ITER-like plasmas with edge-localised pellet fuelling need to be modelled in an integrated approach combining transport codes for the plasma core and Scrape-Off Layer (SOL) with a pellet code based on first principles assumptions. This would allow to properly account for strong interactions between the pellet injection and particle deposition behaviour and edge and core transport properties, which together determine the fuelling efficiency and affect the overall plasma performance. At JET, first integrated predictive modelling has been performed for recent pellet experiments with the new High Frequency Pellet Injector (HFPI). With the help of synthetic diagnostics, i.e. tools that convert the simulation data into diagnostic measures, it was possible to draw direct comparisons with measurements from new high resolution pellet diagnostic systems available at JET and to compensate for reduced measurement capabilities in conditions with low signal-to-noise ratio, enabling the study of the E¥B drift of ablated pellet particles, the pellet retention time, the pellet penetration length required for ELM triggering and the interaction between pellet-triggered ELMs and pellet fuelling.