Page 97 - JSOM Winter 2024

For the detection of wooden FBs, overall sensitivity was With a nearly identical study design to ours, Fleming et al.
71.8% (95% CI 50.7–85.7), and specificity was 82.0% (95% evaluated the accuracy of emergency physicians (including
CI 61.1–92.6). The smallest FB (1.0mm) was the most diffi- resident, fellow, and attending physicians with various levels
cult to identify with 40% accuracy and an overall sensitiv- of ultrasound experience) using ultrasound to detect wooden
ity of 44.4% (95% CI 24.6–66.3). The largest FB (10.0mm) FBs measuring 1, 2.5, 5, 7.5, and 10mm, placed at a depth of
13
was identified with 95% accuracy and an overall sensitivity of 10mm and 30°angle, in a chicken tissue model. A total of 50
95% (95% CI 76.4–99.1). emergency physicians had a sensitivity of 48.4% and specificity
of 67.6% in detecting wooden FBs; sensitivity did not change
as object size increased. These results differ significantly from
Discussion
our data, which may be partly because of a depth of 10mm
The use of ultrasound has greatly expanded the capabilities compared with 5mm in our study. It must also be noted that
of both civilian emergency medicine and military medicine, their study allowed no more than 30 seconds of scanning ver-
particularly in austere environments. Once used primarily sus the 3 minutes allotted in our study. Lastly, the better results
11
by radiology and niche subspecialists, ultrasound has now be- in our study may have been influenced by the recent training
come a widely available and versatile tool for military med- rather than being a true indication of the medics’ long-term
ics, specifically SF medics who are often deployed as the sole proficiency with the technique; our study contained didactics
22
provider on the battlefield. Their ability to assess and treat and training immediately prior to performing the tests, whereas
the vast majority of conditions has significant impact on their physicians in the Fleming et al. study relied on earlier training.
mission. In 2008, the Special Operator Level Clinical Ultra- This is contrasted with a study by Nienaber et al. that had 20
sound ( SOLCUS) program was created and incorporated emergency physicians, including trainees, identifying various
into the Joint Special Operations Medical Training Center soft tissue FBs without time constraints, and revealed sensitivi-
20
(JSOMTC) with didactic and hands-on training for Special ties of 85%–96% and specificities of 70%–83%. Despite the
Operations Combat Medics. The curriculum emphasizes increasing use of POCUS training among emergency physicians,
23
trauma evaluations, such as extended, focused assessment in training to identify FBs is probably still very limited, and these
trauma (EFAST), from identifying immediate life-threats to ED physicians may not have been familiar with performing this
identifying long-bone fractures and measuring optic nerve type of scan prior to the study. The depth, time-constraints, and
sheath diameters, with some training in procedural use for ul- variable skills of their sonographers likely contributed to differ-
trasound-guided vascular access and regional anesthesia. As ences in pooled specificity and sensitivity.
23
our wars move further into prolonged casualty care situations
in austere environments with limited ability for evacuation, the Our study did have some limitations. First, the sample size was
ultrasound skills of military medics continue to expand. With relatively small at 20 participants and 200 total scans, which
military servicemembers’ exposure to dangerous environments may affect the generalizability of the findings to larger pop-
and unique injuries, a medics’ ability to identify retained soft ulations. Also, participants in our study had difficulty iden-
tissue FBs may lead to saved lives and successful missions. tifying the 5-mm FB. In a post-test review, it was noted that
while the 5-mm model was at an appropriate depth, it had a
Our primary objective was to evaluate the ability of SF med- significantly steep angle of >45°, which limited its echogenic-
ics to detect wooden FBs of varying sizes by ultrasound in a ity and posterior acoustic shadowing and was likely difficult
chicken model. The results indicate that these medics, with for novice sonographers to detect during their scans. Addi-
minimal training, can effectively use POCUS to detect FBs. tionally, there were various false positives in which multiple
The accuracy rates observed in this study suggest that POCUS participants detected an FB in at least two of the models that
can be a reliable tool in the hands of specially trained non- did not contain any FB. Post-test scans were performed and
physician providers and enable accurate identification of revealed small calcifications that ran along or near a likely
wooden FBs, particularly those measuring ≥7.5mm. Our study tendon sheath or fascial plane within the models. FBs must
revealed sensitivity and specificity of 73% and 76%, respec- be differentiated from anatomic hyperechoic structures, such
tively, in the detection of FBs of all sizes. as bones, tendons, fascia, scar tissue, and other calcifications,
within the body; this study shows that our participants had
Our results are consistent with a large systematic review and difficulty in differentiating anatomic echogenicity from foreign
meta-analysis performed by Davis et al., which found pooled substance. 25–27 Lastly, the use of chicken thigh models, though
sensitivity of 72% and specificity of 92% across 17 studies eval- previously validated, may also limit generalizability to human
uating soft tissue FB detection by ultrasound. It is likely that soft tissue.
24
our specificity was impacted by false positives, in which several
participants incorrectly identified an anatomic tendinous calcifi- Our findings reveal areas for potential future research. Across
cation as an FB in two models that had not been manipulated at all these studies evaluating wooden FBs, data seem to vary de-
all by our study team. Similar detection rates were identified in pending on FB depth. Future studies may need to focus on
12
a prospective observational study by Driskell et al. Their study detection of wooden FBs at varying depths rather than just of
consisted of 28 army medics (non-special forces) that completed different sizes. Additionally, participants in our study demon-
540 ultrasound scans to detect wooden FBs measuring 1, 2, and strated good accuracy in detection after a 1-hour lecture on
3mm at a depth of 10mm and a 45° angle in chicken thigh mod- soft tissue FB detection. Emergency trainees, military medics,
els. Their results were shown to have a sensitivity of 73% and and other novice sonographers would likely benefit from short
specificity of 78% and overall had better accuracy at very small courses or ultrasound curriculums. Refresher courses for mili-
FB sizes and likely with less ultrasound familiarity than the SF tary special operations medics would likely assist in achieving
medics in our study. However, Driskell et al. did not impose a better detection rates to maintain skill retention among their
time limit on the scanning of each model, and it was noted that non-trauma medical skills. It may also be beneficial to directly
participants had a smaller surface area of chicken thigh to scan. compare the diagnostic accuracy of SF medics using POCUS

SF Medics: Identifying Wooden Foreign Bodies by POCUS | 95