Introduction Characterising the shapes, dimensions and overall numbers of fragments produced by explosive devices is important for determining methods of potential mitigation, such as personal armour. The aim of this investigation was to compare the mass of excised fragments with that predicted from CT to ascertain the validity of using such an approach to measure retained fragments for multiple body areas using CT alone.

Method 27 retained fragments excised from consecutive patients treated at a US Role 3 Medical Treatment Facility in Afghanistan were examined. Each fragment was measured in three dimensions and the mass was obtained to estimate the density and thereby probable composition. These same excised fragments were identified radiologically and their predicted masses calculated and compared with the known masses with a paired t-test. The total numbers of retained fragments in each of four body areas for 20 casualties were determined radiographically and the mass of the largest fragment in each body region estimated.

Results Excised fragments were most commonly metallic (17/27, 63%), with masses ranging from 0.008 to 37.6 g. Mean mass predicted from CT was significantly different from than that measured (p=0.133), with CT underestimating true mass by 5%–17%. 889/958 (93%) retained fragments appeared metallic on imaging, with the most commonly affected body areas being the torso and upper extremity (45% of casualties).

Conclusions Predicting the mass of metallic fragments from CT was possible with an error margin of up to 5%, but was less accurate for non-metallic fragments such as stone. Only 3% of fragments were removed through debridement or purposeful excision; these were not just the largest or most superficial. This suggests that future retrospective analysis of the dimensions and predicted masses of retained fragments in larger casualty cohorts of service personnel is potentially feasible within a small margin of error.