Introduction

Esophageal cancer affects more than 450 000 people worldwide and its incidence is rapidly increasing¹. In patients with resectable non-metastatic esophageal cancer, neoadjuvant chemoradiotherapy followed by surgery is the gold standard treatment with curative intent^2,3. Gastric cancer is also a major global health problem representing the fifth most common disease and the third most frequent cause of cancer related deaths worldwide, with 723 000 deaths per year corresponding to 8.8% of all cancer related deaths⁴.

Over the past decade, the approach to these conditions has changed, firstly due to the positive impact of neoadjuvant treatment, and secondly due to the advent of different tools for better preoperative staging⁵. Nowadays, endoscopic ultrasound (EUS) is considered the gold standard procedure for preoperative locoregional staging of gastric and esophageal cancer^6,9. Several studies evaluating the accuracy of EUS have reported that it is useful to assess the depth of tumor invasion (T stage) with an accuracy greater than 90%. Although its accuracy to assess lymph node involvement (N stage) seems to be lower, previous studies have demonstrated excellent results^7,8. EUS allows for selecting patients who would benefit from neoadjuvant treatment followed by curative intent surgery or who are candidates for local resection.

Unfortunately, Argentina is one of the countries with the highest incidence of esophageal cancer (7.5 and 2.5 per 100 000 inhabitants in men and women, respectively) and gastric cancer (7.5 and 3 per 100 000 inhabitants in men and women, respectively) in South America⁹, and many of these patients cannot access to EUS for preoperative assessment. The aim of this study was to evaluate the usefulness of EUS for staging EAC and GAC and selecting patients who are candidates for neoadjuvant therapy, as compared with the previous stage before the implementation of EUS in a surgical center in Argentina.

Material and methods

We conducted an observational and retrospective analysis of prospectively gathered data between January 2013 and January 2019 at the Hospital Alemán de Buenos Aires. All the patients with EAC and GAC undergoing curative intent surgery were included, except for those with criteria of unresectable cancer (distant metastases, invasion of the aorta or spinal cord) or who underwent emergency surgery.

The sample was divided in four groups considering the organ and the use EUS for preoperative staging: G1, patients with esophageal cancer and EUS; G2, esophageal cancer without EUS; G3, gastric cancer with EUS; and G4, gastric cancer without EUS. All the patients in the four groups underwent computed tomography (CT) scan, positron emission tomography (PET-CT) or both tests for staging. Patients in G1 and G2 with T2 stage or greater or ≥ N1 stage were defined as candidates for neoadjuvant chemoradiotherapy, and those in G3 and G4 T2 stage or greater or ≥ N1 stage were considered candidates for perioperative chemotherapy.

For EUS staging, patients were evaluated after an 8-hour fast. EUS was performed using a 360-degree radial scanning echo endoscope (Olympus GF-UE160AL®) with frequencies of 7.5 MHz and 12 MHz. All the patients were interviewed by the anesthesiologist. The procedure was performed with continuous monitoring, under sedation or general anesthesia in patients with risk factors. All the procedures were performed by the same operator with the patient in the left lateral decubitus position.

The variables included in the analysis were age, sex, indication of neoadjuvant therapy, lymph node metastases in the surgical specimen, overall survival and disease-free survival.

All the variables were compared among the study groups.

Results

During the period analyzed, 89 surgeries for EAC and GAC were performed; 50 of these patients underwent EUS staging.

Most patients with EAC were men (70% in G1 and 83% in G2), while GAC was more common in women (65% in G3 and 62% in G4). More patients with EAC had American Society of Anesthesiologists (ASA) grade ≥ III than those with GAC. Patients’ general characteristics are displayed in Table 1.

Dysphagia was the most common symptom in EAC (21 patients, 53%) while weight loss was more common in GAC (20 patients, 49%).

Twenty-three (75%) patients in G1 and 2 (20%) patients in G2 received neoadjuvant therapy (P ≤ 0.005). Of the patients undergoing preoperative staging, 76% and 40% corresponded to IIB stage or greater in G1 and G2, respectively. Lymph node metastases were observed in 9 (30%) of surgical specimens in G1 and in 60% in G2 (P ≤ 0.005) (Table 2).

In G3 and G4, 40% and 7% of the patients, respectively, received perioperative chemotherapy (P ≤ 0.005). Seventy percent of the patients in G3 and 31% in G4 corresponded to IIA stage or greater. Lymph node metastases were observed in 9 (45%) surgical specimens in G3 and in 13 (45%) in G4 (Table 3).

