For the purposes of prioritization within risk assessments that focus on specific media (air, water, soil) should physical-chemical properties (e.g., Henry’s law constant for air, water solubility for water, adsorption coefficient for soil/sediment) be used to screen out negligible contributors to total exposure? If so, what threshold values should be considered as potential cut-offs?

Yes, but I honestly don't have specific cutoffs to recommend.

All have to be taken into account. They are all relevant but priorization needs to be undertaken and communicated to proceed. Cannot exclude upfront since we do not know beforehand what is important or relevant. Approach must be open to all possible results.
If data are missing, then these should be filled but with justification to be transparent – and have the possibility to reassess.
It shall be noted that most of the constants are derived from models and vary by orders of magnitude from the experiment. Further, they correlate with the exposure medium or matrix (temperature, pH, pressure, presence of particles/solvents/ions). Scenarios must be realistic.

Such an approach could be used to deprioritize certain subgroups from consideration in detailed risk assessment. For example, fluorinated substances that are not emitted to water and have high Henry's law constants are unlikely to be present in water. They could therefore be depriortized for consideration in aquatic risk asessments. I don't want to propose specific cutoff values because this could lead to oversimplifications. A reasoned scientific argument that is peer-reviewed would be my suggestion.

But even then surprises may arise given the unusual behaviour of PFAS. Such highly persisitent substances will accumulate and effects that are currently unknown may be discovered in the future. Therefore, such an approach should only be used for prioritization/deprioritization for risk assessments and not as a basis for exclusion of all chemicals management actions.

Unable to provide an answer to this question.

I don't know the answer but it would not be hard to find out. The process of selecting cutoff values of physical-chemical properties is relatively straightforward.

A risk assessment scenario (the set of algebraic equations that characterize dose and risk to an individual) is created for the exposures an individual receives from a specific pathway. The scenario establishes the toxicity of the PFAS, the duration and intensity of the individual’s exposure, doses from other sources of exposure, and acceptable levels of risk. The exposure assessment portion of the scenario establishes the relationship of the physical-chemical property of interest to the predicted dose. All inputs to the scenario (other than the physical-chemical parameter in question) that are uncertain, vary across individuals or across chemicals are identified. The values of these inputs are set using reasonable worst-case assumptions that reflect the population of interest and the chemicals under consideration.

Once this is done the value of the physical-chemical property that corresponds to the acceptable level is solved for algebraically.

Note it may be necessary to link values of certain physical-chemical properties if they are correlated. For example when solving for a safe value of molecular weight the impacts of molecular weight on vapor pressure, water solubility, and ability to cross a membrane should be accounted for in the scenario.

I do not think so. We have to focus on the big issues and big picture at this stage, rather than draining our energy and resources to negligible contributors of exposure. Once we have established a platform for risk assessment, further refinements in future can consider those negligible contributors. I would say if a source/medium contributes an exposure of >10% of the total exposure or at the reference dose, then we should consider that pathway for risk assessment.

Given the ampiphilic nature of PFAS, these types of approaches may not lead to usable data for such prioritization approaches. Further, as more data are collected on understudied PFAS, underlying assumptions may be challenged. For example, PFMOAA, a PFAS with an oxygen ether linkage and only three carbons (sometimes termed a "short-chain PFAS") should have a very short biological half-life in humans, an assumption about short-chain PFAS. However, it was recently detected in blood of people living near a fluoromanufacturing facility in China (Yao et al., 2020), which suggests that its half-life in humans may not be "short."

The approach described in the question is a reasonable way to prioritize based on potential for exposure, but does not account for toxicity. For example, a minor contributor to total PFAS exposure may be a major contributor to toxicity. Selection of potential cutoff is not a scientific decision per se, it is a risk management decision. In risk assessments that I'm familiar with, values of 90, 95 and 99% have been assumed.

Unless we know all potential degradation byproducts/metabolites of each PFAS (which we won't in the foreseeable future), screening out PFAS may not be appropriate.

Insufficient expertise on my part to offer an informed opinion.

Other than if there is no exposure,there is no risk.

Physical-chemical properties should not be considered by themselves. It is doubtful that a single cut-point identified may result in the same results if applied to different mixtures of PFAS chemicals.

Within the context of mixtures risk assessment, prioritization should be based on the risk-based contribution of a mixture component to total risk. Exposure is but of the two determinants of risk; this issue focuses solely on exposure, which is a complicating factor, as discussed below. This is a complicated question, in that risk is based on dose absorbed from the medium and distributed to the internal organs. Within an environmental medium, exposure in terms of applied dose will be uniform, balanced among components by their concentration in the environmental medium. The identified physical-chemical factors seem to relate more to movement of a PFAS chemical within or among environmental matrices, rather than presenting a challenge to estimating differential exposure (absorption) from an environmental matrix.

However, considering the impact of physical-chemical properties on the exposure alone (here, assumed to be represented by physical-chemical properties) may be accomplished, but will be accompanied by much uncertainty. Nonetheless, this approach may be further considered.

If it to be assumed that screening out those components that contribute minimally to total PFAS exposure is to be undertaken, a scheme in which the total exposure is characterized (determined, estimated, etc) on the basis of both molecular mass and on the basis of the total number of moles comprising the mixture exposure. Next, some fraction (e.g., 80%) of the total PFAS mixture to be accounted-for should be identified. Because of the divergent differences in molecular weight, consideration should be given whether to account for percent contribution on a mass or molar basis. Because some mixtures work by JE Simmons (US EPA) and colleagues has indicated for drinking water disinfection byproducts that an evaluation based on a molar measure of exposure rather than a mass-based exposure provides a better estimation, care should be taken in this decision. Once a decision has been reached as to what fraction of the total PFAS mixture should be accounted-for, an analysis of the molar contribution to this fraction for each component should be determined.

Inasmuch as different PFAS mixtures exposures may contain components with substantially divergent physical-chemical and toxicological properties, it may be presumed that the exclusion cut-point for a given physical-chemical property will vary among mixtures.

Because water solubility impacts a drinking water-based exposure and because an adult human consumes approximately 2 liters per day, a benchmark drinking water concentration (perhaps in a potency-adjusted molar concentration; toxicity equivalents) may be developed on the basis of a toxicity value (e.g., oral RfD value) established for a PFAS chemical thought to exert a toxic potency representative of other chemical with which it may be group. Such an approach to de-select PFAS chemical in a mixture contained in an environmental medium may best be completed on a case-by-case basis. It is doubtful that a pre-selected cut-point for a physical chemical property will be accompanied with a level of confidence to justify its application tor PFAS mixtures comprising different components, at different concentrations.