Page 171 - Sentinel lymph node biopsy in oral cavity cancer - Inne J. den Toom

measurements in tongue tumours (r=0.88). On average, they showed a minimal and clinically acceptable overestimation of 0.5 mm on ultrasonography. Other imaging modalities, as MRI or CT-scans appeared to measure less accurate in particularly thin tumours. These findings were confirmed by a more recent meta-analysis, again showing higher correlation between histopathological thickness and ultrasonography compared to MRI (r=0.96 vs. 0.88).47 Given these better results, its easier applicability and lower costs intraoral ultrasonography could be the approach of choice determining clinical DOI. In the last part of this thesis we would like to focus on new possibilities using SLNB and also on ways to actually improve the procedure itself. All SLNB literature showing high sensitivity and negative predictive value rates are based on patients with primary early stage oral cancer and a previously untreated neck (e.g. no neck dissection and/ or (chemo)radiotherapy). Previously treatment of the neck may disrupt lymphatics and most likely therefore alters lymphatic drainage patterns. Despite the relatively common local recurrences and second primary tumours in oral squamous cell carcinoma, only one study reported about the accuracy of SLNB in 22 patients with a previously treated neck.48 In that study, patients with ipsi- or bilateral neck treatment had a SLN detection rate of 83% and unexpected lymphatic drainage patterns were observed in 67% of the patients. Due to the increased experience using SLNB in general, during the last years patients with a previously treated neck are staged by SLNB as well. In a collaboration of three Dutch head and neck centers we retrospectively analyzed the accuracy of SLNB in 53 patients with a previously treated neck and evaluated the lymphatic drainage patterns in 43 patients in this cohort (chapter 6). We presented an imaging detection rate of 87% and a surgical detection rate of 93%. This is relatively low compared to previously untreated patients with identification rates of 97-98%.7,10,49 Interestingly, all patients without identified SLNs had radiotherapy in history, sometimes combined with surgery. In contrary, patients with a prior SLNB had no lower identification rate, suggesting that SLNB ensures less damage to lymphatic vessels compared to radiotherapy. In this study, three patients showed a positive SLNB and only one patient showed regional recurrence during follow-up (false negative SLNB), resulting in a 75% sensitivity and 98% negative predictive value. The low number of regional metastases could be potentially explained by a close follow-up scheme after treatment of their first primary tumour, also reflected by a high number of pT1 tumours in this study. Due to only 4 patients with regional metastases, it might be prematurely to conclude that 9 SLNB is reliable in pretreated patients. However, a negative predictive value of 98% suggest that SLNB is at least promising and further research is necessary to determine its reliability. 169