Milea Timbergen

163 The retrieved articles were assessed for potential inclusion by the first and second author based on the review of title and abstract. Next, full-text articles were evaluated in accordance with the predetermined inclusion criteria and exclusion criteria for this systematic literature review (listed in Table 1). Data Extraction Data was collected by two researchers (MJMT, AWS) using a pre-defined Excel sheet statin the year of publication, the first author, the journal, the publication title, whether the publication fulfilled the inclusion criteria, the inclusion period, the type of study, the total number of participants, the number of participants receiving AS, the number of patients with familial adenomatous polyposis (FAP)/Gardner syndrome, the number of primary tumours, and the number of recurrent tumours. Of the AS group, the following variables were extracted: the reported mean/median follow-up (range, interquartile range [IQR], 95% confidence interval [CI]), the reported median/mean age (with range or IQR), the sex distribution, the tumour sites, the number of patients with progressive disease (PD), stable disease (SD), partial response (PR), complete response (CR), the number of patients who shifted to active treatment, reasons for shifting to an active form of treatment, and whether RECIST were used for determination of these outcomes. For responses not evaluated by RECIST but by using similar terms, tumour response was categorised based on the RECIST categories; PD, SD, PR and CR. PD included the terms ‘increase’, ‘evolution’, ‘enlarged’, SD included the terms ‘stable’, ‘arrested’, and ‘non-progressive’, PR included the terms ‘decreased’, ‘regressed’, ‘disease free survival’, ‘responding disease’, and ‘spontaneous remission’, and CR included the terms ‘disappeared’, and ‘complete regression’. Not specified (NS) was used in case a variable was missing. Tumour sites were classified as: the extremity/girdle region (including upper extremity, lower extremity, shoulder, buttock, thigh and hip), intra-abdominal (including mesenteric), trunk (including paraspinal and thoracic wall), abdominal wall, head/neck region and other (including inguinal region and not further specified sites). When age and follow-up (in months of years) were reported for each individual patient, the median age and median follow-up with range were extracted and calculated from these data. A shift to ‘active treatment’ was defined as ‘ceasing active surveillance’. The following therapies were categorised as ‘active treatments’: systemic treatment (including hormonal treatment, chemotherapy, and tyrosine kinase inhibitors), surgery, radiotherapy, combination 6