Endoscopic laryngopharyngeal surgery (ELPS) and endoscopic submucosal dissection (ESD) are minimally invasive endoscopic techniques for the treatment of head and neck tumors. ELPS is performed by otolaryngologists under gastrointestinal endoscopic guidance, whereas ESD is conducted by gastroenterologists with laryngeal exposure provided by otolaryngologists. Both procedures are performed collaboratively by gastroenterologists and otolaryngologists under general anesthesia, enabling organ preservation and favorable therapeutic outcomes. This study aimed to compare the clinical outcomes of ESD and ELPS for superficial head and neck tumors.

A total of 120 patients with 129 lesions who underwent endoscopic treatment for head and neck tumors between June 2014 and December 2024 were retrospectively analyzed. Parameters assessed included (1) patient demographics and lesion characteristics, (2) short-term therapeutic outcomes, (3) procedural accidents, and (4) recurrence rates.

(1) Among the 120 patients, 110 were male and 10 female, with a median age of 69 years (range, 45–82). Lesions were located in the hypopharynx (n = 111), oropharynx (n = 9), and larynx (n = 9). Macroscopic types were 0-Is (n = 10), 0-IIa (n = 25), 0-IIb (n = 82), and 0-IIc (n = 12), with a median tumor size of 15 mm (range, 3–70). Pathological diagnoses included pTis (n = 47), pT1 (n = 59), pT2 (n = 18), pT3 (n = 3), and dysplasia (n = 2). (2) ESD was performed in 33 patients with 36 lesions, and ELPS in 88 patients with 93 lesions. The median resection speed, calculated as (long diameter × short diameter) / procedure time (mm²/min), was 13.9 (range, 2.2–32.7) in the ESD group and 11.3 (range, 1.1–73.9) in the ELPS group, with no significant difference (p = 0.38). The R0 resection rate was significantly higher in the ESD group (88.9%, 32/36) than in the ELPS group (48.4%, 45/93) (p < 0.0001). In the ESD group, horizontal margin indeterminate (HMX) status was observed in four lesions, due to tumor or dysplasia near the margin (n = 2), thermal denaturation (n = 1), and specimen incision (n = 1). In the ELPS group, HM1, HMX, VM1, and VMX statuses were noted in 15, 32, 2, and 6 lesions, respectively. Causes of HMX included tumor or dysplasia near the margin (n = 16), specimen crush or epithelial detachment (n = 8), thermal denaturation (n = 7), and specimen incision (n = 1). VMX was attributed to thermal denaturation (n = 5) and specimen crush (n = 1). (3) Procedural accidents occurred in seven patients (21.2%) in the ESD group, including aspiration pneumonia (n = 2), tongue injury (n = 2), laryngeal edema (n = 1), hypoglossal nerve palsy (n = 1), and vocal cord paralysis (n = 1). In the ELPS group, complications were observed in 16 patients (18.2%), including aspiration pneumonia (n = 6), laryngeal edema (n = 5), dysphagia (n = 2), free jejunal perforation (n = 1), tongue/oral floor swelling (n = 1), and prerenal failure (n = 1). There was no significant difference in the incidence of procedural accidents between the ESD and ELPS groups (p = 0.71). (4) In the ELPS group, local residual or recurrent lesions were observed in five cases (5.4%) and distant metastases in four cases (4.3%), whereas no such events occurred in the ESD group. There were no significant differences between the two groups in the rates of local residual or recurrent lesions (p = 0.16) and distant metastases (p = 0.21).

Although resection speed, incidence of procedural accidents, and frequency of residual or recurrent lesions were comparable between ESD and ELPS, the significantly higher R0 resection rate observed in the ESD group suggests that ESD may offer superior curative potential in the management of superficial head and neck tumors.