High Risk of Colorectal Cancer After High‐Grade Dysplasia in Inflammatory Bowel Disease Patients

ABSTRACT

Background

There are limited data on colorectal cancer (CRC) risk after high‐grade dysplasia in inflammatory bowel disease.

Aims

To determine the long‐term CRC and neoplasia risk after a first diagnosis of high‐grade dysplasia in inflammatory bowel disease, and to assess utilisation of high‐grade dysplasia treatment strategies over the past three decades.

Methods

In this nationwide retrospective cohort study, patients with colonic inflammatory bowel disease and high‐grade dysplasia diagnosis between 1991 and 2021 were extracted from the Dutch nationwide pathology databank (PALGA). The primary outcome was the cumulative incidence of metachronous CRC. Kaplan–Meier curves were used to show proctocolectomy‐free survival per decade.

Results

CRC was diagnosed in 348 of 1220 patients (28.5%). Of these, 204 (16.7%) were diagnosed with CRC within 6 months after the first high‐grade dysplasia diagnosis and were considered synchronous patients. Metachronous CRC was diagnosed in 144 of 1016 patients (14.2%) after a median of 3.6 years. The 1‐, 5‐ and 10‐year cumulative incidences of metachronous CRC after high‐grade dysplasia were 2.9%, 9.9% and 15.5%, respectively. The 1‐, 5‐ and 10‐year cumulative incidences of metachronous neoplasia were 18.3%, 51.2% and 68.0%, respectively. Proctocolectomy‐free survival after high‐grade dysplasia decreased over time.

Conclusions

The risk of synchronous and metachronous CRC after a diagnosis of high‐grade dysplasia underlines the high‐risk profile of this subgroup of patients with inflammatory bowel disease. The possible advantages of colon‐sparing treatment should be balanced with the higher risk of metachronous CRC and the subsequent need for stringent endoscopic surveillance.

The advantages of colon‐sparing treatment for high‐grade dysplasia should be weighed against the higher risk of metachronous colorectal cancer and the subsequent need for stringent endoscopic surveillance.

1. Introduction

Patients with inflammatory bowel disease (IBD) are at increased risk of developing colorectal cancer (CRC) compared to the general population [1, 2]. Chronic inflammation is considered the most important driver of carcinogenesis in IBD. Endoscopic surveillance is recommended to detect and remove colorectal neoplasia (CRN) and reduce CRC‐related morbidity and mortality [3]. Colonic high‐grade dysplasia (HGD), with reported incidence rates of 1.0%–3.5% in IBD, is the highest‐risk precursor of CRC [4, 5, 6, 7].

Although the risk of CRC after HGD is elevated in IBD patients, this observation is based on scarce and historical data, resulting in widely varying CRC rates (0.0%–50.0%) [8, 9, 10]. Historically, colectomy was recommended for HGD given the high risk of invisible synchronous CRN. The implementation of advanced endoscopic imaging techniques has made previously invisible lesions detectable, possibly resulting in a lower advanced CRN risk [11, 12, 13, 14]. This, together with advances in endoscopic resection techniques, facilitated a shift towards endoscopic treatment strategies to avoid colectomy [9, 10, 15, 16]. In addition, there is limited information on CRC risk factors after developing HGD, hampering risk assessment in this specific high‐risk group. Consequently, the optimal surveillance and treatment strategy for HGD in IBD remains uncertain.

In this nationwide study in IBD patients, we aimed to [1] determine the long‐term risk of CRC and CRN after index HGD diagnosis, defined as the first diagnosis of colorectal HGD [2], identify risk factors for metachronous CRC and CRN and [3] assess utilisation of HGD treatment strategies over the past three decades.

