Can ChatGPT Analyze Lung Cancer?

Christopher Collins
3 min readMay 27, 2024
ChatGPT analyses lung scan image

The rapid advancements in artificial intelligence (AI) have opened up numerous possibilities in the medical field, especially in the early detection and diagnosis of diseases such as lung cancer. Among these AI tools, ChatGPT, developed by OpenAI, is primarily known for its language processing capabilities. However this experiment shows that ChatGPT can also be surprisingly accurate in image categorization tasks when given a structured dataset.

Experiment Overview

To evaluate ChatGPT’s capability in analyzing lung cancer, I conducted an experiment using a dataset from Kaggle, specifically the “Chest CTScan Images” dataset by Mohamed Hany.

This dataset contains CT scan images categorized into four types: adenocarcinoma, large cell carcinoma, squamous cell carcinoma, and normal.

The goal of the experiment was to see if ChatGPT could correctly classify new CT scan images based on the training data provided.

Dataset Preparation

Downloading and Extracting Data: The dataset was downloaded from Kaggle and extracted to my local machine. Within the extracted data, there was a folder named “Test” containing four subfolders representing different categories of lung cancer and normal tissues:

--

--