Role of endoscopic ultrasonography in determining indications for endoscopic resection of early gastric cancer

Introduction:

Endoscopic resection (ER) of early gastric cancer (EGC) is used for cases fitting criteria that predict minimal risk of lymph nodal metastasis: (1) well-differentiated adenocarcinoma (tub1) or moderately differentiated adenocarcinoma (tub2) limited to mucosa without ulcerative change (UL (-)); (2) tub1 or tub2 limited to mucosa with ulcerative change (UL (+)) (size, ≤30mm), (3) tub1 or tub2, with <500µm of submucosal invasion (sm1) (size, ≤30mm).

Purpose:

To assess the accuracy of endoscopic ultrasonography (EUS) in determining depth of invasion and indications for ER.

Subjects and Methods:

From 2006 to 2010, 300 lesions diagnosed as EGC by EUS were endoscopically or surgically resected at our institution. The macroscopic features were as follows: elevated (n=140); flat or depressed and UL (-) (n=124); depressed and UL (+) (n=36). EMR was performed for 18 lesions; ESD, for 196; and surgery, for 86. EUS-determined depth of invasion was classified as follows: EUS-M (lesion confined to 1st and 2nd sonographic layers); EUS-SM1 (lesion with a distinctly irregular boundary between 2nd and 3rd sonographic layers); and EUS-SM2 (lesion with changes in 3rd sonographic layer).

Results:

Histology	m	sm1	sm2	mp	ss	Total	PPV
EUS
M	198	12	5	0	0	215	92.0% (198/215)
SM1	10	5	6	0	0	21	23.8% (5/21)
SM2	25	7	25	6	1	64	39.0% (25/64)
Total	233	24	36	6	1	300

Unsuspected ulcer scar (2) and cystic change (10) beneath the lesion were detected by EUS in 12 lesions.

The accuracy of EUS was 76.0% (228/300); 30 lesions (10.0%) were underdiagnosed, and 42 (14.0%), overdiagnosed. Misdiagnosis were mainly caused by the wrong evaluation of ulcerative change (25), and by the presence of cystic change beneath the lesion (7).

The positive predictive value (PPV) of EUS-M was 92.0% (198/215); EUS-SM1, 23.8% (5/21); and EUS-SM2, 39.0% (25/64). Of the lesions in the EUS-M and EUS-SM1 groups, 95.3% (225/236) belonged to the “m” or “sm1” histological category. Of the “m” lesions, 10.7% (25/233) were misdiagnosed as EUS-SM2; and surgery was performed for 24 lesions. Of these 24, 4 lesions were elevated; 4, flat or depressed and UL (-); 16, depressed and UL (+).

For all macroscopic types, almost all lesions of the EUS-M and EUS-SM1 groups belonged to “m” or “sm1”: elevated, 94.2% (114/121); UL (-), 98.0% (101/103); and UL (+), 84.6% (11/13). The PPVs of the EUS-SM2 group were lower in the UL (+) type (4.3%) than in the elevated (57.8%) and UL (-) (61.9%) types.

Conclusions:

EUS aids in determining indications for ER of EGC. In case of ulcerative changes, it is difficult to evaluate EUS-SM2-type lesions.