PURPOSE: Τo estimate the discrepancy rate between colposcopy, punch biopsy histology and surgical specimen histology as well as the positive (PPV) and negative predictive value (NPV) of colposcopic diagnosis for high-grade squamous intraepithelial lesions (HGSIL).METHODS: A prospective study was conducted during the period of 2012-2016. Αll cases in which surgical treatment had been applied and histopathological diagnosis of those surgical specimens was available were included. Cases in which ablation was performed and cases with incomplete data or conservative approach were excluded. Primary outcome was the agreement rate between histologic diagnosis of surgical specimen, histologic diagnosis of punch biopsy and colposcopic diagnosis according to REID Colposcopic Index. PPV and NPV of colposcopy and biopsy to diagnose HGSIL were also assessed.RESULTS: Τhere were 120 cases meeting our inclusion criteria, while biopsy was obtained in 104 cases. Mean age of women was 32.7 ± 9.0. Colposcopic diagnosis was CIN2 in 65 cases, CIN3 in 11 cases, CIN1 or less in 44 cases. Τhe level of agreement was fair between colposcopy-surgical specimen histology (κ value 0.443), fair between colposcopy-punch biopsy (κ value 0.34) and moderate between punch biopsy-cone specimen histology (κ value 0.443). PPV of colposcopy to detect HGSIL was 72.3%, while NPV was only 47.7%.CONCLUSION: Punch biopsy and surgical specimen histology present the highest agreement between the different diagnostic procedures. Colposcopy presented satisfying PPV for HGSIL cases, but its NPV was poor. In contrary, punch biopsy was characterized by both satisfying PPV and NPV for HGSIL cases.