Artificial intelligence and machine learning algorithms for early detection of skin cancer in community and primary care settings: a systematic review.

Skin cancers occur commonly worldwide. The prognosis and disease burden are highly dependent on the cancer type and disease stage at diagnosis. We systematically reviewed studies on artificial intelligence and machine learning (AI/ML) algorithms that aim to facilitate the early diagnosis of skin cancers, focusing on their application in primary and community care settings. We searched MEDLINE, Embase, Scopus, and Web of Science (from Jan 1, 2000, to Aug 9, 2021) for all studies providing evidence on applying AI/ML algorithms to the early diagnosis of skin cancer, including all study designs and languages. The primary outcome was diagnostic accuracy of the algorithms for skin cancers. The secondary outcomes included an overview of AI/ML methods, evaluation approaches, cost-effectiveness, and acceptability to patients and clinicians. We identified 14 224 studies. Only two studies used data from clinical settings with a low prevalence of skin cancers. We reported data from all 272 studies that could be relevant in primary care. The primary outcomes showed reasonable mean diagnostic accuracy for melanoma (89·5% [range 59·7-100%]), squamous cell carcinoma (85·3% [71·0-97·8%]), and basal cell carcinoma (87·6% [70·0-99·7%]). The secondary outcomes showed a heterogeneity of AI/ML methods and study designs, with high amounts of incomplete reporting (eg, patient demographics and methods of data collection). Few studies used data on populations with a low prevalence of skin cancers to train and test their algorithms; therefore, the widespread adoption into community and primary care practice cannot currently be recommended until efficacy in these populations is shown. We did not identify any health economic, patient, or clinician acceptability data for any of the included studies. We propose a methodological checklist for use in the development of new AI/ML algorithms to detect skin cancer, to facilitate their design, evaluation, and implementation.

Can we reliably automate clinical prognostic modelling? A retrospective cohort study for ICU triage prediction of in-hospital mortality of COVID-19 patients in the Netherlands.

BACKGROUND: Building Machine Learning (ML) models in healthcare may suffer from time-consuming and potentially biased pre-selection of predictors by hand that can result in limited or trivial selection of suitable models. We aimed to assess the predictive performance of automating the process of building ML models (AutoML) in-hospital mortality prediction modelling of triage COVID-19 patients at ICU admission versus expert-based predictor pre-selection followed by logistic regression. METHODS: We conducted an observational study of all COVID-19 patients admitted to Dutch ICUs between February and July 2020. We included 2,690 COVID-19 patients from 70 ICUs participating in the Dutch National Intensive Care Evaluation (NICE) registry. The main outcome measure was in-hospital mortality. We asessed model performance (at admission and after 24h, respectively) of AutoML compared to the more traditional approach of predictor pre-selection and logistic regression. FINDINGS: Predictive performance of the autoML models with variables available at admission shows fair discrimination (average AUROC = 0·75-0·76 (sdev = 0·03), PPV = 0·70-0·76 (sdev = 0·1) at cut-off = 0·3 (the observed mortality rate), and good calibration. This performance is on par with a logistic regression model with selection of patient variables by three experts (average AUROC = 0·78 (sdev = 0·03) and PPV = 0·79 (sdev = 0·2)). Extending the models with variables that are available at 24h after admission resulted in models with higher predictive performance (average AUROC = 0·77-0·79 (sdev = 0·03) and PPV = 0·79-0·80 (sdev = 0·10-0·17)). CONCLUSIONS: AutoML delivers prediction models with fair discriminatory performance, and good calibration and accuracy, which is as good as regression models with expert-based predictor pre-selection. In the context of the restricted availability of data in an ICU quality registry, extending the models with variables that are available at 24h after admission showed small (but significantly) performance increase.

Contribution to unravel variability in bowhead whale songs and better understand its ecological significance.

Since the first studies on bowhead whale singing behaviour, song variations have been consistently reported. However, there has been little discussion regarding variability in bowhead whale singing display and its ecological significance. Unlike the better studied humpback whales, bowhead whales do not appear to share songs at population level, but several studies have reported song sharing within clusters of animals. Over the winter season 2013-2014, in an unstudied wintering ground off Northeast Greenland, 13 song groups sharing similar hierarchical structure and units were identified. Unit types were assessed through multidimensional maps, showing well separated clusters corresponding to manually labelled units, and revealing the presence of unit subtypes. Units presented contrasting levels of variability over their acoustic parameters, suggesting that bowhead whales keep consistency in some units while using a continuum in values of frequency, duration and modulation parameters for other unit types. Those findings emphasise the need to account for variability in song analysis to better understand the behavioural ecology of this endangered species. Additionally, shifting from song toward units or phrase-based analysis, as it has been suggested for humpback whales, offers the opportunity to identify and track similarities in songs over temporal and geographical scales relevant to population monitoring.

