Concordance of endoscopic ultrasonography-guided fine needle aspiration diagnosis with the final diagnosis in subepithelial lesions

Diagnostic yield of endoscopic and EUS-guided biopsy techniques in subepithelial lesions of the upper GI tract: a systematic review

BACKGROUND AND AIMS

Obtaining adequate tissue samples in subepithelial lesions (SELs) remains challenging. Several biopsy techniques are available, but a systematic review including all available techniques to obtain a histologic diagnosis of SEL is lacking. The aim of this study was to evaluate the diagnostic yield and adverse event rates of endoscopic biopsies, EUS-guided FNA (EUS-FNA), EUS-guided fine-needle biopsy (FNB) (EUS-FNB), and mucosal incision-assisted biopsy (MIAB) for SELs in the upper GI tract.

METHODS

A search strategy in multiple databases was performed. The primary outcome was diagnostic yield, defined as the percentage of procedures in which histology was obtained and resulted in a definitive histopathologic diagnosis. Secondary outcome measures included reported procedure-related adverse events, which were graded according to the AGREE (Adverse Events in Gastrointestinal Endoscopy) classification.

RESULTS

A total of 94 original articles were included. Studies were classified per endoscopic technique to obtain histopathology. This resulted in 8 included studies for endoscopic biopsy methods, 55 studies for EUS-FNA, 33 studies for EUS-FNB, and 26 studies for MIAB. Pooled rates for diagnostic yield were 40.6% (95% confidence interval [CI], 30.8-51.2) for endoscopic biopsy, 74.6% (95% CI, 69.9-78.7) for EUS-FNA, 84.2% (95% CI, 80.7-87.2) for EUS-FNB, and 88.2% (95% CI, 84.7-91.1) for MIAB. Reported procedure-related adverse events graded AGREE II or higher were 2.8% to 3.9% for endoscopic biopsies, 1.0% to 4.5% for EUS-FNA, .9% to 7.7% for EUS-FNB, and 1.9% to 7.9% for MIAB.

CONCLUSIONS

Based on the available evidence, MIAB and EUS-FNB seem to be most effective in terms of achieving a high diagnostic yield, with similar rates of adverse events.

Differentiating Gastrointestinal Stromal Tumors from Leiomyomas Using a Neural Network Trained on Endoscopic Ultrasonography Images

BACKGROUND

Endoscopic ultrasonography (EUS) is crucial to diagnose and evaluate gastrointestinal mesenchymal tumors (GIMTs). However, EUS-guided biopsy does not always differentiate gastrointestinal stromal tumors (GISTs) from leiomyomas. We evaluated the ability of a convolutional neural network (CNN) to differentiate GISTs from leiomyomas using EUS images. The conventional EUS features of GISTs were also compared with leiomyomas.

PATIENTS AND METHODS

Patients who underwent EUS for evaluation of upper GIMTs between 2010 and 2020 were retrospectively reviewed, and 145 patients (73 women and 72 men; mean age 54.8 ± 13.5 years) with GISTs (n = 109) or leiomyomas (n = 36), confirmed by immunohistochemistry, were included. A total of 978 images collected from 100 patients were used to train and test the CNN system, and 384 images from 45 patients were used for validation. EUS images were also evaluated by an EUS expert for comparison with the CNN system.

RESULTS

The sensitivity, specificity, and accuracy of the CNN system for diagnosis of GIST were 92.0%, 64.3%, and 86.98% for the validation dataset, respectively. In contrast, the sensitivity, specificity, and accuracy of the EUS expert interpretations were 60.5%, 74.3%, and 63.0%, respectively. Concerning EUS features, only higher echogenicity was an independent and significant factor for differentiating GISTs from leiomyomas (p < 0.05).

CONCLUSIONS

The CNN system could diagnose GIMTs with higher accuracy than an EUS expert and could be helpful in differentiating GISTs from leiomyomas. A higher echogenicity may also aid in differentiation.

Natural History of Asymptomatic Esophageal Subepithelial Tumors of 30 mm or Less in Size

BACKGROUND

No definite guidelines for the management of small esophageal subepithelial tumors (SETs) have been established, because there are limited data and studies on their natural history. We aimed to assess the natural history and propose optimal management strategies for small esophageal SETs.

METHODS

Patients diagnosed as esophageal SETs ≤ 30 mm in size between 2003 and 2017 using endoscopic ultrasound (EUS) with a minimal follow-up of 3 months were enrolled, and their esophagogastroduodenoscopy (EGD) and EUS were retrospectively reviewed.

