Neural networks for classification of cervical vertebrae maturation: a systematic review

OBJECTIVE

To assess the accuracy of identification and/or classification of the stage of cervical vertebrae maturity on lateral cephalograms by neural networks as compared with the ground truth determined by human observers.

MATERIALS AND METHODS

Search results from four electronic databases (PubMed [MEDLINE], Embase, Scopus, and Web of Science) were screened by two independent reviewers, and potentially relevant articles were chosen for full-text evaluation. Articles that fulfilled the inclusion criteria were selected for data extraction and methodologic assessment by the QUADAS-2 tool.

RESULTS

The search identified 425 articles across the databases, from which 8 were selected for inclusion. Most publications concerned the development of the models with different input features. Performance of the systems was evaluated against the classifications performed by human observers. The accuracy of the models on the test data ranged from 50% to more than 90%. There were concerns in all studies regarding the risk of bias in the index test and the reference standards. Studies that compared models with other algorithms in machine learning showed better results using neural networks.

CONCLUSIONS

Neural networks can detect and classify cervical vertebrae maturation stages on lateral cephalograms. However, further studies need to develop robust models using appropriate reference standards that can be generalized to external data.

Quantitative Patient-Reported Experience Measures Derived From Natural Language Processing Have a Normal Distribution and No Ceiling Effect

BACKGROUND AND OBJECTIVES

Patient-reported experience measures have the potential to guide improvement in health care delivery. Many patient-reported experience measures are limited by the presence of strong ceiling effects that limit their analytical utility.

METHODS

We used natural language processing to develop 2 new methods of evaluating patient experience using text comments and associated ordinal and categorical ratings of willingness to recommend from 1390 patients receiving specialty or nonspecialty care at our offices. One method used multivariable analysis based on linguistic factors to derive a formula to estimate the ordinal likelihood to recommend. The other method used the meaning extraction method of thematic analysis to identify words associated with categorical ratings of likelihood to recommend with which we created an equation to compute an experience score. We measured normality of the 2 score distributions and ceiling effects.

RESULTS

Spearman rank-order correlation analysis identified 36 emotional and linguistic constructs associated with ordinal rating of likelihood to recommend, 9 of which were independently associated in multivariable analysis. The calculation derived from this model corresponded with the original ordinal rating with an accuracy within 0.06 units on a 0 to 10 scale. This score and the score developed from thematic analysis both had a relatively normal distribution and limited or no ceiling effect.

CONCLUSIONS

Quantitative ratings of patient experience developed using natural language processing of text comments can have relatively normal distributions and no ceiling effect.

Children with psoriasis and COVID-19: factors associated with an unfavourable COVID-19 course, and the impact of infection on disease progression (Chi-PsoCov registry)

BACKGROUND

The COVID-19 pandemic has raised questions regarding the management of chronic skin diseases, especially in patients on systemic treatments. Data concerning the use of biologics in adults with psoriasis are reassuring, but data specific to children are missing. Moreover, COVID-19 could impact the course of psoriasis in children.

OBJECTIVES

The aim of this study was therefore to assess the impact of COVID-19 on the psoriasis of children, and the severity of the infection in relation to systemic treatments.

METHODS

We set up an international registry of paediatric psoriasis patients. Children were included if they were under 18 years of age, had a history of psoriasis, or developed it within 1 month of COVID-19 and had COVID-19 with or without symptoms.

RESULTS

One hundred and twenty episodes of COVID-19 in 117 children (mean age: 12.4 years) were reported. The main clinical form of psoriasis was plaque type (69.4%). Most children were without systemic treatment (54.2%); 33 (28.3%) were on biologic therapies, and 24 (20%) on non-biologic systemic drugs. COVID-19 was confirmed in 106 children (88.3%) and 3 children had two COVID-19 infections each. COVID-19 was symptomatic for 75 children (62.5%) with a mean duration of 6.5 days, significantly longer for children on non-biologic systemic treatments (P = 0.02) and without systemic treatment (P = 0.006) when compared with children on biologics. The six children who required hospitalization were more frequently under non-biologic systemic treatment when compared with the other children (P = 0.01), and particularly under methotrexate (P = 0.03). After COVID-19, the psoriasis worsened in 17 cases (15.2%). Nine children (8%) developed a psoriasis in the month following COVID-19, mainly a guttate form (P = 0.01).

DISCUSSION

Biologics appear to be safe with no increased risk of severe form of COVID-19 in children with psoriasis. COVID-19 was responsible for the development of psoriasis or the worsening of a known psoriasis for some children.

