A prediction model for lymph node metastasis in early-stage gastric cancer: Toward tailored lymphadenectomy

Enhancing Lymph Node Metastasis Risk Prediction in Early Gastric Cancer Through the Integration of Endoscopic Images and Real-World Data in a Multimodal AI Model

Objectives: The accurate prediction of lymph node metastasis (LNM) and lymphovascular invasion (LVI) is crucial for determining treatment strategies for early gastric cancer (EGC). This study aimed to develop and validate a deep learning-based clinical decision support system (CDSS) to predict LNM including LVI in EGC using real-world data. Methods: A deep learning-based CDSS was developed by integrating endoscopic images, demographic data, biopsy pathology, and CT findings from the data of 2927 patients with EGC across five institutions. We compared a transformer-based model to an image-only (basic convolutional neural network (CNN)) model and a multimodal classification (CNN with random forest) model. Internal testing was conducted on 449 patients from the five institutions, and external validation was performed on 766 patients from two other institutions. Model performance was assessed using the area under the receiver operating characteristic curve (AUC), probability density function, and clinical utility curve. Results: In the training, internal, and external validation cohorts, LNM/LVI was observed in 379 (12.95%), 49 (10.91%), 15 (9.09%), and 41 (6.82%) patients, respectively. The transformer-based model achieved an AUC of 0.9083, sensitivity of 85.71%, and specificity of 90.75%, outperforming the CNN (AUC 0.5937) and CNN with random forest (AUC 0.7548). High sensitivity and specificity were maintained in internal and external validations. The transformer model distinguished 91.8% of patients with LNM in the internal validation dataset, and 94.0% and 89.1% in the two different external datasets. Conclusions: We propose a deep learning-based CDSS for predicting LNM/LVI in EGC by integrating real-world data, potentially guiding treatment strategies in clinical settings.

Latest guideline of endoscopic submucosal dissection of early gastric cancer may not be suitable for Chinese patients: retrospective study findings from two centers

BACKGROUND

To analyze the diagnostic efficiency of the four absolute endoscopic submucosal dissection (ESD) indications for lymph node metastasis (LNM) of Chinese patients with early gastric cancer (EGC).

METHODS

We retrospectively analyzed EGC patients who underwent radical D2 gastrectomy from January 2019 to December 2022. We evaluated the rate of LNM, false-negative rate, and negative predictive value of the four ESD indications.

RESULTS

Of enrolled 2722 EGC patients, 388 (14.3%) patients presented LNM. Tumor size > 2 cm, ulceration, submucosal invasion, undifferentiated type, and lymphovascular invasion were independent risk factors of LNM in patients with EGC. 1062 (39%) cases of EGC conformed to the four EDS indications; however, 4% of them had LNM. 451 cases were fully in accord with the fourth ESD indication (undifferentiated intramucosal carcinoma without ulceration and a maximum lesion diameter of ≤ 2 cm), and 35 of them had LNM, with a false-negative rate (FNR) of 9.02% and a negative predictive value (NPV) of 92.24%. There was significant difference among the four indications in terms of the rate of LNM (1.0% vs 1.5% vs 1.3% vs 7.8%, P < 0.001), FNR (1.03% vs 0.52% vs 0.26% vs 9.02%, P < 0.001), and NPV (98.99% vs 98.53% vs 98.75% vs 92.24%, P < 0.001).

CONCLUSION

Overall, the fourth ESD indication was associated with a high rate of LNM compared to the other three indications. Thus, it might not be safe to classify it as an absolute indication in Chinese patients with EGC.

Feasibility of Regional Lymphadenectomy for Stomach-Preserving Surgery in Early Gastric Cancer Omitting Sentinel Node Navigation: A Post Hoc Analysis of the SENORITA Trial

BACKGROUND

Sentinel node navigation (SNN) has been known as the effective treatment for stomach-preserving surgery in early gastric cancer; however, SNN presents several technical difficulties in real practice.

OBJECTIVE

This study aimed to evaluate the feasibility of regional lymphadenectomy omitting SNN, using the post hoc analysis of a randomized controlled trial.

METHODS

Using data from the SENORITA trial that compared laparoscopic standard gastrectomy with lymphadenectomy and laparoscopic SNN, 237 patients who underwent SNN were included in this study. Tumor location was divided into longitudinal and circumferential directions. According to the location of the tumor, the presence or absence of lymph node (LN) metastases between sentinel and non-sentinel basins were analyzed. Proposed regional LN stations were defined as the closest area to the primary tumor. Sensitivities, specificities, positive predictive values, and negative predictive values (NPV) of SNN and regional lymphadenectomy were compared.

