Proximal-To-Distal Gradient and Its Implications in Differentiation and Prognosis in Colorectal Neuroendocrine Neoplasms: The RGETNE-HERACLES Study

BACKGROUND: Colorectal neuroendocrine neoplasms (CR-NENs) are traditionally classified into midgut and hindgut derivatives, a binary classification that potentially oversimplifies the underlying developmental continuum. We investigated whether CR-NEN phenotype and prognosis follow a continuous proximal-to-distal gradient to identify the optimal stratification strategy. METHODS: We analyzed 1,158 patients from the RGETNE registry. Anatomical location was modeled as a continuous variable using restricted cubic splines to dissect spatial trends in proliferation (Ki67), differentiation, and survival. Prognostic performance of the continuous spatial model was formally compared with categorical approaches (right/left/rectum and midgut/hindgut) using the Akaike Information Criterion (AIC). RESULTS: A distinct proximal-to-distal gradient was observed. The continuous spline-based model provided the superior fit for Ki-67 (R2 = 0.494), outperforming categorical approaches. Anatomical position was an independent prognostic factor, with risk increasing distally in localized disease (HR 1.79; 95% CI 1.31-2.45 for rectum vs. cecum) and significant mortality differences in metastatic disease (p < 0.001). Adjuvant benefit was also heterogeneously distributed, being concentrated in distal segments. Despite the biological continuum, the embryological midgut-hindgut dichotomy yielded the most efficient statistical fit (lowest AIC) for predicting clinical recurrence. CONCLUSIONS: CR-NENs exhibit a topographical biological continuum. Continuous modeling is optimal for molecular research and characterizing tumor aggressiveness. However, for clinical risk stratification and trial design, the traditional discrete classifications (midgut-hindgut or anatomical segments) remain the most robust, parsimonious, and practically applicable surrogates.