Efectiveness of population based colorectal cancer screening programebin shifing cancer stage to earlier disease

of

population-based

colorectal

cancer

screening

iew
ed

Effectiveness

programme in shifting cancer stage to earlier disease

Chi-Ching Law a, Connie H.N. Wong b, Patrick S.K. Chong c, Oscar W.K.

Mang b,*, Albert W.H. Lam c, Michelle M.Y. Chak c, Rachel S.P. Lee b, Keziah

a

pe
er
re
v

N.T. Yip c, Kam-Hung Wong a,b, Rita K.W. Ho c

Department of Clinical Oncology, Queen Elizabeth Hospital, Hong Kong,

China

b

Hong Kong Cancer Registry, Hospital Authority, Hong Kong, China

c

Department of Health, Hong Kong, China

* Corresponding author at: Hong Kong Cancer Registry, Hospital
Authority, Hong Kong.

ot

E-mail address: (Oscar W.K. Mang).

Word count of Abstract: 249

Pr

ep

rin

tn

Word count of manuscript: 2,996

1

This preprint research paper has not been peer reviewed. Electronic copy available at: />

iew
ed

ABSTRACT
Background: This is the first evaluation study to assess the demographic

characteristics of the colorectal cancer (CRC) cases detected in the prevalent
round of the population-based Colorectal Cancer Screening Programme

(CRCSP) in Hong Kong and to explore the effectiveness of the programme on
the stage distribution of CRC.

pe
er
re
v

Methods: This study covered the period between 28 September 2016 and 31

December 2018. Information on CRC diagnosis, age and stage at diagnosis
were retrieved and reviewed by the Hong Kong Cancer Registry (HKCaR). The
CRC detection rate among participants of the CRCSP and incidence rate
among the Hong Kong general population were calculated respectively, along
with the odds ratio (OR) to measure the strength of association and quantify the
effect of CRCSP on stage shift.

Results: The CRC detection rate among participants of the CRCSP during the

ot

study period was 736.0/100,000, whereas the overall CRC incidence rate
among general population of similar age groups was 393.7/100,000. The OR

tn

of 3.91 at stage I dropped to 0.54 at stage IV for all ages and both sexes.
Meanwhile, the OR dropped from 2.24 to 1.62 with increasing age.

rin

Conclusion: The present study has demonstrated the initial impact of the
CRCSP on shifting the stage at diagnosis towards earlier stage. The benefit of

ep

stage-shift was similar for all ages from 60 to 77 in both sexes and seems to
increase with younger age. Given the stage-dependent survival outcomes, this
stage-shift could lead to a reduction in CRC-associated mortality in Hong Kong

Pr

in future.

2

This preprint research paper has not been peer reviewed. Electronic copy available at: />

Colorectal cancer
Population-based screening programme
Staging

Pr

ep

rin

tn

ot

pe
er
re
v

Sex/gender difference

iew
ed

Keywords:

3

This preprint research paper has not been peer reviewed. Electronic copy available at: />

iew
ed

1. Introduction
Colorectal cancer (CRC) is a major public health problem. The Global
Burden of Disease Study showed that the incidence of CRC increased greatly

from 0.83 million in 1990 to 1.83 million in 2017 [1]. In Hong Kong, CRC has

become the commonest cancer and the second ‘cancer killer’. According to the
Hong Kong Cancer Registry, the number of CRC new cases increased from

pe
er
re
v

2,022 in 1990 to 5,634 in 2018, accounting for 16.6% of all new cancer cases
diagnosed in 2018. Meanwhile, the number of deaths attributed to CRC also
increased from 889 in 1990 to 2,314 in 2018, accounting for 15.9% of all cancer
deaths in 2018 [2].

Population-based CRC screening programmes have been introduced in
many countries [3, 4]. Colorectal cancer screening aims to mitigate the risk of
CRC mortality through detection of cancer at earlier stages, and to reduce the

tn

adenomatous polyp.

ot

incidence and mortality of CRC through detection and removal of precancerous

The stage of CRC at diagnosis has a major impact on 5-year survival rates.

rin

Overall survival is more favorable in patients detected by screening
colonoscopy when compared to patients detected after the appearance of

ep

clinical symptoms [5]. Studies evaluating the CRC screening programmes in
Western populations have consistently shown the phenomenon of stage shift

Pr

towards earlier stage in CRC cases detected in screening as compared with
those non-screening diagnosed cases [3, 6]. Moreover, studies have
demonstrated that screening programmes could lead to reduce incidence and

4

This preprint research paper has not been peer reviewed. Electronic copy available at: />

iew
ed

mortality of CRC [7, 8].

In Hong Kong, a government-subsidised, population-based Colorectal

Cancer Screening Programme (CRCSP) has been launched by the Department
of Health (DH) since September 2016. Participants first receive a Faecal

Immunochemical Test (FIT), and a colonoscopy will be arranged if the FIT result

pe
er
re
v

is positive. Participants with negative FIT result will receive FIT re-screening
every two years until the age of 75 [9].

The present study aimed to assess the demographic characteristics of the
CRC cases detected in the prevalent round of the CRCSP in Hong Kong and
to evaluate the impact of the programme on the stage distribution of CRC.

