A quick review of FAST
Thanks to Alan for his look at the utility of FAST exams, incorporating a pioneering study of FAST as well as some more recent work. The three papers he picked are below:
- Rozycki et al. Surgeon-Performed Ultrasound for the Assessment of Truncal Injuries: Lessons Learned From 1540 Patients Ann Surg. 1998;228(4):557-67.
- Miller et al. Not so Fast J Trauma. 2003;54:52–60.
- Schnuriger et al. The accuracy of FAST in relation to grade of solid organ injuries: A retrospective analysis of 226 trauma patients with liver or splenic lesion. BMC Medical Imaging 2009, 9:3.
For a review of these papers, and some resources for FAST, see below!
The Rozycki paper was a review of 1540 prospectively enrolled patients with blunt abdominal trauma (BAT) or precordial wounds, excluding patients in extremis. Surgeons who completed a training course and 50 supervised exams performed the ultrasound exams. Focusing on the 1197 normotensive BAT cases, if the FAST was positive the patient got a CT. Negative FASTS led to repeat scans and observation (remember, this was the 90’s. Also if scan was equivocal or patient deteriorated, they’d get CT or DPL). They found 16 false-negative FAST exams and 2 false-positives (there were 1129 true negatives and 50 true positives). This led to a calculation of 75.7% sensitivity and 99.8% specificity, and positive LR of 25 with negative LR of 0.014.
What were those false negatives? For the most part, understandable injuries such as low-grade spenic and liver lacs and bowel injuries that we know FAST isn’t great at detecting. 15 out of these 16 survived and all had indications for further imaging or DPL.Of the 30 BAT patients with hypotension, they had no false-negatives and no false positives — 8 had true positives with generally higher grade solid organ injuries and they had 22 true-negatives – though it’s a small sample that precludes meaningful calculations of likelihood ratios, can’t argue that FAST was 100% sensitive and specific in the setting of BAT with hypotension.
The second paper, by Miller, was entitled “Not So Fast” — sadly that was the most clever part of the study. Miller’s group looked at 359 FAST exams performed by surgery residents with unclear levels of training with confirmatory CT within an hour (hypotensive patients and inadequate studies were excluded). They found 22 patients with false-negative FASTs (mostly low-to-mid-grade splenic and liver lacs, bowel perfs and mesenteric injuries). They also went to great lengths to document incidental findings by CT such as polyps and gallstones. Most of these false-negatives and incidentalomas were not treated operatively.
Not only was FAST not designed to catch this stuff, but their estimate of FAST sensitivity was hurt because 5 patients with positive FAST ended up deteriorating and going to the OR, skipping CT.
Their completely unjustified conclusion, that CT since superior to FAST in normotensive patients, negative FASTs should be followed by CT, was critiqued nicely in the panel discussion section by Rozycki. It’s a long except, but really worth it:
I have no questions for the authors, just suggestions. First, redo your hypothesis. Second, clarify the entrance criteria by which the patients underwent FAST examination to ensure that it was indicated……Third, determine the associated injuries of the 22 patients whose examination showed false-negative results and apply the recommendations of Chiu and Ballard for the CT scan. Fourth, restructure your practice on the basis of accepted standards so the FAST examination is used appropriately in your institution.
The FAST is an extension of the physical examination, not its replacement. Ultrasound, diagnostic peritoneal lavage, and CT scanning have excellent roles in the evaluation of the injured patient. Their individual or combined use is discussed in multiple studies and should be applied appropriately so that sound clinical judgment is not overwhelmed by enthusiasm for technology.
The last paper, by Schnuriger, was a review of 226 hemodynamically stable Swiss patients with known liver or splenic lacs seen on CT. CT was done for impaired sensorium or unclear abdominal findings, and Schnuriger’s group went back and looked at the result of the FAST study, which was performed at some point during the workup by a senior radiology resident. How many FASTs were negative on these patients? Well, a fair amount — 45 of the 226 patients with known injuries had negative initial FAST exams, for a sensitivity of 80%.
We couldn’t figure out how they calculated specificity or likelihood ratios, though, because there were no true negatives or false negatives; inclusion criteria guaranteed every one of the 226 patients had some spleen or liver injury. The authors also go on to make conclusions about the role of FAST in unstable patients, even though this study had none. And, for a recent paper, these authors continue the troubling trend of studying shortcomings of FAST that were acknowledged fifteen years ago. But they do find that high-grade solid organ injuries, the kind that tend to get operated on, are more likely to yield positive FAST — grade IV and V lacs had 100% sensitivity here.
The Miller and Schnuriger papers, written by surgeons and radiologists, are part of a new and troubling trend — some folks aretrying make FAST into something it’s not, and then saying it does a lousy job at it. So when our colleagues in surgery or radiology argue that FAST is worthless, or that we shouldn’t make decisions based on it, well, remember the words of the first author, the surgeon Rozycki:
A major concern with the use of any diagnostic modality is the potential for missed injuries. Can the FAST examination consistently detect intra-abdominal injury? Of course not, it was not developed to do that. From the earliest studies to those most recently in print, the FAST is defined as a focused examination that sequentially surveys the pericardium and abdomen for the presence or absence of blood. Therefore, it is difficult to compare its results to those of the CT scan because a FAST examination does not readily identify intraparenchymal or retroperitoneal injuries. This is not new, as it was recognized as a drawback of the FAST examination in many of the sentinel investigations.
It’s also worth remembering the ACEP clinical policy on blunt abdominal trauma –Their level B recommendation, based on a nice lit review (that including Rozycki’s paper), was that FAST is useful as an initial screening examination to detect hemoperitoneum in blunt abdominal trauma patients.
Be sure to check out the images and tutorial on FAST (and other ultrasound topics) at SinaiEM.us. Also some good stuff at sonoguide.com.
Posted
on Sunday, January 24th, 2010 at 6:39 pm by Nick. Filed under
Trauma, Ultrasound, Journal Club, Radiology, Blog.
You may post a comment.
The core controversies surrounding the FAST center on exactly what you note in the write up:
1. Does +FAST= need for laparotomy, or does +FAST=hemoperitoneum? Using the first criteria it’s a crappy test, almost as bad as DPL, which has a 17% non-therapeutic laparotomy rate when used alone as an indicator for OR. The second criteria is much more reasonable, specially in an era of non-operative strategies for splenic and liver injuries.
2. Who is doing the study? There is a huge difference between attending radiologists performing FAST exams with $300,000 Acuson Sequoia ultrasound machines compared to surgical residents using first-generation SonoSite 180 machines. Unfortunately, the latter is the reality at most centers and is worth studying for the purposes of generalizability. Typically, “pioneer” studies of new technologies, techniques, etc. are performed by lite operators, and the same results will not be found when the rest of the world first adopts the techniques. A combination of learning curve and technologic advancement brings everyone back up to a higher level over time.
Comment by BN on January 24th, 2010 at 6:41 pm