Artificial intelligence in neuropathology: deep learning-based assessment of tauopathy.

Accumulation of abnormal tau in neurofibrillary tangles (NFT) occurs in Alzheimer disease (AD) and a spectrum of tauopathies. These tauopathies have diverse and overlapping morphological phenotypes that obscure classification and quantitative assessments. Recently, powerful machine learning-based approaches have emerged, allowing the recognition and quantification of pathological changes from digital images. Here, we applied deep learning to the neuropathological assessment of NFT in postmortem human brain tissue to develop a classifier capable of recognizing and quantifying tau burden. The histopathological material was derived from 22 autopsy brains from patients with tauopathies. We used a custom web-based informatics platform integrated with an in-house information management system to manage whole slide images (WSI) and human expert annotations as ground truth. We utilized fully annotated regions to train a deep learning fully convolutional neural network (FCN) implemented in PyTorch against the human expert annotations. We found that the deep learning framework is capable of identifying and quantifying NFT with a range of staining intensities and diverse morphologies. With our FCN model, we achieved high precision and recall in naive WSI semantic segmentation, correctly identifying tangle objects using a SegNet model trained for 200 epochs. Our FCN is efficient and well suited for the practical application of WSIs with average processing times of 45 min per WSI per GPU, enabling reliable and reproducible large-scale detection of tangles. We measured performance on test data of 50 pre-annotated regions on eight naive WSI across various tauopathies, resulting in the recall, precision, and an F1 score of 0.92, 0.72, and 0.81, respectively. Machine learning is a useful tool for complex pathological assessment of AD and other tauopathies. Using deep learning classifiers, we have the potential to integrate cell- and region-specific annotations with clinical, genetic, and molecular data, providing unbiased data for clinicopathological correlations that will enhance our knowledge of the neurodegeneration.

Artificial intelligence-derived neurofibrillary tangle burden is associated with antemortem cognitive impairment.

Tauopathies are a category of neurodegenerative diseases characterized by the presence of abnormal tau protein-containing neurofibrillary tangles (NFTs). NFTs are universally observed in aging, occurring with or without the concomitant accumulation of amyloid-beta peptide (Aß) in plaques that typifies Alzheimer disease (AD), the most common tauopathy. Primary age-related tauopathy (PART) is an Aß-independent process that affects the medial temporal lobe in both cognitively normal and impaired subjects. Determinants of symptomology in subjects with PART are poorly understood and require clinicopathologic correlation; however, classical approaches to staging tau pathology have limited quantitative reproducibility. As such, there is a critical need for unbiased methods to quantitatively analyze tau pathology on the histological level. Artificial intelligence (AI)-based convolutional neural networks (CNNs) generate highly accurate and precise computer vision assessments of digitized pathology slides, yielding novel histology metrics at scale. Here, we performed a retrospective autopsy study of a large cohort (n = 706) of human post-mortem brain tissues from normal and cognitively impaired elderly individuals with mild or no Aß plaques (average age of death of 83.1 yr, range 55-110). We utilized a CNN trained to segment NFTs on hippocampus sections immunohistochemically stained with antisera recognizing abnormal hyperphosphorylated tau (p-tau), which yielded metrics of regional NFT counts, NFT positive pixel density, as well as a novel graph-theory based metric measuring the spatial distribution of NFTs. We found that several AI-derived NFT metrics significantly predicted the presence of cognitive impairment in both the hippocampus proper and entorhinal cortex (p < 0.0001). When controlling for age, AI-derived NFT counts still significantly predicted the presence of cognitive impairment (p = 0.04 in the entorhinal cortex; p = 0.04 overall). In contrast, Braak stage did not predict cognitive impairment in either age-adjusted or unadjusted models. These findings support the hypothesis that NFT burden correlates with cognitive impairment in PART. Furthermore, our analysis strongly suggests that AI-derived metrics of tau pathology provide a powerful tool that can deepen our understanding of the role of neurofibrillary degeneration in cognitive impairment.

Antemortem detection of Parkinson's disease pathology in peripheral biopsies using artificial intelligence.

