Nonconventional dysplasia in patients with inflammatory bowel disease and colorectal adenocarcinoma: a case-cohort study

Abstract

Background and Aims

Patients with inflammatory bowel disease (IBD) face increased risk of colorectal cancer (CRC). While the natural history of conventional dysplastic precursor lesions has been well-studied, the neoplastic potential of recently described nonconventional (NC) IBD-associated colonic mucosal lesions is unclear. We aimed to assess the incidence of antecedent NC lesions in patients with IBD who developed CRC.

Methods

A case-cohort study was performed to include patients with a diagnosis of IBD with or without CRC who underwent at least 2 surveillance endoscopic procedures at our institution between 1/1/2007 and 5/31/2023. NC lesions included serrated change and indefinite for dysplasia. Detection rates pre- and post-introduction of high-definition (HD) surveillance colonoscopy were compared.

Results

In total, 87 patients with IBD and CRC and 200 patients with IBD without CRC were identified. Of the cases, a majority had ulcerative colitis (n = 52, 60%), most commonly with extensive involvement (n = 46, 89%). Conventional (hazard ratio [HR] 2.18, 95% confidence interval [CI] 1.34-3.52) and NC (HR 2.28, 95% CI 1.59-3.26) lesions were associated with increased risk of CRC. Conventional lesions in the post-HD era appeared to have a stronger association with CRC (HR 2.79, 95% CI 1.62-4.77) than NC lesions (HR 1.62, 95% CI 0.86-3.06).

Conclusions

Both conventional and NC lesions seem to be associated with increased risk of CRC. Conventional lesions are more strongly associated with CRC than NC lesions in the post-HD era, but misclassifications in the pre-HD era may have resulted in a biased increased risk estimate for NC lesions.

1. Introduction

Inflammatory bowel disease (IBD) is a group of chronic inflammatory disorders primarily involving the luminal gastrointestinal (GI) tract. IBD affects about 1.6 million Americans and 70 000 new cases of IBD are diagnosed in the United States each year.¹ Ongoing chronic inflammation, among other risk factors, contributes to development of a dreaded complication, namely colorectal cancer (CRC). Long-standing IBD has a 2- to 3-fold increased risk of CRC.² Despite decreases in CRC in recent years, colitis-associated CRC is still a leading cause of mortality and reason for colectomy in patients with IBD.² Therefore, early detection, histologic grading, and removal of premalignant precursor lesions play a critical role in preventing the development of CRC.³

Although there are numerous publications on the morphologic criteria of defining conventional (intestinal type) dysplasia, the most well-recognized form of dysplasia, several single and multi-institutional studies have described various different types of nonconventional (NC) dysplasia. Some of these lesions are distinctly uncommon and may lack traditional morphologic features of dysplasia, making them more difficult to identify.⁴ For example, a diagnosis of “indefinite for dysplasia” is often made if nuclear atypia is limited to the crypt base without surface involvement. This is largely based on the assumption that true dysplasia does not maintain the capacity for maturation, as dysplastic cells migrate toward the surface epithelium.⁵

Choi et al.⁶ have defined at least 7 distinct NC morphologic patterns including hypermucinous, goblet cell-deficient, crypt cell/terminal differentiation, Paneth cell differentiation, sessile serrated lesion (SSL)-like, traditional serrated adenoma (TSA)-like, and serrated not otherwise specified.⁶ Harpaz et al.⁴ proposed a different classification system encompassing 3 broad categories—intestinal dysplasia, gastric dysplasia, and mixed intestinal-gastric dysplasia.⁴ Emerging literature suggests that NC lesions may have an elevated risk of progression to advanced neoplasia (high-grade dysplasia, HGD, or CRC) and often show molecular alterations characteristic of advanced neoplasia.^5,7–9 There is significant variability in the classification and interpretation of these pathologic findings.^10–12 Practicing pathologists may not be familiar with NC dysplasia and may report potential NC lesions as benign or reactive. These factors limit the ability of surveillance colonoscopies to adequately prevent patients with IBD from developing CRC.^6,13,14

Prior studies^15–17 suggesting NC lesions are preneoplastic or associated with subsequent CRC have been limited by small sample size. Therefore, the outcomes of patients with IBD and NC lesions remain unclear. We aimed to assess the incidence and hazards of antecedent NC lesions in patients with IBD who developed CRC.

