The Challenge: “Can you please rule out torsion?” is a common request ED teams have of their radiologists and gynecologists. Unfortunately, a straightforward answer to this question is rare. The diagnosis of adnexal torsion is full of uncertainty and to make matters worse, we humans are terrible at communicating uncertainty.
Indeed, there are pathognomonic sonographic findings of torsion– whirlpool sign and/or absent flow in the setting of an enlarged, edematous ovary. But certainty is rare. Thus, many reports hedge that “torsion cannot be ruled out.”
We acknowledge that the radiologist interpreting the images is not at fault for this uncertainty. The issue is that the tool itself is imperfect. Ultrasound, as a test, is great at “ruling things in,” but quite mediocre at “ruling things out.” And torsion, as we know, is a surgical diagnosis. However, going to the OR means subjecting a patient to the potential risks of surgical complications, “tying up” healthcare resources, and is expensive.
The Crux: Most imaging is helpful to “rule in,” NOT “rule out” a diagnosis. The complexity and uncertainty of pelvic ultrasonography in the evaluation of women with acute pelvic pain in and of itself is challenging. On top of that, how do we best communicate the uncertainty of NOT seeing something – like looking for a black cat in a dark room — is it even there?
The language we use in ultrasound reports can further complicate the situation. This is especially true when images are interpreted out of context, and a broad differential diagnosis offered. There are incidental findings in asymptomatic patients that warrant further evaluation in the outpatient setting, and there are others that require emergent evaluation in the correct clinical context. A cyst or mass may be an incidentaloma. Torsion is not.
The verbiage used in reports carries significant weight in clinical decision-making and management. When humans read “cannot rule out xyz,” they usually interpret this (for better or worse) as “xyz should be ruled out.” And so, we would love to start a conversation about the linguistics of report-writing for female pelvic ultrasonography.
Cases: Here are a couple of clinical scenarios that illustrate our concern:
- A 14-year-old female patient presents to the ED with right upper abdominal and flank pain. Her ultrasound was performed to evaluate the kidneys (area where pain was originating from), however, it also demonstrated an enlarged right ovary (4.9 cm in largest dimension). The report reads “intermittent torsion cannot be excluded.” We agree; it in fact cannot. However, now intermittent torsion MUST be excluded, and we are consulted.
The patient’s lower abdominal examination is benign. Our suspicion for torsion is exceedingly low. However, ultrasound cannot rule out torsion, only surgery can rule out torsion. Now this teenager has been given an additional, unrelated stressor (“your ovary can die”) that was unlikely to ever have significant medical repercussions. To top it off, the report recommends a follow-up scan at 8–12 weeks for what appears to be a physiologic hemorrhagic corpus luteum — an additional expense and time taken from the patient and her family to follow-up.
Final Diagnosis: Pyelonephritis
- A 43-year-old female patient scheduled for hysterectomy later in the month presents to the ED with persistent left lower quadrant pain that has been present for several weeks. She has a long-standing history of fibroids and was diagnosed with a 5-cm anechoic left ovarian cyst 2 months ago. Ultrasonography re-demonstrates a leiomyomatous uterus and the left ovary was not visualized.
The report reads “torsion of the left adnexa could not be excluded.” Agreed, it in fact cannot. However, the reason why it cannot be excluded is not that the ovary was not visualized. Additionally, torsion could not be excluded because ultrasound is NOT a test to exclude torsion.
On examination, there was focal tenderness and point-of-care ultrasonography confirmed its location over a pedunculated fibroid (likely degenerating). An overnight, unscheduled diagnostic laparoscopy, in this case, would’ve resulted in a reassuring adnexal evaluation and possible myomectomy, not the procedure the patient truly needed (a laparoscopic hysterectomy).
Final Diagnosis: Degenerating fibroid (noted on hysterectomy later that week)
Prior to COVID-19, healthcare overutilization and defensive medicine were problematic. Now, with limited resources and increased demand, the burden is even higher. ER providers, gynecologists, and radiologists must work in tandem to:
- prioritize imaging studies when relevant,
- report in clear, objective language in the context of the clinical scenario, and
- prioritize emergency and inpatient consultations.
Why does this matter?
Most imaging is helpful to “rule in,” not “rule out” a diagnosis. Language and semantics may significantly affect management, especially in the context of less experienced providers. For patients, it may mean the difference between an unscheduled abdominal surgery or observation. In our prior commentary (1) we referenced the language used by our obstetrics colleagues wherein they acknowledge the limitations of the imaging modality and thus, we suggest the following modification to the current style of reporting:
“Ultrasound is not intended to rule out ovarian torsion.”
We understand that this suggestion, for some, is a change of established practice patterns and we would love to hear your thoughts. Please leave comments below or tweet at @StethoscopeOn and @Dmitry_Fridman to continue the conversation!
- Meljen V, Fridman D. Gynecologist’s Perspective: Semantics of “Ruling Out” Ovarian Torsion. J Ultrasound Med 2020; 39:1013. Available at: https://onlinelibrary.wiley.com/doi/10.1002/jum.15175.
Vivienne Meljen, MD, is a resident, and Dmitry Fridman, MD, PhD, is an Assistant Professor of Obstetrics and Gynecology, in the Department of Obstetrics and Gynecology at Duke University Health System in Durham, North Carolina, USA.
Interested in learning more about gynecologic ultrasound? Check out the following posts from the Scan: