You're right on the graph. The graph in the article is very misleading, especially the N=4000 non-smokers.
If you look the table in the source paper, 53.4% of 296 Asian American Female (Single Group) lung-cancer patients are non-smokers.
Also, share of lung-cancer patients being non-smoker vs. smoker being higher in Asian american females, does not necessarily mean they're more likely to get lung-cancer. This assumes the non-smoking vs. smoking population is the same in teh general population, which it isn't.
