Feasibility and Accuracy of Artificial Intelligence-Assisted Sponge Cytology for Community-Based Esophageal Squamous Cell Carcinoma Screening in China

Artificial intelligence enhances the management of esophageal squamous cell carcinoma in the precision oncology era

Esophageal squamous cell carcinoma (ESCC) is the most common histological type of esophageal cancer with a poor prognosis. Early diagnosis and prognosis assessment are crucial for improving the survival rate of ESCC patients. With the advancement of artificial intelligence (AI) technology and the proliferation of medical digital information, AI has demonstrated promising sensitivity and accuracy in assisting precise detection, treatment decision-making, and prognosis assessment of ESCC. It has become a unique opportunity to enhance comprehensive clinical management of ESCC in the era of precision oncology. This review examines how AI is applied to the diagnosis, treatment, and prognosis assessment of ESCC in the era of precision oncology, and analyzes the challenges and potential opportunities that AI faces in clinical translation. Through insights into future prospects, it is hoped that this review will contribute to the real-world application of AI in future clinical settings, ultimately alleviating the disease burden caused by ESCC.

Burden of digestive system diseases in China and its provinces during 1990-2019: Results of the 2019 Global Disease Burden Study

BACKGROUND

Evaluating the impact of digestive system diseases is vital for devising effective prevention strategies. However, comprehensive reports on the burden of digestive system diseases in China are lacking. Our study aimed to provide an overview of the burden and trends of digestive system diseases from 1990 to 2019 in China and its provinces.

METHODS

This cross-sectional study utilized the Global Disease Burden Study 2019 to estimate the incidence, mortality rate, disability-adjusted life years (DALYs), years of life disability, years of life lost, and changes in the burden of digestive diseases across 31 Chinese provinces from 1990 to 2019. The analysis of disease burden primarily examines the characteristics of sub-disease distribution, time trends, age distribution, and sex distribution. Additionally, we compared provincial age-standardized DALYs for digestive diseases with the expected rates based on the socio-demographic index (SDI).

RESULTS

In 2019, there were 499.2 million cases of digestive system diseases in China, resulting in 1,557,310 deaths. Stomach cancer, colon and rectal cancer, and esophageal cancer are the top three diseases associated with mortality and DALY related to digestive system diseases. Meanwhile, cirrhosis and other chronic liver diseases, gastroesophageal reflux disease, and gallbladder and biliary diseases are the top three kinds of diseases with the highest prevalence among digestive system diseases. The risk of gastric cancer sharply increases among men after the age of 40 years, leading to a significant disparity in burden between men and women. As the SDI increased, the DALYs associated with digestive system diseases in China and its provinces showed a downward trend.

CONCLUSION

Our study highlights the inverse correlation between DALYs associated with digestive system diseases and the SDI, providing valuable insights that can assist public health officials in the estimation of the disease burden in this area.

Risk and pathologic factors of recurrence after endoscopic resection for superficial esophageal squamous cell carcinoma: a systematic review and meta-analysis

BACKGROUND AND AIMS

The risk and pathologic factors of recurrence after endoscopic resection (ER) for superficial esophageal squamous cell carcinoma (ESCC) are inconsistent across studies. We evaluated this in a systematic review and meta-analysis.

METHODS

The data of recurrence in such patients were extracted from all studies. Risk ratios (RRs) were combined using random-effects meta-analysis to assess pooled recurrence rate and pathologic risk factors. Relapse-free survival was combined using the Kaplan-Meier method to estimate the relationship between various pathologic factors and recurrence time.

RESULTS

We identified 26 studies, with a total of 5100 patients and 321 with recurrences (pooled rate, 6.2%). The risk of recurrence was significantly higher in positive vertical margin (RR, 4.51; 95% confidence interval [CI], 2.16-9.44), positive horizontal margin (RR, 2.54; 95% CI, 1.57-4.13), lymphovascular invasion (LVI) (RR, 2.33; 95% CI, 1.75-3.11; P < .001), lymphatic invasion (LI) (RR, 2.24; 95% CI, 1.24-4.06), and tumor invading submucosa of ≤200 μm (SM1) (RR, 1.71; 95% CI, 1.32-2.21, compared to muscularis mucosa). Patients with LI (hazard ratio, 2.47; 95% CI, 1.24-4.90; P = .02) and LVI (HR, 2.36; 95% CI, 2.22-4.59; P = .0006) tended to have earlier recurrence after ER.

