Application of artificial intelligence to digital-rapid on-site cytopathology evaluation during endoscopic ultrasound-guided fine needle aspiration: A proof-of-concept study

Tumor specific protein 70 targeted tumor cell isolation technology can improve the accuracy of cytopathological examination

OBJECTIVES

Although existing cytopathological examination is considered essential for the diagnosis of malignant serous effusions, its accuracy is pretty low. Tumor specific protein 70 (SP70), which is highly expressed on human tumor cell membrane, was identified in our previous study. This study aimed to explore whether SP70 targeted tumor cell isolation technology with immunomagnetic beads can improve the accuracy of cytopathological examination.

METHODS

Cytopathological analysis with SP70 targeted tumor cell isolation technology was used in this study. In total, 255 cases were enrolled. Serous effusions were analyzed by both existing cytopathological examination and the new cytopathological analysis concurrently.

RESULTS

The sensitivities of existing cytopathological examination and the new cytopathological analysis were 51.26 % and 85.43 %, respectively, while the specificities were 100 % for both. This new cytopathological analysis demonstrated a higher interobserver agreement with malignant diagnosis than the existing cytopathological examination (kappa coefficient: 0.720 vs. 0.316, p<0.001). In addition, it achieved superior diagnostic efficacy for malignancy differentiation compared to existing cytopathological examination (AUC: 0.927 vs. 0.756, p<0.001). The follow-up results showed that 74 malignant cases with final clinical diagnosis were positive only with the new cytopathological analysis. Among these cases, there were 58 negative and 16 atypical by the existing cytopathological examination. In these malignant cases, 74.3 % (55/74) had been confirmed to have serosa metastasis based on radiographic evidence, and 73.7 % (28/38) harbored tumor hotspot mutations.

CONCLUSIONS

As illustrated in this work, cytopathological analysis with SP70 targeted tumor cell isolation technology can improve the accuracy of existing cytopathological examination prominently.

Diagnostic accuracy of rapid on-site evaluation in subtyping lung cancer via bronchoscopic biopsy

Background

Rapid on-site evaluation (ROSE) is a valuable technique for ensuring the adequacy of specimens during bronchoscopic biopsy; however, its diagnostic utility in lung cancer pathological classification has yet to be comprehensively assessed.

Objective

To evaluate the diagnostic utility of ROSE in lung cancer and its accuracy in classifying lung cancer pathological types.

Methods

A retrospective analysis was performed on 510 consecutive patients who underwent bronchoscopic biopsy with concurrent ROSE between March and July 2023. ROSE diagnoses were compared with the final pathological diagnoses to access concordance. Sensitivity analyses were conducted to evaluate concordance across cancer subtypes, lesion locations, and patient demographics. The diagnostic accuracy of ROSE in classifying lung cancer subtypes-specifically small cell lung cancer (SCLC), non-small cell lung cancer (NSCLC), squamous cell carcinoma (SCC), and adenocarcinoma (AC)-was systematically evaluated.

Results

Overall concordance between ROSE and the final pathological diagnoses was 93.92% (479/510), with near-perfect agreement (k = 0.87, 95% CI: 0.83-0.92). The accuracy of ROSE in distinguishing malignant from benign lesions was significantly lower in central lesions (89.05%) compared to peripheral lesions (95.66%; p = 0.010), and in AC (89.91%) versus SCC (100%; p = 0.0027). ROSE showed high accuracy, sensitivity, specificity, negative predictive value (NPV), and positive predictive value (PPV) for distinguishing SCLC (95.32%, 87.50%, 97.30%, 96.86%, and 89.09%) and NSCLC (92.45%, 92.34%, 92.86%, 75.36%, and 98.09%). For SCC and AC, they were 84.91%, 89.32%, 80.73%, 88.89%, and 81.42% vs 79.72%, 69.72%, 90.29%, 73.81%, and 88.37%, respectively.