RESULTS

Of 275 esophageal SETs in 262 patients, the initial size was < 10 mm, 10-20 mm, and 20-30 mm in 104 (37.8%), 105 (38.2%), and 66 (24.0%) lesions, respectively. Only 22 (8.0%) SETs showed significant changes in size and/or echogenicity and/or morphology at a median of 40 months (range, 4-120 months). Tissues of 6 SETs showing interval changes were obtained using EUS-guided fine needle aspiration biopsy; 1 was identified as a gastrointestinal stromal tumor (GIST) and was surgically resected, while the other 5 were leiomyomas and were regularly observed. Eight SETs showing interval changes were resected surgically or endoscopically without pathological confirmation; 1 was a GIST, 2 were granular cell tumors, and the other 5 were leiomyomas.

CONCLUSION

Regular follow-up with EGD or EUS may be necessary for esophageal SETs ≤ 30 mm in size considering that small portion of them has a possibility of malignant potential. When esophageal SETs ≤ 30 mm show significant interval changes, pathological confirmation may precede treatment to avoid unnecessary resection.

Use of Artificial Intelligence in the Prediction of Malignant Potential of Gastric Gastrointestinal Stromal Tumors

BACKGROUND AND AIMS

This study aimed to investigate whether AI via a deep learning algorithm using endoscopic ultrasonography (EUS) images could predict the malignant potential of gastric gastrointestinal stromal tumors (GISTs).

METHODS

A series of patients who underwent EUS before surgical resection for gastric GISTs were included. A total of 685 images of GISTs from 55 retrospectively included patients were used as the training data set for the AI system. Convolutional neural networks were constructed to build a deep learning model. After applying the synthetic minority oversampling technique, 70% of the generated images were used for AI training and 30% were used to test AI diagnoses. Next, validation was performed using 153 EUS images of 15 patients with GISTs. In addition, conventional EUS features of 55 patients in the training cohort were evaluated to predict the malignant potential of GISTs and mitotic index.

RESULTS

The overall sensitivity, specificity, and accuracy of the AI system for predicting malignancy risk were 83%, 94%, and 82% in the training dataset, and 75%, 73%, and 66% in the validation cohort, respectively. When patients were divided into low-risk and high-risk groups, sensitivity, specificity, and accuracy increased to 99% in the training dataset and 99.7%, 99.7%, and 99.6%, respectively, in the validation cohort. No conventional EUS features were found to be associated with either malignant potential or mitotic index (P > 0.05).

CONCLUSIONS

AI via a deep learning algorithm using EUS images could predict the malignant potential of gastric GISTs with high accuracy.

Subepithelial esophageal tumors: a single-center review of resected and surveilled lesions

BACKGROUND AND AIMS

Subepithelial esophageal tumors (SETs) are frequent incidental findings. Although symptomatic tumors are surgically or endoscopically resected, there is no consensus on the management of asymptomatic esophageal leiomyomas.

METHODS

Appropriate International Classification of Diseases, Ninth Revision codes followed by medical record review were used to identify patients with SETs from January 1992 to March 2017, with abstraction of basic demographics, surveillance intervals, and mortality. Patients were contacted to complete a phone questionnaire to assess follow-up as well as the validated Brief Esophageal Dysphagia Questionnaire (BEDQ).

RESULTS

Eighty-four leiomyomas, 5 leiomyosarcomas, 13 GI stromal tumors (GISTs), and 4 granular cell tumors were found. Among patients with leiomyomas, 58 (69%) were resected and 26 (31%) were followed under surveillance. Resected esophageal leiomyomas were larger than those under surveillance (49.7 mm vs 17.9 mm, P < .003). Esophageal leiomyoma growth during surveillance was only .5 mm over a mean 70-month follow-up (range, 4-288). No malignant transformation was seen, and only 2 patients required subsequent resection. The phone survey was completed by 35 patients and revealed minimal symptom burden, with only 2 patients (15%) under esophageal leiomyoma surveillance reporting symptoms (BEDQ score ≥10) over a mean 96.7-month follow-up. EUS had high diagnostic accuracy with a positive predictive value of 68% and a negative predictive value of 100% for leiomyomas or GISTs compared with surgical pathology.

CONCLUSIONS

EUS demonstrated high diagnostic accuracy in resected SETs. Esophageal leiomyomas demonstrate minimal growth or symptomatic progression. Therefore, long-term EUS surveillance of small asymptomatic esophageal leiomyomas may be unnecessary.

