Non-visualization of the ovary on CT or ultrasound in the ED setting: utility of immediate follow-up imaging.

Purpose: In the setting of female patients in the ED receiving both a CT scan and ultrasound of the pelvis, we compared the studies to assess the diagnostic value of non-visualization of the ovary in predicting the absence of ovarian disease.

Methods: A retrospective review of the electronic medical record identified adult female non-pregnant patients who received both a CT abdomen and pelvis and a transabdominal and transvaginal pelvic ultrasound within a 48-h period through the emergency department. The imaging studies were blindly reviewed to evaluate the ovaries on each CT and US independently. Ovaries were characterized as normal, non-visualized, or abnormal requiring follow-up. All cases of non-visualized ovaries were compared to the appearance on the corresponding exam for each patient, and assessed for agreement (normal or non-visualized) or disagreement (abnormal). Subsequently, the electronic medical record was used as a reference standard to assess the presence or absence of ovarian pathology over the subsequent ninety days. Statistical analysis was performed utilizing a McNemar test with a 95% confidence interval.

Results: A total of 181 patients were included in the study (age range 18-75). Of these, 36 patients had non-visualization of one ovary (21 patients) or both ovaries (15 patients) on either study. On CT, 18 ovaries were non-visualized, with no cases of disagreement on ultrasound or subsequent short-term follow-up (NPV 100%). On ultrasound, 39 ovaries were non-visualized, with three of these adnexae appearing abnormal on CT (NPV 92%): one with an enlarged ovary and two with hydrosalpinx. No ovarian abnormalities were evident on chart follow-up. Significant correlation of agreement between normal and non-visualized pairs on CT and US was observed (p = 0.025, McNemar). There was a combined NPV of 94% regardless of modality for the absence of ovarian pathology in the case of non-visualization. The most common abnormal imaging findings associated with non-visualized ovaries were uterine fibroids and hemoperitoneum.

Conclusion: The absence of detection of the ovary on pelvic US or CT is highly predictive of the lack of ovarian abnormality on short-term follow-up, and does not typically require additional imaging to exclude ovarian disease.