Conclusion

ROSE effectively differentiates benign from malignant lesions and accurately classifies SCLC and NSCLC during bronchoscopic biopsy. While useful for preliminary subtyping of SCC and AC, its reduced sensitivity for AC and challenges in central lesion evaluation limit its utility as a standalone diagnostic tool. ROSE remains critical for optimizing biopsy workflows and reducing repeat procedures.

Endoscopic ultrasound-guided tissue sampling: European Society of Gastrointestinal Endoscopy (ESGE) Technical and Technology Review

This Technical and Technology Review from the European Society of Gastrointestinal Endoscopy (ESGE) represents an update of the previous document on the technical aspects of endoscopic ultrasound (EUS)-guided sampling in gastroenterology, including the available types of needle, technical aspects of tissue sampling, new devices, and specimen handling and processing. Among the most important new recommendations are:ESGE recommends end-cutting fine-needle biopsy (FNB) needles over reverse-bevel FNB or fine-needle aspiration (FNA) needles for tissue sampling of solid pancreatic lesions; FNA may still have a role when rapid on-site evaluation (ROSE) is available.ESGE recommends EUS-FNB or mucosal incision-assisted biopsy (MIAB) equally for tissue sampling of subepithelial lesions ≥20 mm in size. MIAB could represent the first choice for smaller lesions (<20 mm) if proper expertise is available.ESGE does not recommend the use of antibiotic prophylaxis before EUS-guided tissue sampling of solid masses and EUS-FNA of pancreatic cystic lesions.

Updated techniques and evidence for endoscopic ultrasound-guided tissue acquisition from solid pancreatic lesions

Endoscopic ultrasound-guided tissue acquisition (EUS-TA), including fine-needle aspiration (EUS-FNA) and fine-needle biopsy (EUS-FNB), has revolutionized specimen collection from intra-abdominal organs, especially the pancreas. Advances in personalized medicine and more precise treatment have increased demands to collect specimens with higher cell counts, while preserving tissue structure, leading to the development of EUS-FNB needles. EUS-FNB has generally replaced EUS-FNA as the procedure of choice for EUS-TA of pancreatic cancer. Various techniques have been tested for their ability to enhance the diagnostic performance of EUS-TA, including multiple methods of sampling at the time of puncture, on-site specimen evaluation, and specimen processing. In addition, advances in next-generation sequencing have made comprehensive genomic profiling of EUS-TA samples feasible in routine clinical practice. The present review describes updates in EUS-TA sampling techniques of pancreatic lesions, as well as methods for their evaluation.

Role of Artificial Intelligence for Endoscopic Ultrasound

Endoscopic ultrasound (EUS) is widely used for the diagnosis of biliopancreatic and gastrointestinal tract diseases, but it is one of the most operator-dependent endoscopic techniques, requiring a long and complex learning curve. The role of artificial intelligence (AI) in EUS is growing as AI algorithms can assist in lesion detection and characterization by analyzing EUS images. Deep learning (DL) techniques, such as convolutional neural networks, have shown great potential for tumor identification; the application of AI models can increase the EUS diagnostic accuracy, provide faster diagnoses, and provide more information that can be helpful also for a training program.

Facing the challenges of autoimmune pancreatitis diagnosis: The answer from artificial intelligence

Current diagnosis of autoimmune pancreatitis (AIP) is challenging and often requires combining multiple dimensions. There is a need to explore new methods for diagnosing AIP. The development of artificial intelligence (AI) is evident, and it is believed to have potential in the clinical diagnosis of AIP. This article aims to list the current diagnostic difficulties of AIP, describe existing AI applications, and suggest directions for future AI usages in AIP diagnosis.

A semi-supervised convolutional neural network for diagnosis of pancreatic ductal adenocarcinoma based on EUS-FNA cytological images

BACKGROUND

The cytological diagnostic process of EUS-FNA smears is time-consuming and manpower-intensive, and the conclusion could be subjective and controversial. Moreover, the relative lack of cytopathologists has limited the widespread implementation of Rapid on-site evaluation (ROSE) presently. Therefore, this study aimed to establish an AI system for the detection of pancreatic ductal adenocarcinoma (PDAC) based on EUS-FNA cytological images.

