To measure is to know

Quantitative imaging is the future of diagnostic radiology

Prof. Dr. Paul M. Parizel, MD, PhD, FRANZCR

David Hartley Chair of Radiology, Royal Perth Hospital & University of Western Australia, Western Australia

Director, Western Australian NIF node, National Imaging Facility (NIF)

Chair, Clinical Radiology Research Committee. Royal Australian and New Zealand College of Radiologists (RANZCR)

Visiting Professor, University of Antwerp (UA), Belgium

Diagnostic imaging is a cornerstone of modern medicine because it provides a non-invasive way to establish a diagnosis, plan and guide treatment, monitor and follow-up a wide range of medical conditions, and offer early disease detection (screening). Technological developments, including X-rays, CT, MR, US, PET and SPECT, help to identify the type and location of a disease process, to assess the severity (staging and grading), and determine the response to treatment.

However, image interpretation is subject to human error. Radiologists get overburdened, become distracted and tired, overlook lesions, and make diagnostic errors. Subjective differences in interpretation are exacerbated by individual experience, training, acuity in perception, but also by factors such as fatigue or distraction. The reliability of a radiologic diagnosis or measurement is impacted by inter- and intra-observer variability, e.g., in the follow-up and measurement of tumours or nodules. The lack of standardised, reproducible imaging protocols is an additional confounding factor which limits the reliability of diagnostic interpretation. Similarly, the use of non-standardised, free-text prose in radiology reports often causes communication issues, inconsistencies, and renders data analysis for research almost impossible. These variations in interpretation and reporting influence clinical decisions and potentially affect treatment plans.

By contrast, quantitative imaging (QI) tools extract quantifiable metrics from medical scans, and provide objective, consistent, data-driven insights that surpass the subjective limitations of traditional, qualitative imaging. Artificial intelligence (AI) tools yield more precise diagnoses and follow-up assessments, personalised treatments, and enhanced medical research. Examples of measurable data points include tumour volume, tissue density, or metabolic activity. 

· Improved diagnostic accuracy. QI can identify subtle changes and abnormalities that may not be visible to the human eye, lowering the risk of false positives and negatives. In oncology, quantitative analysis of tumour characteristics can provide a more complete assessment of disease progression than simply relying on diameter alone.

· Treatment response assessment. The ability to precisely measure physiological changes allows radiologists to determine if a patient is responding to a particular therapy. For instance, in patients with a chronic disease, such as relapsing remitting multiple sclerosis, who are treated with disease-modifying drugs, volumetric measurements of demyelinating plaques and grey matter provide a detailed assessment of treatment success or failure.

· Tailored therapies. QI enables precision medicine by helping to identify subgroups of patients who will respond best to a specific treatment. This data can inform clinical decisions, such as altering a patient's chemotherapy regimen or planning surgery for a stroke patient. 

The proliferation of quantitative data from medical images provides the rocket fuel for the next generation of artificial intelligence (AI) and radiomics in medicine. 

· AI-powered analysis. AI algorithms can analyse the vast datasets produced by quantitative imaging to detect abnormalities, measure subtle changes over time, and even predict disease progression more efficiently than a human can.

· Quantitative imaging biomarkers. Radiomics, the process of extracting hundreds of quantitative features from medical images, is a key part of this future. By correlating these features with clinical outcomes, researchers can identify new quantitative imaging biomarkers for various diseases, from cancer to Alzheimer's. 

Quantitative imaging addresses the problem of standardisation that plagues conventional imaging. For instance, cross-sectional imaging data may vary significantly based on the equipment manufacturer, software version, or imaging protocols used. 

· Standardised, reproducible and comparable data.  Initiatives such as the Quantitative Imaging Biomarkers Alliance (QIBA) in the USA or the National Imaging Facility (NIF) in Australia establish the standards and protocols necessary to ensure quantitative measurements are reliable and reproducible across different devices and institutions.

· Metrological rigor. With metrology at its core, QI gives clinicians confidence in the data they use for critical patient decisions. For large-scale multi-centre clinical trials, this standardisation is critical to maximise the statistical power and validate new treatments. 

The shift to quantitative imaging can yield benefits for the wider healthcare system.

· Cost reduction. Improved standardisation and higher accuracy can reduce the need for repeat scans and potentially eliminate unnecessary invasive procedures like biopsies.

· Increased accessibility. Robust, quantitative imaging techniques support the development of new, more accessible technologies, such as point-of-care ultra-low-field MRI scanners.

· Research opportunities. Quantitative, standardised imaging data is essential for building large, robust datasets that can accelerate research and validate new medical imaging algorithms. 

Quantitative Imaging (QI) opens new avenues to improve the diagnosis, management and follow-up of our patients. Rather than fighting this inexorable evolution, we should embrace QI technology as the future of diagnostic radiology. As the expression goes: “tomorrow belongs to those who prepare for it today”. Therefore, let us welcome the support that QI can give us to deliver the best possible care to our patients.

