Christian Eusemann, Ph.D.
The present and future of spectral imaging
October 09, 2018
By Christian Eusemann
Decades before it debuted in the clinical setting, the concept of spectral imaging – the acquisition of two or more different X-ray spectra – intrigued the scientific community.
Sir Godfrey Hounsfield first mentioned spectral imaging in the scientific literature in 1973, a year after he invented CT. But while the first clinical CT system capable of spectral imaging – also known as dual energy CT – became available in 1987, clinical adoption didn’t occur until 2005. That’s when the first generation of dual-source CT systems emerged, enabling fast, simultaneous acquisition of spectral data at radiation dose levels similar to conventional CT.
After 30 years of gradual momentum, spectral imaging has thrived over the past decade. Published papers proliferated from less than 50 in 2006 to nearly 350 in 2015, according to PubMed, and spectral imaging’s presence has expanded far beyond the academic institutions that once used it exclusively. In the past two years, adoption has increased rapidly in large, nonacademic hospitals – specifically, in their emergency departments (EDs), where up to 60 percent of all CT scans are conducted in spectral mode. Over the next three years, midsized (200- to 400-bed) hospitals are similarly expected to acquire more CT systems with spectral imaging capabilities.
The value of data obtained from two different energy spectra is multifold. In ED and trauma departments, where speed is critical, spectral imaging can rapidly visualize subtle fractures and bone marrow edema in patients who don’t have time for – or easy access to – a magnetic resonance imaging (MR) scan. Additionally, spectral imaging can improve iodine visualization in the GI and GU tracts for better visualization of solid organ injury or ischemia (i.e., ischemic bowel). Spectral imaging also enables easier, more accurate identification and visualization of pulmonary embolisms, especially at suboptimal contrast enhancement. Finally, it can help clinicians characterize incidental lesions as non-enhancing (i.e., cysts) or enhancing (i.e., requiring additional workup). Outside of the ED, spectral imaging has been widely adopted for multiple applications in radiology and oncology (e.g., improving lesion visualization and characterization).
Despite these benefits, spectral imaging was not widely adopted for many years due to workflow issues. Fortunately, medical device manufacturers have invested heavily in automating the postprocessing of spectral data. These companies are also migrating spectral imaging to more accessibly priced CT systems.
But for all its inherent appeal and increasing accessibility, today’s spectral imaging technology is still limited to the acquisition of only two different sets of image data. The next logical extension of spectral imaging is photon counting, a technology being developed and installed by multiple manufacturers as a prototype in a select few academic institutions. Photon-counting detectors not only register (or count) each photon in an X-ray beam, but they also measure its energy, enabling the sorting of photons into different energy bins.
That ability to gauge each photon’s energy output represents true multi-energy imaging, and it could have profound implications. While current spectral imaging technology is sufficient to characterize some materials in the human body and to quantify iodine enhancement, new contrast materials with additional absorption peaks could one day enable simultaneous imaging of multiple contrast agents. For example, clinicians could highlight and tag a specific lesion with one agent and enhance the vasculature feeding it with another agent.
Exactly when this next iteration of spectral imaging will advance beyond the prototype stage is unclear. But this much is certain: Spectral imaging will one day become THE imaging standard for CT, delivering significant added value in terms of quantitative and functional information beyond the advanced structural imaging of today – all to improve patient outcomes.
About the author: Christian Eusemann, Ph.D., is vice president of collaborations at Siemens Healthineers North America.