A Machine Learning System to Indicate Diagnosis of Idiopathic Pulmonary Fibrosis Non-Invasively in Challenging Cases.

Radiologic usual interstitial pneumonia (UIP) patterns and concordant clinical characteristics define a diagnosis of idiopathic pulmonary fibrosis (IPF). However, limited expert access and high inter-clinician variability challenge early and pre-invasive diagnostic sensitivity and differentiation of IPF from other interstitial lung diseases (ILDs). We investigated a machine learning-driven software system, Fibresolve, to indicate IPF diagnosis in a heterogeneous group of 300 patients with interstitial lung disease work-up in a retrospective analysis of previously and prospectively collected registry data from two US clinical sites. Fibresolve analyzed cases at the initial pre-invasive assessment. An Expert Clinical Panel (ECP) and three panels of clinicians with varying experience analyzed the cases for comparison. Ground Truth was defined by separate multi-disciplinary discussion (MDD) with the benefit of surgical pathology results and follow-up. Fibresolve met both pre-specified co-primary endpoints of sensitivity superior to ECP and significantly greater specificity (p = 0.0007) than the non-inferior boundary of 80.0%. In the key subgroup of cases with thin-slice CT and atypical UIP patterns (n = 124), Fibresolve's diagnostic yield was 53.1% [CI: 41.3-64.9] (versus 0% pre-invasive clinician diagnostic yield in this group), and its specificity was 85.9% [CI: 76.7-92.6%]. Overall, Fibresolve was found to increase the sensitivity and diagnostic yield for IPF among cases of patients undergoing ILD work-up. These results demonstrate that in combination with standard clinical assessment, Fibresolve may serve as an adjunct in the diagnosis of IPF in a pre-invasive setting.

A new continuous noninvasive finger cuff device (Vitalstream) for cardiac output that communicates wirelessly via bluetooth or Wi-Fi.

BACKGROUND: The new noninvasive Vitalstream (VS) continuous physiological monitor (Caretaker Medical LLC, Charlottesville, Virginia), allows continuous cardiac output by a low pump-inflated, finger cuff that pneumatically couples arterial pulsations via a pressure line to a pressure sensor for detection and analysis. Physiological data are communicated wirelessly to a tablet-based user interface via Bluetooth or Wi-Fi. We evaluated its performance against thermodilution cardiac output in patients undergoing cardiac surgery. METHODS: We compared the agreement between thermodilution cardiac output to that obtained by the continuous noninvasive system during cardiac surgery pre and post-cardiac bypass. Thermodilution cardiac output was performed routinely when clinically indicated by an iced saline cold injectate system. All comparisons between VS and TD/CCO data were post-processed. In order to match the VS CO readings to the averaged discrete TD bolus data, the averaged CO readings of the ten seconds of VS CO data points prior to a sequence of TD bolus injections was matched. Time alignment was based on the medical record time and the VS time-stamped data points. The accuracy against reference TD measurements was assessed via Bland-Altman analysis of the CO values and standard concordance analysis of the ΔCO values (with a 15% exclusion zone). RESULTS: Analysis of the data compared the accuracy of the matched measurement pairs of VS and TD/CCO VS absolute CO values with and without initial calibration to the discrete TD CO values, as well as the trending ability, i.e., ΔCO values of the VS physiological monitor compared to those of the reference. The results were comparable with other non-invasive as well as invasive technologies and Bland-Altman analyses showed high agreement between devices in a diverse patient population. The results are significant regarding the goal of expanding access to effective, wireless and readily implemented fluid management monitoring tools to hospital sections previously not covered because of the limitations of traditional technologies. CONCLUSION: This study demonstrated that the agreement between the VS CO and TD CO was clinically acceptable with a percent error (PE) of 34.5 to 38% with and without external calibration. The threshold for an acceptable agreement between the VS and TD was considered to be below 40% which is below the threshold recommended by others.

The application of a neural network to predict hypotension and vasopressor requirements non-invasively in obstetric patients having spinal anesthesia for elective cesarean section (C/S).