2.1 Study population

ot

2. Material and methods

tn

This was a retrospective cohort study covering the period between 28
September 2016 and 31 December 2018. In this period, the CRCSP was

initiated in different phases: start from September 2016 for Hong Kong

rin

residents born in the years 1946 to 1948, in 2017 for residents born in the years
1946 to 1955, and in 2018 for residents born in the years 1942 to 1957. The

ep

cohort consisted of de-identified tabular data extracted from DH, which included
all participants who enrolled the CRCSP with interpreted FIT results in the

Pr

corresponding periods. As of the end of December 2018, over 115,000 Hong
Kong residents enrolled the CRCSP. Of those FIT-positive participants who
were subsequently diagnosed with CRC, information on CRC diagnosis, date

5

This preprint research paper has not been peer reviewed. Electronic copy available at: />

Kong Cancer Registry (HKCaR).

iew
ed

and stage of disease at diagnosis were retrieved and reviewed by the Hong

Individuals who participated and detected to have CRC in the CRCSP

were compared with those in the rest of the target population. The population
at risk information was obtained from the Hong Kong Census and Statistics

pe
er
re
v

Department and was used to calculate the CRC incidence rate [10]. Annual

counts of the population data by sex and corresponding single years of age (in
aged 68–70 in 2016; aged 62–68 in 2017; and aged 61–62 and 73–76 in 2018)
were selected as the target population. Eligible persons who did not enroll in
the CRCSP were classified as non-screened individuals. All CRC cases
diagnosed outside the CRCSP would be matched with the relevant birth cohorts
by the HKCaR within the same period of time as “HKCaR cohort”. The
demographic data and stage at diagnosis of the two cohorts were evaluated

ot

based on the database of the HKCaR.

tn

Cancer stage classification of the CRC was based on the 7th edition of the
American Joint Committee on Cancer (AJCC) staging system [11]. The stages

rin

at diagnosis were categorised as stage 0, I, II, III, and IV. The stage at diagnosis

was classified as “unknown” if there was inadequate available information for

ep

cancer staging.

Pr

2.2 Statistical analyses
Data analyses for this study were performed using SPSS version 26 for

Window (IBM Corporation). Proportion of age at diagnosis and cancer stage

6

This preprint research paper has not been peer reviewed. Electronic copy available at: />

iew
ed

distribution between CRCSP-detected and HKCaR cohort, and further stratified
by gender would be compared with the Chi-square (χ2) test. All statistical tests
were two-sided and a p-value less than 0.05 was considered statistically
significant. Age-specific CRC detection rates were calculated for each of the

populations: participants of CRCSP and the rest of the target population of
Hong Kong, along with the odds ratio (OR) of CRC among participants of the

pe
er

re
v

CRCSP compared with those not yet participated with 95% confidence intervals
(CIs) to measure the strength of association. CRC detection per 100,000 and
OR with 95% CIs of CRC at different stages would be likewise calculated for

CRCSP-detected and HKCaR cohorts in order to quantify the effect of CRCSP
on stage shift. All results were then stratified by age group and gender.

3. Results

From September 2016 to December 2018, a total of 1,217,800 Hong Kong

ot

residents were eligible for inclusion into the CRCSP. As of 31 December 2018,
a total of 115,223 Hong Kong residents participated in the CRCSP with

tn

interpreted FIT results, constituting around 9.5% of the target population. The
ratio of men to women on CRCSP participants was 1:1.2. As the programme

rin

commenced in late September 2016, there were some participants who were
due for the second round of FIT screening starting from end September 2018.

ep

Among 848 participants detected with CRC by the programme, 17 participants
were detected in the second round. Compared with the rest of the target

Pr

population, there were 4,341 CRC patients detected within the same period of
time,

after

matching

with

relevant

birth

cohorts.

The

demographic

characteristics of the two groups are summarised in Table 1.

7

This preprint research paper has not been peer reviewed. Electronic copy available at: />

iew
ed

For those individuals participated in the CRCSP, the CRC detection rate
during the study period was 736.0/100,000, whereas the overall CRC incidence
rate was 393.7/100,000 for not participated individuals (Table 1). The male-to-

female ratio of CRC detected from the CRCSP and from the HKCaR cohort was
1.8:1 and 1.7:1, respectively. More men were detected with CRC even though

pe
er
re
v

more women participated the programme.

Regarding the distribution of cancer stage, it was observed that earlier
stages (stage 0 to II) were identified from the CRCSP-detected cohort (61.0%),
comparing to those from the HKCaR cohort (41.5%). In particular, 33.8% of
CRC detected in the CRCSP belonged to stage I compared with 16.2% in the
HKCaR cohort. Besides, more advanced stage (stage IV) was found in HKCaR

ot

cohort (23.3%) than in the CRCSP-detected cohort (6.7%).

Table 2 summarises the distribution of stage at diagnosis between men

tn

and women. Both men and women in the CRCSP-detected cohort were
significantly more likely to be diagnosed with earlier stages (stage 0 to II) than

rin

in the HKCaR cohort (men: 59.3% vs 44.3%, p<0.001; women: 64.5% vs 46.6%,
p<0.001), while the percentage of advanced stage (stage IV) in the HKCaR

ep

cohort were more than triple that of CRCSP-detected cohort for both men and

Pr

women (men: 26.2% vs 7.4%; women: 24.0% vs 5.5%).