2. Materials and methods

This retrospective case-cohort study was approved by the Mayo Clinic Institutional Review Board. Using a bioinformatics tool to search the electronic medical record, all adult (≥18 years old) patients were identified using ICD-9 and ICD-10 diagnostic codes for IBD with or without CRC who underwent at least 2 surveillance endoscopic procedures at our institution between 1/1/2007 and 5/31/2023. This was followed by manual review of individual patient charts as permitted by Minnesota Health Records Act, 144.295, which allows patients to opt out of retrospective research.

2.1. Case-cohort definitions

For patients with IBD and CRC, date of diagnosis, endoscopic evaluation, and surgical intervention were recorded including the type of surgical intervention performed. Location, size, number, and grade of adenocarcinoma were also noted. The comparison group was identified from a random sample of patients seen during the same time period with a diagnosis of IBD. No CRC cases were randomly selected as members of the comparison group (Figure 1).

Figure 1.

Screening of patients for study inclusion. Utilizing an institutional bioinformatics search tool, all adult (>18 years old) patients were identified using ICD-9 and ICD-10 codes for diagnoses of Crohn’s disease, ulcerative colitis, and colorectal cancer who had a valid research authorization on file. A total of 4738 patients were excluded. *Of the 200 controls randomly selected as a subcohort, none were members of the case group.

2.2. Patient population

The initial data search identified 211 350 patients with a suspected diagnosis of IBD. From this search, 200 patients were randomly selected for the subcohort, which served as the control group (none of the IBD cases in the subcohort were a CRC case). From the initial data search of 211 350 patients with suspected IBD, 4825 were identified with a suspected concomitant diagnosis of CRC. After manual review, 87 patients with IBD and CRC were included in the final analysis.

In patients with a confirmed IBD and/or CRC diagnosis, we then abstracted relevant demographic, clinical, endoscopic, and histologic outcomes, including location and number of conventional and NC lesions predating diagnosis of CRC. The date of first visit for IBD diagnosis (index) was recorded. A total of 4738 patients were excluded for the following reasons: no confirmed IBD or CRC diagnosis on record; CRC diagnosis before 1/1/2007; subtotal colectomy or ileal pouch anal anastomosis before 1/1/2007; and less than 2 endoscopic procedures performed at our institution (Figure 1).

2.3. IBD and endoscopic characteristics

IBD was diagnosed based on clinical diagnostic criteria as per the treating gastroenterologist and review of the medical record for confirmation of characteristic endoscopic, radiographic, and/or histologic findings, and then subcategorized as having Crohn’s disease (CD), ulcerative colitis (UC), or indeterminate colitis. The date of IBD diagnosis was defined by the earliest pathologic confirmation by our institution.

Related complications such as presence of backwash ileitis or perianal disease, personal history of primary sclerosing cholangitis (PSC), cholangiocarcinoma, and/or liver transplantation, and prior surgical intervention predating 1/1/2007 were abstracted.

Presence of dysplasia and/or mucosal abnormality before 1/1/2007 was noted including location, morphology, size, treatment, and histopathology of the most advanced neoplastic lesion. Presence of low-grade dysplasia (LGD) and HGD were also recorded.

Endoscopic variables included dates of first and last endoscopic procedure, total number of endoscopic procedures within the predefined study period, indication for endoscopic procedure, endoscopic and/or histopathologic evidence of colonic inflammation (including if surveillance and/or random biopsies were obtained), and severity of colonic inflammation if present. Endoscopic severity of colonic inflammation was classified as quiescent, mild, moderate, or severe as documented in procedure reports. At our institution, 4 quadrant surveillance biopsies (totaling 32 biopsies) are obtained during every surveillance colonoscopy, in addition to targeted biopsy and polypectomy.

