Assess agreement between proxy respondents (caregivers) and children/adolescents related to the impact of cancer on children's/adolescents' health-related quality of life, with respect to anxiety and worry issues. A cross-sectional study was conducted among 83 Brazilian children/adolescents, of both genders, diagnosed with cancer, aged 5-18 years and their proxy respondents. Anxiety and worry were assessed through items of the instrument Pediatric Quality of Life Inventory™ Cancer Module Scale. Participants were recruited from the pediatric hematology/oncology centers at two public hospitals. All individuals were receiving medical care. Descriptive statistics were performed as well as a weighted kappa coefficient, Spearman's correlation coefficient, Wilcoxon signed-rank test and Bland-Altman plots. The magnitude of the difference between the mean scores obtained from children/adolescents and that of their proxy respondents was evaluated through effect size. The proxy respondents underestimated the feelings of worry among children (8-12 years) (p < 0.001; effect size 0.71) and overestimated adolescents' (13-18 years) treatment anxiety (p < 0.05; effect size 0.57). The comparison between the three age groups (5-7, 8-12, 13-18 years) showed a tendency for children/adolescents to report increasing feelings of worry as they got older. In the 'treatment anxiety' subscale, there was a tendency for proxy respondents to present higher mean scores, revealing that proxy respondents believed the children's/adolescents' treatment anxiety decreased as they aged. Discrepancies between the reports of children/adolescents and their proxy respondents were observed. Children's/adolescents' reports should not be ignored nor replaced by proxy reports; both reports should be analyzed together.