Exploring the dynamics of symptom networks in esophageal cancer patients during chemotherapy: a cross-lagged panel network analysis.

PURPOSE: This study used cross-lagged panel network analysis to examine longitudinal symptom networks in patients with esophageal cancer undergoing chemotherapy, aiming to identify stage-specific influential symptoms and directed temporal associations across treatment stages.

METHOD: This longitudinal study was conducted from January to September 2025 in four Grade III Level A hospitals in Henan, China. Patients receiving chemotherapy for esophageal cancer were assessed at three time points using the M.D. Anderson Symptom Inventory Gastrointestinal Cancer Module (MDASI-GI). Cross-lagged panel network models were applied to estimate symptom-to-symptom temporal associations and quantify symptom influence using network centrality indices.

RESULTS

A total of 339 participants completed all three assessments. During the T1→T2 interval, S1 (Pain) and S18 (Difficulty Swallowing) exhibited the highest out-expected influence (out-EI = 1.825 and 1.717, respectively), with the strongest temporal association observed from S1 (Pain) to S6 (Shortness of Breath) (β = 0.282). During the T2→T3 interval, S2 (Fatigue) and S3 (Nausea) emerged as the most influential symptoms (out-EI = 2.172 and 1.937, respectively), with the strongest association identified from S2 (Fatigue) to S6 (Shortness of Breath) (β = 0.454).

CONCLUSIONS

This study identified stage-specific influential symptoms and directed temporal associations within symptom networks among patients with esophageal cancer undergoing chemotherapy. Pain and difficulty swallowing were most influential early, whereas fatigue and nausea became more influential later. These findings support a network-based, stage-sensitive perspective on symptom management and may inform time-specific supportive care prioritization.