In a word, no.
QtQuick pre-5.0 are extremely premature and uninteresting in terms of technology and what you are actually able to achieve with it. You would not see much improvement in terms of results or effort.
Since 5.0, QtQuick has drastically leaped forward to become a well suited successor to GUI development with Qt. But, as you said, most of the DCC software are still on 4.8, so what now?
If we ignore for now the fact that each new increment of most DCC software will build upon Qt 5.6 and above, as per the vfx-platform specification, with pyblish-qml the solution was interprocess communication.
pyblish-qml runs in an independent Python interpreter with PyQt5, and visualises data coming from Maya and others by communicating remotely, in this case by making RPC calls to and from the GUI.
The decision was made so as to not depend on software featuring an integration with Qt and build one GUI fit for all. The vast majority of development time in this case was spent external to the software in which it was used, so it made sense long-term at the time.
Now, in retrospect, whether it was the right decision or not is questionable, as there has been numerous other unforeseen issues with way it has to manage it's interprocess communication. Some of it due to inexperience, some of it due to mere complexity.
Apart from all this, I have personal interest in the technology, OpenGL and "edgy" graphical user interface design (I want this) which is also a strong motivator.
So if you're asking whether you should go with widgets or go with QML for your next project, my recommendation would be to go with widgets for hack-and-slash tools, for tools you need many artists and TDs to understand and develop, and go for QML when you (a) need to unleash the artist in you and/or (b) when you expect to build and maintain it long term. Widgets are dead and no longer developed, at some point this will no longer be a choice.