Nurturing diversity and inclusion in AI in Biomedicine through a virtual summer program for high school students.

Artificial Intelligence (AI) has the power to improve our lives through a wide variety of applications, many of which fall into the healthcare space; however, a lack of diversity is contributing to limitations in how broadly AI can help people. The UCSF AI4ALL program was established in 2019 to address this issue by targeting high school students from underrepresented backgrounds in AI, giving them a chance to learn about AI with a focus on biomedicine, and promoting diversity and inclusion. In 2020, the UCSF AI4ALL three-week program was held entirely online due to the COVID-19 pandemic. Thus, students participated virtually to gain experience with AI, interact with diverse role models in AI, and learn about advancing health through AI. Specifically, they attended lectures in coding and AI, received an in-depth research experience through hands-on projects exploring COVID-19, and engaged in mentoring and personal development sessions with faculty, researchers, industry professionals, and undergraduate and graduate students, many of whom were women and from underrepresented racial and ethnic backgrounds. At the conclusion of the program, the students presented the results of their research projects at the final symposium. Comparison of pre- and post-program survey responses from students demonstrated that after the program, significantly more students were familiar with how to work with data and to evaluate and apply machine learning algorithms. There were also nominally significant increases in the students' knowing people in AI from historically underrepresented groups, feeling confident in discussing AI, and being aware of careers in AI. We found that we were able to engage young students in AI via our online training program and nurture greater diversity in AI. This work can guide AI training programs aspiring to engage and educate students entirely online, and motivate people in AI to strive towards increasing diversity and inclusion in this field.

A Systematic Review of Time Series Classification Techniques Used in Biomedical Applications.

Background: Digital clinical measures collected via various digital sensing technologies such as smartphones, smartwatches, wearables, and ingestible and implantable sensors are increasingly used by individuals and clinicians to capture the health outcomes or behavioral and physiological characteristics of individuals. Time series classification (TSC) is very commonly used for modeling digital clinical measures. While deep learning models for TSC are very common and powerful, there exist some fundamental challenges. This review presents the non-deep learning models that are commonly used for time series classification in biomedical applications that can achieve high performance. Objective: We performed a systematic review to characterize the techniques that are used in time series classification of digital clinical measures throughout all the stages of data processing and model building. Methods: We conducted a literature search on PubMed, as well as the Institute of Electrical and Electronics Engineers (IEEE), Web of Science, and SCOPUS databases using a range of search terms to retrieve peer-reviewed articles that report on the academic research about digital clinical measures from a five-year period between June 2016 and June 2021. We identified and categorized the research studies based on the types of classification algorithms and sensor input types. Results: We found 452 papers in total from four different databases: PubMed, IEEE, Web of Science Database, and SCOPUS. After removing duplicates and irrelevant papers, 135 articles remained for detailed review and data extraction. Among these, engineered features using time series methods that were subsequently fed into widely used machine learning classifiers were the most commonly used technique, and also most frequently achieved the best performance metrics (77 out of 135 articles). Statistical modeling (24 out of 135 articles) algorithms were the second most common and also the second-best classification technique. Conclusions: In this review paper, summaries of the time series classification models and interpretation methods for biomedical applications are summarized and categorized. While high time series classification performance has been achieved in digital clinical, physiological, or biomedical measures, no standard benchmark datasets, modeling methods, or reporting methodology exist. There is no single widely used method for time series model development or feature interpretation, however many different methods have proven successful.

Immunogenomic profiling and pathological response results from a clinical trial of docetaxel and carboplatin in triple-negative breast cancer.

