Artificial intelligence in oncology: Current status and possibilities (Review)

TL;DR

This review explores how AI is transforming cancer care through improved diagnosis, prognosis, and treatment, while highlighting challenges and future directions.

Contribution

A comprehensive synthesis of current AI applications in oncology and identification of barriers to clinical adoption.

Findings

AI models like CNNs and transformers show expert-level performance in lesion detection and survival prediction.

Translation of AI into clinical practice is hindered by dataset bias, lack of standardization, and regulatory issues.

Emerging AI techniques offer potential solutions to current limitations in generalizability and interpretability.

Abstract

Artificial intelligence (AI) is increasingly reshaping oncology by enhancing diagnostic accuracy, improving prognostication and enabling personalized treatment planning. The present review aimed to critically synthesize the contemporary landscape of AI applications across cancer imaging, digital pathology, clinical outcome prediction, chemotherapy and radiotherapy. Recent advances in machine learning and deep learning, particularly convolutional neural networks and transformer-based architectures, have demonstrated robust performance in lesion detection, tumour grading, survival prediction and treatment optimization, in several instances approaching or exceeding expert-level accuracy. Despite these advances, translation into routine clinical practice remains limited due to dataset bias, limited generalizability, the lack of standardized data protocols, insufficient interpretability and regulatory barriers. Ethical challenges related to fairness, transparency and equitable access are especially relevant in low- and middle-income countries. Emerging frontiers, including multimodal AI, foundation models, federated learning, and explainable AI, provide potential solutions to these challenges. Multidisciplinary collaboration, rigorous prospective validation and robust ethical governance will be essential to realize the full potential of AI in advancing precision oncology and improving global cancer outcomes.

Figures