2. Materials and Methods

2.1. Design and Outcomes

We performed a nationwide retrospective cohort study to assess the following outcomes after index HGD diagnosis in IBD patients:

1. 1, 5 and 10 year cumulative incidences of metachronous CRC and CRN (including indefinite for dysplasia (IND), low‐grade dysplasia (LGD), HGD and CRC)

2. Hazard ratios (HRs) of clinical and disease characteristics associated with metachronous CRC and CRN (any grade)

3. Applied HGD treatment strategies and proctocolectomy free survival per decade.

2.2. Study Population

We used the Dutch nationwide pathology databank (PALGA) to identify IBD patients with a diagnosis of HGD between 1991 and 2021. PALGA has complete nationwide coverage since 1991, as previously described [17, 18, 19]. We included the following search terms for IBD: ‘ulcerative colitis’, ‘Crohn's disease’, ‘indeterminate colitis’, and ‘chronic idiopathic inflammatory bowel disease’, and for HGD: ‘high‐grade dysplasia’, ‘carcinoma in situ’ and ‘severe dysplasia’ located in the colon or rectum. All patients with colonic IBD and a histological diagnosis of HGD with available histological follow‐up data were included. Patients with familial CRC syndromes or a diagnosis of CRC prior to index HGD diagnosis were excluded.

2.3. Data Collection

Data extracted from PALGA included sex, age at index HGD diagnosis, IBD type, IBD duration at index HGD diagnosis, history of post‐inflammatory polyps (irrespective of number or location), strictures, primary sclerosing cholangitis, prior CRN, prior bowel surgery and academic versus non‐academic follow‐up. In addition, we extracted data on index HGD, including location (right colon, transverse colon or left colon), lesion shape (visible or invisible), multifocality and treatment strategy (type of endoscopic resection technique or surgical bowel resection). All endoscopic or surgical procedures within 6 months after index HGD diagnosis were considered treatment of index HGD. Proctocolectomy was defined as resection of the entire colon and rectum. Moreover, follow‐up data on synchronous and metachronous CRN, including date, grade (IND, LGD, HGD or CRC), and treatment strategy were collected. Synchronous CRC was defined as a histological diagnosis of CRC within 6 months after index HGD diagnosis. All lesions diagnosed > 6 months after index HGD were considered metachronous CRN. Patients with synchronous CRC were excluded from further analysis of metachronous CRC and CRN.

2.4. Statistical Analysis

Categorical and continuous variables were presented as proportions with percentages and medians with interquartile ranges (IQR), respectively. Differences between groups were assessed with the chi‐square test, Student's t test, or nonparametric alternatives. A 2‐tailed p value of < 0.05 was considered statistically significant. The Bonferroni method was applied to correct for multiple testing where appropriate. Cumulative incidences of overall development of CRC, metachronous CRC, and CRN after index HGD diagnosis were shown in 1 minus Kaplan–Meier curves. Patients were censored if no event occurred at the end of follow‐up, defined as the date of proctocolectomy or the last pathology report. We performed a timeframe analysis comparing cumulative incidences of patients with index HGD diagnosis prior to 2010 to patients with index HGD diagnosis in 2010 or later, using the log‐rank test. The cut‐off point was set on 2010 because of the implementation of high‐definition equipment and dye‐based and virtual chromoendoscopy around this year, and the implementation of CRC surveillance in the Dutch IBD guideline in 2008 [20, 21, 22, 23]. We performed a sensitivity analysis excluding metachronous IND since IND may potentially represent reactive changes rather than true neoplasia. Associations with metachronous CRC and CRN (any grade) were assessed using Cox proportional hazards model and were presented as (adjusted) HRs with 95% confidence intervals (CI). Potential associations and confounders were selected based on clinical relevance [24]. The proportional hazards assumption was tested for all variables by visual inspection of log‐minus‐log survival plots and adding time‐dependent covariates to the model. A p value of < 0.05 was considered statistically significant. Missing data on IBD duration, lesion shape, and location were imputed by performing multiple imputations by chained equations (MICE), using predictive mean matching. Missing data were assumed to be missing at random. Variables included in the imputation model were sex, age at index HGD, IBD type, strictures, post‐inflammatory polyps, prior non‐advanced CRN, prior bowel surgery, year of index HGD diagnosis, multifocal index HGD, academic versus non‐academic follow‐up and treatment strategy of index HGD. Cumulative proctocolectomy‐free survival after index HGD per decade was compared using the log‐rank test and was shown in Kaplan–Meier curves. Statistical analyses were performed with SPSS version 29. Incidence rates of CRC and CRN (any grade) with 95% CIs were determined using OpenEpi software [25].

2.5. Ethical Considerations

This study was approved by the institutional review board of the Radboud University Medical Center (2024‐17445) and the scientific committee of PALGA (LZV2022‐80).

