Battle of the axes: simulation-based assessment of fine needle aspiration biopsies for thyroid nodules

Despite the ubiquitous use of thyroid UGFNAs, to date there has been a paucity of literature comparing the short-axis and long-axis techniques. Proponents of the long-axis technique argue that the improved visualization of the entire needle track deems it easier to perform an UGFNA successfully; whereas, those in support of the short-axis technique note that it requires less hand–eye coordination and less challenging needle angulation [15, 16]. Both techniques provide comparable adequacy of cytologic samples for superficial nodules [17]. This study is the first to compare trainee performance in thyroid UGFNA and provide a visual analysis for the learning curves; furthermore, it demonstrates the possible utility of simulators in thyroid UGFNA training.

FNA performance by technique

This study highlights a high first-pass successful biopsy attempt rate; which outlines that, to begin with, successful UGFNA is technically possible for the majority of participants. The long-axis technique resulted in a higher overall success rate compared to short-axis (89.8% vs 80.7%, respectively), however this did not meet statistical significance. A trend towards higher odds of successful biopsy using the long-axis technique with no difference in procedure duration was observed (OR = 2.2, p = 0.095). This may suggest improved success with this technique, but could not be confirmed with the available data. The ability to visualize the entire needle is a potential explanation for the improved biopsy success with the long-axis technique.

Of note, two of the FNA trials using the short-axis technique had an undetermined outcome due to the difficulty in visualizing the needle tip clearly during evaluation. This emphasizes a known shortcoming in this technique, as the suboptimal visualization makes it challenging to distinguish if a nodule was pass-pointed or missed. Kandil et al. compared outcomes of thyroid UGFNA using both techniques in patients [17]. They reported an increased diagnostic accuracy (95%) of the UGFNAs performed with the long-axis technique particularly in nodules deeper than 3 cm compared to short-axis (p = 0.01). Lesions superficial to 3 cm had comparable yields using both techniques. Studies comparing long-axis and short-axis techniques for peripheral venous access in a simulated setting similarly concluded that the long-axis technique was preferred due to the superior visualization of the entire catheter length, preventing the complication of posterior wall punctures [18, 19].

If a participant was initially unsuccessful in their first attempt, our study found that they required an average of 2.7 ± 1.3 re-attempts to achieve success. It is preferable to minimize the number of biopsy attempts to minimize procedure time and potential risk to surrounding structures. In a systematic review investigating FNA complications, Polyoz et al. reported an overall low risk of complications post-FNA; however, outlined complications including: infection, hematoma, recurrent laryngeal nerve palsy, tracheal puncture, and needle track seeding [20].

Comfort levels

In our post-study survey, participants reflected on their improvement and comfort with thyroid UGFNA biopsies. The majority of the trainees (64%) felt that the long-axis technique was easier/no different to learn, and 71% reported that the long-axis was easier to perform. The post survey analysis suggests that although the long-axis may appear slightly more challenging at the beginning, it becomes the easier technique to use with successive attempts. This is in keeping with previous studies that have demonstrated positive effects of simulator training on resident comfort in performing thyroid UGFNA. Davis et al. assessed residents’ comfort levels in performing thyroid FNAs among 12 senior general surgery residents using a Blue Phantom Ultrasound Central Line Training model© with an embedded olive to simulate a nodule. At baseline, 62% felt “not comfortable” with the procedure; however, after practicing the aspiration technique, this was reduced to 0%. This further demonstrates that residents’ comfort in performing thyroid FNA is positively influenced through training on thyroid simulators [21].

Learning curves

Learning curves were explored visually to gain further insight into UGFNA performance. We noted an increased variability in the performance of novice learners (PGY1-3) when compared to participants in their PGY4 year (Fig. 3a). The degree of unpredictability in PGY 1–3 decreases as trial number increases, suggesting an improvement in their UGFNA and skill acquisition. There was a positive learning trend for novice learners as repeated trials of UGFNAs led to improved speed of biopsy completion. In contrast, the more experienced PGY4 residents, who have performed numerous UGFNA throughout their residency training, had plateauing learning curves irrespective of the FNA trial number. The increased efficiency with repetition supports the use of thyroid simulators to provide an opportunity for additional practice with completing thyroid UGFNAs for residents early in their training.

Simulation model: construct and face validity

Previous literature has also explored and demonstrated the benefit of ultrasound simulation, including Blue Phantom Models ©, in medical education [9, 21,22,23,24,25]. Construct validity is an important step in simulator evaluation and learning curve analysis, the model used in this study successfully differentiated between novice learners and experts (Additional file 1). The model had adequate face validity as the majority of residents responded that the simulator was an accurate representation of a thyroid biopsy experience (Additional file 1).

Limitations

This study has important limitations that restrict the ability to draw broad conclusions based on its findings including limited sample size. Future studies to validate those findings in a larger group of trainees with a retention of skill analysis are needed. Furthermore, to accommodate the surgical residents’ schedules, the time intervals between study trials were variable. The simulated nature of our study design also limits generalization to clinical settings. Studies assessing predictive validity of the model can address that. Future studies will be aimed at correlating simulation findings with clinical performance.

There were drawbacks specific to the simulation model that are relevant. The simulated thyroid model left track marks from previous FNA attempts with continued use. This meant that participants had a visualized needle track to follow, which was not reflective of clinical practice. The authors are currently in the process of creating a more cost-effective, novel thyroid model to address this flaw.

留言 (0)

沒有登入
gif