Aims:Rectal Neuroendocrine Tumors (R-NETs) carry a risk of lymph node metastasis even when smaller than 10 mm, and detecting small R-NETs endoscopically is clinically important. R-NETs are generally recognized as yellowish submucosal tumors; however, the reason why they appear yellow has not been elucidated. In this study, we analyzed R-NETs at our institution to evaluate the diagnostic utility of color characteristics.

Methods:From April 2021 to September 2025, we identified lesions ≤5 mm that had not undergone pre-treatment biopsy and were observed using the Olympus X-1 system (CF-XZ1200). From stored endoscopic images, we extracted tumor and surrounding mucosal areas and measured color values (RGB) at nine points each to obtain mean color values.Analysis 1: We evaluated diagnostic capability based on color differences between tumor and surrounding mucosa. RGB values were converted to the CIE-Lab color space, and color differences (ΔE) were calculated.Analysis 2: We assessed diagnostic utility incorporating pathological findings, specifically the depth of tumor extension from the submucosa into the mucosa.

Results:A total of 32 patients with 35 lesions were included (16 males, 19 females; mean age 56.3 ± 13.1 years). The mean tumor size was 3.5 ± 0.9 mm.Analysis 1: Five endoscopists evaluated whether the color of each tumor could be visually distinguished from the surrounding mucosa. In 16 lesions (46%), differentiation was not possible (non-CD group), whereas in 19 lesions (54%), it was possible (CD group). Although tumor size did not differ between groups, ΔE values showed a significant difference(P<0.01).Analysis 2: Pathological invasion depth was classified as follows: no invasion into mucosa (MI-1): 8 lesions (23%), <1/3 mucosal invasion (MI-2): 9 lesions (26%), ≥1/3 mucosal invasion (MI-3): 18 lesions (51%). While tumor size did not differ among MI categories, ΔE was significantly higher in MI-1 vs MI-3(P<0.01).A nominal logistic regression analysis using color difference as the dependent variable revealed that MI classification was significantly associated with endoscopic color recognition (p < 0.01).

Conclusions:In small R-NETs, the yellowish appearance is not attributable to tumor size but instead to the reverse invasion of tumor cells from the submucosa into the mucosa.