Stage distribution of CRC across different age groups in both women and

men is summarised in Table 3. The odds of CRC detected by the CRCSP were

8

This preprint research paper has not been peer reviewed. Electronic copy available at: />

iew
ed

88% higher than those CRC diagnosed outside the CRCSP for all ages

between 60 to 77 and both sexes (OR = 1.88, 95% CI = 1.74-2.02). This is

consistent with the prevalent round of CRC screening in detecting individual

with prevalent cancer. The OR of being detected with CRC among different age
groups for both sexes was highest at 2.24 (95% CI = 1.97-2.56) among 60 to
65 year olds, and was 1.64 (95% CI = 1.46-1.83) among 66 to 69 year olds and

pe
er
re
v

1.62 (95% CI = 1.40-1.87) among 70 to 77 year olds.

For men, the OR of being diagnosed with CRC was 2.12 (95% CI = 1.932.32) inside the CRCSP compared with the HKCaR cohort. For women, the
corresponding OR was 1.66 (95% CI = 1.47-1.88) inside the CRCSP compared
with the HKCaR cohort.

Furthermore, the ORs increase with younger age at diagnosis for both

ot

sexes. For men, it ranged from 2.60 (95% CI = 2.20-3.07) among 60 to 65 year
olds to 1.71 (95% CI = 1.42-2.05) among 70 to 77 year olds. For women,

tn

similarly, the highest OR was observed among 60 to 65 year olds (OR = 2.01,
95% CI = 1.62-2.49). As such, CRC detection rate was higher among

rin

participants of CRCSP as compared to the general population with greater

ep

differences observed in younger age group.

For all ages and both sexes, the ORs of being detected with CRC in the

CRCSP-detected cohort decreased with increasing stages, when compared to

Pr

the HKCaR cohort. The OR of 3.91 at stage I dropped to 0.54 at stage IV. The
CRCSP-detected men had 2.96 times increase in chance of being detected

9

This preprint research paper has not been peer reviewed. Electronic copy available at: />

iew
ed

with stage I than those from the HKCaR cohort, whilst, 3.09 times increase in
chance for the CRCSP-detected women. Moreover, male participants of the

CRCSP had 35% decrease in chance of being detected CRC with stage IV than
those from the HKCaR cohort; while, female participants of CRCSP had 58%

decrease in chance of being detected stage IV CRC than those from the

HKCaR cohort. Therefore, it implies the screening programme has worked as

pe
er
re
v

expected and detected more CRC in an earlier stage.

As shown in Table 3, for stage I CRC detected in men, the OR of being
detected CRC by the CRCSP ranged from 3.20 to 4.05 across age groups 6065, 66-69 and 70-77 as compared with the HKCaR cohort. For stage I detected
in women, the OR of being detected CRC by the CRCSP ranged from 3.00 to
5.25 across the three age groups. On the contrary, in the reason of small
number of stage IV CRC patients in CPCSP (n=57), for both sexes, the OR of

ot

being detected stage IV CRC by the CRCSP decreased from 0.65 among 60
to 65 year olds to 0.20 among 70 to 77 year olds as compared with the HKCaR

tn

cohort. As a result, the screening programme was able to detect more CRC in

rin

an earlier stage and younger age.

4. Discussion

ep

The present study has demonstrated the impact of the CRCSP on shifting the
stage at diagnosis towards earlier stage. In the prevalent round of this FIT-

Pr

based CRC screening programme, CRC detected by CRCSP had almost 3
times increase in chance to be detected with stage I than those in the HKCaR
cohort. The CRCSP-detected CRC was 46% decrease in chance to be stage

10

This preprint research paper has not been peer reviewed. Electronic copy available at: />

iew
ed

IV disease than those from the HKCaR cohort. This means when joining the
CRCSP, the participants have more chance to be detected CRC at early stage
so that treatment can be initiated earlier with better survival.

The stage shift found in our study is consistent with that observed in the
prevalent round of bowel cancer screening programmes in other localities. In

pe
er
re

v

the report of the initial impact of Australian faecal occult blood test (FOBT)-

based National Bowel Cancer Screening Program, screening-detected CRC
was diagnosed at an earlier stage than symptomatic CRC (stage I, 40% vs 14%;
stage IV, 3% vs 15%, respectively) [12]. In the nationwide cohort study in
Denmark evaluating the prevalence round of its FIT-based CRC screening
program, the relative risk of being diagnosed at any stage decreased with
increasing stage in invited individuals compared with not yet invited individuals.
The relative risk of being diagnosed with stage I was 3.39, with stage II was

ot

2.16, with stage III was 1.37 and with stage IV was 0.92 [3]. Another casecontrol study in England compared the stage distribution of the CRC diagnosed

tn

in the prevalent round of the national bowel cancer screening programme with
that before the introduction of the programme. A significant shift towards earlier

rin

stage disease was demonstrated in the screened group [6].

ep

Previous studies have suggested gender difference in the effectiveness

of CRC screening. The number needed to screen to detect one CRC is higher

in women than in men, which can be partly explained by the relatively higher

Pr

CRC incidence in men than women at the screening ages [13]. The FOBT/FIT
is less sensitive in women than in men, probably explained by the finding of

11

This preprint research paper has not been peer reviewed. Electronic copy available at: />

iew
ed

significantly less haemoglobin in faeces in women than in men [14-16].

