From 'No Treatment Available' to 'No Evidence of Disease': A Review of GitLab Co-founder’s AI-Assisted Cancer Journey

According to monitoring by 1M AI News, Sid Sijbrandij, co-founder and executive chairman of GitLab, recently shared his complete experience of using AI to combat osteosarcoma at the OpenAI Forum event, alongside geneticist Jacob Stern. Sid was diagnosed with spinal osteosarcoma at the end of 2022, with a tumor measuring 6 centimeters in diameter. After undergoing surgery, spinal fusion, radiation therapy, and high-intensity chemotherapy in 2023, the cancer recurred in 2024, with standard treatment options largely exhausted and his primary physician unable to recommend effective next steps. Sid subsequently stepped down from his daily responsibilities to form a personal medical team, led by former 10x Genomics executive Jacob Stern, taking full control of his treatment decisions in a ‘founder mode.’ The team adopted an ‘extreme diagnostics’ strategy, conducting comprehensive analyses of the tumor through single-cell sequencing, DNA/RNA sequencing, targeted imaging, and organoid testing, generating a total of 25TB of data. Single-cell sequencing revealed high expression of the FAP protein in the tumor, leading the team to find an experimental radioactive ligand therapy targeting FAP in Germany. After two treatments, the tumor necrosis reached 60%, reduced in size by 20%, and detached from the spinal dura, allowing surgeons to operate again and remove it. AI acted as an accelerator throughout the process. Jacob directly input the tumor RNA sequencing data into GPT-4 for analysis, where AI identified the subsequently crucial B7H3 target and characterized the dynamic features of the tumor immune microenvironment. The team also developed an Agent system that automatically conducts literature searches, hypothesis generation, and bioinformatics analysis upon inputting natural language questions, costing about $20 per run and taking 30 minutes to analyze and generate reports on 600,000 single-cell data points. AI was also used to assist in screening candidate neoantigens when designing personalized mRNA tumor vaccines. Jacob noted that while AI cannot replace specialists, it empowers him to be a qualified interlocutor with specialists, allowing him to quickly engage and advance collaboration in areas where he is not an expert. Sid stated that after targeted radioactive treatment and surgery, he is currently in a ‘no evidence of disease’ state, with available treatment options expanded from zero to thirty. The team is working to commercialize this approach, having incubated companies such as Thalus (gene expression analysis and target discovery) and Arden (personalized treatment for complex immune diseases), aiming to lower the barriers to personalized treatment.