Introduction

The number of individuals living with a history of cancer has increased steadily worldwide over several decades [1]. This rise is driven by population growth and ageing, with more people reaching cancer-prone ages, as well as by earlier detection through screening and improved diagnostics, which increase both incidence and survival. Advances in cancer treatment have further contributed to improved survival, resulting in a growing population of cancer survivors.

In Switzerland, approximately 48,000 new cancer diagnoses are recorded annually, with a continuing upward trend [2]. At the same time, survival has improved substantially for many cancer types: currently, 62% of men and 64% of women are alive five years after diagnosis. As a result, cancer survivorship has increasingly become a long-term or permanent condition. This shift has drawn attention to survivors’ quality of life and their specific healthcare needs [3]. However, healthcare systems have not yet fully adapted to the complex physical, psychological, informational and supportive care needs of cancer survivors, who frequently report persistent symptoms, psychological distress, and challenges related to work, finances, access to high-quality information and help in processing it [4-6].

Reliable estimates of the number of cancer survivors, particularly long-term survivors, are essential for healthcare planning, resource allocation and policy-making. Most countries, including Switzerland, routinely report limited-duration prevalence, typically restricted to diagnoses within one, five or ten years before a reference date [7]. The 1-year limited-duration includes patients immediately after their diagnoses, many of whom are in intensive care; the 1- to 5-year prevalence includes patients having regular check-ups and maybe additional treatment; and the 5- to 10-year prevalence includes ‘long-term survivors’, who are mainly under surveillance for recurrence.

The complete cancer prevalence, or lifetime cancer prevalence, quantifies all individuals alive at a given time who have ever been diagnosed with cancer. This estimate is still not part of routine cancer surveillance in most countries, including Switzerland, because cancer registration often does not span the full lifespan of older populations and because it requires complex modelling to estimate missed cancer diagnoses before registration started, as well as the expected survival of these individuals. Complete cancer prevalence is needed to estimate the number of ‘very long-term survivors’ with ten or more years since diagnosis, a group in which impairments in quality of life may persist or worsen [6].

In Switzerland, the first reporting of complete cancer prevalence dated to 2013 [8] and 2014 [9], both for date of reference 31 December 2010, with predictions to dates of reference 2015 and 2020, and have not been updated since then. This study addresses this gap by providing complete cancer prevalence estimates in Switzerland for the end of 2025. We also describe complete cancer prevalence trends over the past two decades for major cancer types in an international context.

Methods

For this population-based study, the Swiss national cancer dataset of the National Agency for Cancer Registration (NACR) was used. In Switzerland, cancer cases are collected and registered separately by cantonal cancer registries. The NACR is legally mandated to coordinate cancer registration in Switzerland, to collect the data recorded in the cantonal cancer registries and the Childhood Cancer Registry, to harmonise their collection and coding, and to ensure data quality. The NACR also processes the data, evaluates it, makes it available to the Federal Statistical Office (FSO) and to researchers and is responsible for national health reporting on cancer. Population coverage by cancer registration in Switzerland has progressively expanded since 1970 and has become complete nationwide for diagnoses since 2020 (figure S1 in the appendix). In 2020, the Cancer Registration Act (CRA) came into force in Switzerland, obliging all private and public healthcare institutions to report cancer cases to the corresponding cancer registry [10].

Complete cancer prevalence was estimated by modelling temporal trends in cancer incidence and survival, with backward extrapolation to incorporate cases diagnosed prior to the start of registration and their survival up to the prevalence date.

Diagnoses

Primary malignant cancer diagnoses from 1980 to 2021 were selected from the Swiss national cancer dataset. Analysis groups were all cancer types together (excluding non-melanoma skin cancer) and 26 specific cancer types, each stratified by sex (table 1). People diagnosed with multiple cancers belonging to different reporting groups were included separately in each group, whereas for multiple diagnoses within a single reporting group, only the first-occurring diagnosis was considered. Diagnoses based on DCO (Death Certificate Only) and diagnoses after death were excluded.

Population

The Swiss population from 1980 to 2023, as well as predictions for future population in 2024 and 2025, stratified by canton, age and sex were provided by the FSO. For predictions, we employed the “reference” scenario for future growth [11].

Estimation of x-year limited duration prevalence (LDPx) by the counting method

Cancer prevalence is the number of people alive at a given reference date who were previously diagnosed with cancer. Limited-duration prevalence (LDPx) represents the number of people at the reference date who were diagnosed with cancer within the past 𝑥 years (e.g. 𝑥 = 1,2,5 or 10 years). For example, the LDP10 for 31 December 2025 comprises all people alive at that date, and diagnosed between 1 January 2016 and 31 December 2025.

There is active vital status follow-up implemented in Switzerland. Vital status is based on information provided to the cantonal cancer registries by the Central Compensation Office, cantonal census registries/inhabitant control and the FSO. For people who were lost to follow-up or were without active vital status follow-up, the probability of still being alive at the reference date, conditional on the length of observed survival, was estimated as described by Gigli et al. [12]. Survival probabilities were estimated by flexible parametric models using the stpm2 command written for the Stata® language (version 11, 2009) [13]. In brief, flexible parametric models are modified Weibull models where the time after diagnosis is incorporated as a restricted cubic spline function. LDPx was estimated separately for the predominantly German-speaking and the predominantly French-/Italian-speaking part of Switzerland in order to adjust for heterogeneous registration coverage. Counts specific for cancer type, attained age, sex and language region were extrapolated from partially covered language regions to full language regions. Swiss prevalence counts were estimated as the sum of both extrapolated language regions. Prevalence proportions for the whole of Switzerland were estimated as the weighted sum of observed language region-specific proportions. Weights were the population ratios of language region with the whole of Switzerland.

More comprehensive descriptions of the methods used to estimate limited-duration prevalence and to project prevalence to future reference dates (see below) are provided in Chapter 4 of the official documentation on the statistical methods employed by the NACR, the FSO and the Childhood Cancer Registry for calculating the main cancer epidemiological indicators [14].

