Influence of a nationwide colorectal cancer screening program on the incidence of synchronous colorectal peritoneal metastases

Abstract

This study aims to assess the influence of colorectal cancer screening on the incidence of synchronous colorectal peritoneal metastases (CPM). Patients diagnosed with CPM between 2009 and 2022 were selected from the Netherlands Cancer Registry. Crude rates of the observed and expected CPM incidence were calculated per 100,000 individuals and compared. Expected incidence was extrapolated from the incidence in the years prior to screening invitation. In total, 9,238 patients with CPM were included. For the screen‐eligible population (55–75 years), the observed CPM incidence increased from 5.1 (2009) to 7.0 before screening initiation (2013) (Annual percent change [APC] 8.2%, p = .014). Since the start of screening, the observed CPM incidence stabilized: 8.8 (2014) to 8.9 (2022) (APC −1.0%, p = .159). Within the total population, the observed CPM incidence before screening showed an increase from 3.6 (2009) to 4.0 (2013) (APC 3.4%, p = .050). Since screening, a decrease in CPM incidence was observed from 4.4 (2014) to 3.5 (2022) (APC −2.2%, p = .010). The observed and expected number of CPM differed significantly in the screen‐eligible population (6,437 observed vs. 7,992 expected individuals; p < .001) and in the overall Dutch population (9,238 observed vs. 10,440 expected individuals; p < .001). In conclusion, a declining trend was observed in the incidence of CPM in the Dutch population since the start of colorectal cancer screening. The observed incidence was lower compared to the expected incidence, both in the screen‐eligible population and in the overall population. These findings suggest that screening results in a decrease of patients diagnosed with CPM possibly resulting in an improved survival of colorectal cancer patients.

What's new?

Colorectal cancer screening has been shown to decrease the incidence of the disease and allow detection at an earlier stage. Here, the authors also show that, since the introduction of the colorectal cancer screening program in the Netherlands in 2014, the incidence of synchronous peritoneal metastases in colorectal cancer patients has stabilized in the screening‐eligible population, despite the expectation of a further increase. Moreover, the incidence of colorectal peritoneal metastases in the total Dutch population shows a decreasing trend. These results suggest screening as a means for both improving early diagnosis and preventing advanced disease.

1. INTRODUCTION

Colorectal cancer (CRC) is the third leading cause of cancer and the second deadliest with almost 1 million casualties per year worldwide. CRC is notorious for its propensity to metastasize, of which the peritoneum is one of the most frequently affected sites. ¹ , ² Approximately 20% of patients with CRC present with metastases at the time of initial diagnosis ³ with colorectal peritoneal metastases (CPM) being associated with the poorest prognosis. ⁴ , ⁵ Approximately 10% of patients with CRC will be diagnosed with CPM during the course of their disease. Unfortunately, treatment options with curative intent are very limited up to this date. ⁴

To combat the current disease burden of CRC, screening was introduced. Twenty‐four countries in the European Union had initiated or were establishing screening programs by 2015. ⁶ The fecal immunochemical test (FIT) is the most commonly used screening tool. ⁶ FIT screens for the presence of hemoglobin in collected stool samples which may be associated with (occult) blood loss in the gastrointestinal tract. ⁷ In the Netherlands, FIT‐based screening has been conducted biennially since 2014, including individuals aged 55–75 years. ⁸ The screening program in the Netherlands was gradually introduced by dividing the population into cohorts, based on their year of birth. Inclusion of all eligible patients was reached in 2019 and has been repeated ever since. ⁹

Screening for CRC primarily aims to detect asymptomatic premalignant colorectal polyps. Subsequent removal of these screen‐detected polyps has been shown to effectively decrease the incidence of CRC. ¹⁰ , ¹¹ Besides, screening also detects CRC at an earlier stage which may result in a more favorable outcome. ¹² The effect of screening on the incidence of CPM has never been investigated before. As CPM is a highly lethal manifestation of CRC, prevention of the occurrence of CPM would be highly desirable. This nationwide population‐based study aimed to investigate the effect of screening on the incidence of synchronous CPM both in the screen‐eligible population (aged 55–75 years) as well as in the total Dutch population.

2. METHODS

2.1. Data sources

For this nationwide cohort study, data was extracted from the Netherlands Cancer Registry (NCR). The NCR contains all diagnosed malignancies, which are registered by trained, hospital‐independent data managers. This database contains, amongst others, information about topography, morphology, and classification of the tumor. The International Classification of Disease–Oncology was used to classify data of anatomical site and morphology of the primary tumor and its metastases. ¹³ To register the tumor and nodular stages of each diagnosed tumor, the Tumor Node Metastasis classification that was valid at the time of diagnosis was used for staging the primary tumor. ¹⁴ The clinical tumor and nodular stages were used when there was no information available on the pathological stage.

Data from the Central Bureau of Statistics was used to identify the absolute population per cohort from 2009 to 2022.

2.2. Study population and definitions

All patients diagnosed with CPM between January 2009 and December 2022 were included. Patients who were diagnosed with an appendiceal tumor, a neuro‐endocrine tumor, or a tumor with a different histology than adenocarcinoma, were excluded. Additionally, tumors that were left in situ, and did neither have a pathological nor a clinical Tumor Node Metastasis stage were also not considered for further analysis.

According to the International Classification of Disease–Oncology classification, the included primary tumors were divided into four categories, (1) right‐sided colon (C18.0, C18.2, C18.3, C18.4), (2) left‐sided colon (C18.5, C18.6, C18.7), (3) colon, not otherwise specified (C18.8, C18.9), and (4) rectum (C19.9, C20.9). If any of the following codes were registered under metastasis for the included patients, it was considered to be peritoneal metastases (PM): C48.0, C48.1, C48.2, C48.4, C48.8, C16.0, C16.1, C16.2, C16.3, C16.5, C16.6, C16.8, C16.9, C17.0, C17.1, C17.2, C17.3, C17.8, C17.9, C18.0, C18.1, C18.2, C18.3, C18.4, C18.6, C18.7, C18.8, C18.9, C19.9, C20.9, C21.8, C23.9, C26.9, C48.0, C48.1, C48.2, C48.4, C48.8, C49.4, C49.5, C52.9, C54.3, C54.8, C54.9, C55.9, C56.9, C57.0, C57.1, C57.2, C57.3, C57.4, C57.8, C66.9, C67.0, C67.1, C67.4, C67.8, C67.9, C76.2. Metastases registered at other locations than listed above were considered systemic metastases.

Patients with PM were subdivided into cohorts, comprising the invited birth years in the Dutch national screening program. An overview of the newly invited birth years and their corresponding cohorts is displayed in Table 1. Thus, all cohorts together contain the total screen‐eligible population, aged 55–75 years old. Individuals aged 18–54 years and patients aged ≥76 years were not included in the screening program. Patients diagnosed with CPM from these age categories are therefore referred to as non‐screen eligible patients. An additional overview of the CPM incidence in the total Dutch population was provided as well, which consists of the screen‐eligible population and the non‐screen‐eligible population.

TABLE 1.

Division of cohorts for colorectal screening in the Netherlands, based on the first year of invitation per year of birth.

Cohort	Birth year
2014	1938, 1939, 1947, 1949, 1951
2015	1940, 1946, 1948, 1950, 1952, 1954
2016	1941, 1945, 1953, 1955, 1957
2017	1942, 1944, 1956, 1958, 1960
2018	1943, 1959, 1961, 1963
2019	1962, 1964
2020	1965
2021	1966
2022	1967

2.3. Statistical analyses

For all cohorts, crude and cumulative incidence rates were calculated. Crude incidence was calculated by dividing the number of individuals diagnosed with CPM per cohort, by the total population in the Netherlands in the corresponding calendar year. The annual incidence of CPM was displayed per 100,000 individuals. For the entire Dutch population and for each screening cohort, trend lines were generated based on the CPM incidence in the years prior to the start of national screening program (2009–2013). The trend lines were created by fitting a linear regression line using the natural logarithm of the incidence rate with the years prior to screening initiation as regression variable.

The observed incidence of CPM was calculated for the entire Dutch population for all cohorts and for the non‐screen eligible population from 2009 to 2022. This was achieved by dividing the observed number of CPM patients per cohort by the number of individuals at risk per cohort in 2009. Following, the expected incidence was calculated for CPM incidence by dividing the amount of expected CPM patients per cohort by the number of individuals at risk per cohort in 2009. The expected CPM incidence was based on the generated trend lines.

The annual percent change (APC) was calculated for the screen‐eligible patients and the entire Dutch population in the years prior to screening (2009–2013) and for the years after screening initiation (2014–2022).

Finally, the observed and expected incidence of CPM were compared using Chi‐square tests. Calculated p‐values were two‐sided and were considered statistically significant if <.05. All statistical analyses were performed with SAS statistical software (SAS system 9.4, SAS Institute, Cary, NC, USA).

3. RESULTS

3.1. Study population

Between 2009 and 2022, 178,460 patients were diagnosed with CRC in the Netherlands, of whom 9,238 (5%) were diagnosed with synchronous CPM. Of the patients with CPM, 6,437 (70%) patients were screen‐eligible, and 2,801 (30%) patients were not eligible for screening based on their age. Baseline characteristics of the total CRC population, the screen‐eligible population, and the non‐screen‐eligible population are presented in Table 2. In general, most patients were male (56%) and had right‐sided colon cancer (34%) with non‐mucinous, non‐signet ring cell adenocarcinoma as histology (91%). The majority of patients were diagnosed after the start of the screening program (65%).

TABLE 2.

Baseline characteristics of all CRC patients diagnosed from 2009 to 2022 and the study population, patients diagnosed with CPM in 2009–2022.

	Total population of CRC patients (n = 178,460)	Synchronous CPM, screen‐eligible (n = 6,437)	Synchronous CPM, non‐eligible for screening (n = 2,801)
Mean age (years ± SD)	69 ± 11	n.a.	73 ± 17
< 55	n.a.	n.a.	43 ± 7
55–75	n.a.	66 ± 6	n.a.
> 75	n.a.	n.a.	83 ± 4

	n (%)	n (%)	n (%)
Sex
Male	99,529 (56)	3,326 (52)	1,289 (46)
Female	78,931 (44)	3,111 (48)	1,512 (54)
Primary tumor location
Right‐sided colon	60,537 (34)	3,300 (51)	1,355 (48)
Left‐sided colon	58,417 (33)	2,053 (32)	937 (33)
Rectum	56,414 (32)	743 (12)	328 (18)
Colon, NOS	3,092 (2)	341 (5)	181 (6)
Tumor histology
Adenocarcinoma	161,897 (91)	5,014 (78)	2,143 (77)
Mucinous adenocarcinoma	14,604 (8)	1,018 (16)	463 (17)
Signet ring cell carcinoma	1,959 (1)	405 (6)	195 (7)
Tumor differentiation
Good/moderate	126,986 (88)	2,621 (65)	1,009 (63)
Poor/undifferentiated	18,039 (12)	1,425 (35)	592 (37)
Unknown	33,435	2,391	1,200
Tumor stage
T0‐3	141,461 (85)	2,170 (43)	838 (41)
T4	25,179 (15)	2,838 (57)	1,222 (59)
Unknown	11,820	1,429	741
Nodal stage
N0	99,720 (60)	1,451 (26)	582 (26)
N1/N2	67,336 (40)	4,056 (74)	1,646 (74)
Unknown	11,404	930	568
Time of diagnosis
2009–2013	61,606 (34)	1,928 (30)	1,289 (46)
2014–2022	116,854 (65)	4,509 (70)	1,512 (54)
Peritoneal metastases
Yes	9,238 (5)	n.a.	n.a.
No	169,222 (95)	n.a.	n.a.

Abbreviations: CRC, colorectal cancer; CPM, colorectal peritoneal metastases; n.a., not applicable; SD, standard deviation.

3.2. Incidence of synchronous CPM in the screen‐eligible patients

Figure 1 provides an overview of CPM in the group of screen‐eligible patients in the Netherlands between 2009 and 2022. In this group, an increase in CPM incidence is observed in the period from 2009 with an incidence of 5.1 per 100,000 to 7.0 per 100,000 in 2013 (APC of 8.2%, p = .014). From the start of the screening in 2014 onward, an overall stable trend is observed in the incidence of CPM (APC of −1.0%, p = .159). Incidence in 2014 was 8.8 per 100,000 individuals and 8.9 per 100,000 individuals in 2022. Based on the observed incidence of CPM from 2009 to 2013, the incidence was expected to rise from 2014 onward with an expected incidence of 7.9 per 100,000 individuals in 2014 and 16.3 per 100,000 individuals in 2022.

FIGURE 1.

Incidence of CPM in the Dutch population eligible for screening (55–75 years old), expressed per 100,000 individuals. CPM, colorectal peritoneal metastases.

Table 3 provides an overview of the observed and expected cumulative CPM incidence in the Netherlands, with a subdivision of the screen‐eligible population into screening cohorts, which is based on the first invitation year per year of birth (Table 1). The observed and expected CPM incidence differed significantly in the screen‐eligible population (1938–1967) with an observed incidence of 6,437 individuals versus 7,993 expected individuals (p < .001). Stratified by screening cohorts, based on the first year of invitation to participate in screening, cohorts 2014– 2018 show a significant difference in observed and expected CPM incidence. Cohorts 2019–2022 do not differ significantly in observed and expected CPM incidence (Table 3). Figure 2 gives a visual representation of the observed and expected incidence of all screening cohorts.