3. Results

3.1. Patients

Our PALGA search yielded 2641 patients, of whom 1220 were available for inclusion (Figure 1). The majority of patients were male (66.3%) and had ulcerative colitis (62.8%) (Table 1). The median age at index HGD diagnosis was 61 years. Index HGD was treated endoscopically in 890 (73.0%) patients and surgically in 329 (26.9%) patients, including 67 (5.5%) with proctocolectomy (Table 2). Patients had a median of 8 colorectal pathology reports available. The cumulative follow‐up of the entire cohort from index HGD until the end of follow‐up was 8426 patient‐years, with a median follow‐up of 5.2 years (95% CI 1.9–10.2), including 1034 subjects with a follow‐up of ≥ 1 year, 627 of ≥ 5 years and 312 of ≥ 10 years.

FIGURE 1.

Flowchart patient selection. CRC, colorectal cancer; HGD, high‐grade dysplasia; IBD, inflammatory bowel disease; PALGA, Dutch Nationwide pathology databank.

TABLE 1.

Baseline characteristics of n = 1220 HGD patients.

Characteristics	No CRC (n = 872)	Synchronous CRC (n = 204)	Metachronous CRC (n = 144)	Total (n = 1220)
Male sex, n [%]	580 [66.5]	131 [68.1]	131 [64.2]	809 [66.3]
Disease, n [%]
Ulcerative colitis	548 [62.8]	113 [55.4]	105 [72.9]	766 [62.8]
Crohn's disease	222 [25.5]	64 [31.4]	30 [20.8]	316 [25.9]
IBD‐unclassified	102 [11.7]	27 [13.2]	9 [6.3]	138 [11.3]
Post‐inflammatory polyps, n [%]	191 [21.9]	49 [24.0]	46 [31.9]	286 [23.4]
Strictures, n [%]	79 [9.1]	47 [23.0]	23 [16.0]	149 [12.2]
Primary sclerosing cholangitis, n [%]	49 [5.6]	14 [6.9]	7 [4.9]	70 [5.7]
Prior neoplasia	235 [26.9]	51 [25.0]	45 [31.3]	331 [27.1]
Indefinite for dysplasia	11 [1.3]	4 [2.0]	2 [1.4]	17 [1.4]
Low‐grade dysplasia	203 [23.3]	43 [21.1]	35 [24.3]	281 [23.0]
Unknown non‐advanced grade	21 [2.4]	4 [2.0]	8 [5.6]	33 [2.7]
Prior bowel surgery, n [%]	54 [6.3]	19 [9.3]	8 [4.9]	81 [6.6]
Ileocecal resection	15 [1.7]	3 [1.5]	2 [1.4]	20 [1.6]
Segmental resection	15 [1.7]	3 [1.5]	1 [0.7]	19 [1.6]
Right hemicolectomy	7 [0.8]	1 [0.5]	0 [0.0]	8 [0.7]
Left hemicolectomy	1 [0.1]	0 [0.0]	0 [0.0]	1 [0.1]
Subtotal or total colectomy	16 [1.8]	12 [5.9]	5 [3.5]	33 [2.7]
Age at index HGD diagnosis, median [IQR]	62 [52–69]	61 [48–69]	58 [46–68]	61 [50–69]
IBD duration at index HGD diagnosis in years, median [IQR] a	16.8 [8.6–25.0]	20.2 [12.7–27.9]	17.1 [9.6–25.0]	17.5 [9.3–25.5]
Location index HGD, n [%] a
Ascending colon	155.7 [17.9]	42.7 [20.9]	21.9 [15.2]	220.3 [18.1]
Transverse colon	85.5 [9.8]	26.7 [13.1]	14.8 [10.3]	127.0 [10.5]
Descending colon	603.6 [69.2]	127.1 [62.3]	104.1 [72.3]	834.8 [68.4]
Multiple segments	27.3 [3.1]	7.5 [3.7]	3.3 [2.3]	38.0 [3.1]
Visible index HGD, n [%] a
Visible	816.6 [93.6]	188.1 [92.2]	122.7 [85.2]	1127.4 [92.4]
Invisible	55.4 [6.4]	15.9 [7.8]	21.3 [14.8]	92.6 [7.6]
Multifocal index HGD, n [%]	266 [30.5]	52 [25.5]	38 [26.4]	356 [29.2]
Indefinite for dysplasia	5 [0.6]	1 [0.5]	1 [0.6]	7 [0.6]
Low‐grade dysplasia	174 [20.0]	31 [15.2]	21 [14.6]	226 [18.5]
High‐grade dysplasia	87 [10.0]	20 [9.8]	16 [11.1]	123 [10.1]
Year of index HGD diagnosis, median [IQR]	2008 [2002–2012]	2010 [2004–2015]	2002 [1995–2009]	2007 [2001–2012]
Procedures after treatment of index HGD, median [IQR]	3 [2–6]	3 [1–4]	6 [4–11]	3 [2–6]
Follow‐up after index HGD in years, median [IQR]	5.5 [1.8–10.2]	3.3 [1.3–7.1]	7.1 [2.7–14.9]	5.2 [1.9–10.2]
Academic follow‐up, n [%]	368 [42.2]	111 [54.4]	82 [56.9]	563 [46.0]