Sperm whale long-range echolocation sounds revealed by ANTARES, a deep-sea neutrino telescope.

Despite dedicated research has been carried out to adequately map the distribution of the sperm whale in the Mediterranean Sea, unlike other regions of the world, the species population status is still presently uncertain. The analysis of two years of continuous acoustic data provided by the ANTARES neutrino telescope revealed the year-round presence of sperm whales in the Ligurian Sea, probably associated with the availability of cephalopods in the region. The presence of the Ligurian Sea sperm whales was demonstrated through the real-time analysis of audio data streamed from a cabled-to-shore deep-sea observatory that allowed the hourly tracking of their long-range echolocation behaviour on the Internet. Interestingly, the same acoustic analysis indicated that the occurrence of surface shipping noise would apparently not condition the foraging behaviour of the sperm whale in the area, since shipping noise was almost always present when sperm whales were acoustically detected. The continuous presence of the sperm whale in the region confirms the ecological value of the Ligurian sea and the importance of ANTARES to help monitoring its ecosystems.

Listening to the Deep: live monitoring of ocean noise and cetacean acoustic signals.

The development and broad use of passive acoustic monitoring techniques have the potential to help assessing the large-scale influence of artificial noise on marine organisms and ecosystems. Deep-sea observatories have the potential to play a key role in understanding these recent acoustic changes. LIDO (Listening to the Deep Ocean Environment) is an international project that is allowing the real-time long-term monitoring of marine ambient noise as well as marine mammal sounds at cabled and standalone observatories. Here, we present the overall development of the project and the use of passive acoustic monitoring (PAM) techniques to provide the scientific community with real-time data at large spatial and temporal scales. Special attention is given to the extraction and identification of high frequency cetacean echolocation signals given the relevance of detecting target species, e.g. beaked whales, in mitigation processes, e.g. during military exercises.

Indications for the tracking of elevated nitrogen levels through the fungal route in a soil food web.

The objective of the present study was to determine the effects of elevated N in dead organic matter on the growth of fungi and to establish the consequences for the development of microbivores. Therefore, three fungal species were cultured on Scots pine litter differing in N content. The growth of the soil fungal species Trichoderma koningii, Penicillium glabrum and Cladosporium cladosporioides was directly influenced by the N content (ranging from 1.25 to 2.19% N) of the substrate. For all three fungal species maximum growth was highest at intermediate N content (1.55%) of the substrate. The fungivorous collembolan Orchesella cincta reached highest asymptotic body mass when fed with C. cladosporioides, grown on litter medium with intermediate N content (1.55%). The growth of O. cincta was lower when fed with C. cladosporioides from litter medium with the highest N content (2.19%). Similar results were obtained in mesocosm experiments in which pine litter with three levels of N (1.11, 1.78, 2.03% N) was used as substrate for the fungi. On litter with the highest N content (2.03%) hyphal length and asymptotic body mass of O. cincta were reduced. The results show that the N content of the substrate determines the growth of both fungi and fungivores, and suggest that elevated levels of N in soil track through the fungal part of the soil food web.

Characterization of inhibitory innervation in porcine colonic circular muscle.

Effects of stimulation of intramural nerves in the circular smooth muscle layer of the porcine colon (Sus scrofa domestica) were studied using the sucrose-gap technique. Electrical field stimulation of the preparation, superfused with Krebs solution at 21 degrees C, induced a transient hyperpolarization of the smooth muscle cell membrane. This hyperpolarization was an inhibitory junction potential (IJP). The responses obtained from circular muscle originating from either the centripetal or centrifugal gyri of the ascending colon did not differ significantly. The IJP was characterized as being mediated by intramural, nonadrenergic, noncholinergic (NANC) nerves. The amplitude and latency of the IJP changed linearly with temperature (15-25 degrees C: +1 mV and -0.1 s per degree Celsius, respectively) reflecting a temperature-dependent synchronization of transmitter release. The membrane resistance decreased during the IJP. The IJP amplitude decreased or increased during conditioning hyperpolarizations or depolarizations, respectively, and reversed at membrane potentials about 30 mV more negative than the resting membrane potential. Potassium conductance blocking agents, barium (1 mM), tetraethylammonium chloride (TEA, 20 mM), 4-aminopyridine (4-AP, 5 mM), apamin (1 microM), and aminacrine (10(-4) M) added to the superfusion medium increased the membrane resistance. Only barium, TEA, and apamin depolarized the smooth muscle cell membrane. The IJP amplitude decreased in the presence of aminacrine and apamin to 75 and 35%, respectively, suggesting that apamin-sensitive Ca2+-activated K+ channels are involved in this response. ATP, adenosine, and related adenine nucleotides in concentrations up to 10(-3) M did not mimic the IJP. Superfusion with ATP for 15 min revealed a gradually increasing attenuation by up to 20% of the IJP. This might suggest that the release of neurotransmitter from intramural NANC nerves is modulated presynaptically via purinoceptors. Exogenously applied vasoactive intestinal polypeptide (VIP) in concentrations of 10(-9) to 10(-4) M did not affect the preparation. Also at elevated temperatures (up to 35 degrees C), VIP (10(-7) to 10(-4) M) did not cause measurable effects. It is concluded that the inhibitory mediator of the intramural NANC nerves present in the circular muscle layers of the porcine colon is neither a purine nor VIP.

