Search

PurposeReal world evidence is crucial to understanding the diffusion of new oncologic therapies, monitoring cancer outcomes, and detecting unexpected toxicities. In practice, real world evidence is challenging to collect rapidly and comprehensively, often requiring expensive and time-consuming manual case-finding and annotation of clinical text. In this Review, we summarise recent developments in the use of artificial intelligence to collect and analyze real world evidence in oncology.MethodsWe performed a narrative review of the major current trends and recent literature in artificial intelligence applications in oncology.ResultsArtificial intelligence (AI) approaches are increasingly used to efficiently phenotype patients and tumors at large scale. These tools also may provide novel biological insights and improve risk prediction through multimodal integration of radiographic, pathological, and genomic datasets. Custom language processing pipelines and large language models hold great promise for clinical prediction and phenotyping.ConclusionsDespite rapid advances, continued progress in computation, generalizability, interpretability, and reliability as well as prospective validation are needed to integrate AI approaches into routine clinical care and real-time monitoring of novel therapies.

The African claw-toed frog, Xenopus laevis, is a well-established laboratory model for the biology of vertebrate oogenesis, fertilisation, and development at embryonic, larval, and metamorphic stages. For ovulation, X. laevis females are usually injected with chorionic gonadotropin, whereupon they lay typically hundreds to thousands of eggs in a day. After being rested for a minimum of three months, animals are re-used. The literature suggests that adult females can lay much larger numbers of eggs in a short period. Here, we compared the standard single ovulation protocol with a double ovulation protocol, in which females were ovulated, then re-ovulated after seven days and then rested for three months before re-use. We quantified egg number, fertilisation rate (development to cleavage stage), and corticosterone secretion rate as a measure of stress response for the two protocol groups over seven 3-month cycles. We found no differences in egg number-per-ovulation or egg quality between the groups and no long-term changes in any measures over the 21-month trial period. Corticosterone secretion was elevated by ovulation, similarly for the single ovulation as for the first ovulation in the double-ovulation protocol, but more highly for the second ovulation (to a level comparable to that seen following shipment) in the latter. However, both groups exhibited the same baseline secretion rates by the time of the subsequent cycle. Double ovulation is thus transiently more stressful/demanding than single ovulation but within the levels routinely experienced by laboratory X. laevis. Noting that stress hormone corticosterone/cortisol secretion is linked to physiological processes, such as ovulation, that are not necessarily harmful to the individual, we suggest that the benefits of a doubling in egg yield-per-cycle per animal without loss of egg quality or signs of acute or long-term harm may outweigh the relatively modest and transient corticosterone elevation we observed. The double ovulation protocol therefore represents a potential new standard practice for promoting the 3Rs (animal use reduction, refinement and replacement) mission for Xenopus research.

BACKGROUND: Studies that examined sex differences in first-episode patients consistently show that males compared to females have poor premorbid adjustment, earlier age of onset, worse clinical characteristics, and poorer outcomes. However, little is known about potential mediators that could explain these sex differences. METHODS: Our sample consisted of 137 individuals with first episode schizophrenia (males, n = 105; 77%) with a mean age of 22.1(s.d. = 4.1) years and mean education of 12.5(s.d. = 1.7) years. At entry, patients were within 2 years of their first psychotic episode onset. Baseline assessments were conducted for premorbid adjustment, symptoms, cognitive functioning, insight, and at 6-months for role and social functioning. RESULTS: Males as compared to females had poorer premorbid adjustment across several key developmental periods (p < 0.01), an earlier age of onset [M = 20.3(3.3) v. 22.8(5.6), p = 0.002], more negative symptoms (p = 0.044), poorer insight (p = 0.031), and poorer baseline and 6-month role (p = 0.002) and social functioning (p = 0.034). Several of these variables in which males showed impairment were significant predictors of 6-month role and social functioning. Premorbid adjustment and insight mediated the relationship between sex and role and social functioning at 6-months, but not negative symptoms. DISCUSSION: Males compared to females were at lower levels across several key premorbid and clinical domains which are strongly associated with functional outcome supporting the hypothesis that males might have a more disabling form of schizophrenia. The relationship between sex with role and social functioning was mediated through premorbid adjustment and insight suggesting pathways for understanding why females might have a less disabling form of schizophrenia.