Our study found that the benefit of CRC screening in term of stage shifting

is similar in both women and men. Significant difference in the stage distribution
between CRCSP detected versus non-CRCSP detected CRC was found in

both gender groups. Besides, the degree of benefit in stage shifting towards

pe
er
re
v

earlier stage appears similar across gender. In men, greater proportion of stage
I was observed in CRCSP detected vs non-CRCSP detected CRC. Similarly,

the proportion of stage I was higher in CRCSP detected vs non-CRCSP
detected CRC in women. On the other hand, lower proportion of stage IV in
men was found in CRCSP detected vs non-CRCSP detected CRC. Likewise,
the proportion of stage IV in women was lower in CRCSP detected vs nonCRCSP detected CRC. The stage-shift benefit of the CRCSP is maintained

ot

across different age groups in both women and men.

Stage at diagnosis is the key prognosticator for survival in CRC. In

tn

England, the 5-year relative survival of CRC patients diagnosed during 1996 –
2002 was 93.2% for Dukes A, 77.0% for Dukes B, 47.7% for Dukes C, and only

rin

6.6% for Dukes D [17]. Similarly in the United States, the 5-year relative survival
rates for colon cancer patients diagnosed between 2010 and 2016 was 91% for

ep

localised stage, 72% for regional stage, and only 14% in distant stage [18]. In
Hong Kong, the 5-year relative survival rates of CRC patients that were
diagnosed in 2010-2017 were: 95.7% in stage I, 87.3% in stage II, 68.7% in

Pr

stage III and 9.3% in stage IV [19].

12

This preprint research paper has not been peer reviewed. Electronic copy available at: />

iew
ed

Stage shifting of CRC detected by screening can translate to survival
benefit. A northern Italian study showed that CRC was detected at earlier

stages in screening individuals, with 73% of screen-detected CRC in stage I

and II compared with 43% and 40% of non-screening and pre-screening

detected CRC, respectively. Importantly, CRC detected in screening had better
survival outcome, with significantly lower cumulative 5-year mortality rate in

pe
er
re
v

screening compared to non-screening or pre-screening detected CRC (19%
versus 37% and 41%) [20]. Another colorectal cancer guaiac faecal occult

blood test screening pilot studies in the United Kingdom found a trend towards
earlier stage of CRC detected in screening pilot areas versus non-screening
pilot areas. The mortality rates for CRC were significantly lower in the
populations invited for screening [21].

The treatment intensity of CRC depends considerably on cancer stage at

ot

diagnosis [22, 23]. After colonic resection, adjuvant chemotherapy is not
indicated in stage I, whilst adjuvant chemotherapy is recommended in stage III

tn

with an aim to improve survival. For rectal cancer, early stage I disease can be
treated by surgery alone, whilst more locally advanced disease requires the

rin

addition of preoperative radiotherapy +/- chemotherapy. Unlike earlier stage,
inoperable stage IV CRC is not curable and the treatment intent becomes

ep

palliative. The mainstay treatments for metastatic CRC include chemotherapy
and targeted therapy. Stage shift to early stage obviates the need of adjuvant
chemotherapy or radiotherapy, and the associated treatment-related toxicities

Pr

in colorectal cancer patients.

13

This preprint research paper has not been peer reviewed. Electronic copy available at: />

iew
ed

Treatment cost of CRC increases significantly with increasing stage at
diagnosis. A Norwegian study utilised a model approach to estimate the

expected lifetime treatment cost of CRC for an average 70-year-old patient. The
expected lifetime treatment cost for a 70-year-old CRC patient was found to be

dependent on the stage at diagnosis (CRC stage I €26,630, II €38,130, III

€56,800, and IV €69,890) [24]. Another Spanish cost analysis of CRC treatment

pe
er
re
v

by stage found that the mean long-term cost per patient ranged from €20,708
(in situ) to €47,681 (stage III), with increasing estimated costs with more

advanced stages up to stage III [25]. Systematic reviews have demonstrated
the cost-effectiveness of CRC screening compared with no screening [26, 27].
The modelling study on the National Bowel Cancer Screening Program in
Australia demonstrated that the 2-yearly FIT screening program is highly costeffective, mainly due to the cancer treatment costs averted in the costeffectiveness analysis [28]. In a local study on Chinese population in Hong

ot

Kong, the incremental cost-effectiveness ratio presented $20,542/LYs and
$3155/QALYs gained for annual I-FOBT, and $19,838/LYs gained and

tn

$2976/QALYs gained for biennial I-FOBT compared with no screening [29]. It
is anticipated that the stage-shift found in the present study could contribute to

rin

reduced overall cost and improved cost-effectiveness of treatments for CRC

ep

patients.

