Quality-adjusted time without symptom or toxicity and quality-adjusted progression-free survival of first-line maintenance niraparib in patients with advanced ovarian cancer

Le résumé fourni par la source

Objective: We aimed to develop an artificial intelligence (AI)based comprehensive serum glycopeptide spectra analysis (CSGSA-AI) method in combination with convolutional neural network (CNN) to detect aberrant glycans in serum samples of patients with epithelium ovarian cancer (EOC).Methods: A total of 97 serum samples were collected from patients with early-stage EOC at the time of ovarian mass detection prior to the initiation of any treatment (stage I).The non-EOC control group (n=254) comprised both healthy women (n=220) and women with gynecologic benign diseases (n=34).We used AlexNet, the latest CNN-based technology, as a discrimination tool for CSGSA to identify early-stage EOC.To facilitate CNN training, we converted numerical data of glycopeptide expression to 2D barcode images and let CNN learn and distinguish early-stage EOC.CNN was trained using 60% samples and validated using 40% samples.To further enhance the learning efficacy and diagnostic performance of CNN, we added cancer antigen 125 (CA125) and HE4 information into the 2D barcode by changing the color (multicolored model).Results: The sensitivity, specificity, positive predictive value, negative predictive value and area under the curve (AUC) of CSGSA-AI were 79%, 96%, 89%, 92%, and 88%, respectively.When CNN was trained with 2D barcodes colored on the basis of serum levels of CA125 and HE4 (multicolored model), AUC of 95% was achieved.Conclusion: CSGSA-AI has the potential to be a useful tool for diagnosis of early-stage epithelial ovarian cancer. Oral (OO8) Epithelial Ovarian Cancer including Borderline Tumor

Ce résumé expose les affirmations des auteurs. BNTIC ne l’interprète pas comme une validation indépendante des résultats.