Abbreviations: CRC, colorectal cancer; HGD, high‐grade dysplasia; IBD, inflammatory bowel disease; IQR, interquartile range.

^a Pooled value of multiple imputation datasets. IBD duration at index HGD: 81.7% missing, location index HGD: 9.9% missing, Visible index HGD: 8.5% missing.

TABLE 2.

Treatment index HGD within 6 months after index HGD diagnosis.

Treatment index HGD	No CRC (n = 872)	Synchronous CRC (n = 204)	Metachronous CRC (n = 144)	Total (n = 1220)
Endoscopic treatment	724 [83.0]	32 [15.7]	134 [93.1]	890 [73.0]
EMR	17 [1.9]	3 [1.5]	1 [0.7]	21 [1.7]
ESD	12 [1.4]	9 [4.4]	1 [0.7]	22 [1.8]
Hybrid EMR/ESD	1 [0.1]	0 [0.0]	1 [0.7]	2 [0.2]
EFTR	1 [0.1]	0 [0.0]	1 [0.7]	2 [0.2]
Unspecified	693 [79.5]	20 [9.8]	130 [90.3]	843 [69.1]
Surgery	148 [17.0]	172 [84.3]	10 [6.9]	329 [26.9]
Ileocecal resection	10 [1.1]	3 [1.5]	1 [0.7]	14 [1.1]
Segment resection	30 [3.4]	59 [28.9]	0 [0.0]	89 [7.3]
Right hemicolectomy	28 [3.2]	30 [14.7]	5 [3.5]	63 [5.2]
Left hemicolectomy	4 [0.5]	7 [3.4]	0 [0.0]	11 [0.9]
Subtotal or total colectomy	47 [4.7]	40 [19.6]	4 [2.8]	85 [7.0]
Proctocolectomy	35 [4.0]	33 [16.2]	0 [0.0]	67 [5.5]

Abbreviations: CRC, colorectal cancer; EFTR, endoscopic full‐thickness resection; EMR, endoscopic mucosal resection; ESD, endoscopic submucosal dissection; HGD, high‐grade dysplasia.

3.2. Overall Risk of Developing CRC and Synchronous CRC

CRC was diagnosed in 348 of 1220 patients (28.5%) after a median of 0.3 years, corresponding with an incidence rate of 5.2 (95% CI 4.7–5.8) per 100 patient‐years. The 1‐, 5‐ and 10‐year cumulative incidences of CRC after HGD were 20.2% (231/1143), 28.1% (283/1006) and 34.9% (310/889), respectively (Figure 2). Of these, 204 patients (16.7%) were diagnosed with CRC within 6 months after index HGD diagnosis and were considered synchronous CRC patients. CRC was found in the same colonic segment as index HGD in 257 patients (73.9%), including 190 synchronous CRC patients (93.1%). Of these 257 patients, 13 represented local recurrence after macroscopically and microscopically radical resection (R0) of index HGD (all metachronous CRC at least 12 months after index HGD). 20 patients had incomplete (R1 or R2) resection. Most other cases represented either sampling error or invisible index HGD. Of note, in 17/257 cases the resection margin was unknown.

FIGURE 2.

