Next-generation wireless systems must significantly improve performance to enable all envisioned 6G applications, which require an increase in end-user data rate by 100, a thousand-fold multiplication of total data volume and sub-ms latency. The largely untouched THz range ([0.1-10] THz) is a prime candidate to unlock these features. Yet, to obtain THz phased arrays, groundbreaking antenna systems based on fundamentally novel topologies are key, as conventional arrays scaled beyond 100 GHz suffer from large metal losses and sizes. In this project, a new, groundbreaking methodology for designing fully dielectric continuous transverse stub (CTS) arrays while completely avoiding metals is conceived, fed by quasi-optical lenses. A unique CTS topology fulfilling stringent requirements in terms of broad bandwidth, high efficiency and large field of view will be explored to meet the needs of 6G applications. Lenses also allow for a highly compact design capable of forming narrow beams over large scanning angles and wide bandwidth. Finally, by omitting metal, this novel methodology allows to save energy and to design more durable high-performance components by drastically reducing the metal losses, covering the gap between typical millimeter-wave antenna and photonic systems, and direct integration of THz ICs. Eventually, the proposed innovations will lead to low-profile THz systems, providing low-loss high-gain broadband performance and allowing highly efficient compact integration.