METHODS

We collected 3213 unified magnification images of pancreatic cell clusters from 210 pancreatic mass patients who underwent EUS-FNA in four hospitals. A semi-supervised CNN (SSCNN) system was developed to distinguish PDAC from Non-PDAC. The internal and external verifications were adopted and the diagnostic accuracy was compared between different seniorities of cytopathologists. 33 images of "Atypical" diagnosed by expert cytopathologists were selected to analyze the consistency between the system and definitive diagnosis.

RESULTS

The segmentation indicators Mean Intersection over Union (mIou), precision, recall and F1-score of SSCNN in internal and external testing sets were 88.3%, 93.21%,94.24%, 93.68% and 87.75%, 93.81%, 93.14%, 93.48% successively. The PDAC classification indicators of the SSCNN model including area under the ROC curve (AUC), accuracy, sensitivity, specificity, PPV and NPV in the internal testing set were 0.97%, 0.95%, 0.94%, 0.97%, 0.98%, 0.91% respectively, and 0.99%, 0.94%, 0.94%, 0.95%, 0.99%, 0.75% correspondingly in the external testing set. The diagnostic accuracy of senior, intermediate and junior cytopathologists was 95.00%, 88.33% and 76.67% under the binary diagnostic criteria of PDAC and non-PDAC. In comparison, the accuracy of the SSCNN system was 90.00% in the dataset of man-machine competition. The accuracy of the SSCNN model was highly consistent with senior cytopathologists (Kappa = 0.853, P = 0.001). The accuracy, sensitivity and specificity of the system in the classification of "atypical" cases were 78.79%, 84.20% and 71.43% respectively.

CONCLUSION

Not merely tremendous preparatory work was drastically reduced, the semi-supervised CNN model could effectively identify PDAC cell clusters in EUS-FNA cytological smears which achieved analogically diagnostic capability compared with senior cytopathologists, and showed outstanding performance in assisting to categorize "atypical" cases where manual diagnosis is controversial.

TRIAL REGISTRATION

This study was registered on clinicaltrials.gov, and its unique Protocol ID was PJ-2018-12-17.

Challenges and Opportunities in Cytopathology Artificial Intelligence

Artificial Intelligence (AI) has the potential to revolutionize cytopathology by enhancing diagnostic accuracy, efficiency, and accessibility. However, the implementation of AI in this field presents significant challenges and opportunities. This review paper explores the current landscape of AI applications in cytopathology, highlighting the critical challenges, including data quality and availability, algorithm development, integration and standardization, and clinical validation. We discuss challenges such as the limitation of only one optical section and z-stack scanning, the complexities associated with acquiring high-quality labeled data, the intricacies of developing robust and generalizable AI models, and the difficulties in integrating AI tools into existing laboratory workflows. The review also identifies substantial opportunities that AI brings to cytopathology. These include the potential for improved diagnostic accuracy through enhanced detection capabilities and consistent, reproducible results, which can reduce observer variability. AI-driven automation of routine tasks can significantly increase efficiency, allowing cytopathologists to focus on more complex analyses. Furthermore, AI can serve as a valuable educational tool, augmenting the training of cytopathologists and facilitating global health initiatives by making high-quality diagnostics accessible in resource-limited settings. The review underscores the importance of addressing these challenges to harness the full potential of AI in cytopathology, ultimately improving patient care and outcomes.

