Dynamic Network Analysis of Symptom Profiles and Their Expected Impact on Patients with Colorectal Cancer Receiving Postoperative Adjuvant Chemotherapy
摘要
Chemotherapy causes patients to experience a series of side effects. Dynamic network analysis plays a significant role in elucidating the synergistic effects among symptoms. This study attempts to analyze the dynamic correlation among symptoms during the early, middle and late stages of chemotherapy, and to clarify their expected influence on colorectal cancer (CRC) patients.
MethodsA longitudinal study was conducted from December 2022 to December 2024. General information was collected before chemotherapy, and symptoms were assessed using the Chinese version of the Memorial Symptom Assessment Scale (MSAS-C) at the early stage (T1), the middle stage (T2), and the end of chemotherapy (T3). Cross-Lagged Panel Network (CLPN) analysis was applied to construct dynamic symptom networks. Out-expected influence (out-EI) and in-expected influence (in-EI) were used to identify highly predictive and predicted symptoms in dynamic networks. Linear regression was used to verify the effect of symptoms with high out-EI on symptom burden at the late stage.
ResultsA total of 380 patients were included in the analysis. From the early to middle stage of chemotherapy (T1 → T2), change in food tastes was the core predictor (out-EI was 0.846). From the middle to late stage of chemotherapy (T2 → T3), feeling drowsy was the core predictor (out-EI was 0.713). Numbness or tingling in hands and feet exhibited the highest in-EI (in-EI was 0.436 in T1 → T2, 0.522 in T2 → T3). After adjusting for confounding variables, symptoms with the highest out-EI had a significant influence on symptom burden at the late stage (P < 0.001).
ConclusionMedical staff are encouraged to focus on taste change in the early stage of chemotherapy, and strengthen the evaluation of feeling drowsy in the middle stage. For those who feel numbness or tingling in hands and feet, it is necessary to analyze other symptoms that may contribute to their occurrence.