Projection of LDP10 for reference date 31 December 2025

LDP10 was estimated for the reference date 31 December 2021 with the counting method based on 21 of 26 Swiss cantons (about 74% Swiss population coverage) providing diagnoses from 2012 to 2021 at the time of analysis (figure S1 in the appendix). The 21 cantons and their regional language (German unless otherwise noted) were: Appenzell Ausserrhoden (AR), Appenzell Innerrhoden (AI), Basel-Landschaft (BL), Basel-Stadt (BS), Fribourg (FR – French), Geneva (GE – French), Glarus (GL), Graubünden (GR), Jura (JU – French), Luzern (LU), Neuchâtel (NE – French), Nidwalden (NW), Obwalden (OW), Sankt Gallen (SG), Thurgau (TG), Ticino (TI – Italian), Uri (UR), Valais (VS – French), Vaud (VD – French), Zürich (ZH) and Zug (ZG). The data from the cancer registries of Aargau (AG), Bern (BE), Solothurn (SO), Schwyz (SZ) and Schaffhausen (SH) was not used because they began operating after 2012.

Projection of LDP10 to reference date 31 December 2025 required separate projection of incidence rates and survival proportions for the years 2022 to 2025 as suggested by Pisani et al. [15]. Incidence projection was conducted with age-period models implemented as Stata® (version 3.1, 1993) commands predmap and outmap written by Dyba and Hakulinen [16]. The projection is based on observed incidence trends within the period 2012–2021, and takes the predicted population growth until 2025 into account. Restricted cubic spline regression was applied to observed and projected incidences in 10-year age groups for calendar years 2012 to 2025 in order to interpolate them for single years of age.

Future survival proportions were estimated by applying the period method to patients with very recent dates of follow-up (2017 to 2021), a validated approach of survival prediction [17, 18]. The survival functions were estimated with flexible parametric models using the stpm2 command of the Stata® language [19]. Models included the covariate age at diagnosis as linear term. Proportional hazards were assumed. Survival probabilities at 0.5, 1.5, …, 9.5 years after diagnosis for single years of age from age 0 to age 95 were predicted.

The projected incidence counts and survival probabilities for single years of age and for each calendar year were processed for future LDP10 counts as described by Pisani et al. [15].

Modelled completeness index (Rmod)

We applied the completeness index method [20], which has been validated extensively with US SEER data [21]. The method has also recently been endorsed for the European Cancer Information System (ECIS) [22]. Completeness indexes (or R indexes) specific for dates of reference and LDPx duration indicate which fraction of complete cancer prevalence the LDPx represents. For example, an R index of 0.6 (or 60%) for LDP10 indicates that the LDP10 count represents 60% of the complete cancer prevalence count for the same date of reference.

We used the software COMPREV version 3.0.29 [23], developed by the US National Cancer Institute, to estimate R indexes based on modelled trends in incidence and survival. Incidence and survival trend data was based on diagnoses from all Swiss cantons registering cancer cases (figure S1 in the appendix).

Incidence trends were estimated in a 2-step process. First, aggregated incidence rates were calculated for eight 5-year periods (1982–1986, …, 2017–2021) and nineteen 5-year age groups (0–4, …, 90–94). Methodology details are provided in chapter 2 of [14]. Second, the incidence rates were modelled using Poisson regression on the number of cases, adjusting for 6th degree polynomial for centred (using reference age 55) and scaled (using scale factor 30) age group (for example, for the age group 40–44, the middle age 42 would be used, and converted to (42 – 55)/30 or -0.43), as well as birth cohort (derived as the difference of middle age and middle period minus the minimum observed birth cohort 1890) and its squared and cubed terms in order to adjust for risk trends across different birth cohorts (see supplementary appendix of the EUROCARE-6 report [1]):

ncases(agec)=β0+β1agec+β2agec2+...+β6agec6+γ1coh+γ2coh2+γ3coh3+offset(log(∑py)) (1)

with agec as the centred and scaled age at diagnosis, coh as the birth cohort and using the sum of population years at risk (py) as an offset term. The intercept β0 (representing the oldest birth cohort) and 6 age coefficients were exported to COMPREV (in a format described at https://surveillance.cancer.gov/help/comprev/file-formats/data-source-files). The rational for using the incidence function of the oldest birth cohort and other analytical details are described by Capocaccia et al. [20] and Merrill et al. [24].

Survival trends were estimated by first calculating relative survival rates for eight incidence periods (1980–1985, 1986–1990, 1991–1995, 1996–2000, 2001–2005, 2006–2011, 2012–2016, 2017–2021), five age groups (0–44, 45–54, 55–64, 65–74, 75+) and survival durations up to 10 years after diagnosis. The relative survival rates were then fitted with Weibull mixture cure models using non-linear regression (see supplementary appendix of the EUROCARE-6 report [1]):

S(t,x-t,y)=((1-C)+C*exp(-(scale*(x-t))shape))shape (2)

where S(t, x-t, y) is the period-specific relative survival probability from age at diagnosis t until attained age x, C is the proportion not cured of cancer, t is the middle age at diagnosis in the age group, reference age at diagnosis is 65, y is the middle year in the incidence period, reference period is 1997, and scale and shape are parameters of the Weibull distribution. The coefficients for attained age, period, scale and shape were exported to COMPREV.

The incidence modelling with equation 1 and the survival modelling with equation 2 were done in R (version 4.5.2; 2025-10-31).

The fitted parameters for both models as well as their variance-covariance matrix were processed by COMPREV in order to obtain completeness indexes Rmod for LDP10 at reference dates 31 December 2005, 31 December 2015, 31 December 2021 and 31 December 2025. Upper and lower 95% confidence limits for Rmod were derived in COMPREV using the delta method.