CONCLUSIONS

The recurrence rate of superficial esophageal squamous cell carcinoma after ER is acceptable. Patients with positive margins, LVI, LI, and SM1 need to pay significant attention to the risk of recurrence. LI and VI should be evaluated separately. (PROSPERO CRD42023406309.).

Genome-wide methylation profiling identified methylated KCNA3 and OTOP2 as promising diagnostic markers for esophageal squamous cell carcinoma

BACKGROUND

Early detection of esophageal squamous cell carcinoma (ESCC) can considerably improve the prognosis of patients. Aberrant cell-free DNA (cfDNA) methylation signatures are a promising tool for detecting ESCC. However, available markers based on cell-free DNA methylation are still inadequate. This study aimed to identify ESCC-specific cfDNA methylation markers and evaluate the diagnostic performance in the early detection of ESCC.

METHODS

We performed whole-genome bisulfite sequencing (WGBS) for 24 ESCC tissues and their normal adjacent tissues. Based on the WGBS data, we identified 21,469,837 eligible CpG sites (CpGs). By integrating several methylation datasets, we identified several promising ESCC-specific cell-free DNA methylation markers. Finally, we developed a dual-marker panel based on methylated KCNA3 and OTOP2 , and then, we evaluated its performance in our training and validation cohorts.

RESULTS

The ESCC diagnostic model constructed based on KCNA3 and OTOP2 had an AUC of 0.91 [95% CI: 0.85-0.95], and an optimal sensitivity and specificity of 84.91% and 94.32%, respectively, in the training cohort. In the independent validation cohort, the AUC was 0.88 [95% CI: 0.83-0.92], along with an optimal sensitivity of 81.5% and specificity of 92.9%. The model sensitivity for stage I-II ESCC was 78.4%, which was slightly lower than the sensitivity of the model (85.7%) for stage III-IV ESCC.

CONCLUSION

The dual-target panel based on cfDNA showed excellent performance for detecting ESCC and might be an alternative strategy for screening ESCC.

Development and status quo of digestive endoscopy in China: An analysis based on the national census in 2013 and 2020

Background and Objectives

Technique and practice of digestive endoscopy are undergoing speedy development all over the world. This study aimed to evaluate its status quo and development in China.

Methods

All hospitals performing digestive endoscopy in mainland China participated in the national census in 2013 and 2020. Retrospective data of hospitals, endoscopists, volumes, and qualities were collected via an online structured questionnaire, and its accuracy and rationality were verified by logical tests and manual reviews. Data from other countries were used to compare with that of China.

Results

From 2012 to 2019, the number of hospitals performing digestive endoscopy increased from 6,128 to 7,470 (1.22-fold), in which primary healthcare played a minor role. The median hospitals per 100,000 inhabitants per provincial region increased from 0.49 (IQR, 0.39-0.57) to 0.55 (IQR, 0.49-0.63). The endoscopists increased from 26,203 to 39,638 (1.51-fold), but their average workload even expanded. Overall volume increased from 28.8 million to 44.5 million (1.55-fold), and most types of endoscopic procedures recorded a high growth rate. Contrastingly, the specific utilization rates were low and paled in comparison with some developed countries. Nationwide, regional utilization rates showed a significant correlation with GDP per capita (P <0.001). Overall qualities of digestive endoscopy were excellent, but certain results of quality indicators posed a huge challenge, such as the detection rates of adenoma and early cancers.

Conclusions

Impressive progress has been made in digestive endoscopy with rapidly expanding economy in China. However, primary healthcare, utilization rates, and income-related inequality of regional services were needed to be improved to promote public health better.