Artificial Intelligence in Pancreatic Imaging: A Systematic Review

The rising incidence of pancreatic diseases, including acute and chronic pancreatitis and various pancreatic neoplasms, poses a significant global health challenge. Pancreatic ductal adenocarcinoma (PDAC) for example, has a high mortality rate due to late-stage diagnosis and its inaccessible location. Advances in imaging technologies, though improving diagnostic capabilities, still necessitate biopsy confirmation. Artificial intelligence, particularly machine learning and deep learning, has emerged as a revolutionary force in healthcare, enhancing diagnostic precision and personalizing treatment. This narrative review explores Artificial intelligence's role in pancreatic imaging, its technological advancements, clinical applications, and associated challenges. Following the PRISMA-DTA guidelines, a comprehensive search of databases including PubMed, Scopus, and Cochrane Library was conducted, focusing on Artificial intelligence, machine learning, deep learning, and radiomics in pancreatic imaging. Articles involving human subjects, written in English, and published up to March 31, 2024, were included. The review process involved title and abstract screening, followed by full-text review and refinement based on relevance and novelty. Recent Artificial intelligence advancements have shown promise in detecting and diagnosing pancreatic diseases. Deep learning techniques, particularly convolutional neural networks (CNNs), have been effective in detecting and segmenting pancreatic tissues as well as differentiating between benign and malignant lesions. Deep learning algorithms have also been used to predict survival time, recurrence risk, and therapy response in pancreatic cancer patients. Radiomics approaches, extracting quantitative features from imaging modalities such as CT, MRI, and endoscopic ultrasound, have enhanced the accuracy of these deep learning models. Despite the potential of Artificial intelligence in pancreatic imaging, challenges such as legal and ethical considerations, algorithm transparency, and data security remain. This review underscores the transformative potential of Artificial intelligence in enhancing the diagnosis and treatment of pancreatic diseases, ultimately aiming to improve patient outcomes and survival rates.

Endoscopic Ultrasound-Guided Pancreatic Tissue Sampling: Lesion Assessment, Needles, and Techniques

Endoscopic ultrasound (EUS)-guided tissue sampling includes the techniques of fine needle aspiration (FNA) and fine needle biopsy (FNB), and both procedures have revolutionized specimen collection from the gastrointestinal tract, especially from remote/inaccessible organs. EUS-FNB has replaced FNA as the procedure of choice for tissue acquisition in solid pancreatic lesions (SPLs) across various society guidelines. FNB specimens provide a larger histological tissue core (preserving tissue architecture) with fewer needle passes, and this is extremely relevant in today's era of precision and personalized molecular medicine. Innovations in needle tip design are constantly under development to maximize diagnostic accuracy by enhancing histological sampling capabilities. But, apart from the basic framework of the needle, various other factors play a role that influence diagnostic outcomes, namely, sampling techniques (fanning, aspiration or suction, and number of passes), collection methods, on-site evaluation (rapid, macroscopic, or visual), and specimen processing. The choice taken depends strongly on the endoscopist's preference, available resources at the disposal, and procedure objectives. Hence, in this review, we explicate in detail the concepts and available literature at our disposal on the topic of EUS-guided pancreatic tissue sampling to best guide any practicing gastroenterologist/endoscopist in a not-to-ideal set-up, which EUS-guided tissue acquisition technique is the "best" for their case to augment their diagnostic outcomes.

Advancements in biliopancreatic endoscopy - A comprehensive review of artificial intelligence in EUS and ERCP

The development and implementation of artificial intelligence (AI), particularly deep learning (DL) models, has generated significant interest across various fields of gastroenterology. While research in luminal endoscopy has seen rapid translation to clinical practice with approved AI devices, its potential extends far beyond, offering promising benefits for biliopancreatic endoscopy like optical characterization of strictures during cholangioscopy or detection and classification of pancreatic lesions during diagnostic endoscopic ultrasound (EUS). This narrative review provides an up-to-date of the latest literature and available studies in this field. Serving as a comprehensive guide to the current landscape of AI in biliopancreatic endoscopy, emphasizing technological advancements, main applications, ethical considerations, and future directions for research and clinical implementation.

Viability of Whole-Slide Imaging for Intraoperative Touch Imprint Cytological Diagnosis of Axillary Sentinel Lymph Nodes in Breast Cancer Patients

BACKGROUND

Whole-slide imaging (WSI) is a promising tool in pathology. However, the use of WSI in cytopathology has lagged behind that in histology. We aimed to evaluate the utility of WSI for the intraoperative touch imprint cytological diagnosis of axillary sentinel lymph nodes (SLNs) in breast cancer patients.

