Frequency and Outcomes of BI-RADS Category 3 Assessments in Patients With a Personal History of Breast Cancer: Full-Field Digital Mammography Versus Digital Breast Tomosynthesis.

BACKGROUND. Studies establishing the validity of BI-RADS category 3 excluded patients with personal history of breast cancer (PHBC). Use of category 3 in patients with PHBC may be impacted not only by this population's increased breast cancer risk, but also by adoption of digital breast tomosynthesis (DBT) over full-field digital mammography (FFDM). OBJECTIVE. The purpose of this article was to compare the frequency, outcomes, and additional characteristics of BI-RADS category 3 assessments between FFDM and DBT in patients with PHBC. METHODS. This retrospective study included 14,845 mammograms in 10,118 patients (mean age, 63 years) with PHBC who had undergone mastectomy and/or lumpectomy. Of these, 8422 examinations were performed by FFDM from October 2014 to September 2016, and 6423 examinations by FFDM with DBT from February 2017 to December 2018, after interval conversion of the center's mammography units. Information was extracted from the EHR and radiology reports. FFDM and DBT groups were compared in the entire sample and among index category 3 lesions (i.e., earliest category 3 assessment per lesion). RESULTS. The frequency of category 3 assessment was lower for DBT than FFDM (5.6% vs 6.4%; p = .05). DBT, compared with FFDM, showed a lower malignancy rate for category 3 lesions (1.8% vs 5.0%; p = .04), higher malignancy rate for category 4 lesions (32.0% vs 23.2%; p = .03), and no difference in malignancy rate for category 5 lesions (100.0% vs 75.0%; p = .24). Analysis of index category 3 lesions included 438 and 274 lesions for FFDM and DBT, respectively. For category 3 lesions, DBT, compared with FFDM, showed lower PPV3 (13.9% vs 36.1%; p = .02) and a more frequent mammographic finding of mass (33.2% vs 23.1%; p = .003). CONCLUSION. The malignancy rate for category 3 lesions in patients with PHBC was less than the accepted limit (2%) for DBT (1.8%), but not FFDM (5.0%). A lower malignancy rate for category 3 lesions but higher malignancy rate for category 4 lesions for DBT supports more appropriate application of category 3 assessment in patients with PHBC through use of DBT. CLINICAL IMPACT. These insights may help establish whether category 3 assessments in patients with PHBC are within benchmarks for early detection of second cancers and reduction of benign biopsies.

Use of machine-learning algorithms to determine features of systolic blood pressure variability that predict poor outcomes in hypertensive patients.

BACKGROUND: We re-analyzed data from the Systolic Blood Pressure Intervention Trial (SPRINT) trial to identify features of systolic blood pressure (SBP) variability that portend poor cardiovascular outcomes using a nonlinear machine-learning algorithm. METHODS: We included all patients who completed 1 year of the study without reaching any primary endpoint during the first year, specifically: myocardial infarction, other acute coronary syndromes, stroke, heart failure or death from a cardiovascular event (n = 8799; 94%). In addition to clinical variables, features representing longitudinal SBP trends and variability were determined and combined in a random forest algorithm, optimized using cross-validation, using 70% of patients in the training set. Area under the curve (AUC) was measured using a 30% testing set. Finally, feature importance was determined by minimizing node impurity averaging over all trees in the forest for a specific feature. RESULTS: A total of 365 patients (4.1%) reached the combined primary outcome over 37 months of follow-up. The random forest classifier had an AUC of 0.71 on the testing set. The 10 most significant features selected in order of importance by the automated algorithm included the urine albumin/creatinine (CR) ratio, estimated glomerular filtration rate, age, serum CR, history of subclinical cardiovascular disease (CVD), cholesterol, a variable representing SBP signals using wavelet transformation, high-density lipoprotein, the 90th percentile of SBP and triglyceride level. CONCLUSIONS: We successfully demonstrated use of random forest algorithm to define best prognostic longitudinal SBP representations. In addition to known risk factors for CVD, transformed variables for time series SBP measurements were found to be important in predicting poor cardiovascular outcomes and require further evaluation.

Automatic analysis of medical dialogue in the home hemodialysis domain: structure induction and summarization.

Spoken medical dialogue is a valuable source of information for patients and caregivers. This work presents a first step towards automatic analysis and summarization of spoken medical dialogue. We first abstract a dialogue into a sequence of semantic categories using linguistic and contextual features integrated in a supervised machine-learning framework. Our model has a classification accuracy of 73%, compared to 33% achieved by a majority baseline (p<0.01). We then describe and implement a summarizer that utilizes this automatically induced structure. Our evaluation results indicate that automatically generated summaries exhibit high resemblance to summaries written by humans. In addition, task-based evaluation shows that physicians can reasonably answer questions related to patient care by looking at the automatically generated summaries alone, in contrast to the physicians' performance when they were given summaries from a naïve summarizer (p<0.05). This work demonstrates the feasibility of automatically structuring and summarizing spoken medical dialogue.