This retrospective cohort study enrolled patients with pathologically confirmed RNETs undergoing endoscopic resection between 2013 and 2024 at the Department of Gastroenterology at Qilu Hospital of Shandong University. Patients were excluded if they had neuroendocrine carcinomas or incomplete clinicopathological documentation. Institutional Review Board approval for this retrospective report was obtained from the ethics committees of Qilu Hospital, Shandong University (No. KYLL-202306-009). Informed consent was obtained from all participants. In our department, all RNETs scheduled for endoscopic submucosal dissection (ESD) treatment were first evaluated for size via endoscopy and for depth and staging via endoscopic ultrasound (EUS). Before ESD, patients routinely undergo radiological imaging examinations such as CT or MRI to rule out lymph node or distant metastasis. Patients who met the criteria for endoscopic resection (size smaller than 2 cm, EUS stage limited to the T1 stage, and without radiologically detected lymph node or distant metastasis) underwent ESD treatment (6). The lesions were completely removed and sent for pathological examination.

Clinical, pathological, and endoscopic data were retrospectively collected from the inpatient medical record system of the Department of Gastroenterology at Qilu Hospital of Shandong University and the Medikang Digital Gastrointestinal Endoscopy Workstation. Demographic parameters, including gender, age, body mass index (BMI), smoking, and alcohol consumption, were systematically documented. Clinicopathological characteristics such as tumor size, number of lesions, anatomical localization, endoscopic resection method, and histopathological characteristics, as well as treatments, were also recorded. Tumor size was determined through histopathological measurement.

Endoscopic specimens were evaluated by pathologists, and immunohistochemical staining and special staining were performed to confirm the tumor size, the involvement of the excised margins, depth of invasion, and the presence of lymphovascular invasion. Each tumor was graded based on its Ki-67 index according to the WHO criteria (G1, <3%; G2, 3%–20%; and G3, >20%). Histologically complete (R0) resection was defined as the microscopic absence of tumor cells at the resection margin. Incomplete resection (R1) was defined as the presence of tumor tissue in the resection margin (vertical or lateral). Indeterminate resection (Rx) was defined as a margin status that could not be assessed because of electrocautery artifacts and inappropriate orientation or fragmentation. For standard diagnostic markers, NETs were classified as positive if they were detected in clumps in the majority of cells, while NETs expressed in scattered single cells (<5%) were classified as focally positive/negative (further referred to as negative).

SSTR2 expression was initially classified into four categories according to the HER2 scoring scheme as proposed by Kasajima et al. (7). For further analyses, all tumors with an SSTR2 2+/3+ score were assigned to the SSTR2-positive category, while SSTR2 1+ and completely negative NETs were assigned to the SSTR2-negative category. During routine clinicopathological examination post-tumor resection, the most locally advanced tumor was identified as the “leading primary NET” and was evaluated regarding its Ki-67 index in order to determine the tumor grade according to WHO criteria and the size of lesions, referring to the definitions in previous literature (8). The leading primary NET served as the reference for our comparative analyses of all multiple tumors.

The follow-up strategy was determined based on the pathological grading and the status of the resection margins. Endoscopy and radiological follow-ups were recommended based on the stages and grades of the lesions. In our center, for RNETs smaller than 1 cm, no follow-up was recommended; for those that are 1 to <2 cm, G2 tumor, or Rx resection, endoscopy and radiological follow-ups at 6 and 12 months were recommended, then as clinically indicated. Among the cases pathologically evaluated as R1 resection or lymphovascular invasion, additional surgery, including lymph node dissection or other salvage treatments, was recommended. The follow-up time was from endoscopic treatment to the latest outpatient visit. The follow-up results were based on the latest outpatient medical record or electronic examination report. Local recurrence was defined as NETs diagnosed at the same site 6 months after the initial resection, whereas metachronous lesions were defined as NETs detected at a site different from the initial rectal tumor 6 months after the initial resection. Disease-free survival (DFS) was defined as the duration from tumor resection to locoregional recurrence, developing metachronous lesions, or distant metastasis.

Normally distributed continuous variables are presented as the mean with standard deviation (SD) and compared using Student’s t-test. Non-normally distributed continuous variables were reported as median with an interquartile range (IQR) and compared using the Mann–Whitney U test. Categorical variables were compared using the χ² test or Fisher’s exact test.

