Risk Factors and Prediction Model for Non-curative Resection of Early Gastric Cancer With Endoscopic Resection and the Evaluation

Development and validation of a predictive model for submucosal fibrosis in patients with early gastric cancer undergoing endoscopic submucosal dissection: experience from a large tertiary center

BACKGROUND

Submucosal fibrosis is associated with adverse events of endoscopic submucosal dissection (ESD). The present study mainly aimed to establish a predictive model for submucosal fibrosis in patients with early gastric cancer (EGC) undergoing ESD.

METHODS

Eligible patients with EGC, identified at Qilu Hospital of Shandong University from April 2013 to December 2023, were retrospectively included and randomly split into a training set and a validation set in a 7:3 ratio. Logistic regression analyses were used to pinpoint the risk factors for submucosal fibrosis. A nomogram was developed and confirmed using receiver operating characteristic (ROC) curves, calibration plots, Hosmer-Lemeshow (H-L) tests, and decision curve analysis (DCA) curves. Besides, a predictive model for severe submucosal fibrosis was further conducted and tested.

RESULTS

A total of 516 cases in the training group and 220 cases in the validation group were recruited. The nomogram for submucosal fibrosis contained the following items: tumour location (long axis), tumour location (short axis), ulceration, and biopsy pathology. ROC curves showed high efficiency with an area under the ROC of 0.819 in the training group, and 0.812 in the validation group. Calibration curves and H-L tests indicated good consistency. DCA proved the nomogram to be clinically beneficial. Furthermore, the four items were also applicable for a nomogram predicting severe fibrosis, and the model performed well.

CONCLUSION

The predictive models, initially constructed in this study, were validated as convenient and feasible for endoscopists to predict submucosal fibrosis and severe fibrosis in patients with EGC undergoing ESD.

Clinical outcome of non-curative endoscopic submucosal dissection for early gastric cancer

Background

Early gastric cancer (EGC) is defined as cancer cells confined to the mucosal or submucosal layer, irrespective of size or presence of lymph node metastasis. The recent EGC endoscopic submucosal dissection (ESD) and endoscopic mucosal resection (EMR) guidelines (2021 Japan Gastroenterological Endoscopy Society (JGES) guidelines, 2nd edition) revised the concept from "endoscopic curative/non-curative resection" (NCR) to "endoscopic curability (eCura)". Under this, eCuraA and eCuraB signify curative resections (CRs), while eCuraC (including eCuraC-1 and eCura-C2) indicate NCRs. This study retrospectively analyzes clinical and pathological data from EGC patients who underwent endoscopic resection, assessing the long-term clinical outcomes in a substantial cohort after undergoing NCR.

Methods

We retrospectively analyzed clinical and pathological data from 443 EGC patients, encompassing 478 lesions, who received endoscopic treatment. The long-term clinical outcomes of patients who underwent NCR were statistically evaluated. Characteristics of the NCR group were compared with those of the surgical group, employing single- and multi-factor logistic regression analyses to identify risk factors that necessitate further surgical intervention. Prognostically, the Kaplan-Meier method and Log-Rank test determined the impact of risk factors on recurrence-free survival post-surgery in NCR patients. Differences were assessed using a method incorporating statistically significant differences in the multi-factor Cox regression analysis, evaluating the hazard ratio (HR) for disease recurrence following NCR.

Results

In this study, 443 EGC cases were pathologically diagnosed, comprising a total of 478 lesions. Of these, 127 cases underwent non-curative endoscopic resection, resulting in a NCR rate of 24.4%. Long-term follow-up was achieved for 117 (92.12%) patients. The metastasis/recurrence rate at 6 months stood at 23.1%. Multivariate Cox regression analysis identified lesion size ≥2.0 and <3 cm [P=0.02, HR =0.12, 95% confidence interval (CI): 0.02-0.67], presence of ulceration (P=0.03, HR =5.48, 95% CI: 1.23-24.33), lymphatic invasion (P=0.05, HR =17.51, 95% CI: 1.07-286.23), positive vertical margins (P=0.09, HR =3.77, 95% CI: 0.81-17.53), and flat macroscopic morphology (P=0.048, HR =4.8, 95% CI: 1.01-22.73) as independent risk factors for recurrence-free survival post non-curative endoscopic resection in EGC patients.

