Censored-Data-Aware Screening of Hazardous Metal(loid)s in Wild Mushrooms from Urban and Peri-Urban Green Spaces: Multielement Fingerprints and Adult Exposure Assessment

Wild mushrooms collected from urban and peri-urban green spaces may represent environmentally exposed biological matrices for hazardous metal(loid)s, but their interpretation is complicated by taxonomic heterogeneity, tissue partitioning and left-censored analytical data. This study evaluated As, Be, Cd, Co, Cr, Cu, Mn, Ni, Pb, V and Zn in wild mushrooms from Leicester/Leicestershire, UK, using a reconstructed fruiting-body-level dataset designed to avoid double counting paired cap/stem records. The central question was whether taxon- and tissue-aware multielement profiling alters screening-level exposure interpretation compared with a pooled edible-relevant summary. In total, 121 fruiting-body-level units were evaluated on a dry-weight basis. Cu, Mn and Zn were detected in all units, while Cd and Pb were detected in 117/121 and 106/121 units, respectively. V, Cr and Ni were detected in 84/121, 70/121 and 64/121 units, whereas As, Co and Be showed substantial censoring. Be was not interpreted distributionally. Central concentrations selected under the censored-data framework were 1.856 mg kg−1 for Cd, 71.142 mg kg−1 for Cu, 14.318 mg kg−1 for Mn, 3.874 mg kg−1 for Ni, 2.171 mg kg−1 for Pb, 0.104 mg kg−1 for V and 97.506 mg kg−1 for Zn. Taxon-level profiles showed marked heterogeneity, and paired tissue analysis in Agaricus bitorquis showed higher cap than stem concentrations for Cd, Cu, Pb and Zn. The primary PCA was restricted to well-detected elements (Cd, Cu, Mn, Pb, V and Zn), explaining 73.92% of total variance across the first three components. Sensitivity analyses showed that the overall PCA configuration remained stable across alternative treatments of censored observations and after exclusion of flagged extreme records, although some loading magnitudes showed moderate variation. Under the 300 g fresh-weight week−1 scenario, Cd represented 47.3% of the EFSA tolerable weekly intake of 2.5 µg kg−1 bw week−1 in the pooled edible-relevant population (N = 43) and 50.9% in Agaricus bitorquis; however, taxon-specific screening changed both the magnitude and ranking of the principal threshold-based drivers, with Cu becoming the leading driver for Coprinopsis atramentaria and Marasmius oreades. None of the primary or exploratory park-level soil–mushroom associations remained statistically significant after false-discovery-rate correction in the harmonised eight-park analysis. Overall, the findings partially support the working hypothesis: taxon- and tissue-aware profiling improved screening-level interpretation, whereas park-level soil context did not provide robust evidence of prediction or transfer. The fruiting-body-level, censored-data-aware framework provides a transparent basis for exposure screening, although uncertainties related to speciation, bioaccessibility, moisture assumptions and analytical interferences remain.