Below is an unofficial translation of a document submitted by Dr. Shoichiro Tsugane from the National Cancer Center during the fourth session of the Thyroid Examination Assessment Subcommittee of the Prefectural Oversight Committee Meeting for the Fukushima Health Management Survey, held on November 11, 2014. The document outlined the estimated prevalence (or, more precisely, the estimated number of prevalent cases) of pediatric thyroid cancer in Fukushima Prefecture, calculated by Dr. Kota Katanoda, a staff member for National Cancer Center. This was also published here.
*****
The Estimated Number of Prevalent Cases of Thyroid Cancer in Fukushima
Prefecture
Shoichiro Tsugane (National Cancer Center)
November 11, 2014
Background
During the second session of “Thyroid Examination Assessment Subcommittee” held on March 2, 2014, I commented on a few points regarding evaluation of frequency of thyroid cancer diagnosis as part of the thyroid examination, especially the inappropriateness of comparison with incidence rate data (refer to the handout 4). Considering that frequency of diagnosis during a cross-sectional examination could reflect early diagnosis of thyroid cancer cases which would be clinically diagnosed in the future, it seemed more appropriate to try to compare with the number of thyroid cancer cases estimated with cumulative incidence risk based on incidence data. I requested the calculations be done by the Division of Surveillance (person in charge: Kota Katanoda, Section Head of the Epidemiology and Statistics Section), Center for Cancer Control and Information Services at National Cancer Center.
Methods and Results
See attachment below, The estimated number of prevalent cases of thyroid cancer in Fukushima
Prefecture.
Discussion
- As the participation rate in the thyroid examination is 80%, it is necessary to consider the number of cases in the screened population to be about 80% of the estimate.
- Based on cancer incidence rate from 2001 to 2010 (national estimates), the number of thyroid cancer cases clinically diagnosed by age 18 in Fukushima Prefecture will be 2.1 (male 0.5, female 1.6). In the screened population, the estimate will be 1.7 (male 0.4, female 1.3). (An accurate estimate will require the number of screening participants by age). If all of the 104 (male 36, female 68) cases confirmed or suspicious of malignancy were to be diagnosed with thyroid cancer, the occurrence would be 61 times (male 90, female 52) the estimated number of thyroid cancer cases.
- Assuming there would be no more thyroid cancer cases detected in the screened population in the future (i.e. all future potential thyroid cancer cases were already detected), this screening has detected all the cases of thyroid cancer to be clinically diagnosed by age 35 (age when the number of diagnosed cases will exceed 100). Most of them are estimated to be diagnosed after age 20.
- According to the mortality statistics from the 2011 demographics, probability of dying from thyroid cancer before age 40 (cumulative mortality risk) is 0.00036% (3.6 in 1,000,000) for male and 0.00032% (3.2 in 1,000,000) for female. That is, about 1 in 300,000 of the screening participants. Thus, it is an extremely rare event to die from thyroid cancer even in the absence of early detection due to screening.
- The current situation in Fukushima Prefecture where over 100 cases of thyroid cancer have been diagnosed in those aged 18 or younger is thought to be due to either excess occurrence due to some cause or diagnosis of many (latent) cancers which could not be clinically diagnosed in the future or which might not be fatal (i.e. overdiagnosis). This is difficult to interpret solely on the basis of an additional cases from early diagnosis of thyroid cancer which might be clinically diagnosed one to several years later (i.e. screening effect). Also, the number of thyroid cancer cases which could avoid death due to early diagnosis would be at most one.
- In regards to excess occurrence, unlike in the case of acute infection, it is known that a certain number of years is required for a causative factor to lead to carcinogenesis. Thus it is difficult to interpret that the occurrence rate of thyroid cancer diagnosed up to 2014 was increased due to some factor added after the 2011 accident.
- Meanwhile, there is sufficient probability of overdiagnosis as has definitely been observed in adult thyroid cancer, and also there is a precedence of neuroblastoma mass screening in children (refer to handout from the second session). A scenario (note: shown in the graph below) is anticipated where many of thyroid cancer currently diagnosed would either grow very slowly, remain the same in size, or begin to shrink.
Note: modified from Welch and Black, shown below. Two additional captions are inserted for the
vertical axis: 1) “Screening with high sensitivity” immediately above “Abnormal
cell” and 2) “Detectable with screening” further above.
(A modified version of Figure 1 from Welch and Black).
Conclusion
This thyroid screening examination was implemented with good intentions on the assumption that “More testing brings more comfort,” and “Early diagnosis is good and causes no harm.” However, it is necessary to share a common understanding that screening with high precision/sensitivity conducted in asymptomatic, healthy individuals could bring about many detriments such as overdiagnosis and related treatments and complications, as well as physical and emotional burden as a consequence of the reduced QOL, the confirmatory examination necessitated by false-positive results, and the primary examination itself.