The diagnosis of Parkinson's disease (PD) is challenging at all stages due to variable symptomatology, comorbidities, and mimicking conditions. Postmortem assessment remains the gold standard for a definitive diagnosis. While it is well recognized that PD manifests pathologically in the central nervous system with aggregation of α-synuclein as Lewy bodies and neurites, similar Lewy-type synucleinopathy (LTS) is additionally found in the peripheral nervous system that may be useful as an antemortem biomarker. We have previously found that detection of LTS in submandibular gland (SMG) biopsies is sensitive and specific for advanced PD; however, the sensitivity is suboptimal especially for early-stage disease. Further, visual microscopic assessment of biopsies by a neuropathologist to identify LTS is impractical for large-scale adoption. Here, we trained and validated a convolutional neural network (CNN) for detection of LTS on 283 digital whole slide images (WSI) from 95 unique SMG biopsies. A total of 8,450 LTS and 35,066 background objects were annotated following an inter-rater reliability study with Fleiss Kappa = 0.72. We used transfer learning to train a CNN model to classify image patches (151 × 151 pixels at 20× magnification) with and without the presence of LTS objects. The trained CNN model showed the following performance on image patches: sensitivity: 0.99, specificity: 0.99, precision: 0.81, accuracy: 0.99, and F-1 score: 0.89. We further tested the trained network on 1230 naïve WSI from the same cohort of research subjects comprising 42 PD patients and 14 controls. Logistic regression models trained on features engineered from the CNN predictions on the WSI resulted in sensitivity: 0.71, specificity: 0.65, precision: 0.86, accuracy: 0.69, and F-1 score: 0.76 in predicting clinical PD status, and 0.64 accuracy in predicting PD stage, outperforming expert neuropathologist LTS density scoring in terms of sensitivity but not specificity. These findings demonstrate the practical utility of a CNN detector in screening for LTS, which can translate into a computational tool to facilitate the antemortem tissue-based diagnosis of PD in clinical settings.

Left ventricular systolic and diastolic function in healthy adult bonnet macaques (Macaca radiata). New echocardiographic indices in Old World monkeys.

OBJECTIVES: Macaques are used in cardiovascular and metabolic research. We determined echocardiographic-derived reference values of left ventricular (LV) systolic and diastolic function in healthy adult bonnet macaques (Macaca radiata). METHODS: Transthoracic echocardiography was performed during ketamine sedation in 83 (67% female) healthy monkeys (age 7-26 years). RESULTS: Technically adequate studies were obtained in all subjects and required 10.1 +/- 1.3 min of scanning time. Age correlated inversely with the following Doppler indices: E (r = -0.44, p < 0.001), E/A (r = -0.26, p = 0.02), E' (r = -0.45, p < 0.001, E'/A' (r = -0.44, p < 0.001), E/E' (r -0.25, p = 0.03), S' (r = -0.33, p = 0.003), Vp (r = -0.26, p = 0.049). LV mass was more strongly correlated with crown rump length (r = 0.72, p < 0.001) and body surface area (r = 0.70, p < 0.001) than with body mass index (r = 0.47, p < 0.001) and weight (r = 0.63, p < 0.001). CONCLUSIONS: This study demonstrates echocardiography is feasible for characterizing LV function. Age-related changes in Doppler indices in primates are similar to those in humans. LV mass is more closely related to fat-free mass indices. We provide reference values for LV systolic and diastolic function in adult bonnet macaques across the captive life span.

Alternatively spliced human tissue factor promotes tumor growth and angiogenesis in a pancreatic cancer tumor model.

INTRODUCTION: Tissue Factor (TF) expression is observed in many types of cancer, associated with more aggressive disease, and thrombosis. Alternatively-spliced human tissue factor (asHTF) has recently been identified in which exon 5 is deleted. asHTF is soluble due to the substitution of the transmembrane and cytoplasmic domains of exon 6 with a unique COOH-terminal domain. MATERIALS AND METHODS: We examine the expression and function of asHTF and full-length Tissue Factor ((FL)TF) in six human pancreatic cancer cells. Further, we transfected asHTF, (FL)TF, and control expression vectors into a non-expressing, human pancreatic cancer line (MiaPaCa-2). We studied the procoagulant activity of asHTF and (FL)TF and the effect on tumor growth in mice. RESULTS: asHTF is expressed in 5 of 6 human pancreatic cancer cell lines, but not in normal human fibroblasts, nor the MiaPaCa-2 line. (FL)TF conferred procoagulant activity, but asHTF did not. Transfected cells were injected subcutaneously in athymic mice. Interestingly, compared with control transfection, (FL)TF expression was associated with reduced tumor growth (mean 7 mg vs 85 mg), while asHTF-expression was associated with enhanced tumor growth (mean 389 mg vs. 85 mg). asHTF expression resulted in increased mitotic index and microvascular density. CONCLUSIONS: These data suggests that asHTF expression promotes tumor growth, and is associated with increased tumor cell proliferation and angiogenesis in vivo. Our results raise a new perspective on the understanding of the relationship between TF expression and cancer growth, by showing a dissociation of the procoagulant activity of (FL)TF and the cancer-promoting activity of asHTF.