METHODS

Glass slides from touch imprint cytology of 480 axillary SLNs were scanned using two different WSI scanners. The intra- and interobserver concordance, accuracy, possible reasons for misdiagnosis, scanning time, and review time for three cytopathologists were compared between WSI and light microscopy (LM).

RESULTS

A total of 4320 diagnoses were obtained. There was substantial to strong intraobserver concordance when comparing reads among paired LM slides and WSI digital slides (κ coefficient ranged from 0.63 to 0.88, and concordance rates ranged from 94.58% to 98.33%). Substantial to strong interobserver agreement was also observed among the three cytopathologists (κ coefficient ranged from 0.67 to 0.85, and concordance rates ranged from 95.42% to 97.92%). The accuracy of LM was slightly higher (average of 98.06%) than that of WSI (averages of 96.81% and 97.78%). The majority of misdiagnoses were false negative diagnoses due to the following top three causes: few cancer cells, confusing cancer cells with histiocytes, and confusing cancer cells with lymphocytes.

CONCLUSIONS

This study is the first to address the feasibility of WSI in touch imprint cytology. The use of WSI for intraoperative touch imprint cytological diagnosis of SLNs is a practical option when experienced staff are not available on-site.

The Evolving Role of Artificial Intelligence in Gastrointestinal Histopathology: An Update

Significant advances in artificial intelligence (AI) over the past decade potentially may lead to dramatic effects on clinical practice. Digitized histology represents an area ripe for AI implementation. We describe several current needs within the world of gastrointestinal histopathology, and outline, using currently studied models, how AI potentially can address them. We also highlight pitfalls as AI makes inroads into clinical practice.

Advances and efficacy in specimen handling for endoscopic ultrasound-guided fine needle aspiration and biopsy: A comprehensive review

Endoscopic ultrasound-guided fine needle aspiration and biopsy have significantly evolved since they offer a minimally invasive approach for obtaining pathological specimens from lesions adjacent to or within the intestine. This paper reviews advancements in endoscopic ultrasound-guided fine needle aspiration and biopsy techniques and devices, emphasizing the importance of handling specimens for diagnostic accuracy. Innovations of fine needle biopsy needles with features like side holes and Franseen shapes have enhanced histological sampling capabilities. Techniques for specimen handling, including rapid on-site evaluation and macroscopic on-site evaluation, play pivotal roles in assessing sample adequacy, thereby influencing diagnostic outcomes. The utility of artificial intelligence in augmenting rapid on-site evaluation and macroscopic on-site evaluation, although still in experimental stages, presents a promising avenue for improving procedural efficiency and diagnostic precision. The choice of specimen handling technique is dependent on various factors including endoscopist preference, procedure objectives, and available resources, underscoring the need for a comprehensive understanding of each method's characteristics to optimize diagnostic efficacy and procedural safety.

Cutting edge of endoscopic ultrasound-guided fine-needle aspiration for solid pancreatic lesions

This article provides an extensive review of the advancements and future perspectives related to endoscopic ultrasound-guided tissue acquisition (EUS-TA) for the diagnosis of solid pancreatic lesions (SPLs). EUS-TA, including fine-needle aspiration (EUS-FNA) and fine-needle biopsy (EUS-FNB), has revolutionized the collection of specimens from intra-abdominal organs, including the pancreas. Improvements in the design of needles, collection methods, and specimen processing techniques have improved the diagnostic performance. This review highlights the latest findings regarding needle evolution, actuation number, sampling methods, specimen evaluation techniques, application of artificial intelligence (AI) for diagnostic purposes, and use of comprehensive genomic profiling (CGP). It acknowledges the rising use of Franseen and fork-tip needles for EUS-FNB and emphasizes that the optimal number of actuations requires further study. Methods such as the door-knocking and fanning techniques have shown promise for increasing diagnostic performance. Macroscopic on-site evaluation (MOSE) is presented as a practical rapid specimen evaluation method, and the integration of AI is identified as a potentially impactful development. The study also underscores the importance of optimal sampling for CGP, which can enhance the precision of cancer treatment. Ongoing research and technological innovations will further improve the accuracy and efficacy of EUS-TA.

