A Bonferroni correction for multiple comparisons was performed. To maintain an overall type I error rate of 0.05, each test selleck chem Pacritinib of association with complete pathologic response was considered significant if P<0.005. All variables significant in univariate analysis were entered into a multiple logistic regression model. Statistical interactions between significant variables were tested. All analyses were carried out with SAS V9 (The SAS Institute, Cary, NC, USA). RESULTS Clinicopathological features Patient characteristics are summarised in Table 1. Thirty-three (31.7%) patients had a complete pathologic tumour response to preoperative HDREB. Seventy-one (68.3%) were found to have residual carcinoma, including 35 patients (34.3%) who were considered as partial responders due to the presence of microfoci of residual carcinoma.

Ninety-six (94.1%) patients were preoperatively staged as cT3. pT stage was available for 80 patients, of which 31 (38.8%) were pT0, 12 (15.0%) were downstaged to pT1, 17 (21.3%) were pT2 and 20 cases (25.0%) were pT3. Table 1 Clinicopathological characteristics of rectal cancer patients treated with high-dose-rate brachytherapy Selection of cut-off scores based on ROC curve analysis Cut-off scores were determined to be 50% for p53, 20% for VEGF, Bcl-2 and EGFR and 10% for APAF-1. Tumours with scores above the obtained cut-off values were considered positive for the expression of the protein. The corresponding AUCs (95% CI) are listed in Table 2. AUCs were the largest for EGFR (0.66 (0.54�C0.78)) and VEGF (0.64 (0.51�C0.

77) indicating that the discriminatory power for complete response was the greatest for these two markers. Table 2 ROC curve-derived cut-off scores, area under the curve (AUC) and association of protein expression with complete pathologic response Univariate analysis The association of protein expression with complete pathologic response (Table 2) demonstrated that negative VEGF expression (P-value=0.004, OR (95% CI)=0.23 (0.09�C0.63)) and EGFR positivity (P-value=0.003, OR (95% CI)=5.78 (1.85�C18.07)) were significantly associated with complete tumour response after correction for multiple comparisons while p53, APAF-1 and Bcl-2 demonstrated no predictive ability for the outcome. Loss of VEGF expression was associated with more than a 4.3 times greater chance of complete tumour response compared with VEGF-positive tumours, while EGFR positivity resulted in a 5.

78 times increased odds of complete tumour regression. Representative immunostains of VEGF and EGFR are illustrated in Figure 1. Figure 1 Representative immunostains of VEGF (A) and EGFR (B) from pretreatment rectal tumour biopsies. Multivariable analysis VEGF and Cilengitide EGFR were entered into multivariable analysis. Eighty-eight tumours could be evaluated, of which 27 (31%) had a complete pathologic response. Loss of VEGF (P-value=0.009; OR (95% CI)=0.24 (0.08�C0.69)) and positive EGFR (P-value=0.