PURPOSE: Patients with triple-negative breast cancer (TNBC) who do not achieve pathological complete response (pCR) following neoadjuvant chemotherapy have a high risk of recurrence and death. Molecular characterization may identify patients unlikely to achieve pCR. This neoadjuvant trial was conducted to determine the pCR rate with docetaxel and carboplatin and to identify molecular alterations and/or immune gene signatures predicting pCR. EXPERIMENTAL DESIGN: Patients with clinical stages II/III TNBC received 6 cycles of docetaxel and carboplatin. The primary objective was to determine if neoadjuvant docetaxel and carboplatin would increase the pCR rate in TNBC compared to historical expectations. We performed whole-exome sequencing (WES) and immune profiling on pre-treatment tumor samples to identify alterations that may predict pCR. Thirteen matching on-treatment samples were also analyzed to assess changes in molecular profiles. RESULTS: Fifty-eight of 127 (45.7%) patients achieved pCR. There was a non-significant trend toward higher mutation burden for patients with residual cancer burden (RCB) 0/I versus RCB II/III (median 80 versus 68 variants, p 0.88). TP53 was the most frequently mutated gene, observed in 85.7% of tumors. EGFR, RB1, RAD51AP2, SDK2, L1CAM, KPRP, PCDHA1, CACNA1S, CFAP58, COL22A1, and COL4A5 mutations were observed almost exclusively in pre-treatment samples from patients who achieved pCR. Seven mutations in PCDHA1 were observed in pre-treatment samples from patients who did not achieve pCR. Several immune gene signatures including IDO1, PD-L1, interferon gamma signaling, CTLA4, cytotoxicity, tumor inflammation signature, inflammatory chemokines, cytotoxic cells, lymphoid, PD-L2, exhausted CD8, Tregs, and immunoproteasome were upregulated in pre-treatment samples from patients who achieved pCR. CONCLUSION: Neoadjuvant docetaxel and carboplatin resulted in a pCR of 45.7%. WES and immune profiling differentiated patients with and without pCR. TRIAL REGISTRATION: Clinical trial information: NCT02124902, Registered 24 April 2014 & NCT02547987, Registered 10 September 2015.

Associations of CDH1 germline variant location and cancer phenotype in families with hereditary diffuse gastric cancer (HDGC).

INTRODUCTION: Hereditary diffuse gastric cancer (HDGC) is a cancer syndrome associated with variants in E-cadherin (CDH1), diffuse gastric cancer and lobular breast cancer. There is considerable heterogeneity in its clinical manifestations. This study aimed to determine associations between CDH1 germline variant status and clinical phenotypes of HDGC. METHODS: One hundred and fifty-two HDGC families, including six previously unreported families, were identified. CDH1 gene-specific guidelines released by the Clinical Genome Resource (ClinGen) CDH1 Variant Curation Expert Panel were applied for pathogenicity classification of truncating, missense and splice site CDH1 germline variants. We evaluated ORs between location of truncating variants of CDH1 and incidence of colorectal cancer, breast cancer and cancer at young age (gastric cancer at <40 or breast cancer <50 years of age). RESULTS: Frequency of truncating germline CDH1 variants varied across functional domains of the E-cadherin receptor gene and was highest in linker (0.05785 counts/base pair; p=0.0111) and PRE regions (0.10000; p=0.0059). Families with truncating CDH1 germline variants located in the PRE-PRO region were six times more likely to have family members affected by colorectal cancer (OR 6.20, 95% CI 1.79 to 21.48; p=0.004) compared with germline variants in other regions. Variants in the intracellular E-cadherin region were protective for cancer at young age (OR 0.2, 95% CI 0.06 to 0.64; p=0.0071) and in the linker regions for breast cancer (OR 0.35, 95% CI 0.12 to 0.99; p=0.0493). Different CDH1 genotypes were associated with different intracellular signalling activation levels including different p-ERK, p-mTOR and ß-catenin levels in early submucosal T1a lesions of HDGC families with different CDH1 variants. CONCLUSION: Type and location of CDH1 germline variants may help to identify families at increased risk for concomitant cancers that might benefit from individualised surveillance and intervention strategies.

Transcriptomic profiling and quantitative high-throughput (qHTS) drug screening of CDH1 deficient hereditary diffuse gastric cancer (HDGC) cells identify treatment leads for familial gastric cancer.

