Table 17.
Overview of GPTs for medical imaging applications. The asterisks (*) indicate terms that are either not present in the original paper or do not apply in this context.
| References | ROI | Modality | Dataset | GPT version |
|---|---|---|---|---|
| Li et al (2023c) | * | x-ray | MIMIC-CXR, VQA-RAD, MS-CXR | GPT-4-V |
| Rau et al (2023) | * | * | Private case file, ACR protocol | GPT-3.5-Turbo, GPT-4 |
| Buckley et al (2023) | Skin, eyes, mouth | Radiographic images, ocular images, oral images | NEJM image challenge cases, NEJM clinicopathological conferences | GPT-4-V |
| Horiuchi et al (2024) | Rain, spine, head and neck | * | ‘Case of the Week’ by American Journal of Neuroradiology | GPT-4 |
| Wagner and Ertl-Wagner (2023) | Neuro, pediatrics, gastrointestinal, genitourinary, cardiac, musculoskeletal, breast, chest wall and pulmonary | * | 88 Questions | GPT-3 |
| Wu et al (2023) | 17 human body systems | x-ray, CT, MRI, PET, DSA, mammography, ultrasound, pathology | Online samples from 17 human body systems | GPT-4-V |
| Lim et al (2023) | Gastrointestinal tract | EGD, colonoscopy, ERCP, EUS | * | GPT-3 |
| Jeblick et al (2023) | Knee, head | CT, MRI | Fictitious radiology reports | GPT-3.5 |
| Barash et al (2023) | * | * | Institutional clinical notes | GPT-4 |
| (Son et al 2023) | Skeletal system | Bone Scan | Institutional bone scans | GPT-3.5 |
| (Jiang et al 2023) | Chest | x-ray | OPEN-I, MIMIC-CXR | GPT-3 |
| Li et al (2023a) | Abdomen, pelvis, neurological, musculoskeletal | x-ray, ultrasound, CT, MRI | Institutional data (400 radiology report) | GPT-3 |
| Rao et al (2023a) | Breast | * | * | GPT-3.5 |
| Rao et al (2023b) | Breast | * | * | GPT-3.5, GPT-4 |
| Guillen-Grima et al (2023) | * | * | Questions from spanish medical residency entrance examination (MIR) | GPT-3.5, GPT-4 |
| Rosen and Saban (2023) | * | * | Institutional 97 patient cases | GPT-3 |
| McCarthy et al (2023) | * | * | Material from society of interventional radiology (sir) patient center website | GPT-3 |
| Sorin et al (2023) | Eyes and ocular structures | Ocular imaging | 40 institutional ocular images | GPT-4-V |
| Chen et al (2023) | Lung | x-ray | Kaggle COVID-19 Lung x-ray dataset | GPT-4-V |
| Yan et al (2023b) | Fifteen object of interest | 11 modalities including microscopy, dermascope, x-ray, CT, ECHO, angiography, ultrasound, MRI, PET | VQA-RAD, PMC-VQA, PathVQA | Kaggle COVID-19 lung x-ray dataset |
| Li et al (2023a) | Chest | x-ray | Chest film database | GPT-4-V |
| Russe et al (2023) | Bone fracture | * | AOOTA fracture and dislocation classification compendium | GPT-3.5, GPT-4 |
| Elkassem and Smith (2023) | * | * | * | GPT-3.5 |
| Schmidt et al (2024) | Knee joint | MRI | * | GPT 3.5 |
| Kelly et al (2023) | Brain | MRI | Institutional dataset for multiple sclerosis | GPT-4-V |
| Naik et al (2023) | Breast | Mammography, ultrasound, MRI | Institutional data of bilateral breast cancer | GPT-3.5 |
| Lyu et al (2023) | Brain | CT, MRI | Atrium health wake forest baptist clinical database | GPT-3, GPT-4 |
| Nazario-Johnson et al (2023) | * | * | Presentation of a 65 year-old patient | GPT-3.5 |
| Doshi et al (2023) | * | MRI, CT, ultrasound, mammogram | * | GPT-3, GPT-4 |