Milea Timbergen

174 The influence of tumour site on initiation of active surveillance Frequent reported tumour sites (available in sixteen studies) were the extremities/girdles (n = 273 patients, median percentage of incidence in studies 31% [IQR: 3-68%]), the abdominal wall (n = 253 patients, median percentage of incidence in studies 9% [IQR: 0-37%]), and the trunk (n = 153 patients, median percentage of incidence in studies 17% [IQR: 0-37%]). Intra-abdominal (n = 60) and head/neck (n = 15) tumours were less common, with a median percentage of incidence in studies of 0% (0-8%) and 0% (IQR: 0-4%) respectively. From a total of 1480 patients receiving AS, the tumour sites were not specified in 726 (49%) of patients (Table 2). Cassidy et al. described that patients with abdominal wall tumours were often managed with AS (61%), whereas those with chest wall and intra-abdominal tumours more often received active treatment (80% and 60%, respectively) 31 . Fiore et al. also described that patients who received AS commonly had abdominal wall tumours (p < 0.0001) compared with patients who received other treatments 14 , whilst Park et al. found no difference in tumour sites between groups managed with AS or surgery 30 . The influence of the tumour site on disease stabilisation, progression or a change of the treatment strategy No differences in risk of progression during AS were found between abdominal wall tumours and other sites (p = 0.53) by Turner et al. 38 nor on a chance of spontaneous stabilisation among axial sites or extremity tumours (p = 0.148) by Kim et al. 37 (Table 5). The 5-year PFS of primary cases managed with AS of trunk/thoracic wall tumours and abdominal wall tumours was similar (53.9% [SE: 16.2%] versus 52.5%, [SE: 14.3%]) in the study from Fiore et al. 14 . Van Houdt et al. concluded that upper-extremity and chest wall DTF tumours have the highest percentage of progression (39% and 47%, respectively), although this difference was not significant compared with other locations 36 . Cassidy et al. described that tumours located paraspinal or flank were more commonly associated with a change in treatment than abdominal wall tumours (p = 0.01), but no differences were found comparing extremity, intra-abdominal or abdominal wall tumours 31 . Van Houdt et al concluded that there was no difference in initiation of active treatment between upper extremity and chest wall DTF (p = 0.36) 36 . This is in line with the findings of Colombo et al. who did not identify the tumour site as a predicting factor for progression and/or change in the treatment strategy among tumour sites (p = 0.926) 9 . No single conclusion could be reached regarding tumour site and the success or failure of the AS approach because of the heterogeneity of the cohorts of included studies. 6