Even the WoS Spark implementation is in a self custodial grey area. Not only are they serving as a Lightning Service Provider in which you must trust them for channel management and channel balance security, you're also pushing trust onto the security of the Spark sidechain. Usually with self-custodial lightning you have to monitor for malicious peer activity manually or via watchdog nodes but with Spark its much more difficult due to the abstraction. 

Spark is essentially a layer 3 at this point. Which begs the question, if we're already trusting a sidechain AND lightning channels (two huge security leaps), why would one use Spark over WBTC or TBTC on an EVM chain or BTCX on Zano? Most of the global south already has adopted USDT/C on Tron and EVM chains, it'd be easier UX for those users to use wrapped BTC on their existing wallets. 

Also privacy is basically 0 from Spark implementations. A far cry from the great sender privacy of a private LND node behind Tor. 

Overall point here is, we can add in a bunch of convoluted trust points in order to make lightning work, but it's just kicking the can. Lightning is only as secure and scalable as the base chain and since the community has agreed to ensure L1 remains slow, expensive and transparent, it severely handicaps Lightning.