The limitation in the present study is the availability of staging information,

which is the main challenge of this study. The stage was unknown in 8.2% of

Pr

the HKCaR cohort compared with 0.4% of the CRCSP-detected cohort. Since
the CRCSP could detect earlier stage CRC patients, these patients would more

14

This preprint research paper has not been peer reviewed. Electronic copy available at: />

iew

ed

likely undergo surgery yielding pathological staging. Hence, there is a risk of
information bias due to the more missing data on overall stage in the HKCaR
cohort compared with the CRCSP-detected cohort.

Secondly, the CRCSP was launched by phase, and has further extended

its coverage to Hong Kong residents aged 50-75 (those born in the years 1944

pe
er
re
v

to 1970) starting from 1 January 2020. Data of CRCSP participants aged 50-59
was not included in this study.

In conclusion, our study has demonstrated the stage-shifting towards
earlier stage at diagnosis of CRC detected in the prevalent round of the CRCSP
in Hong Kong. The benefit of stage-shift was similar in both women and men,
and maintained across different age groups included in this study. Given the
stage-dependent survival outcomes of CRC, it is anticipated that this stage-shift

tn

Sources of funding

ot

could lead to a reduction in CRC-associated mortality in Hong Kong in future.

This research received no specific grant from any funding agency in the

rin

public, commercial, or not-for-profit sectors. The work was supported by the

ep

Department of Health and Hong Kong Cancer Registry.

Ethics approval

Pr

The study was approved by the Department of Health and its Ethics

Committee [LM 77/2021].

15

This preprint research paper has not been peer reviewed. Electronic copy available at: />

Concept or design: Chi-Ching Law.

iew
ed

Author’s contributions

Acquisition of data: Oscar W.K. Mang, Connie H.N. Wong, Keziah N.T.
Yip.

Analysis or interpretation of data: Chi-Ching Law, Connie H.N. Wong,
Oscar W.K. Mang.

pe
er
re
v

Drafting of the manuscript: Chi-Ching Law, Connie H.N. Wong.

Critical revision for important intellectual content: Chi-Ching Law, Patrick
S.K. Chong, Connie H.N. Wong, Oscar W.K. Mang, Albert W.H. Lam, KamHung Wong, Rita K.W. Ho.

Final approval of the manuscript: all authors.

CRediT authorship contribution statement

Chi-Ching Law: Conceptualization, Writing-Original draft preparation,

ot

Writing – Review & Editing. Connie H.N. WONG: Methodology, Validation,

tn

Formal analysis, Writing-Original draft preparation, Writing – Review & Editing.

Patrick S.K. Chong: Writing – Review & Editing. Oscar W.K. Mang:
Methodology, Formal analysis, Resources, Writing – Review & Editing,

rin

Visualization, Project administration. Albert W.H. Lam: Writing – Review &
Editing. Michelle M.Y. Chak: Writing – Review & Editing. Rachel S.P. Lee:

ep

Investigation, Data curation, Validation. Keziah N.T. Yip: Resources. KamHung Wong: Supervision, Writing – Review & Editing. Rita K.W. Ho:

Pr

Supervision, Writing – Review & Editing.

16

This preprint research paper has not been peer reviewed. Electronic copy available at: />

iew
ed

Declaration of Competing Interest
The authors have no conflicts of interest to disclose.

Disclaimer

The contents are solely the responsibility of the authors and do not

pe
er
re
v

necessarily represent the official views of the Department of Health and

Pr

ep

rin

tn

ot

Hospital Authority.

17

This preprint research paper has not been peer reviewed. Electronic copy available at: />

Table 1. Characteristics of target population, n (%)
CRCSP-enrolled
No.
%

Target population
No.

%

9.5

1,102,577

90.5

1,217,800

100

8,938
39,156
38,316
28,813

7.8
34.0
33.3
25.0

154,762
232,544
313,284
401,987

14.0
21.1
28.4

36.5

163,700
271,700
351,600
430,800

13.4
22.3
28.9
35.4

51,867
63,356

45.0
55.0

551,133
551,444

50.0
50.0

603,000
614,800

49.5
50.5

848
114,375

0.7
99.3

4,341
1,098,236

0.4
99.6

5,189
1,212,611

0.4
99.6

Pr

ep

rin

tn

ot

CRCSP, Colorectal Cancer Screening Programme.

iew
ed

115,223

pe
er
re
v

All
Year of Birth
1942 – 1945
1946 – 1949
1950 – 1953
1954 – 1957
Gender
Men
Women
Colorectal cancer
Yes
No

Not yet enrolled
No.
%

18

This preprint research paper has not been peer reviewed. Electronic copy available at: />

Table 2. Distribution of age at diagnosis, sex and stage among those diagnosed with colorectal cancer
No.

