8.1 DSA Principles, Timing & Image Acquisition

Key Takeaways

  • Mask frames should precede contrast arrival at the target and match fill frames in projection and position.

  • In the ideal model, the log of mask intensity divided by fill intensity gives the change in line attenuation.

  • Frame rate describes time sampling. Dose also depends on output per exposure, duration and repeats.

  • Check which action the console delay actually controls. Choose the image phases and duration for the diagnostic task.

Last updated: November 2026

8.1 DSA Principles, Timing & Image Acquisition

Clinical Core: Compare a mask acquired without contrast in the structure of interest with the corresponding contrast-filled views. Logarithmic subtraction suppresses matched background. The result depends on alignment, exposure, contrast arrival and the quality of the acquired images.

Mask, fill and log subtraction

The mask records background before the target opacifies. Systems can use a suitable individual frame or average selected frames. The fill or contrast series shows the same field as contrast passes through it. Check for patient motion, table or gantry changes, and residual contrast in the mask. A native image is a source view before subtraction. Both mask and fill have native frames.

In an illustrative single-energy model, assume the same incident intensity I0I_0, matched background line attenuation BB, negligible scatter and correctly responding detector measurements. Let CC be the change in line attenuation from the opacified lumen relative to the mask. Then:

Imask=I0eโˆ’B,Ifill=I0eโˆ’(B+C)I_{\mathrm{mask}}=I_0e^{-B},\qquad I_{\mathrm{fill}}=I_0e^{-(B+C)}

Taking logs gives:

lnโกImaskโˆ’lnโกIfill=C\ln I_{\mathrm{mask}}-\ln I_{\mathrm{fill}}=C

Equivalently, for attenuation values A=โˆ’lnโก(I/I0)A=-\ln(I/I_0), this convention uses Afillโˆ’AmaskA_{\mathrm{fill}}-A_{\mathrm{mask}}. If an implementation subtracts in the opposite order the signal changes sign. Display polarity also depends on the presentation mapping.

Worked comparison. Consider two rays through backgrounds of different thickness. For the first, mask and fill are 400 and 200 in proportional intensity units. For the second, they are 80 and 40. The raw differences are 200 and 40. Each log ratio is lnโก(2)\ln(2), about 0.693, because each intensity halves. Log subtraction therefore isolates the same added attenuation under the assumptions. In a real acquisition, differences in scatter, spectrum, noise, positioning, and receptor response can leave background signal or change vessel conspicuity. An apparently clear subtraction does not make low-signal or clipped detail recoverable.

Timing and contrast phases

Timing or phaseAcquisition question
BaselineDo the selected mask frames show the target without contrast, with the same view and similar position as the fill frames?
ArterialIs the run sampling the inlet, branches and early lesion filling quickly enough? A shunt may require resolving early venous opacification.
ParenchymalIs tissue blush or transition through the organ of interest visible when it matters to the study?
Venous or delayedAre the drainage route, delayed filling, collaterals or distal runoff included when needed?

The delay label on a console can refer to the time before an exposure or an injection trigger, depending on how the devices are coupled. Check the actual order, transit to the imaged anatomy, and which frame or frames will serve as the mask. For example, the Philips Azurion coupled-operation instructions specify that contrast is not injected until exposure starts in the described operation. A delayed target can still be unopacified after injection begins elsewhere. As a timing exercise, if acquisition starts at 0.0 s, injection at 0.4 s and local contrast first arrives at 1.1 s, a stable frame at 0.8 s can serve as a mask in this simplified sequence. The 0.4 s injection start alone does not settle mask suitability. If available, check the native images. Match contrast rate, volume and pressure to the catheter and setup described in ยง11.1.

Temporal sampling and dose

Acquisition frame rate is the number of acquired images per second. Fluoroscopy pulse rate is the number of X-ray pulses per second. Compare the relevant exposure count and per-exposure output. They coincide only if the mode acquires one frame per pulse. Increasing frame rate helps sample rapid changes more closely. Prefer the lowest rate adequate for the task and retain detail where early transitions, rapid blood flow or motion matter. Slow delayed filling can sometimes be sampled less frequently. There is no single rate for an entire vascular territory.

Worked comparison. Assume a hypothetical 15 s run with one pulse per frame at unchanged output per pulse. Using 4 frames/s throughout yields 60 frames. Using 4, 2 and 1 frames/s for successive 5 s intervals yields 20+10+5=3520+10+5=35 frames. There are 25 fewer exposures, about 42% fewer. Actual dose depends on automatic technique and total acquisitions, not just the frame count. Real cerebral angiography programs have used variable sampling, including a 4-to-2-to-1 schedule. Choose the delay, frame rate and duration for the imaging question and system.

Loading diagram...
DSA Acquisition, Logarithmic Transformation & Subtraction Mechanics
Test Your Knowledge

Under the stated matched, single-energy model, both mask intensities are twice their fill intensities despite different background transmission. What does the log-ratio model predict?

A

The same change in line attenuation, because the mask-to-fill ratios match.

B

A stronger log signal solely from the brighter background.

C

Zero signal because both rays contain a background.

D

The same raw intensity difference on both rays.

Test Your Knowledge

Two candidate 10-second runs each use one X-ray pulse per acquired frame at the same output per pulse. A slower diagnostic phase can be sampled adequately at 2 rather than 5 frames/s for the full comparison. What changes?

A

Fifty exposures become forty because both runs last 10 seconds.

B

The count falls from 50 to 20. Under the stated output assumption this is a 60% reduction in modeled X-ray output.

C

The dose always rises because the rate is lower.

D

Only display refresh changes. Both runs have 50 exposures.

Test Your Knowledge

A local DSA protocol records valid baseline frames before contrast arrives at the selected artery. The injector is changed to a different console. What must the team verify?

A

Always enter 0.5 s under the X-ray delay label.

B

Require the mask to capture the peak bolus.

C

Verify the imaging and injection sequence and confirm that suitable mask frames precede opacification at the target.

D

Match injector ramp time to the display refresh rate.

Sections you finish are checked off in the contents.