Inference of dynamic systems from noisy and sparse data via manifold-constrained Gaussian processes

Parameter estimation for nonlinear dynamic system models, represented by ordinary differential equations (ODEs), using noisy and sparse data, is a vital task in many fields. We propose a fast and accurate method, manifold-constrained Gaussian process inference (MAGI), for this task. MAGI uses a Gaussian process model over time series data, explicitly conditioned on the manifold constraint that derivatives of the Gaussian process must satisfy the ODE system. By doing so, we completely bypass the need for numerical integration and achieve substantial savings in computational time. MAGI is also suitable for inference with unobserved system components, which often occur in real experiments. MAGI is distinct from existing approaches as we provide a principled statistical construction under a Bayesian framework, which incorporates the ODE system through the manifold constraint. We demonstrate the accuracy and speed of MAGI using realistic examples based on physical experiments.

Dynamic systems, represented as a set of ordinary differential equations (ODEs), are commonly used to model behaviors in scientific domains, such as gene regulation (1), biological rhythms (2), spread of disease (3), ecology (4), etc. We focus on models specified by a set of ODEsx˙(t)=dx(t)dt=f(x(t),θ,t), t∈[0,T],[1]where the vector x(t) contains the system outputs that evolve over time t and θ is the vector of model parameters to be estimated from experimental/observational data. When f is nonlinear, solving x(t) given initial conditions x(0) and θ generally requires a numerical integration method, such as Runge–Kutta.Historically, ODEs have mainly been used for conceptual or theoretical understanding rather than data fitting as experimental data were limited. Advances in experimental and data collection techniques have increased the capacity to follow dynamic systems closer to real time. Such data will generally be recorded at discrete times and subject to measurement error. Thus, we assume that we observe y(τ)=x(τ)+ϵ(τ) at a set of observation time points τ with error ϵ governed by noise level σ. Our focus here is inference of θ given y(τ), with emphasis on nonlinear f where specialized methods that exploit a linear structure (e.g., refs. 5 and 6), are not generally applicable. We shall present a coherent, statistically principled framework for dynamic system inference with the help of Gaussian processes (GPs). The key to our method is to restrict the GPs on a manifold that satisfies the ODE system: Thus, we name our method MAGI (manifold-constrained Gaussian process inference). Placing a GP on x(t) facilitates inference of θ without numerical integration, and our explicit manifold constraint is the key idea that addresses the conceptual incompatibility between the GP and the specification of the ODE model, as we shall discuss shortly when overviewing our method. We show that the resulting parameter inference is computationally efficient, statistically principled, and effective in a variety of practical scenarios. MAGI particularly works in the cases when some system component(s) is/are unobserved. To the best of our knowledge, none of the current available software packages that do not use numerical integration can analyze systems with unobserved component(s).     

室上性宽QRS波群心动过速的临床分析

目的对自发室上性宽QRS波群心动过速的临床特征、心电图特点进行分析,并用Brugada4步诊断法与室速进行鉴别诊断,评价其敏感性及特异性。方法选择入院时为心动过速的患者,符合窦性P波消失、心室率>100bpm且QRS波群时限≥0.12s,心脏电生理检查明确诊断后,采用单盲法回顾性分析12导联体表心电图的形态学特征,独立做出诊断后,计算Brugada4步诊断法的特异性和敏感性并评价其实用性,同时分析其临床特征及血液动力学改变。结果入选50例室上性宽QRS波群心动过速患者,46例诊断为室上速伴束支阻滞或心室内差异性传导,特异性92%,敏感性100%;心室率<180bpm组与心室率>180bpm组,两组的平均收缩压、平均舒张压及平均动脉压(平均动脉压=舒张压 1/3脉压)均有差别,心室率较快组血压较低,p值<0.05,差别有统计学意义。结论Brugada法对室上性宽QRS波群心动过速的鉴别诊断有重要意义,正确的诊断和及时的处理对患者至关重要。     

