Perspectives on Machine Learning-assisted Plasma Medicine Towards Automated Plasma Treatment
Abstract : Cold atmospheric plasmas (CAPs) have shown great promise for medical applications through their synergistic chemical, electrical, and thermal effects, which can induce therapeutic outcomes. However, safe and reproducible plasma treatment of complex biological surfaces poses a major hurdle to the widespread adoption of CAPs for medical applications. Predictive modeling of the mutual interactions between the plasma and biological surfaces and, thus, systematic approaches to quantify and predict plasma treatment outcomes remain largely elusive due to the lack of mechanistic understanding of plasma-surface interactions that can span across vastly different length-and time-scales. In addition, real-time sensing capabilities in biomedical CAP devices are often limited, which can be detrimental to plasma treatment due to the intrinsic plasma and surface variability during the treatment, as well as sensitivity to external perturbations.
All of these challenges can make reproducible and effective plasma treatment of biological surfaces difficult to realize, which is further compounded by errors due to human operation of hand-held CAP devices. Machine learning and data-driven approaches can be particularly useful in addressing these challenges in three major ways: (i) data-driven modeling of hard-to-model plasma-surface interactions and plasma treatment outcomes; (ii) learning data analytics for plasma and surface diagnostics in real-time; and (iii) developing predictive controllers that enable reliable and effective CAP treatments. This paper discusses the promise of machine learning to accelerate plasma medicine research in these areas, toward machine learning-assisted and automated CAP treatment of complex biological surfaces.
? The reason for that is not really clear. One reason could be that in most cases dental problems or diseases are not that wearing or life-threatening as non-healing chronic wounds or cancers are.
? Therefore, the risk-benefit-balance may be assessed in dentistry with different emphasis compared to other medical fields.
? However, this kind of reservation should be removed at large because it was specifically proven for oral mucosa, too, that no long-term side effects caused by CAP treatment exist (Jablonowski et al., 2019).
? Moreover, for most of the problems in oral medicine more or less satisfactory treatment options are existing.
? ML lends itself well to handling large amounts of data with a plethora of different measurable properties/characteristics (i.e., features), including heterogeneous and unstructured datasets. While ML methods can be employed across different fields, it is generally advantageous to tailor ML algorithms to field-specific data analysis problems.
? Although considering each sub-aim (i.e., prediction of treatment outcomes, real-time diagnostics, and learning-based control) independently is critical to examining the impact of ML on different components of the problem, future work must also direct its efforts to combining the independently-developed tools in a comprehensive automated strategy.
As physical evaluation criteria, plasma/gas temperature, thermal output, optical emission spectra and irradiance measurements in the range between 200 and 900 nm, current flows (patient leakage current), and gas emission are proposed.
Biological evaluation criteria comprise the in vitro determination of inactivation of specified microorganisms as well as viability tests of eukaryotic cell cultures in vitro.
The detection of chemical species generated by CAP treatment of aqueous liquid is additionally proposed to roughly evaluate the composition and extent of ROS and RNS generated by the plasma device.
? The main advantage of IL over RL is that there is no need for explicit programming or designing reward functions specific to the task, which makes it particularly appropriate to addressing potentially conflicting and hard-tomodel control goals, since the entire learning process can be reduced to providing demonstrations
? Quantitative and real-time diagnosis of the plasma and plasma-surface interactions is a key challenge in plasma medicine.
? To this end, multivariate statistical methods have been widely employed for plasma diagnostics due to the several advantages that they offer over moreconventional methods.
|