Trajectories of Pan-Immune-Inflammation Value during neoadjuvant immunochemotherapy predict short- and long-term postoperative outcomes in esophageal squamous cell carcinoma: a group-based trajectory analysis

Background While the Pan-Immune-Inflammation Value (PIV) is an established predictor of prognosis following neoadjuvant immunochemotherapy (nICT) in patients with esophageal squamous cell carcinoma (ESCC), the association between longitudinal PIV trajectories and post-surgical outcomes remains unknown. This study aimed to investigate the relationship between dynamic changes in PIV during nICT and short- and long-term outcomes after nICT and surgery. Methods This retrospective study included 324 ESCC patients. Group-based trajectory modeling (GBTM) was employed to identify distinct longitudinal trajectories of PIV during nICT. These trajectory groups were then compared based on pathological response and recurrence-free survival (RFS). Chi-square test and Firth’s penalized logistic regression were used to evaluate pCR, while Kaplan-Meier curves and Cox regression were utilized for survival analysis. Results Three distinct PIV trajectories were identified: the Medium Ascent-Descent Group (Medium Group; n=82), the Low Decline-Rise-Decline Group (Low Group; n=123), and the High Stable Group (High Group; n=119). In the Cox regression analysis, compared with the Low Group, the Medium and High Groups had higher hazards of recurrence or death. Kaplan-Meier analysis also demonstrated that the Low Group had a significantly better RFS than the other two groups. Furthermore, a significant difference was observed in pathological complete response (pCR) among the three trajectory groups (P = 0.01), with the Low Group achieving the highest rate at 25.2%. In multivariable Firth logistic regression, PIV trajectory remained an independent predictor of pCR (both Medium and High Groups showed significantly lower odds of achieving pCR vs. Low Group, P < 0.05). Conclusion Distinct longitudinal trajectories of PIV during nICT are significantly associated with pathological response and RFS in ESCC patients. The dynamic changes of PIV have the potential to serve as a valuable tool for prognostic stratification.