The Reporting Quality of Machine Learning Studies on Pediatric Diabetes Mellitus: Systematic Review

BACKGROUND

Diabetes mellitus (DM) is a major health concern among children with the widespread adoption of advanced technologies. However, concerns are growing about the transparency, replicability, biasedness, and overall validity of artificial intelligence studies in medicine.

OBJECTIVE

We aimed to systematically review the reporting quality of machine learning (ML) studies of pediatric DM using the Minimum Information About Clinical Artificial Intelligence Modelling (MI-CLAIM) checklist, a general reporting guideline for medical artificial intelligence studies.

METHODS

We searched the PubMed and Web of Science databases from 2016 to 2020. Studies were included if the use of ML was reported in children with DM aged 2 to 18 years, including studies on complications, screening studies, and in silico samples. In studies following the ML workflow of training, validation, and testing of results, reporting quality was assessed via MI-CLAIM by consensus judgments of independent reviewer pairs. Positive answers to the 17 binary items regarding sufficient reporting were qualitatively summarized and counted as a proxy measure of reporting quality. The synthesis of results included testing the association of reporting quality with publication and data type, participants (human or in silico), research goals, level of code sharing, and the scientific field of publication (medical or engineering), as well as with expert judgments of clinical impact and reproducibility.

RESULTS

After screening 1043 records, 28 studies were included. The sample size of the training cohort ranged from 5 to 561. Six studies featured only in silico patients. The reporting quality was low, with great variation among the 21 studies assessed using MI-CLAIM. The number of items with sufficient reporting ranged from 4 to 12 (mean 7.43, SD 2.62). The items on research questions and data characterization were reported adequately most often, whereas items on patient characteristics and model examination were reported adequately least often. The representativeness of the training and test cohorts to real-world settings and the adequacy of model performance evaluation were the most difficult to judge. Reporting quality improved over time (r=0.50; P=.02); it was higher than average in prognostic biomarker and risk factor studies (P=.04) and lower in noninvasive hypoglycemia detection studies (P=.006), higher in studies published in medical versus engineering journals (P=.004), and higher in studies sharing any code of the ML pipeline versus not sharing (P=.003). The association between expert judgments and MI-CLAIM ratings was not significant.

CONCLUSIONS

The reporting quality of ML studies in the pediatric population with DM was generally low. Important details for clinicians, such as patient characteristics; comparison with the state-of-the-art solution; and model examination for valid, unbiased, and robust results, were often the weak points of reporting. To assess their clinical utility, the reporting standards of ML studies must evolve, and algorithms for this challenging population must become more transparent and replicable.

A review on modeling tumor dynamics and agent reward functions in reinforcement learning based therapy optimization

Revolutionary changes of deep reinforcement learning are leading to high-performing intelligent solutions in multiple fields, including healthcare. At the moment, chemotherapy and radiotherapy are common types of treatments for cancer, however, both therapies are usually radical procedures with undesirable side effects. There is an increasing number of evidence that patient-based optimal schedule has a significant impact in increasing efficiency and survival, and reducing side effects during both therapies. To apply artificial intelligence in therapy optimization, an adequate model of tumor growth incorporating the effect of the treatment is mandatory. A method on training a controller for dosage and scheduling, reinforcement learning can be applied, where a well-chosen agent rewarding function is key to achieve optimal behavior. In this survey paper, some selected tumor growth models, reinforcement learning based solutions and especially agent reward functions are reviewed and compared, providing a summary on state of the art approaches.

Deep Metric Learning Using Negative Sampling Probability Annealing

Multiple studies have concluded that the selection of input samples is key for deep metric learning. For triplet networks, the selection of the anchor, positive, and negative pairs is referred to as triplet mining. The selection of the negatives is considered the be the most complicated task, due to a large number of possibilities. The goal is to select a negative that results in a positive triplet loss; however, there are multiple approaches for this-semi-hard negative mining or hardest mining are well-known in addition to random selection. Since its introduction, semi-hard mining was proven to outperform other negative mining techniques; however, in recent years, the selection of the so-called hardest negative has shown promising results in different experiments. This paper introduces a novel negative sampling solution based on dynamic policy switching, referred to as negative sampling probability annealing, which aims to exploit the positives of all approaches. Results are validated on an experimental synthetic dataset using cluster-analysis methods; finally, the discriminative abilities of trained models are measured on real-life data.

Switching reciprocity on and off in a magneto-optical x-ray scattering experiment using nuclear resonance of α-(57)Fe foils

Reciprocity is when the scattering amplitude of wave propagation satisfies a symmetry property, connecting a scattering process with an appropriate reversed one. We report on an experiment using nuclear resonance scattering of synchrotron radiation, which demonstrates that magneto-optical materials do not necessarily violate reciprocity. The setting enables us to switch easily between reciprocity and its violation. In the latter case, the exhibited reciprocity violation is orders of magnitude larger than achieved by previous wave scattering experiments.

[Late post-tracheostomy airway complications in patients with central nervous system injuries]

At decannulation or often weeks later some of the patients undergone long-term mechanical ventilation through tracheostoma reveal symptoms of airway stenosis. Posttracheostomy airway complications detected during the rehabilitation of 34 patients with central nervous system injury are presented. Endoscopic examinations (total 130) were performed with Olympus BF B3 and OES 20-type fiberscopes and Friedel-type rigid bronchoscopes under local or general anaesthesia. Stridor (20 cases) and possible decannulation (10 cases) were the main indications of the first endoscopic examinations. Stridor was caused by tracheal (12 cases) and laryngeal (8 cases) stenosis. With six patients decannulation was possible after the first endoscopy. Removal of granulomas and sutures was successful in 8 cases. Rigid bronchoscopic dilatations were temporarily successful in 7 cases, but in five of them it had to be repeated on emergency. Further therapeutic interventions were: recannulation (3 cases), retracheostomy (4 cases), implantation of silicon T-endoprothesis (5 cases), tracheal resection (4 cases). Three patients are discussed in details. Eliminating airway complications largely contributed to rehabilitation of the nervous and musculoskeletal system, improved the quality of life, and in acute cases it was a life-saving measure. In order to prevent the development of late airway complications or detect them in time, it is recommended to have an endoscopic protocol following every tracheostomy: 1. at the time of decannulation 2. at closing the trachea and 3. 2-3 months after decannulation.