GPTs
I introduce my GPTs.
I introduce my GPTs.
The TNM Classification GPT is an innovative tool designed to assist users in determining the TNM (Tumor, Node, Metastasis) stage of cancer based on clinical or radiological information. By inputting relevant clinical or imaging findings, users can receive a structured TNM classification output—streamlining the process of oncologic staging.
Preliminary Assessment of TNM Classification Performance for Pancreatic Cancer in Japanese Radiology Reports using GPT-4. Japanese Journal of Radiology (2025)
This GPT is based on a five-level categorization system for COVID-19 pneumonia CT findings originally implemented by radiologists at Tokyo Women’s Medical University (TWMU). Retrospective validation was conducted using natural language processing (NLP), and the framework was later extended by replacing NLP with GPT technology. The current version, as of September 2025, operates on the GPT-5 engine. Although widespread vaccination against COVID-19 has made typical CT abnormalities of COVID-19 pneumonia far less common, this GPT remains as a commemoration of humanity’s successful effort to overcome the pandemic.
Category Classification for Lung Computed Tomography of COVID-19 by Natural Language Processing in Japanese Radiology Report, TWMU Journal, Vol. 7:109–114, 2023
Radiology DDx GPT is a specialized AI assistant designed to support radiologists in making accurate and efficient differential diagnoses. When given a description of imaging findings, the model generates the three most likely diagnoses and explains the reasoning behind each suggestion, highlighting the relevant imaging features and clinical context. To ensure the highest level of diagnostic accuracy, the system automatically translates non-English inputs into English for processing, and then translates the results back into the original language for the user. By combining clear explanations with language-aware processing, Radiology DDx GPT not only accelerates the diagnostic workflow but also enhances confidence, consistency, and learning in radiologic practice.
Radiologic diagnosis itself is a complex process that begins with detecting abnormalities on images, describing them with precise terminology, and then using these findings to infer the underlying pathology. Such work requires proper training, and must be carried out by certified radiologists. In addition, the use of GPTs for actual clinical decision-making is prohibited by policy; this tool is intended solely for academic research and educational demonstration purposes.