2.4. Classification of conventional and NC lesions

When neoplasia was found on endoscopic evaluation, the location, size (if available), and type of neoplasia were noted. Conventional dysplasia was defined as tubular adenoma, tubulovillous adenoma, villous adenoma, SSL, and TSA. SSL and TSA were included as conventional lesions as these are often visible lesions. NC lesions included serrated change and indefinite for dysplasia as documented in pathology reports. As the definition of indefinite for dysplasia can be variable pathologically, this was grouped with serrated change as these are distinctly different than conventional lesions.

2.5. Statistical analysis

The effect of NC dysplastic precursor lesions on the risk of CRC was estimated through Cox regression models. To account for the case-cohort study design, Prentice weights were used. Cases entered the risk set instantaneously at the time of their diagnosis based on Barlow et al.¹⁸

For these models, the cumulative number of conventional and NC lesions was updated at each endoscopic procedure and modeled as a time-dependent variable. Lesions discovered during the endoscopic procedure at which CRC was discovered were not included in these cumulative counts (only lesions known prior to CRC). Due to the skewed nature of the number of lesions variable, a square-root transformation was applied. As a result, the hazard ratio (HR) can be interpreted as the expected change in risk for a change in number of lesions from 1 perfect square to the next.

Dysplasia found on surveillance and/or random biopsies were also included for modeling. For each of the 4 colon segment locations (cecum/ascending, transverse, descending, and rectosigmoid), any history of LGD, HGD, and other non-CRC dysplasia was recorded. As HGD was rare, LGD and HGD were combined into a single variable for modeling.

A “location-level” modeling approach looked to model more precise associations by using location-level data with location-level outcomes (CRC). For the location-level models, each of the 4 locations had a separate baseline hazard for the Cox model and a robust variance estimator was used to account for multiple observations per subject. Age and sex were chosen a priori as covariates. Other covariates (year of first surveillance, surveillance prior to 2007, IBD duration, PSC, cholangiocarcinoma, and liver transplant) were chosen due to group differences between cases and controls. As study power was determined by the number of cases that occurred during the study period, a comparison group with a minimum of 200 patients was required. Continuous variables were compared using the Wilcoxon rank sum test. Discrete variables were tested with the Pearson Chi-squared test. All hypotheses’ tests are 2-tailed with a 0.05 Type I error rate. All analyses were conducted using R software (R Foundation for Statistical Computing) version 4.2.2.

3. Results

3.1. Demographics and IBD characteristics

Of the cases, 37% were female with a median age of 51 years at first surveillance endoscopic procedure (Table 1). A majority had UC (60%) with extensive colitis being the most common type (89%). Fifteen (29%) of UC had backwash ileitis and 25 (29%) cases had PSC (Table 2).

Table 1.

Patient characteristics.

	Control (N = 200)	Case (N = 87)	P-value
Age at first surveillance endoscopic procedure Median (Q1, Q3)	42.6 (29.5, 56.8)	50.6 (35.7, 64.2)	.002a
Sex Female	95 (47.5%)	32 (36.8%)	.093b
IBD duration at first surveillance, y Median (Q1, Q3)	5.1 (0.7, 15.8)	10.7 (3.6, 23.4)	.003a
Years on surveillance Median (Q1, Q3)	5.2 (1.8, 8.9)	6.8 (3.0, 9.6)	.126a
Years followed pre-HD Median (Q1, Q3)	2.7 (0.0, 6.9)	5.0 (2.3, 7.6)	<.001a
Years followed in HD Median (Q1, Q3)	6.5 (1.4, 7.4)	3.7 (0.3, 5.7)	<.001a
Total number of surveillance endoscopic procedures Median (Q1, Q3)	3.0 (2.0, 5.0)	4.0 (3.0, 7.0)	<.001a
Nonconventional lesions over all follow-up
0	126 (63.0%)	31 (35.6%)
1	34 (17.0%)	13 (14.9%)
2	16 (8.0%)	13 (14.9%)
3	5 (2.5%)	10 (11.5%)
4+	19 (9.5%)	20 (23.0%)	<.001c
Conventional lesions over all follow-up
0	153 (76.5%)	52 (59.8%)
1	28 (14.0%)	17 (19.5%)
2	8 (4.0%)	9 (10.3%)
3	4 (2.0%)	5 (5.7%)
4+	7 (3.5%)	4 (4.6%)	.009c

Abbreviations: HD, high-definition; IBD, inflammatory bowel disease.

