Nondiagnostic Bariatric Abdominal CT: Photon Starvation, Truncation, and Arms at the Sides
A 47-year-old man weighing 178 kg (body mass index 52) is scanned for acute abdominal pain using a 64-row scanner with a 50 cm maximum field of view. He cannot raise his arms because of shoulder pain and a recent clavicle fracture, so both arms rest along his sides. The technologist selects a standard adult abdomen/pelvis protocol at 120 kVp with automatic exposure control, a noise index of 10, and a maximum tube current of 500 mA. The mA reaches its ceiling for most of the acquisition and the reported CTDIvol is 22 mGy, yet the reconstructed 3 mm images show severe dark streaking between the hips, graininess throughout the soft tissues, and truncation of the lateral body wall. The patient's arms produce dense bands across the upper abdomen. The radiologist says the images are nondiagnostic for small bowel and mesenteric pathology.
1. Identify the artifacts on these images and explain the mechanism of each. Why does raising the tube current not solve the problem?
Read the worked answer
Three separate problems are superimposed, and they have different fixes.
- Photon starvation streaks between the hips: in a body mass index of 52 the lateral dimensions attenuate so much that very few photons reach the detectors in projections through the pelvis, and electronic noise dominates those views; reconstruction smears them into dark bands. Raising mA would help, but it is already maxed at 500 mA, so the ceiling itself is the problem.
- Out-of-field (truncation) artifact: the arms lying along the body extend beyond the 50 cm reconstruction field of view. Data outside the field are not used, so the reconstruction algorithm, which assumes the body ends at the edge, produces bands and a distorted lateral body wall. Dense arms also create the obvious streak bands across the upper abdomen.
- Graininess: simply too few photons for the requested noise index at this attenuation, compounded by thin 3 mm slices.
This is not a single-technique failure; changing only one variable will not fix it.
2. List, in order of impact, the protocol and positioning changes you would make, and justify the ranking.
Read the worked answer
Ranked by impact:
- Get the arms out of the anatomy. Even with limited shoulder motion, raise them overhead on a support or position them forward; this removes both the truncation and most of the streak banding and usually achieves more than any reconstruction change.
- Increase kVp to 140. Attenuation falls and penetration improves; in a very large patient the higher output per mAs often allows total mAs to drop enough that CTDIvol rises little or even falls. Beam-hardening artifact and photon starvation both improve.
- Use a large-bore or wide-FOV scanner if available, select the large-patient bowtie filter, and use the widest reconstruction field of view that still includes the skin line.
- Reconstruct thicker slices (5 mm) with a smoother kernel and iterative reconstruction to reduce noise; keep thin slices for reformats if needed.
- Set a realistic noise index - chasing a noise index of 10 in a 178 kg patient at 3 mm will only max out the tube.
- Consider oral contrast or an alternative modality if the question remains small bowel pathology and images stay inadequate.
3. How do you judge whether the study is diagnostic, and what would you do before considering a repeat?
Read the worked answer
Diagnostic adequacy is a judgement about the clinical question, not a global pass or fail.
- Measure rather than guess: place regions of interest in the liver, aorta and the noisiest soft tissue, record the standard deviation as noise, and compute contrast-to-noise ratio for the structure in question. Document that the mA ceiling was reached and the CTDIvol delivered.
- Decide against the question: if the question is mesenteric ischemia or small bowel obstruction, the mesentery and bowel wall must be interpretable; if they are not, the study is nondiagnostic for that question even if the liver looks acceptable.
- Before repeating: repeating the same parameters will produce the same images and add dose. A repeat is only justified with a changed strategy - arms repositioned, 140 kVp, thicker slices, large-bore scanner - and after discussing cumulative dose with the radiologist.
- Communicate: tell the referring team what is and is not answerable, and suggest an alternative (ultrasound, MRI, or clinical observation) when CT cannot answer the question. Record the discussion.