%

χ2

p-valueb

30.9
43.8
25.4

1,518
1,690
1,133

35.0
38.9
26.1

7.73

0.021

540
308

63.7

36.3

2,725
1,616

62.8
37.2

0.25

0.617

48
287
182
271
57
3

5.7
33.8
21.5
32.0
6.7
0.4

182
704
914
1,172

1,012
357

4.2
16.2
21.1
27.0
23.3
8.2

203.85

<0.001

162
242
136

30.0
44.8
25.2

956
1,051
718

35.1
38.6
26.3

8.07

0.018

29
166
124
179
40

5.4
30.9
23.0
33.3
7.4

109
446
555
737
656

4.4
17.8
22.2
29.4
26.2

108.94

<0.001

100
129
79

32.5
41.9
25.6

562
639
415

34.8
39.5
25.7

0.76

0.685

6.2
39.4
18.9
30.0
5.5

73
258

359
435
356

4.9
17.4
24.2
29.4
24.0

103.68

<0.001

ot

19
121
58
92
17

pe
er
re
v

262
371
215

tn

All
Age at diagnosis
60 – 65
66 – 69
70 – 77
Gender
Men
Women
Stage
0
I
II
III
IV
Unknownc
Men
Age at diagnosis
60 – 65
66 – 69
70 – 77
Stage
0
I
II
III
IV
Women

Age at diagnosis
60 – 65
66 – 69
70 – 77
Stage
0
I
II
III
IV

HKCaRa

iew
ed

CRCSP-detected
No.
%

Pr

ep

rin

CRCSP, Colorectal Cancer Screening Programme; HKCaR, Hong Kong Cancer Registry.
a
Excluded those detected from CRCSP in the HKCaR cohort.
b

Chi-square test (χ2) was performed with significant level of 0.05. All statistical tests were two-sided.
c
Unknown cases were excluded in statistical analysis.

19

This preprint research paper has not been peer reviewed. Electronic copy available at: />

d
e

Table 3. Number of colorectal cancer detected per 100,000, number of colorectal cancer incidence per 100,000 and odds ratio with 95% confidence intervals of
colorectal cancer at different stages for CRCSP-detected vs HKCaR cohorts by gender and age group

All
Total
Age at diagnosis
60 – 65
66 – 69
70 – 77
Men
Total
Age at diagnosis
60 – 65
66 – 69
70 – 77
Women
Total

66 – 69

r
P

70 – 77

Stage I:
CRCSP-detected /
HKCaRa
OR (95% CI)

Stage II:
CRCSP-detected /
HKCaRa
OR (95% CI)

Stage III:
CRCSP-detected /
HKCaRa
OR (95% CI)

Stage IV:
CRCSP-detected /
HKCaRa
OR (95% CI)

736.0 / 393.7
1.88 (1.74-2.02)

249.1 / 63.9

3.91 (3.41-4.48)

158.0 / 82.9
1.91 (1.63-2.24)

235.2 / 106.3
2.22 (1.94-2.53)

49.5 / 91.8
0.54 (0.41-0.70)

701.6 / 313.9
2.24 (1.97-2.56)
921.6 / 564.9
1.64 (1.46-1.83)
571.5 / 354.3
1.62 (1.40-1.87)

216.9 / 50.4
4.31 (3.35-5.54)
308.0 / 87.6
3.52 (2.85-4.37)
218.0 / 61.9
3.53 (2.72-4.56)

133.9 / 68.4
1.96 (1.45-2.64)
201.2 / 117.3
1.72 (1.35-2.19)
135.6 / 72.6

1.87 (1.38-2.53)

254.4 / 82.1
3.10 (2.48-3.88)
283.2 / 158.8
1.79 (1.46-2.19)
164.8 / 93.8
1.76 (1.34-2.31)

48.2 / 74.2
0.65 (0.40-1.04)
82.0 / 131.7
0.62 (0.44-0.89)
15.9 / 81.0
0.20 (0.09-0.44)

1,041.1 / 494.4
2.12 (1.93-2.32)

320.0 / 80.9
3.96 (3.32-4.74)

239.1 / 100.7
2.38 (1.96-2.89)

345.1 / 133.7
2.59 (2.20-3.05)

77.1 / 119.0
0.65 (0.47-0.89)

1,024.0 / 396.7
2.60 (2.20-3.07)
1,332.7 / 703.3
1.91 (1.66-2.20)
760.2 / 446.8
1.71 (1.42-2.05)

259.2 / 65.6
3.96 (2.81-5.59)
451.6 / 111.7
4.05 (3.11-5.28)
240.4 / 75.3
3.20 (2.26-4.53)

221.2 / 81.7
2.71 (1.89-3.88)
302.9 / 142.5
2.13 (1.58-2.86)
190.1 / 90.2
2.11 (1.45-3.06)

391.9 / 102.1
3.85 (2.91-5.09)
396.5 / 198.7
2.00 (1.54-2.59)
251.6 / 120.7
2.09 (1.51-2.89)

88.5 / 98.8

0.90 (0.52-1.54)
121.2 / 166.0
0.73 (0.47-1.13)
22.4 / 105.8
0.21 (0.08-0.57)

191.0 / 46.8
4.09 (3.29-5.07)

91.5 / 65.1
1.41 (1.07-1.86)

145.2 / 78.9
1.84 (1.47-2.31)

26.8 / 64.6
0.42 (0.26-0.68)

185.8 / 35.4
5.25 (3.61-7.64)
190.1 / 63.5
3.00 (2.09-4.31)
197.6 / 48.4
4.09 (2.78-6.01)