^aKruskal-Wallis rank sum test.

^bPearson’s Chi-squared test.

^cTrend test for ordinal variables.

Table 2.

Inflammatory bowel disease characteristics.

	Control (N =  = 200)	Case (N =  = 87)	P-value
Subtype			.481
CD	89 (44.5%)	32 (36.8%)
UC	105 (52.5%)	52 (59.8%)
IC	6 (3.0%)	3 (3.4%)
Maximal endoscopic extent of CD in past			.861
Ileal	20 (22.5%)	7 (21.9%)
Ileocolonic	56 (62.9%)	19 (59.4%)
Colonic	13 (14.6%)	6 (18.8%)
Maximal endoscopic extent of UC in past			.0321
Proctitis	10 (9.5%)	1 (1.9%)
Left-sided	19 (18.1%)	4 (7.7%)
Extensive	74 (70.5%)	46 (88.5%)
Maximal endoscopic extent of IC in past
Ileocolonic	3 (1.5%)	2 (2.3%)	.631
Colonic	3 (1.5%)	1 (1.1%)	.821
Perianal disease presentb	25 (28.1%)	16 (50.0%)	.0251
Backwash ileitis present	8 (7.6%)	15 (28.8%)	<.001a
PSC present	15 (7.5%)	25 (28.7%)	<.001a
History of cholangiocarcinoma	3 (1.5%)	6 (6.9%)	.021
Prior liver transplant	5 (2.5%)	11 (12.6%)	<.001a
Family history of colorectal cancer	24 (12.0%)	12 (13.8%)	.891
Current/former tobacco use	57 (28.5%)	24 (27.6%)	.871
Previous surgery (prior to 1/1/2007)			.251
Ileocecal resection	14 (56.0%)	2 (25.0%)
Colon segment resection	2 (8.0%)	2 (25.0%)
Hemicolectomy	2 (8.0%)	2 (25.0%)
Other	7 (28.0%)	2 (25.0%)

Abbreviations: CD, Crohn’s disease; IC, indeterminate colitis; PSC, primary sclerosing cholangitis; UC, ulcerative colitis.

^aPearson’s Chi-squared test.

^bPatients with UC and IC were not included.

3.2. Characteristics of dysplasia prior to 1/1/2007

Six (7%) cases had documented dysplasia prior to the start of the study period with indefinite for dysplasia being the most common baseline lesion morphology (50%) (Supplement 1). If prior dysplasia is added as an exclusion criteria (which removes 6 cases and 2 controls), the estimated effects for NC lesions, conventional lesions, and invisible lesions are similar.

3.3. CRC characteristics

For the endpoint of CRC, most had surgical resection (89%) with most common locations being rectosigmoid (45%) and right (41%) colon (Supplement 2). At the time of the endoscopy prior to CRC diagnosis, a mean of 3.5 NC lesions (median 2, IQR 0-3) and a mean of 0.9 conventional lesions (median 0, IQR 0-1) were identified per patient.

In 38 patients who developed CRC in the rectosigmoid colon, 24% had at least 1 conventional lesion, 48% had at least 1 NC lesion, and 29% had invisible LGD/HGD in the same colon segment predating CRC diagnosis. In 32 patients who developed CRC in the right colon, 38% had at least 1 conventional lesion, 31% had at least 1 NC lesion, and 38% had invisible LGD/HGD in the same colon segment predating CRC diagnosis. While attempts were made to classify IBD-related LGD in a prior or actively involved colon segment, it is not always possible to distinguish sporadic adenomas from polypoid LGD.

