Biomarkers discovery for endometrial cancer: A graph convolutional sample network method

Integrated machine learning identifies a cellular senescence-related prognostic model to improve outcomes in uterine corpus endometrial carcinoma

Background

Uterine Corpus Endometrial Carcinoma (UCEC) stands as one of the prevalent malignancies impacting women globally. Given its heterogeneous nature, personalized therapeutic approaches are increasingly significant for optimizing patient outcomes. This study investigated the prognostic potential of cellular senescence genes(CSGs) in UCEC, utilizing machine learning techniques integrated with large-scale genomic data.

Methods

A comprehensive analysis was conducted using transcriptomic and clinical data from 579 endometrial cancer patients sourced from the Cancer Genome Atlas (TCGA). A subset of 503 CSGs was assessed through weighted gene co-expression network analysis (WGCNA) alongside machine learning algorithms, including Gaussian Mixture Model (GMM), support vector machine - recursive feature elimination (SVM-RFE), Random Forest, and eXtreme Gradient Boosting (XGBoost), to identify key differentially expressed cellular senescence genes. These genes underwent further analysis to construct a prognostic model.

Results

Our analysis revealed two distinct molecular clusters of UCEC with significant differences in tumor microenvironment and survival outcomes. Utilizing cellular senescence genes, a prognostic model effectively stratified patients into high-risk and low-risk categories. Patients in the high-risk group exhibited compromised overall survival and presented distinct molecular and immune profiles indicative of tumor progression. Crucially, the prognostic model demonstrated robust predictive performance and underwent validation in an independent patient cohort.

Conclusion

The study emphasized the significance of cellular senescence genes in UCEC progression and underscored the efficacy of machine learning in developing reliable prognostic models. Our findings suggested that targeting cellular senescence holds promise as a strategy in personalized UCEC treatment, thus warranting further clinical investigation.

ECPC-IDS: A benchmark endometrial cancer PET/CT image dataset for evaluation of semantic segmentation and detection of hypermetabolic regions

BACKGROUND

Endometrial cancer is one of the most common tumors in the female reproductive system and is the third most common gynecological malignancy that causes death after ovarian and cervical cancer. Early diagnosis can significantly improve the 5-year survival rate of patients. With the development of artificial intelligence, computer-assisted diagnosis plays an increasingly important role in improving the accuracy and objectivity of diagnosis and reducing the workload of doctors. However, the absence of publicly available image datasets restricts the application of computer-assisted diagnostic techniques.

METHODS

In this paper, a publicly available Endometrial Cancer PET/CT Image Dataset for Evaluation of Semantic Segmentation and Detection of Hypermetabolic Regions (ECPC-IDS) are published. Specifically, the segmentation section includes PET and CT images, with 7159 images in multiple formats totally. In order to prove the effectiveness of segmentation on ECPC-IDS, six deep learning semantic segmentation methods are selected to test the image segmentation task. The object detection section also includes PET and CT images, with 3579 images and XML files with annotation information totally. Eight deep learning methods are selected for experiments on the detection task.

RESULTS

This study is conduct using deep learning-based semantic segmentation and object detection methods to demonstrate the distinguishability on ECPC-IDS. From a separate perspective, the minimum and maximum values of Dice on PET images are 0.546 and 0.743, respectively. The minimum and maximum values of Dice on CT images are 0.012 and 0.510, respectively. The target detection section's maximum mAP values on PET and CT images are 0.993 and 0.986, respectively.

CONCLUSION

As far as we know, this is the first publicly available dataset of endometrial cancer with a large number of multi-modality images. ECPC-IDS can assist researchers in exploring new algorithms to enhance computer-assisted diagnosis, benefiting both clinical doctors and patients. ECPC-IDS is also freely published for non-commercial at: https://figshare.com/articles/dataset/ECPC-IDS/23808258.