HUMAnN for Low Biomass Scenario

I have a cancer tissue data set but I have only seen HUMAnN used for typical data sets such as gut microbiome. Almost all samples in my data set have zero pathways detected and a couple of samples have one or two pathways. The typical number of species per sample detected by MetaPhlAn is about thirty. Is there any low-biomass analysis for HUMAnN which allows for portions of the pathway to be missing in the data but for the pathway to still be inferred and quantitated nonetheless. I can’t find any journal article about HUMAnN being used in a low-biomass scenario, such as human cancer tissue.