Python Dataframe rowwise min and max with NaN values

My dataframe has nans. But I am trying to find row wise min and max. How do I find it.

df = pd.DataFrame({"A":[12,NaN,13],"B":[45,12,65],"C":[63,78,NaN]})

df= 
    A    B     C
0   12   45    63
1   NaN  12   78
2   13   65    NaN

I am tring to find min and max in each row and difference between them in column A B and C. My code:

poacols = ['A','B','C']
df['dif'] = np.nanmax(df[poacols])-np.nanmin(df[poacols])

Present output:

df['dif'] = 
0    66
1    66
2    66

Expected output:

df['dif'] = 
0    51
1    66
2    52