Identification of Idiopathic Pulmonary Fibrosis and Prediction of Disease Severity via Machine Learning Analysis of Comprehensive Metabolic Panel and Complete Blood Count Data.

BACKGROUND: Diagnosis of idiopathic pulmonary fibrosis (IPF) typically relies on high-resolution computed tomography imaging (HRCT) or histopathology, while monitoring disease severity is done via frequent pulmonary function testing (PFT). More reliable and convenient methods of diagnosing fibrotic interstitial lung disease (ILD) type and monitoring severity would allow for early identification and enhance current therapeutic interventions. This study tested the hypothesis that a machine learning (ML) ensemble analysis of comprehensive metabolic panel (CMP) and complete blood count (CBC) data can accurately distinguish IPF from connective tissue disease ILD (CTD-ILD) and predict disease severity as seen with PFT. METHODS: Outpatient data with diagnosis of IPF or CTD-ILD (n = 103 visits by 53 patients) were analyzed via ML methodology to evaluate (1) IPF vs CTD-ILD diagnosis; (2) %predicted Diffusing Capacity of Lung for Carbon Monoxide (DLCO) moderate or mild vs severe; (3) %predicted Forced Vital Capacity (FVC) moderate or mild vs severe; and (4) %predicted FVC mild vs moderate or severe. RESULTS: ML methodology identified IPF from CTD-ILD with AUCTEST = 0.893, while PFT was classified as DLCO moderate or mild vs severe with AUCTEST = 0.749, FVC moderate or mild vs severe with AUCTEST = 0.741, and FVC mild vs moderate or severe with AUCTEST = 0.739. Key features included albumin, alanine transaminase, %lymphocytes, hemoglobin, %eosinophils, white blood cell count, %monocytes, and %neutrophils. CONCLUSION: Analysis of CMP and CBC data via proposed ML methodology offers the potential to distinguish IPF from CTD-ILD and predict severity on associated PFT with accuracy that meets or exceeds current clinical practice.

A multi-institutional evaluation of active surveillance for low risk prostate cancer.

PURPOSE: For select men with low risk prostate cancer active surveillance is more often being considered a management strategy. In a multicenter retrospective study we evaluated the actuarial rates and predictors of remaining on active surveillance, the incidence of cancer progression and the pathological findings of delayed radical prostatectomy. MATERIALS AND METHODS: A cohort of 262 men from 4 institutions met the inclusion criteria of age 75 years or younger, prostate specific antigen 10 ng/ml or less, clinical stage T1-T2a, biopsy Gleason sum 6 or less, 3 or less positive cores at diagnostic biopsy, repeat biopsy before active surveillance and no treatment for 6 months following the repeat biopsy. Active surveillance started on the date of the second biopsy. Actuarial rates of remaining on active surveillance were calculated and univariate Cox regression was used to assess predictors of discontinuing active surveillance. RESULTS: With a median followup of 29 months 43 patients ultimately received active treatment. The 2 and 5-year probabilities of remaining on active surveillance were 91% and 75%, respectively. Patients with cancer on the second biopsy (HR 2.23, 95% CI 1.23-4.06, p = 0.007) and a higher number of cancerous cores from the 2 biopsies combined (p = 0.002) were more likely to undergo treatment. Age, prostate specific antigen, clinical stage, prostate volume and number of total biopsy cores sampled were not predictive of outcome. Skeletal metastases developed in 1 patient 38 months after starting active surveillance. Of the 43 patients undergoing delayed treatment 41 (95%) are without disease progression at a median of 23 months following treatment. CONCLUSIONS: With a median followup of 29 months active surveillance for select patients appears to be safe and associated with a low risk of systemic progression. Cancer at restaging biopsy and a higher total number of cancerous cores are associated with a lower likelihood of remaining on active surveillance. A restaging biopsy should be strongly considered to finalize eligibility for active surveillance.

Prediction of lung cancer immunotherapy response via machine learning analysis of immune cell lineage and surface markers.

BACKGROUND: Although advances have been made in cancer immunotherapy, patient benefits remain elusive. For non-small cell lung cancer (NSCLC), monoclonal antibodies targeting programmed death-1 (PD-1) and programmed death ligand-1 (PD-L1) have shown survival benefit compared to chemotherapy. Personalization of treatment would be facilitated by a priori identification of patients likely to benefit. OBJECTIVE: This pilot study applied a suite of machine learning methods to analyze mass cytometry data of immune cell lineage and surface markers from blood samples of a small cohort (n= 13) treated with Pembrolizumab, Atezolizumab, Durvalumab, or Nivolumab as monotherapy. METHODS: Four different comparisons were evaluated between data collected at an initial visit (baseline), after 12-weeks of immunotherapy, and from healthy (control) samples: healthy vs patients at baseline, Responders vs Non-Responders at baseline, Healthy vs 12-week Responders, and Responders vs Non-Responders at 12-weeks. The algorithms Random Forest, Partial Least Squares Discriminant Analysis, Multi-Layer Perceptron, and Elastic Net were applied to find features differentiating between these groups and provide for the capability to predict outcomes. RESULTS: Particular combinations and proportions of immune cell lineage and surface markers were sufficient to accurately discriminate between the groups without overfitting the data. In particular, markers associated with the B-cell phenotype were identified as key features. CONCLUSIONS: This study illustrates a comprehensive machine learning analysis of circulating immune cell characteristics of NSCLC patients with the potential to predict response to immunotherapy. Upon further evaluation in a larger cohort, the proposed methodology could help guide personalized treatment selection in clinical practice.