BACKGROUND: Neural networks are increasingly used to assess physiological processes or pathologies, as well as to predict the increased likelihood of an impending medical crisis, such as hypotension. METHOD: We compared the capabilities of a single hidden layer neural network of 12 nodes to those of a discrete-feature discrimination approach with the goal being to predict the likelihood of a given patient developing significant hypotension under spinal anesthesia when undergoing a Cesarean section (C/S). Physiological input information was derived from a non-invasive blood pressure device (Caretaker [CT]) that utilizes a finger cuff to measure blood pressure and other hemodynamic parameters via pulse contour analysis. Receiver-operator-curve/area-under-curve analyses were used to compare performance. RESULTS: The results presented here suggest that a neural network approach (Area Under Curve [AUC] = 0.89 [p < 0.001]), at least at the implementation level of a clinically relevant prediction algorithm, may be superior to a discrete feature quantification approach (AUC = 0.87 [p < 0.001]), providing implicit access to a plurality of features and combinations thereof. In addition, the expansion of the approach to include the submission of other physiological data signals, such as heart rate variability, to the network can be readily envisioned. CONCLUSION: This pilot study has demonstrated that increased coherence in Arterial Stiffness (AS) variability obtained from the pulse wave analysis of a continuous non-invasive blood pressure device appears to be an effective predictor of hypotension after spinal anesthesia in the obstetrics population undergoing C/S. This allowed us to predict specific dosing thresholds of phenylephrine required to maintain systolic blood pressure above 90 mmHg.

Reassessing public opinion of captive cetacean attractions with a photo elicitation survey.

BACKGROUND: Captive cetacean attractions are growing in number globally, their operators citing entertainment, education, and conservation as benefits. Those for and against developing such attractions claim public support. Previous public opinion research, however, shows little consensus, partly due to the introduction of biases in study design that influence participants' responses. Those involved in, or concerned with, developing and licensing these attractions need to better understand what drives the lack of consensus to take socially-acceptable decisions. METHODS: We reviewed previous research on public opinion of cetacean captivity, noting possible sources of bias. Survey question wording can be a major source of introduced bias, so we used an open-ended photo elicitation approach. We showed tourists in the Turks Caicos Islands (N = 292) images of a marine mammal park (MMP) killer whale show and a swim-with-the-dolphins (SWTD) attraction and asked for their qualitative comments on the potential development of each. They also indicated how likely they would be to visit each on a Likert scale. RESULTS: Respondents were generally against visiting MMP killer whale shows, with 60.9% not likely to visit. SWTD attractions were more popular; 60.3% were likely to visit. For SWTD, USA residents were more likely to visit; older respondents and those staying in all-inclusive resorts were less likely. Those staying in all-inclusive resorts were also less likely to visit MMP killer whale shows. The great majority of qualitative comments centred on either entertainment value or animal welfare concerns. There were very few, if any, comments on the education or conservation value of these attractions. DISCUSSION: Our findings contradict several previous studies on public opinion of captive cetaceans that did not use photo elicitation. The support shown for MMP killer whale shows in this survey was well below that claimed by studies conducted on behalf of captive cetacean attraction operators. Opposition to SWTD was also noticeably lower than that found in surveys conducted with wild cetacean tourism participants. This difference can likely be attributed to the different survey populations and settings, but this variation is also very likely attributable to researcher-introduced bias. While photo selection can introduce bias, photo elicitation reduces reliance on pre-scripted questions and responses, and seems to effectively reduce other forms of bias. Allowing open-ended responses, where participants responded to an image, seems to have given a more representative understanding of what is at the forefront of the public's mind than closed questioning. These conclusions, among others made in this study, suggest that development decisions for captive cetacean attractions are being made on imprecise data. Going forward, data collected via responder-led, open-ended, bias-minimising approaches should at least be considered when informing such decisions.

Continuous Non-invasive finger cuff CareTaker® comparable to invasive intra-arterial pressure in patients undergoing major intra-abdominal surgery.