3.4. Effect of conventional versus NC lesions on the development of CRC

With location-specific data, both conventional (HR 2.18, 95% confidence interval [CI]s 1.34-3.52) and NC (HR 2.28, 95% CI 1.59-3.26) lesions were associated with increased risk of CRC. The risk was not significantly different between these 2 classes of lesions (HR 1.05, 95% CI 0.57, 1.92, P = .89) (Table 3). We also fit a model with an interaction between the number of conventional and NC lesions (HR 0.64, 95% CI 0.39-1.08, P = .09). Without distinguishing between conventional and NC lesions, the adjusted association with CRC was 1.95 (95% CI 1.43, 2.65; P < .001).

Table 3.

Adjusted Cox model for effect of nonconventional versus conventional lesions on the development of colorectal adenocarcinoma.

Effect	Location level
	HR (95% CI)	P-value
Cumulative conventional lesion counta	2.18 (1.34, 3.52)	.002
Cumulative nonconventional lesion counta	2.28b (1.59, 3.26)	<.001
Age (per 5 years)	1.11 (1.01, 1.23)	.04
Female	0.89 (0.45, 1.73)	.72
Year of first surveillance endoscopic procedure	1.14 (1.03, 1.26)	.01
Surveillance endoscopic procedure prior to 2007	0.75 (0.33, 1.70)	.48
IBD duration, years	1.00 (0.99, 1.02)	.87
PSC	3.62 (1.23, 10.67)	.02
Cholangiocarcinoma	1.89 (0.21, 10.34)	.46
Liver transplant	0.97 (0.21, 4.51)	.97
History of invisible low/high-grade dysplasia	32.8 (14.3, 75.3)	<.001
History of other invisible dysplasia	2.20 (0.95, 5.11)	.07

Abbreviations: CI, confidence interval; HR, hazard ratio; IBD, inflammatory bowel disease; PSC, primary sclerosing cholangitis.

^aSquare-root transformation was used. 

^bThe effect of nonconventional lesions versus conventional lesions was HR 1.05 (95% CI 0.57, 1.92; P = .89).

3.5. Effect of conventional versus NC lesions in the pre-HD (prior to 2015) versus post-HD (2015 to present) era

HD surveillance endoscopy was introduced at our center on 1/1/2015. Conventional lesions identified in the post-HD era were significantly associated with risk of CRC (HR 2.79, 95% CI 1.63-4.77), whereas NC lesions were not (HR 1.62, 95% CI 0.86-3.06). In the pre-HD era, these findings were reversed (Table 4). The model accounting for HD era did not perform significantly better than the model without (P = .64).

Table 4.

Cox model for effect of nonconventional versus conventional lesions in the pre (prior to 2015) versus post (2015 to present) high-definition endoscopic era.

Effect	Location level
	HR (95% CI)	P-value
Cumulative conventional lesion count (pre-HD)a	1.79 (0.84, 3.81)	.13
Cumulative conventional lesion count (post-HD)a	2.79 (1.62, 4.77)	<.001
Cumulative nonconventional lesion count (pre-HD)a	2.39 (1.60, 3.57)	<.001
Cumulative nonconventional lesion count (post-HD)a	1.62 (0.86, 3.06)	.13
Age (per 5 years)	1.12 (1.01, 1.24)	.03
Female	0.85 (0.43, 1.67)	.64
Year of first surveillance endoscopic procedure	1.14 (1.02, 1.27)	.03
Surveillance endoscopic procedure prior to 2007	0.72 (0.31, 1.68)	.45
IBD duration, years	1.00 (0.98, 1.02)	.96
PSC	3.63 (1.24, 10.64)	.02
Cholangiocarcinoma	1.65 (0.29, 9.41)	.57
Liver transplant	1.06 (0.23, 4.96)	.94
History of invisible low-/high-grade dysplasia	36.4 (16.1, 82.3)	<.001
History of other invisible dysplasia	2.06 (0.87, 4.86)	.10

Abbreviations: CI, confidence interval; HR, hazard ratio; IBD, inflammatory bowel disease; PSC, primary sclerosing cholangitis.

^aSquare-root transformation was used.