Conclusions

The recurrence/metastasis rate in patients who underwent NCR is notably higher compared to the control group. Significant prognostic risk factors include tumor size ≥2.0 and <3 cm, positive vertical margins, lymphatic invasion, and flat type (one of pathological gross classification). Patients in the eCuraC-2 category of NCR should consider further surgical intervention. The necessity for additional surgical intervention in these patients warrants further investigation.

Comparison of endoscopic submucosal dissection outcomes between early gastric cardiac and non-cardiac cancers: a retrospective single-center study

OBJECTIVES

This study aims to compare the efficacy of endoscopic submucosal dissection (ESD) between early gastric cardiac cancer (EGCC) and early gastric non-cardiac cancer (EGNCC), and investigate associated risk factors for non-curative resection.

METHODS

Early gastric cancer (EGC) patients who underwent ESD from January 2015 to September 2020 in Beijing Friendship Hospital were consecutively enrolled. The clinical, histopathological and endoscopic data were retrospectively analyzed. The study was registered in Chinese Clinical Trial Registry (ChiCTR1800017117).

RESULTS

Among 500 patients with 534 EGC lesions, 117 patients with 118 lesions were allocated to the EGCC group, and 383 patients with 416 lesions to the EGNCC group. The rates of en bloc resection, complete resection and curative resection in the EGCC group were 97.5%, 78.8% and 71.2%, respectively, significantly lower than those in the EGNCC group (99.8%, 94.5% and 90.4%, p = .010, <.001 and <.001). Among non-curative resected lesions, EGCC had more cases in both endoscopic curability (eCura) C-1 and C-2 groups than EGNCC (10.2% and 18.6% vs. 2.4% and 7.2%, p < .001). Multivariate analysis showed that tumor size (OR 2.393, 95% CI 1.388-4.126) and submucosal invasion (OR 11.498, 95% CI 3.759-35.175) were risk factors for non-curative resection in the EGCC group. For EGCC larger than 3 cm, none achieved curative resection, 86.7% were classified as eCura C-2 and 46.7% exhibited deep submucosal infiltration.

CONCLUSIONS

The curative resection rate of ESD for EGCC was lower than that for EGNCC. ESD for EGCC larger than 3 cm should be cautiously considered.

Endoscopic Resection of Early Gastric Cancer and Pre-Malignant Gastric Lesions

Early gastric cancer comprises gastric malignancies that are confined to the mucosa or submucosa, irrespective of lymph node metastasis. Endoscopic resection is currently pivotal for the management of such early lesions, and it is the recommended treatment for tumors presenting a very low risk of lymph node metastasis. In general, these lesions consist of two groups of differentiated mucosal adenocarcinomas: non-ulcerated lesions (regardless of their size) and small ulcerated lesions. Endoscopic submucosal dissection is the technique of choice in most cases. This procedure has high rates of complete histological resection while maintaining gastric anatomy and its functions, resulting in fewer adverse events than surgery and having a lesser impact on patient-reported quality of life. Nonetheless, approximately 20% of resected lesions do not fulfill curative criteria and demand further treatment, highlighting the importance of patient selection. Additionally, the preservation of the stomach results in a moderate risk of metachronous lesions, which underlines the need for surveillance. We review the current evidence regarding the endoscopic treatment of early gastric cancer, including the short-and long-term results and management after resection.

Systemic immune-inflammation index in predicting non-curative resection of endoscopic submucosal dissection in patients with early gastric cancer

BACKGROUND AND PURPOSE

Although endoscopic submucosal dissection (ESD) is considered standard treatment for early gastric cancer (EGC), patients with non-curative resection (NCR) of ESD may still require gastrectomy. The systemic immune-inflammation index (SII) showed great potential in predicting the prognosis of gastric cancer patients. This study aims to investigate the predictive validity of SII of NCR in EGC patients.

METHODS

We reviewed data from EGC patients who underwent ESD in the past. The relationship between SII and clinicopathologic features was investigated. We used Receiver operating characteristic curves to compare the predictive values of NCR between SII and other inflammation indices. Binary logistic analysis was used to identify independent risk factors for NCR. These factors were then used to construct a predictive nomogram.

RESULTS

SII was associated with larger tumor size, male gender, older age, submucosal invasion, and a greater risk of NCR. SII showed better predictivity of NCR than platelet-to-lymphocyte ratio (PLR) and neutrophil-to-lymphocyte ratio (NLR). SII [odds ratio (OR) = 1.003, P = 0.001], NLR (OR = 1.520, P = 0.029), PLR (OR = 1.009, P = 0.010), upper stomach tumors (OR = 16.393, P < 0.001), poorly differentiated type (OR = 29.754, P < 0.001), ulceration (OR = 4.814, P = 0.001), and submucosal invasion (OR = 48.91, P < 0.001) were independent risk factors for NCR. The nomogram model based on these factors exhibited superior concordance and accuracy.

