Development of a Deep Learning Algorithm for the Histopathologic Diagnosis and Gleason Grading of Prostate Cancer Biopsies: A Pilot Study.

BACKGROUND: The pathologic diagnosis and Gleason grading of prostate cancer are time-consuming, error-prone, and subject to interobserver variability. Machine learning offers opportunities to improve the diagnosis, risk stratification, and prognostication of prostate cancer. OBJECTIVE: To develop a state-of-the-art deep learning algorithm for the histopathologic diagnosis and Gleason grading of prostate biopsy specimens. DESIGN, SETTING, AND PARTICIPANTS: A total of 85 prostate core biopsy specimens from 25 patients were digitized at 20× magnification and annotated for Gleason 3, 4, and 5 prostate adenocarcinoma by a urologic pathologist. From these virtual slides, we sampled 14803 image patches of 256×256 pixels, approximately balanced for malignancy. OUTCOME MEASUREMENTS AND STATISTICAL ANALYSIS: We trained and tested a deep residual convolutional neural network to classify each patch at two levels: (1) coarse (benign vs malignant) and (2) fine (benign vs Gleason 3 vs 4 vs 5). Model performance was evaluated using fivefold cross-validation. Randomization tests were used for hypothesis testing of model performance versus chance. RESULTS AND LIMITATIONS: The model demonstrated 91.5% accuracy (p<0.001) at coarse-level classification of image patches as benign versus malignant (0.93 sensitivity, 0.90 specificity, and 0.95 average precision). The model demonstrated 85.4% accuracy (p<0.001) at fine-level classification of image patches as benign versus Gleason 3 versus Gleason 4 versus Gleason 5 (0.83 sensitivity, 0.94 specificity, and 0.83 average precision), with the greatest number of confusions in distinguishing between Gleason 3 and 4, and between Gleason 4 and 5. Limitations include the small sample size and the need for external validation. CONCLUSIONS: In this study, a deep learning-based computer vision algorithm demonstrated excellent performance for the histopathologic diagnosis and Gleason grading of prostate cancer. PATIENT SUMMARY: We developed a deep learning algorithm that demonstrated excellent performance for the diagnosis and grading of prostate cancer.

Antimicrobial peptide LL-37 and recombinant human mannose-binding lectin express distinct age- and pathogen-specific antimicrobial activity in human newborn cord blood in vitro.

Background: There is a need to prevent and treat infection in newborns. One approach is administration of antimicrobial proteins and peptides (APPs) such as LL-37, a membrane-active cathelicidin antimicrobial peptide, and mannose-binding lectin (MBL), a pattern-recognition protein that binds to microbial surface polysaccharides resulting in opsonization and complement activation. Low plasma/serum levels of LL-37 and of MBL have been correlated with infection and exogenous administration of these agents may enhance host defense. Methods: The antimicrobial activity of LL-37 (15 µg/ml) or rMBL (0.5, 2 and 10 µg/ml) was tested in hirudin-anticoagulated preterm and term human cord blood (N = 12-14) against Staphylococcus aureus (SA) USA 300 (2x10 4 CFU/ml), Staphylococcus epidermis (SE) 1457 (2x10 4 CFU/ml) and Candida albicans (CA) SC5314 (1x10 4 CFU/ml). After incubation (1, 45, or 180 min), CFUs were enumerated by plating blood onto agar plates. Supernatants were collected for measurement of MBL via ELISA. Results: Preterm cord blood demonstrated impaired endogenous killing capacity against SA and SE compared to term blood. Addition of LL-37 strongly enhanced antimicrobial/antifungal activity vs SA, SE and CA in term blood and SE and CA in preterm blood. By contrast, rMBL showed modest fungistatic activity vs CA in a sub-analysis of term newborns with high basal MBL levels. Baseline MBL levels varied within preterm and term cohorts with no correlation to gestational age. In summary, exogenous LL-37 demonstrated significant antimicrobial activity against SA, SE and CA in term and SE and CA in preterm human blood tested in vitro. rMBL demonstrated modest antifungal activity in term cord blood of individuals with high baseline MBL levels. Conclusions: To the extent that our in vitro results predict the effects of APPs in vivo, development of APPs for prevention and treatment of infection should take into account host age as well as the target pathogen.