you have the right to ask such important question. To run ANOVA, one should calculate the minimum smaple size needed in each group. In your case, even with large effect size, the group with 10 cases is definitly less than the needed sample. Thus the result is high likely will include type II error.
Assuming moderate effect size (.25) then the minimum sample size per group should be 53 case.
As you did not have such sample size, then the alternative is to run "Kruskal Wallis test" which is a nonparametric test can tolerate the small sample in the gourps.