A multi-institutional evaluation of active surveillance for low risk prostate cancer.

PURPOSE: For select men with low risk prostate cancer active surveillance is more often being considered a management strategy. In a multicenter retrospective study we evaluated the actuarial rates and predictors of remaining on active surveillance, the incidence of cancer progression and the pathological findings of delayed radical prostatectomy. MATERIALS AND METHODS: A cohort of 262 men from 4 institutions met the inclusion criteria of age 75 years or younger, prostate specific antigen 10 ng/ml or less, clinical stage T1-T2a, biopsy Gleason sum 6 or less, 3 or less positive cores at diagnostic biopsy, repeat biopsy before active surveillance and no treatment for 6 months following the repeat biopsy. Active surveillance started on the date of the second biopsy. Actuarial rates of remaining on active surveillance were calculated and univariate Cox regression was used to assess predictors of discontinuing active surveillance. RESULTS: With a median followup of 29 months 43 patients ultimately received active treatment. The 2 and 5-year probabilities of remaining on active surveillance were 91% and 75%, respectively. Patients with cancer on the second biopsy (HR 2.23, 95% CI 1.23-4.06, p = 0.007) and a higher number of cancerous cores from the 2 biopsies combined (p = 0.002) were more likely to undergo treatment. Age, prostate specific antigen, clinical stage, prostate volume and number of total biopsy cores sampled were not predictive of outcome. Skeletal metastases developed in 1 patient 38 months after starting active surveillance. Of the 43 patients undergoing delayed treatment 41 (95%) are without disease progression at a median of 23 months following treatment. CONCLUSIONS: With a median followup of 29 months active surveillance for select patients appears to be safe and associated with a low risk of systemic progression. Cancer at restaging biopsy and a higher total number of cancerous cores are associated with a lower likelihood of remaining on active surveillance. A restaging biopsy should be strongly considered to finalize eligibility for active surveillance.

Estimating Propensity Parameters Using Google PageRank and Genetic Algorithms.

Stochastic Boolean networks, or more generally, stochastic discrete networks, are an important class of computational models for molecular interaction networks. The stochasticity stems from the updating schedule. Standard updating schedules include the synchronous update, where all the nodes are updated at the same time, and the asynchronous update where a random node is updated at each time step. The former produces a deterministic dynamics while the latter a stochastic dynamics. A more general stochastic setting considers propensity parameters for updating each node. Stochastic Discrete Dynamical Systems (SDDS) are a modeling framework that considers two propensity parameters for updating each node and uses one when the update has a positive impact on the variable, that is, when the update causes the variable to increase its value, and uses the other when the update has a negative impact, that is, when the update causes it to decrease its value. This framework offers additional features for simulations but also adds a complexity in parameter estimation of the propensities. This paper presents a method for estimating the propensity parameters for SDDS. The method is based on adding noise to the system using the Google PageRank approach to make the system ergodic and thus guaranteeing the existence of a stationary distribution. Then with the use of a genetic algorithm, the propensity parameters are estimated. Approximation techniques that make the search algorithms efficient are also presented and Matlab/Octave code to test the algorithms are available at http://www.ms.uky.edu/~dmu228/GeneticAlg/Code.html.

The PDS-XADS reference accelerator and its radioprotection issues.

At the start-up of the FP5 project PDS-XADS, the main initial specifications for the accelerator system (e.g. beam energy, beam intensity, beam profile, their stability and the accelerator availability and reliability) have been defined by Working Package 1 (WP1), 'Global Coherence', in connection with the other WPs of the project. From this, WP3 ('The Accelerator') has assessed the main requirements and the corresponding technical answers. A reference solution, based on a linear superconducting accelerator with its associated doubly achromatic beam line has been worked out to some detail. For high reliability, the proposed design is intrinsically fault tolerant, relying on highly modular 'de-rated' components associated with a fast digital feedback system. The proposed solution also appears to be robust in terms of operational aspects like maintenance and radioprotection. A programme for the remaining R&D required has been elaborated and is proposed within the FP6 project EUROTRANS.

Involvement of cytolytic and non-cytolytic T cells in the control of feline immunodeficiency virus infection.

The appearance of non-cytolytic T cells that suppressed feline immunodeficiency virus (FIV) replication in vitro, and FIV-specific cytotoxic T cell (CTL) responses was compared in a group of seven, specific pathogen free (SPF) domestic cats following primary infection with the Glasgow(8) isolate of FIV (FIV(GL-8)). FIV proviral burdens were quantified in the blood and lymphoid tissues by real-time PCR. Non-cytolytic T cell suppression of FIV replication was measured by co-cultivating lymphoblasts prepared from the cats at different time-points during infection with FIV-infected MYA-1 cells in vitro. Non-cytolytic suppressor activity was detected as early as 1 week after infection, and was evident in all the lymphoid tissues examined. Further, this activity was present in subpopulations of T cells in the blood with normal (CD8(hi)) or reduced (CD8(lo)) expression of the CD8 molecule, and temporal modulations in non-cytolytic suppressor activity were unrelated to the circulating CD8(+) T cell numbers. Virus-specific CTL responses, measured by (51)Cr release assays, were not detected until 4 weeks after infection, with the emergence of FIV-specific effector CTLs in the blood. Throughout infection the response was predominantly directed towards FIV Gag-expressing target cells, and by 47 weeks after infection CTL responses had become localised in the lymph nodes and spleen. The results suggest that both non-cytolytic T cell suppression of FIV replication and FIV-specific CTL responses are important cellular immune mechanisms in the control of FIV replication in infected asymptomatic cats.