Evaluation of prognostic risk models for postoperative pulmonary complications in adult patients undergoing major abdominal surgery: a systematic review and international external validation cohort study

BACKGROUND

Stratifying risk of postoperative pulmonary complications after major abdominal surgery allows clinicians to modify risk through targeted interventions and enhanced monitoring. In this study, we aimed to identify and validate prognostic models against a new consensus definition of postoperative pulmonary complications.

METHODS

We did a systematic review and international external validation cohort study. The systematic review was done in accordance with the Preferred Reporting Items for Systematic Reviews and Meta-Analyses guidelines. We searched MEDLINE and Embase on March 1, 2020, for articles published in English that reported on risk prediction models for postoperative pulmonary complications following abdominal surgery. External validation of existing models was done within a prospective international cohort study of adult patients (≥18 years) undergoing major abdominal surgery. Data were collected between Jan 1, 2019, and April 30, 2019, in the UK, Ireland, and Australia. Discriminative ability and prognostic accuracy summary statistics were compared between models for the 30-day postoperative pulmonary complication rate as defined by the Standardised Endpoints in Perioperative Medicine Core Outcome Measures in Perioperative and Anaesthetic Care (StEP-COMPAC). Model performance was compared using the area under the receiver operating characteristic curve (AUROCC).

FINDINGS

In total, we identified 2903 records from our literature search; of which, 2514 (86·6%) unique records were screened, 121 (4·8%) of 2514 full texts were assessed for eligibility, and 29 unique prognostic models were identified. Nine (31·0%) of 29 models had score development reported only, 19 (65·5%) had undergone internal validation, and only four (13·8%) had been externally validated. Data to validate six eligible models were collected in the international external validation cohort study. Data from 11 591 patients were available, with an overall postoperative pulmonary complication rate of 7·8% (n=903). None of the six models showed good discrimination (defined as AUROCC ≥0·70) for identifying postoperative pulmonary complications, with the Assess Respiratory Risk in Surgical Patients in Catalonia score showing the best discrimination (AUROCC 0·700 [95% CI 0·683-0·717]).

INTERPRETATION

In the pre-COVID-19 pandemic data, variability in the risk of pulmonary complications (StEP-COMPAC definition) following major abdominal surgery was poorly described by existing prognostication tools. To improve surgical safety during the COVID-19 pandemic recovery and beyond, novel risk stratification tools are required.

FUNDING

British Journal of Surgery Society.

Abstract CT149: A phase 1/2a first-in-human study of AFM24, a CD16A/epidermal growth factor (EGFR) bispecific Innate Cell Engager (ICE®), in patients with locally advanced or metastatic EGFR-expressing solid tumors: Preliminary findings from the dose-escalation phase

AFM24 is a first-in-class, tetravalent, bispecific ICE® that binds to EGFR on tumor cells and CD16A on natural killer (NK) cells and macrophages, inducing antibody-mediated cytotoxicity and antibody-mediated phagocytosis, respectively. Preclinical in vitro and in vivo studies have demonstrated AFM24 can induce killing of EGFR+ solid tumor cell lines, independently of EGFR mutational status; toxicology testing in cynomolgus monkeys revealed a favorable safety profile. An ongoing phase 1/2a study (NCT04259450) is evaluating the safety, efficacy, immunogenicity, pharmacokinetic (PK) and pharmacodynamic responses of AFM24 in patients with locally advanced or metastatic, treatment refractory solid tumors that are known to express EGFR. AFM24 is administered intravenously Q1W until disease progression, intolerable toxicity, investigator’s discretion or patient withdrawal. As of 29 Oct 2021, 29 patients (median [range] age 58 [29-81] years; number of prior therapies 4 [2-8]), predominantly with colorectal- (CRC, 16/29; 10 KRAS mutant) and non-small cell lung cancer (NSCLC, 7/29; 6 EGFR mutant), were treated with AFM24 across six dose levels (14-480 mg flat). The median number of AFM24 doses administered was 8, range 1-29. The most frequently reported (≥20% of patients) AFM24-related treatment-emergent adverse events (TEAEs) were infusion-related reactions (IRR, 20/29), nausea (7/29) and headache (6/29). There were no on-study deaths; two patients had serious TEAEs (one Grade 3 IRR and one Grade 2 hypoxia) and five patients had transient and reversible Grade 3-4 TEAEs (two Grade 3 IRRs, one Grade 3 hypertension, three ≥Grade 3 lymphocytopenia) attributed to AFM24, respectively. There was one dose-limiting toxicity at 40 mg (Grade 3 IRR). Best objective response was stable disease in 8/24 response-evaluable patients; three patients had stable disease for ≥4 months (two CRC, one NSCLC). Dose-proportional PK between 320 and 480 mg indicated saturation of target-mediated elimination. CD16A receptor occupancy (CD16ARO) on circulating blood cells was correlated with exposure and seemed to level off at 480 mg. Estimated intra-tumoral AFM24 concentrations and CD16ARO were in the range associated with maximum tumor cell killing in vitro. TNF-α, IFN-γ and IL-10 increased over time; other cytokines including IL-6 showed only transient increases. AFM24 has a well-managed safety profile and shows pharmacodynamic activity at doses of 320-480 mg. In parallel to continued dose escalation, expansion in disease specific cohorts has been launched at 480 mg. Other studies are evaluating AFM24 in combination with atezolizumab, and in combination with autologous NK cells holding the potential to activate the innate immune response to fight EGFR+ cancer.

