i find your interpretation about unscear's assertions subjective. unscear as an organization have rejected lnt. it was resolved in assembly about a decade ago

i find your interpretation about unscear's assertions subjective. unscear as an organization have rejected lnt. it was resolved in assembly about a decade ago that it's just plain unreliable for extrapolating to low and even medium doses. if you look in reports you'll find that conclusions are never based on lnt, and when sources using lnt are incorporated they're always reevaluated against other models and often specifically discredited as overestimating low dose health effects, like this little bit from the 2019 report: 472. Due to the limited statistical power, it is generally difficult to detect a departure from linearity in the dose–response relationship for circulatory diseases. As an alternative to the standard approach based on a single preferred model, Schöllnberger et al. [S2] applied a multi-model inference approach to the dose–response analysis for mortality from cardiovascular disease in the LSS cohort. A combined set of plausible models fitted equally well. With this combined set of models, they found a dose response that was about one third weaker than that with the LNT model at doses below 0.6 Gy. The dose response at higher doses was stronger. More recently, the multi-model inference approach was applied to mortality in the LSS cohort from cerebrovascular and heart disease [S3]. For cerebrovascular disease, the dose–response curve estimated by multi-model inference was below the LNT model at low to medium doses (0–1.4 Gy) while at higher doses, a higher risk was estimated compared to the LNT fit. Similarly, a sublinear dose response was found for heart disease at doses of 0 to 3 Gy. The estimated confidence bounds indicated no conclusive answer for an increased risk below 0.75 Gy for cerebrovascular disease and 2.6 Gy for heart disease.