• 2019-07
  • 2019-08
  • 2019-09
  • 2019-10
  • 2019-11
  • 2020-03
  • 2020-07
  • 2020-08
  • Note that global PC incidence and mortality have been


    Note that global PC incidence and mortality have been reported to correlate directly with country-specific levels of socioeconomic development [17] which seems in K 252a to the higher PC incidence we found in eastern Europe compared to the rest of Europe. We found that survival for PC and BTC improved modestly over time, except for PC in eastern Europe. Changing diagnostic techniques, changing morphology mix, and improved treatments are all possible explanations. The affirmation of ultrasound-guided fine-needle aspiration biopsy of the pancreas [24] and endoscopic ultrasound imaging of the gallbladder [25] over the last 2–3 decades are likely to have improved diagnostic accuracy and staging, resulting in more effective treatments and plausibly better survival. In the USA, over the period 1974-2013, the incidence of better prognosis morphologies, like endocrine PC, increased, while that of poor prognosis forms (e.g. poorly specified PC) decreased in both sexes [19]. Over an overlapping period (1981-2010), 1-year and 5-year RS for PC increased in the USA [26]. Increasing incidence of endocrine PCs (over 1995–2007) has also been reported in Europe [27,28] and may have contributed to the modestly improving survival we found in the present study. Use gemcitabine-based chemotherapy may also have contributed to improved survival during the study period, but FOLFIRINOX was introduced only the end of the period [29,30] and so is unlikely to have had any effect. As regards substitution of outdated ESP with RESP to age-standardise incidence and mortality rates, this produced practically no changes in incidence and mortality rankings of European regions, but did result in considerably higher ASRi and ASRm – that are perhaps more accurate than those afforded by the old standardisation – and raising the possibility that the lethality rankings of these cancers might change if mortality rates of all cancers were compared using the RESP. ASRi and ASRm estimated with RESP in southern Europe were similar to RESP-standardised rates for Italy in the year 2007: ASRi and ASRm for PC of 20 and 18 in men, and 16 and 14 in women; with corresponding rates for BTC of 8–9 in men and 6–7 in women [31]. RESP-standardised incidence was higher for PC in northern and central than southern Italy, while BTC incidence was higher in the south. Use of a single standard population for age standardisation is necessary to allow comparisons between studies, countries, and over time. Since EUROSTAT is adopting RESP, it is advisable that all European countries and studies involving European populations adopt RESP without delay. As far as we are able to ascertain, the present study is the first to examine incidence, mortality, and survival together for PC and BTC patients across Europe. Although it was necessary to use two datasets to obtain all three indicators, it is important to analyse the three population-based indicators together so as to better interpret overall trends [32]. The fact that nine CRs included in the survival analyses were excluded from the incidence/mortality analyses raises possibility of a study limitation due to dataset compatibility. Because of under-estimation of PC incidence due to registration problems in the Swedish CR [18,33] – which as a large national registry would have had major effects on estimates for northern Europe – we excluded this CR from the incidence/mortality analyses. However little error is expected by including Sweden in the survival analyses, since RS estimates were in line with those of other northern European countries, which do not have registration problems [9]. It is possible that incomplete registration was a problem for other CRs. This was mitigated by performing the analyses at the regional instead of the CR or country level, since, if data quality for most registries in a European region is adequate, inclusion of a small (non-national) CRs with incomplete registration will have little effect on the overall accuracy of incidence/mortality estimates.