Windowing & Reconstruction Planes
- A CT scan stores a single set of numbers (Hounsfield units); windowing is just deciding which slice of that number range becomes black-to-white on your screen.
- One scan, many windows: lung, soft tissue, and bone windows are different views of the same data — you didn't re-scan anything, you re-shaded it.
- Window width = how much contrast (narrow = high contrast, fewer shades; wide = low contrast, more shades). Window level = where the middle gray sits.
- Reconstruction planes (axial, coronal, sagittal) reslice that 3D block of data from different angles — like cutting a loaf of bread crosswise, lengthwise, or down the middle.
- Reformats are nearly free and ridiculously useful; if a finding is confusing on one plane, look at another before you panic.
Here's a thing that took me embarrassingly long to internalize: when a radiologist flicks a chest CT from a gray blur into a crisp picture of the ribs, then into a picture of the lungs, they did not load three different scans. They loaded one. They just changed the lighting. Once that clicks, half the magic of CT stops being magic and starts being a dimmer switch you control.
What a CT actually stores
A CT scanner doesn't save a picture. It saves a grid of numbers — one number per tiny cube of you — measuring how much each cube soaked up the X-ray beam. That number is the Hounsfield unit (HU), an attenuation scale anchored so that water is 0 and air is around −1000. Dense bone runs up into the high hundreds or beyond; fat sits below water in the negatives.
The catch: your monitor and your eyeballs are pathetic compared to that range. A screen shows only a couple hundred shades of gray, and the human eye can only tell apart a few dozen at a glance. The data spans thousands of HU. So you cannot see all of it at once, ever. You have to pick which part of the number range you care about right now.
That picking is windowing. (If the HU scale itself feels shaky, it's worth a quick detour through attenuation and radiographic contrast — this page is the practical sibling of that one.)
Windowing: width and level, in plain English
Two knobs. That's it.
- Window level (WL) is the center of your gray ramp — the HU value that comes out as middle gray. Slide it up to look at denser stuff, down to look at airy stuff.
- Window width (WW) is how wide a band of HU you stretch across black-to-white. Everything below the band is pure black; everything above is pure white; the band in between gets the grays.
Think of it like sunglasses with an adjustable tint. The level is where you point your gaze; the width is how much glare you let through.
Narrow width = high contrast, big visible differences between tissues that are close in density, but anything outside the band blows out to black or white. Wide width = low contrast, gentle gradients, more tissues on screen at once but subtler.
A narrow window is a microscope for one neighborhood of densities; a wide window is a wide-angle lens that shows everyone but flatters no one.
Why one scan needs several windows
Different tissues live in wildly different parts of the HU scale, so no single window shows them all well. The classic presets exist because each tunes the dimmer to a different problem:
| Window | Roughly tuned for | What pops | What's sacrificed |
|---|---|---|---|
| Lung | very low densities (air vs. soft tissue) | lung markings, nodules, pneumothorax | the mediastinum looks like a black hole |
| Soft tissue / mediastinal | densities near water | organs, vessels, fluid, masses | lung detail and fine bone vanish |
| Bone | very high densities | cortex, fracture lines, trabeculae | soft tissue washes out to gray mush |
| Brain | a narrow band near water | subtle gray–white difference, early stroke, blood | everything else, deliberately |
The numbers behind these presets vary by scanner and by what you're hunting, so I won't pretend there's one sacred value — the concept (narrow band aimed at the tissue you care about) is the part that matters.
A finding can be flat-out invisible in the wrong window. A small lung nodule can hide in soft-tissue window; a subtle rib fracture can vanish on lung window; an acute brain bleed can wash out to nondescript gray on a wide window if you forget to switch back to a narrow brain window. "I didn't see it" sometimes just means "I never windowed for it." Always look at the tissue you're worried about in its own window.
Reconstruction planes: reslicing the loaf
Now the other half. Modern CT collects a true 3D block of data — a stack of thin slices so fine they behave like a solid volume. Because it's volumetric, the computer can cut that block at any angle after the fact. This is multiplanar reconstruction (MPR).
The three standard cuts:
- Axial — slices crosswise, like cutting a banana into coins. This is usually the raw acquisition plane.
- Coronal — slices front-to-back into a standing-up, face-on view, like a slice of bread pulled from the loaf.
- Sagittal — slices left-to-right down the midline, the profile view.
The payoff is that some findings only make sense in the right plane. A long structure that wanders in and out of axial slices — the aorta, a ureter, the spinal canal — is far easier to follow when you reslice it lengthwise. A vertebral compression or a question of "is this above or below that" practically answers itself on sagittal.
Reformats are essentially free and don't add any radiation — they're computed from data you already have. So reaching for a coronal or sagittal view when an axial image is confusing isn't fancy; it's just good habit. The thinner the original slices, the cleaner the reformats look.
A quick honesty note: windowing and reconstruction are different verbs. Windowing changes how the numbers are shaded; reconstruction changes the geometry you slice along (and, at a deeper level, the math that builds the image — see CT physics and reconstruction). You routinely do both: pick a plane, then dial the window for the tissue you're chasing.
The one habit to keep
If you remember nothing else: the image on the screen is a choice, not a fact. The scan holds far more than any single view reveals. When something looks weird, or when you're hunting a specific structure, your first move isn't to doubt the scanner — it's to change the window or change the plane. The same instinct carries over to reading the chest CT, where flipping between lung, soft-tissue, and bone windows on every study is just part of the routine. You're not looking at a picture. You're looking through one.