Friday, May 17, 2024

Journal Club: Nature Reviews Cancer: AI for Cancer Researchers (Perez-Lopez 2024)


[AI Corner - let's see how AI would summarize this paper!]

"A Guide to Artificial Intelligence for Cancer Researchers," by Raquel Perez-Lopez et al., published in Nature Reviews Cancer, provides a comprehensive overview of how AI can be leveraged in cancer research. It emphasizes the practical applications of AI, aiming to demystify its usage for non-computational cancer researchers. The article covers several key areas, including AI for image analysis, natural language processing (NLP), and drug discovery.

Key Points:

AI's Evolution in Cancer Research:

AI has transitioned from a specialized tool to a widely accessible resource that enhances research productivity and enables new discoveries.

Researchers can use off-the-shelf software for basic applications or develop custom AI solutions for more advanced needs.

Understanding Deep Learning:

The article explains three types of deep learning: supervised, unsupervised, and reinforcement learning.

Supervised learning involves training models on labeled data, while unsupervised learning finds patterns in unlabeled data. Reinforcement learning involves agents learning optimal behaviors through interaction with their environment.

AI in Biomedical Image Analysis:

Deep learning has revolutionized image analysis, with convolutional neural networks (CNNs) and transformers being central to advancements.

AI tools can detect tumors in medical images, analyze genetic data, and assist in histopathology by distinguishing between benign and malignant cells.

AI for Language:

Large language models (LLMs) like GPT-4 have advanced NLP, enabling tasks such as summarizing text, translating, and generating new text.

These models are used in cancer research to process clinical notes, research papers, and generate structured data from unstructured text.

Emerging Uses of AI:

AI is increasingly applied in drug discovery, predicting protein structures, and identifying potential drug candidates.

Multimodal AI, which integrates various data types (images, text, genomic data), is becoming a significant trend, offering more comprehensive analysis capabilities.

Challenges and Future Directions:

The article highlights the need for better data integration, explainable AI models, and standards for data sharing.

Collaboration between AI researchers and clinicians is crucial for developing clinically relevant AI tools.


This reveiw will be a valuable resource for cancer researchers looking to incorporate AI into their work. It provides practical guidance on how to start using AI tools, the types of AI applications in cancer research, and the future potential of AI in this field. Researchers may find the explanations of deep learning and the practical examples of AI applications particularly helpful.