Developing a Cancer Digital Twin: Supervised Metastases Detection From Consecutive Structured Radiology Reports.

The development of digital cancer twins relies on the capture of high-resolution representations of individual cancer patients throughout the course of their treatment. Our research aims to improve the detection of metastatic disease over time from structured radiology reports by exposing prediction models to historical information. We demonstrate that Natural language processing (NLP) can generate better weak labels for semi-supervised classification of computed tomography (CT) reports when it is exposed to consecutive reports through a patient's treatment history. Around 714,454 structured radiology reports from Memorial Sloan Kettering Cancer Center adhering to a standardized departmental structured template were used for model development with a subset of the reports included for validation. To develop the models, a subset of the reports was curated for ground-truth: 7,732 total reports in the lung metastases dataset from 867 individual patients; 2,777 reports in the liver metastases dataset from 315 patients; and 4,107 reports in the adrenal metastases dataset from 404 patients. We use NLP to extract and encode important features from the structured text reports, which are then used to develop, train, and validate models. Three models-a simple convolutional neural network (CNN), a CNN augmented with an attention layer, and a recurrent neural network (RNN)-were developed to classify the type of metastatic disease and validated against the ground truth labels. The models use features from consecutive structured text radiology reports of a patient to predict the presence of metastatic disease in the reports. A single-report model, previously developed to analyze one report instead of multiple past reports, is included and the results from all four models are compared based on accuracy, precision, recall, and F1-score. The best model is used to label all 714,454 reports to generate metastases maps. Our results suggest that NLP models can extract cancer progression patterns from multiple consecutive reports and predict the presence of metastatic disease in multiple organs with higher performance when compared with a single-report-based prediction. It demonstrates a promising automated approach to label large numbers of radiology reports without involving human experts in a time- and cost-effective manner and enables tracking of cancer progression over time.

Substantial Targeting Advantage Achieved by Pulmonary Administration of Colistin Methanesulfonate in a Large-Animal Model.

Colistin, administered as its inactive prodrug colistin methanesulfonate (CMS), is often used in multidrug-resistant Gram-negative pulmonary infections. The CMS and colistin pharmacokinetics in plasma and epithelial lining fluid (ELF) following intravenous and pulmonary dosing have not been evaluated in a large-animal model with pulmonary architecture similar to that of humans. Six merino sheep (34 to 43 kg body weight) received an intravenous or pulmonary dose of 4 to 8 mg/kg CMS (sodium) or 2 to 3 mg/kg colistin (sulfate) in a 4-way crossover study. Pulmonary dosing was achieved via jet nebulization through an endotracheal tube cuff. CMS and colistin were quantified in plasma and bronchoalveolar lavage fluid (BALF) samples by high-performance liquid chromatography (HPLC). ELF concentrations were calculated via the urea method. CMS and colistin were comodeled in S-ADAPT. Following intravenous CMS or colistin administration, no concentrations were quantifiable in BALF samples. Elimination clearance was 1.97 liters/h (4% interindividual variability) for CMS (other than conversion to colistin) and 1.08 liters/h (25%) for colistin. On average, 18% of a CMS dose was converted to colistin. Following pulmonary delivery, colistin was not quantifiable in plasma and CMS was detected in only one sheep. Average ELF concentrations (standard deviations [SD]) of formed colistin were 400 (243), 384 (187), and 184 (190) mg/liter at 1, 4, and 24 h after pulmonary CMS administration. The population pharmacokinetic model described well CMS and colistin in plasma and ELF following intravenous and pulmonary administration. Pulmonary dosing provided high ELF and low plasma colistin concentrations, representing a substantial targeting advantage over intravenous administration. Predictions from the pharmacokinetic model indicate that sheep are an advantageous model for translational research.

Dynamics of IL-4 and IL-13 expression in the airways of sheep following allergen challenge.

BACKGROUND: IL-4 and IL-13 play a critical yet poorly understood role in orchestrating the recruitment and activation of effector cells of the asthmatic response and driving the pathophysiology of allergic asthma. The house dust mite (HDM) sheep asthma model displays many features of the human condition and is an ideal model to further elucidate the involvement of these critical Th2 cytokines. We hypothesized that airway exposure to HDM allergen would induce or elevate the expression profile of IL-4 and IL-13 during the allergic airway response in this large animal model of asthma. METHODS: Bronchoalveolar lavage (BAL) samples were collected from saline- and house dust mite (HDM)- challenged lung lobes of sensitized sheep from 0 to 48 h post-challenge. BAL cytokines (IL-4, IL-13, IL-6, IL-10, TNF-α) were each measured by ELISA. IL-4 and IL-13 expression was assessed in BAL leukocytes by flow cytometry and in airway tissue sections by immunohistology. RESULTS: IL-4 and IL-13 were increased in BAL samples following airway allergen challenge. HDM challenge resulted in a significant increase in BAL IL-4 levels at 4 h compared to saline-challenged airways, while BAL IL-13 levels were elevated at all time-points after allergen challenge. IL-6 levels were maintained following HDM challenge but declined after saline challenge, while HDM administration resulted in an acute elevation in IL-10 at 4 h but no change in TNF-α levels over time. Lymphocytes were the main early source of IL-4, with IL-4 release by alveolar macrophages (AMs) prominent from 24 h post-allergen challenge. IL-13 producing AMs were increased at 4 and 24 h following HDM compared to saline challenge, and tissue staining provided evidence of IL-13 expression in airway epithelium as well as immune cells in airway tissue. CONCLUSION: In a sheep model of allergic asthma, airway inflammation is accompanied by the temporal release of key cytokines following allergen exposure that primarily reflects the Th2-driven nature of the immune response in asthma. The present study demonstrates for the first time the involvement of IL-4 and IL-13 in a relevant large animal model of allergic airways disease.