First clinical evaluation of the CELL-DYN® 3200 haematology analyser

A prototype of the CELL-DYN 3200® haematology analyser was evaluated in a tertiary care hospital laboratory. Precision, effects of sample ageing, linearity, carry-over, and com-parability of cellular blood counts and five-part leucocyte differentiation were determined in accordance with the ICSH guidelines for the evaluation of blood cell analysers; the results were satisfactory for all parameters tested: haemoglobin concentration, RBC, MCV, WBC, platelet count, and counts of neutrophils, lymphocytes, monocytes, and eosinophils. Two-hundred and forty-seven routine blood samples were used for the comparability studies. The cellular blood count results from the CELL-DYN 3200® and the Bayer Diagnostic H-1 systems corresponded closely (correlation coefficient r > 0.96 for all parameters). For 201 samples without an instrument-generated suspect flag the same was true with regard to the dif-ferential parameters, although somewhat lower correlation was observed for monocyte counts (r = 0.88). Comparisons to 400-cell microscopic differentials gave similar results (r > 0.93 for neutrophil, lymphocyte and eosinophil counts). Our results suggest that the CELL-DYN 3200® analyser will serve the needs for automated blood cell counting and differential leucocyte counting in a tertiary care hospital laboratory.     

CEUS定量分析在肾脏肿瘤诊断中的应用

肾脏肿瘤发病率逐年升高,恶性肿瘤发病率更是居高不下。CEUS定量分析在肾脏肿瘤的诊断及疗效评价中发挥着重要作用,有助于鉴别肾脏肿瘤良恶性。本文对CEUS定量分析技术在肾脏占位诊断中的应用进行综述。     

乳腺CEUS预测模型诊断乳腺恶性病灶的观察者间一致性:多中心研究

目的 探讨拉伸指数模型DWI鉴别诊断乳腺良恶性病变的价值。方法 收集58例乳腺病变患者,共63个病灶（良性33个,恶性30个）,行多b值DWI及动态增强MRI （DCE-MRI）扫描。计算ADC、扩散分布指数（DDC）和扩散异质性指数（α）值,并生成时间-信号强度曲线（TIC）。比较良恶性病变间各参数差异,采用ROC曲线评价各参数诊断效能。结果 恶性病变ADC、DDC和α分别为（1.01±0.19）×10^-3 mm²/s、（0.89±0.23）×10^-3 mm²/s和0.75±0.09,良性病变分别为（1.41±0.27）×10^-3 mm²/s、（1.49±0.29）×10^-3 mm²/s和0.87±0.07,恶性病变均低于良性病变（P均<0.01）。各参数中DDC曲线下面积（AUC）最大（AUC=0.958）,最佳诊断界值1.22×10^-3 mm²/s,敏感度和特异度分别为96.67%、81.82%,DDC与TIC联合所得AUC为0.976,对应敏感度和特异度分别为93.33%、93.94%。结论 拉伸指数模型DWI参数DDC、α能够鉴别诊断乳腺良恶性病变,DDC与TIC联合的诊断效能高于ADC和DCE。     

影像学骶髂关节炎的鉴别诊断需密切结合临床特征

[目的]分析影像学骶髂关节炎的误诊病例和典型病例,提高对影像学骶髂关节炎包括强直性脊柱炎(AS)、致密性骨炎(OC)及跨专科疾病的认识.[方法](1)筛选出诊断明确,资料齐全的腰背痛或伴骶髂关节炎的误诊病例104例进行分析;(2)收集确诊的OC病例11例和AS病例50例,从临床表现、体格检查、影像学、实验室检查方面进行对比分析.[结果](1)104例误诊病例包括感染性疾病(29.81％)、骨关节疾病(26.92％)、内分泌代谢疾病(22.12％)、血液系统疾病(10.58％)和肿瘤(8.65％);(2)OC和AS的病例特点分析:OC组均为女性,均有下腰痛,大部分活动后加重,部分患者可有与AS相似的夜间痛和晨僵,Schober试验均阴性,CRP和ESR多正常,HLA-B27均阴性,骶髂关节X线和MRI显示OC的改变.AS组41例男性,9例女性,均有炎性腰痛表现,部分病人Schober试验阳性,CRP和ESR多升高,HLA-B27多阳性,骶髂关节X线和MRI显示AS的改变.[结论]影像学显示骶髂关节炎和伴有慢性腰背痛的疾病鉴别诊断需结合临床年龄、病史、症状、体征、实验室检查以及影像学检查等综合分析,注意跨专科疾病的诊断和鉴别,避免误诊.     