Kaplan–Meier curves of the overall cumulative incidence of development of CRC after index HGD diagnosis (n = 1220, the orange striped line), the cumulative incidence of metachronous CRC (n = 1016, the blue line), and the cumulative incidence of metachronous neoplasia (n = 1016, the green dotted line). CRC, colorectal cancer; CRN, colorectal neoplasia; HGD, high‐grade dysplasia.

3.3. Cumulative Incidence of Metachronous CRC

Metachronous CRC was diagnosed in 144 of 1016 patients (14.2%) after a median of 3.6 years, corresponding with an incidence rate of 2.2 (95% CI 1.8–2.5) per 100 patient‐years. The 1‐, 5‐ and 10‐year cumulative incidences of metachronous CRC after HGD were 2.9% (27/939), 9.9% (79/802) and 15.5% (106/684), respectively (Figure 2).

Our timeframe analysis showed an incidence rate of 2.5 (95% CI 1.8–3.5) per 100 patient‐years for patients with index HGD diagnosis in 2010 or later, which is similar to the incidence rate of patients with index HGD before 2010 (Figure S1A, p = 0.34).

3.4. Cumulative Incidence of Metachronous CRN (Any Grade)

Metachronous CRN developed in 638 patients (62.8%) after a median of 2.2 years, corresponding with an incidence rate of 15.7 (95% CI 14.5–16.9) per 100 patient‐years (Figure 2). Of these, 12 patients developed IND as the highest grade of metachronous CRN, 243 LGD, 227 HGD and 144 CRC. The 1‐, 5‐ and 10‐year cumulative incidences of metachronous CRN were 18.3% (174/952), 51.2% (456/891) and 68.0% (577/848), respectively. Twenty‐five of 60 patients (41.7%) with a residual rectum after subtotal colectomy developed metachronous CRN, including two patients with IND as the maximum grade of CRN, 7 with LGD, 7 with HGD and 9 with CRC.

Our sensitivity analysis excluding metachronous IND showed an incidence rate of 15.1 (95% CI 14.0–16.3) per 100 patient‐years. The 1‐, 5‐ and 10‐year cumulative incidences were 17.8% (169/951), 49.9% (443/888) and 66.8% (563/843).

Our timeframe analysis showed an incidence rate of 14.0 (95% CI 12.7–15.3) per 100 patient‐years for patients with index HGD diagnosis before 2010, and 21.8 (95% CI 18.9–25.0) per 100 patient‐years for patients with index HGD diagnosis in 2010 or later (Figure S1B, p = < 0.01).

3.5. Associations With Metachronous CRC and CRN (Any Grade)

Post‐inflammatory polyps (aHR 2.03, 95% CI 1.43–2.88, p < 0.01), prior IND or LGD (aHR 1.50, 95% CI 1.01–2.22, p = 0.04), disease duration of 20 years or longer (aHR 1.28, 95% CI 1.23–1.33, p < 0.01), invisible index HGD (aHR 2.25, 95% CI 1.34–3.78, p < 0.01), academic follow‐up (aHR 1.49, 95% CI 1.05–2.12, p = 0.03) and endoscopic versus surgical treatment (aHR 2.31, 95% CI 1.17–4.56, p = 0.02) were independent risk factors for metachronous CRC after index HGD diagnosis (Table 3).

TABLE 3.

Univariable and multivariable hazard ratios for metachronous CRC for n = 1016 patients.

