All posts by The Trauma Pro

New Technology: Using AI To Interpret Pelvic X-rays

April 16, 2021 The Trauma Pro Leave a comment

Look out, radiologists! The computers are coming for you!

Radiologists use their extensive understanding of human anatomy and combine it with subtle findings they see on x-ray shadow pictures. In doing this, they can identify a wide variety of diseases, anomalies, and injuries. But as we have seen with vision systems and game playing (think chess), computers are getting pretty good at doing this as well.

Is it only a matter of time until computer artificial intelligence (AI) starts reading x-rays? Look at how good they already are at interpreting EKGs. The trauma group at Stanford paired up with the Chang Gung Memorial Hospital in Taiwan to test the use of AI for interpreting images to identify a specific set of common pelvic fractures.

The Stanford group used a deep learning neural network (XCeption) to analyze source x-rays (standard A-P pelvis images) from Chang Gung. These x-rays were divided into training and testing cohorts. The authors also applied different degrees of blurring, brightness, rotation, and contrast adjustment to the training set in order to help the AI overcome these issues when interpreting novel images.

The AI interpreted the test images with a very high degree of sensitivity, specificity, accuracy, and predictive values, with all of them over 0.90. The algorithms generated a “heat map” that showed the areas that were suspicious for fracture. Here are some examples with the original x-ray on the left and the heat map on the right:

The top row shows a femoral neck fracture, the middle row an intertrochanteric fracture, and the bottom row another femoral neck fracture with a contralateral implant. All were handily identified by the AI.

AI applications are usually only as good as their training sets. In general, the bigger the better so they can gain a broader experience for more accurate interpretation. So it is possible that uncommon, subtle fractures could be missed. But remember, artificial intelligence is meant to supplement the radiologist, not replace him or her. You can all breathe more easily now.

This technology has the potential for broader use in radiographic interpretation. In my mind, the best way to use it is to first let the radiologist read the images as they usually do. Once they have done this, then turn on the heat map so they can see any additional anomalies the AI has found. They can then use this information to supplement the initial interpretation.

Expect to see more work like this in the future. I predict that, ultimately, the picture archiving and communications systems (PACS) software providers will build this into their product. As the digital images are moving from the imaging hardware to the digital storage media, the AI can intercept it and begin the augmented interpretation process. The radiologist will then be able to turn on the heat map as soon as the images arrive on their workstation.

Stay tuned! I’m sure there is more like this to come!

Reference: Practical computer vision application to detect hip fractures on pelvic X-rays: a bi-institutional study. Trauma Surgery and Acute Care Open 6(1), http://dx.doi.org/10.1136/tsaco-2021-000705.

Philosophy

Do I Have To Call My Trauma Team For Incoming Transfers?

April 13, 2021 The Trauma Pro Leave a comment

I had a great question sent in by a reader last week:

Some trauma centers receive a number of transfers from referring hospitals. Much of the time, a portion of the workup has already been done by that hospital. If the patient meets one or more of your trauma activation criteria, do you still need to activate your team when they arrive?

And the answer is: sometimes. But probably not that often.

Think about it. The reason you should be activating your team is that you suspect the patient may have an injury that demands rapid diagnosis and treatment. The purpose of any trauma activation is speed. Rapid evaluation. Fast lab results. Quick access to CT scan or OR. If a significant amount of time has already passed (transported to an outside hospital, worked up for an hour or two, then transported to you), then it is less likely that a trauma activation will benefit the patient.

There are four classes of trauma activation criteria. I’ll touch on each one and the need to activate in a delayed fashion if present, in priority order.

Physiologic. If there is a significant disturbance in vital signs while in transit to you (hypotension, tachycardia, respiratory problems, coma), then you must activate. Something else is going on that needs to be corrected as soon as the patient arrives. And remember the two mandatory ACS criteria that fall into this category: respiratory compromise/need for an emergent airway, and patients receiving blood to maintain vital signs. But a patient who needed an airway who is already intubated and no longer compromised does not need to be a trauma activation.
Anatomic. Most simple anatomic criteria (e.g. long bone or pelvic fractures) do not need a trauma activation unless the patient is beginning to show signs of physiologic compromise. However, anatomic criteria that require rapid treatment or access to the OR (proximal amputations, mangled or pulseless extremities, spinal cord injury) should be activated.
Mechanism. Most of the vague mechanistic criteria (falls, pedestrian struck, vehicle intrusion) do not require trauma activation after transfer to you. But once again, if the mechanism suggests a need for further rapid diagnosis or treatment (penetrating injury to abdomen), then activate.
Comorbidities. This includes underlying diseases, extremes of age, and pregnancy. In general, these will not require trauma activation after they arrive.

Bottom line: In many cases, the patient transferred in from another hospital will not need to be a trauma activation, especially if they have been reasonably assessed there. The patient should be rapidly eyeballed by your emergency physicians, and if there is any doubt about their condition, activate then.

However, if little workup was done at the outside hospital (my preference), and the injuries are “fresh” (less than a few hours old), then definitely call your team.