Difficulty in distinguishing malignant gastric lymphoma from advanced gastric cancer: Focusing on endoscopic findings of the Borrmann type

Malignant gastric lymphoma (MGL) accounts for a small proportion (upto 5%) of gastric malignancies. However, unlike for advanced gastric cancer (AGC) that requires surgical treatment, the standard treatments for MGL are chemotherapy and radiotherapy. Hence, the initial impression of the endoscopist is critical for the differential diagnosis and for planning future treatment. The purpose of this study was to assess the endoscopic diagnostic accuracy and the possibility of distinguishing between AGC and MGL depending on the endoscopist''s experience.A total of 48 patients who had MGL, and 48 age and sex-matched patients who had AGC were assessed by endoscopic review at a tertiary referral hospital between June 2008 and February 2017. Two endoscopic specialists reviewed the endoscopic findings and divided these diagnoses into 5 groups: Borrmann type (1, 2, 3, and 4) and early gastric cancer-like type. After this, 7 experts and 8 trainees were asked to complete a quiz that was comprised of 6 images for each of the 96 cases and to provide an endoscopic diagnosis for each case. The test results were analyzed to assess the diagnostic accuracy according to the pathologic results, endoscopic subgroups, and endoscopists’ experience. For inter-observer agreement was calculated with Fleiss kappa values.The overall diagnostic accuracy of endoscopic findings by the experts was 0.604 and that by the trainees was 0.493 (P = .050). There was no significant difference in the diagnosis according to the final pathology (lymphoma cases, 0.518 vs 0.440, P = .378; AGC cases, 0.690 vs 0.547, P = .089, respectively). In the subgroup analysis, the experts showed significantly higher diagnostic accuracy for the endoscopic Borrmann type 4 subgroup, including lymphoma or AGC cases, than the trainees (P = .001). Inter-observer agreement of final diagnosis (Fleiss kappa, 0.174) and endoscopic classification groups (Fleiss kappa, 0.123–0.271) was slightly and fair agreement.The experts tended to have a higher endoscopic diagnostic accuracy. Distinguishing MGL from AGC based on endoscopic findings is difficult, especially for the beginners. Even if the endoscopic impression is AGC, it is important to consider MGL in the differential diagnosis.     

A partial nucleated differential cell count of the bone marrow aspirate that is independent of peripheral blood dilution

In the bone marrow (BM) nucleated differential cell count (NDC), myeloblasts are enumerated as a percentage of total nucleated cells, which are inevitably diluted with peripheral blood nucleated cells (PBNC) during BM aspiration. We propose a partial NDC (PNDC) comprising only immature haemopoietic cells capable of division, i.e. myeloblasts, promyelocytes, myelocytes and erythroblasts. We show that the myeloid : erythroid (M : E) ratio of the PNDC remains approximately constant in progressively dilute aliquots of BM aspirates. We determined the PNDC in 22 healthy subjects and investigated the effect of peripheral blood dilution on disease stratification of 66 BM aspirates with myelodysplastic syndromes (MDS). NDC and PNDC myeloblast counts were compared and the equivalent PNDC myeloblast counts for NDC myeloblast threshold counts of 5, 10 and 20% were derived. Reclassification of MDS samples with the PNDC resulted in a change in disease category in 33.3% of 51 MDS samples with NDC myeloblast counts ranging from 3 to 26%. The PNDC is independent of PBNC dilution and can be determined in dilute BM samples. It alters the disease category in a significant proportion of BM aspirates with MDS and has the potential to better stratify MDS to improve clinical outcomes and treatment.