Empirical completeness indexes (Remp)

The empirical completeness index is a simple approximation of the Rmod index as derived with modelling. Estimating Remp was relatively straightforward and relied on fewer assumptions than Rmod. Its main purpose was to provide a validation check for Rmod. LDP10, LDP36 and LDP41 were estimated by the counting method for reference date 31 December 2021 based on nine Swiss cantons – AI, AR, BL, BS, GE, NE, SG, VD, ZH – which provided diagnoses from 1981 to 2021 (figure S1 in the appendix). The empirical completeness index Remp, specific for LDP10, was estimated as

Remp=LDP10LDP41 (3)

when the following criteria were met: LDP41/LDP36 ≤1.05, i.e. when LDP appeared to have reached a maximum and LDP41 can be regarded as an approximation of complete cancer prevalence, and LDP41 ≥20 cases.

Complete prevalence estimation using completeness indexes (empirical or modelled)

Complete prevalence counts were estimated by

CP=LDP10R (4)

using LDP10 and R indexes specific for cancer reporting group, sex, age and date of reference. Upper and lower 95% confidence limits for complete cancer prevalence were derived by applying the delta method to the logit transformation of complete cancer prevalence. Complete prevalence proportions are counts per 100,000 people alive at the date of reference.

Complete cancer prevalence estimated with Swiss Rmod was preferred but data models did not converge for every reporting group, in which case either the R index for EUROCARE-6 (EC-6) data was used [1], if the cancer definition was the same as in EC-6, or the empirical R index Swiss Remp, if the cancer definition deviated from that of EC-6. The EC-6 R indexes were calculated with COMPREV based on modelled incidence and survival trend parameters for the pool of 29 European countries including Switzerland kindly provided by the EUROCARE-6 study group (see Acknowledgements).

Completeness index validation

Swiss modelled indexes Rmod for 31 December 2021 were validated by comparing them with Swiss empirical indexes Remp for the same date. Both indexes overlapped in almost every case, except for head/neck cancer in men and for lung cancer (figure S2 in the appendix). For cancer reporting groups without Swiss Rmod, Swiss Remp were compared to the EUROCARE-6 indexes and the overlap justified their use as proxies for missing Swiss Rmod (figure S3 in the appendix).

Average annual percentage change (AAPC)

The AAPC of prevalence counts were calculated as

AAPCx=100*((CxC0)1x-1)(5)

with the number of years x between beginning and end, the counts C0 at the beginning and Cx at the end.

The Stata® and R codes used to generate the results are available on request.

Table 1: Definition of reporting groups for 27 types of cancer. In 35 sex-specific cancer reporting groups, the R index was based on Swiss data (CH) and in 11 on EUROCARE-6 (EC-6) data. Empirical R indexes are labelled “Yes” if they met the acceptance criteria listed in Methods.

Cancer reporting groupICD-10*Data source for modelled R, in menData source for modelled R, in womenSwiss empirical R acceptable, in menSwiss empirical R acceptable, in women
All cancer types#C00–C43, C45–C97CHCHYesYes
Head, neck#C00–C14CHCH (Remp)$YesYes
OesophagusC15 EC-6EC-6YesYes
StomachC16CHCHYesYes
Colon, rectum, anusC18–C21CHCHYesYes
LiverC22CHCHYesYes
GallbladderC23–C24EC-6EC-6YesYes
PancreasC25CHEC-6YesYes
LarynxC32EC-6EC-6YesYes
LungC33–C34CHCHYesYes
Bone#C40–C41––NoNo
Skin melanoma C43CHCHYesNo
Soft tissue#C47, C49CHCHNoYes
Breast (female)C50–CH–Yes
Cervix uteriC53–EC–6–No
Corpus uteriC54–C55–CH–Yes
OvaryC56–CH–No
ProstateC61CH–Yes–
TestisC62CH–No–
Kidney, other urinary organsC64–C66, C68CHEC–6YesYes
Urinary bladder#C67CHCHYesYes
Brain#C70–C72CHCHNoNo
ThyroidC73CHCHNoNo
Hodgkin lymphoma (HL)C81CHEC–6NoNo
Non-Hodgkin lymphoma (NHL)#C82–C86, C96CHCHYesYes
Multiple myelomaC90EC–6CHYesYes
LeukaemiaC91–C95CHCHYesYes

* International Classification of Diseases 10th revision [reference = https://icd.who.int/browse10/2019/en].

# Cancer definition in EUROCARE-6 (EC-6) differs from this report.

$ None of the modelled R indexes were adequate, and the Swiss empirical R index was applied.

Results

We estimated the complete prevalence for 46 sex-specific cancer reporting groups using 3 types of R indexes (table 1). In 34 reporting groups, R indexes were modelled with Swiss data, but in 11 groups, Swiss data models did not converge and the R index of the EUROCARE-6 (EC-6) study was applied instead, and in 1 group (head and neck tumours in women), the Swiss empirical R index was used, because Swiss data models did not converge and the cancer definition deviated from that of EC-6. For every reporting group where EC-6 R indexes were applied, we found that they overlapped with Swiss empirical R indexes (figure S3 in the appendix).

As of the end of 2025, a considerable number of people in Switzerland were estimated to be living with a past diagnosis of malignant cancer: about 242,260 men or 5.4% of the male population (table 2) and about 253,040 women or 5.6% of the female population (table 3). It is estimated that 55% of the affected men (table 2) and 51% of the affected women (table 3) were aged 70 or over. About 96,350 (40%) of the affected men will have had a past diagnosis of prostate cancer and about 110,320 (44%) of the affected women breast cancer. Second and third in proportion are skin melanoma (11% in men and in women) and colorectal cancer (10% in men and 9% in women). About 27% of the Swiss male population aged 80 or over will have been diagnosed with malignant cancer, 48% of them with prostate cancer. Among women, 19% of those aged 80 or over will have been diagnosed with malignant cancer, 46% of them with breast cancer. For individuals aged below 50, 0.7% men will have had a cancer diagnosis, predominantly testicular cancer (22%) or skin melanoma (12%). In women aged below 50, the corresponding estimates were 1.1%, predominantly breast cancer (31%) or skin melanoma (12%).

