Ultrasound has been shown to be superior to supine chest x-ray in the diagnosis of pneumothorax, with one recent systematic review demonstrating 91% sensitivity using ultrasound compared to 50% using chest x-ray.1 CT scan remains the gold standard in diagnosis but is often not feasible in unstable trauma patients. Ultrasound is recommended by ATLS guidelines for use in trauma patients as part of the eFAST protocol. There are three main described ultrasound findings in pneumothorax: lung sliding, B-lines, and the lung point. While B-mode (2D mode) is commonly described, many resources also suggest the use of M-mode (motion mode).
This study evaluates whether the addition of M-mode to B-mode impacts the sensitivity, specificity, and accuracy of bedside ultrasound in the diagnosis of lung sliding. Previous studies have evaluated the accuracy of M-mode on cadaveric subjects2, but no previous studies have investigated the accuracy of M-mode + B-mode compared to B-mode alone in live human subjects.
Does adding M-mode to B-mode improve accuracy in diagnosing pneumothorax?
Avila, J et al. Does the Addition of M-Mode to B-Mode Ultrasound Increase the Accuracy of Identification of Lung Sliding in Traumatic Pneumothoraces?. J Ultrasound Med, 37: 2681-2687
Does the addition of M-mode to B-mode improve accuracy in identifying lung sliding? Does this vary by ultrasound experience and level of training?
Methods & Study Design
Emergency Physicians including residents, fellows, and attending physicians
Hemithorax anterior lung field ultrasound scans were performed on 15 patients who had a unilateral pneumothorax confirmed by CT scan. B-mode and corresponding M-mode images were obtained for each patient, with one scan on each side, producing scans of 30 lungs. These images were incorporated into a 30-question quiz in which respondents were asked to identify the presence or absence of lung sliding. One version of the quiz contained B-mode clips alone and one version contained B-mode and M-mode clips for each lung. Respondents were randomized to one of the two quizzes. The quiz was sent to EM residency directors for distribution. One hundred forty physicians responded and were randomized.
Sensitivity, specificity, and accuracy of the diagnosis of lung sliding, and association with respondent ultrasound experience and level of training.
Overall, the addition of M-Mode to B-Mode resulted in unchanged sensitivity, 93.1% vs 93.2%, improved specificity from 89.8% to 96% (P < 0.0001), and improved accuracy from 91.5% to 94.5% (p=0.0091).
In subgroup analysis, there was no significant difference in accuracy, sensitivity, or specificity when adding M-mode for physicians with more than 250 ultrasound scans previously performed. For physicians with less than 250 total scans previously performed, use of B-mode + M-mode increased accuracy from 88.2% to 94.4% (P = 0.001) and increased specificity from 87.0% to 97.2% (P < 0.0001) compared with B-mode alone. For resident physicians, the addition of M-mode to B-mode significantly improved accuracy from 89.6% to 94.6% (P = 0.0016) and specificity from 87.9% to 95.9% (P < 0.001) for resident physicians. There was no significant improvement for fellows and attending physicians.
Strengths and Limitations
The authors describe methods in detail, including how the ultrasound scans were performed, number of sites scaned, and the type of machine, probe, and settings used. They also collected detailed information on level of ultrasound experience which helps generalize results among emergency physicians with varying levels of ultrasound experience. Ultrasound results were compared to the gold standard of CT scan.
The survey was sent out to residency program directors to distribute to residents, fellows, and attendings, which excludes the large number of practicing emergency physicians in the community. Community physicians may have different levels of experience and formal training with ultrasound and would be an important group to include in terms of study generalizability. Additionally, the sample size was relatively small (140 total participants) and included many more residents (92) than fellows/attendings (48). The images used also did not capture the absence or presence of B-lines, which could also impact interpretation and management. This study evaluated interpretation only and did not evaluate image acquisition, which could impact the outcomes measured and would be more helpful for practical application. Finally, there may be a difference in clinical significance between pneumothorax diagnosed with x-ray or bedside ultrasound versus CT scan– CT may identify more smaller and less clinically relevant pneumothoraces which may be missed on ultrasound.
“The addition of M-mode images to B-mode clips aids in the accurate diagnosis of lung sliding by emergency physicians. The subgroup analysis showed that the benefit of M-mode US disappears after emergency physicians have performed more than 250 US examinations.”
The addition of M-mode to B-mode can improve accuracy in identifying lung sliding when evaluating for pneumothorax when performed by emergency physicians with less training or ultrasound experience. Given this benefit, more junior physicians could be encouraged to add M-mode to their evaluation for pneumothorax, especially as the additional image acquisition required is relatively quick.
The Bottom Line
Adding M-mode to B-mode when using ultrasound to evaluate for pneumothorax improved accuracy amongst emergency physicians with less US experience.
This post was written by Julie Westover, MS4 at UCSD School of Medicine, with editing by Ben Liotta, MD and Amir Aminlari, MD.
Avila, J., Smith, B., Mead, T., Jurma, D., Dawson, M., Mallin, M. and Dugan, A. (2018), Does the Addition of M-Mode to B-Mode Ultrasound Increase the Accuracy of Identification of Lung Sliding in Traumatic Pneumothoraces?. J Ultrasound Med, 37: 2681-2687. https://doi.org/10.1002/jum.14629
1. Alrajhi K, Woo MY, Vaillancourt C. Test characteristics of ultrasonography for the detection of pneumothorax: a systematic review and analysis. Chest 2012; 141:703–708.
2. Adhikari S, Zeger W, Wadman M, Walker R, Lomneth C. Assessment of a human cadaver model for training emergency medicine res- idents in the ultrasound diagnosis of pneumothorax. Biomed Res Int 2014; 2014:724050