It likely is highly dependent on the cell line you are testing. It's also important to consider the difference between cells surviving and cells growing in serum-free conditions.

The environmental demands of cancer cells are not necessarily much different from their somatic counterparts. The key difference is that cancer cells (1) grow when they should not and/or (2) do not die when they should.

I agree with John Hardy Lockhart, If you are culturing cancer stem cells spheroids (in-vitro culture) that require serum-free medium supplemented with many growth factors then it is ok. Otherwise, Serum is vitally important as a source of growth and adhesion factors and its absence may affect cell integrity and such long-term serum-free environment will cause internal stress.

Hello Jalal Mosayebi Amroabadi

It will depend on the cancer cell type. One peculiar characteristics of cancer cells is that the growth factor requirements of many cancer cells are reduced compared to their normal counterparts, contributing to the unregulated proliferation of cancer cells both in vitro and in vivo. In some cases, cancer cells produce growth factors that stimulate their own proliferation. Such abnormal production of a growth factor by a responsive cell leads to continuous autostimulation of cell division (autocrine growth stimulation), and the cancer cells are therefore less dependent on growth factors from other physiologically normal sources such as serum.

In some other cases, the reduced growth factor dependence of cancer cells results from abnormalities in intracellular signaling systems, such as unregulated activity of growth factor receptors or other proteins (e.g., Ras proteins or protein kinases) as elements of signal transduction pathways leading to cell proliferation.

So, you should try culturing the cancer cells (of your interest) in vitro in serum free media yourself because there is no definite answer to your question. For instance, HepG2 (human liver cancer cell line) can survive in serum free media for 72 hrs. which may vary in case of other cancer cell lines.

Best.