Predictive value of tumor marker index for pathological complete response following neoadjuvant chemoradiotherapy in locally advanced rectal cancer

Abstract

Background: Pathological complete response (pCR) after neoadjuvant chemoradiotherapy (nCRT) in locally advanced rectal cancer (LARC) is associated with favorable outcomes; however, pre-treatment biomarkers that reliably predict pCR remain limited. We evaluated whether the Tumor Marker Index (TMI; geometric mean of normalized CEA and CA19-9) predicts pCR and examined its relationship with recurrence and survival.

Methods: This single-center retrospective study included 123 stage III LARC patients treated with nCRT followed by total mesorectal excision (2015-2022). The primary endpoint was pCR (ypT0N0). Secondary endpoints were progression-free survival (PFS) and overall survival (OS). Receiver operating characteristic (ROC) analysis assessed TMI discrimination and identified a cut-off. Predictors of pCR were evaluated by univariable and multivariable logistic regression. Systemic inflammatory markers (NLR, PLR, SII) were also analyzed.

Results: Median age was 67 years; 61.0% were male. pCR occurred in 18.7% (23/123). Patients with pCR had lower baseline CEA and TMI. TMI predicted pCR (AUC 0.633, 95% CI 0.509-0.756; p = 0.036); the Youden cut-off ≤ 0.79 yielded sensitivity 86.96% and NPV 92.31%. In multivariable analysis (outcome coded as non-pCR), comorbidity (aOR 3.871; p = 0.035), cT3 vs. cT2 (aOR 4.447; p = 0.026) and higher TMI (per unit; aOR 3.343; p = 0.036) independently increased the odds of non-pCR, whereas a longer RT-surgery interval (per week) reduced it (aOR 0.940; p = 0.016). NLR/PLR/SII were not independent predictors. Recurrence was lower in low-TMI patients (20.5% vs. 57.5%, p < 0.001). pCR was associated with longer PFS (91.6 vs. 62.2 months; log-rank p = 0.012) but not OS (81.9 vs. 64.2 months; p = 0.165).

Conclusions: Pre-treatment TMI is an independent predictor of pCR and correlates with lower recurrence in LARC after nCRT. Given its high sensitivity/NPV at the identified threshold, TMI may support organ-preservation discussions and guide treatment intensification strategies; prospective validation is warranted.

Keywords: CA19-9; CEA; Neoadjuvant chemoradiotherapy; Pathological complete response; Rectal cancer; Systemic immune-inflammation index; Tumor marker index.

Fig. 1

ROC curve for Tumor Marker Index (TMI) in predicting pCR. The area under the curve (AUC) was 0.633(95% CI 0.509–0.756; p = 0.036). The optimal cut-off ≤ 0.79 yielded sensitivity 86.96%, specificity 36.00%, PPV 23.81%, NPV 92.31% LR + 1.36, LR– 0.36, and accuracy 45.53

Fig. 2

Overall survival (OS) according to pathological complete response. Kaplan–Meier curves comparing OS between the pCR and non-pCR groups. Mean OS was 81.86 months (95% CI 68.87–94.84) in the pCR group and 64.24 months (95% CI 57.59–70.88) in the non-pCR group (log-rank p = 0.165)

Fig. 3

Progression-free survival (PFS) according to pathological complete response. Kaplan–Meier curves comparing PFS between the pCR and non-pCR groups. Mean PFS was 91.57 months (95% CI 81.77–101.37) in the pCR group and 62.16 months (95% CI 54.98–69.35) in the non-pCR group (log-rank p = 0.012)