Citation Format: Anthony El-Khoueiry, Juanita Lopez, Omar Saavedra, Mark Awad, Jacob Thomas, Crescens Tiu, Elena Garralda, Bettina Rehbein, Gabriele Hintzen, Kerstin Pietzko, Christa Raab, Erich Rajkovic, Paulien Ravenstijn, Michael Emig. A phase 1/2a first-in-human study of AFM24, a CD16A/epidermal growth factor (EGFR) bispecific Innate Cell Engager (ICE®), in patients with locally advanced or metastatic EGFR-expressing solid tumors: Preliminary findings from the dose-escalation phase [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2022; 2022 Apr 8-13. Philadelphia (PA): AACR; Cancer Res 2022;82(12_Suppl):Abstract nr CT149.

Tanaka score predicts surgical intensive care admission following abdominal wall reconstruction

PURPOSE

With the advancement of abdominal wall reconstruction, more complex hernia patients are undergoing repairs that may require a postoperative surgical intensive care unit (SICU) admission. The volume ratio (VR) of the hernia sac to the abdominal cavity is an easily applied method to quantify disease severity and the ensuing physiologic insult. This study aimed to predict postoperative SICU admission using VR and other preoperative variables.

METHODS

A single-center retrospective review was conducted for patients undergoing large abdominal hernias (width ≥ 18 cm) repaired from September 2014 to October 2019. Patient demographics, comorbidities, abdominal surgical history, and VR were analyzed through univariate and multivariable methods to identify predictors of SICU admission within the first two postoperative days. A predictive model was generated and validated.

RESULTS

Of 434 patients meeting inclusion criteria, 127(29%) required a SICU admission within the first two postoperative days. VR was significantly higher in SICU patients (Median 30.6% [IQR 14.4-59.0] vs. 10.6% [IQR 4.35-23.6], P < 0.001). Male sex, history of chronic obstructive pulmonary disease, prior component separation, recurrent incisional hernia, hernia grade 3, and VR showed higher odds of SICU admission. When validated on a testing dataset, these variables showed strong SICU admission predictions, with an area under the curve, sensitivity, and specificity of 0.82, 81.7% and 68.5%, respectively.

CONCLUSIONS

The volume ratio in combination with preoperatively available variables can reliably predict postoperative SICU admission following abdominal wall reconstruction. Anticipating such events preoperatively allows for bed space allocation as well as optimizing postoperative care.

Tofacitinib in the treatment of refractory eczemas - a case series

Janus Kinase inhibitors have been shown to be effective in the treatment of various dermatoses such as alopecia areata, psoriasis, vitiligo and atopic dermatitis. The Janus kinase-signal transducer and activator of transcription (JAK-STAT) pathway is a common junction for the signaling of several dermatologically significant cytokines and the inhibition of this pathway by JAK inhibitors may be broadly useful in the treatment of various other dermatological disorders. A case series is presented of twelve patients (M:6, F:6) with a mean age 47.1 years, presenting with various eczematous dermatoses (histopathology showing spongiotic dermatitis) who had failed to respond to conventional therapy. These patients were prescribed tofacitinib. All 12 patients improved and 10 patients had improvement to clear or almost clear (physician global assessment score 0/1) after 1 month of therapy. No hematological or biochemical abnormalities were noted during the treatment period. Infections were the most common adverse events observed. Conventional treatment options like systemic steroids and steroid sparing agents are the cornerstone in the treatment of chronic eczematous disorders, however, in patients who do not respond to conventional agents, tofacitinib may represent a viable treatment option.

