Accuracy of Endoscopic Ultrasonography for Determining the Treatment Method for Early Gastric Cancer

Background. Endoscopic resection (ER) for early gastric cancer (EGC) is a minimally invasive and curative treatment. The value of endoscopic ultrasonography (EUS) in determining the therapeutic strategy for EGC was assessed in this study. Materials and Methods. Pretreatment EUS was performed on 406 EGCs. The lesions were divided into the histological categories m/sm1 and sm2. The EUS-determined depths of invasion were classified as EUS-M/SM1, EUS-SM2, and EUS-MP or deeper. An analysis of the factors influencing the EUS-based depth determination was then conducted. Results. Most (92.8%) of the EUS-M/SM1 group belonged to the m/sm1 histological category. Ulcerated lesions, tumor size of larger than 2 cm, and the use of an ultrasound endoscope were independently associated with misdiagnosis of the depth of EGC by EUS. The ulcerated lesions had a significantly higher probability of overestimation. Conclusions. EUS is a useful method for determining the therapeutic strategy for EGC. Special attention should be paid not to overestimate the depth of cancer invasion when determining the ulcerated lesions and the type of curative procedure to be used.


Introduction
Endoscopic resection (ER) for the treatment of early gastric cancer (EGC) is currently accepted as a minimally invasive and curative treatment. According to the Japanese Gastric Cancer Treatment Guidelines [1], the indication for ER is a mucosal lesion, less than 2 cm in size, without ulceration. However, the guidelines have also expanded the indications for ER into the following categories that have very low possibilities of lymph node metastasis [1,2]: (1) differentiated, mucosal cancer lesions, larger than 2 cm, without ulcerative findings [UL(−)]; (2) differentiated, mucosal cancer lesions ≤3 cm in size, with ulcerative findings [UL(+)]; (3) undifferentiated, mucosal cancer lesions ≤2 cm, UL(−); (4) differentiated lesions ≤3 cm in size, with submucosal invasion of less than 500 µm (sm1). Therefore, accurate determination of the depth of gastric cancer invasion is increasingly important to the determination of the therapeutic strategy.
Endoscopic ultrasonography (EUS) is one of the diagnostic methods for determining the depth of gastric cancer invasion. In this study, EUS was evaluated for its utility in determining the depth of gastric cancer invasion, and for its necessity in determining therapeutic strategy.

Materials and Methods
Pretreatment EUS was performed on 406 EGCs with histologically proven mucosal and submucosal cancer lesions between January 2006 and December 2009 at Kyoto Second Red Cross Hospital (Kyoto, Japan). Endoscopic mucosal resection (EMR) was performed on 18 lesions; endoscopic submucosal dissection (ESD) was performed on 202; 186 lesions were treated surgically. The results of these treatments were retrospectively reviewed.
Based on the classification system of the Japanese Gastric Cancer Association [3], the locations of the stomach lesions were divided into the upper, middle, and lower thirds of the stomach; each lesion was classified as either differentiated or undifferentiated, based on a histological assessment. The macroscopic features of the lesions were diagnosed by endoscopic findings and were classified as elevated type (0-I and 0-IIa), UL(−) type (0-IIb and 0-IIc without ulcerative  findings), and UL(+) type (0-IIc with ulcerative findings, and 0-III) according to the classifications described in Table 1 and based on whether or not they were ulcerated. The depth of cancerous invasion was histologically classified as follows: lesion confined to mucosal layer (m); <500 µm invasion into the submucosal layer (sm1); >500 µm deep invasion into the submucosal layer (sm2). The lesions were divided into the histological m/sm1 group, for which ER may be suitable, and the sm2 group, for which surgery was indicated.
EUS was used to determine the depth of cancer invasion. Two EUS devices were used: the ultrasound probe (US-probe; UM-3R, 20 MHz, Olympus, Tokyo, Japan) was selected for use with smaller or flat lesions, and the ultrasound endoscope (US-endoscope; GF-UM-2000, Olympus) was selected for use with larger or deep, depressed lesions. The EUS-determined depths of invasion were classified as lesions with no abnormality in the submucosal layer or smooth tapering of the submucosal layer (EUS-M/SM1) (Figures 1(a) and 1(b)); lesions with irregularity of the submucosal layer (EUS-SM2) (Figure 1(c)); lesions with an abrupt interruption of the submucosal or deeper layer (EUS-MP or deeper).
The location, macroscopic features, tumor size, histological type, and EUS type were analyzed to determine if they influenced the EUS diagnosis of the depth of cancer invasion. "Dr. SPSS II for Windows" was used for statistical analysis. A Chi-square test was used for the univariate analyses, and logistic regression was used for multivariate analyses.