Mean disease-free survival was 20 (6-72) months in G1, 25 (6-72) months in G2, 32 (6-72) months in G3 and 26 (6-56) months in G4. Overall survival was 32 (6-72) months in G1, 36 (6-72) months in G2, 40 (6- 70) months in G3 and 37 (6-72) months in G4 (Fig. 1, 2, 3, and 4).

Discussion

Neoadjuvant treatment is widely recommended in locally advanced esophageal and gastric cancers. For this reason, EUS has become a very useful tool for preoperative locoregional staging to correctly select patients who would benefit from this treatment followed by surgery and those who are candidates for local resections.

There are plenty of publications supporting EUS as the modality with the highest accuracy for locoregional staging of esophageal cancer compared with CT scan or PET-CT. A review of 21 studies reported an accuracy of 84% for EUS to predict depth of tumor invasion¹⁰. Other studies have reported that the accuracy of EUS for T staging with most ^7.5-12 MHz radial scanning transducers ranges from 75 to 92% compared with CT scan with an accuracy of 42-60%¹¹. However, the accuracy of EUS for staging of esophageal cancer varies according to the depth of tumor invasion (T stage)¹². EUS is more reliable for staging T3 and T4 tumors, with an accuracy of 89-94% and 88-100%, respectively, but is lower for T1 and T2 tumors, with an accuracy of 75-84% and 64-85%, respectively¹⁰. In particular, T2 lesions seem to be most challenging because they are more likely to be over-staged¹³. Rice et al.¹⁴ reported that EUS could differentiate T1 and T2 lesions from T3 or T4 lesions with 87% accuracy, 82% sensitivity and 91% specificity.

EUS represents one of the most accurate modalities available for examining regional lymph nodes to assess metastasis. Endosonographic features predictive of malignancy include size greater than 1 cm, rounded contour, sharply demarcated borders, and hypoechoic structure. When the four features are present, malignancy can be predicted with 80% accuracy; yet, only a few lymph nodes will have the four characteristics¹⁵. The overall accuracy of EUS for N staging is 75-80% compared with 51-74% of CT scan¹⁰ and 37-90% of PTE-CT¹⁶. In a prospective study of 75 patients with recent diagnosis of esophageal cancer, Lowe et al.¹⁷ reported that the sensitivity and specificity of EUS, CT and PET-CT to detect lymph node involvement were 86% and 67%, 84% and 67% and 82% and 60%, respectively.

In our series, the use of EUS increased the number of patients with EAC who received chemotherapy (G1 75% vs. G2 20%) and, as expected, reduced the number of patients with lymph node metastasis in the pathology report and increased survival.

EUS staging of gastric cancer starts with the assessment of the primary tumor, paying particular attention to the depth of tumor invasion, as it changes the outcome and treatment. EUS, CT scan and magnetic resonance imaging (MRI) are the imaging tests used for staging; EUS has shown the best accuracy for T and N staging¹⁸. Kuntz et al.¹⁹ evaluated the three modalities and reported that EUS had the greatest accuracy for T staging (73%) and N staging (87%). The lowest accuracy for T staging was observed with MRI (48%), while MRI and CT scan showed similar performance for N staging (69% and 65%, respectively). Kwee et al.²⁰ reported that although multidetector CT and MRI have improved the accuracy for T staging, EUS remains the first-choice imaging modality for locoregional staging. However, CT and MRI are the best studies for evaluating distant disease; therefore, the complementary role of cross-sectional imaging and EUS in staging of esophageal and gastric cancer is of utmost importance.

We understand the limitations of our study due to its retrospective nature. In our series of patients with GAC we observed a higher number of indications of perioperative chemotherapy in patients undergoing EUS staging (G3 40% vs. G4 7%) which is not reflected in the absolute numbers of lymph node metastases, probably due to the difference in the sample size. In terms of overall survival and disease-free survival, those patients treated with perioperative chemotherapy had better outcome as expected but not statistically significant.

Nowadays, there are not many centers in Argentina that routinely implement EUS for staging of EAC and GAC. We believe that this is partly due to the difficulty in training young endoscopists in the use of this complex tool. The cost of the equipment, materials and their adequate maintenance is another limitation due to the fragile economic situation of our country. We did not find any evidence of the current use of EUS in Argentina for this disease in the literature, but if we want to expand its use, we should start by training our endoscopists in different reference centers in the country and thus increase the access of the population to this practice.

Conclusions

The use of EUS as a tool for locoregional staging in selected patients with EAC and GAC plays a very important role in the selection of patients who will undergo neoadjuvant treatment before curative intent surgery.