CZT Camera SPECT/CT Finds Hidden Gastrointestinal Bleeding That CT Misses

When a patient arrives at the emergency department with visible gastrointestinal bleeding, every minute counts. Multidetector computed tomography, or MDCT, has become the standard first-line tool for locating the source of such hemorrhage, guiding clinicians toward endoscopy, embolization, or surgery. Yet a substantial number of these scans come back negative, leaving physicians with a bleeding patient and no anatomical target. A new study published in BMC Medical Imaging suggests that a next-generation nuclear medicine technique could rescue many of these diagnostic dead ends.

Researchers at Soonchunhyang University Cheonan Hospital in South Korea evaluated the diagnostic performance of single-photon emission computed tomography combined with computed tomography, SPECT/CT, using technetium-99m-labeled red blood cells and a full-ring 360-degree cadmium-zinc-telluride, or CZT, camera. The retrospective analysis covered 65 patients who presented with overt gastrointestinal bleeding between April 2021 and December 2024 and whose initial multiphasic MDCT scans showed no evidence of active extravasation. The findings indicate that this radionuclide approach can detect bleeding sites that conventional cross-sectional imaging misses, with particularly strong results in the small bowel and at surgical anastomosis sites.

The technology behind the study represents a meaningful departure from traditional gamma cameras. Conventional scintigraphy relies on sodium iodide crystals coupled to photomultiplier tubes, which convert gamma photons emitted by the radiotracer into electrical signals. CZT detectors, by contrast, directly convert gamma photons into electron-hole pairs within the semiconductor crystal itself. This direct conversion yields superior energy resolution and spatial resolution, allowing the camera to distinguish true photon signals from scatter more effectively. The full-ring 360-degree configuration also means the detectors completely surround the patient, enabling simultaneous acquisition from all angles rather than the stepwise rotation required by older systems, which shortens acquisition times and improves count statistics.

The imaging protocol in the study followed a two-stage design. Patients first underwent dynamic planar imaging with a dual-head gamma camera, capturing sequential frames that could reveal the moment radiolabeled blood escaped from a vessel into the bowel lumen. This was followed by delayed static SPECT/CT acquisition using the CZT camera, which fused functional information about tracer distribution with anatomical detail from the CT component. The fusion is critical: while planar images can show that bleeding is occurring somewhere in the abdomen, SPECT/CT localizes the extravasation to a specific bowel segment, which is what surgeons and endoscopists actually need.

Technetium-99m-labeled red blood cells work on a simple but elegant principle. A sample of the patient’s own red blood cells is labeled with the gamma-emitting radionuclide technetium-99m and then reinjected into the bloodstream. Because the labeled cells remain within the vascular compartment, any accumulation of signal outside blood vessels indicates active hemorrhage. Unlike iodinated contrast used in CT angiography, which requires a sufficiently rapid bleeding rate at the exact moment of scanning, labeled red blood cells circulate for hours, allowing intermittent or slow bleeding to be captured over an extended window. This is precisely the scenario in which MDCT fails most often.

The results were striking. Of the 65 enrolled patients, 49, or 75.4 percent, were ultimately diagnosed with a gastrointestinal bleeding focus through definitive clinical workup. Against this reference standard, technetium-99m-labeled red blood cell SPECT/CT achieved a sensitivity of 73.5 percent, a specificity of 93.8 percent, a positive predictive value of 97.3 percent, a negative predictive value of 53.6 percent, and an overall accuracy of 78.5 percent. The high positive predictive value is particularly noteworthy: when the scan did show a bleeding site, it was almost always correct, making it a powerful tool for directing targeted interventions.

The performance in specific clinical scenarios was even more impressive. The technique accurately identified bleeding in 21 of 26 patients, or 80.8 percent, who were diagnosed with small bowel bleeding. The small bowel has long been the most challenging region for bleeding localization because its length, mobility, and overlapping loops confound both endoscopy and conventional imaging. Push enteroscopy and capsule endoscopy can visualize the mucosa but offer limited therapeutic reach, and CT often cannot distinguish bleeding within small bowel loops from adjacent structures. A nuclear medicine scan that reliably flags small bowel hemorrhage fills a genuine clinical gap.

Equally remarkable was the detection of bleeding from surgical anastomosis sites. All five patients in the cohort who bled from the connection point of a prior surgical procedure were correctly identified by the SPECT/CT scan, a 100 percent detection rate. Anastomotic bleeding is a feared complication after gastrointestinal surgery, and localizing it can be difficult because postoperative anatomy distorts normal landmarks and inflammation can mimic extravasation on CT. The functional nature of the red blood cell scan, which depends on physiology rather than anatomy, appears to cut through this ambiguity.

The relatively low negative predictive value of 53.6 percent deserves careful interpretation. A negative SPECT/CT scan does not reliably exclude ongoing bleeding, likely because hemorrhage may be truly intermittent, may have temporarily stopped, or may occur at a rate too slow to produce detectable extravasation even over the extended imaging window. Clinicians should therefore treat a negative scan as inconclusive rather than reassuring, and continue to pursue alternative diagnostic pathways if the clinical picture suggests active hemorrhage. Conversely, the exceptional specificity and positive predictive value mean that a positive scan can be acted upon with confidence.

The study’s authors conclude that in patients with overt gastrointestinal bleeding and negative MDCT findings, technetium-99m-labeled red blood cell SPECT/CT could be helpful for detecting bleeding sites, particularly in cases with suspected small bowel bleeding or bleeding from an anastomosis site. As CZT detector technology continues to mature and spread through nuclear medicine departments worldwide, the combination of faster acquisition, better resolution, and fused functional-anatomical localization may reshape the diagnostic algorithm for one of medicine’s most urgent and elusive emergencies. For patients bleeding from a source that no CT can see, this full-ring camera may finally provide the answer.

Subject of Research: Diagnostic performance of technetium-99m-labeled red blood cell SPECT/CT with a 360-degree CZT camera in patients with overt gastrointestinal bleeding and negative multidetector CT findings.

Article Title: Diagnostic performance of 99mTc-labeled red blood cell SPECT/CT using a 360º CZT camera in patients with overt gastrointestinal bleeding and negative findings on multidetector CT

Article References: Yoo, I. D., Lee, S. M., Hong, S.-P., Kim, J. E., & Lee, J. W. (2026). Diagnostic performance of 99mTc-labeled red blood cell SPECT/CT using a 360º CZT camera in patients with overt gastrointestinal bleeding and negative findings on multidetector CT. BMC Medical Imaging. https://doi.org/10.1186/s12880-026-02737-x

Image Credits: AI Generated

DOI: 10.1186/s12880-026-02737-x

Keywords: gastrointestinal bleeding, SPECT/CT, CZT camera, technetium-99m, red blood cell scintigraphy, multidetector CT, small bowel bleeding, nuclear medicine, diagnostic imaging, anastomotic bleeding, radionuclide imaging, hemorrhage detection