Patient characteristics
The median age of the 41 patients was 67 years; the patient characteristics are detailed in Table 1. Macroscopically, protruding tumors were detected in 31 patients, and the median maximum tumor diameter was 20 mm. ESD and surgery were performed as initial treatments in 13 and 28 patients, respectively. The histological types in 21, 17, and 3 patients were well differentiated (tub1), moderately differentiated (tub2), and poorly differentiated (por), respectively. Clinicopathological characteristics did not differ significantly between patients with SSBE and with LSBE.
Histopathological findings
Table 2 shows the histological type and number of patients with LVI on H-E staining and immunostaining for D2-40/CD31 according to the depth of invasion. Overall, 21 and 20 patients had pT1a and pT1b lesions, respectively, and 12 of the 21 patients with pT1a had DMM lesions. The depth of SM infiltration in endoscopic resection exceeded 200 μm in 3 patients, and the depths were 400, 800, and 1,300 μm, respectively. These patients were diagnosed as SM2. Among them, 2 patients underwent additional surgery, which revealed no residual cancer or lymph node metastases. The remaining patient preferred follow-up; no relapse has been observed during the 3-year follow-up after ESD. The incidence of tub2 and por histological subtypes increased with the depth of invasion.
In 7 patients, LVI positivity was noted using H-E-stained specimens alone (positivity rate: 17.1%), and the depth of invasion was evaluated to be SM1 or deeper. LVI was found in 10 patients (positivity rate: 24.4%) who were additionally diagnosed with LVI positivity on immunostaining for D2-40 and CD31. The LVI-positivity rates in SMM, LPM, DMM, SM1, SM2, and SM3 lesions were 0, 0, 0, 75, 28.6, and 55.6%, respectively. Overall, between H-E staining alone and immunostaining, LVI was consistently absent in 85.4% (35/41) cases. LVI was additionally detected on immunostaining in cases with SM1 (Fig. 2), in which the lymphatic endothelial cells were very thin near the tumor margin (site where LVI diagnosis is relatively easy), making recognition difficult, and in cases with SM3 (Fig. 3), in which the tumor volume was large, making the identification of LVI at the site of tumor infiltration impossible.
The patients in the SM group were sub-divided into two groups based on the depth of infiltration as follows: <500 μm and ≥500 μm. The former subgroup had 5 patients with SM1 lesions (4 and 1 underwent surgery and endoscopic resection, respectively), with a depth of infiltration of 400 μm. The latter subgroup included 15 patients. LVI was present in 3 (60%) and 7 (46.7%) patients. No specific pattern of distribution of LVI sites was observed.
Lymph node metastasis and relapse
Table 3 shows the pathological findings in 30 surgically treated patients with superficial cancer, including 2 who underwent additional treatment after ESD. In total, 3/41 (7.3%) patients showed lymph node metastases. These patients had protruding SSBE-derived lesions, invading at least up to SM2 (depth of infiltration: >1,000 µm). The tumor diameters were ≥25 mm, and they contained poorly-differentiated components. LVI was identified in 2 of 3 patients.
Overall survival rate and relapse-free survival
There were no significant differences in overall, disease-specific, and relapse-free survival between patients with T1a and T1b disease (Fig. 4). However, the recurrence rate was slightly higher among patients with T1b lesions. Relapse occurred in 3 patients with T1b disease, with a median follow-up of 46 months. In all 3 patients, LVI was present, the depth of invasion was evaluated to be at least SM2, poorly differentiated components were observed, and the tumor diameter was ≥20 mm. Among them, 1 patient died of primary disease. The 3-year disease-specific survival rate in those with T1a and T1b disease was 100% and 95.0%, respectively.