RESUMO
RATIONALE AND OBJECTIVES: Identifying intrahepatic cholangiocarcinoma (iCCA) patients who are at high risk for early recurrence (ER) can guide personalized treatment strategy and improve survival. This study aimed to investigate the value of preoperative MRI, especially diffusion-weighted imaging, in predicting ER, including in patients receiving neoadjuvant therapy. MATERIALS AND METHODS: This study included 175 pathologically-confirmed iCCA patients who underwent curative resection (114 men, 61 women; mean age 59.0 ± 9.56 years). MRI features, particularly apparent diffusion coefficient (ADC), were analyzed and compared between ER and non-ER cases. Survival analyses of ER were evaluated using Cox regression and Kaplan-Meier analysis. RESULTS: ER occurred in 54.3% (95/175) of patients. Multivariate logistic regression analysis identified tumor ADC as the only independent predictor of ER (odds ratio = 0.034, P < 0.001), with AUCs of 0.758 (95%CI 0.664, 0.836) in the testing cohort and 0.779 (95%CI 0.622, 0.893) in the validation cohort. The optimal ADC threshold was 1.273 × 10-3 mm2/s. Tumor ADC was comparable to the AJCC 8th staging system in predicting ER (AUC 0.758 vs 0.650 in testing cohort and 0.779 vs 0.661 in validation cohort). Multivariate Cox analysis identified high tumor burden score (HR = 1.109, P = 0.009), non-smooth margin (HR = 2.265, P = 0.008) and tumor ADC (HR = 0.111, P < 0.001) as independent risk factors for ER. Lower ADC values were linked to shorter RFS in both testing and validation cohorts (P < 0.001 and 0.0219), as well as in patients receiving neoadjuvant therapy (P = 0.003). CONCLUSION: Preoperative MRI, particularly ADC, can help predict ER in iCCA, regardless of the application of neoadjuvant therapy, comparable to the AJCC 8th staging system.
RESUMO
MOTIVATION: Accurate ADMET (an abbreviation for 'absorption, distribution, metabolism, excretion and toxicity') predictions can efficiently screen out undesirable drug candidates in the early stage of drug discovery. In recent years, multiple comprehensive ADMET systems that adopt advanced machine learning models have been developed, providing services to estimate multiple endpoints. However, those ADMET systems usually suffer from weak extrapolation ability. First, due to the lack of labelled data for each endpoint, typical machine learning models perform frail for the molecules with unobserved scaffolds. Second, most systems only provide fixed built-in endpoints and cannot be customized to satisfy various research requirements. To this end, we develop a robust and endpoint extensible ADMET system, HelixADMET (H-ADMET). H-ADMET incorporates the concept of self-supervised learning to produce a robust pre-trained model. The model is then fine-tuned with a multi-task and multi-stage framework to transfer knowledge between ADMET endpoints, auxiliary tasks and self-supervised tasks. RESULTS: Our results demonstrate that H-ADMET achieves an overall improvement of 4%, compared with existing ADMET systems on comparable endpoints. Additionally, the pre-trained model provided by H-ADMET can be fine-tuned to generate new and customized ADMET endpoints, meeting various demands of drug research and development requirements. AVAILABILITY AND IMPLEMENTATION: H-ADMET is freely accessible at https://paddlehelix.baidu.com/app/drug/admet/train. SUPPLEMENTARY INFORMATION: Supplementary data are available at Bioinformatics online.