The progress of quantitative MRI towards biomarkers

MRI has historically generally been used qualitatively, to discern pathologies. However, MRI also has the functionality of assessing physical measurements, for example, of tissue volumes. An ability that has not gone unnoticed by clinicians and researchers working with MRI.¹

Qualitative versus quantitative MRI

In conventional qualitative magnetic resonance imaging (MRI), the image contrast is relative and diagnostic decisions are based on the interpretation of a radiologist.¹ Although qualitative images of this kind are essential for informing clinical decisions about individual patients, but they may be more limited when it comes to comparing patients or sets of timepoints as different scanners have different sensitivities and operate under different conditions.¹ Scanner sensitivity also drifts over time, with results subtly affected by servicing and software updates.¹ Moreover with qualitative MRI, analyses are generally not provided with an associated measurement of uncertainty.¹

Quantitative MRI, as the name suggests, allows for the measurement of various properties from the MR image.² The quantitative measurement process then allows scanned parameters to be traced to verifiable physical quantities, theoretically providing greater consistency of analysis.¹ This implies that quantitative MRI may be a very useful technique for clinical studies as images can be produced and analyzed using consistent criteria.¹ Indeed, quantitative MRI, following the Response Evaluation Criteria in Solid Tumors or RECIST criteria are common endpoints in clinical trials, and are increasingly filtering into routine practice.² MRI also facilitates a more comprehensive quantification of tumor characteristics, the tumor’s perfusion, cellularity, metabolism and protein deposition. This can enrich the understanding of the tumor beyond conventional anatomic information.² These measurements can also be used as surrogates, or biomarkers, for pathophysiological processes.² For example, apparent diffusion coefficient (ADC) has been used to enhance molecular subtyping of breast cancer; or increases in tumor T₁ relaxation time may indicate a better response to anti-angiogenic therapy in ovarian cancer; to mention just two among many other potentially useful biomarkers.³ Quantitative MRI is also a key process in combining multiple measures in order to predict clinical outcomes.² Nevertheless, quantitative MRI is not without considerable challenges limiting its clinical acceptance. For instance because of the longer acquisition times required to combine different sequence parameters on MRI protocols can make quantitative MRI unfeasible in clinical practice.³ There are also issues of generalizability; the ability to translate methods and models across institutions.⁴

Turning quantitative MRI into biomarkers

Quantitative MR imaging doesn’t automatically translate into useable biomarkers. To qualify as a reliable biomarker, the quantitative MRI measurement needs to be thoroughly demonstrated to be accurate, precise, repeatable, and reproducible. Moreover, it also needs to be practical and cost-effective for use in most conventional MRI settings.²

At a technical level, for a measurement to be meaningful it must be accompanied by a description of the associated uncertainty of that measurement.¹ Without this assessment of uncertainty, values cannot be meaningfully compared as there is no quantification of the significance of differences in values.¹ And as both data the acquisition and processing steps are potential sources of bias and uncertainty, quantitative MRI requires careful calibration and testing of MRI instruments and image analysis tools in order for the quantitative measurements to be applicable as biomarkers.¹ Although quantitative MRI has the potential to become vital in the oncology workflow, poor reproducibility is still limiting its translation into clinical practice.⁵

What still needs to be done

Quantitative MRI has advanced considerably, particularly in the field of oncology in recent years, with many techniques now commonly used in clinical trials and others that have shown promise in preliminary studies.² However, to make the best use of quantitative MRI, consensus still needs to be found about the methodologies involved in data acquisition and analysis. The biomarkers proposed also need to be shown to be repeatable and reproducible in well-designed studies.²

References

1. Cashmore MT, McCann AJ, Wastling SJ, McGrath C, Thornton J, Hall MG. Clinical quantitative MRI and the need for metrology. Br J Radiol 2021;94:20201215.
2. Abramson RG, Arlinghaus LR, Dula AN, Quarles CC, Stokes AM, Weis JA, Whisenant JG, Chekmenev EY, Zhukov I, Williams JM, Yankeelov TE. MR imaging biomarkers in oncology clinical trials. Magn Reson Imaging Clin N Am 2016;24:11-29.
3. Ding H, Velasco C, Ye H, Lindner T, Grech-Sollars M, O'Callaghan J, Hiley C, Chouhan MD, Niendorf T, Koh DM, Prieto C, Adeleke S. Current applications and future development of magnetic resonance fingerprinting in diagnosis, characterization, and response monitoring in cancer. Cancers (Basel) 2021;13.
4. Keenan KE, Delfino JG, Jordanova KV, Poorman ME, Chirra P, Chaudhari AS, Baessler B, Winfield J, Viswanath SE, deSouza NM. Challenges in ensuring the generalizability of image quantitation methods for MRI. Med Phys 2022;49:2820-35.
5. Kendrick J, Francis R, Hassan GM, Rowshanfarzad P, Jeraj R, Kasisi C, Rusanov B, Ebert M. Radiomics for identification and prediction in metastatic prostate cancer: a review of studies. Front Oncol 2021;11:771787.