NMR and MS reveal characteristic metabolome atlas and optimize esophageal squamous cell carcinoma early detection

Metabolic changes precede malignant histology. However, it remains unclear whether detectable characteristic metabolome exists in esophageal squamous cell carcinoma (ESCC) tissues and biofluids for early diagnosis. Here, we conduct NMR- and MS-based metabolomics on 1,153 matched ESCC tissues, normal mucosae, pre- and one-week post-operative sera and urines from 560 participants across three hospitals, with machine learning and WGCNA. Aberrations in 'alanine, aspartate and glutamate metabolism' proved to be prevalent throughout the ESCC evolution, consistently identified by NMR and MS, and reflected in 16 serum and 10 urine metabolic signatures in both discovery and validation sets. NMR-based simplified panels of any five serum or urine metabolites outperform clinical serological tumor markers (AUC = 0.984 and 0.930, respectively), and are effective in distinguishing early-stage ESCC in test set (serum accuracy = 0.994, urine accuracy = 0.879). Collectively, NMR-based biofluid screening can reveal characteristic metabolic events of ESCC and be feasible for early detection (ChiCTR2300073613).

New US capsule endoscopy for superficial and submucosal imaging of the esophagus: the first-in-human study

BACKGROUND AND AIMS

EUS is essential in diagnosing and staging of esophageal subepithelial lesions and tumors. However, EUS is invasive, relies on highly trained endoscopists, and typically requires sedation. The newly developed US capsule endoscopy (USCE), which incorporates both white-light and US imaging modalities into a tethered capsule, is a minimally invasive method for obtaining superficial and submucosal information of the esophagus. This study aimed to assess the feasibility and safety of this USCE system.

METHODS

Twenty participants were enrolled: 10 healthy volunteers and 10 patients with esophageal lesions indicated for EUS. Participants first underwent USCE and subsequently EUS within 48 hours. The primary outcome was the technical success rate of USCE. Secondary outcomes were safety, visualization of the esophagus, and comfort assessment.

RESULTS

The technical success rate of USCE was 95% because 1 patient failed to swallow the capsule. No adverse events were observed. The esophagus was well visualized, and all lesions were detected under USCE optical mode in 19 participants. For healthy volunteers, the US images of normal esophageal walls were all characterized by differentiated 7-layer architecture under both USCE and EUS. For 9 patients, the features of esophageal lesions were recognized clearly under USCE, and presumptive diagnoses derived from USCE were all consistent with those from EUS. Most participants preferred USCE to EUS.

CONCLUSIONS

The novel USCE is feasible and safe to observe the esophageal mucosa and acquire submucosal information, which has the potential to be widely used in the clinic. (Clinical trial registration number: NCT05054933.).

Improving outcomes in patients with oesophageal cancer

The care of patients with oesophageal cancer or of individuals who have an elevated risk of oesophageal cancer has changed dramatically. The epidemiology of squamous cell and adenocarcinoma of the oesophagus has diverged over the past several decades, with a marked increase in incidence only for oesophageal adenocarcinoma. Only in the past decade, however, have molecular features that distinguish these two forms of the disease been identified. This advance has the potential to improve screening for oesophageal cancers through the development of novel minimally invasive diagnostic technologies predicated on cancer-specific genomic or epigenetic alterations. Surgical techniques have also evolved towards less invasive approaches associated with less morbidity, without compromising oncological outcomes. With improvements in multidisciplinary care, advances in radiotherapy and new tools to detect minimal residual disease, certain patients may no longer even require surgical tumour resection. However, perhaps the most anticipated advance in the treatment of patients with oesophageal cancer is the advent of immune-checkpoint inhibitors, which harness and enhance the host immune response against cancer. In this Review, we discuss all these advances in the management of oesophageal cancer, representing only the beginning of a transformation in our quest to improve patient outcomes.

Machine learning-based automated sponge cytology for screening of oesophageal squamous cell carcinoma and adenocarcinoma of the oesophagogastric junction: a nationwide, multicohort, prospective study

BACKGROUND

Oesophageal squamous cell carcinoma and adenocarcinoma of the oesophagogastric junction have a dismal prognosis, and early detection is key to reduce mortality. However, early detection depends on upper gastrointestinal endoscopy, which is not feasible to implement at a population level. We aimed to develop and validate a fully automated machine learning-based prediction tool integrating a minimally invasive sponge cytology test and epidemiological risk factors for screening of oesophageal squamous cell carcinoma and adenocarcinoma of the oesophagogastric junction before endoscopy.

