Practically, can the meta be changed for the better?

Currently, the meta seems to be fairly set in stone. Sometimes I see pro teams in matches doing some interesting things, like having a support Sona that roams instead of staying bottom.

Is the current meta definitely the best way to win games? If you are looking to optimally scale each member of your team, then it seems like it is, but there may be more to it. Besides very special cases, should the meta of sending 2 bot, 1 mid, and 1 top be messed with?

I think there are two important points to make in answering your question.

• Can I individually change the meta for the better?
• Is the current meta the optimal way to play the game?

The answer to both of these is no.

League of Legends is a team game. If you're playing by yourself in Solo Queue or normal games, it's going to be impossible to change the meta. Maybe every once in a while you'll run across a team willing to try something slightly unconventional, but for the most part, you'll be paired with four other people who want to just run a standard game. Instalocking AP mid is something we're going to have to deal with until the meta changes on its own.

More interesting is the second question, about whether the current meta is the optimal way to play. It is not, but the reason for that is that there is no optimal way to play. That is what metagame means.

Firstly, note that the North American meta is different from the Chinese, European, Korean, and other regional metas. Each tends to evolve individually because there is so little cross-play. Why do pro teams from different regions have success with utterly different team compositions? Because they're playing in different metagames. The metagame is simply what is currently popular, strategy-wise, that you can expect other teams to do. What that means is that, if you find and pick a strategy that is particularly strong against the current metagame, then on average, you will do better than other teams.

In League of Legends, this has turned into a kind of rock paper scissors meta scene, at least in North America. Originally, the metagame revolved around a lot of bruisers and AoE damage. This helps catch individuals off guard and kill them quickly, making for lopsided teamfights. A counter developed against this was the idea of the hypercarry, where one individual champ (for example, Kog'Maw), would be the entire focus of the team, and every other champion on the team would make it their life mission to make sure Kog'Maw didn't die. Kog'Maw and a few other champions could build in such a way that it didn't matter how much health or defense you had, he was still going to kill you rather quickly. If he got farmed up, there was nothing you could do. That is, until the meta evolved again, and became very assassin heavy. Champions like Nocturne and Talon would dive in and kill Kog'Maw instantly, often dying in the process. What that did, however, was leave 4 players on the hypercarry team who hadn't built any sort of damage, and couldn't deal with the oncoming threat. The counter for this team composition just happens to be to go back to a bunch of bruisers, catching people (the assassin, now) off guard and making for lopsided team fights.

This is a very slow shift, however. One person won't make a difference. One individual pro team can barely make a dent in this process, even with all the publicity they get.

Rather than try to dictate the meta, try to understand the meta. Figure out why the meta is the way it is, and that can help you set the trend in the right direction to counter it. If you notice that there are a lot of Kog'Maw or other hypercarries getting picked recently, maybe try going for an assassin to pick him off.