RESULTS

Metastasis to non-sentinel basins with tumor-free in sentinel basins was observed in one patient (0.4%). The rate of LN metastasis to non-regional LN stations without regional LN metastasis was 2.5% (6/237). The sensitivity and NPV of SNN were found to be significantly higher than those of regional lymphadenectomy (96.8% vs. 80.6% [p = 0.016] and 99.5% vs. 97.2% [p = 0.021], respectively).

CONCLUSIONS

This study showed that regional lymphadenectomy for stomach-preserving surgery, omitting SNN, was insufficient; therefore, SNN is required in stomach-preserving surgery.

Predicting lung cancer's metastats' locations using bioclinical model

Background

Lung cancer is a global leading cause of cancer-related deaths, and metastasis profoundly influences treatment outcomes. The limitations of conventional imaging in detecting small metastases highlight the crucial need for advanced diagnostic approaches.

Methods

This study developed a bioclinical model using three-dimensional CT scans to predict the spatial spread of lung cancer metastasis. Utilizing a three-layer biological model, we identified regions with a high probability of metastasis colonization and validated the model on real-world data from 10 patients.

Findings

The validated bioclinical model demonstrated a promising 74% accuracy in predicting metastasis locations, showcasing the potential of integrating biophysical and machine learning models. These findings underscore the significance of a more comprehensive approach to lung cancer diagnosis and treatment.

Interpretation

This study's integration of biophysical and machine learning models contributes to advancing lung cancer diagnosis and treatment, providing nuanced insights for informed decision-making.

Metastasis patterns and prognosis in young gastric cancer patients: A propensity score‑matched SEER database analysis

BACKGROUND

Whether young patients with metastatic gastric cancer (GC) had distinct metastasis patterns and survival outcomes from older patients remains controversial. The aim of the present study was to explore the metastasis patterns and prognostic factors in young patients and evaluate the survival outcome in comparison to their older counterparts.

MATERIALS AND METHODS

We identified patients with metastatic GC in the surveillance, epidemiology, and end results (SEER) database from 2010 to 2015. The patients were divided into two groups based on age at diagnosis: younger (≤40 years old) and older (>40 years old). We employed the chi-squared test to compare the clinicopathological characteristics between the two age groups. Furthermore, we conducted survival analyses using Kaplan-Meier and Cox regression analyses. To balance disparities in baseline characteristics, we employed propensity score matching (PSM).

RESULTS

We identified 5,580 metastatic GC patients from the SEER database, with 237 (4.2%) classified as younger and 5343 (95.8%) as older patients. A total of 237 pairs of patients were generated after adjustment by PSM. Patients in the younger group exhibited a higher proportion of bone-only metastases and a lower proportion of liver-only metastases compared with patients in the older group. Multivariate Cox regression analysis demonstrated that youth was an independent protective factor for overall survival (OS) before and after PSM, but not for gastric cancer-specific survival (GCSS). Among the younger group, patients with liver-only metastasis demonstrated the best prognosis, whereas patients with lung-only metastasis exhibited significantly worse survival outcomes compared with liver-only metastases, even comparable to that of bone metastasis.

CONCLUSIONS

Compared with the older group, the metastatic GC patients in the younger group exhibited more aggressive tumors but better prognoses. The metastasis pattern and its effect on the prognosis of GC varied by age group.

Development and validation of a predictive model associated with lymph node metastasis of gastric signet ring carcinoma patients

The risk factors for lymph node metastasis (LNM) in patients with gastric signet ring cell carcinoma (GSRC) have not been well-defined. This study was designed to prognosticate LNM in patients with GSRC by constructing and verifying a nomogram. A total of 2789 patients with GSRC from the Surveillance, Epidemiology, and End Results (SEER) database and Yijishan Hospital of Wannan Medical College (YJS) were retrospectively reviewed. A predictive model was established using logistic regression based on the SEER cohort. The performance of the model was evaluated using the concordance index (C-index) and decision curve analysis (DCA). In addition, its robustness was validated using the YJS cohort. Four independent predictors of LNM were identified in the SEER cohort. Next, a nomogram was constructed by incorporating these predictors. The C-index were 0.800 (95% confidence interval [CI] = 0.781-0.819) and 0.837 (95% CI = 0.784-0.890) in the training and external validation cohorts, respectively. The outcomes of DCA supported good clinical benefits. The proposed model for evaluating the LNM in patients with GSRC can help to avoid the misdiagnosis risk of N-stage, assist to screen the population suitable for neoadjuvant therapy and help clinicians to optimize clinical decisions.