Table 2: Complete prevalence by attained age in men (date of reference: 31 December 2025). The predicted male population on 31 December 2025 was 4,501,331 in total, distributed among age groups as follows: 2,726,106 (0–49 years); 624,769 (50–59); 555,300 (60–69); 363,488 (70–79); 231,668 (80+).

Attained age: 0–49Attained age: 50–59Attained age: 60–69Attained age: 70–79Attained age: 80+All attained ages
Count#Proportion$Count#Proportion$Count#Proportion$Count#Proportion$Count#Proportion$Count#Proportion$
Head, neckn5111912582012463443276676117357498732194
95% CI188; 13947; 511010; 1566162; 2512151; 2819387; 5082360; 3240649; 8911113; 2706480; 11687688; 9918171; 220
Oeso­phagusn6522353850491612168479207189642
95% CI9; 4490; 16176; 31528; 50430; 59277; 107527; 712145; 196357; 641154; 2771675; 214637; 48
Stomachn221855489104118713033581339578445799
95% CI83; 5893; 22431; 71269; 114888; 1220160; 2201131; 1501311; 4131062; 1689458; 7294008; 495789; 110
Colon, rectum, anusn12984826434235269949718019757347317223; 738527
95% CI426; 395416; 1451941; 3600311; 5764447; 242801; 11246212; 82981709; 22835777; 93452494; 403421,110; 26,693469; 593
Livern177638261728131746205382165241554
95% CI35; 8841; 32290; 50346; 81624; 850112; 153637; 872175; 240255; 574110; 2482066; 282446; 63
Gallbladdern492901417031227621767671216
95% CI13; 1910; 761; 13210; 21135; 21424; 39188; 27352; 75128; 24355; 105610; 83214; 18
Pancreasn160637961618111583160336145207546
95% CI52; 4872; 18295; 48747; 78524; 72894; 131490; 695135; 191214; 52792; 2271804; 238740; 53
Larynxn71323938582105753207548237219349
95% CI10; 5040; 19174; 32728; 52490; 69188; 125639; 888176; 244385; 780166; 3371916; 251043; 56
Lungn3351213682193141566394510853147135811; 936265
95% CI17; 65871; 242940; 1992151; 3192629; 3751473; 6763343; 4654920; 12802128; 4654919; 200910; 201; 13; 965227; 310
Skin melanoman226383412366059891079710419546704289426; 184582
95% CI1229; 416945; 1533211; 5295514; 8485031; 7129906; 12846156; 81981694; 22555367; 83752317; 361523; 612; 29; 035525; 645
Soft tissuen5392047175564102591163603260276962
95% CI423; 68915; 25372; 59660; 95465; 68484; 123499; 701137; 193493; 737213; 3182523; 303856; 67
Prostaten1190446301100922; 177399436; 277998030; 40613; 12596; 3522141
95% CI0; 27,573,9860; 1,011,4793298; 12,037528; 192718,118; 27,1463263; 488831,074; 42,3528549; 11,65120,812; 44,4228984; 19,17579,475; 116,8121766; 2595
Testisn4249156427568440517292338643114049216; 053357
95% CI3693; 4889135; 1791312; 13,926210; 2229933; 17,581168; 316619; 289,2505; 79; 5760; 207,991,8400; 89,780,1364775; 53,974106; 1199
Kidneyn836311334213254245831328622573111110; 417231
95% CI333; 209812; 771046; 1701167; 2722190; 2950394; 5312732; 3590752; 9881939; 3416837; 14749242; 11,741205; 261
Bladdern18876029616222922819775323413968465188
95% CI6; 60390; 222383; 94861; 1521341; 1962242; 3532477; 3208681; 8832675; 39091155; 16877477; 9583166; 213
Brainn8253056590527953459512956239053
95% CI709; 96026; 35480; 66677; 107445; 62480; 112266; 44673; 12347; 35420; 1532163; 264148; 59
Thyroidn77728903145929167737203404174374983
95% CI449; 134316; 49526; 155084; 248566; 1524102; 274361; 150499; 41454; 299523; 12932687; 523260; 116
HLn123945921147855154587162277119387886
95% CI1099; 139840; 51712; 1191114; 191638; 1144115; 206419; 823115; 226118; 65051; 2813422; 439676; 98
NHLn1453531895303271348931138572601112311; 775262
95% CI856; 246531; 901471; 2440235; 3912288; 3218412; 5792688; 3606740; 9922033; 3328878; 143710,557; 13,134235; 292
Multiple myeloman1756416678701571049289676292318671
95% CI37; 8241; 30300; 57748; 92727; 1041131; 187896; 1227247; 338490; 932211; 4022778; 365462; 81
Leukaemian15875811661871524274178649115856847648170
95% CI1133; 222442; 82917; 1484147; 2381297; 1790234; 3221564; 2038430; 5611281; 1961553; 8466913; 8460154; 188
All cancers n19; 48171531,742508158,04710,45369,92819,23863,06627,223242,2645382
95% CI7739; 49,038284; 179923,353; 43,1453738; 690649,730; 67,7548956; 12; 20160,981; 80,18816,777; 22,06147,521; 83,69620,512; 36,128213,913; 274,3724752; 6095

# Absolute number of people alive on the date of reference in the indicated age range with a history of the cancer indicated.

$ Relative number of people alive on the date of reference in the indicated age range with a history of the cancer indicated, expressed as per 100,000 people.

CI: confidence interval; HL: Hodgkin lymphoma; NHL: non-Hodgkin lymphoma.

