10519 Background: An MCED test (Galleri, GRAIL, LLC, Menlo Park, CA) intended to complement recommended screening has been in clinical use since 04/2021. Here, we report RW clinical experience across age, sex, ordering site, and use case with the initial ~53,000 tests. Methods: This cell-free DNA-based MCED test uses a targeted methylation assay and a machine learning classification algorithm to detect a cancer signal (CS) and predict CS origin (CSO). This report includes tests returned from 04/20/2021 to 12/31/2022 on individuals aged ≥22 years (yrs) and excludes tests from clinical studies and sites limiting external data sharing, and repeat tests. Systematic collection of outcomes for cases with a “CS detected” (CSD) result was completed for a limited subset and continues via a rigorously controlled quality assurance (QA) program. Results: Tests were ordered and processed from across all US states (results returned, 98.9%; mean turnaround time, 6.7 business days). Among the 53,134 tests with results returned, CSD rate (CSDR) was was 1.0% (95% CI, 0.9-1.0; 510/53134), generally higher in males (1.1% [1.0-1.2; 313/29201]) vs females (0.8% [0.7-0.9; 197/23933]), and comparable to expected CSDR (males 1.07%; females 0.96%) as modeled based on MCED test performance and cancer incidence from SEER. CSDR increased with age (Table), which was a significant predictor of CSDR (p < 2e-13). In males and females, 67.4% and 61.9% of CSOs represented cancers without (w/o) recommended population screening, respectively. Early data from the QA program from an initial limited subset of CSD cases showed that a CSD result was associated with a diagnosis of invasive cancer across multiple cancers (eg, anus, breast, esophagus, head and neck, liver/bile duct, lymphoma, ovary, pancreas, plasma cell neoplasm, prostate, sarcoma), including stage I and II cancers. Conclusions: RW experience with the MCED test was consistent with previous large-scale clinical studies with an average CSDR of 1.0%, which increased with age. The test detected a CS and predicted CSO across multiple cancers, including early-stage cancers and cancers w/o recommended screening. This indicates that the MCED test can reliably detect a CS, which is essential to support population screening. Follow-up of CSD cases is ongoing through the QA program and will allow for future reporting of RW outcomes. [Table: see text]
Introduction: Liver cancer is the third leading cause of cancer-related deaths in the world, with liver cancer incidence and related deaths increasing in the United States. Although there is no guideline-recommending screening for liver cancer for the general population, the American Association for the Study of Liver Diseases recommends liver cancer surveillance in individuals with cirrhosis through ultrasound imaging, with or without alpha-fetoprotein (AFP) testing, every 6 months. A multi-cancer early detection (MCED) test was developed that detects a shared cancer signal from methylation patterns of cell-free DNA in blood. A ‘cancer signal detected’ result (positive result) is reported with 1 or 2 cancer signal origin (CSO) prediction(s). Here, an early-stage liver cancer case is presented to review the diagnostic pathway guided by a positive MCED test result. Case Description/Methods: A 62-year-old White male (BMI: 25.2 kg/m2) with chronic hepatitis B (low viral load since 9 years ago; no active therapy) underwent MCED testing. He reported moderate alcohol use (2-3 drinks/day) and that he formerly smoked over 10 years ago. He had a prostatectomy for prostate cancer 2 years prior to using the test. Three days after a blood sample was collected for analysis by the MCED test, computed tomography (CT) of the abdomen was performed as follow-up on a 1.9 cm indeterminate nodule without interval change for 1 year; CT of the abdomen showed a 1.9x1.9x1.9 cm left hepatic lobe nodule. Ten days after the CT scan, a positive MCED test result (top-CSO prediction=Liver/Bile duct; second-CSO prediction=Lung) was reported to the provider (Day 1). The result was communicated to the patient the next day (Day 2). The positive MCED test result resulted in aggressive evaluation of the mass. Biopsy of the liver nodule revealed hepatocellular carcinoma, moderately differentiated (Day 10). The cancer was subsequently classified as stage II. The mass was completely resected (Day 52). Prior to surgery, AFP level was 29.3 ng/mL (Day 13); after surgery, AFP level was 5.7 ng/mL (Day 66). The patient currently has no clinical symptoms (normal appetite and no weight loss, pain, or jaundice; Day 234). Discussion: The MCED test detected a cancer signal and predicted a liver CSO for an individual with stage II liver cancer. An indeterminate nodule was found in the patient 1 year prior to MCED test use with no resolution. Neither the CT scan results nor AFP level, which was lower than the AFP threshold (400 ng/mL) potentially indicative of hepatocellular carcinoma, would have led to this cancer being worked up without the MCED test. In conjunction with a positive MCED test result with a top-CSO prediction of liver, an aggressive workup took place resulting in diagnostic resolution within 1 month. Early-stage hepatocellular carcinoma has better prognosis and outcomes than late-stage disease. The use of the MCED test guided aggressive diagnostic workup and potentially improved outcomes for this patient. Citation Format: Martin Poliak. A case of stage II hepatocellular carcinoma diagnosed using a multi-cancer early detection test. [abstract]. In: Proceedings of the AACR Special Conference: Precision Prevention, Early Detection, and Interception of Cancer; 2022 Nov 17-19; Austin, TX. Philadelphia (PA): AACR; Can Prev Res 2023;16(1 Suppl): Abstract nr P060.
Research in Artificial Intelligence (AI) has focused mostly on two extremes: either on small improvements in narrow AI domains, or on universal theoretical frameworks which are often uncomputable, or lack practical implementations. In this paper we attempt to follow a big picture view while also providing a particular theory and its implementation to present a novel, purposely simple, and interpretable hierarchical architecture. This architecture incorporates the unsupervised learning of a model of the environment, learning the influence of one's own actions, model-based reinforcement learning, hierarchical planning, and symbolic/subsymbolic integration in general. The learned model is stored in the form of hierarchical representations which are increasingly more abstract, but can retain details when needed. We demonstrate the universality of the architecture by testing it on a series of diverse environments ranging from audio/visual compression to discrete and continuous action spaces, to learning disentangled representations. MotivationDespite the fact that strong AI capable of handling a diverse set of human-level tasks was envisioned decades ago, and there has been significant progress in developing AI for narrow tasks, we are still far away from having a single system which would be able to learn with efficiency and generality comparable to human beings or animals. While practical research has focused mostly on small improvements in narrow AI domains, research in the area of Artificial General Intelligence (AGI) has tended to focus on frameworks of truly general theories, like AIXI [1], Causal Entropic Forces [2], or PowerPlay [3]. These are usually uncomputable, incompatible with theories of biological intelligence, and/or lack practical implementations. Another class of algorithm that can be mentioned encompasses systems that are usually somewhere on the edge of cognitive architectures and adaptive general problem-solving systems. Examples of such systems are: the Non-Axiomatic Reasoning System [4], Growing Recursive Self-Improvers [5], recursive data compression architecture [6], OpenCog [7], Never-Ending Language Learning [8], Ikon Flux [9], MicroPsi [10], Lida [11] and many others [12]. These systems usually have a fixed structure with adaptive parts and are in some cases
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.
hi@scite.ai
10624 S. Eastern Ave., Ste. A-614
Henderson, NV 89052, USA
Copyright © 2024 scite LLC. All rights reserved.
Made with 💙 for researchers
Part of the Research Solutions Family.