Radiologists can detect evidence of intimate partner violence before women disclose it and help facilitate earlier interventions to reduce cycles of abuse, according to a December 3 presentation at RSNA.
In a review of imaging from 1,322 cases, radiologic evidence of intimate partner violence was often present months or years before it was first self-reported, noted presenter Patrick Lenehan, MD, PhD, a clinical fellow at Massachusetts General Hospital in Boston.
“Our overall takeaway and hope is that awareness of characteristic imaging patterns among radiologists can help us to detect [intimate partner violence] actually before disclosure and facilitate earlier interventions that can help break cycles of abuse for these patients,” Lenehan said.
Intimate partner violence is a serious public health concern, with recent U.S. Centers for Disease Control and Prevention (CDC) estimates suggesting up to 40% of women in the U.S. have experienced some form of it -- physical, sexual, or psychological -- in their lifetime. While it can be difficult to detect in the clinical setting, previous work has demonstrated that radiologists can raise suspicion for physical intimate partner violence before it is self-reported, Lenehan noted.
In this study, Lenehan and colleagues aimed to validate this observation on an expanded cohort. The study population consisted of women who reported partner violence to Mass General’s violence prevention support program between 2013 and 2018 (“cases”; n = 1,322 patients with 48,792 imaging studies) and age- and race-matched women who did not report violence (“controls”; n = 6,693 patients with 102,752 imaging studies).
Two emergency radiologists reviewed each study and indicated whether intimate partner violence was suspected based on observed injuries and overall findings. Rates of suspected partner violence between cases and controls were compared using a crude odds ratio and Fisher-exact test.
In addition, the potential for early detection was assessed by summarizing the distribution of time between the date of initial self-reporting and the date on which a radiologist first raised suspicion of partner violence.
According to the results, suspicion was raised on a total of 433 out of all 151,544 (0.29%) radiology reports reviewed, including 250 reports from cases and 183 reports from controls. At the patient level, intimate partner violence was suspected by radiologists at any time for 159 of 1,437 cases versus 129 of 6,693 controls (11.1% vs. 1.9%).
Among the 159 cases for which partner violence was suspected, the first radiologic evidence occurred a median of one day before initial self-reporting (IQR: 1,865 days), while 64 of these cases (40.3%) were suspected by radiologists at least six months prior to self-reporting, Lenehan said.
In addition, upper and lower extremity fractures were among the most common injury patterns that raised suspicion for both groups, while facial fractures contributed disproportionately for cases (17 of 250 reports, 6.8%) compared with controls (2 of 183 reports, 1.1%; p = 0.004).
“For facial injuries, we do see that there's a significant enrichment towards more of these injuries being observed in the case population preceding the initial suspicion of [intimate partner violence],” Lenehan said.
Ultimately, the study builds on previous work by researchers at Mass General, as well as Brigham and Women’s Hospital Trauma Imaging Research and Innovation Center, Lenehan noted.
“[The study] demonstrates the potential for radiologists to detect physical or nonphysical [intimate partner violence] and also provides some additional insight by including a control cohort so that we were able to assess specificity and apply this in a more translationally relevant way,” he concluded.
For our full coverage of this year’s meeting, visit our RADCast.
