Related Experiment Videos
Changes in postoperative symptom communities and network structure in patients with oral squamous cell carcinoma: A
Shumin Nie1,2, Chunqin Liu3, Yuqing Zhang1,2
1Shandong Provincial Hospital Affiliated to Shandong First Medical University, Jinan, Shandong Province, China.
Objective:
To describe changes in symptom communities and contemporaneous symptom network structure across four postoperative time points in patients with oral squamous cell carcinoma (OSCC) undergoing flap reconstruction.
Methods:
This prospective longitudinal study included 251 participants with complete data. Symptoms were assessed on postoperative days 1 (T1), 3 (T2), 7 (T3), and 30 (T4) using the Chinese version of the MD Anderson Symptom Inventory-Head and Neck Module. Gaussian graphical models with EBICglasso regularization were estimated separately at each time point. Symptom communities were identified using principal component analysis (PCA), with exploratory graph analysis (EGA) performed as a sensitivity analysis. Node strength and bridge strength were summarized. Paired Network Comparison Tests with 2000 permutations compared networks across time points; global-strength comparisons were Bonferroni-adjusted.
Results:
In the PCA solution, three, three, four, and five symptom communities were identified at T1, T2, T3, and T4, respectively; the sensitivity analysis yielded four, three, three, and six communities, indicating methodological sensitivity. Choking cough, somnolence, problems with mucus, and vomiting had the highest node strength at T1, T2, T3, and T4, respectively. Problems with mucus had the highest bridge strength at T1 and T2, lack of appetite at T3, and distress at T4. Global strength differed between T2 and T3 and between T2 and T4 after Bonferroni correction.
Conclusions:
These descriptive, hypothesis-generating findings provide a complementary view of postoperative symptom interrelationships and may support stage-specific symptom assessment. They do not establish causal mechanisms or validated intervention targets.