Development and Validation of Models to Predict Lymph Node Metastasis in Early Gastric Cancer Using Logistic Regression and Gradient Boosting Machine Methods

PURPOSE

To identify important features of lymph node metastasis (LNM) and develop a prediction model for early gastric cancer (EGC) using a gradient boosting machine (GBM) method.

MATERIALS AND METHODS

The clinicopathologic data of 2556 patients with EGC who underwent gastrectomy were used as training set and the internal validation set (set 1) at a ratio of 8:2. Additionally, 548 patients with EGC who underwent endoscopic submucosal dissection (ESD) as the initial treatment were included in the external validation set (set 2). The GBM model was constructed, and its performance was compared with that of the Japanese guidelines.

RESULTS

LNM was identified in 12.6% (321/2556) of the gastrectomy group (training set & set 1) and 4.3% (24/548) of the ESD group (set 2). In the GBM analysis, the top five features that most affected LNM were lymphovascular invasion, depth, differentiation, size, and location. The accuracy, sensitivity, specificity, and the area under the receiver operating characteristics of set 1 were 0.566, 0.922, 0.516, and 0.867, while those of set 2 were 0.810, 0.958, 0.803, and 0.944, respectively. When the sensitivity of GBM was adjusted to that of Japanese guidelines (beyond the expanded criteria in set 1 [0.922] and eCuraC-2 in set 2 [0.958]), the specificities of GBM in sets 1 and 2 were 0.516 (95% confidence interval, 0.502-0.523) and 0.803 (0.795-0.805), while those of the Japanese guidelines were 0.502 (0.488-0.509) and 0.788 (0.780-0.790), respectively.

CONCLUSION

The GBM model showed good performance comparable with the eCura system in predicting LNM risk in EGCs.

Machine learning applications in upper gastrointestinal cancer surgery: a systematic review

BACKGROUND

Machine learning (ML) has seen an increase in application, and is an important element of a digital evolution. The role of ML within upper gastrointestinal surgery for malignancies has not been evaluated properly in the literature. Therefore, this systematic review aims to provide a comprehensive overview of ML applications within upper gastrointestinal surgery for malignancies.

METHODS

A systematic search was performed in PubMed, EMBASE, Cochrane, and Web of Science. Studies were only included when they described machine learning in upper gastrointestinal surgery for malignancies. The Cochrane risk-of-bias tool was used to determine the methodological quality of studies. The accuracy and area under the curve were evaluated, representing the predictive performances of ML models.

RESULTS

From a total of 1821 articles, 27 studies met the inclusion criteria. Most studies received a moderate risk-of-bias score. The majority of these studies focused on neural networks (n = 9), multiple machine learning (n = 8), and random forests (n = 3). Remaining studies involved radiomics (n = 3), support vector machines (n = 3), and decision trees (n = 1). Purposes of ML included predominantly prediction of metastasis, detection of risk factors, prediction of survival, and prediction of postoperative complications. Other purposes were predictions of TNM staging, chemotherapy response, tumor resectability, and optimal therapy.

CONCLUSIONS

Machine Learning algorithms seem to contribute to the prediction of postoperative complications and the course of disease after upper gastrointestinal surgery for malignancies. However, due to the retrospective character of ML studies, these results require trials or prospective studies to validate this application of ML.

Establishment of a nomogram for predicting lymph node metastasis in patients with early gastric cancer after endoscopic submucosal dissection

Background

Endoscopic submucosal dissection (ESD) has been accepted as the standard treatment for the appropriate indication of early gastric cancer (EGC). Determining the risk of lymph node metastasis (LNM) is critical for the following treatment selection after ESD. This study aimed to develop a predictive model to quantify the probability of LNM in EGC to help minimize the invasive procedures.

Methods

A total of 952 patients with EGC who underwent radical gastrectomy were retrospectively reviewed. LASSO regression was used to help screen the potential risk factors. Multivariate logistic regression was used to establish a predictive nomogram, which was subjected to discrimination and calibration evaluation, bootstrapping internal validation, and decision curve analysis.

Results

Results of multivariate analyses revealed that gender, fecal occult blood test, CEA, CA19-9, histologic differentiation grade, lymphovascular invasion, depth of infiltration, and Ki67 labeling index were independent prognostic factors for LNM. The nomogram had good discriminatory performance, with a concordance index of 0.816 (95% CI 0.781–0.853). The validation dataset yielded a corrected concordance index of 0.805 (95% CI 0.770–0.842). High agreements between ideal curves and calibration curves were observed.