Table 3: Complete prevalence by attained age in women (date of reference: 31 December 2025). The predicted female population on 31 December 2025 was 4,557,006 in total, distributed among age groups as follows: 2,628,566 (0–49 years); 624,072 (50–59); 563,248 (60–69); 410,397 (70–79); 330,723 (80+).

Attained age: 0–49Attained age: 50–59Attained age: 60–69Attained age: 70–79Attained age: 80+All attained ages
  Count#Proportion$Count#Proportion$Count#Proportion$Count#Proportion$Count#Proportion$Count#Proportion$
Head, neckn366146179910481861177287896271412491
95% CI78; 17173; 65364; 104658; 168762; 1441135; 256852; 1626208; 396488; 1647147; 4983213; 529371; 116
Oeso­phagusn221781218934216531655067015
95% CI0; 99800; 38034; 1805; 29128; 28023; 50147; 31736; 7772; 37722; 114475; 94410; 21
Stomachn217836859578103683166870263271760
95% CI73; 6463; 25246; 55139; 88433; 77377; 137523; 893127; 218615; 1231186; 3722276; 324250; 71
Colon, rectum, anusn15495925014014569811640815618225248723.253510
95% CI561; 428221; 1631764; 3546283; 5683698; 5645657; 10025382; 76291311; 18596551; 10–3271981; 312320.448; 26.441449; 580
Livern10941352221338240581624985819
95% CI8; 15140; 5857; 3219; 51128; 35523; 63153; 37537; 9160; 44318; 134550; 133912; 29
Gallbladdern351671114225221542377270315
95% CI7; 1770; 742; 1087; 17109; 18619; 33181; 27044; 66179; 31454; 95601; 82213; 18
Pancreasn172735857565100625152400121211947
95% CI55; 5362; 20268; 47843; 77470; 67983; 121525; 743128; 181265; 60380; 1821828; 245840; 54
Larynxn241671114125180441243753512
95% CI0; 71130; 27123; 1944; 3177; 25914; 4699; 32624; 8027; 5718; 173317; 9037; 20
Lungn45617153224533215903745912263279611.686256
95% CI32; 64451; 2451078; 2178173; 3492807; 3930498; 6983185; 4404776; 10731764; 3925534; 11879949; 13.727218; 301
Skin melanoman3424130526384365211158659416076232188428.034615
95% CI2229; 525985; 2004190; 6611671; 10595329; 7980946; 14175392; 80641314; 19654584; 84721386; 256224.939; 31,513547; 692
Soft tissuen486193315339069411100444134206145
95% CI368; 64214; 24243; 45039; 72298; 51053; 91321; 52678; 128331; 594100; 1801818; 233640; 51
Breastn889333817,019272726,073462929,517719228–8208714110–3212421
95% CI5257; 15,043200; 57214,055; 20,6082252; 330222,092; 30,7713922; 546324,609; 35,4045996; 862720,823; 39,8886296; 12,06198,103; 124,0612153; 2722
Cervix uterin8913413342141665296145235412573806599145
95% CI730; 108828; 411090; 1633175; 2621284; 2161228; 384958; 2201233; 536422; 3738128; 11305185; 8400114; 184
Corpus uterin4771815312453186566448510934869147214,548319
95% CI104; 21944; 831141; 2054183; 3292674; 3797475; 6743769; 5337918; 13013425; 69211036; 209312,557; 16,855276; 370
Ovaryn6742610991761558277169041214014246421141
95% CI417; 108916; 41889; 1358142; 2181307; 1856232; 3301394; 2048340; 499925; 2120280; 6415649; 7298124; 160
Kidneyn5252053185992176133832613734154759104
95% CI77; 35673; 136280; 100945; 162681; 1445121; 257975; 1836238; 447815; 2316246; 7003546; 638778; 140
Bladdern10341652642275732178987298240853
95% CI0; 73–4150; 279333; 8285; 133209; 85337; 152455; 1176111; 287522; 1865158; 5641548; 374634; 82
Brainn6042336458418743137615547185441
95% CI491; 74319; 28286; 46146; 74337; 52060; 92233; 42157; 10357; 42417; 1281617; 212635; 47
Thyroidn260599273843927534892090509133540411–521253
95% CI2175; 312183; 1192344; 3198376; 5122275; 3333404; 5921011; 4319246; 1052212; 841764; 25458877; 14,953195; 328
HLn10434067410856099471115335101308368
95% CI902; 120534; 46451; 100872; 161363; 86564; 154310; 71676; 175159; 70748; 2142616; 363457; 80
NHLn9793712692032021359258763025237639380206
95% CI435; 220317; 84886; 1817142; 2911618; 2525287; 4482170; 3084529; 7521899; 3353574; 10148142; 10,805179; 237
Multiple myeloman104427444577103773188520157224849
95% CI7; 14500; 55165; 45327; 73443; 75379; 134635; 942155; 229350; 772106; 2331859; 271941; 60
Leukaemian124847763122973173117728711983625360118
95% CI790; 197330; 75533; 109385; 175757; 1251134; 222959; 1445234; 352887; 1618268; 4894612; 6229101; 137
All cancern28,291107638,070610057,87710,27665,82416,03962,97519,042253,0365553
95% CI15,030; 53,250572; 202629,617; 48,9354746; 784147,898; 69,9348504; 12,41654,483; 79,52713,276; 19,37844,658; 88,80513,503; 26,852221,235; 289,4084855; 6351

# Absolute number of people alive on the date of reference in the indicated age range with a history of the cancer indicated.

$ Relative number of people alive on the date of reference in the indicated age range with a history of the cancer indicated, expressed as per 100,000 people.

CI: confidence interval; HL: Hodgkin lymphoma; NHL: non-Hodgkin lymphoma.

