Page 56 - Sentinel lymph node biopsy in oral cavity cancer - Inne J. den Toom

Chapter 3 The overall sensitivity of detecting occult metastasis was comparable between the END and SLNB patients (84% vs. 81%, p = 0.612). Both groups had a similar negative predictive value (NPV) (93%, p = 1.000) (Table 2). Because of dissimilarity in pT staging, we separately analysed the accuracy for pT1 and pT2. In the SLNB cohort a trend towards a higher, though not significantly different, sensitivity was observed for pT2 tumours compared to pT1 tumours (88% vs. 76%, p = 0.075). In the END cohort, pT2 tumours also showed a higher sensitivity in comparison to pT1 tumours (90% vs. 70%, p = 0.010). NPVs did not differ significantly regarding pT stage within the groups. No significant differences were found for sensitivity and NPV between the SLNB and END when corrected for pT stage (Table 2). Floor of mouth tumours In total 131 (27%) of the SLNB and 133 (34%) of the END patients had a tumour located in the FOM. SLNB had a lower sensitivity (63% vs. 92%, p = 0.006) and NPV (90% vs. 97%, p = 0.057) compared to END. The SLNB had a higher (but not significantly) sensitivity (86% vs. 80%, p = 0.315) and NPV (95% vs. 92%, p = 0.250) compared to END for other (non-FOM) anatomical locations. When comparing FOM tumours with other non-FOM locations within the SLNB group, there was a lower sensitivity (63% vs. 86%, p = 0.008) and NPV (90% vs. 95%, p = 0.113). In contrast, within the END group at most a trend towards a higher sensitivity (92% vs. 80%, p = 0.114) and a higher NPV (97% vs. 92%, p = 0.130) was observed for FOM tumours compared to other anatomical locations. Of the 11 FOM patients with a false negative SLNB, 64% (7/11) had a regional recurrence in level I. In 3 FOM END patients, 1 patient (33%) had a regional recurrence in level I, the remaining patients had a regional recurrence in level II or higher. Five years disease specific survival The DSS was significantly longer for SLNB pT1 patients (96%) compared to END pT1 (90%, p = 0.008), SLNB pT2 (90%, p = 0.001) and END pT2 (86%, p < 0.001) patients. No significant differences in DSS were seen between the other groups. After the Bonferroni correction, the SLNB pT1 had still a significant longer DSS compared to the other groups: END pT1 (p = 0.048), SLNB pT2 (p = 0.006) and END pT2 (p < 0.001) (Figure 2A). We furthermore analysed the differences between DSS of the END and SLNB groups divided by anatomical location (FOM vs. other locations, Figure 2B). SLNB staged patients with a FOM tumour had a longer DSS compared to END FOM patients (98% vs. 87%, p = 0.021). The other (non-FOM) SLNB patients had longer DSS compared 54