Abstract
Objective: Artificial intelligence (AI) use and computational approaches have transformed the discovery of plant-derived bioactive chemicals for
cancer treatment. This review synthesized evidence on AI-based approaches, including network pharmacology, molecular docking, and machine
learning, for identifying and validating anticancer plant-derived phytochemicals, focusing on their molecular targets and signaling pathways. Data
Sources: PRISMA 2020 guidelines were followed. PubMed, Scopus, Web of Science, and Google Scholar were searched between October 2020 and
July 2025. Study Selection: Studies with at least one AI or computational method, investigating plant-derived phytochemicals for anticancer effects,
reporting quantitative computational outcomes, and offering free full-text access in English were included. Review articles, conference abstracts,
non-plant studies, and non-English publications were excluded. Data Extraction: Two reviewers extracted data on predefined variables. An
adapted PROBAST tool was used to assess risk of bias for in silico studies and modified SYRCLE tools for in vitro and in vivo studies, with
certainty of evidence evaluated using the GRADE method. Data Synthesis: Eighteen studies satisfied the inclusion criteria. The most frequently
discovered bioactive chemical was quercetin, which targets breast cancer, hepatocellular carcinoma (HCC), and lung cancer via AKT1, EGFR, TP53,
and MAPK signaling pathways. The majority of in silico studies have a high risk of bias, while in vitro and in vivo studies were rated as having
some concerns. The overall certainty of evidence was low. Conclusion: AI-based computational techniques have become critical tools for
discovering plant-derived compounds and identifying multi-target mechanisms involving AKT1, EGFR, and PI3K/Akt signaling.