The application of artificial intelligence for Rapid On-Site Evaluation during flexible bronchoscopy

Background

Rapid On-Site Evaluation (ROSE) during flexible bronchoscopy (FB) can improve the adequacy of biopsy specimens and diagnostic yield of lung cancer. However, the lack of cytopathologists has restricted the wide use of ROSE.

Objective

To develop a ROSE artificial intelligence (AI) system using deep learning techniques to differentiate malignant from benign lesions based on ROSE cytological images, and evaluate the clinical performance of the ROSE AI system.

Method

6357 ROSE cytological images from 721 patients who underwent transbronchial biopsy were collected from January to July 2023 at the Tangdu Hospital, Air Force Medical University. A ROSE AI system, composed of a deep convolutional neural network (DCNN), was developed to identify whether there were malignant cells in the ROSE cytological images. Internal testing, external testing, and human-machine competition were used to evaluate the performance of the system.

Results

The ROSE AI system identified images containing lung malignant cells with the accuracy of 92.97% and 90.26% on the internal testing dataset and external testing dataset respectively, and its performance was comparable to that of the experienced cytopathologist. The ROSE AI system also showed promising performance in diagnosing lung cancer based on ROSE cytological images, with accuracy of 89.61% and 87.59%, and sensitivity of 90.57% and 94.90% on the internal testing dataset and external testing dataset respectively. More specifically, the agreement between the ROSE AI system and the experienced cytopathologist in diagnosing common types of lung cancer, including squamous cell carcinoma, adenocarcinoma, and small cell lung cancer, demonstrated almost perfect consistency in both the internal testing dataset (κ  =   0.930 ) and the external testing dataset (κ  =   0.932 ).

Conclusions

The ROSE AI system demonstrated feasibility and robustness in identifying specimen adequacy, showing potential enhancement in the diagnostic yield of FB. Nevertheless, additional enhancements, incorporating a more diverse range of training data and leveraging advanced AI models with increased capabilities, along with rigorous validation through extensive multi-center randomized control assays, are crucial to guarantee the seamless and effective integration of this technology into clinical practice.

American Society of Cytopathology Telecytology validation recommendations for rapid on-site evaluation (ROSE)

Telecytology has multiple applications, including rapid onsite evaluation (ROSE) of fine-needle aspiration (FNA) specimens. It can enhance cytopathology practice by increasing productivity, reducing costs, and providing subspecialty expertise in areas with limited access to a cytopathologist. However, there are currently no specific validation guidelines to ensure safe practice and compliance with regulations. This initiative, promoted by the American Society of Cytopathology (ASC), intends to propose recommendations for telecytology implementation. These recommendations propose that the validation process should include testing of all hardware and software, both separately and as a whole; training of all individuals who will participate in telecytology with regular competency evaluations; a structured approach using retrospective slides with defined diagnoses for validation and prospective cases for verification and quality assurance. Telecytology processes must be integrated into the laboratory's quality management system and benchmarks for discrepancy rates between preliminary and final diagnoses should be established and monitored. Special attention should be paid to minimize discrepancies that downgrade malignant cases to benign (false positive on telecytology). Currently, billing and reimbursement codes for telecytology are not yet available. Once, they are, recommendation of the appropriate usage of these codes would be a part of the recommendations. These proposed guidelines are intended to be a resource for laboratories that are considering implementing telecytology. These recommendations can help to ensure the safe and effective use of telecytology and maximize its benefits for patients.

