The Portable Chest That Keeps Coming Back
A 180-bed hospital's ICU uses a wireless DR panel on a mobile unit. Portable AP chests are done at 100 cm SID, 85 kVp, and a fixed 3.2 mAs chosen by the technologist rather than AEC, with the system's exposure index target set at 400. Over four weeks, 22 of 60 portable chests were repeated, and the technologists' stated reason was consistently 'too noisy' or 'too light.' The deviation index values on the repeat images cluster at -3.4, -3.1, -2.9, -3.6 and -3.2. The radiologist has also complained that the images look 'washed out on the monitor.' At the monthly meeting, a senior technologist proposes raising kVp to 100 and dropping mAs to 2.5 for all portable chests, arguing that this will 'cut patient dose and fix the noise.'
1. What do the deviation index values tell you about detector exposure, and does the senior technologist's proposal address the problem?
Read the worked answer
The deviation index is logarithmic: DI = 10 log10(EI / EI_target). A DI of about -3 means EI is roughly 10^-0.3, i.e. approximately half the target value. The receptor is receiving about half the photons it was set up for, so the mottle is quantum noise from underexposure, not a processing fault.
The proposal works in the wrong direction. Dropping 3.2 mAs to 2.5 mAs removes about 22% of the photons before any other change; raising kVp to 100 increases transmission through the chest, so the net detector exposure is unpredictable and the noise will very likely be worse, not better. It also lowers subject contrast. The proposal treats EI as a dose meter; EI describes exposure at the detector. If the noise level the radiologist wants genuinely requires fewer photons than EI 400 implies, the target should be formally revalidated for that body part, not swapped ad hoc.
2. State the corrective action, and list what you would verify before concluding that technique is at fault.
Read the worked answer
Corrective action: increase mAs in a stepwise fashion (for example 3.2 to 4.5 to 5 mAs), changing one variable at a time, and re-check DI on at least 10 consecutive patients before accepting the new technique.
Before blaming technique, verify the system:
- Correct EI target and correct body-part/protocol menu selected for portable chest.
- Collimation such that the entire EI region of interest is covered by anatomy; unattenuated panel area distorts the histogram and the EI.
- Grid in place and correctly oriented, and no air gap or added material between patient and panel.
- SID actually 100 cm, and the tube angle not shortening it.
- Detector gain/flat-field calibration current, and the panel not shared or recalibrated since the last QC. Only after those are excluded is the mAs genuinely the variable to change.
3. DI is now centred near 0 and the noise complaints have stopped, but the radiologist still calls the images 'washed out' on the reporting monitor. What do you do, and what do you not do?
Read the worked answer
Separate acquisition from display. EI/DI describe how many photons reached the detector; brightness and contrast on the monitor come from the presentation (window/level, VOI LUT) and from the monitor itself.
Do not change technique to darken an image that is correctly exposed - doing so re-introduces underexposure and puts the DI back where it started.
Investigate the display chain: check the default chest window applied by the workstation, view the same study on a second workstation, confirm the diagnostic monitor has a current DICOM GSDF calibration, check ambient lighting and screen cleanliness, and confirm the technologist workstation and reporting workstation are not applying different LUTs. Escalate to the PACS administrator and the QC physicist if the display is out of tolerance.