A high percentage of those estimated to have ever had a malignant cancer at the end of 2025, about 40% of men and almost 50% of women, belong to the group of very long-term survivors, i.e. the diagnosis will have occurred 10 or more years in the past (see “All cancer types” in figures 1 and 2). Around a quarter, 24%, of the affected men and 22% of the affected women were estimated to be long-term survivors, i.e. the diagnosis will have occurred between 5 and 10 years previously. About 17% of affected men and 13% of affected women will have had a recent diagnosis (less than two years previously). Cancer types in figures 1 and 2 are sorted by the proportion of very long-term survivors and, as expected, cancers with a good prognosis and being diagnosed at relatively young ages are at the top: testis cancer in men and cervical cancer in women, with over 70% very long-term survivors. In pancreatic cancer, on the other hand, many men (29%) and women (27%) will have had their diagnosis less than a year previously.

Figure
Figure 1: Percentage distribution of patient groups by time since diagnosis in men (all attained ages combined). The date of reference is 31 December 2025. Sorting from top to bottom is by decreasing proportion of very long-term survivors (>10 years).
Figure
Figure 2: Percentage distribution of patient groups by time since diagnosis in women (all attained ages combined). The date of reference is 31 December 2025. Sorting from top to bottom is by decreasing proportion of very long-term survivors (>10 years).

Table 4 shows the change in complete cancer prevalence for the last 20 years in seven leading cancer types. Based on our incidence and survival trend models, we also derived R indexes for LDP10 estimates published previously in the context of Swiss national cancer monitoring with dates of reference 31 December 2005 and 31 December 2015. Figure 3 shows the huge increase in prevalent cancer patients over time as a bar chart.

In men, the number of affected cancer patients on 31 December 2025 is expected to be twice that of 31 December 2005, with an average increase of 3.6% per year (AAPC in table 4). In women, there is a 1.8-fold increase in number, or 3.0% AAPC since 2005. Although the numbers involved are small, the largest observed AAPC of 11.4% between 2005 and 2025 concerned prostate cancer in men under 50 years of attained age (table 4). In women, the largest increase was seen for those over 80 years of age with a history of lung cancer (AAPC of 10.2% between 2005 and 2025). For lung cancer, skin melanoma and breast cancer, the age group 80+ expressed the largest AAPC values (between 4.3% and 10.2%). For colorectal and corpus uteri cancer, the youngest age group expressed the highest AAPC (between 2.9% and 4.9%). Increases were also substantial for thyroid cancer at every attained age (between 3.3% and 11.3%). There was, in contrast, very little change over time for cancer of the corpus uteri (between 0.2% and 2.3%), and only moderate increases for colon, rectum and anus combined (between 1.4% and 4.9%).

Figure
Figure 3: Trend in complete cancer prevalence counts of all patients with malignant primary cancer by attained age in Switzerland for the last 20 years. Dates of reference are 31 December 2005, 31 December 2015 and 31 December 2025.

Table 4: Trend in complete cancer prevalence by attained age in Switzerland for the last 20 years in seven major cancers. Dates of reference are 31 December 2005, 31 December 2015 and 31 December 2025. The percentage refers to the relative contribution of each attained age group to the total. The average annual percentage change (AAPC) is a measure of the increase in complete prevalence counts over time.

31 Dec 2005    31 Dec 2015    31 Dec 2025    
   0–4950–5960–6970–7980+All ages0–4950–5960–6970–7980+All ages0–4950–5960–6970–7980+All ages
Colon, rectum, anusMenn555165939735353406915,609760204749227102656621,3981298264352697180734723,737
%3.6%10.6%25.5%34.3%26.1%100.0%3.6%9.6%23.0%33.2%30.7%100.0%5.5%11.1%22.2%30.2%31.0%100.0%
AAPC#------3.2%2.1%2.2%2.9%4.9%3.2%4.3%2.4%1.4%1.5%3.0%2.1%
Women n598158129924652532815,151922206838755892703519,7931549250145696408822523,252
%3.9%10.4%19.7%30.7%35.2%100.0%4.7%10.5%19.6%29.8%35.5%100.0%6.7%10.8%19.6%27.6%35.4%100.0%
AAPC#------4.4%2.7%2.6%2.4%2.8%2.7%4.9%2.3%2.1%1.6%2.2%2.2%
LungMenn1908261510196387853672219202402297917348256335136831413945314711,936
%3.5%15.4%28.1%36.6%16.4%100.0%2.7%11.1%29.1%36.1%21.0%100.0%2.8%11.5%26.3%33.1%26.4%100.0%
AAPC#------1.5%1.1%4.8%4.3%7.0%4.4%2.9%2.6%3.7%3.6%6.6%4.1%
Women n25966288678537929712829681799213710706257456153233213745263211,686
%8.7%22.3%29.8%26.4%12.7%100.0%4.5%15.5%28.7%34.2%17.1%100.0%3.9%13.1%28.4%32.0%22.5%100.0%
AAPC#------0.9%3.9%7.3%10.5%10.9%7.7%2.9%4.3%6.8%8.1%10.2%7.1%
Skin me­la­nomaMenn1678179123392069135092271902297138034510360216,7882263412359897104670426,184
%18.2%19.4%25.3%22.4%14.6%100.0%11.3%17.7%22.7%26.9%21.5%100.0%8.6%15.7%22.9%27.1%25.6%100.0%
AAPC#------1.3%5.2%5.0%8.1%10.3%6.2%1.5%4.3%4.8%6.4%8.3%5.4%
Women n2836241724542300208912,0963132394741484271405619,5543424526365216594623228,034
%23.4%20.0%20.3%19.0%17.3%100.0%16.0%20.2%21.2%21.8%20.7%100.0%12.2%18.8%23.3%23.5%22.2%100.0%
AAPC#------1.0%5.0%5.4%6.4%6.9%4.9%0.9%4.0%5.0%5.4%5.6%4.3%
BreastWomenn533411,61518,54616,54312,31964,357661315,51921,38326,25219,82789,594889317,01926,07329,51728,820110,321
%8.3%18.0%28.8%25.7%19.1%100.0%7.4%17.3%23.9%29.3%22.1%100.0%8.1%15.4%23.6%26.8%26.1%100.0%
AAPC#------2.2%2.9%1.4%4.7%4.9%3.4%2.6%1.9%1.7%2.9%4.3%2.7%
Corpus uteriWomenn301116027353947389812,040366143527914515451313,620477153131864485486914,548
%2.5%9.6%22.7%32.8%32.4%100.0%2.7%10.5%20.5%33.2%33.1%100.0%3.3%10.5%21.9%30.8%33.5%100.0%
AAPC#------2.0%2.2%0.2%1.4%1.5%1.2%2.3%1.4%0.8%0.6%1.1%1.0%
Prosta­teMenn137238510,93618,24812,21043,916256310615,80827,99418,03165,1951190630122,17736,27730,40696,352
%0.3%5.4%24.9%41.6%27.8%100.0%0.4%4.8%24.2%42.9%27.7%100.0%1.2%6.5%23.0%37.7%31.6%100.0%
AAPC#------6.4%2.7%3.8%4.4%4.0%4.0%11.4%5.0%3.6%3.5%4.7%4.0%
ThyroidMenn401290290199107128866980063458327029567779039297374043749
%31.1%22.6%22.5%15.5%8.3%100.0%22.6%27.1%21.4%19.7%9.1%100.0%20.7%24.1%24.8%19.7%10.8%100.0%
AAPC#------5.3%10.7%8.1%11.3%9.7%8.7%3.4%5.8%6.0%6.8%6.9%5.5%
Womenn1372104510017846064808236424292070156494193682605273827532090133511,521
%28.5%21.7%20.8%16.3%12.6%100.0%25.2%25.9%22.1%16.7%10.0%100.0%22.6%23.8%23.9%18.1%11.6%100.0%
AAPC#------5.6%8.8%7.5%7.2%4.5%6.9%3.3%4.9%5.2%5.0%4.0%4.5%
All cancerMenn13,25615,94930,17935,90523,219118,50916,33421,71343,96656,10937,471175,59319,48131,74258,04769,92863,066242,264
%11.2%13.5%25.5%30.3%19.6%100.0%9.3%12.4%25.0%32.0%21.3%100.0%8.0%13.1%24.0%28.9%26.0%100.0%
AAPC#------2.1%3.1%3.8%4.6%4.9%4.0%1.9%3.5%3.3%3.4%5.1%3.6%
Womenn18,04823,53734,09335,17730,190141,04522,82632,93643,79051,81045,393196,75428,29138,07057,87765,82462,975253,036
%12.8%16.7%24.2%24.9%21.4%100.0%11.6%16.7%22.3%26.3%23.1%100.0%11.2%15.0%22.9%26.0%24.9%100.0%
AAPC#------2.4%3.4%2.5%3.9%4.2%3.4%2.3%2.4%2.7%3.2%3.7%3.0%

