Automated Identification of Breast Cancer Relapse in Computed Tomography Reports Using Natural Language Processing.

Purpose: Breast cancer relapses are rarely collected by cancer registries because of logistical and financial constraints. Hence, we investigated natural language processing (NLP), enhanced with state-of-the-art deep learning transformer tools and large language models, to automate relapse identification in the text of computed tomography (CT) reports.

Methods: We analyzed follow-up CT reports from patients diagnosed with breast cancer between January 1, 2005, and December 31, 2014.

Learning traps and change blindness in dynamic environments.

Learning traps arise when early experience leads to a false belief about the reward structure of the environment which, in turn, leads to avoidance of rewarding options. Previous work on the negative effects of such traps has focused on static learning environments. The current work examines an additional negative effect of learning traps in dynamic environments-blindness to change in the features that predict decision outcomes.

The myelin water imaging transcriptome: myelin water fraction regionally varies with oligodendrocyte-specific gene expression.

Identifying sensitive and specific measures that can quantify myelin are instrumental in characterizing microstructural changes in neurological conditions. Neuroimaging transcriptomics is emerging as a valuable technique in this regard, offering insights into the molecular basis of promising candidates for myelin quantification, such as myelin water fraction (MWF). We aimed to demonstrate the utility of neuroimaging transcriptomics by validating MWF as a myelin measure.

Changing your mind about the data: Updating sampling assumptions in inductive inference.

When people use samples of evidence to make inferences, they consider both the sample contents and how the sample was generated ("sampling assumptions"). The current studies examined whether people can update their sampling assumptions - whether they can revise a belief about sample generation that is discovered to be incorrect, and reinterpret old data in light of the new belief. We used a property induction task where learners saw a sample of instances that shared a novel property and then inferred whether it generalized to other items.

Management of ovarian and breast cancer risk in non-BRCA HBOC pathogenic variant carriers in a large California health care system.

Purpose: To describe breast and ovarian cancer risk reduction strategies in the clinical management of women who test positive for non-BRCA hereditary breast and ovarian cancer (HBOC) pathogenic variants compared to those who test positive for pathogenic BRCA variants or have negative germline panel testing.

Methods: Examination of imaging and preventive surgeries in women undergoing HBOC genetic testing from 1/1/2015 to 12/31/2018, with follow up to 03/31/2020 in Kaiser Permanente Northern California.

Results: A total of 13,271 tests which included HBOC genes were identified.

Streamlining genetic testing for women with ovarian cancer in a Northern California health care system.

Objective: Referral to Genetics for pre-testing counseling may be inefficient for women with ovarian cancer. This study assesses feasibility of gynecologic oncologists directly offering genetic testing.

Methods: A prospective pilot study was conducted at two gynecologic oncology hubs in an integrated healthcare system from May 1 to November 6, 2019.

Prenatal detection of congenital heart disease.

Objectives: To define current frequency of prenatal detection of congenital heart disease (CHD), factors affecting prenatal detection, and its influence on postnatal course.

Study Design: We prospectively identified all fetuses and infants < or =6 months of age with major CHD at 3 referral centers in Northern California over 1 year; we obtained prenatal and demographic data, reviewed prenatal ultrasound (US) and postnatal records, and used logistic regression to analyze maternal, fetal, and prenatal-care provider risk factors for prenatal diagnosis.

Results: Ninety-eight of 309 infants with major CHD had prenatal diagnosis (36% accounting for 27 pregnancy terminations); 185 infant-families participated in the postnatal survey, and although 99% had prenatal US, only 28% were prenatally diagnosed.