Biography

Stefan Sandström is CEO of Biosector, a Japan‑based commercialisation specialist for life sciences. For nearly two decades in Japan, he has helped non‑Japanese companies establish and scale with Japan’s pharma/biopharma industries. He has facilitated business with 32 of Japan’s largest pharmaceutical companies. His focus is on “what is next for the industry” and applied, auditable AI that shortens cycle time and improves right‑first‑time. A frequent speaker at CGT and bioprocess meetings, he blends training in chemical engineering and medicine with hands‑on go‑to‑market execution to make advanced therapies manufacturable and accessible. Originally from Sweden, he lives in the Tokyo–Yokohama area and works across Asia, Europe, and North America to help companies convert scientific breakthroughs into dependable patient access.

Abstract

CGT will not fulfil their social promise until it runs like an industry, not an experiment. The most ethical act we can perform as a field is a dose delivered on time, right‑first‑time. This keynote sets out a practical blueprint for scaling reliability and fairness into CGT operations without straying into pricing or regulation.

I will frame ethics through operational efficacy: cycle time, right-first-time, centre throughput, and chain-of-identity/chain-of-custody integrity. Using field‑tested playbooks, I will outline how to standardise what should be identical across sites (data models, SOP packs, closed/automated steps, tech‑transfer kits) and how to localise intelligently (people, site readiness, and centre workflows). We will examine hybrid networks that combine centralized facilities with point‑of‑care capability, and the objective thresholds at which bedside becomes the ethical and operational choice.

Applied AI is treated as a GMP‑grade asset, not a slide deck. I will show minimal, auditable AI use-cases that measurably shorten cycle time and reduce failure rates: capacity and slot scheduling for autologous flows, real-time deviation triage, and in-process signals that inform faster, safer decisions. The emphasis is on trust‑by‑design: clear data lineage, monitored model performance, and change control. This means that AI changes the schedule or makes a decision only when it is demonstrably safe to do so.

The goal is simple: make reliability routine. When CGT becomes reliable and repeatable, access grows, costs per treated patient trend down, and the field earns the trust implied by the conference theme: Ethics and Efficacy.