Accuracy of EUS and Risk Factors for Misdiagnosis of the Depth of Cancer Invasion.
Of the 406 lesions evaluated, 52 were located in the upper third of the stomach; 45, in the middle third; 309, in the lower third. Morphologically, 152 lesions were classified as the elevated type; 171, the UL(−) type; 83 as the UL(+) type. Histologically, 314 lesions were the differentiated type and 92 lesions were the undifferentiated type. The US-probe was used to evaluate 298 lesions, and the US-endoscope was used in the remaining 108 lesions.
Depending on the macroscopic features, the tumor size, and the histological type, the accuracy of EUS varied widely. The accuracy also varied depending on the ultrasound instrument that was used, with the US-probe and USendoscope having accuracies of 85.2% and 45.3%, respectively. The univariate analysis showed that the accuracy was significantly lower for the UL(+), the tumor size of larger than 2 cm, and the undifferentiated types of lesions as well as for those diagnosed with the US-endoscope (Table 3). Multivariate analysis of these 4 factors showed that the UL(+) type (OR 8.573; 95% CI 4.632-15.867), the tumor size of larger than 2 cm (OR 2.071; 95% CI 1.149-3.731), and the use of US-endoscope (OR 2.472; 95% CI 1.330-4.593) were independently associated with misdiagnosis of the depth of EGC by EUS (Table 4).
In these risk factors for misdiagnosis, the UL(+) type and the use of US-endoscope had a significantly higher probability of overestimation (Table 5).
According to previous reports, lesions with ulcerous changes [8,10] or lesions of the depressed [6,11] or undifferentiated types [6,9] or tumor size of larger than 3 cm [9] or lesions located in the upper third of the stomach [5,11] were associated with incorrect depth determinations by EUS. These reported results are similar to our study.   Another report showed that the accuracy of US-probe was significantly higher than that of US-endoscope [4]. In our study, one of the risk factors for misdiagnosis of the depth of EGC was also associated with the use of the USendoscope. The US-probe is particularly suitable for the determination of the depth of EGC because the frequency of the US-probe is higher than that of the US-endoscope, allowing the US-probe to have a higher resolution within the shallower layers. However, the selection of the type of ultrasound instrument used to make the depth determination was based on the endoscopic appearance of the tumor, such as its size, height of elevation, and depth of depression. The US-probe was used for smaller lesions or lesions with shallower depressions that were easy to diagnose as mucosal cancer, whereas the US-endoscope was used for lesions with a deep ulceration that were difficult to distinguish between a benign fibrosis and a cancerous invasion. This selection bias could explain why the accuracy of the US-probe was higher than that of the US-endoscope.

Therapeutic Strategy of EGC.
In this study, most of the EUS-M/SM1 group lesions belonged to the histological m/sm1 category (Table 2: 92.8%, 260/280). Therefore, ER is appropriate for lesions determined to be EUS-M/SM1. On the other hand, the EUS-SM2 group included many histological m/sm1 lesions (Table 2: 56.5%, 56/99) for which ER, especially ESD, might be a curative treatment. However, most of these lesions have ulcerative changes (Table 6) which are predictive of difficult dissections during ESD. Therefore, although ESD might be considered for lesions determined to be EUS-SM2 or deeper, surgery is also an appropriate treatment for these lesions.

Conclusions
EUS is a useful tool for determining the therapeutic strategy for EGCs. However, EUS is not the best method to correctly determine the depth of the EGC invasion, in the cases of UL(+) lesions or tumor size of larger than 2 cm. Special attention should be paid not to overestimate the depth of cancer invasion when determining the UL(+) lesions and the type of curative procedure to be used. ER should be performed for lesions classified as EUS-M/SM1, whereas surgery is an appropriate treatment for EUS-SM2 lesions.