Keywords:
Trauma, Embolisation, CT-Angiography, Catheter arteriography, Interventional vascular, Pelvis
Authors:
G. Harisis, J. Lee, W. Clements, G. S. Goh; Melbourne/AU
DOI:
10.1594/ranzcr2018/R-0036
Results
Patient data
Table 1 provides a summary of patient data from the different scanning methods between the first and second 3 year periods.
The median time to DSA following the initial trauma CT (rather than mean) was used as a better representation of the central value.
Radiologist accuracy in identifying bleeding of any source
Table 2 provides a summary of individual radiologist accuracy in the diagnosis of active pelvic bleeding of any source.
Sensitivity for the identification of active bleeding was high for both radiologists while specificity was comparatively low.
This was the case for both DB-CT and PV-CT.
Figures 2 and 3 show examples of active bleeding identified on PV-CT and DB-CT.
Radiologist accuracy in identifying arterial bleeding
Table 3 provides a summary of individual radiologist accuracy in the diagnosis of active pelvic bleeding that was identified as being arterial in origin and hence would necessitate formal DSA +/- embolisation.
Sensitivity for the identification of active arterial bleeding was variable between the reviewers,
but overall lower than that for identification of bleeding in general.
This was the case for both DB-CT and PV-CT as evidenced by the overlapping confidence intervals.
NB:
True positive |
Defined as the reviewer calling ‘arterial bleeding’ or ‘can’t tell’ on CT with no evidence of active arterial extravasation,
AV fistula or severe vasospasm of the artery in question
|
True negative |
Defined as the reviewer calling ‘venous bleeding’ or ‘no bleeding’ on CT with subsequent DSA showing no evidence of active arterial extravasation,
AV fistula or severe vasospasm of the artery in question
|
False positive |
Defined as the reviewer calling ‘arterial bleeding’ or ‘can’t tell’ on CT with no evidence of active arterial extravasation,
AV fistula or severe vasospasm of the artery in question
|
False negative |
Defined as the reviewer calling ‘venous bleeding’,
‘no bleeding’ or ‘can’t tell’ on CT with subsequent DSA showing evidence of active arterial extravasation,
AV fistula or severe vasospasm of the artery in question
|
Radiologist inter-observer agreement
Table 4 summarises the inter-observer agreement for the diagnosis of arterial bleeding in PV-CT vs DB-CT based on the un-weighted kappa statistic.
Agreement was marginally better for DB-CT.