METHODS

For this multicohort prospective study, we enrolled participants aged 40-75 years undergoing upper gastrointestinal endoscopy screening at 39 tertiary or secondary hospitals in China for model training and testing, and included community-based screening participants for further validation. All participants underwent questionnaire surveys, sponge cytology testing, and endoscopy in a sequential manner. We trained machine learning models to predict a composite outcome of high-grade lesions, defined as histology-confirmed high-grade intraepithelial neoplasia and carcinoma of the oesophagus and oesophagogastric junction. The predictive features included 105 cytological and 15 epidemiological features. Model performance was primarily measured with the area under the receiver operating characteristic curve (AUROC) and average precision. The performance measures for cytologists with AI assistance was also assessed.

FINDINGS

Between Jan 1, 2021, and June 30, 2022, 17 498 eligible participants were involved in model training and validation. In the testing set, the AUROC of the final model was 0·960 (95% CI 0·937 to 0·977) and the average precision was 0·482 (0·470 to 0·494). The model achieved similar performance to consensus of cytologists with AI assistance (AUROC 0·955 [95% CI 0·933 to 0·975]; p=0·749; difference 0·005, 95% CI, -0·011 to 0·020). If the model-defined moderate-risk and high-risk groups were referred for endoscopy, the sensitivity was 94·5% (95% CI 88·8 to 97·5), specificity was 91·9% (91·2 to 92·5), and the predictive positive value was 18·4% (15·6 to 21·6), and 90·3% of endoscopies could be avoided. Further validation in community-based screening showed that the AUROC of the model was 0·964 (95% CI 0·920 to 0·990), and 92·8% of endoscopies could be avoided after risk stratification.

INTERPRETATION

We developed a prediction tool with favourable performance for screening of oesophageal squamous cell carcinoma and adenocarcinoma of the oesophagogastric junction. This approach could prevent the need for endoscopy screening in many low-risk individuals and ensure resource optimisation by prioritising high-risk individuals.

FUNDING

Science and Technology Commission of Shanghai Municipality.

Study protocol for artificial intelligence-assisted sponge cytology as pre-endoscopy screening for early esophegeal squmaous epithelial lesions in China

Background

Endoscopic screening is the widely accepted screening strategy for esophageal squmaous cell carcinoma (ESCC). However, massive endoscopic screening is expensive and not cost-efficient, and novel pre-endoscopy detection used as a preliminary screening method arouses new concerns. We are planning to launch an artificial intelligence (AI) assisted sponge cytology for detecting esophageal squmaous high-grade intraepithelial neoplasia (HGIN) and above lesions. The aim of this trail is to investigate the efficiency of AI-assisted sponge cytology in population-based screening of early esophageal squmaous epithelial lesions.

Methods

The study will be prospectively conducted in five regions with a high prevalence of ESCC. AI-assisted sponge cytology and endoscopic examination will be sequentially performed. Based on our previous data, at least 864 patients with esophageal HGIN and above lesions are needed to achieve enough statistical power. And, a calculated 112,500 individuals with high risks of ESCC will be recruited. In the first stage, each 24,000 participants who meet the inclusion criteria will be recruited on a voluntary basis. Setting pathological results as standard reference, diagnostic threshold and according performance of AI-assisted detection will be evaluated. A prediction model will be constructed by co-analyzing cytological results and relevant risk factors. Then, an external validation cohort will be used for validation of the model efficiency. Also, cost-efficiency analysis will be performed. This study protocol was registered on chineseclinicaltrial.gov (ChiCTR1900028524).

Discussion

Our study will determine whether this AI-assisted sponge cytology can be used as an effective pre-endoscopy detection tool for large-scale screening for ESCC in high-risk areas.

Artificial intelligence assisted cytological detection for early esophageal squamous epithelial lesions by using low-grade squamous intraepithelial lesion as diagnostic threshold

BACKGROUND

Manual cytological diagnosis for early esophageal squamous cell carcinoma (early ESCC) and high-grade intraepithelial neoplasia (HGIN) is unsatisfactory. Herein, we have introduced an artificial intelligence (AI)-assisted cytological diagnosis for such lesions.