A mouse model to evaluate the impact of species, sex, and lipid load on lymphatic drug transport.

PURPOSE: To establish a lymph-cannulated mouse model, and use the model to investigate the impact of lipid dose on exogenous and endogenous lipid recruitment, and drug transport, into the lymph of males versus females. Finally, lymphatic transport and drug absorption in the mouse were compared to other pre-clinical models (rats/dogs). METHODS: Animals were orally or intraduodenally administered 1.6 mg/kg halofantrine in low or high (14)C-lipid doses. For bioavailability calculation, animals were intravenuosly administered halofantrine. Lymph or blood samples were taken and halofantrine, triglyceride, phospholipid and (14)C-lipid concentrations measured. RESULTS: Lymphatic lipid transport increased linearly with lipid dose, was similar across species and in male/female animals. In contrast, lymphatic transport of halofantrine differed markedly across species (dogs>rats>mice) and plateaued at higher lipid doses. Lower bioavailability appeared responsible for some species differences in halofantrine lymphatic transport; however other systematic differences were involved. CONCLUSIONS: A contemporary lymph-cannulated mouse model was established which will enable investigation of lymphatic transport in transgenic and disease models. The current study found halofantrine absorption and lymphatic transport are reduced in small animals. Future analyses will investigate mechanisms involved, and if similar trends occur for other drugs, to establish the most relevant model(s) to predict lymphatic transport in humans.

Fatty acid binding proteins: potential chaperones of cytosolic drug transport in the enterocyte?

PURPOSE: Several poorly water-soluble drugs have previously been shown to bind to intestinal (I-FABP) and liver fatty acid binding protein (L-FABP) in vitro. The purpose of this study was to examine the potential role of drug binding to FABPs on intestinal permeability and gut wall metabolism in vivo. METHODS: The intestinal permeability of ibuprofen, progesterone and midazolam (which bind FABPs) and propranolol (which does not) was examined using an autoperfused recirculating permeability model in control rats and rats where FABP levels were upregulated via pre-feeding a fat-rich diet. RESULTS: The intestinal permeability of drugs which bind FABPs in vitro was increased in animals where FABP levels were upregulated by prefeeding a high fat diet. The gut wall metabolism of midazolam was also reduced in animals with elevated FABP levels. CONCLUSIONS: Consistent with their role in the cellular transport of endogenous lipophilic substrates, FABPs appear to facilitate the intracellular disposition of drug molecules that bind FABPs in vitro. Drug binding to FABPs in the enterocyte may also attenuate gut wall metabolism in a manner analogous to the reduction in hepatic extraction mediated by drug binding to plasma proteins in the systemic circulation.

Lymphatic transport of Methylnortestosterone undecanoate (MU) and the bioavailability of methylnortestosterone are highly sensitive to the mass of coadministered lipid after oral administration of MU.

The contribution of lymphatic transport to the oral bioavailability of methylnortestosterone (M) after oral administration of the lipophilic prodrug methylnortestosterone undecanoate (MU) has been evaluated, and the sensitivity of lymphatic MU transport to lymphatic lipid transport has been investigated. M and MU were administered intravenously and orally to greyhound dogs to determine absolute bioavailability after oral dosing of MU. MU was also administered orally with differing quantities of food (lipid) to lymph duct-cannulated greyhound dogs to investigate the relative roles of lymph versus blood transport on M bioavailability and the effect of lipid load on systemic exposure. The relationship between lymphatic lipid and MU transport was further investigated in anesthetized rats. The oral bioavailability of M after administration of MU was found to be highly dependent on coadministration of food, and the bioavailability of M increased approximately 700% in fed versus fasted animals. In both cases, lymph diversion resulted in negligible systemic exposure of M, indicating almost complete dependence on lymphatic transport of MU for systemic exposure of M. Lymphatic transport of MU was even more highly dependent on the quantity of coadministered lipid and increased more than 50-fold with increasing lipid load. Therefore, increasing the quantity of food or lipid coadministered with MU stimulated a significant increase in the lymphatic transport of MU and systemic exposure of M. The lipid sensitivity of lymphatic transport of MU is significantly higher than previously observed for more metabolically stable compounds, suggesting a role for coadministered lipid in promoting avoidance of enterocyte-based cleavage of MU.

In vitro activity of anidulafungin against Candida albicans biofilms.

We tested the activity of anidulafungin against 30 Candida albicans isolates. The planktonic MICs for 50 and 90% of the isolates tested (MIC(50) and MIC(90)) were < or =0.03 and 0.125 microg/ml, respectively (MIC range, < or =0.03 to 2 microg/ml). The sessile MIC(50) and MIC(90) were < or =0.03 and < or =0.03 microg/ml, respectively (MIC range, < or =0.03 to >16 microg/ml).