Can you see what I see?

When imaging children, we are all about the ALARA (as low as reasonably achievable) approach. One of the best ways to do this is not to use radiation. Unfortunately, not all of us can be Casey Parker so that we might need some help with our ultrasound skills.

An enterprising study group in Italy has examined whether remote guidance of paediatricians by trained radiologists might help. Here is the paper…

Zennaro F, Neri E, Nappi F, Grosso D, Triunfo R, Cabras F, et al. (2016) Real-Time Tele-Mentored Low Cost “Point-of-Care US” in the Hands of Paediatricians in the Emergency Department: Diagnostic Accuracy Compared to Expert Radiologists. PLoS ONE 11(10): e0164539. doi:10.1371/journal.pone.0164539

What did they set out to do?

They wanted to find out if a paediatrician guided remotely by a radiologist was as accurate as a radiologist at the bedside. Each child entered into the study had three scans within the hour.

• First, by remote control (TELE-POC)
• Secondly, by the same radiologist but at the bedside (UNBLIND RAD)
• Finally by an independent radiologist (BLIND RAD)

They used low-cost, off-the-shelf equipment and open-source software to transmit the images from the ultrasound machine to a radiographer in another room, who was looking at the scan in real time and telling them where to move the probe. The paediatricians were not total novices, however, as they had undertaken a rigorous two-hour course on how to switch on the machine (and look for pertinent findings).

What sort of conditions did they look at?

Their convenience sample looked at children between the ages of 0 and 18 (!) years of age who presented to their Paeds ED during working hours. They didn’t want to study children who were acutely unwell as it would not have been ethical to subject them to three scans. Instead, they chose children for whom they felt an ultrasound might help them turn an equivocal diagnosis into a firm one. They chose eight conditions to look at – and this is where it gets interesting…

• Traumatic abdomen (i.e. a FAST scan)
• Suspected appendicitis
• Suspected intussusception
• Suspected hypertrophic pyloric stenosis
• Suspected pulmonary infection
• Hip pain
• Soft tissues swelling
• Non-specific abdominal pain

What they don’t tell you is the sensitivity or specificity of ultrasound in making the diagnoses.

FAST in kids is as good as flipping a coin with only 50% sensitivity. At least ultrasound has a sensitivity of around 88% and a sensitivity of 94% for detecting acute appendicitis in children when performed by radiologists.

What outcomes were they looking at?

They were looking at three main outcomes.

• Did the radiologist reach the same diagnosis if he performed the scan him/herself or via telemedicine?
• Did the blinded radiologist come to the same conclusion as the tele-mentored paediatrician?
• Did the scans performed by the blinded and unblinded radiologist tally up?

What sort of numbers did they get?

During the six-month study period, 59 children were possible candidates for inclusion, and 52 were included (some refused consent). Eight of these kids were examined for two or three potential conditions. This meant that there were a potential 62 clinical conditions for scanning. Somehow, this works out as 170 ultrasound scans. Here is the distribution of scans taken directly from the paper.

By far, the majority of scans were performed for traumatic injury to the abdomen. If you want to hear Ross Fisher’s rant, listen to St Emlyns’s podcast.

What were their results?

Take a look at Table 3 from the paper.

Of the 170 scans performed, 155 were negative. Understandably, the experts found every positive finding in the tele-mentored group, so there were no false positives. There were three false negatives – a minimal peri-hepatic effusion missed by the mentored paediatrician and a case of intermittent intussusception only spotted by the blinded radiologist.  They go on to derive a Kappa of 0.93 for inter-rater reliability, but with such small numbers of positive findings, this is not generalisable to the real world.

It is also interesting to note that the positive findings include an effusion in the Pouch of Douglas and a distended bladder.

If the detection of free fluid is only 50% sensitive at best, then its inclusion as a positive diagnosis makes little sense.

What were their conclusions?

They concluded that POCUS performed by ED physicians with real-time telementoring produced reliable and timely diagnoses. However, they acknowledged that some of the rarer conditions were not as amenable to study.

What does this mean in the real world?

I like the premise that an expert, sitting in a room across the campus, state, country or even in another continent, could help guide my hand and help me make ultrasound-based diagnoses. Haney Mallemat spoke of the same at SMACCDub.

Intuitively it makes sense, but this study does not provide much evidence for its benefit. No mention of the potential improvement in clinical outcomes for the children studied is made. I am also concerned that over half of the patients had a type of FAST scan for traumatic abdominal injury. We know that FAST in kids is as useful as tossing a coin, and in adults, it has very little place in the haemodynamically stable patient who will end up going to CT anyway.

Will this study change my practice? It is unlikely to. I have utilized a similar approach in my adult retrieval work, using telehealth to look at bedside images, and perhaps in the age of smartphones, this will become more ubiquitous. If the clinician at the bedside can obtain good images then a short video clip can make selling the patient to a receiving unit so much easier.

For the future

I’d like to see the study done on a different patient group. How about a study comparing tele-mentored emergency physicians looking for fractures with radiologists and against the gold-standard radiograph (that would have been performed anyway)?

References

Zennaro F, Neri E, Nappi F, Grosso D, Triunfo R, Cabras F, et al. (2016) Real-Time Tele-Mentored Low Cost “Point-of-Care US” in the Hands of Paediatricians in the Emergency Department: Diagnostic Accuracy Compared to Expert Radiologists. PLoS ONE 11(10):  e0164539. doi:10.1371/journal.pone.0164539

Fox, J.C., Boysen, M., Gharahbaghian, L., Cusick, S., Ahmed, S.S., Anderson, C.L., Lekawa, M. and Langdorf, M.I., 2011. Test characteristics of focused assessment of sonography for trauma for clinically significant abdominal free fluid in pediatric blunt abdominal trauma. Academic emergency medicine, 18(5), pp.477-482.

Scaife ER, Rollins MD, Barnhart DC, Downey EC, Black RE, Meyers RL, Stevens MH, Gordon S, Prince JS, Battaglia D, Fenton SJ. The role of focused abdominal sonography for trauma (FAST) in pediatric trauma evaluation. Journal of pediatric surgery. 2013 Jun 30;48(6):1377-83.