Propensity score matching (PSM) was adopted for comparing the disease-free survival analysis between the solitary and multiple rectal neuroendocrine groups and to minimize selection bias. The matching ratio was 4:1, and the caliper value was set to 0.05. The nearest-neighbor method was used for PSM analysis. Survival curves were generated using the Kaplan–Meier method and were compared using the log-rank test. A two-tailed p-value <0.05 was considered statistically significant. All statistical analyses were conducted using SPSS 25.0 (SPSS Inc., Chicago, IL, USA) and R version 4.4.0 (R Foundation for Statistical Computing, Vienna, Austria).

A total of 205 patients were considered for inclusion in this study, including 189 patients with solitary RNETs (group A) and 16 patients with multiple RNETs (group B). However, 100 patients in group A and one patient in group B were excluded because of incomplete clinicopathological data. Finally, 15 patients with multiple RNETs and 89 patients with single RNETs were included in the study. The multifocal group contained 35 lesions, including 12 patients with two lesions, two with three lesions, and one with five lesions. After preoperative endoscopic evaluation, all lesions were found to be smaller than 2 cm. Ultrasonography showed that all lesions were confined to the T1 stage, which met the indications for endoscopic treatment. All patients underwent ESD as the primary treatment ( Table 1 ).

Endoscopic evaluation revealed predominantly flat-type submucosal yellowish-white protrusions (80%) without ulceration. Red vascular patterns were observed in 6.7% (1/15) of cases.

The median age of the patients was 50 years (IQR, 44.0–53.5). Ten patients were male (66.7%), and five patients were female. The median tumor size was 0.6 cm (IQR, 0.5–0.7 cm). Macroscopically, 80% of the tumors were of the flat type. Most tumor invasion was confined to the submucosal layer in all patients, and no lymphovascular invasion was identified ( Figure 1 ). More detailed clinicopathological characteristics of the multiple rectal NETs are presented in Table 2 .

The clinical differences between the groups are summarized in Table 2 . Macroscopically, group B was more likely to contain flat-type tumors compared to group A (80.0% vs. 52.8%, p < 0.05). However, there were no significant differences with respect to the presence of solitary and multiple tumors in other features. In the matched cohort (Group A, n = 32; Group B, n = 14), the selected variables were balanced across the two groups ( Table 2 ). In group B, 21 lesions were identified among 14 patients.

Table 3 summarizes the prognostic factors in the matched solitary and multiple NET groups. The SSTR2 positivity rate was significantly higher in group A compared to group B (84.4% vs. 57.1%, p < 0.05). Conversely, the positivity rate of CgA was significantly higher in group B than in group A (57.1% vs. 21.9%, p < 0.05). However, there were no significant differences in platelet-to-lymphocyte ratio (PLR) and neutrophil-to-lymphocyte ratio (NLR) between the two groups (p > 0.05).

The histological complete resection rate (R0) of multiple RNETs was 68.6% (24/35), 11.4% (4/35) R1 resection (two lesions with positive horizontal margins and two with positive vertical margins), and 20% (7/35) Rx resection. As for the safety of ESD procedures, none of the 15 patients in group B had intraoperative complications, including bleeding or perforation.

The overall median follow-up period was 20 months (range, 3–60). None of the patients in group A developed recurrence or metastasis. One of the 15 patients in group B developed local recurrence with local lymph node metastasis (LNM). This patient had two tumors (G1 and G2), each measuring less than 10 mm in diameter, with Rx margins and no muscular invasion. After endoscopic resection, the patient was treated with growth inhibitor analogs.

Additionally, metachronous rectal NETs were diagnosed in two patients with both G1 lesions, one of whom also had local LNM. The patient with LNM did not receive further treatment or evaluation. One patient with postoperative pathology showing G2 in all lesions and R0 margins has undergone three follow-up colonoscopies to date, none of which indicated recurrence or metachronous lesions. However, somatostatin receptor scintigraphy revealed multiple bone concentrations suggestive of bone metastasis, and the patient was subsequently treated with somatostatin analogs. The clinicopathological characteristics of these patients are detailed in Table 4 . There was a significant difference in DFS between the two groups (p = 0.0003) ( Figure 3 ).