A sinogram has two distinct meanings in medicine, and which one applies depends on the context. In medical imaging physics, a sinogram is the raw data collected by a CT or PET scanner before it gets converted into the cross-sectional images doctors actually look at. In clinical radiology, a sinogram (also called sinography) is a diagnostic procedure where contrast dye is injected into a wound or fistula tract to map its path through tissue. Both uses share a name but serve very different purposes.
The CT and PET Sinogram: Raw Scanner Data
When you lie inside a CT scanner, an X-ray source rotates around your body and a detector on the opposite side records how much radiation passes through at each angle. Before those measurements become the familiar slice images, they exist as a two-dimensional data map called a sinogram. The horizontal axis represents the detector position (essentially where along your body the X-ray landed), and the vertical axis represents the rotation angle of the scanner, typically spanning from -90 to 90 degrees.
The name comes from what this data looks like. A single dense point in your body, like a small bone or a metal implant, traces out a sine wave pattern across the sinogram as the scanner rotates around it. Every structure in the scan creates its own overlapping sine curve, producing a striped, wave-like image that looks nothing like anatomy to the untrained eye.
PET scanners generate sinograms too, though the data acquisition works differently. Instead of an external X-ray source, PET detects pairs of gamma rays emitted from a radioactive tracer inside your body. The raw coincidence data gets organized into sinograms or projection views for storage and processing before reconstruction into images.
How Sinograms Become Medical Images
The process of converting a sinogram into a usable image is called reconstruction. The most established method, filtered back projection, works in two steps. First, the raw data is passed through a mathematical filter that sharpens it and prevents blurring. Second, the filtered data from every angle is “projected back” across the image plane, and where all those projections overlap, the original anatomy emerges as a cross-sectional picture.
Think of it like this: if you shined a flashlight through a tree from hundreds of angles and recorded the shadow each time, you could work backward from all those shadow profiles to figure out the shape of the trunk and branches. That’s essentially what back projection does with X-ray data.
Modern scanners also use a technique called iterative reconstruction, which repeatedly refines the image by comparing a simulated sinogram to the real one and adjusting until they match. This approach handles noise better, which is why it’s become standard on newer CT systems and allows for lower radiation doses.
Why Engineers and Physicists Care About Sinograms
Radiologists almost never look at sinograms directly. The images they interpret have already been reconstructed. But for medical physicists and scanner engineers, sinograms are essential for quality control and troubleshooting. Problems that are hard to spot in a finished image often show clear signatures in sinogram space.
Metal implants are a good example. When X-rays hit dense metal, they lose energy unevenly across different wavelengths, a phenomenon called beam hardening. In the sinogram, metal shows up as regions of abnormally high attenuation that create nonlinear biases in the data. In the final image, these appear as bright streaks radiating outward from the metal object. Patient motion during the scan also leaves characteristic patterns in the sinogram, appearing as discontinuities or shifts in the sine wave traces.
Artificial intelligence is increasingly being applied directly to sinogram data. Deep learning models can correct for metal artifacts, fill in missing data from low-dose scans, and even identify anatomy or detect abnormalities without ever reconstructing a traditional image. Research has shown that all the diagnostic information present in a finished CT image also exists in the raw sinogram. Neural networks trained on sinogram data can perform anatomy identification and pathology detection, sometimes with performance comparable to models trained on reconstructed images.
The Clinical Sinogram: Mapping a Wound Tract
The other type of sinogram is a completely different procedure. When a patient has a chronic wound, ulcer, or surgical site that keeps draining fluid, doctors sometimes need to know exactly where the underlying channel (called a sinus tract) goes. A clinical sinogram involves injecting diluted iodinated contrast dye into the opening of the wound, then taking X-ray images to trace the path of the tract through soft tissue and bone.
The contrast fills the channel and makes it visible on X-ray, revealing how deep the tract extends, whether it connects to bone, and whether any abscesses or cavities are present along its path. In one documented case, a sinogram of a draining heel ulcer revealed that the sinus tract extended from the skin surface through subcutaneous tissue all the way into the calcaneus (heel bone), confirming osteomyelitis, a bone infection, that guided surgical planning.
This type of sinogram is particularly valuable in settings where MRI or CT scanners aren’t available. It requires only basic X-ray equipment and contrast dye, making it a practical diagnostic tool in resource-limited areas. Combined with plain radiographs, sinography provides a reliable way to assess chronic infections, wound complications, and fistula anatomy when advanced imaging isn’t an option.
Which Meaning Applies to You
If you encountered the term in the context of CT physics, image reconstruction, or scanner technology, you’re dealing with the raw data format. If you saw it mentioned in a clinical report about a wound, fistula, or chronic infection, it refers to the contrast dye procedure. The two share a name because both involve the concept of tracing paths, one mathematical, one anatomical, but they exist in entirely different branches of medical practice.