4. Discussion

Both conventional and NC lesions seem to be associated with increased risk of CRC in patients with IBD. Conventional lesions identified in the post-HD era were numerically more strongly associated with CRC than NC lesions from the same era (HRs 2.79 vs 1.62). Most of the literature pertaining to IBD-related dysplasia is in reference to conventional or intestinal-type dysplasia, which is the most common form of dysplasia. However, several patterns of NC dysplasia have been recently described in IBD. Although their clinicopathologic and molecular features are not fully characterized, in part due to the rarity of and likely underrecognition of these subtypes, identification of these lesions is becoming increasingly important as they often present as invisible or flat lesions, and some appear to have higher malignant potential in comparison to conventional dysplasia.^5,7

The natural history of NC dysplastic precursor lesions and relation to future risk of CRC is unclear. In a multicenter cohort study, Choi et al.¹⁵ retrospectively identified 106 dysplastic or serrated lesions in 58 patients with long-standing colonic IBD who ultimately developed CRC. Forty-five percent of these patients had NC dysplasia, and nearly half of these had only NC dysplasia without typical dysplasia findings. This may suggest NC dysplasia alone can either progress to CRC or serve as a biomarker for the development of CRC.¹⁵

Choi et al.¹⁶ further investigated the natural history of a subset of NC dysplastic lesions. They examined 126 cases of hypermucinous, goblet cell-deficient, and crypt cell dysplasia from 97 patients with IBD. Remarkably, 60% of these lesions were associated with subsequent detection of HGD or adenocarcinoma at the site of previous biopsy or in the same colonic segment with a mean follow-up time of 12 months. This suggests some NC dysplasia phenotypes can progress to CRC. However, given the short follow-up time, it is possible these HGD, or adenocarcinoma, are more prevalent than incident.¹⁶

Lee et al.¹⁷ investigated potential mechanisms underlying the possible association of NC dysplasia in 168 patients with IBD and CRC. NC dysplasia was more likely to develop HGD or CRC in the same colonic segment than conventional dysplasia on follow-up (38% vs 19%, P < .001). This increased risk was associated with increased rates of aneuploidy and presentation as flat/invisible lesions in NC dysplasia compared to conventional dysplasia. While this cohort study only included 7 cases of serrated lesions, limiting the ability to describe the natural history of these specific NC dysplasia findings, subsets of serrated lesions have previously been described to have high rates of aneuploidy.¹⁷

In our study, cases had longer IBD duration (10.7 vs 5.1 years), more time on surveillance (6.8 vs 5.2 years), and more intensive surveillance (4 vs 3 endoscopic procedures during study period) compared to controls. With more frequent surveillance in the cases given the length of IBD duration along with endoscopic resection of both conventional and NC lesions, this may be contributing to the prevention of and decreased incidence of CRC over time. Cases had significantly less follow-up years in the HD era in comparison to controls (3.7 vs 6.5 years, P < .001), consistent with increased efficacy of HD colonoscopy in reduction of CRC.

Prior studies have been unclear as to whether NC in the presence of conventional lesions drives the risk of progression to CRC or if NC lesions themselves harbor similar molecular aberrations to that of CRC. It has been described that subtypes of NC dysplasia may be more likely than conventional types to be aneuploid and develop HGD or CRC. Next-generation sequencing to probe molecular characterization of serrated lesions has found frequent mutations in CRC-associated tumor-suppressor TP53 and oncogenes KRAS/BRAF, suggesting these lesions may have important implications in risk of developing CRC.^8,9 Selected DNA markers known to be present in advanced IBD neoplasia can also be detected in tissue from patients with IBD with small adenomas and serrated lesions.⁸ This can be relevant to potential future applications of tissue testing for IBD surveillance. Additional molecular studies are needed to support a causal relationship between NC lesions and CRC. In our study, NC lesion effect relies little on the presence of conventional lesions (HR 0.64, 95% CI 0.39-1.08).