BACKGROUND: Patients with hereditary diffuse gastric cancer (HDGC), a cancer predisposition syndrome associated with germline mutations of the CDH1 (E-cadherin) gene, have few effective treatment options. Despite marked differences in natural history, histopathology, and genetic profile to patients afflicted by sporadic gastric cancer, patients with HDGC receive, in large, identical systemic regimens. The lack of a robust preclinical in vitro system suitable for effective drug screening has been one of the obstacles to date which has hampered therapeutic advances in this rare disease. METHODS: In order to identify therapeutic leads selective for the HDGC subtype of gastric cancer, we compared gene expression profiles and drug phenotype derived from an oncology library of 1912 compounds between gastric cancer cells established from a patient with metastatic HDGC harboring a c.1380delA CDH1 germline variant and sporadic gastric cancer cells. RESULTS: Unsupervised hierarchical cluster analysis shows select gene expression alterations in c.1380delA CDH1 SB.mhdgc-1 cells compared to a panel of sporadic gastric cancer cell lines with enrichment of ERK1-ERK2 (extracellular signal regulated kinase) and IP3 (inositol trisphosphate)/DAG (diacylglycerol) signaling as the top networks in c.1380delA SB.mhdgc-1 cells. Intracellular phosphatidylinositol intermediaries were increased upon direct measure in c.1380delA CDH1 SB.mhdgc-1 cells. Differential high-throughput drug screening of c.1380delA CDH1 SB.mhdgc-1 versus sporadic gastric cancer cells identified several compound classes with enriched activity in c.1380 CDH1 SB.mhdgc-1 cells including mTOR (Mammalian Target Of Rapamycin), MEK (Mitogen-Activated Protein Kinase), c-Src kinase, FAK (Focal Adhesion Kinase), PKC (Protein Kinase C), or TOPO2 (Topoisomerase II) inhibitors. Upon additional drug response testing, dual PI3K (Phosphatidylinositol 3-Kinase)/mTOR and topoisomerase 2A inhibitors displayed up to >100-fold increased activity in hereditary c.1380delA CDH1 gastric cancer cells inducing apoptosis most effectively in cells with deficient CDH1 function. CONCLUSION: Integrated pharmacological and transcriptomic profiling of hereditary diffuse gastric cancer cells with a loss-of-function c.1380delA CDH1 mutation implies various pharmacological vulnerabilities selective to CDH1-deficient familial gastric cancer cells and suggests novel treatment leads for future preclinical and clinical treatment studies of familial gastric cancer.

Seizures detection using multimodal signals: a scoping review.

Introduction. Epileptic seizures are common neurological disorders in the world, impacting 65 million people globally. Around 30% of patients with seizures suffer from refractory epilepsy, where seizures are not controlled by medications. The unpredictability of seizures makes it essential to have a continuous seizure monitoring system outside clinical settings for the purpose of minimizing patients' injuries and providing additional pathways for evaluation and treatment follow-up. Autonomic changes related to seizure events have been extensively studied and attempts made to apply them for seizure detection and prediction tasks. This scoping review aims to depict current research activities associated with the implementation of portable, wearable devices for seizure detection or prediction and inform future direction in continuous seizure tracking in ambulatory settings.Methods. Overall methodology framework includes 5 essential stages: research questions identification, relevant studies identification, selection of studies, data charting and summarizing the findings. A systematic searching strategy guided by systematic reviews and meta-analysis (PRISMA) was implemented to identify relevant records on two databases (PubMed, IEEE).Results. A total of 30 articles were included in our final analysis. Most of the studies were conducted off-line and employed consumer-graded wearable device. ACM is the dominant modality to be used in seizure detection, and widely deployed algorithms entail Support Vector Machine, Random Forest and threshold-based approach. The sensitivity ranged from 33.2% to 100% for single modality with a false alarm rate (FAR) ranging from 0.096 to 14.8 d-1. Multimodality has a sensitivity ranging from 51% to 100% with FAR ranging from 0.12 to 17.7 d-1.Conclusion. The overall performance in seizure detection system based on non-cerebral physiological signals is promising, especially for the detection of motor seizures and seizures accompanied with intense ictal autonomic changes.

A randomized phase 2 study of neoadjuvant carboplatin and paclitaxel with or without atezolizumab in triple negative breast cancer (TNBC) - NCI 10013.

