Symptom clusters and network analysis in lung cancer patients receiving taxane-based chemotherapy: a comprehensive...

This study aimed to clarify the topological structure, core symptoms, and inter-symptom association patterns of the symptom network in lung cancer patients receiving taxane-based chemotherapy and to provide a basis for formulating precise symptom management strategies.

A convenience sampling method was used to enroll 315 hospitalized lung cancer patients who received taxane-based chemotherapy (paclitaxel, albumin-bound paclitaxel, docetaxel) in a Grade III-A hospital in Shanghai from January 2023 to June 2024. Data on demographics, physiological, psychological, and symptomatic variables were collected using a general information questionnaire, the Chemotherapy-Induced Peripheral Neuropathy Assessment Tool (CIPNAT), the Pittsburgh Sleep Quality Index (PSQI), the Psychological Capital Questionnaire (PCQ), and the European Organization for Research and Treatment of Cancer Quality of Life Questionnaire-Core 30 (EORTC QLQ-C30). A symptom network was constructed using the graphical LASSO (least absolute shrinkage and selection operator) based on the Extended Bayesian Information Criterion (EBIC) (EBIC-glasso) algorithm. Centrality analysis was conducted to identify core symptoms; the bootstrap method was used to verify the network accuracy and stability, and the Network Comparison Test (NCT) was applied to analyze differences in network structure between two age groups (≤ 65 years vs. > 65 years).

Among the 315 patients, 86.35% were male, with a median age of 68 years (interquartile range (IQR): 60.50-72.00 years), and 58.73% were aged over 65 years. Three pairs of strongly correlated symptoms were identified in the symptom network: optimism and hope (r = 0.625), symptom frequency and bothersomeness (r = 0.603), and sleep efficiency and sleep duration (r = 0.522). The network was dominated by positive connections and exhibited high global connectivity, indicating that symptoms tended to co-occur and mutually reinforce each other. Centrality analysis showed that fatigue (QLQ7) was the key hub node in the network, with the highest strength centrality (2.735), closeness centrality (2.078), and betweenness centrality (3.944). The frequency of chemotherapy-induced peripheral neuropathy (CIPNAT3) was the driver node of the network, with the highest expected influence (EI = 1.417). The network showed good stability, with a correlation stability (CS) coefficient of 0.673 for strength centrality and expected influence. Subgroup analysis by age revealed no significant differences in network structure (M = 0.240, p = 0.347) or global connectivity (12.516 vs. 12.418, p = 0.802) between the two age groups.

The symptom network of lung cancer patients receiving taxane-based chemotherapy exhibits a tightly interconnected characteristic. Fatigue is the core hub symptom, and the frequency of chemotherapy-induced peripheral neuropathy is the key driver symptom. Additionally, the network structure shows age universality. Clinically, interventions can be prioritized for the aforementioned core symptoms to achieve efficient management of the overall symptom cluster.