Thorax

Delayed Hemothorax In Older Adults: Real Or Not?

April 6, 2021 The Trauma Pro Leave a comment

I came across an interesting paper in the Journal of Trauma & Acute Care Surgery Open recently. I always read these articles a bit more critically, though, because the peer review process just doesn’t feel quite the same to me as the more traditional journal process. But maybe it’s just me.

In this paper, the authors decided to look at the incidence of delayed hemothorax because “emerging evidence suggests HTX in older adults with rib fractures may experience subtle hemothoraces that progress in a delayed fashion over several days.” They cite two references to back up this rationale.

They retrospectively reviewed records from two busy US Level I trauma centers for adults age 50 or older who were diagnosed with delayed hemothorax (dHTX). Delayed was defined as 48 hours or more after initial chest CT showed either a minimal or trace HTX. The authors went on to analyze the characteristics and demographics of the patients involved.

Here are the factoids:

A total of 14 older adults experienced dHTX after rib fractures, an overall incidence of 1.3% (!)
About half were diagnosed during the initial hospitalization for the fractures
All patients had multiple fractures, with an average of 6 consecutive ones; four had a flail chest
One third progressed from a trace HTX, two thirds had a completely negative initial chest CT
Only one third were taking anticoagulants or anti-platelet agents
Patients with multiple fractures, posteriorly located, and displaced were most likely to develop dHTX

The authors concluded that “delayed progression and delayed development of HTX among older adults with rib fractures require wider recognition.”

Bottom line: Really? First, I looked at the papers cited by the authors as the rationale for doing this study. They each found dHTX in about 10% of patients, but their definition was very broad: any fluid visible on upright chest x-ray. Furthermore, the patients were not really “older” either. Average age was around 50.

So I’m not sure yet whether this is a problem, especially with the low incidence of 1.3%. This study doesn’t come right out and state how many patients they reviewed to find their 14, but it can be calculated to be 14 / 1.3% = 1,177. This incidence is only one tenth of that found in the two studies cited. Seems relatively uncommon, and half were discovered while the patients were still in the hospital. Thus only 0.65% sought readmission for chest discomfort or difficulty breathing.

This study required chest CT for rib fracture diagnosis. Is all that radiation (and possibly contrast) really necessary? And did these patients get another chest CT to delineate the pathology? More radiation?

Overall, this paper was not very helpful to me. Yes, I have seen patients come back days or weeks later with a hemothorax that was not seen during their first visit. It’s just that this study raises many more questions that should have been easily answered in the discussion. But they weren’t.

Given that only about a half of a percent of rib fracture patients develop delayed hemothorax after discharge, it is probably prudent to provide information to the patient recommending they see their practitioner if they develop any symptoms days or weeks later. And a simple chest x-ray should do.

Reference: Complication to consider: delayed traumatic hemothorax in older adults. Trauma Surgery & Acute Care Open 2021;6:e000626. doi: 10.1136/tsaco-2020-000626.

Philosophy

What Is The Curbside Consult? And The ELEVENTH Law of Trauma!

April 2, 2021 The Trauma Pro Leave a comment

Surgeons, I’m sure you’ve had an experience something like this at some point:

You happen to be wandering through the emergency department and one of your Emergency Medicine colleagues approaches you and says, “Hey, I ‘ve got this patient I’m seeing that I just want to run by you…”

How should you deal with this? They want a quick tidbit of information to help them decide what to do with the patient. Can they send them home, or should they “formally” consult you?

It’s important to look at the pros and cons of this practice. First the pros:

It’s direct. You’re right there. No phone calls, no paging.
It’s quick. Just a quick description of the problem, and a prompt answer. Then everyone can get on with their business.

But then there are the cons:

Situational accuracy. The consultee has not seen the patient, so the information they have been given was filtered through the consulter. Any number of cognitive biases are possible, so the real story may not be exactly as it seems.
Interpretation of the recommendation. Other cognitive biases are also possible as the consulter acts on and implements the recommendations of the consulter. Have they really been followed?
Lack of documentation. This is the biggest problem with a curbside consult. The consultee may act without documenting the source of the recommendation. Or, they may document that they spoke with Dr. Consultee. In either case, one or the other may be hung out to dry, so to speak.

Consider what happens if there is a complication in the care of that patient. There is no way to really determine what was said during that conversation a week or two years later. It boils down to recollections and may end up as a he said … she said situation. And in the worst case scenario, if such a case were to enter the medicolegal arena, there is no official record that any recommendation was made or followed. It’s a very easy case for the plaintiff’s attorney to prevail.

So this leads to my new Eleventh Law of Trauma:

“Work not documented is work not done“

Bottom line: There is no such thing as a curbside consult! The consultee should say, “I’d better take a look at this patient, why don’t you officially consult me?”

In doing this, the consulter gets to use their own clinical and cognitive skills, and thus render a real opinion based on first hand experience. The consultee gets the most accurate recommendations possible, and they are noted in the record so there is no room for misinterpretation. And finally, there is good documentation from both that will stand up in a court of law if needed.

Home of the Trauma Professional's Blog