# Average annual percentage change since date of reference: 31 December 2005.

Discussion

This study estimated the number of individuals living with a previous cancer diagnosis in Switzerland at the end of 2025 and examined the evolution of complete cancer prevalence over the past two decades (2005–2025).

Cancer prevalence is determined by cancer incidence, cancer-specific survival, general life expectancy and the population age structure. While the interpretation of incidence and survival statistics are relatively straightforward, prevalence is inherently more complex. Depending on the reasons why cancer prevalence is high or low, it might indicate either a problem or a success. High cancer prevalence may reflect improved survival and treatment outcomes, increased incidence due to higher exposure to risk factors, demographic changes such as population ageing or a combination of these factors. Consequently, cancer prevalence alone is difficult to interpret aetiologically and is primarily used in health services research to quantify cancer burden for healthcare planning, resource allocation and characterisation of the survivorship population.

We estimated complete cancer prevalence in Switzerland using statistical modelling based on observed cancer incidence and survival data from 1980 to 2021. The completeness index (R index) method has become a key approach for estimating complete prevalence, particularly when cancer registry data cover a limited time span [20]. The R index is a cancer-, sex- and time-specific correction factor that converts limited-duration prevalence into complete prevalence by accounting for diagnoses occurring before the start of registration.

This method is currently the most extensively validated approach for estimating complete cancer prevalence in Europe [25] and the United States [21], notably through large collaborative efforts like the EC-6 project [26]. It is implemented in software tools such as COMPREV [23], promoted by institutions including the US National Cancer Institute and SEER Program [21], facilitating its application without advanced statistical expertise. Despite this, routine implementation of complete prevalence in national cancer reporting remains limited.

Recently, the EC-6 Working Group published complete cancer prevalence estimates for 29 European countries, including Switzerland [1]. Incidence and survival data up to 2013 from 61 national or regional cancer registries (including 4 of the 13 Swiss registries) were pooled to model R indexes, which were then applied to registry-specific limited-duration prevalence estimates extrapolated to a reference date of 1 January 2020. EC-6 R indexes were compared with those derived from pooled US-SEER cancer registries [21] and found to be comparable for most cancer sites, except where known differences in incidence and prognosis exist (e.g. corpus uteri, thyroid, stomach or brain) [26]. The authors concluded that, for many cancer types, applying R indexes derived from foreign populations, such as the pool of European or US-SEER registries, approximates the “gold standard” of country-specific statistical modelling. However, they caution that for cancers with marked geographical heterogeneity in incidence or prognosis, area-specific R indexes may be required, and that the validity of foreign R indexes decreases when applied to limited-duration prevalence intervals shorter than 25 years [26].

We modelled Swiss-specific R indexes because cancer incidence and survival patterns in Switzerland may differ from those of the European pooled population, and because we used R indexes for relative short limited-duration prevalence intervals (LDP10) to maximise the number of contributing Swiss cantons. An additional advantage of modelling R indexes nationally is the full control over cancer reporting group definitions and temporal specificity, both of which are essential for the routine integration of complete prevalence into Swiss cancer surveillance.

