Enhanced diagnostic precision of colorectal sessile lesions with blue laser imaging and JNET classification: a multicenter retrospective study in Chinese cohorts

Feng Wu; Wenxin Tan; Panpan Liu; Weiguang Qiao; Tongyin Xing

doi:10.3389/fonc.2025.1534922

Enhanced diagnostic precision of colorectal sessile lesions with blue laser imaging and JNET classification: a multicenter retrospective study in Chinese cohorts

Front Oncol. 2025 May 30:15:1534922. doi: 10.3389/fonc.2025.1534922. eCollection 2025.

Authors

Feng Wu¹, Wenxin Tan², Panpan Liu², Weiguang Qiao², Tongyin Xing²

Affiliations

¹ Department of Critical Care Medicine, Nanfang Hospital, Southern Medical University, Guangzhou, China.
² Department of Gastroenterology, Nanfang Hospital, Southern Medical University, Guangzhou, China.

Abstract

Background: Colorectal cancer (CRC), a leading global malignancy, underscores the need for precise endoscopic diagnosis. Blue Laser Imaging (BLI), a novel endoscopic technology enhancing mucosal surface visualization, combined with the Japan NBI Expert Team (JNET) classification, has shown promise in characterizing colorectal lesions. However, its diagnostic performance in Chinese populations and the impact of endoscopist experience remain underexplored.

Methods: In this multicenter, retrospective study, 131 colorectal sessile lesions were enrolled. The lesions' characteristics were assessed by both expert and trainee endoscopists, utilizing magnified BLI in combination with the JNET classification system to establish diagnostic predictions. This approach allowed for a comparative evaluation of diagnostic accuracy between experienced and less experienced practitioners.

Results: Pathological diagnoses confirmed 2 hyperplastic/sessile serrated lesions (HP/SSL), and 70 low-grade dysplasia (LGD) among the 131 lesions. There were 36 high-grade dysplasia (HGD), 16 superficial submucosal invasive cancers (m-SMs), and 7 deep submucosal invasive cancers (SM-d) demonstrated. The performance metrics for expert and trainee endoscopists in evaluating JNET type 2A(LGD) were as follows: expert endoscopists demonstrated a sensitivity, specificity, positive predictive value (PPV), negative predictive value (NPV), and accuracy of 93%, 93.3%, 94.3%, 91.8%, and 93.1%, respectively; trainee endoscopists showed a sensitivity, specificity, PPV, NPV, and accuracy of 64.6%, 63.5%, 72.9%, 54.1%, and 64.1%, respectively(p<0.01). For JNET type 2B(HGD/m-SMs), expert endoscopists exhibited a sensitivity, specificity, PPV, NPV, and accuracy of 88.2%, 91.3%, 86.5%, 92.4%, and 90.1%, respectively; trainee endoscopists showed a sensitivity, specificity, PPV, NPV, and accuracy of 59.6%, 73.4%, 73.4%, and 67.9%, respectively(p<0.01).

Conclusions: BLI-JNET provides high diagnostic accuracy for colorectal sessile lesions in expert endoscopists, validating its clinical utility. However, trainee endoscopists exhibited significantly low accuracy, underscoring the need for structured training. The proportion of HGD/m-SMs in JNET type 2B lesions within the Chinese cohort (88.2%) was significantly higher than that reported in Japanese data (Kobayashi et al., 2019), highlighting the need to optimize classification systems by incorporating region-specific characteristics.

Keywords: Japan NBI Expert Team (JNET) classification; blue laser imaging (BLI); colorectal cancer; colorectal sessile lesions; pathological prediction.