Hossains et al reported that females with AML showed
Hossains et al. reported that females with AML showed a significant survival advantage over males , while our results showed no such differences after adjusting for attributing factors. The difference in outcomes may be attributable to the different study population; while their study population comprised all AML cases, ours included only 5+y survivors of all site cancer. As most mortality in childhood cancer tends to occur in the first years following initial diagnosis, it may be that after 5 years sex differences in survival disappear. However, Storm et al.  in their study relating to AML patients' outcomes suggested that the superior outcomes of females over males may be explained by molecular, genetic and other biological disease characteristics that affect survival across AML types . Importantly, our data showed that while in the general population females have better survival than males, in our cohort females lost their survival advantage compared to males.
Conflicts of interest
Introduction Epidermal Growth Factor Receptor (EGFR) mutations are present in subsets of patients with non-squamous non-small cell lung cancer (NSCLC), and associated with sensitivity to EGFR tyrosine kinase inhibitors (TKIs). Since 2013, standard oncology practice guidelines recommend EGFR (-)-Bicuculline methiodide testing of non-squamous NSCLC to select advanced-stage patients for first-line treatment with EGFR-TKIs [, , ]. However, practical barriers such as accessibility, costs, insufficient quantity or quality of tissue specimen available for testing and variable testing referral practices may limit full uptake of testing by all eligible patients [, , ]. Incomplete testing uptake may have an impact on estimation of mutation prevalence. Accurate estimation of the prevalence of EGFR mutation is necessary to inform policy and practice. Many studies have assessed the mutation prevalence in different settings and patient groups. A recent meta-analysis pooled the results of 456 studies published to June 2013, and reported that the prevalence of EGFR mutation was 32.3% (95% CI 30.9%–33.7%) with a higher prevalence observed in Asians, females, non-smokers and patients with adenocarcinoma . Similar findings were observed in earlier meta-analyses [, , ]. However, testing was incomplete and selective in many studies, which may bias the estimates of mutation prevalence. For example, the “mutMapII” study, that reviewed 151 articles published to June 2014 and created a global map of EGFR mutation frequency, reported wide variations in mutation prevalence between studies, even when grouped by geographic region and country . This may be explained, at least in part, by differing patient characteristics and mutation testing practices. Our previous analyses [6,11], based on a population-based cohort of 1857 patients diagnosed with non-squamous NSCLC in northern New Zealand between January 2010 and April 2014, showed that only 27% of patients were tested, of whom 22% were mutation positive. The latter is likely to be an overestimate as patients with a known higher chance of harbouring an EGFR mutation were more often tested. This analysis therefore investigated the impact of incomplete testing on the mutation prevalence previously reported, using the data from an expanded study population involving 2701 patients diagnosed up to December 2015. Specific research objectives are : to examine trends in the uptake of EGFR mutation testing in patients with non-squamous NSCLC in New Zealand , to develop a composite metric that quantifies the influences of demographic and clinico-pathological factors on the testing uptake, and  to estimate the prevalence of EGFR mutation if all patients were tested.