Patient and lesion characteristics
Among the consecutive patients with superficial ESCC who underwent ESD, 398 elderly patients with 505 superficial lesions were enrolled. Prognoses and recurrences were evaluated in 381 elderly patients (96%). The clinicopathological characteristics of all the patients are shown in Supplementary Tables 1 and 2.
ESD Outcomes
The short-term outcomes are presented in Table 1. The mean procedure time was 77.1 minutes, the en bloc resection rate was 98%, and the complete en bloc resection rate was 95%. Severe fibrosis was observed in 86 lesions (17%). Endoscopic balloon dilatation was performed in 82 patients (16%) because of postoperative stenosis. In addition, perforation, pneumonia, and delayed bleeding were observed in 19 (4%), 15 (3%), and 5 (1%) patients, respectively. All cases were resolved with conservative treatment, and there were no deaths related to ESD. Pathological diagnoses included 380 (75%), 61 (12%), and 64 (13%) EP/LPM, MM, and SM lesions, respectively. Forty-four lesions (9%) had positive lymphatic invasion, and 22 (4%) had positive venous invasion. Noncurative resections by ESD (lymphovascular invasion positive and/or pT1b-SM) were observed with 76 (15%) lesions.
<Table 1>
Prognoses after ESD
The prognoses of 381 patients (96%) were investigated (average follow-up period of 71.1±37.3 months). Of the 76 patients diagnosed with noncurative resection based on the pathological findings from ESD specimens, 40, 7, 1, and 6 were additionally treated with CRT, radiotherapy (RT), chemotherapy, and surgery, respectively, and 22 were followed up without additional treatment.
The causes of death in these patients are shown in Supplementary Table 3. In total, 104 patients died during the observation period; seven of them died due to ESCC. Four of these seven patients were diagnosed with an invasive depth of pT1a without lymphovascular invasion based on the post-ESD pathological findings and were followed up without additional treatment. The diagnoses were pT1a-EP in one case, pT1a-LPM in one case, and pT1a-MM in two cases. Lymph node metastasis from the metachronous carcinoma of the cervix recurred in the patient diagnosed with pT1a-EP, and lymph node metastasis from the primary carcinoma recurred in the other patients. Contrastingly, three of the seven patients who died due to ESCC were diagnosed with invasive depth pT1a with lymphovascular invasion or pT1b based on the post-ESD pathological findings. Two of these three patients were treated with CRT after ESD, and the other patient was followed up without additional treatment.
The most common cause of death from other diseases was cancer, excluding esophageal cancer, pneumonia, and cardiac disease. The breakdown of deaths due to other cancers was as follows: lung cancer, n=7 (19%); oral cancer, n=6 (16%); liver cancer, n=4 (11%); colorectal cancer, n=4 (11%); pancreatic cancer, n=4 (11%); bladder cancer, n=3 (8%); gastric cancer, n=2 (5%); malignant lymphoma cancer, n=2 (5%); myelodysplastic syndromes, n=2 (5%); duodenal cancer, n=1 (3%); renal cancer, n=1 (1%); and occult primary cancer, n=1 (1%).
OS, according to the prognostic factors, is summarized in Table 2. Univariate analysis with log-rank tests showed that age, ASA-PS, Alb, PNI, GNRI, NLR, history of advanced cancer excluding esophageal cancer, and tumor location were significantly associated with impaired survival. However, multivariate analysis showed that age (hazard ratio: 1.02; 95% confidence interval (CI): 1.00–1.04, p=0.0458) and ASA-PS class III (hazard ratio: 1.27; 95% CI 1.01–1.59, p=0.0392) were independently associated with OS.
OS and DSS rates of the elderly patients with ESCC after ESD are shown in Figure 1. No difference in OS and DSS rates was observed between the <75-year-old and ≥75-year-old groups (Figure 1a, 1b). Contrastingly, the OS rate in the ASA-PS class III was significantly lower than that in ASA-PS class I/II (p<0.0001), although no difference in DSS rate was observed between both groups (Figure 1c, 1d). Furthermore, a significantly lower OS rate was observed in the high-risk follow-up group than in the low-risk follow-up and high-risk additional treatment groups (p<0.01). However, no significant difference in DSS was observed between the three groups (Figure 1e, 1f). Contrastingly, a significantly lower DSS rate was observed in the high-risk group of patients in the ASA-PS class III than in the low-risk group of patients in the ASA-PS class III (p<0.05). Furthermore, no difference in DSS was observed between the high-risk, additional treatment group and the high-risk, follow-up group (Figure 1g, 1h).
<Table 2>