The term 'main-sequence' is not an absolute 'temporal' interpretation of a typical star. i.e. they do not all start in the top left and work their way down the main-sequence line to the bottom right of the H-R diagram as an evolution.
The H-R diagram is a distribution plot - a snapshot of observation of all stars taken in an almost infinitely small scale of time compared with the life time of a star. It simply plots the distribution of stars based on what is observed of their brightness and surface temperature.
(Think of it like taking a snapshot of all the aircraft in the world. Most would be either on the ground clustered at airports or in the air.)
However, there is a strong correlation between the mass of the star and it's position on the H-R diagram. It's also the mass of a star which is the greatest factor in determining the actual temporal evolution of that star. There is also a correlation between the life-span of a star and it's position on the H-R diagram. It's the mixing/transposing of these facts where many people confuse the interpretation of the H-R diagram.
Blue stars (progenitors of Blue-Supergiants) are only considered very rare main sequence. They are hot, bright and short lived. i.e. the population density of these stars does not occupy the dense population defining main-sequence other than at the limits of the distribution pattern. Although the dividing line on the Hertspung -Russell diagram looks small, they are actually well differentiated because the luminosity scale is logarithmic.
This is what wiki has to say:
" O type main sequence stars and the most massive of the B type blue-white stars become supergiants. Because of their extreme masses they have short lifespans of 30 million years down to a few hundred thousand years.
[3] They are mainly observed in young galactic structures such as
open clusters, the arms of
spiral galaxies, and in
irregular galaxies. They are less abundant in spiral galaxy bulges, and are rarely observed in
elliptical galaxies, or
globular clusters, which are composed mainly of old stars.Supergiants develop when massive main sequence stars run out of hydrogen in their cores. They then start to expand, just like lower mass stars, but unlike lower mass stars, they begin to fuse helium in the core almost immediately. This means that they do not increase their luminosity as dramatically as lower mass stars and they progress nearly horizontally across the HR diagram to become red supergiants. Also unlike lower mass stars, red supergiants are massive enough to fuse elements heavier than helium so they do not puff off their atmospheres as planetary nebulae when their helium becomes depleted. Furthermore, they cannot lose enough mass to form a white dwarf, so will leave behind a neutron stars or black hole remnant, usually after a core collapse supernova explosion.
Stars more massive than about 40M
☉ cannot expand into a red supergiant. They burn too quickly and lose their outer layers too quickly, so they reach the blue supergiant stage, or perhaps yellow hypergiant, and then return to become hotter stars. The most massive stars, above about 100M
☉, hardly move at all from their position as O main sequence stars. These stars convect so efficiently that they mix hydrogen from the surface right down to the core. They continue to fuse hydrogen until it is almost entirely depleted throughout the star, then very rapidly evolve through a series of stages of very similar hot and luminous stars, If supergiants, slash stars, WNh stars, WN stars, and possibly WC or WO stars. They are expected to explode as supernovae but it is not clear how far they evolve before this happens. The existence of these supergiants still burning hydrogen in their cores may necessitate a slightly more complex definition of supergiant: a massive star with increased size and luminosity due to fusion products building up, but still with some hydrogen remaining."