TABLE 3.

Observed and expected cumulative CPM incidence in the Netherlands in the period 2009–2022.

Invitation cohort (birth years)	Number of individuals at risk (2009)	Observed amount of CPM 2009–2022	Expected amount of CPM 2009–2022	Observed cumulative CPM incidence	Expected cumulative CPM incidence	p‐Value
	n	n	n	%	%
Total Dutch population	16,485,787	9,238	10,441	0.056	0.06	<.001
Total screening eligible population (1938–1967)	5,911,802	6,437	7,993	0.11	0.14	<.001
2014 (1938, 1939, 1947, 1949, 1951)	930,513	1318	1,871	0.14	0.20	<.001
2015 (1940, 1946, 1948, 1950, 1952, 1954)	761,681	1,647	2,455	0.22	0.32	<.001
2016 (1941, 1945, 1953, 1955, 1957)	998,339	1,131	1,771	0.11	0.18	<.001
2017 (1942, 1944, 1956, 1958, 1960)	1,019,630	1,126	1,487	0.11	0.15	<.001
2018 (1943, 1959, 1961, 1963)	899,983	696	814	0.08	0.09	.002
2019 (1962, 1964)	517,651	244	268	0.05	0.05	.289
2020 (1965)	266,112	98	89	0.04	0.03	.510
2021 (1966)	261,342	102	92	0.04	0.04	.473
2022 (1967)	256,551	75	65	0.03	0.03	.448

Note: The bold values are the values that are statistically significant. The other p‐values do not prove a statistical significance.

Abbreviation: CPM, colorectal peritoneal metastases.

FIGURE 2.

Incidence CPM per screening cohort, expressed per 100,000 individuals. (A) Incidence of CPM for cohort 2014–2017 per 100,000 individuals. (B) Incidence of CPM for cohort 2018–2022 per 100,000 individuals. CPM, colorectal peritoneal metastases.

3.3. Incidence of synchronous CPM in the total population

Figure 3 represents the observed incidence of CPM incidence in the entire Dutch population, per 100,000 individuals. An initial increasing, non‐significant, trend is observed from 2009 until 2013, with an observed incidence of 3.6 per 100,000 individuals in 2009 compared to an incidence of 4.0 per 100,000 individuals in 2013 (APC of 3.4%, p = .050). This increase is followed by a significant decrease in the period of 2014–2022, with an observed incidence of 4.4 per 100,000 individuals in 2014 and an observed incidence of 3.5 per 100,000 individuals in 2022 (APC of −2.2%, p = .010). Based on the incidence of 2009–2013, the incidence was expected to rise with an expected incidence of 4.2 per 100,000 individuals in 2014 to 5.2 patients per 100,000 individuals in 2022. In the total Dutch population, 9238 individuals with CPM were diagnosed versus an expected incidence of 10,441 individuals with CPM (Table 3, p < .001).

FIGURE 3.

Incidence of CPM in the entire Dutch population, expressed per 100,000 individuals. CPM, colorectal peritoneal metastases.

4. DISCUSSION

Since the introduction of CRC screening in the Netherlands in 2014, the incidence of synchronous PM in CRC patients has stabilized in the screen‐eligible population whereas a further increase was anticipated. The incidence of CPM in the total Dutch population shows a decreasing trend since screening has been introduced.

CRC is one of the most common and deadliest malignancies worldwide and it was expected that its incidence would further rise if no action was undertaken. ¹⁵ Worldwide, several countries have been setting up screening programs to alleviate this rising burden of CRC. At this moment, CRC screening programs have been proven to be cost‐effective, to lower the disease burden, and to decrease morbidity and mortality of CRC. ⁶ , ¹¹ , ¹⁶ Results displayed by the present study suggest that the introduction of a countrywide screening program using FIT, complementary to its known effects in early disease stages, also aids to prevent advanced stage disease in the form of CPM. ¹¹ , ¹⁷ Based on the observed increasing incidence of CPM in the years prior to screening, a further increase was expected. However, the screen‐eligible population has shown a clear stabilization since screening was initiated, while the entire Dutch population even displays a decrease in CPM incidence. This decrease could be attributed to the fact that when individuals participate in screening, an additional benefit from screening participation can also be observed in the years after being screen‐eligible, above 75 years old. ¹⁸ , ¹⁹

Previous studies have shown that after an initial increase in CRC incidence at the start of the screening program, an overall decreasing trend is observed in all CRC stages. ¹¹ , ²⁰ Detecting CRC in earlier stages and with lower tumor stages is a known effect of screening; this can also be argued to be the reason why a decrease in the CPM incidence is observed as PM is associated with T4 tumors. ²¹ For this matter, screening should not be seen as a measure to diagnose advanced‐stage CRC but also as a means to prevent it.