Characteristics	HR univariable [95% CI]	p value	HR multivariable [95% CI]	p value
Male sex	1.00 [0.71–1.42]	0.99	—	—
Disease1
Ulcerative colitis	1.19 [0.79–1.79]	0.40	1.19 [0.79–1.79]	0.40
IBD‐unclassified (ref Crohn's disease)	0.66 [0.31–1.39]	0.27	0.66 [0.31–1.39]	0.27
Post‐inflammatory polyps2	2.03 [1.43–2.88]	< 0.01	2.03 [1.43–2.88]	< 0.01
Strictures3	1.68 [1.08–2.63]	0.02	1.54 [0.95–2.51]	0.08
Primary sclerosing cholangitis4	0.95 [0.44–2.03]	0.89	0.94 [0.44–2.01]	0.87
Prior neoplasia5	1.70 [1.19–2.43]	< 0.01	1.50 [1.01–2.22]	0.04
Prior bowel surgery6	1.21 [0.59–2.47]	0.61	0.92 [0.41–2.05]	0.83
Age at index HGD diagnosis, per year increase7	1.00 [0.99–1.02]	0.86	1.00 [0.99–1.01]	0.95
Disease duration at index HGD diagnosis in years, per year increase8	1.02 [1.00–1.03]	0.10	1.02 [1.00–1.03]	0.10
Disease duration at index HGD diagnosis of 20 years or longer8 (vs. < 20 years)	1.28 [1.23–1.33]	< 0.01	1.28 [1.23–1.33]	< 0.01
Year index HGD diagnosis ≥ 2010	1.23 [0.81–1.86]	0.34	—	—
Right sided index HGD9	0.92 [0.56–1.50]	0.74	0.94 [0.57–1.54]	0.79
Invisible index HGD10	2.19 [1.34–3.60]	< 0.01	2.25 [1.34–3.78]	< 0.01
Multifocal index HGD11	0.98 [0.67–1.41]	0.89	0.91 [0.63–1.33]	0.64
Academic follow‐up12	1.71 [1.23–2.38]	< 0.01	1.49 [1.05–2.12]	0.03
Endoscopic treatment of index HGD13 (vs. surgical treatment)	1.76 [0.92–3.35]	0.09	2.31 [1.17–4.56]	0.02

Note: Multivariable model with adjustment for:¹ sex and PSC; ²sex and age at index HGD diagnosis; ³sex, IBD type, disease duration, and invisible index HGD; ⁴sex and IBD type; ⁵sex, IBD type, PIPs, strictures, PSC, age at index HGD diagnosis, and disease duration; ⁶sex, IBD type, PIPs, strictures, PSC, age at index HGD diagnosis, disease duration, year of index HGD diagnosis; ⁷sex, PSC, disease duration; ⁸sex and PSC; ⁹IBD type, PSC and prior bowel surgery; ¹⁰PIPs, strictures, PSC, age at index HGD, disease duration, year of index HGD diagnosis; ¹¹sex, IBD type, PIPs, strictures, PSC, prior neoplasia, prior bowel surgery; ¹²invisible index HGD, PSC, prior bowel surgery, treatment of index HGD, strictures, prior neoplasia, multifocal index HGD; ¹³sex, IBD type, PIPs, strictures, PSC, prior neoplasia, prior bowel surgery, age at index HGD, disease duration, year of index HGD, location, invisible index HGD, multifocal index HGD, academic follow‐up.

Abbreviations: HGD, high‐grade dysplasia; HR, hazard ratio; IBD, inflammatory bowel disease; PIPs, post‐inflammatory polyps; PSC, primary sclerosing cholangitis.

In addition, post‐inflammatory polyps (aHR 1.36, 95% CI 1.13–1.64, p < 0.01), primary sclerosing cholangitis (aHR 1.59, 95% CI 1.15–2.20, p < 0.01), prior IND or LGD (aHR 1.38, 95% CI 1.14–1.68, p < 0.01), age at index HGD diagnosis (aHR 1.01, 95% CI 1.01–1.02, p < 0.01), IBD duration at index HGD (aHR 1.01, 95% CI 1.00–1.01, p = 0.04), year of index HGD diagnosis in 2010 or later (HR 1.47, 95% CI 1.23–1.75, p < 0.01), multifocal index HGD (aHR 1.25, 95% CI 1.05–1.49, p = 0.01), academic follow‐up (aHR 1.25, 95% CI 1.06–1.48, p < 0.01) and endoscopic treatment of index HGD (aHR 2.64, 95% CI 1.93–3.60, p < 0.01) were independent risk factors for metachronous CRN (Table S1). Moreover, male sex (HR 0.78, 95% CI 0.66–0.93, p < 0.01) was a protective factor for metachronous CRN.

