT-PCR for HPV16 E6/E7. We are continuing to study new cases by RT-PCR and IHC. Comparison of the ISH data (Ventana and Dako probes, read blinded by multiple observers) with the RT-PCR data is on-going.
Conclusions: The performance of p16 as assessed by IHC on TMAs, using the above cut-offs, indicates that this test has good specificity, but poor sensitivity. HNSCC with no p16 staining, or staining below these cut-off values, will still require additional HPV testing.
Category: Head & Neck
Tuesday, March 23, 2010 8:00 AM
Platform Session: Section G, Tuesday Morning