Extra-mucosal enucleation is still a safe and feasible treatment option of giant esophageal leiomyomas

Esophageal leiomyoma (EL) is rare but still the most common benign tumor of the esophagus. Extra-mucosal enucleation (EME) is the treatment of choice. Many recent reports have favored esophageal resection for giant Esophageal leiomyomas (ELs). The consequence of esophageal resection is well known and it would be radical to consider it as a preferred treatment for giant EL since most of them are still benign. We share case series of five giant ELs managed by EME, avoiding a mucosal breach and hence avoiding potentially morbid esophageal resections.

The role of EUS and EUS-guided FNA in the management of subepithelial lesions of the esophagus: A large, single-center experience

BACKGROUND

Referral for endosonographic evaluation of subepithelial lesions seen in the gastrointestinal (GI) tract is fairly common. Although rarely studied separately in details, esophageal lesions have some unique differences from other GI sites and might deserve some special considerations regarding follow-up and management.

MATERIALS AND METHODS

All cases referred for endoscopic ultrasound (EUS) evaluation of subepithelial esophageal lesions at Bezmialem University Hospital, a tertiary center in Istanbul, Turkey were retrospectively reviewed. Data were collected for patient and lesion characteristics as well as for pathology results and follow-up if available. Lesions were subcategorized according to their size, location, and final diagnosis.

RESULTS

A total of 164 EUS examinations were identified. In 22.5% of cases, the lesion could not be identified by EUS. Of the remaining cases, 57.6% had a lesion larger than 1 cm in size. Extramural compression was the diagnosis in 12% and leiomyoma in around 60%. Thirteen patients had follow-up examinations with only two showing an increase in size after 12 months. Sixty-five EUS-guided fine needle aspirations (EUS-guided FNAs) were performed, with around 50% having nondiagnostic samples and 94% of the remaining samples confirming the presumptive diagnosis.

CONCLUSIONS

The majority of subepithelial lesions in the esophagus are benign with extremely low malignancy potential. EUS examinations performed for lesions smaller than 2 cm as well as FNAs taken from lesions smaller than 3 cm might have minimal impact on their ultimate management and outcome. More than one FNA pass should be attempted in order to improve the yield.

The Role of Endoscopic Ultrasound in the Diagnostic Assessment of Subepithelial Lesions of the Upper Gastrointestinal Tract

Introduction

The identification of subepithelial lesions is a relatively frequent finding at endoscopy however their natural history is not well known. Our aim was to analyze the role of endoscopic ultrasound (EUS) in the diagnostic approach of subepithelial lesions of the upper gastrointestinal tract.

Methods

Retrospective study which included 324 patients undergoing upper radial EUS for evaluation of subepithelial lesions from 2008 to 2014. The EUS features, presumptive diagnosis and management decision were analyzed.

Results

324 patients included, 60% with gastric subepithelial lesions, 28% oesophageal and 12% from the duodenum. Based on EUS features the presumptive diagnosis was: 25% gastrointestinal stromal tumor, 21% lipoma, 19% leiomyoma, 17% pancreatic rest, 7% submucosa cysts, 1% granular cell tumors, 1% carcinoids, 1% mucosa lesions and 8% not defined. After EUS the suggested approach was no follow-up in 45%, follow-up with re-examination with EUS in 35% and additional tissue sampling or endoscopic/surgical resection in 20%. The latter was based on EUS features of risk at the diagnosis (53%), such as size ≥2 cm, hypoechogenicity, heterogeneity, lobulation, calcifications, cystic component and regional adenopathies; impossibility to define a presumptive diagnosis (39%) or EUS features change at follow-up (8%). The combination of multiple features correlated with a higher probability of this recommended strategy (p < 0.001), in 100% when 4 or more features were present. Among the 33 patients who underwent fine needle aspiration, in 66% the result was inconclusive. During follow-up, none of the patients who were managed with surveillance radial EUS presented complications.

Conclusion

EUS is the method of choice in the study of subepithelial lesions of the upper gastrointestinal tract, in most cases defining a diagnosis. The need for a definitive diagnosis or therapeutic approaches can be based on ultrasound risk features, presented, in the majority, at presentation. This study shows that EUS is capable of safely and accurately define those subepithelial lesions that can be managed only with surveillance ultrasound while waiting for better results with fine needle aspiration.