National cancer incidence is estimated using the incidence/mortality ratio in countries with local incidence data: is this estimation correct?