Let us have a look at all the options one at a time.
In Option A, we can see that the reverse may not be true as the increase in the risk for lung cancer may not necessarily mean an increase in cigarette smoked in a day. For example, high risk of lung cancer may be due to high exposure to asbestos dust too.
In Option B, again we can see that the reverse may not be true as an increase in the height of an infant does not necessarily mean that the age of the infant is increasing too. For example the infant may have a rapid gain in height even if the age is not increasing as rapidly.
In Option C, too, that an increase in the amount of pollution in a city does necessarily mean that the number of vehicles in the city has increased. For example, this increase in pollution may be due to the establishment of a high pollution causing industry in the city or in it's vicinity.
Likewise, in Option D, an increase in the density of water does not necessarily mean that the concentration of salt in the water has increased.
Only in Option E do we see a possible reverse dependence happening because an increase in the phone bill amount does usually mean an increase in the number of calls made by the cell phone.
So, in the given list of Options only in Option E can we reverse the dependent and independent variables while keeping the interpretation of the slope meaningful.
