What is the most effective way to backtest ICT entry models on lower timeframes?

Backtesting on 1-minute or 5-minute charts requires tracking real-time session parameters, spreads, and market structure context. How do you structure your backtesting data to get accurate results?