Comparison of Swiss-specific and EC-6 modelled R indexes (figure S4 in the appendix) revealed smaller Swiss R indexes for colorectal, lung and prostate cancer. A smaller R index indicates higher incidence in age groups not captured by the LDP10, or better survival, or both. According to the Swiss Cancer Report 2021 [27], 5-year relative survival in Switzerland ranked at least second best when compared with nine other countries in Europe for colorectal and lung cancer, but ranked only eighth for prostate cancer. For prostate cancer, differences in incidence rates might contribute more to the observed difference in R index than differences in survival, because Switzerland has experienced a pronounced increase in prostate cancer since 2011, largely driven by early-stage disease [28].

Approximately 5.5% of the Swiss population in 2025 was estimated to be living with a past cancer diagnosis, a proportion very similar to that of other high-income countries, including Germany (2017: 5.6%) [29], Italy (2014: 5.8%) [30], the United Kingdom (predicted to 2030: 5.8%) [31] and the United States (2022: 5.4%) [21]. This estimate is also similar to the EC-6 estimate for Switzerland in 2020 (5.2%) [1], although the latter included non-malignant brain and bladder tumours.

The number of individuals living with a malignant cancer diagnosis in Switzerland has increased 1.9-fold over the past 20 years, from 259,600 in 2005 to 495,300 in 2025. This increase is slightly lower than the 2.1-fold increase observed in the first Swiss publication of 139,700 in 1990 to 289,800 in 2010 [8]. Overall, this amounts to a 3.5-fold increase between 1990 and 2025, corresponding to an average annual increase of approximately +3.7%. Over the same period, the Swiss population grew only 1.4-fold, from 6,712,300 in 1990 to about 9,058,300 in 2025, equivalent to an annual growth of about +0.86%.

To put these trends into context, we compared them to trends in the US. The SEER Cancer Statistical Report estimated complete prevalence for date of reference 1999 with 8,928,100 patients in a total population of about 267,829,100 [32], and for 2022 with about 18,000,110 patients in a total population of about 333,340,000 [21]. This corresponds to an annual increase in prevalence of 3.1%, compared with a 0.96% annual increase in population. Thus, cancer prevalence has been growing slightly faster in Switzerland than in the US over the past two to three decades. Different age-specific trends in population size and structure, cancer incidence, cancer survival and competing all-cause mortality are the most likely explanatory factors [33].

The female-to-male ratio in the affected population in Switzerland balanced out over time: it was 1.19 in 2005, 1.12 in 2015 and 1.04 in 2025. It is also more balanced as compared with the EC-6 estimate for the pool of European countries: 1.18 in 2020 [1].

A vast majority of affected men and women are long-term and very long-term survivors, with 24% of men and 22% of women diagnosed with cancer 5–10 years ago, and 40% of men and 49% of women diagnosed more than 10 years ago. Long-term cancer survivors have needs that extend beyond medical follow-up and include psychosocial, economic and spiritual dimensions. Addressing these needs requires new models of survivorship care. In Switzerland, consensus-based recommendations for cancer survivorship care were recently published to establish a common framework for structured, high-quality care [3].

Strengths and limitations

This study has several strengths. The underlying data is of high quality, supported by regular quality checks at both cantonal registry and National Agency for Cancer Registration (NACR) levels. Proportions of diagnoses based on information from death certificates were small and case ascertainment completeness sufficient [34]. Swiss R indexes were derived from modelled incidence and survival trends spanning 40 years of observed data. Extensive validation was performed by comparing modelled Swiss R indexes, as well as indexes based on EC-6 incidence and survival data, with Swiss empirical R indexes based on 41-year limited-duration prevalence estimates in cantons with long-standing cancer registration (see supplement figures 2 and 3). Despite wide confidence intervals due to small sample sizes, empirical and modelled R indexes showed strong agreement in almost every case, except for head/neck cancer in men and in lung cancer.

Several limitations should be noted. For certain cancer types, Swiss-specific R indexes could not be modelled, including head and neck (women), oesophagus, gallbladder, pancreas (women), larynx, bone, cervix uteri, kidney (women) and Hodgkin lymphoma (women). As some of these cancers tend to affect younger people, the age group specifications used may not have been optimal. For these cancers, we had to rely on R indexes of the EC-6 study (if the cancer definitions permitted it) or by using an empirical Swiss R index (if the acceptance criteria were met), which may have introduced greater biases. Further methodological work is planned but was beyond the scope of the present study.

Data of some Swiss cantons was unavailable for early diagnosis years, which might have biased the incidence and survival trends used in modelling. Sensitivity analyses comparing incidence and survival trends between cantons covering the whole period 1980 to 2021, and trends derived from all cantons, no matter when registration started, showed only minor differences in incidence rates (see figure S5 in the appendix for major cancer types) and survival proportions (see figure S6 in the appendix for major cancer types). We therefore concluded that bias due to incomplete registration coverage is likely to be small.

Conclusions

This study provides up-to-date information on the number and proportion of cancer patients alive in Switzerland. The number of cancer survivors has substantially increased during past decades. With the baby boomer generation (those born 1946–1964) entering their 8th decade of life, the burden of cancer is bound to increase further and challenge the Swiss healthcare system. Given the huge numbers of long-term cancer survivors, addressing cancer survivors’ needs has become a high priority. Our study can serve to provide regular and systematic prevalence estimates and develop evidence-based strategies for cancer survival in Switzerland.

Data sharing statement

The dataset used in this work is held by the National Agency for Cancer Registration (NACR). The NACR data is not publicly available but can be obtained upon reasonable request. Data in anonymised or aggregated form can be made available to third parties for the purposes of research. Research projects are regulated by the Human Research Act and require, if data is not used in anonymised or aggregated form, informed consent from patients in addition to approval from the relevant ethics committee. Exceptionally, the relevant ethics committee can authorise transmission of non-anonymised data. Data intended for publication must always be anonymised.