Why MRIninja
Why MRIninja
The Idea Behind MRI.NINJA
MRI.NINJA was created from the founder’s desire to remain constantly updated on global trends in MRI examination protocols and to promote a more evidence-based approach to protocol design.
Historically, MRI protocol planning has often been affected by significant variability. Different national traditions, local schools of thought, departmental habits, vendor-specific sequence implementations, and differences in magnetic field strength have all contributed to marked heterogeneity in how similar examinations are performed across institutions. As a result, standardisation has frequently been limited and confusion has often persisted.
The project was born from the intention to study these topics in depth, to support professional growth, and to help organise MRI services on the basis of scientific evidence rather than routine habits or informal assumptions. The goal was to build a system capable of combining completeness, transparency, and clinical justification.
Years ago, this process would have required extensive manual review of the scientific literature to identify the most reliable and comprehensive sources describing recommended MRI protocols. Modern AI tools now make it possible to accelerate the most time-consuming phases of this work.
MRI.NINJA therefore uses a structured multi-step methodology based on strict research instructions provided to advanced AI systems, with emphasis on high-level specialist literature rather than generic public-facing sources. Content is organised into predefined sections and weighted according to Evidence-Based Medicine principles, so that higher-quality sources have greater influence than lower-level evidence.
For this reason, MRI.NINJA pages are not simple copy-and-paste AI answers to generic prompts such as “What is the standard lumbar spine MRI protocol?”. They are the result of a controlled synthesis process designed to maximise relevance, depth, and scientific consistency.
Another core objective is long-term maintainability. The platform is being developed so that pages can be efficiently updated over time as new evidence, technologies, and clinical practices evolve.
MRI.NINJA is a non-profit, vendor-independent project and is not supported by any MRI vendor. Donations help cover the costs of hosting, maintenance, technical development and long-term preservation of the platform, without influencing editorial decisions or protocol recommendations.
The added value remains human expert review. Everything that is published is intended to be re-read, checked, refined, and corrected by experienced professionals in the field in order to minimise possible AI-related inaccuracies. Because this review process requires time, some pages may temporarily remain in a pre-verification state; when this occurs, it is clearly indicated within the page itself.
MRI.NINJA currently focuses on clinical MRI protocols performed on 1.5T and 3T systems. Low-field and mid-field MRI systems are outside the present scope of the project.
Why images are intentionally limited
Unlike traditional radiology resources, MRIninja intentionally uses very few images. This is a deliberate editorial choice, but it also reflects the practical limits of the project.
MRIninja content is primarily built through structured, high-depth, guided research workflows designed to extract, compare, and organise technical and clinical knowledge from scientific and professional sources. These workflows are focused on evidence synthesis, protocol logic, acquisition strategy, and decision-making principles. They do not inherently generate or include diagnostic imaging material.
Although images could theoretically be added afterwards, this would require manual case selection, validation, anonymisation, image preparation, annotation, and contextual explanation for each example. That process is extremely time-consuming and is not currently compatible with the scale and available resources of the project.
For this reason, MRIninja prioritises text-based, high-density technical and clinical explanations over image collections. The goal is to train structured reasoning, protocol awareness, and technical decision-making rather than to reproduce a conventional radiology atlas.
Images may be included in the future when they provide clear added value and when they can be selected, anonymised, and contextualised with adequate quality control.
Core Principles
No universal MRI protocol exists. Every examination must be adapted to the clinical question, the patient, the scanner, the available time and the local workflow.
Curated content matters. MRIninja is designed as a controlled knowledge base rather than a generic answer generator.
Evidence orientation is mandatory. Each protocol is linked to references and a transparent estimate of the strength of evidence.
Practical workflow is part of quality. Recommendations are written for real departments where patient cooperation, scanner availability and acquisition time matter.
Vendor-neutral language improves clarity. Sequence names are standardised and vendor equivalents are kept in the References section.
Scenario-based protocols are more useful than generic templates. Master protocols provide the base; child protocols document what changes for specific clinical indications.
Updates must be visible. Changelog and last-updated dates make content evolution transparent.
Searchable keywords help navigation. Keywords connect related anatomical, technical and clinical topics across the site.
Human review is essential. AI can accelerate literature discovery, but trust requires expert review and clinical judgement.
Continuous improvement is the method. MRIninja is expected to evolve as guidelines, techniques and evidence change.
Contributors
Andrea Forneris — Founder and clinical direction.
AI-assisted editorial workflow under expert human supervision.