METHODS

Low-grade squamous intraepithelial lesion or worse was set as the diagnostic threshold for AI-assisted diagnosis. The performance of AI-assisted diagnosis was evaluated and compared to that of manual diagnosis. Feasibility in large-scale screening was also assessed.

RESULTS

AI-assisted diagnosis for abnormal cells was superior to manual reading by presenting a higher efficiency for each slide (50.9 ± 0.8 s vs 236.8 ± 3.9 s, p = 1.52 × 10^-76 ) and a better interobserver agreement (93.27% [95% CI, 92.76%-93.74%] vs 65.29% [95% CI, 64.35%-66.22%], p = 1.03 × 10^-84 ). AI-assisted detection showed a higher diagnostic accuracy (96.89% [92.38%-98.57%] vs 72.54% [65.85%-78.35%], p = 1.42 × 10^-14 ), sensitivity (99.35% [95.92%-99.97%] vs 68.39% [60.36%-75.48%], p = 7.11 × 10^-15 ), and negative predictive value (NPV) (97.06% [82.95%-99.85%] vs 40.96% [30.46%-52.31%], p = 1.42 × 10^-14 ). Specificity and positive predictive value (PPV) were not significantly differed. AI-assisted diagnosis demonstrated a smaller proportion of participants of interest (3.73%, [79/2117] vs.12.84% [272/2117], p = 1.59 × 10^-58 ), a higher consistence between cytology and endoscopy (40.51% [32/79] vs. 12.13% [33/272], p = 1.54 × 10^- 8), specificity (97.74% [96.98%-98.32%] vs 88.52% [87.05%-89.84%], p = 3.19 × 10^-58 ), and PPV (40.51% [29.79%-52.15%] vs 12.13% [8.61%-16.75%], p = 1.54 × 10^-8 ) in community-based screening. Sensitivity and NPV were not significantly differed. AI-assisted diagnosis as primary screening significantly reduced average cost for detecting positive cases.

CONCLUSION

Our study provides a novel cytological method for detecting and screening early ESCC and HGIN.

Feasibility of using P16 methylation as a cytologic marker for esophageal squamous cell carcinoma screening: A pilot study

BACKGROUND

Early diagnosis and treatment of esophageal squamous cell dysplasia (ESCdys) and esophageal squamous cell carcinoma (ESCC) could significantly reduce the incidence and mortality of ESCC. This pilot study aimed to investigate whether P16/CDKN2A methylation could serve as a cytologic biomarker for early detection of ESCdys and ESCC.

METHODS

Paired esophageal biopsy and cytology specimens (exfoliated cells) were obtained from subjects at different stages of ESCC development. The methylation status of P16 gene in these two specimen types was determined using a 115-bp MethyLight assay. Categorical data were compared by the Chi-square test. Logistic regression was performed to assess adjusted odds ratios of P16 methylation associated with ESCC and ESCdys. Prediction models for identifying individuals at risk of ESCC and high-grade ESCdys (high-grade intraepithelial neoplasia, HGIN) were developed by multivariable logistic regression. Diagnostic performance was evaluated using receiver operating characteristic (ROC) analysis. Internal validation of the prediction models was performed using the 1000-bootstrap resample.

RESULTS

A total of 105 subjects with diagnoses ranging from normal mucosa through ESCC were included in this study. An increase in P16 methylation frequency was observed with increasing severity of esophageal lesions (p for trend <0.001). In the adjusted logistic regression models, P16 methylation in cytology specimens was positively associated with ESCC and ESCdys risk, whereas P16 methylation in biopsy specimens was only associated with a higher risk of developing ESCC. The predictive capacity of base model I (AUC, 0.816) for ESCC and HGIN was significantly increased by adding P16 methylation in cytology specimens (model III; AUC, 0.882; p = 0.043), but not P16 methylation in biopsy specimens (model II; AUC, 0.850; p = 0.225). Bootstrap validation showed optimism-corrected AUC of 0.789 for model I, 0.822 for model II, and 0.854 for model III.

CONCLUSION

P16 methylation as a cytologic marker was associated with the ESCC development and has the potential for application in minimally invasive ESCC screening.