Anal squamous cell carcinomas (ASCC) represent the most common histologic entity among anal cancers. Oncogenic human papillomavirus (HPV) types play an etiological role in a large proportion of ASCC as indicated by the detection of HPV DNA in up to 90% of cases frequently accompanied by overexpression of the cell cycle regulator protein p16INK4A on immunohistochemistry (IHC). By analogy to head and neck squamous cell carcinomas, it has been suggested that HPV DNA and p16INK4A status might be of prognostic relevance in ASCC patients. However, the reported survival rates for ASCC patients, stratified by these two markers, differ among the published studies.
We aimed to determine the prognostic relevance of oncogenic HPV DNA and p16INK4A status among all published literature in a systematic review and meta-analysis.
A broad search string was designed to identify all published studies analyzing p16INK4A expression by IHC and providing survival data in patients diagnosed with ASCC. Overall survival (OS) was analyzed performing Cox Regression including p16INK4A IHC, HPV DNA status and clinical data as covariates. Authors were contacted to obtain lacking information or data.
16 studies were found to be eligible for inclusion in the final analysis. From 8 of them we obtained the individual patient records comprising 666 ASSC cases. 84.3% of 555 ASCC tested positive for oncogenic HPV DNA. 81.8% of 658 ASCC demonstrated overexpression of p16INK4A on IHC. Patients with ASCC demonstrating p16INK4A overexpression had a significantly longer median OS than patients without p16INK4A overexpression (36 vs. 28 months (m), respectively), hazard ratio (HR)=0.42 (95% confidence interval (CI), 0.30-0.61) in a pooled analysis. Patients with HPV DNA-positive ASCC also demonstrated a significantly better median OS compared to HPV DNA-negative ASCC patients (39 vs. 26 m, respectively), HR=0.39 (95% CI, 0.27-0.57). Hazard ratios for gender (female vs. male), T-stage (T3/4 vs. T1/2), N-stage (N+ vs. N0), M-stage (M+ vs. M0), HIV status (positive vs. negative) and age were 0.42 (95% CI, 0.30-0.59), HR=2.88 (95% CI, 2.06-4.04), HR=1.92 (95% CI, 1.37-2.70), HR=3.29 (95% CI, 1.60-6.02), HR=1.30 (95% CI, 0.66-2.33), HR=1.02 (95% CI, 1.01-1.03), respectively.
p16INK4A overexpression and oncogenic HPV DNA are detected in a large proportion of ASCC patients and predict better survival compared to p16INK4A- or oncogenic HPV DNA-negative ASCC patients. The obtained data will be further analyzed in multivariate analyses. In the future, we will use digitized Kaplan-Meiers to meta-analyze the 16 studies and use available individual patient data from 8 studies to validate statistical methods.