Upon contact with a biomaterial, cells and surrounding tissues respond in a manner dictated by the physicochemical and mechanical properties of the material. Traditionally, cellular responses are monitored using invasive analytical methods that report the expression of genes or proteins. These analytical methods involve assessing commonly used markers for a predefined readout, masking the actual situation induced in the cells. Hence, a broader expression profile of the cellular response should be envisioned, which technically limits up scaling to higher throughput systems. However, it is increasingly recognized that morphometric readouts, obtained non-invasively, are related to gene expression patterns. Here, we introduced distinct surface roughness to three PLA surfaces, by exposure to oxygen plasma of different duration times. The response of mesenchymal stromal cells was compared to smooth untreated PLA surfaces without the addition of differentiation agents. Morphological and genome wide expression profiles revealed underlying cellular changes which was hidden for the commonly used gene markers for osteo-, chondro- and adipogenesis. Using 3 morphometric parameters, obtained by high content imaging, we were able to build a classifier and discriminate between oxygen plasma-induced modified sheets and non-modified PLA sheets where evaluating classical candidates missed this effect. This approach shows the feasibility to use noninvasive morphometric data in high-throughput systems to screen biomaterial surfaces indicating the underlying genetic biomaterial-induced changes.