Size and Macroscopic Type of Type B2 Vessel Areas in JES Classification for Predicting Invasion Depth: A Multicenter Prospective Study

Objectives To investigate whether considering the B2 vessel area (B2VA) size and macroscopic type improves diagnostic performance. Methods In total, 136 esophageal squamous cell carcinoma (ESCCs) with B2 but not B3 vessels were prospectively enrolled. Tumor invasion depth was assessed using the standard Japan Esophageal Society (JES) classification and a “B2VA‐based” method that incorporates B2VA size and macroscopic type. Lesions with B2VA ≤ 4 mm (B2‐narrow) were classified as cEP/LPM; B2VA ≥ 10 mm (B2‐wide) or type 0–I or 0–III within the B2VA were classified as cSM2–3; intermediate lesions were classified as cMM/SM1. All lesions underwent endoscopic resection or esophagectomy with histopathological confirmation. The primary endpoint was diagnostic accuracy, and positive likelihood ratios (+LRs) were calculated for each endoscopic feature. Results Among the 136 ESCCs, 27.2% were pEP/LPM, 49.3% were pMM/SM1, and 23.5% were pSM2–3. The overall diagnostic accuracy was 49.3% for the standard JES evaluation and 50.0% for the B2VA‐based evaluation ( p = 1.00). B2VA size correlated with pathological depth, with median values of 3, 5, and 10 mm for pEP/LPM, pMM/SM1, and pSM2–3, respectively. B2‐narrow predicted pEP/LPM (+LR 2.26), whereas B2‐wide predicted pSM2–3 (+LR 3.45), and type 0–I within the B2VA predicted pSM2–3 (+LR 4.47). Combination of B2‐wide and 0–I demonstrated the highest predictive value for pSM2–3 (+LR 9.75). Conclusions B2VA‐based evaluation incorporating B2VA size and macroscopic type did not improve the diagnostic accuracy compared with that of the standard JES evaluation for predicting invasion depth in ESCC with B2 vessels. Trial Registration Registry and registration number: UMIN000051145