Dynamic changes in symptom clusters and their influencing factors in oral cancer patients throughout the treatment cycle: a longitudinal study based on principal component analysis and linear mixed-effects modeling.

purposeThis study examines the symptom clusters present in patients with oral cancer and assesses the impact of age, sex, tumor site, and clinical stage as four covariates on their longitudinal trajectories.

methodsThis study collected data from 108 patients with oral cancer treated with surgery followed by chemoradiotherapy at a single center across six time points: 24 h before surgery, 7 days after surgery, 24 h before discharge, 24 h before initiation of CRT, within 24 h after completion of CRT, and 2 weeks after completion of CRT. PCA was performed on pooled observations from all time points to extract symptom clusters and LME modeled their temporal trajectories. The analysis accounted for age, sex, tumor site, and clinical stage as covariates, incorporating collinearity testing and model diagnostics.

resultsFive clusters were identified: Oropharyngeal-Systemic Discomfort (OSD), Upper GI-Sleep Disruption (USD), Oral Sensory Abnormality (OSA), Neuro-Constipation (NC), and Emotional-Somatic Reaction (ESR). OSD followed a quadratic (increase-then-decrease) trend, USD initially declined before rising, and OSA increased continuously. Clinical stage was significantly associated with OSD and USD trajectories, whereas age primarily related to OSA. Although change rates were comparable across clusters, higher baseline burdens predicted less improvement in USD and OSA, with a similar trend observed for NC. Diagnostics supported robust fit for OSD, USD, and OSA models, while NC and ESR findings warrant cautious interpretation.

conclusionCombining PCA and LME effectively delineates symptom dynamics in oral cancer. Age and clinical stage serve as key factors, supporting the potential for early high-risk patient identification and personalized supportive care.