3.6. Treatment Modalities Per Decade

Patients with index HGD diagnosis in 2011–2020 were older, with longer disease duration at index HGD diagnosis, and more often a diagnosis of Crohn's disease, prior neoplasia, post‐inflammatory polyps and multifocal index CRN, compared to previous decades (Table S2). Endoscopic treatment was performed in 191 (78.6%) patients with diagnosis of index HGD in 1991–2000, versus 392 (74.1%) in 2001–2010 and 268 (66.5%) in 2011–2020 (Figure S2, p < 0.01). Proctocolectomy was performed in 162 (15.9%) patients after a median of 5.3 (IQR 1.7–10.2) years after index HGD diagnosis. Proctocolectomy‐free survival was lower in patients with index HGD diagnosis in 2011–2020 compared to previous decades after 7 years of follow‐up (Figure 3, p = 0.04).

FIGURE 3.

Proctocolectomy free survival after index HGD diagnosis per decade. HGD, high‐grade dysplasia.

4. Discussion

In this nationwide cohort study, including 1220 IBD patients with colorectal HGD, the 10‐year cumulative incidences of metachronous CRC and CRN were 15.5% and 68.0%, respectively. Synchronous CRC was diagnosed in one out of six patients. Post‐inflammatory polyps, prior IND or LGD, academic follow‐up and endoscopic treatment of index HGD were associated with metachronous CRC and CRN.

Our study yielded an incidence rate of metachronous CRC of 14.2% after a median of 3.6 years after HGD. In addition, we showed an incidence of overall development of CRC of 5.2 per 100 patient‐years. A prior study from our group showed a combined HGD and CRC incidence rate of 11.5% after a median of 2.3 years after index advanced neoplasia (n = 81 HGD, n = 108 CRC) depending on treatment strategy [9]. Other reported smaller cohorts (maximum n = 48) included metachronous CRC rates after index HGD varying from 0.0% to 50.0% after a median of 2.8–15.0 years, and an overall CRC rate of 23 per 100 patient‐years [4, 8, 10]. These variations might be the result of differences in the definition of metachronous CRC, study population (e.g., academic vs. non‐academic) and most importantly, sample size. We found a high synchronous CRC rate of 16.7%, in line with previous studies that reported synchronous CRC rates of 11.4%–33.3% [4, 8]. This high risk of a CRC diagnosis within 6 months after HGD diagnosis indicates the need for clinical awareness and the need for careful endoscopic assessment. Indeed, recent studies showed that endoscopic surveillance compliant with primary quality indicators reduces HGD and CRC risk, and that post‐colonoscopy CRCs were associated with poor bowel preparation and high‐risk factors for CRC [26, 27]. Of note, we frequently found cases in which there was a high endoscopic suspicion for CRC, followed by a histopathological diagnosis of HGD. Moreover, 9 out of 10 synchronous CRCs were found in the same colonic segment as the index HGD lesion. This observation emphasises the risk of sampling error and the challenging histopathological differentiation between HGD and CRC in IBD. Indeed, a previous study from our group showed an increased metachronous CRC risk after confirmation of histopathological HGD diagnosis by a second pathologist [28].

Our timeframe analysis showed an increased incidence rate of metachronous CRN over time, which may reflect advances in endoscopic imaging techniques and lesion detection. The incidence rate of metachronous CRC did not change. Although a population‐based study from Scandinavia showed a decreasing risk for CRC in UC between 1969 and 2017, this was all CRC rather than the metachronous CRC in our study [1]. In addition, a recent Dutch study showed no change in time to CRC in UC patients undergoing colectomy between 2013 and 2022 [29].

We found that post‐inflammatory polyps, prior IND or LGD, academic follow‐up, and endoscopic versus surgical treatment of index HGD were associated with metachronous CRC and CRN, which is in line with previous studies on CRN risk in IBD [9, 19, 24]. Post‐inflammatory polyps likely result from chronic active inflammation and may therefore serve as a surrogate marker for disease severity, as was shown in two recent studies [30, 31]. In contrast to previous reports, male sex was not identified as a risk factor in our cohort [19, 24].

We found an increased proportion of surgical treatment of index HGD and a decreasing proctocolectomy‐free survival after index HGD over time. This could be explained by the longer disease duration, as well as the higher prior non‐advanced neoplasia and multifocal index HGD rate in the last decade, making surgical treatment a more appropriate strategy. This, in turn, might be the result of expanding therapeutic options for IBD, delaying or even preventing colectomies performed for refractory disease. In line, a recent Dutch study showed an increased proportion of CRC in ulcerative colitis patients undergoing colectomy over time, although the absolute number of colonic resections did not change [32].