CONCLUSION

SII could be considered a simple and effective predictor of NCR of ESD in EGC patients.

Nomogram for pre-procedural prediction of non-curative endoscopic resection in patients with early gastric cancer

BACKGROUND

Non-curative resection (non-CR) after endoscopic submucosal dissection (ESD) requires additional surgery due to the possibility of lymph node metastasis (LNM). Therefore, it is important to accurately predict the risk of non-CR to avoid unnecessary preoperative procedures. Thus, we aimed to develop and verify a nomogram to predict the risk of non-CR prior to ESD.

METHODS

Patients who underwent ESD for early gastric cancer (EGC) were divided into CR and non-CR groups based on the present ESD criteria. The pre-procedural factors, such as endoscopic features, radiologic findings, and pathology of the lesion, were compared between the groups to identify the risk factors associated with non-CR. A nomogram was developed using multivariate analysis, and its predictive value was assessed using an external validation group.

RESULTS

Among 824 patients, 682 were curative (82.7%) and 142 were non-curative (17.3%). By comparing two groups, endoscopic features including redness, whitish mucosal change, fold convergence, and large lesion size; histologic features such as moderately or poorly differentiated or signet ring cell carcinoma; and abnormal CT findings including non-specific lymph node enlargement and fold thickening were identified as significant predictors of non-CR. The nomogram was developed based on these predictors and showed good predictive performance in the external validation, with an area under the curve of 0.87.

CONCLUSIONS

We developed a nomogram to predict the risk of non-CR prior to ESD. These predictive factors in addition to the existing ESD criteria can help provide the best treatment option for patients with EGC.

Construction of a nomogram for preoperative prediction of the risk of lymph node metastasis in early gastric cancer

Background

The status of lymph node metastasis (LNM) in patients with early gastric cancer (EGC) is particularly important for the formulation of clinical treatment. The purpose of this study was to construct a nomogram to predict the risk of LNM in EGC before operation.

Methods

Univariate analysis and logistic regression analysis were used to determine the independent risk factors for LNM. The independent risk factors were included in the nomogram, and the prediction accuracy, discriminant ability and clinical practicability of the nomogram were evaluated by the receiver operating characteristic curve (ROC), calibration curve and clinical decision curve (DCA), and 100 times ten-fold cross-validation was used for internal validation.

Results

33 (11.3%) cases of AGC were pathologically confirmed as LNM. In multivariate analysis, T stage, presence of enlarged lymph nodes on CT examination, carbohydrate antigen 199 (CA199), undifferentiated histological type and systemic inflammatory response index (SIRI) were risk factors for LNM. The area under the ROC curve of the nomogram was 0.86, the average area under the ROC curve of the 100-fold ten-fold cross-validation was 0.85, and the P value of the Hosmer-Lemeshow test was 0.60. In addition, the clinical decision curve, net reclassification index (NRI) and Integrated Discriminant Improvement Index (IDI) showed that the nomogram had good clinical utility.

Conclusions

We found that SIRI is a novel biomarker for preoperative prediction of LNM in EGC, and constructed a nomogram for preoperative prediction of the risk of LNM in EGC, which is helpful for the formulation of the clinical treatment strategies.

Establishment and Validation of a Pathologic Upgrade Prediction Nomogram Model for Gastric Low-Grade Intraepithelial Neoplasia Patients After the Eradication of Helicobacter pylori

BACKGROUND

As yet, there is no unified method of treatment for the evaluation and management of gastric low-grade intraepithelial neoplasia (LGIN) worldwide.

METHODS

Patients with gastric LGIN who had been treated with Helicobacter pylori eradication were gathered retrospectively. Based on several relevant characteristics described and analyzed by LASSO regression analysis and multivariable logistic regression, a prediction nomogram model was established. C-index, the area under the receiver operating characteristic curve (AUC), calibration plot, and decision curve analysis (DCA) were adopted to evaluate the accuracy and reliability of the model.

RESULTS

A total of 309 patients with LGIN were randomly divided into the training groups and the validation groups. LASSO regression analysis and multivariable logistic regression identified that 6 variables including gender, size, location, borderline, number, and erosion were independent risk factors. The nomogram model displayed good discrimination with a C-index of .765 (95% confidence interval: .702-.828). The accuracy and reliability of the model were also verified by an AUC of .764 in the training group and .757 in the validation group. Meanwhile, the calibration curve and the DCA suggested that the predictive nomogram had promising accuracy and clinical utility.