Conclusions

The nomogram is clinically useful for predicting LNM after ESD in EGC, which is beneficial to identifying patients who are at low risk for LNM and would benefit from avoiding an unnecessary gastrectomy.

Feasibility and Tolerance of Apatinib plus PD-1 Inhibitors for Previously Treated Advanced Gastric Cancer: A Real-World Exploratory Study

Background

Apatinib is established to be the standard of care as third-line therapy for patients with previously treated advanced gastric cancer (GC). Programmed cell death protein 1 (PD-1) blockades also exhibited promising efficacy and safety for patients with treatment-refractory advanced GC.

Objective

This study was to explore the feasibility and tolerance of apatinib plus PD-1 inhibitors for patients with previously treated advanced GC.

Methods

This study was performed as a real-world study; patients with advanced GC who were treated with previous systemic chemotherapy were screened retrospectively. Eligible patients were administered with apatinib combined with PD-1 blockade treatment. Efficacy of the patients was assessed with the change of target lesion using radiological evidence according to RECIST 1.1 criteria, and follow-up was carried out regularly. A safety profile was collected and documented during the combination treatment. Univariate analysis based on baseline characteristic subgroup was implemented in univariate analysis to identify the potential factor that might contribute to progression-free survival (PFS).

Results

Between August 2018 and October 2021, a total of 39 patients with advanced GC or gastroesophageal junction adenocarcinoma participated in this study consecutively and all the patients were available for efficacy and safety assessment. The best overall response during apatinib plus PD-1 blockade administration exhibited that PR was observed in 8 patients, SD was noted in 19 patients, and PD was found in 12 patients, which yielded an ORR of 20.5% (95% CI: 9.3%-36.5%), and DCR was 69.2% (95% CI: 52.4%-83.0%). Furthermore, the relatively enough follow-up had resulted in the mature PFS and overall survival (OS) data, suggesting that the median PFS of the 39 patients with advanced GC was 3.9 months (95% CI: 2.74-5.06). Additionally, the median OS of the 39 patients with advanced GC was 7.8 months (95% CI: 4.82-10.78). Furthermore, the most common adverse reactions of the 39 patients who received apatinib plus PD-1 blockades treatment were fatigue (61.5%), nausea and vomiting (56.4%), diarrhea (48.7%), hypertension (46.2%), hand-foot syndrome (38.5%), and rash (28.2%). Furthermore, performance status was independently associated with PFS of apatinib plus PD-1 inhibitor combination administration in baseline characteristic subgroup analysis.

Conclusion

Apatinib plus PD-1 inhibitors exhibited promising effectiveness and acceptable tolerance for previously treated advanced GC preliminarily. And this conclusion should be confirmed in clinical trials in the future.

External validation of the BEST-J score and a new risk prediction model for ESD delayed bleeding in patients with early gastric cancer

Background

Delayed bleeding is an important adverse event after gastric endoscopic submucosal dissection (ESD). We aimed to externally validate the Bleeding after ESD Trend from Japan (BEST-J) score and subsequently develop a risk prediction model for bleeding in Chinese patients with early gastric cancer (EGC) after ESD.

Methods

The clinical data of patients who underwent ESD for EGC in Beijing Friendship Hospital from June 2013 to December 2019 were collected retrospectively. The BEST-J score was evaluated according to the clinical data. Through univariate and multivariate logistic regression analyses of the clinical data, the factors affecting delayed bleeding were identified, and a new risk prediction model for bleeding was established. Receiver operating characteristic (ROC) curves were used to evaluate the predictive value of the two prediction models.

Results

A total of 444 patients with EGC undergoing ESD were included, of whom 27 patients had delayed bleeding (6.1%). Multivariate logistic regression analysis showed that a history of smoking (P = 0.029), tumor size > 20 mm (P = 0.022), intraoperative use of hemoclips (P = 0.025), resection of multiple tumors (P = 0.027), and prolongation of activated partial thromboplastin time (APTT) (P = 0.020) were independent influencing factors for delayed bleeding. ROC curve analysis showed that the areas under the curves (AUCs) were different between the BEST-J score and the newly built prediction model (0.624 vs. 0.749, P = 0.012).

Conclusions

The BEST-J score has moderately good discrimination for Chinese patients with EGC. However, for patients with EGC without severe comorbidities, the new risk prediction model may predict delayed bleeding better than the BEST-J score.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12876-022-02273-2.

