Generating colloquial radiology reports with large language models.

OBJECTIVES: Patients are increasingly being given direct access to their medical records. However, radiology reports are written for clinicians and typically contain medical jargon, which can be confusing. One solution is for radiologists to provide a "colloquial" version that is accessible to the layperson. Because manually generating these colloquial translations would represent a significant burden for radiologists, a way to automatically produce accurate, accessible patient-facing reports is desired. We propose a novel method to produce colloquial translations of radiology reports by providing specialized prompts to a large language model (LLM). MATERIALS AND METHODS: Our method automatically extracts and defines medical terms and includes their definitions in the LLM prompt. Using our method and a naive strategy, translations were generated at 4 different reading levels for 100 de-identified neuroradiology reports from an academic medical center. Translations were evaluated by a panel of radiologists for accuracy, likability, harm potential, and readability. RESULTS: Our approach translated the Findings and Impression sections at the 8th-grade level with accuracies of 88% and 93%, respectively. Across all grade levels, our approach was 20% more accurate than the baseline method. Overall, translations were more readable than the original reports, as evaluated using standard readability indices. CONCLUSION: We find that our translations at the eighth-grade level strike an optimal balance between accuracy and readability. Notably, this corresponds to nationally recognized recommendations for patient-facing health communication. We believe that using this approach to draft patient-accessible reports will benefit patients without significantly increasing the burden on radiologists.

Imaging uterine lipoleiomyomas: A case series and review of the literature.

Background: Lipoleiomyomas are uncommon uterine lesions containing adipose and smooth muscle tissue. They have a variable presentation and are usually found incidentally on imaging or post-hysterectomy tissue analysis. Given their low prevalence, there is a dearth of literature describing imaging characteristics for uterine lipoleiomyomas. In this image-rich case series, we summarize an example of an initial presentation as well as present ultrasound, computed tomography (CT), and magnetic resonance imaging (MRI) findings for 36 patients. Case presentation: We present the detailed clinical course of a representative patient evaluated for uterine lipoleiomyoma and describe imaging findings seen in another 35 patients. This includes ultrasound findings from 16 patients, CT findings from 25 patients, and MRI findings from 5 patients. Among the 36 total patients, symptoms at the time of diagnosis were variable but often included abdominal or pelvic pain; however, most patients were asymptomatic, and the lipoleiomyomas were incidentally discovered on imaging. Conclusions: Uterine lipoleiomyomas are rare and benign tumors with variable presentations. Ultrasound, CT, and MRI findings can assist in diagnosis. Findings on ultrasound typically include well-circumscribed hyperechoic and septated lesions with minimal to no internal blood flow. CT shows fat-containing either homogeneous or heterogeneous circumscribed lesions depending on their ratio of fat and smooth muscle tissue. Lastly, on MRI, uterine lipoleiomyomas commonly appear heterogenous with loss of signal on fat-suppressed sequences. These imaging findings are highly specific for lipoleiomyomas, and familiarity with these findings may reduce unnecessary and potentially invasive procedures.

Artificial Intelligence for Improved Hepatosplenomegaly Diagnosis.

Hepatosplenomegaly is commonly diagnosed by radiologists based on single dimension measurements and heuristic cut-offs. Volumetric measurements may be more accurate for diagnosing organ enlargement. Artificial intelligence techniques may be able to automatically calculate liver and spleen volume and facilitate more accurate diagnosis. After IRB approval, 2 convolutional neural networks (CNN) were developed to automatically segment the liver and spleen on a training dataset comprised of 500 single-phase, contrast-enhanced CT abdomen and pelvis examinations. A separate dataset of ten thousand sequential examinations at a single institution was segmented with these CNNs. Performance was evaluated on a 1% subset and compared with manual segmentations using Sorensen-Dice coefficients and Pearson correlation coefficients. Radiologist reports were reviewed for diagnosis of hepatomegaly and splenomegaly and compared with calculated volumes. Abnormal enlargement was defined as greater than 2 standard deviations above the mean. Median Dice coefficients for liver and spleen segmentation were 0.988 and 0.981, respectively. Pearson correlation coefficients of CNN-derived estimates of organ volume against the gold-standard manual annotation were 0.999 for the liver and spleen (P < 0.001). Average liver volume was 1556.8 ± 498.7 cc and average spleen volume was 194.6 ± 123.0 cc. There were significant differences in average liver and spleen volumes between male and female patients. Thus, the volume thresholds for ground-truth determination of hepatomegaly and splenomegaly were determined separately for each sex. Radiologist classification of hepatomegaly was 65% sensitive, 91% specific, with a positive predictive value (PPV) of 23% and an negative predictive value (NPV) of 98%. Radiologist classification of splenomegaly was 68% sensitive, 97% specific, with a positive predictive value (PPV) of 50% and a negative predictive value (NPV) of 99%. Convolutional neural networks can accurately segment the liver and spleen and may be helpful to improve radiologist accuracy in the diagnosis of hepatomegaly and splenomegaly.

A Case Report on Managing Hemolysis, Elevated Liver Enzymes, and Low Platelets (HELLP Syndrome) During a Rapid Response Code.

HELLP syndrome (hemolysis, elevated liver enzymes, and low platelets) is a relatively rare condition that can complicate pregnancies. We present a 31-year-old gravida 5 para 0 female at 37-3/7 weeks gestation who presented with sudden onset severe epigastric pain, shortness of breath, diaphoresis, bradycardia, and elevated blood pressure. A Rapid Response Team (RRT) was called, and the patient was treated with IV magnesium in addition to blood pressure control medications. Laboratory results confirmed the diagnosis of HELLP syndrome. HELLP syndrome has the potential for major thrombotic complications and should be considered. Early consideration and evaluation of thrombotic complications amongst other common diagnoses are important for prompt treatment and optimization of maternal and fetal well-being.

An Atypical Case of Chilaiditi Syndrome.

Chilaiditi's sign is a rare radiological finding in which a portion of the colon or small intestine is interposed between the liver and right hemidiaphragm. We present a 28-year-old male who came to the emergency room with nausea and vomiting. A computed tomography scan without contrast of the abdomen and pelvis showed a new focus of air in the perihepatic region, suggesting a pneumoperitoneum or a loop of bowel. Exploratory laparotomy was pursued but deferred after a multi-disciplinary review of the imaging. A decision was made to pursue conservative management with a diagnosis of Chilaiditi syndrome. This case illustrates the importance of maintaining a broad differential when approaching a patient with abdominal distress and possible pneumoperitoneum, especially when the clinical picture does not align with radiological findings. Early consideration of Chilaiditi syndrome is important to minimize unnecessary surgical intervention such as laparotomy or further endoscopic intervention, which may lead to potential complications such as perforation, bowel wall ischemia, or respiratory failure.

Salpingovesical fistula mimicking an enterovaginal fistula.

Fallopian tube fistula with the bladder can mimic an enterovaginal fistula. A 34-year-old woman presented with continuous urinary incontinence after hysterectomy. A cystogram confirmed a vesicovaginal fistula and a possible additional intestinal communication. Further imaging, however, ruled out an enterovaginal fistula and diagnosed a fallopian tube prolapse with salpingovesicovaginal fistula. This case demonstrates the importance of multiple imaging modalities in identifying and clearly delineating the anatomy of gynecologic fistulous connections. The case illustrates the fact that while salpingovesical fistula is a rare complication of hysterectomy, it is an important consideration in one's differential diagnosis.