Objective: This study assessed the accuracy of linear endoscopic ultrasound (EUS) to diagnose submucosal (SM) invasion and compared linear EUS with mini-probe EUS in suspected early gastric cancer (EGC) patients.
Methods: Patients diagnosed with biopsy-verified suspected EGC were analyzed retrospectively. All cases were examined by linear EUS or miniprobe EUS for preoperative diagnosis of invasion depth and underwent endoscopic or surgical treatment for radical resection. The invasion depth evaluated by EUS and pathology were categorized as no invasion of SM and invasion of SM or deeper. The diagnosis of EUS was compared with postoperative pathology results.
Results: A total of 105 patients were included in the final analysis. The overall prediction accuracy of linear EUS (n = 57) for SM invasion in suspected EGC was higher than that of mini-probe EUS (n = 48), although no statistically significant differences were noted (82.5% vs 72.9%, p = 0.344). The negative predictive value (NPV) of linear EUS was significantly higher than that of mini-probe EUS (100% vs 82.8%, p = 0.037). The binary logistic regression analysis showed that tumor size (p = 0.036), the presence of ulceration (p < 0.001) and EUS type (p = 0.027) were independent risk factors for the diagnosis of SM invasion by EUS. The area under the receiver operating curve (ROC) was 0.889 and 0.719 for linear and mini-probe EUS, respectively.
Conclusion: Linear EUS diagnosed suspected EGC for SM invasion with a higher accuracy than mini-probe EUS. In addition, large and ulcerative lesions may lead to overestimation.