dicom-rs-transformer: Enterprise-Grade DICOM Transformation & Compliance Engine

Blazingly fast, memory-safe DICOM dataset transformation, anonymization, and extraction powered by Rust and dicom-rs.

Executive Overview

dicom-rs-transformer is an open-source Rust library and Model Context Protocol (MCP)-enabled CLI designed for healthcare data engineering teams, AI researchers, and medical imaging systems developers.

It simplifies batch DICOM data processing, patient identity de-identification, metadata extraction, and multi-cloud storage pipelines—delivering sub-millisecond per-file transformation speeds with complete memory safety.

Key Product Highlights

⚡ Blazingly Fast & Lightweight

Built on the Rust dicom-rs ecosystem (v0.10.0). Zero garbage collection overhead, minimal CPU footprint, and instant startup times for batch processing millions of medical images.

🛡️ Flexible PHI De-Identification & Anonymization Pipelines

De-identify patient datasets and build DICOM PS 3.15 Annex E compliant confidentiality pipelines. Strip Protected Health Information (PHI), generate cryptographically secure pseudo-IDs, re-key UIDs, and export structured JSON audit maps linking original datasets to anonymized output.

📜 Dual Specification Engine (JSON DSL & Line Scripts)

Define complex transformation rules using declarative JSON DSL for web APIs and microservices, or human-readable Line-by-Line Script Language designed for terminal scripts, batch execution, and interactive REPL consoles.

🤖 Model Context Protocol (MCP) Ready

Integrates directly with AI developer tools (Google Antigravity IDE, Cursor, Claude Desktop) via native MCP discovery (dicom-transformer install-mcp). Allows AI agents to programmatically inspect, anonymize, and manipulate DICOM datasets.

📦 Multi-Format & Stream Support

Native automatic header detection supports both DICOM Part-10 file format (with 128-byte preamble and "DICM" magic headers) and raw DICOM stream datasets (headerless Little Endian explicit/implicit VR streams).

Features at a Glance

Community vs. Enterprise PRO Edition

Code Examples

1. Line Script Example (anonymize_script.txt)

# Batch DICOM Anonymization & Extraction Pipeline
ANONYMIZE NAME="ANON^PATIENT" ID="ANON-ID-1234"
SET InstitutionName "CLINICAL RESEARCH ORG"
SAVE target/output/anonymized/
SAVE_JSON target/output/json/
EXTRACT_PIXELS target/output/pixels/ png
SAVE_MAP target/output/maps/
EXECUTE

2. Rust API Usage

use dicom_rs_transformer::{Action, DicomTransformer, TagSelector, TransformSpec};
use dicom_object::open_file;

fn main() -> Result<(), Box<dyn std::error::Error>> {
    // Build specification
    let mut spec = TransformSpec::new();
    spec.add_action(Action::SetTag {
        selector: TagSelector::Keyword("PatientName".to_string()),
        value: "ANONYMOUS^PATIENT".to_string(),
    });
    spec.add_action(Action::RemoveTag {
        selector: TagSelector::Keyword("PatientAddress".to_string()),
    });

    // Execute transformation
    let transformer = DicomTransformer::new(spec);
    let mut dicom_obj = open_file("input.dcm")?;
    let report = transformer.transform_file(&mut dicom_obj)?;

    println!("Transformed {} tags in {}ms", report.tags_modified, report.duration_ms);
    dicom_obj.write_to_file("output.dcm")?;

    Ok(())
}

Get Started & Enterprise Enquiries

• 🐙 GitHub Repository: https://github.com/gosmart-health/dicom-rs-transformer
• 🌐 Company Website: https://www.gosmart.health
• 📧 Enterprise Licensing & PRO Subscriptions: [email protected]