Data from pT1 rectal adenocarcinoma patients who underwent surgery from January 1990 to December 2011 were retrospectively reviewed. The data included age, sex, lesion size, distance from the anal verge, serum carcinoembryonic antigen (CEA) level, histology type and grade, lymphovascular invasion, depth of invasion, and adjuvant therapy. Exclusion criteria were: (1) preoperative suspicion of lymph node metastasis, (2) age greater than 79 years or less than 19 years, (3) upper rectal cancer, i.e., a lower tumor border located 10 cm or more proximally from the anal verge, (4) noncurative surgery (R1 or R2), (5) preoperative chemoradiation, (6) pathology other than adenocarcinoma, (7) recurrent or metastatic cancer, (8) unclear pathologic data, including depth of tumor or lymphovascular invasion, and (9) less than 6 months of follow-up. This study was approved and was exempted from informed consent by the Institutional Review Board of the Asan Medical Center (approval number: 2018-1135).

The indications and detailed procedures for TAE in our institution were described in our previous study [5]. Indications for TAE include a small (less than 3 cm), movable, palpable with index finger and well to moderately differentiated tumor. The tumor should be at the T1 stage without invading more than 30% of the rectal circumference and without lymph node metastasis on preoperative transrectal ultrasonography, computerized tomography, or magnetic resonance imaging.

Patients had 4 L of polyethylene glycol orally on the day before surgery and fasted from midnight. Surgery was performed under either general or spinal anesthesia. A urinary catheter was inserted after the anesthesia, and the patient’s position was changed to a prone jack-knife position. After dilation and adequate traction of the anal canal, the tumor was identified and excised under direct vision at full thickness with margins of at least 0.5 cm. The rectal wall defect was repaired with absorbable sutures. A single surgical instrument used for transendoscopic microsurgery has not been identified at our institution. Cases of transanal minimally invasive surgery were not included due to the heterogeneity of the procedure and scarcity of cases.

Patients were placed into a TAE or a TME group, depending on the procedure. Five patients that underwent TAE followed by TME because of adverse histologic findings were included in the TME group. Distance from anal verge was dichotomized at under or over 5 cm from the anal verge, which divides the low rectum from the mid rectum [6]. Tumor size was dichotomized at ≥3 cm or <3 cm [7]. The threshold for elevated serum CEA was 6 ng/mL, which is the normal limit at our institution. The depth of invasion into the submucosa was classed as sm1 (upper third), sm2 (middle third), and sm3 (lower third) for sessile tumors (n = 254) [8]. Pedunculated lesions (n = 14) were categorized by Haggitt level [9]. Haggitt levels 3–4 and sm2–3 were grouped as deep submucosal invasions [4, 10]. The favorable histology group had well-differentiated or moderately differentiated adenocarcinomas, whereas the unfavorable histology group contained poorly differentiated or mucinous adenocarcinomas [5, 11]. The risk factors for lymph node metastasis were: defined tumors with any type of deep submucosal invasion, lymphovascular invasion, and unfavorable histology. Recurrence included LR, systemic recurrence, and both recurrences concomitantly.

Patients received standardized postoperative follow-up, including clinical examinations, complete blood counts, blood chemistry tests, serum CEA quantitation, and chest radiography, every 6 months for the first 2 years postoperatively and every 12 months thereafter. They also underwent abdominal and pelvic computed tomography (CT) scanning and transrectal ultrasound (TUS) every 12 months. Colonofiberoscopy or sigmoidofiberoscopy was performed every 6 months after the first postoperative year and then annually thereafter. Patients with suspected recurrence had individualized examinations including specific imaging work-up (e.g., TUS, CT, magnetic resonance imaging, and/or positron emission tomography) or biopsy. Survival status and date of death were confirmed by follow-up data.

Adjuvant chemoradiation therapy was considered in cases within the TME group with lymph node metastasis and within the TAE group that exhibited sm2 or sm3, lymphovascular invasion, unfavorable histologic findings, or close resection margins. Radiation therapy was delivered by a linear accelerator (Clinac 1800 instrument; Varian Medical Systems, Palo Alto, CA, USA). Patients were instructed to lie in a prone position, and 6 to 15 MV of radiation was applied through 3 or 4 fields to the pelvis. Patients were irradiated with 2 Gy, 5 days a week, for a total dose of 50 Gy. The superior border for irradiation was the bottom edge of the fifth lumbar vertebra, and the inferior border was 3 cm caudal to the tumor. The lateral borders were 1.5 to 2.0 cm medial from the inner bony margin of the most lateral portion of the true pelvis. The anterior border was 3 cm ventral to the tumor, and the posterior border was 1 cm from the dorsal margin of the sacrum. The primary tumor, regional lymph nodes, and perirectal fat tissue were included in the target volume [5].

Data were reported as mean ± standard deviation for continuous variables and as frequency (%) for categorical variables. Univariate analyses were conducted to compare patient characteristics of the TAE and TME groups. Pearson chi-square test or Fisher exact test were used for categorical data, and the Student t-test was used for continuous data. The LR, recurrence-free survival (RFS), overall survival (OS), and cancer-specific survival (CSS) rates were calculated using the Kaplan-Meier method, and the log-rank test was used to compare differences among survival curves in univariate analyses. The potential prognostic factors that were screened in univariate analysis were further analyzed by multivariate analysis using a Cox regression model, and hazard ratios (HRs) were calculated with 95% confidence intervals (CIs). A P-value of <0.05 was considered statistically significant for all analyses, and all calculations were carried out using IBM SPSS Statistics ver. 24.0 (IBM Co., Armonk, NY, USA).

Of the 268 pathologic T1 rectal cancer patients, 61 (22.8%) underwent TAE and 207 (77.2%) underwent TME. The TAE group was characterized by tumors that were closer to the anus (P < 0.001), more frequent deep submucosal invasion lesions (P = 0.02), and more frequently infiltration (P = 0.033). Thirty-two patients (52.5%) in the TAE group had risk factors for lymph node metastasis, which was less frequent than for the TME group (P = 0.013) (Table 1). There were no differences between the groups in age, sex ratio, tumor size, lymphovascular invasion, tumor differentiation, and serum CEA level.

The follow-up period had a median of 10.4 years (range, 1.6–20.6 years), and the 10-year OS rate for TAE patients was not different from the TME group (91% vs. 90%) (Table 2). The 10-year CSS rate for TAE was slightly lower than for the TME group (98% vs. 100%), although this result was not statistically significant (P = 0.22) (Fig. 1). There was no difference in the 10-year OS and CSS rates for lymph node metastasis risk factors (Table 2).

Twelve patients had systemic and/or LR, which are described in Table 3. In total, 75% of recurrence cases (n = 9) developed within 5 years postoperatively. The 10-year RFS rate for TAE was lower than that for TME (86% vs. 99%, P < 0.001). A TAE surgical procedure was the only independent risk factor for recurrence (HR, 6.8; 95% CI, 2.0–22.5; P = 0.002) (Table 4).

The 10-year LR rate for TAE was greater than for TME (10% vs. 0%, P < 0.001) (Fig. 2). Among patients with risk factor for lymph node metastasis, the 10-year LR rate for TAE was 16% compared to 1% of TME (P < 0.001) (Table 2). Of the 5 TAE patients with LR, four had salvage surgery but one of the patients eventually died. The patient that did not undergo salvage surgery was lost to follow-up immediately after confirmation of recurrence. The median time from initial surgery to salvage surgery was 18 months, and the mean survival after LR was 12 years. The TAE surgical procedure (HR, 19.1; 95% CI, 2.3–159.8; P = 0.007) was the only independent risk factor for LR (Table 5).