CONCLUSIONS

A predictive nomogram model was constructed and proved to be clinically applicable to identify high-risk groups with possible pathologic upgrade in patients with gastric LGIN. Since it is regarded that strengthening follow-up or endoscopic treatment of high-risk patients may contribute to improving the detection rate or reducing the incidence of gastric cancer, the predictive nomogram model provides a reliable basis for the treatment of LGIN.

Machine learning for lymph node metastasis prediction of in patients with gastric cancer: A systematic review and meta-analysis

Objective

To evaluate the diagnostic performance of machine learning (ML) in predicting lymph node metastasis (LNM) in patients with gastric cancer (GC) and to identify predictors applicable to the models.

Methods

PubMed, EMBASE, Web of Science, and Cochrane Library were searched from inception to March 16, 2022. The pooled c-index and accuracy were used to assess the diagnostic accuracy. Subgroup analysis was performed based on ML types. Meta-analyses were performed using random-effect models. Risk of bias assessment was conducted using PROBAST tool.

Results

A total of 41 studies (56182 patients) were included, and 33 of the studies divided the participants into a training set and a test set, while the rest of the studies only had a training set. The c-index of ML for LNM prediction in training set and test set was 0.837 [95%CI (0.814, 0.859)] and 0.811 [95%CI (0.785-0.838)], respectively. The pooled accuracy was 0.781 [(95%CI (0.756-0.805)] in training set and 0.753 [95%CI (0.721-0.783)] in test set. Subgroup analysis for different ML algorithms and staging of GC showed no significant difference. In contrast, in the subgroup analysis for predictors, in the training set, the model that included radiomics had better accuracy than the model with only clinical predictors (F = 3.546, p = 0.037). Additionally, cancer size, depth of cancer invasion and histological differentiation were the three most commonly used features in models built for prediction.

Conclusion

ML has shown to be of excellent diagnostic performance in predicting the LNM of GC. One of the models covering radiomics and its ML algorithms showed good accuracy for the risk of LNM in GC. However, the results revealed some methodological limitations in the development process. Future studies should focus on refining and improving existing models to improve the accuracy of LNM prediction.

Systematic Review Registration

https://www.crd.york.ac.uk/PROSPERO/, identifier CRD42022320752

Machine Learning Improves the Prediction Rate of Non-Curative Resection of Endoscopic Submucosal Dissection in Patients with Early Gastric Cancer

Non-curative resection (NCR) of early gastric cancer (EGC) after endoscopic submucosal dissection (ESD) can increase the burden of additional treatment and medical expenses. We aimed to develop a machine-learning (ML)-based NCR prediction model for EGC prior to ESD. We obtained data from 4927 patients with EGC who underwent ESD between January 2006 and February 2020. Ten clinicopathological characteristics were selected using extreme gradient boosting (XGBoost) and were used to develop a ML-based model. Dataset was divided into the training and internal validation sets and verified using an external validation set. Sensitivity, specificity, and area under the receiver operating characteristic curve (AUROC) were evaluated. The performance of each model was compared by using the Delong test. A total of 1100 (22.1%) patients were identified as being treated non-curatively with ESD. Seven ML-based NCR prediction models were developed. The performance of NCR prediction was highest in the XGBoost model (AUROC, 0.851; 95% confidence interval, 0.837–0.864). When we compared the prediction performance by the Delong test, XGBoost (p = 0.02) and support vector machine (p = 0.02) models showed a significantly higher performance among the NCR prediction models. We developed an ML model capable of accurately predicting the NCR of EGC before ESD. This ML model can provide useful information for decision-making regarding the appropriate treatment of EGC before ESD.

