medgemma-4b-csrrg-impression

medgemma-4b: Contextualized Structured Radiology Report Generation (Impression)

This model is a fine-tuned version of google/medgemma-4b-it for generating the IMPRESSION section of contextualized structured chest X-ray radiology reports. It was trained using LoRA (Low-Rank Adaptation) on the csrrg_ift_dataset containing instruction-following examples from MIMIC-CXR and CheXpert+ datasets.

Model Description

This model performs Contextualized Structured Radiology Report Generation (CSRRG) for chest X-rays, generating concise impression sections with rich clinical context including patient history, imaging technique, comparison to prior studies, and temporal reasoning.

Key characteristics:

• Generates the IMPRESSION section of radiology reports
• Incorporates clinical history/indication, technique, and comparison to prior studies
• Performs temporal reasoning across multiple examinations
• Produces clinically relevant summaries with contextual awareness
• Fine-tuned with LoRA for parameter-efficient adaptation

Intended Use

Primary Use Cases

• Research on contextualized radiology report generation
• Development of temporal reasoning systems for medical imaging
• Clinical decision support with longitudinal patient data
• Medical AI and multimodal model research
• Educational tools for radiology training

Intended Users

• Medical AI researchers
• Healthcare technology developers
• Clinical informatics specialists
• Radiology departments (research use only)

Out-of-Scope Use

• NOT intended for clinical diagnosis without physician review
• Should not replace human radiologists in clinical practice
• Requires validation before any clinical deployment

Training Details

Training Data

• Dataset: csrrg_ift_dataset (csrrg_ift_dataset_impression subset)
• Training samples: ~405,971 instruction-following examples
• Data sources: MIMIC-CXR and CheXpert+ chest X-ray datasets
• Task format: Instruction fine-tuning with rich clinical context
• Context includes: Clinical history/indication, imaging technique, comparison to prior studies, current and prior images

Training Procedure

Fine-tuning method: LoRA (Low-Rank Adaptation)

LoRA Configuration:

• Rank (r): 32
• Alpha: 64
• Dropout: 0.1
• Target modules: q_proj, k_proj, v_proj, o_proj, gate_proj, up_proj, down_proj

Training hyperparameters:

• Learning rate: 2e-4
• Batch size: 4 per device
• Gradient accumulation steps: 32 (effective batch size: 128)
• Epochs: 1
• Optimizer: AdamW
• Learning rate scheduler: Cosine with 3% warmup
• Precision: bfloat16
• Attention implementation: Flash Attention 2
• Max sequence length: 2048
• Max images per sample: 3

Hardware:

• GPU: NVIDIA H100
• Training framework: HuggingFace Transformers + PEFT

Usage

Loading the Model

from transformers import AutoProcessor, AutoModelForVision2Seq
from PIL import Image
import torch

# Load model and processor
model_name = "erjui/medgemma-4b-csrrg-impression"
model = AutoModelForVision2Seq.from_pretrained(
    model_name,
    trust_remote_code=True,
    torch_dtype=torch.bfloat16,
    device_map="auto"
)
processor = AutoProcessor.from_pretrained("google/medgemma-4b-it", trust_remote_code=True)

# Load chest X-ray images (current and prior studies)
# CSRRG models support multiple images for temporal comparison (max_images_per_sample: 3)
current_image = Image.open("current_xray.jpg")

# Prepare input with clinical context
messages = [
    {
        "role": "system",
        "content": [{"type": "text", "text": "You are an expert radiologist."}]
    },
    {
        "role": "user",
        "content": [
            {
                "type": "text",
                "text": """Analyze the chest X-ray images and write the IMPRESSION section of a radiology report. Provide a concise clinical summary and diagnosis based on the imaging findings. Consider the available clinical contexts when formulating your impression.

=== CLINICAL HISTORY/INDICATION ===
Male patient with leukocytosis and fever, query pneumonia.

=== TECHNIQUE ===
Portable anteroposterior chest radiograph.

=== COMPARISON ===
None.

=== CURRENT IMAGES ==="""
            },
            {"type": "image"}  # Current image (supports multiple images for temporal comparison)
        ]
    }
]

# Process and generate
inputs = processor(images=current_image, text=messages, return_tensors="pt").to(model.device)
outputs = model.generate(**inputs, max_new_tokens=256)
generated_text = processor.decode(outputs[0], skip_special_tokens=True)

print(generated_text)

Expected Output Format

IMPRESSION:
1. Right apical rounded opacity concerning for infection or malignancy.
2. Recommend repeat dedicated AP and lateral chest radiograph, or CT for further evaluation.

Citation

If you use this model, please cite:

@article{kang2025automated,
  title={Automated Structured Radiology Report Generation with Rich Clinical Context},
  author={Kang, Seongjae and Lee, Dong Bok and Jung, Juho and Kim, Dongseop and Kim, Won Hwa and Joo, Sunghoon},
  journal={arXiv preprint arXiv:2510.00428},
  year={2025}
}

Also cite the base model:

@article{sellergren2025medgemma,
  title={Medgemma technical report},
  author={Sellergren, Andrew and Kazemzadeh, Sahar and Jaroensri, Tiam and Kiraly, Atilla and Traverse, Madeleine and Kohlberger, Timo and Xu, Shawn and Jamil, Fayaz and Hughes, C{\'\i}an and Lau, Charles and others},
  journal={arXiv preprint arXiv:2507.05201},
  year={2025}
}

Model Card Authors

Seongjae Kang (erjui)

Model Card Contact

For questions or issues, please open an issue on the model repository.