When regarding every screening cohort separately, the observed and expected CPM incidence show less difference in more recent cohorts. With cohort 2014 until cohort 2018 showing a significant difference between the observed and expected crude incidence rates, such differences were no longer observed in cohorts 2019–2022. This is probably due to a lower number of individuals in these cohorts comprising only one or two birth years in total. Furthermore, these cohorts included a generally younger population as compared to the previous cohorts. Each screening cohort shows a peak in its CPM incidence in the year patients were invited to participate in screening for the first time. After this peak incidence, a stable CPM incidence was observed. The peak CPM incidence was lower for every subsequent cohort after 2014, suggesting that the CPM incidence decrease is influenced by the Dutch screening program. For the total incidence of CRC since the introduction of screening, a peak in the incidence was observed from 2014 to 2016 followed by a slowly decreasing trend. ²⁰ However, no peak in the CPM incidence occurred when all cohorts were included (Figure 1). This could partially be explained by the relatively small proportion of synchronous CPM patients. Furthermore, due to the initiation of the CRC screening program, high proportions of T1 tumors were diagnosed which have probably lead to the observed peak in the general CRC incidence but this does not cause a visible increase in the CPM population. ¹¹

This study has certain limitations. The entire screening population was divided into different cohorts. As 5% of CRC patients are diagnosed with synchronous PM, this results in a relatively low number of included patients per cohort. Besides, the different cohorts comprised of multiple or only one birth year, depending on when patients were invited, which causes the independent cohorts to differ substantially in size. Moreover, participation rate from 2014 to 2022 has slowly been decreasing from approximately 72% to 68%. ²² , ²³ Overall men participate less often in the screening program compared to women, especially in the group of 55–59‐year‐olds. Older invites have the highest participation grade over all the observed years. On average, 93% of individuals participating in the previous screening round, participated again when receiving the next invitation. ²⁴ , ²⁵ , ²⁶ , ²⁷ , ²⁸ , ²⁹ It is important to keep track of this participation rate and evaluate if this decreasing trend persists. As the current results show, screening appears to be a promising tool to detect CRC in its early stages and subsequently aids in preventing advanced‐stage disease such as PM.

5. CONCLUSIONS

This study shows that since the start of nationwide CRC screening, a stabilization of the CPM incidence in the screen‐eligible population has occurred. In the entire population in the Netherlands, even a decrease in the CPM incidence has been observed. These results suggest that nationwide screening through FIT has the potential to decrease the incidence of CPM and as such improve the diagnosis of CRC patients.

AUTHOR CONTRIBUTIONS

Laskarina J. K. Galanos: Conceptualization; investigation; writing – original draft; writing – review and editing; visualization; methodology; software; project administration; data curation; resources; formal analysis. Anouk Rijken: Conceptualization; investigation; writing – review and editing; visualization; methodology; supervision; validation; resources; software. Marloes A. G. Elferink: Writing – review and editing; validation; supervision; resources. Niels F. M. Kok: Validation; writing – review and editing; resources; supervision. Felice N. van Erning: Conceptualization; investigation; writing – review and editing; methodology; validation; visualization; supervision; resources; software; project administration. Ignace H. J. T. De Hingh: Resources; supervision; conceptualization; validation; writing – review and editing; project administration.

CONFLICT OF INTEREST STATEMENT

Dr. I.H.J.T. De Hingh reported an unrestricted research grant from RanD/QPS and Roche, outside the submitted work, paid to the institute. The remaining authors, Dr. L.J.K. Galanos, Dr. A. Rijken, Dr. M.A.G. Elferink, Dr. N.F.M. Kok, and Dr. F.N. van Erning, have nothing to declare.

ETHICS STATEMENT

This study (K22.173) is approved by the privacy review board of the NCR as well as the scientific committee of the Prospective Dutch Colorectal Cancer Cohort (PLCRC).

Galanos LJK, Rijken A, Elferink MAG, Kok NFM, van Erning FN, De Hingh IHJT. Influence of a nationwide colorectal cancer screening program on the incidence of synchronous colorectal peritoneal metastases. Int J Cancer. 2025;157(2):232‐238. doi: 10.1002/ijc.35356

DATA AVAILABILITY STATEMENT

The data that support the findings of this study are available on request from the corresponding author.