Endoscopic submucosal dissection for superficial gastrointestinal lesions: European Society of Gastrointestinal Endoscopy (ESGE) Guideline - Update 2022

ESGE recommends that the evaluation of superficial gastrointestinal (GI) lesions should be made by an experienced endoscopist, using high definition white-light and chromoendoscopy (virtual or dye-based).ESGE does not recommend routine performance of endoscopic ultrasonography (EUS), computed tomography (CT), magnetic resonance imaging (MRI), or positron emission tomography (PET)-CT prior to endoscopic resection.ESGE recommends endoscopic submucosal dissection (ESD) as the treatment of choice for most superficial esophageal squamous cell and superficial gastric lesions.For Barrett's esophagus (BE)-associated lesions, ESGE suggests the use of ESD for lesions suspicious of submucosal invasion (Paris type 0-Is, 0-IIc), for malignant lesions > 20 mm, and for lesions in scarred/fibrotic areas.ESGE does not recommend routine use of ESD for duodenal or small-bowel lesions.ESGE suggests that ESD should be considered for en bloc resection of colorectal (but particularly rectal) lesions with suspicion of limited submucosal invasion (demarcated depressed area with irregular surface pattern or a large protruding or bulky component, particularly if the lesions are larger than 20 mm) or for lesions that otherwise cannot be completely removed by snare-based techniques.ESGE recommends that an en bloc R0 resection of a superficial GI lesion with histology no more advanced than intramucosal cancer (no more than m2 in esophageal squamous cell carcinoma), well to moderately differentiated, with no lymphovascular invasion or ulceration, should be considered a very low risk (curative) resection, and no further staging procedure or treatment is generally recommended.ESGE recommends that the following should be considered to be a low risk (curative) resection and no further treatment is generally recommended: an en bloc R0 resection of a superficial GI lesion with superficial submucosal invasion (sm1), that is well to moderately differentiated, with no lymphovascular invasion, of size ≤ 20 mm for an esophageal squamous cell carcinoma or ≤ 30 mm for a stomach lesion or of any size for a BE-related or colorectal lesion, and with no lymphovascular invasion, and no budding grade 2 or 3 for colorectal lesions.ESGE recommends that, after an endoscopically complete resection, if there is a positive horizontal margin or if resection is piecemeal, but there is no submucosal invasion and no other high risk criteria are met, this should be considered a local-risk resection and endoscopic surveillance or re-treatment is recommended rather than surgery or other additional treatment.ESGE recommends that when there is a diagnosis of lymphovascular invasion, or deeper infiltration than sm1, or positive vertical margins, or undifferentiated tumor, or, for colorectal lesions, budding grade 2 or 3, this should be considered a high risk (noncurative) resection, and complete staging and strong consideration for additional treatments should be considered on an individual basis in a multidisciplinary discussion.ESGE recommends scheduled endoscopic surveillance with high definition white-light and chromoendoscopy (virtual or dye-based) with biopsies of only the suspicious areas after a curative ESD.

The future of endoscopic resection for early gastric cancer

Endoscopic resection for early gastric cancer is recommended when the risk of lymph node metastasis is negligible and should be performed through submucosal dissection due to well-established short- and long-term results. To overcome technical difficulties and decrease adverse events some techniques have been studied. This review outlines current strategies for improving patient selection and highlights innovative techniques that help minimize adverse events. Moreover, we discuss how to improve management after curative and noncurative resections.

Application of Machine Learning Algorithms to Predict Lymph Node Metastasis in Early Gastric Cancer

Objective

This study aimed to establish the best early gastric cancer lymph node metastasis (LNM) prediction model through machine learning (ML) to better guide clinical diagnosis and treatment decisions.

Methods

We screened gastric cancer patients with T1a and T1b stages from 2010 to 2015 in the Surveillance, Epidemiology and End Results (SEER) database and collected the clinicopathological data of patients with early gastric cancer who were treated with surgery at the Second Affiliated Hospital of Nanchang University from January 2014 to December 2016. At the same time, we applied 7 ML algorithms—the generalized linear model (GLM), RPART, random forest (RF), gradient boosting machine (GBM), support vector machine (SVM), regularized dual averaging (RDA), and the neural network (NNET)—and combined them with patient pathological information to develop the best prediction model for early gastric cancer lymph node metastasis. Among the SEER set, 80% were randomly selected to train the models, while the remaining 20% were used for testing. The data from the Second Affiliated Hospital were considered as the external verification set. Finally, we used the AUROC, F1-score value, sensitivity, and specificity to evaluate the performance of the model.

Results

The tumour size, tumour grade, and depth of tumour invasion were independent risk factors for early gastric cancer LNM. Comprehensive comparison of the prediction model performance of the training set and test set showed that the RDA model had the best prediction performance (F1-score = 0.773; AUROC = 0.742). The AUROC of the external validation set was 0.73.

Conclusions

Tumour size, tumour grade, and depth of tumour invasion were independent risk factors for early gastric cancer LNM. ML predicted LNM risk more accurately, and the RDA model had the best predictive performance and could better guide clinical diagnosis and treatment decisions.