69.7 / 55.2
1.26 (0.74-2.15)
117.7 / 92.2
1.28 (0.84-1.94)
86.1 / 54.7

1.58 (0.94-2.65)

153.3 / 62.2
2.47 (1.69-3.59)
190.1 / 118.9
1.60 (1.14-2.24)
86.1 / 66.6
1.29 (0.77-2.16)

18.6 / 49.9
0.37 (0.14-1.01)
49.8 / 97.5
0.51 (0.28-0.94)
10.1 / 56.0
0.18 (0.04-0.74)

464.6 / 231.6
2.01 (1.62-2.49)
583.8 / 426.8
1.37 (1.13-1.66)
400.3 / 260.9
1.54 (1.21-1.96)

p
e

t
o

n

t

ir n

486.1 / 293.0
1.66 (1.47-1.88)

Age at diagnosis
60 – 65

w
e
i
v

All stages:
CRCSP-detected /
HKCaRa
OR (95% CI)

r
e
e
p

e
r

CRCSP, Colorectal Cancer Screening Programme; HKCaR, Hong Kong Cancer Registry; OR, odds ratio; CI, confidence interval.
a Excluded those detected from CRCSP in the HKCaR cohort.

This preprint research paper has not been peer reviewed. Electronic copy available at: />
20


iew
ed

References

[1]

GBD 2017 Colorectal Cancer Collaborators, The global, regional, and national
burden of colorectal cancer and its attributable risk factors in 195 countries and

territories, 1990-2017: a systematic analysis for the Global Burden of Disease

[2]

pe
er
re
v

Study 2017, The Lancet. Gastroenterology & Hepatology. 4(12) (2019) 913-933.
Hong Kong Cancer Registry, Hong Kong Hospital Authority, Hong Kong Cancer
Statistics (2018), 2021. (accessed May 21,
2021.
[3]

M.B. Larsen, S. Njor, P. Ingeholm, B. Andersen, Effectiveness of Colorectal
Cancer Screening in Detecting Earlier-Stage Disease—A Nationwide Cohort

ot

Study in Denmark, Gastroenterology. 155(1) (2018) 99-106.
[4]

C.H. Kubisch, A. Crispin, U. Mansmann, B. Göke, F.T. Kolligs, Screening for

tn

Colorectal Cancer Is Associated With Lower Disease Stage: A Population-Based

[5]

rin

Study, Clinical Gastroenterology and Hepatology. 14(11) (2016) 1612-1618.
A. Wiegering, S. Ackermann, J. Riegel, U.A. Dietz, O. Götze, C.-T. Germer, I.

ep

Klein, Improved survival of patients with colon cancer detected by screening
colonoscopy, International Journal of Colorectal Disease. 31(5) (2016) 1039-

Pr

1045.

[6]

P. Ellul, E. Fogden, C.L. Simpson, C.L.R. Nickerson, B.C. McKaig, E.T.
Swarbrick, A.M. Veitch, Downstaging of colorectal cancer by the national bowel
21

This preprint research paper has not been peer reviewed. Electronic copy available at: />

iew
ed

cancer screening programme in England: first round data from the first centre,
Colorectal Disease. 12(5) (2010) 420-422.
[7]

E.M. Montminy, A. Jang, M. Conner, J.J. Karlitz, Screening for Colorectal Cancer,
The Medical Clinics of North America. 104(6) (2020) 1023-1036.

R.C. Chen, H.M. Chiu, Y.W. Wang, J.C.Y. Fann, A.M.F. Yen, S.Y.H. Chiu, S.L.S.

pe
er
re
v

[8]

Chen, Y.C. Lee, C.Y. Wu, Y.Y. Jou, H.H. Chen, Reduction of Colorectal Cancer
Mortality and Advanced Stage Cancer Incidence After 10 Years of Fecal
Immunochemical Test Screening, Journal of Global Oncology. (2018 (Suppl 2))

206s.
[9]

Department of Health, The Government of the Hong Kong Special Administrative
Colorectal

Cancer

Screening

Programme,

2016.

ot

Region,

(accessed May 21, 2021.

tn

[10] Census and Statistics Department, The Governmet of the Hong Kong Special
Region,

Population

by

Sex

and

Age,

2021.

rin

Administrative

(accessed May 21, 2021.

ep

[11] Edge SB, Byrd DR, Compton CC, Fritz AG, Greene FL, Trotti A, AJCC Cancer
Staging Manual, 7th ed., Springer, NY, 2010.

Pr

[12] S.S. Ananda, S.J. McLaughlin, F. Chen, I.P. Hayes, A.A. Hunter, I.J. Skinner,
M.C.A. Steel, I.T. Jones, I.A. Hastie, N.A. Rieger, S. Shedda, D.J. Compston, P.
Gibbs, Initial impact of Australia's National Bowel Cancer Screening Program,
22

This preprint research paper has not been peer reviewed. Electronic copy available at: />

iew
ed

Medical Journal of Australia. 191(7) (2009) 378-381.

[13] A.J. Quyn, C.G. Fraser, G. Stanners, F.A. Carey, C. Carden, A. Shaukat, R.J.C.
Steele, Uptake trends in the Scottish Bowel Screening Programme and the

influences of age, sex, and deprivation, Journal of Medical Screening. 25(1)

pe
er
re
v

(2018) 24-31.