The findings of this study facilitate further CRC risk classification of this subgroup of IBD patients and may subsequently aid in selecting the appropriate treatment strategy for HGD. This is of clinical relevance, since current international guidelines provide only limited guidance for endoscopic surveillance after HGD, based on historic data. Endoscopic techniques have become more refined over time, facilitating better lesion detection and resection, which may favour endoscopic resection as a treatment option for HGD. However, prior IND or LGD and endoscopic treatment were risk factors for metachronous CRC and CRN in our study, emphasising the need for intensified endoscopic surveillance after neoplasia. Moreover, if high‐quality endoscopic surveillance is not feasible or the patient is at high risk of metachronous CRC, colectomy should be considered.

Our study has several strengths. The nationwide design allowed us to build the largest HGD cohort in IBD to date in the literature. We included over 1200 patients, where the second largest cohort in the literature comprised less than 100 HGD patients. The median follow‐up of 5.2 years was likely sufficient to identify the vast majority of metachronous lesions, if any. Of note, most patients reached the natural end of follow‐up and censorship, where the number of patients with loss to follow‐up was small. We performed a timeframe analysis to assess the possible change in metachronous CRC risk over time. There are also limitations. Data on IBD extent and severity and medical therapy were not available in the pathology database, limiting the ability to estimate the historic inflammatory burden per patient and to differentiate IBD‐related CRN from sporadic CRN. Sporadic CRN is defined as CRN detected in a part of the colon that is not presently or historically affected by endoscopic or histological inflammation. However, in daily practice, it often remains difficult to accurately identify colonic segments that do not show signs of past or present inflammation. In line with that, the prediction calculator published in the most recent BSG guideline is based on a prediction model from an international study with retrospective data from 15 sites in 3731 patients that does not distinguish sporadic from IBD‐related CRN [33, 34]. In addition, established risk factors such as smoking and family history of CRC were not available. Consequently, not all known risk factors for CRN development in IBD could directly be included in the analysis. Hence, these outcomes should be interpreted with caution. Although numbers correspond with previous studies, there might be an underestimation of the presence of post‐inflammatory polyps, strictures, and primary sclerosing cholangitis in our cohort, since these may not always be described in the clinical information of pathology reports [6, 9, 30, 35]. Likewise, the endoscopic resection technique was frequently not specified in pathology reports. In addition, data on IBD duration was incomplete. In order to ensure adequate confounder correction, we performed data imputation in our multivariable analyses.

In conclusion, we found a high risk of synchronous and metachronous CRC after index HGD, underlining the high‐risk profile of this subgroup of IBD patients. The advantage of endoscopic resection of HGD, including prevention of bowel surgery and preserving bowel function, should be carefully balanced with the higher risk of metachronous CRC and the subsequent need for stringent endoscopic surveillance. These data may aid physicians in counselling IBD patients on CRC surveillance after HGD.

Author Contributions

Monica E. W. Derks: conceptualization, data curation, formal analysis, writing – original draft, methodology. Maarten te Groen: conceptualization, methodology, writing – review and editing. Lauranne A. A. P. Derikx: conceptualization, methodology, writing – review and editing. Iris D. Nagtegaal: resources, writing – review and editing. Frank Hoentjen: conceptualization, methodology, writing – review and editing, supervision.

Ethics Statement

This study was approved by the institutional review board of the Radboud University Medical Center (2024–17445) and the scientific committee of PALGA (LZV2022‐80).

Conflicts of Interest

Monica E. W. Derks: none. Maarten te Groen: none. Lauranne A. A. P. Derikx has served on advisory boards or as a speaker for Abbvie, Janssen, Galapagos, and Pfizer, and has received independent research funding from Pfizer. Iris D. Nagtegaal: none. Frank Hoentjen has served on advisory boards or as a speaker for Abbvie, Janssen, MSD, Takeda, Pfizer, Celltrion, Teva, Sandoz, Amgen, and Pendopharm, and has received independent research funding from Janssen, Abbvie, Pfizer, and Takeda.

Supporting information

Data S1: apt70429‐sup‐0001‐supinfo.docx.

Data Availability Statement

Data are available on request from the corresponding author.