*****
Attachment
The Estimated Number
of Prevalent Cases of Thyroid Cancer in
Fukushima Prefecture in 2010
November 4, 2014
Division of Surveillance
Center for Cancer Control and Information Services
National Cancer Center, Japan
(1) Data
utilized
1. National estimates of thyroid cancer incidence (2001 to 2010)
National estimates of cancer incidence based on cancer registries in Japan (Number of thyroid cancer incidence by sex and age, in 5-year age groups)
ganjoho.jp/data/en/professional/statistics/files/cancer_incidence(1975-2010)E.xls
2. National population (2001-2010)
Population estimates by Ministry of Internal Affairs and Communications (National Census population used for National Census years) (Population by age, sex, in 5-year age groups)
ganjoho.jp/data/en/professional/statistics/files/cancer_incidence(1975-2010)E.xls
3. National mortality from all causes (2001-2010)
Vital statistics (Mortality from all causes by sex, age, in 5-year age groups)
[Volume 3] General mortality Table 1-1 Deaths by causes (the list of three-character categories), sex and age.
http://www.e-stat.go.jp/SG1/estat/eStatTopPortalE.do
4. Age 0 population in Fukushima Prefecture (1970-2010)
Population for each year of age estimated using the birth cohort method from the census population determined every 5 years as well as the number of births. (Total population by sex and each year of age. Persons of unknown age distributed proportionally).
ganjoho.jp/data/professional/statistics/statistics05/files/07_all_1970-2015.csv
(2) Method of estimation
Cumulative incidence risk* of thyroid cancer for each year of age was calculated and multiplied by the yearly age 0 population in Fukushima Prefecture, in order to obtain the number of cumulative incident cases for each year of age. Addition of the obtained number from age 0 to the chosen age will yield a total number of the cumulative incident cases up to the chosen age, which is regarded as the number of prevalent cases. Details are shown in steps ①~③.
*Cumulative incidence risk: probability of developing a
certain disease by a certain age
① Calculation of cumulative incidence risk of thyroid cancer (for 5-year age groups) (Figure 1)
Using items 1-3 in section (1), cumulative incidence risk of thyroid cancer by 5-year age groups is calculated. Population consisting of 100 persons of age 0 is assumed, and cumulative incidence risk is determined by counting the number of cancer incidence that occurred after aging the population by 5 years and subtracting the number of deaths (from all causes). (Lifetime Data Anal. 4: 169-186, 1998)
② Calculation of cumulative incidence risk of thyroid cancer (for each specific year of age)
Spline function is applied to the cumulative incidence risk of thyroid cancer for 5-year age groups, calculated in section ①, to estimate the cumulative incidence risk for each specific year of age.
③ Calculation of the number of thyroid cancer cases
The cumulative incidence risk of thyroid cancer for 5-year age groups, calculated in section ①, is multiplied by the respective yearly age 0 population in Fukushima Prefecture from (1)-4 (age 0 population in 2010 x risk at age 0, age 0 population in 2009 x risk at age 1, and so on), to calculate the number of cumulative incidence at each age. Total of the number of cumulative incidence from age 0 to 18 is regarded as the number of thyroid cancer cases up to age 18.
Using items 1-3 in section (1), cumulative incidence risk of thyroid cancer by 5-year age groups is calculated. Population consisting of 100 persons of age 0 is assumed, and cumulative incidence risk is determined by counting the number of cancer incidence that occurred after aging the population by 5 years and subtracting the number of deaths (from all causes). (Lifetime Data Anal. 4: 169-186, 1998)
② Calculation of cumulative incidence risk of thyroid cancer (for each specific year of age)
Spline function is applied to the cumulative incidence risk of thyroid cancer for 5-year age groups, calculated in section ①, to estimate the cumulative incidence risk for each specific year of age.
③ Calculation of the number of thyroid cancer cases
The cumulative incidence risk of thyroid cancer for 5-year age groups, calculated in section ①, is multiplied by the respective yearly age 0 population in Fukushima Prefecture from (1)-4 (age 0 population in 2010 x risk at age 0, age 0 population in 2009 x risk at age 1, and so on), to calculate the number of cumulative incidence at each age. Total of the number of cumulative incidence from age 0 to 18 is regarded as the number of thyroid cancer cases up to age 18.
(3) Results (Figure 2)
The number of thyroid cancer cases for ages 18 and younger
in Fukushima Prefecture as of 2010 was estimated to be 2.0 (male 0.5, female
1.6). It was estimated that the number of cases of both male and female will
exceed 50 at age 31, and 100 at age 35.
(4) Points to consider
- As thyroid cancer incidence rate below age 20 is low, 10-year average data was used from 2001 to 2010. However, as thyroid cancer incidence rate is showing a tendency for long-term increase, 10-year average incidence rate might be underestimation compared to the current incidence rate.
- The methodology used in this estimation assumes 10-year incidence by age group from 2001 to 2010 is experienced by all generations up to age 40 as of 2010. Along with the fact the incidence rate is showing a tendency for long-term increase, this assumption might overestimate incidence rate as the age increases.
- Spline function was used in calculation of cumulative incidence risk at each year of age here, but another method might be used in (2)-① where the population is aged by 1 year (instead of 5 years).
- As thyroid cancer incidence below age 20 is low, national estimates using population-based cancer registry data might be unstable.
- Although national estimates of cancer incidence were based on population-based cancer registries fulfilling a certain quality standards, some degree of incompleteness in registrations might have led to underestimation.
End
Figure 1 Estimated cumulative incidence risk for
thyroid cancer
Figure 2 Estimated
cumulative prevalence of thyroid cancer in Fukushima Prefecture in 2010.