[14] R.J.C. Steele, P. McClements, C. Watling, G. Libby, D. Weller, D.H. Brewster, R.
Black, F.A. Carey, C.G. Fraser, Interval cancers in a FOBT-based colorectal
cancer population screening programme: implications for stage, gender and
tumour site, Gut. 61(4) (2012) 576-581.

[15] P.J. McDonald, J.A. Strachan, J. Digby, R.J.C. Steele, C.G. Fraser, Faecal

ot

haemoglobin concentrations by gender and age: implications for populationbased screening for colorectal cancer, Clinical Chemistry & Laboratory Medicine.

tn

50(5) (2012) 935-940.

rin

[16] E. Arana-Arri, I. Idigoras, B. Uranga, R. Pérez, A. Irurzun, I. Gutiérrez-Ibarluzea,
C.G. Fraser, I. Portillo, E. Group, Population-based colorectal cancer screening

ep

programmes using a faecal immunochemical test: should faecal haemoglobin
cut-offs differ by age and sex?, BMC Cancer. 17(1) (2017) 1-13.

Pr

[17] National Cancer Registration and Analysis Service, London, National Cancer
Registration and Analysis Service, Public Health England. Colorectal Cancer
Survival

by

Stage

-

NCIN

Data

Briefing,

2010.
23

This preprint research paper has not been peer reviewed. Electronic copy available at: />

y_stage. (accessed September 6, 2020.

iew
ed

/>
[18] American Cancer Society, Survival Rates for Colorectal Cancer, 2021.
/>
pe
er
re
v

staging/survival-rates.html. (accessed May 21, 2021.

[19] Hong Kong Cancer Registry, Hong Kong Hospital Authority, The First Report of
Stage-specific Survival of Breast and Colorectal Cancers in Hong Kong, 2020.
(accessed December 8, 2020.

[20] F. Parente, C. Vailati, C. Boemo, E. Bonoldi, A. Ardizzoia, A. Ilardo, F. Tortorella,
D. Cereda, M. Cremaschini, R. Moretti, Improved 5-year survival of patients with

ot

immunochemical faecal blood test-screen-detected colorectal cancer versus

68-72.

tn

non-screening cancers in northern Italy, Digestive & Liver Disease. 47(1) (2015)

rin

[21] P.L. McClements, V. Madurasinghe, C.S. Thomson, C.G. Fraser, F.A. Carey,
R.J.C. Steele, G. Lawrence, D.H. Brewster, Impact of the UK colorectal cancer

ep

screening pilot studies on incidence, stage distribution and mortality trends,
Cancer Epidemiology. 36(4) (2012) e232-e242.

Pr

[22] G. Argilés, J. Tabernero, R. Labianca, D. Hochhauser, R. Salazar, T. Iveson, P.
Laurent-Puig, P. Quirke, T. Yoshino, J. Taieb, E. Martinelli, D. Arnold, Localised
colon cancer: ESMO Clinical Practice Guidelines for diagnosis, treatment and
24

This preprint research paper has not been peer reviewed. Electronic copy available at: />

iew
ed

follow-up, Annals of Oncology. 31(10) (2020) 1291-1305.

[23] R. Glynne-Jones, L. Wyrwicz, E. Tiret, G. Brown, C. Rödel, A. Cervantes, D.

Arnold, Rectal cancer: ESMO Clinical Practice Guidelines for diagnosis,
treatment and follow-up, Annals of Oncology. 28(suppl 4) (2017) iv22-iv40.

pe
er
re
v

[24] P. Joranger, A. Nesbakken, H. Sorbye, G. Hoff, A. Oshaug, E. Aas, Survival and
costs of colorectal cancer treatment and effects of changing treatment strategies:
a model approach, The European Journal of Health Economics. 21(3) (2020)
321-334.

[25] J. Corral, X. Castells, E. Molins, P. Chiarello, J.M. Borras, F. Cots, Long-term
costs of colorectal cancer treatment in Spain, BMC Health Services Research.

ot

16 (2016) 1-8.

[26] S.S. Patel, M.L. Kilgore, Cost Effectiveness of Colorectal Cancer Screening

tn

Strategies, Cancer Control. 22(2) (2015) 248-258.

rin

[27] T. Ran, C.Y. Cheng, B. Misselwitz, H. Brenner, J. Ubels, M. Schlander, CostEffectiveness of Colorectal Cancer Screening Strategies—A Systematic Review,

ep

Clinical Gastroenterology and Hepatology. 17(10) (2019) 1969-1981.

[28] J.B. Lew, D.J.B. St John, X.M. Xu, M.J.E. Greuter, M. Caruana, D.R. Cenin, E.

Pr

He, M. Saville, P. Grogan, V.M.H. Coupé, K. Canfell, Long-term evaluation of
benefits, harms, and cost-effectiveness of the National Bowel Cancer Screening
Program in Australia: a modelling study, The Lancet Public Health. 2(7) (2017)
25

This preprint research paper has not been peer reviewed. Electronic copy available at: />