Atezolizumab with chemotherapy has shown improved progression-free and overall survival in patients with metastatic PD-L1 positive triple negative breast cancer (TNBC). Atezolizumab with anthracycline- and taxane-based neoadjuvant chemotherapy has also shown increased pathological complete response (pCR) rates in early TNBC. This trial evaluated neoadjuvant carboplatin and paclitaxel with or without atezolizumab in patients with clinical stages II-III TNBC. The co-primary objectives were to evaluate if chemotherapy and atezolizumab increase pCR rate and tumor infiltrating lymphocyte (TIL) percentage compared to chemotherapy alone in the mITT population. Sixty-seven patients (ages 25-78 years; median, 52 years) were randomly assigned - 22 patients to Arm A, and 45 to Arm B. Median follow up was 6.6 months. In the modified intent to treat population (all patients evaluable for the primary endpoints who received at least one dose of combination therapy), the pCR rate was 18.8% (95% CI 4.0-45.6%) in Arm A, and 55.6% (95% CI 40.0-70.4%) in Arm B (estimated treatment difference: 36.8%, 95% CI 8.5-56.6%; p = 0.018). Grade 3 or higher treatment-related adverse events occurred in 62.5% of patients in Arm A, and 57.8% of patients in Arm B. One patient in Arm B died from recurrent disease during the follow-up period. TIL percentage increased slightly from baseline to cycle 1 in both Arm A (mean ± SD: 0.6% ± 21.0%) and Arm B (5.7% ± 15.8%) (p = 0.36). Patients with pCR had higher median TIL percentages (24.8%) than those with non-pCR (14.2%) (p = 0.02). Although subgroup analyses were limited by the small sample size, PD-L1-positive patients treated with chemotherapy and atezolizumab had a pCR rate of 75% (12/16). The addition of atezolizumab to neoadjuvant carboplatin and paclitaxel resulted in a statistically significant and clinically relevant increased pCR rate in patients with clinical stages II and III TNBC. (Funded by National Cancer Institute).

Challenges targeting cancer neoantigens in 2021: a systematic literature review.

Introduction: Cancer neoantigens represent important targets of cancer immunotherapy. The goal of cancer neoantigen vaccines is to induce neoantigen-specific immune responses and antitumor immunity while minimizing the potential for autoimmune toxicity. Advances in sequencing technologies, neoantigen prediction algorithms, and other technologies have dramatically improved the ability to identify and prioritize cancer neoantigens. Unfortunately, results from preclinical studies and early phase clinical trials highlight important challenges to the successful clinical translation of neoantigen cancer vaccines.Areas covered: In this review, we provide an overview of current strategies for the identification and prioritization of cancer neoantigens with a particular emphasis on the two most common strategies used for neoantigen identification: (1) direct identification of peptide ligands eluted from peptide-MHC complexes, and (2) next-generation sequencing combined with neoantigen prediction algorithms. We highlight the limitations of current neoantigen prediction pipelines, and discuss broader challenges associated with cancer neoantigen vaccines including tumor purity/heterogeneity and the immunosuppressive tumor microenvironment.Expert opinion: Despite current limitations, neoantigen prediction is likely to improve rapidly based on advances in sequencing, machine learning, and information sharing. The successful development of robust cancer neoantigen prediction strategies is likely to have a significant impact, with the potential to facilitate cancer neoantigen vaccine design.

Transit-Amplifying Cells Coordinate Changes in Intestinal Epithelial Cell-Type Composition.

Renewing tissues have the remarkable ability to continually produce both proliferative progenitor and specialized differentiated cell types. How are complex milieus of microenvironmental signals interpreted to coordinate tissue-cell-type composition? Here, we investigate the responses of intestinal epithelium to individual and paired perturbations across eight epithelial signaling pathways. Using a high-throughput approach that combines enteroid monolayers and quantitative imaging, we identified conditions that enrich for specific cell types as well as interactions between pathways. Importantly, we found that modulation of transit-amplifying cell proliferation changes the ratio of differentiated secretory to absorptive cell types. These observations highlight an underappreciated role for transit-amplifying cells in the tuning of differentiated cell-type composition.

Generation and Quantitative Imaging of Enteroid Monolayers.

