The superficial resistance (from track to track on one polyimide side) may change a lot through the soda action. Washing well with deionized water after the alkali can improve.
Na+ is one of the worst ions for semiconductor reliability - in case you have chips on the printed circuit, be they bare or plastic encapsulated. A heavier alkali, or if possible Ca++, would then be preferable from this aspect.
Could the polyimide degas during the hot pressing? PI is known to release >1% mass vapour under vacuum. Baking it before putting the epoxy, like 1h +125°C, would improve then.