Heat shock protein 27 increases after androgen ablation and plays a cytoprotective role in hormone-refractory prostate cancer.

Heat shock protein 27 (Hsp27) is a chaperone implicated as an independent predictor of clinical outcome in prostate cancer. Our aim was to characterize changes in Hsp27 after androgen withdrawal and during androgen-independent progression in prostate xenografts and human prostate cancer to assess the functional significance of these changes using antisense inhibition of Hsp27. A tissue microarray was used to measure changes in Hsp27 protein expression in 232 specimens from hormone naive and posthormone-treated cancers. Hsp27 expression was low or absent in untreated human prostate cancers but increased beginning 4 weeks after androgen-ablation to become uniformly highly expressed in androgen-independent tumors. Androgen-independent human prostate cancer PC-3 cells express higher levels of Hsp27 mRNA in vitro and in vivo, compared with androgen-sensitive LNCaP cells. Phosphorothioate Hsp27 antisense oligonucleotides (ASOs) and small interference RNA potently inhibit Hsp27 expression, with increased caspase-3 cleavage and PC3 cell apoptosis and 87% decreased PC3 cell growth. Hsp27 ASO and small interference RNA also enhanced paclitaxel chemosensitivity in vitro, whereas in vivo, systemic administration of Hsp27 ASO in athymic mice decreased PC-3 tumor progression and also significantly enhanced paclitaxel chemosensitivity. These findings suggest that increased levels of Hsp27 after androgen withdrawal provide a cytoprotective role during development of androgen independence and that ASO-induced silencing can enhance apoptosis and delay tumor progression.

Role of alternatively spliced tissue factor in pancreatic cancer growth and angiogenesis.

Increased tissue factor (TF) expression is observed in many types of cancer, associated with more aggressive disease, and in thrombosis. The mechanism by which TF promotes tumor growth remains unclear. Anticoagulation has been shown to result in a trend toward improved survival; no direct antitumor effect has been shown in cancer patients. Alternatively spliced tissue factor (asTF) was recently described, in which exon 5 is deleted. Because of a frame-shift in exon 6, the transmembrane and cytoplasmic domains are replaced with a unique COOH-terminal domain, making asTF soluble. Both alternatively spliced human tissue factor (asHTF) and full-length tissue factor (flTF) are expressed in human pancreatic cancer lines and in pancreatic cancer specimens. We studied the role of asHTF and flTF in a mouse model of pancreatic cancer. Although lacking procoagulant activity, asTF promotes primary growth of human pancreatic cancer cells in mice and augments tumor-associated angiogenesis. This body of work suggests a new paradigm for the role of TF in pancreatic cancer: that asHTF contributes to cancer growth, independent of procoagulant activity.

Energy expenditure in chow-fed female non-human primates of various weights.

BACKGROUND: Until now no technology has been available to study energy metabolism in monkeys. The objective of this study was to determine daily energy expenditures (EE) and respiratory quotients (RQ) in female monkeys of various body weights and ages. METHODS: 16 socially reared Bonnet Macaque female monkeys [5.5 +/- 1.4 kg body weight, modified BMI (length measurement from head to base of the tail) = 28.8 +/- 6.7 kg/crown-rump length, m2 and 11.7 +/- 4.6 years] were placed in the primate Enhanced Metabolic Testing Activity Chamber (Model 3000a, EMTAC Inc. Santa Barbara, CA) for 22-hour measurements of EE (kcal/kg) and RQ (VCO2/VO2). All were fed monkey chow (4.03 kcal/g) ad-libitum under a 12/12 hour light/dark cycle. Metabolic data were corrected for differences in body weight. Results were divided into day (8-hours), dark (12 hours) and morning (2-hours) periods. Data analysis was conducted utilizing SPSS (Version 13). RESULTS: Modified BMI negatively correlated with 22-hour energy expenditure in all monkeys (r = -0.80, p < 0.01). The large variability of daily energy intake (4.5 to 102.0 kcal/kg) necessitated division into two groups, non-eaters (< 13 kcal/kg) and eaters (> 23 kcal/kg). There were reductions (p < 0.05) in both 22-hour and dark period RQs in the "non-eaters" in comparison to those who were "eaters". Monkeys were also classified as "lean" (modified BMI < 25) or "obese" (modified BMI > 30). The obese group had lower EE (p < 0.05) during each time period and over the entire 22-hours (p < 0.05), in comparison to their lean counterparts. CONCLUSION: The EMTAC proved to be a valuable tool for metabolic measurements in monkeys. The accuracy and sensitivity of the instrument allowed detection of subtle metabolic changes in relation to energy intake. Moreover, there is an association between a reduction of energy expenditure and a gain in body weight.