The intestinal epithelium is a single layer of cells that plays a critical role in digestion, absorbs nutrients from food, and coordinates the delicate interplay between microbes in the gut lumen and the immune system. Epithelial homeostasis is crucial for maintaining health; disruption of homeostasis results in disorders including inflammatory bowel disease and cancer. The advent of 3D intestinal epithelial organoids has greatly advanced our understanding of the molecular underpinnings of epithelial homeostasis and disease. Recently, we developed an enteroid monolayer (2D) culture system that recapitulates important features of 3D organoids and the in vivo intestinal epithelium such as tissue renewal, representation of diverse epithelial cell types, self-organization, and apical-basolateral polarization. Enteroid monolayers are cultured in microtiter plates, enabling high-throughput experiments. Furthermore, their 2D nature makes it easier to distinguish individual cells by fluorescent microscopy, enabling quantitative analysis of single cell behaviors within the epithelial tissue.Here we describe experimental methods for generating enteroid monolayers and computational methods for analyzing immunofluorescence images of enteroid monolayers. We outline experimental methods for generating enteroid monolayers from freshly isolated intestinal crypts, frozen intestinal crypts, and 3D organoids. Fresh crypts are easily obtained from murine or human intestinal samples, and the ability to derive enteroid monolayers from both frozen crypts and 3D organoids enables genetic modification and/or biobanking of patient samples for future studies. We outline computational methods for identifying distinct epithelial cell types (goblet, stem, EdU+) in immunofluorescence images of enteroid monolayers and, importantly, individual nuclei, enabling truly single cell measurements of epithelial cell behaviors to be made. Taken together, these methods will enable detailed studies of epithelial homeostasis and intestinal disease.

Enteroid Monolayers Reveal an Autonomous WNT and BMP Circuit Controlling Intestinal Epithelial Growth and Organization.

The intestinal epithelium maintains a remarkable balance between proliferation and differentiation despite rapid cellular turnover. A central challenge is to elucidate mechanisms required for robust control of tissue renewal. Opposing WNT and BMP signaling is essential in establishing epithelial homeostasis. However, it has been difficult to disentangle contributions from multiple sources of morphogen signals in the tissue. Here, to dissect epithelial-autonomous morphogenic signaling circuits, we developed an enteroid monolayer culture system that recapitulates four key properties of the intestinal epithelium, namely the ability to maintain proliferative and differentiated zones, self-renew, polarize, and generate major intestinal cell types. We systematically perturb intrinsic and extrinsic sources of WNT and BMP signals to reveal a core morphogenic circuit that controls proliferation, tissue organization, and cell fate. Our work demonstrates the ability of intestinal epithelium, even in the absence of 3D tissue architecture, to control its own growth and organization through morphogen-mediated feedback.

The phospholipase A2 enzyme complex PAFAH Ib mediates endosomal membrane tubule formation and trafficking.

Previous studies have shown that membrane tubule-mediated export from endosomal compartments requires a cytoplasmic phospholipase A(2) (PLA(2)) activity. Here we report that the cytoplasmic PLA(2) enzyme complex platelet-activating factor acetylhydrolase (PAFAH) Ib, which consists of α1, α2, and LIS1 subunits, regulates the distribution and function of endosomes. The catalytic subunits α1 and α2 are located on early-sorting endosomes and the central endocytic recycling compartment (ERC) and their overexpression, but not overexpression of their catalytically inactive counterparts, induced endosome membrane tubules. In addition, overexpression α1 and α2 altered normal endocytic trafficking; transferrin was recycled back to the plasma membrane directly from peripheral early-sorting endosomes instead of making an intermediate stop in the ERC. Consistent with these results, small interfering RNA-mediated knockdown of α1 and α2 significantly inhibited the formation of endosome membrane tubules and delayed the recycling of transferrin. In addition, the results agree with previous reports that PAFAH Ib α1 and α2 expression levels affect the distribution of endosomes within the cell through interactions with the dynein regulator LIS1. These studies show that PAFAH Ib regulates endocytic membrane trafficking through novel mechanisms involving both PLA(2) activity and LIS1-dependent dynein function.