Multidimensional symptom structure of anxiety, depression symptoms and quality of life in postoperative breast cancer survivors: a network analysis and bayesian directed acyclic graphs

Postoperative breast cancer (BC) survivors often experience high and severity of anxiety and depression, which have become an important public issue globally and impact their quality of life (QoL). However, limited evidence focuses on the symptom-level structure and core symptoms of anxiety, depression and QoL in long-term postoperative stage. Therefore, our study aims to reveal the specific symptom network relationship and core symptom of anxiety, depression and QoL, offering new insights into the mechanism underlying QoL in postoperative BC survivors. An institution-based retrospective follow-up study was conducted among 713 postoperative BC patients at the First Affiliated Hospital of Guangxi Medical University between December 2016 and December 2019. The follow-up data were collected from July 2022 to September 2022. Demographic questionnaires, the European Organization for Research and Treatment of Cancer (EORTC QLQ-C30) cancer-specific questionnaire and Quality of Life questionnaire-breast cancer module 23 (QLQ-BR23), the self-Rating Anxiety Scale (SAS), and the self-Rating Depression Scale (SDS). The complex network relationships among symptoms were analyzed using Extended Bayesian Information Criterion Graphical LASSO (EBIC-GLASSO) and Bayesian directed acyclic graphs (DAGs). Mean anxiety and depression scores were 51.00 ± 4.96 and 63.42 ± 3.46, respectively. Among QLQ-C30 domains, social functioning had the highest score (97.03 ± 8.53), while body image was the highest-scoring domain in the BR23 module (95.59 ± 12.18). Network analysis revealed that emotional functioning (QLQ3) was most strongly related to nausea/vomiting (QLQ7) (weight = -0.811), followed by cognitive functioning (QLQ4) with fatigue (QLQ6) (weight = -0.654) and social functioning (QLQ5) with financial difficulties (QLQ14) (weight = -0.544). In the breast cancer–specific domain, sexual functioning (BR2) was closely linked to sexual enjoyment (BR3) (weight = 0.644), whereas body image (BR1) was inversely associated with arm symptoms (BR7) (weight = -0.756). Depression (PSY2) was directly connected with global QoL (QLQ15) (weight = 0.770). Social functioning (QLQ5) had the highest strength centrality (2.489); systemic therapy side effects (BR5), insomnia (QLQ10), and depression (PSY2) showed the greatest expected influence, with the value of 1.928, 1.203, and 0.984, respectively. No significant differences in network structure or global strength were found by subgroup analysis (all P > 0.05). In the Bayesian DAG, anxiety (PSY1) showed conditional dependencies with physical functioning (QLQ1), social functioning (QLQ5), and global QoL (QLQ15). The graph also identified an oriented conditional dependency pattern involving fatigue (QLQ6), cognitive functioning (QLQ4), depression (PSY2), and anxiety (PSY1). This study started from symptoms and provided new insights into the relationship between anxiety and depression and the QoL of BC patients after surgery. Social function and treatment-related side effects occupied key positions in this network, while anxiety seemed to lie between fatigue, cognitive impairment, depression, and overall decline in QoL. These findings indicate that a symptom-oriented approach, targeting core symptoms and bridging symptoms, can help bring broader benefits to the survival process of these population.