Future of Medical Imaging

Rationale for Combining Medical Imaging Modalities

Each imaging modality provides different information. MRI and CT primarily show anatomy, tissue structure, density, and composition. PET provides functional and molecular information such as metabolism and receptor activity. Ultrasound provides low-cost, radiation-free imaging of soft tissue and blood flow, while fNIRS measures cortical oxygenation.

Combining modalities can overcome the limitations of individual techniques. Anatomical imaging can be difficult to interpret without functional information, while functional abnormalities may be difficult to localize without detailed anatomy.

Established Multimodality Combinations

PET/CT

PET/CT combines PET's functional information with CT's anatomical information, improving disease detection, localization, characterization, and monitoring. Combining the examinations also reduces the need to interpret and mentally register separate scans.

From an engineering perspective, PET/CT is practical because CT can acquire images rapidly and provide attenuation correction for PET reconstruction.

PET/MRI

PET/MRI combines MRI's high-quality anatomical and soft-tissue information with PET's quantitative molecular and functional information. It can be particularly useful for the brain, breast, liver, kidney, musculoskeletal system, and anatomically complex regions.

PET/MRI required specialized engineering because conventional PET photomultiplier tubes are sensitive to MRI's magnetic field. The development of solid-state digital PET detectors made integrated PET/MRI systems possible.

Ultrasound + MRI/CT

Ultrasound can be combined with previously acquired MRI or CT data during image-guided procedures. For example, MRI can provide detailed anatomy while live ultrasound provides real-time, radiation-free guidance.

fNIRS + EEG

fNIRS measures changes in cerebral blood oxygenation, providing useful spatial information but limited temporal resolution. EEG measures electrical brain activity with high temporal resolution but more limited spatial localization. Combining them provides complementary information for neuroscience and brain-computer interface research.

Limitations of Less Common Modality Pairings

MRI + CT

MRI and CT are rarely physically integrated because both primarily provide anatomical information. Although their contrast mechanisms differ, the additional benefit may not justify the added complexity, infrastructure, cost, and space requirements.

Ultrasound + PET

Direct PET/ultrasound integration has limited advantages. PET requires radioactive tracers, controlled detector geometry, and longer acquisition times, while ultrasound uses an externally positioned, operator-controlled transducer for real-time imaging. Combining them would add mechanical and workflow complexity with limited additional benefit.

Future Directions in Multimodality Imaging

PET/MRI has significant potential because it combines detailed anatomy, soft-tissue characterization, and molecular information. It may be particularly useful in oncology and neuroscience.

Neurodegenerative disease is another important application. For example, MRI can measure cerebral atrophy while PET can detect molecular processes such as amyloid deposition. Together, these measurements can provide a more complete picture of disease progression.

However, broader clinical adoption depends on cost, availability, clinical validation, and evidence of improved patient outcomes.

The Accessibility Challenge

Creating a single system containing MRI, CT, PET, and ultrasound is unlikely to be practical because each technology has different engineering requirements.

• MRI: Strong magnetic field and specialized infrastructure.
• CT: Rotating X-ray source, detectors, and radiation shielding.
• PET: Radioactive tracers, specialized detectors, and radiation-safety infrastructure.
• Ultrasound: Compact transducers and acoustic electronics.

Combining these systems would increase mechanical, electrical, computational, and infrastructure complexity, along with acquisition and maintenance costs. These challenges are especially important in resource-limited settings.

A More Realistic Approach to Accessible Multimodal Imaging

A more practical approach is to integrate imaging data through software rather than physically combining the hardware. Individual technologies could become smaller, cheaper, and more portable while a computational layer connects their outputs.

Supporting Technologies

• Portable/low-field MRI: Reduces infrastructure requirements and may expand access to MRI.
• Point-of-care ultrasound: Provides portable, real-time imaging where conventional infrastructure is unavailable.
• Wearable fNIRS and EEG: Enables repeated brain monitoring in clinical, ambulatory, or home settings.
• Cross-modality computational integration: Software could register and compare MRI, ultrasound, PET, and other data acquired at different times, creating a longitudinal view of a patient's condition.

This approach can be called a "virtual hybrid" imaging model: instead of combining incompatible hardware, it combines their outputs computationally.

Conclusion

The future of multimodal imaging is unlikely to depend entirely on increasingly complex all-in-one scanners. PET/CT and PET/MRI demonstrate the value of physically combining complementary modalities, but their infrastructure and engineering requirements limit accessibility.

A more scalable approach is to improve each modality independently while using computational systems to integrate their outputs. This could allow MRI, CT, PET, ultrasound, and fNIRS to work together without requiring their hardware to occupy the same system.

The goal of multimodal imaging may not be one machine that performs every scan, but an imaging ecosystem where complementary technologies communicate to provide a more complete view of disease while reducing cost and infrastructure requirements.

About

Joanna Ye,
Biomedical Engineering Major
Boston University Class of '29

Contact

Add a contact form or details here.