3.2 Risk of Bias Assessment
We used RoB2.0 to assess the risk of bias in these articles. Except 2 articles (Kang X, et al; Lu S, et al), other studies have shown that with randomization, so the risk of deviation (SG) from sequence generation was assessed as ”low”. The experimental groups in the two studies were not randomized, so the risk of SG bias in this group was assessed as ”high”. Three studies (Zhang L, et al; Hou J, et al; Li Y, et al) described allocation concealment (AC), participant blindness (BPt), and these were decided “low risk”. The other 54 studies did not describe the treatment course of AC and were therefore considered to be at “unclear risk”. In cancer trials, it is difficult to blind participants. For selective outcome reporting (SOR), the study was assessed as ”low-risk” only if the objectives and outcome measures described in the methods section are in the results section. Our results show TRR symmetry in funnel plots of the 57 studies, indicating a lower risk of publication bias.