Evaluation of GPT-4 ability to identify and generate patient instructions for actionable incidental radiology findings.

OBJECTIVES: To evaluate the proficiency of a HIPAA-compliant version of GPT-4 in identifying actionable, incidental findings from unstructured radiology reports of Emergency Department patients. To assess appropriateness of artificial intelligence (AI)-generated, patient-facing summaries of these findings. MATERIALS AND METHODS: Radiology reports extracted from the electronic health record of a large academic medical center were manually reviewed to identify non-emergent, incidental findings with high likelihood of requiring follow-up, further sub-stratified as "definitely actionable" (DA) or "possibly actionable-clinical correlation" (PA-CC). Instruction prompts to GPT-4 were developed and iteratively optimized using a validation set of 50 reports. The optimized prompt was then applied to a test set of 430 unseen reports. GPT-4 performance was primarily graded on accuracy identifying either DA or PA-CC findings, then secondarily for DA findings alone. Outputs were reviewed for hallucinations. AI-generated patient-facing summaries were assessed for appropriateness via Likert scale. RESULTS: For the primary outcome (DA or PA-CC), GPT-4 achieved 99.3% recall, 73.6% precision, and 84.5% F-1. For the secondary outcome (DA only), GPT-4 demonstrated 95.2% recall, 77.3% precision, and 85.3% F-1. No findings were "hallucinated" outright. However, 2.8% of cases included generated text about recommendations that were inferred without specific reference. The majority of True Positive AI-generated summaries required no or minor revision. CONCLUSION: GPT-4 demonstrates proficiency in detecting actionable, incidental findings after refined instruction prompting. AI-generated patient instructions were most often appropriate, but rarely included inferred recommendations. While this technology shows promise to augment diagnostics, active clinician oversight via "human-in-the-loop" workflows remains critical for clinical implementation.

Subtle vaginal evisceration resulting in small bowel evisceration: a case report.

BACKGROUND: Evisceration of bowel contents through the vagina is a rare event that may be complicated by bowel obstruction. OBJECTIVE: We report a case of vaginal evisceration with small bowel obstruction which, in contrast to previous, more dramatic case reports in the literature, is a more subtle and, in fact, characteristic clinical presentation for this unusual occurrence. CASE REPORT: A 72-year-old woman with a previous history of pelvic surgery presented to the Emergency Department with lower abdominal discomfort and a prolapsing mass from her vagina. She was initially discharged home after bedside reduction of the mass, but returned 48 h later with worsening symptoms. A computed tomography scan on her repeat visit confirmed evisceration of bowel into the vaginal vault with obstruction of distal bowel loops. Surgical and gynecologic services were consulted and the patient underwent partial small bowel resection and vaginal cuff repair in the operating room. CONCLUSION: Early recognition of subtle presentations of vaginal evisceration is crucial for preserving bowel viability and preventing morbidity from bowel ischemia or infarction. Risk factors for this rare condition include postmenopausal status, previous pelvic surgery, and presence of an enterocele.

Diagnosis And Management Of Deep Venous Thrombosis In The Emergency Department.

Although the clinical presentations of deep venous thrombosis are notoriously subtle and nonspecific, risk stratification tools such as the Wells clinical model have improved the efficiency of the diagnostic evaluation. The emergency clinician may be guided down several pathways, including D-dimer assays and/ or ultrasonography. New oral anticoagulants offer alternatives to the traditional heparins and vitamin K antagonists in the treatment of deep venous thrombosis. This review examines the current literature, evidence, and guidelines in the diagnosis and management of deep venous thrombosis. It also explores some of the controversies and developments regarding risk stratification, adjusted D-dimer thresholds,special populations, isolated distal deep venous thrombosis, upper extremity deep venous thrombosis, outpatient treatment, and the new oral anticoagulants.

High-risk chief complaints I: chest pain--the big three.

Chest pain is one of the most frequently seen chief complaints in patients presenting to emergency departments, and is considered to be a "high-risk" chief complaint. The differential diagnosis for chest pain is broad, and potential causes range from the benign to the immediately life-threatening. Although many (if not most) emergency department patients with chest pain do not have an immediately life-threatening condition, correct diagnoses can be difficult to make, incorrect diagnoses may lead to catastrophic therapies, and failure to make a timely diagnosis may contribute to significant morbidity and mortality. Several atraumatic "high-risk" causes of chest pain are discussed in this article, including myocardial infarction and ischemia, thoracic aortic dissection, and pulmonary embolism. Also included are brief discussions of tension pneumothorax, esophageal perforation, and cardiac tamponade.