Nomograms for Predicting the Lymph Node Metastasis in Early Gastric Cancer by Gender: A Retrospective Multicentric Study

Background

Lymph node metastasis (LNM) has a significant impact on the prognosis of patients with early gastric cancer (EGC). Our aim was to identify the independent risk factors for LNM and construct nomograms for male and female EGC patients, respectively.

Methods

Clinicopathological data of 1,742 EGC patients who underwent radical gastrectomy and lymphadenectomy in the First Affiliated Hospital, Second Affiliated Hospital, and Fourth Affiliated Hospital of Anhui Medical University between November 2011 and April 2021 were collected and analyzed retrospectively. Male and female patients from the First Affiliated Hospital of Anhui Medical University were assigned to training sets and then from the Second and Fourth Affiliated Hospitals of Anhui Medical University were enrolled in validation sets. Based on independent risk factors for LNM in male and female EGC patients from the training sets, the nomograms were established respectively, which was also verified by internal validation from the training sets and external validation from the validation sets.

Results

Tumor size (odd ratio (OR): 1.386, p = 0.030), depth of invasion (OR: 0.306, p = 0.001), Lauren type (OR: 2.816, p = 0.000), lymphovascular invasion (LVI) (OR: 0.160, p = 0.000), and menopause (OR: 0.296, p = 0.009) were independent risk factors for female EGC patients. For male EGC patients, tumor size (OR: 1.298, p = 0.007), depth of invasion (OR: 0.257, p = 0.000), tumor location (OR: 0.659, p = 0.002), WHO type (OR: 1.419, p = 0.001), Lauren type (OR: 3.099, p = 0.000), and LVI (OR: 0.131, p = 0.000) were independent risk factors. Moreover, nomograms were established to predict the risk of LNM for female and male EGC patients, respectively. The area under the ROC curve of nomograms for female and male training sets were 87.7% (95% confidence interval (CI): 0.8397–0.914) and 94.8% (95% CI: 0.9273–0.9695), respectively. For the validation set, they were 92.4% (95% CI: 0.7979–1) and 93.4% (95% CI: 0.8928–0.9755), respectively. Additionally, the calibration curves showed good agreements between the bias-corrected prediction and the ideal reference line for both training sets and validation sets in female and male EGC patients.

Conclusions

Nomograms based on risk factors for LNM in male and female EGC patients may provide new insights into the selection of appropriate treatment methods.

Oncologic Feasibility of Proximal Gastrectomy in Upper Third Advanced Gastric and Esophagogastric Junctional Cancer

Purpose

The aim of this study was to investigate the oncologic safety and identify potential candidates for proximal gastrectomy (PG) in upper third advanced gastric cancer (AGC) and esophagogastric junction (EGJ) cancers.

Materials and Methods

Among 5,665 patients who underwent gastrectomy for gastric adenocarcinoma between January 2011 and December 2017, 327 patients who underwent total gastrectomy with standard lymph node (LN) dissection for upper third AGC and Siewert type II EGJ cancers were enrolled. We analyzed the correlation between the metastatic rates of distal LNs (No. 4d, 5, 6, and 12a) around the lower part of the stomach and the clinicopathological characteristics. We identified subgroups with no metastasis to the distal LNs.

Results

The metastatic rate of distal LNs in proximal AGC and Siewert type II EGJ cancers was 7.0% (23 of 327 patients). On multivariate analysis, pathological T stage (P=0.001), tumor size (P=0.043), and middle third invasion (P=0.003) were significantly associated with distal LN metastases. Pathological ‘T2 stage’ (n=88), or ‘T3 stage with ≤5 cm tumor size’ (n=87) showed no metastasis in distal LNs, regardless of middle third invasion. Pathological T3 stage with tumor size > 5 cm (n=61) and T4 stage (n=91) had metastasis in the distal LNs.

Conclusions

In the upper third AGC and Siewert type II EGJ cancer, pathological T2 and small-sized T3 stage groups are possible candidates for PG in cases without distal LN metastasis. Further validation studies are required for clinical application.

Clinicopathological Features and Prognosis of Gastric Cancer in Young Patients

Gastric cancer mainly occurs in middle-aged and older people after their 50s, and the incidence of gastric cancer in younger people is rare. The frequency of gastric cancer that occurred before 40 years of age is 6-8% of all gastric cancer patients, and most of them are over 35 years of age. The prognosis of gastric cancer in young patients is believed to be poor because of more aggressive tumor behaviors and delayed diagnosis, however, there is controversy.