Perfusion reflects how blood reaches tissue and supports cellular function, making it clinically relevant for diagnosis, staging and monitoring across many conditions. Magnetic resonance imaging (MRI) offers several ways to assess perfusion, providing insight into vascularity, microcirculation and treatment effects. Despite these capabilities, routine clinical use remains inconsistent, particularly when moving beyond visual interpretation to quantitative measurements that could function as imaging biomarkers. Variation in acquisition protocols, analysis methods and reporting practices continues to limit reproducibility and confidence, especially across sites and over time. Practice guidance by the European Society for Magnetic Resonance in Medicine and Biology sets out a structured approach to selecting techniques, optimising protocols and integrating perfusion MRI into routine workflows.
Four Established Techniques Define Current Practice
Perfusion MRI in clinical settings is primarily based on four techniques: dynamic susceptibility contrast (DSC) MRI, dynamic contrast-enhanced (DCE) MRI, arterial spin labelling (ASL) MRI and intravoxel incoherent motion (IVIM) MRI. Two methods require intravenous contrast agents, while two offer non-contrast alternatives. Each technique provides different information and faces specific barriers that influence adoption.
DSC-MRI measures signal changes during the passage of a contrast bolus and is most widely used in brain imaging. It supports semi-quantitative assessment and, with additional modelling, more quantitative parameters derived using an arterial input function. Clinically, DSC-MRI is applied in brain tumour evaluation, including grading, monitoring response and differentiating progression patterns. It also plays a role in acute stroke imaging, where perfusion data contribute to identifying tissue at risk.
DCE-MRI relies on dynamic T1-weighted imaging during and after contrast administration. In routine care it is often used qualitatively or semi-quantitatively, supporting lesion detection and characterisation, particularly in breast and liver imaging. Quantitative DCE approaches have also been explored across neurological, oncological, renal and musculoskeletal conditions, although routine quantitative implementation remains limited.
ASL-MRI uses blood water as an endogenous tracer, allowing perfusion assessment without contrast agents. This makes it valuable when contrast is contraindicated. In brain imaging, ASL is used for tumour assessment and for interpreting post-treatment changes, and it is recommended in managing Moyamoya disease. Single-delay implementations are common in clinical practice, while multi-delay approaches can provide additional timing information but are less widely available as standard solutions.
Must Read: Gadolinium Use Remains Central in Head and Neck MRI
IVIM-MRI derives perfusion-related information from diffusion-weighted imaging acquired with multiple b-values. By separating diffusion and perfusion-like components, it offers a non-contrast method linked to microvascular behaviour. Applications have included oncology, stroke and renal imaging, where IVIM has been described as capable of detecting chronic kidney disease and acute graft dysfunction. Clinical uptake remains limited by technical and reproducibility challenges.
Quantification Depends on Protocol Discipline
Visual interpretation of perfusion MRI is often achievable with modest protocol demands, but quantitative use requires greater consistency in acquisition and analysis. Signal behaviour, temporal resolution and modelling assumptions are sensitive to parameter choices, scanner characteristics and physiological factors, which complicates standardisation in busy clinical environments.
For DSC-MRI, guidance highlights the importance of sufficient coverage and temporal resolution to capture bolus dynamics. Recommended parameter ranges support reliable signal changes, while recognising that assumptions underlying perfusion modelling become less valid when the blood-brain barrier is disrupted. In such cases, leakage correction is advised using established technical strategies. Even with optimisation, absolute quantification is affected by technical limitations, making relative measures combined with visual assessment more practical for many clinical decisions.
DCE-MRI protocols must balance temporal resolution, spatial detail and coverage according to the clinical objective. Quantitative analysis requires converting signal to contrast concentration, supported by baseline measurements and correction for technical effects at higher field strengths. It also depends on arterial input function selection and pharmacokinetic modelling. In practice, reproducibility remains a major barrier, with variability introduced by differences in acquisition settings, analysis software and reporting conventions. Identical datasets can yield different outputs depending on the analysis approach, limiting confidence in quantitative results.
ASL-MRI introduces a timing challenge because labelled signal decays, requiring a compromise between arrival time and signal preservation. Recommended protocols define standard labelling and readout approaches, with timing adjustments for haemodynamic impairment and neonatal imaging. Spatial resolution is constrained by signal-to-noise limitations, and interpretation requires familiarity with characteristic artefacts related to delayed arrival or vascular drainage. Technical setup and limited clinician familiarity remain practical barriers.
IVIM-MRI requires careful selection of diffusion parameters, motion management and sufficient averaging to stabilise signal. Clinical adoption is restricted by limited repeatability for some parameters, longer acquisition times and the absence of widely accepted standards for acquisition and analysis.
Workflow Integration and Standardisation Are Central
Across all techniques, the main obstacle to wider adoption is not availability on scanners but consistent integration into clinical workflows. Postprocessing is often fragmented, with analysis performed using third-party tools that vary widely in approach and output. This fragmentation complicates reporting, reduces comparability and limits clinical confidence in quantitative measures.
The guidance emphasises the need for improved standardisation and accessibility, including closer alignment between scanner vendors and external software providers. Multicentre technical studies are identified as essential for establishing reproducibility, while multicentre clinical trials are needed to define clinically meaningful thresholds. Increasing awareness and expertise among clinicians, supported by professional communities and shared practice, is also highlighted as a key driver of adoption. Efforts to improve reproducibility through common acquisition and analysis standards underpin these recommendations.
Perfusion MRI provides complementary contrast-based and non-contrast techniques with established clinical roles and growing relevance for quantitative assessment. Wider routine use remains limited by protocol variation, analysis differences and uneven reproducibility. Structured guidance clarifies how perfusion techniques can be selected, acquired and analysed more consistently within clinical workflows. The immediate value for healthcare professionals tlies in disciplined implementation that strengthens interpretability and comparability, enabling perfusion MRI to contribute more reliably to diagnosis, monitoring and treatment decisions.
Source: European Radiology
Image Credit: iStock