BACKGROUND: Despite increased interest in non-invasive arterial pressure monitoring, the majority of commercially available technologies have failed to satisfy the limits established for the validation of automatic arterial pressure monitoring by the Association for the Advancement of Medical Instrumentation (AAMI). According to the ANSI/AAMI/ISO 81060-2:2013 standards, the group-average accuracy and precision are defined as acceptable if bias is not greater than 5 mmHg and standard deviation is not greater than 8 mmHg. In this study, these standards are used to evaluate the CareTaker® (CT) device, a device measuring continuous non-invasive blood pressure via a pulse contour algorithm called Pulse Decomposition Analysis. METHODS: A convenience sample of 24 patients scheduled for major abdominal surgery were consented to participate in this IRB approved pilot study. Each patient was monitored with a radial arterial catheter and CT using a finger cuff applied to the contralateral thumb. Hemodynamic variables were measured and analyzed from both devices for the first thirty minutes of the surgical procedure including the induction of anesthesia. The mean arterial pressure (MAP), systolic and diastolic blood pressures continuously collected from the arterial catheter and CT were compared. Pearson correlation coefficients were calculated between arterial catheter and CT blood pressure measurements, a Bland-Altman analysis, and polar and 4Q plots were created. RESULTS: The correlation of systolic, diastolic, and mean arterial pressures were 0.92, 0.86, 0.91, respectively (p < 0.0001 for all the comparisons). The Bland-Altman comparison yielded a bias (as measured by overall mean difference) of -0.57, -2.52, 1.01 mmHg for systolic, diastolic, and mean arterial pressures, respectively with a standard deviation of 7.34, 6.47, 5.33 mmHg for systolic, diastolic, and mean arterial pressures, respectively (p < 0.001 for all comparisons). The polar plot indicates little bias between the two methods (90%/95% CI at 31.5°/52°, respectively, overall bias = 1.5°) with only a small percentage of points outside these lines. The 4Q plot indicates good concordance and no bias between the methods. CONCLUSIONS: In this study, blood pressure measured using the non-invasive CT device was shown to correlate well with the arterial catheter measurements. Larger studies are needed to confirm these results in more varied settings. Most patients exhibited very good agreement between methods. Results were well within the limits established for the validation of automatic arterial pressure monitoring by the AAMI.

Global analysis of cellular proteolysis by selective enzymatic labeling of protein N-termini.

Proteolysis is a critical modification leading to alteration of protein function with important outcomes in many biological processes. However, for the majority of proteases, we have an incomplete understanding of both cellular substrates and downstream effects. Here, we describe detailed protocols and applications for using the rationally engineered peptide ligase, subtiligase, to specifically label and capture protein N-termini generated by proteases either induced or added to complex biological samples. This method allows identification of the protein targets as well as their precise cleavage locations. This approach has revealed >8000 proteolytic sites in healthy and apoptotic cells including >1700 caspase cleavages. One can further determine substrate preferences through rate analysis with quantitative mass spectrometry, physiological substrate specificities, and even infer the identity of proteases operating in the cell. In this chapter, we also describe how this experimental method can be generalized to investigate proteolysis in any biological sample.

The DegraBase: a database of proteolysis in healthy and apoptotic human cells.

Proteolysis is a critical post-translational modification for regulation of cellular processes. Our lab has previously developed a technique for specifically labeling unmodified protein N termini, the α-aminome, using the engineered enzyme, subtiligase. Here we present a database, called the DegraBase (http://wellslab.ucsf.edu/degrabase/), which compiles 8090 unique N termini from 3206 proteins directly identified in subtiligase-based positive enrichment mass spectrometry experiments in healthy and apoptotic human cell lines. We include both previously published and unpublished data in our analysis, resulting in a total of 2144 unique α-amines identified in healthy cells, and 6990 in cells undergoing apoptosis. The N termini derive from three general categories of proteolysis with respect to cleavage location and functional role: translational N-terminal methionine processing (∼10% of total proteolysis), sites close to the translational N terminus that likely represent removal of transit or signal peptides (∼25% of total), and finally, other endoproteolytic cuts (∼65% of total). Induction of apoptosis causes relatively little change in the first two proteolytic categories, but dramatic changes are seen in endoproteolysis. For example, we observed 1706 putative apoptotic caspase cuts, more than double the total annotated sites in the CASBAH and MEROPS databases. In the endoproteolysis category, there are a total of nearly 3000 noncaspase nontryptic cleavages that are not currently reported in the MEROPS database. These studies significantly increase the annotation for all categories of proteolysis in human cells and allow public access for investigators to explore interesting proteolytic events in healthy and apoptotic human cells.