Digital cytology part 2: artificial intelligence in cytology: a concept paper with review and recommendations from the American Society of Cytopathology Digital Cytology Task Force

Digital cytology and artificial intelligence (AI) are gaining greater adoption in the cytology laboratory. However, peer-reviewed real-world data and literature are lacking in regard to the current clinical landscape. The American Society of Cytopathology in conjunction with the International Academy of Cytology and the Digital Pathology Association established a special task force comprising 20 members with expertise and/or interest in digital cytology. The aim of the group was to investigate the feasibility of incorporating digital cytology, specifically cytology whole slide scanning and AI applications, into the workflow of the laboratory. In turn, the impact on cytopathologists, cytologists (cytotechnologists), and cytology departments were also assessed. The task force reviewed existing literature on digital cytology, conducted a worldwide survey, and held a virtual roundtable discussion on digital cytology and AI with multiple industry corporate representatives. This white paper, presented in 2 parts, summarizes the current state of digital cytology and AI practice in global cytology practice. Part 1 of the white paper is presented as a separate paper which details a review and best practice recommendations for incorporating digital cytology into practice. Part 2 of the white paper presented here provides a comprehensive review of AI in cytology practice along with best practice recommendations and legal considerations. Additionally, the cytology global survey results highlighting current AI practices by various laboratories, as well as current attitudes, are reported.

Artificial Intelligence in Endoscopic Ultrasonography-Guided Fine-Needle Aspiration/Biopsy (EUS-FNA/B) for Solid Pancreatic Lesions: Opportunities and Challenges

Solid pancreatic lesions (SPLs) encompass a variety of benign and malignant diseases and accurate diagnosis is crucial for guiding appropriate treatment decisions. Endoscopic ultrasonography-guided fine-needle aspiration/biopsy (EUS-FNA/B) serves as a front-line diagnostic tool for pancreatic mass lesions and is widely used in clinical practice. Artificial intelligence (AI) is a mathematical technique that automates the learning and recognition of data patterns. Its strong self-learning ability and unbiased nature have led to its gradual adoption in the medical field. In this paper, we describe the fundamentals of AI and provide a summary of reports on AI in EUS-FNA/B to help endoscopists understand and realize its potential in improving pathological diagnosis and guiding targeted EUS-FNA/B. However, AI models have limitations and shortages that need to be addressed before clinical use. Furthermore, as most AI studies are retrospective, large-scale prospective clinical trials are necessary to evaluate their clinical usefulness accurately. Although AI in EUS-FNA/B is still in its infancy, the constant input of clinical data and the advancements in computer technology are expected to make computer-aided diagnosis and treatment more feasible.

A deep learning model using hyperspectral image for EUS-FNA cytology diagnosis in pancreatic ductal adenocarcinoma

BACKGROUND AND AIMS

Endoscopic ultrasonography-guided fine-needle aspiration/biopsy (EUS-FNA/B) is considered to be a first-line procedure for the pathological diagnosis of pancreatic cancer owing to its high accuracy and low complication rate. The number of new cases of pancreatic ductal adenocarcinoma (PDAC) is increasing, and its accurate pathological diagnosis poses a challenge for cytopathologists. Our aim was to develop a hyperspectral imaging (HSI)-based convolution neural network (CNN) algorithm to aid in the diagnosis of pancreatic EUS-FNA cytology specimens.

METHODS

HSI images were captured of pancreatic EUS-FNA cytological specimens from benign pancreatic tissues (n = 33) and PDAC (n = 39) prepared using a liquid-based cytology method. A CNN was established to test the diagnostic performance, and Attribution Guided Factorization Visualization (AGF-Visualization) was used to visualize the regions of important classification features identified by the model.

RESULTS

A total of 1913 HSI images were obtained. Our ResNet18-SimSiam model achieved an accuracy of 0.9204, sensitivity of 0.9310 and specificity of 0.9123 (area under the curve of 0.9625) when trained on HSI images for the differentiation of PDAC cytological specimens from benign pancreatic cells. AGF-Visualization confirmed that the diagnoses were based on the features of tumor cell nuclei.

CONCLUSIONS

An HSI-based model was developed to diagnose cytological PDAC specimens obtained using EUS-guided sampling. Under the supervision of experienced cytopathologists, we performed multi-staged consecutive in-depth learning of the model. Its superior diagnostic performance could be of value for cytologists when diagnosing PDAC.