Histamine H2-binding on guinea pig cerebral cortex.

The Kd-values of some histamine H2-active compounds, obtained from radio-ligand-binding studies on a homogenate of the guinea-pig cerebral cortex with 3H-tiotidine as the labelled H2-ligand, were compared with the pA2/pD2-value of these compounds on the guinea-pig right atrium and guinea-pig isolated gastric fundus. A good correlation was found between the pKd of the H2-antagonists and their pA2 on the guinea-pig right atrium. A much poorer correlation however was obtained between the pKd of the agonists on the cerebral cortex and their pD2-values on the guinea-pig right atrium and the gastric fundus. This poor correlation between true affinity and H2-activity of the agonists might be explained by spare receptors as a much better correlation was obtained between pKd and pD2 of partial agonists.

Differentiation of functional adrenoceptors in human and guinea pig airways.

The nature of the beta-adrenoceptor population(s) mediating direct smooth muscle relaxation, inhibition of antigen-induced histamine release and inhibition of antigen-induced (leukotriene-mediated) smooth muscle contraction of human and guinea pig central and peripheral airways was investigated. Preferential blockade by beta 1- and beta 2-selective antagonists of the relaxation induced by beta 1- and beta 2-selective agonists, respectively, revealed the guinea pig tracheal smooth muscle relaxation to be mediated by both beta 1- and beta 2-adrenoceptors. Using a highly beta 2-selective antagonist, the NE-induced relaxation was split up biphasically into a beta 1- and a beta 2-component. In contrast, no such differential blockade was observed with the relaxation of the guinea pig lung parenchyma strip, neither with the human tracheal, main bronchus and respiratory bronchiolus smooth muscle, which are all mediated by homogeneous beta 2-adrenoceptor populations. Only in the guinea pig trachea did neuronal and extraneuronal uptake inhibitors produce pronounced left shifts of the NE- and ISO-induced relaxation curves, respectively, suggesting a causal relationship between noradrenergic innervation and the presence of the beta 1-adrenoceptor subpopulation in the airways. Using the same techniques, it was established that inhibition of antigen-induced histamine release from guinea pig lung and tracheal mast cells is mediated by homogeneous beta 2-adrenoceptor populations as well. In contrast to catecholamines, non-catecholamine beta-agonists such as fenoterol, clenbuterol and zinterol had a substantially higher apparent affinity for the inhibition of the anaphylactic (leukotriene-mediated) guinea pig tracheal contraction than for the inhibition of histamine release; the same was true for lung tissue, though the difference was less pronounced. With some non-catecholamine beta-agonists considerable selectivity both in central and peripheral airway preparations was observed for the inhibition of anaphylactic contraction as compared with smooth muscle relaxation.

Comparison of functional beta-adrenoceptor heterogeneity in central and peripheral airway smooth muscle of guinea pig and man.

The nature of the beta-adrenoceptor population(s) mediating the relaxation of guinea pig and human airway smooth muscle was investigated. On the basis of a preferential blockade by beta 1- and beta 2-selective antagonists of the relaxation induced by beta 1- and beta 2-selective agonists, guinea pig tracheal strip relaxation was found to be mediated both by beta 1- and beta 2-adrenoceptors, the relative participation of which depending on the relative affinities of the agonist towards these two receptors. With highly selective antagonists the noradrenaline (NA)-induced relaxation could be split up biphasically into a beta 1- and a beta 2-component. In contrast, no such differential blockade was observed with the guinea pig lung parenchyma strip relaxation which is mediated by a homogenous beta 2-adrenoceptor population. On comparison of the tracheal, the spirally cut main bronchus- and intrapulmonary airway smooth muscle strips it could be shown that both the sensitivity of NA for neuronal uptake and the apparent affinity of the relaxation by NA decreased in the direction of the lung periphery. Using the same techniques it was ascertained that the relaxation of human tracheal smooth muscle (autopsy, obtained within 6 hours after death), main bronchus and intrapulmonary smooth muscle (operation) are mediated by homogenous beta 2-adrenoceptor populations. In addition, neuronal and extraneuronal uptake sites were not operative in these preparations, whether obtained from operation or from autopsy.