Heterogenous susceptibility to R-pyocins in populations of Pseudomonas aeruginosa sourced from cystic fibrosis lungs

Bacteriocins are proteinaceous antimicrobials produced by bacteria which are active against other strains of the same species. R-type pyocins are phage tail-like bacteriocins produced by Pseudomonas aeruginosa. Due to their anti-pseudomonal activity, R-pyocins have potential as therapeutics in infection. P. aeruginosa is a Gram-negative opportunistic pathogen and is particularly problematic for individuals with cystic fibrosis (CF). P. aeruginosa from CF lung infections develop increasing resistance to antibiotics, making new treatment approaches essential. P. aeruginosa populations become phenotypically and genotypically diverse during infection, however, little is known of the efficacy of R-pyocins against heterogeneous populations. R-pyocins vary by subtype (R1-R5), distinguished by binding to different residues on the lipopolysaccharide (LPS). Each type varies in killing spectrum, and each strain produces only one R-type. To evaluate the prevalence of different R-types, we screened P. aeruginosastrains from the International Pseudomonas Consortium Database (IPCD) and from our biobank of CF strains. We found that (i) R1-types were the most prevalent R-type among strains from respiratory sources; (ii) there are a large number of strains lacking R-pyocin genes, and (iii) isolates collected from the same patient have the same R-type. We then assessed the impact of diversity on R-pyocin susceptibility and found a heterogenous response to R-pyocins within populations, likely due to differences in the LPS core. Our work reveals that heterogeneous populations of microbes exhibit variable susceptibility to R-pyocins and highlights that there is likely heterogeneity in response to other types of LPS-binding antimicrobials, including phage.

Comparison of Azure Kinect and optical retroreflective motion capture for kinematic and spatiotemporal evaluation of the sit-to-stand test

BACKGROUND

The sit-to-stand test (STS) is commonly used to evaluate functional capabilities within a variety of clinical populations. Traditionally STS is a timed test, limiting the depth of information which can be gained from its evaluation. The Azure Kinect has the potential to add in-depth analysis to STS. Despite these potential benefits, the recently released (2019) Azure Kinect has yet to be evaluated for its ability to accurately assess STS.

RESEARCH QUESTIONS

Purposes of this work were to compare data captured during STS using both a 12 camera Vicon motion capture system and the Azure Kinect; and to calculate kinematic and spatiotemporal variables related to the four phases of the STS cycle.

METHODS

Spatiotemporal and kinematic measures for STS were simultaneously collected by both devices for 15 participants. Cycle waveforms were compared for right and left hip and knee flexion/extension angular displacement, right and left hip and knee flexion/extension angular velocity, and knee-to-ankle separation ratio. Evaluated discrete outcome variables included: phase time points, maximum knee extension velocity from phases 3 to 4, medial-lateral pelvic sway range, and total time to completion. Waveform summary data were compared using R, R², and RMSE. Discrete variables were analyzed using Spearman's Rank correlation coefficient.

RESULTS

R and R² values between the two systems indicated high levels of correlation (all R values > 0.711, all R² values > 0.660). Although there was an overall high level of agreement between waveform shapes, high RMSE values indicated some minor tracking errors for Kinect within the STS cycle. Spearman's Rank correlation coefficient indicated high levels of correlation between the systems for discrete variables (all R values > 0.89), with the exception of medial-lateral pelvic sway range.

SIGNIFICANCE

The Azure Kinect provides valuable insight into STS movement strategies allowing for improved precision in clinical decision making across multiple clinical populations.

Inpatient Pediatric Palliative Care Consult Requests and Recommendations

INTRODUCTION

Little is known about the requests for and recommendations from inpatient pediatric palliative care (PPC) consults and whether they differ by patient location (ward vs. intensive care unit) or patient type (new vs. established with PPC).

METHODS

Single-center, retrospective cohort study comparing PPC consult requests and recommendations for children who received a PPC consult between January 1, 2018 and June 30, 2019. Comparisons were made by patient location and patient type using bivariate statistics.

RESULTS

Three hundred twenty-seven PPC encounters were evaluated. Symptoms were more likely to be addressed in consults for ward patients than for intensive care unit patients. Compared with established patients, consults for new patients occurred significantly later in the hospital course and were more likely to be for psychosocial support or goals of care.

DISCUSSION

We found variability in PPC consult requests and recommendations that may inform future work and targeted education for primary providers.