Overall, there is little data available in the literature regarding detection rates of conventional and NC lesions with HD endoscopic techniques. We report in this study a difference in the effects of conventional lesions and NC lesions dependent on whether they were identified in the pre- or post-HD era. Though the effects are not significantly different, they are large enough to be hypothesis-generating. Our findings would be consistent with the possibility that conventional lesions are more strongly associated with CRC than NC lesions, but that misclassifications in the pre-HD era may have resulted in a biased increased risk estimate for NC lesions (and decreased risk estimate for conventional lesions). Several other cumulative effects may account for this difference including introduction of HD endoscopy and chromoendoscopy and changes in histopathologic classification of these lesions over time. Use of advanced endoscopic techniques may play a significant role in heightened detection of NC lesions but may mean less overall as it relates to progression of NC lesions to CRC. As NC lesions are most often flat or invisible, and with removal of conventional adenomas, this may result in a falsely elevated NC lesion count and decreased conventional lesion count.

The introduction of HD endoscopy and chromoendoscopy has challenged the use of random biopsies in surveillance of IBD patients, as a majority of neoplasia is now endoscopically visible.^19–24 In a randomized controlled trial by Watanabe et al.,²⁴ similar rates of dysplasia were detected by targeted (11.4%) versus random (9.3%) biopsies, leading some to contend that random biopsies may not be indicated for dysplasia surveillance.^13,25,26 Conversely, random biopsies have been reported to increase yield of dysplasia in PSC-IBD patients even in the absence of endoscopic inflammation.^1,27–30 Our study also describes increased risk of CRC with prior invisible LGD or HGD, further adding to the importance of considering random biopsies in patients with IBD.

There are some limitations to our study. This study was conducted at a tertiary/quaternary referral center which could potentially contribute to referral bias. Our cohort may have a tendency toward more severe outcomes, making these results less generalizable. This study is also retrospective; thus, future prospective research would be important to verify these findings. Due to insufficient sample size and thus being underpowered, it is unclear as to what specific NC subtype may be driving the HR in association with CRC. Additionally, as most practicing pathologists are not familiar with NC dysplasia and due to the significant variability in the classification and interpretation of these lesions, we may be underestimating the true prevalence and incidence of NC lesions and any potential causal relationship with CRC.

In conclusion, both conventional and NC lesions seem to be associated with increased risk of CRC. In the post-HD endoscopic era, we may be better able to detect differences between the endoscopic appearance of conventional and NC lesions. Also, we postulate that in the post-HD era, we are better at identifying conventional lesions, thereby resulting in significance of conventional lesions post-HD compared to NC lesions in pre-HD in this study. This may also be related to simultaneous evolution in the histopathological definition of these types. Further studies to define the molecular characteristics of NC lesions compared to conventional dysplasia and incidence of CRC in longitudinal cohorts are needed. Also, with the advent of artificial intelligence in IBD, discriminatory algorithms to help differentiate lesions will definitely be developed.

Author Contributions

S.A.U.: Literature search, data collection and interpretation, formation of figures and tables, and drafting of manuscript. N.P.: Data collection and critical review of manuscript. H.V.V.: Data collection and critical review of manuscript. B.B.: Data collection and critical review of manuscript. P.M.: Data collection and critical review of manuscript. A.K.S.: Data collection and critical review of manuscript. R.R.: Data collection and critical review of manuscript. K.A.: Data collection and critical review of manuscript. K.G.: Data collection and critical review of manuscript. S.P.A.: Data collection, analysis, and interpretation, and critical review of manuscript. K.N.B.: Data collection, analysis, and interpretation, and critical review of manuscript. D.W.M.: Data collection, analysis, and interpretation, and critical review of manuscript. B.A.K.: Data collection, analysis, and interpretation, and critical review of manuscript. R.J.L.: Data collection, analysis, and interpretation, and critical review of manuscript. J.B.K.: Literature search, study design, data interpretation, and critical review of manuscript. N.C.-P.: Literature search, study design, data interpretation, and critical review of manuscript.

Funding

J.B.K. was supported by CA214679. No other specific funding or grant support received.

Conflicts of Interest

S.A.U.: Discloses no conflicts. N.P.: Discloses no conflicts. H.V.V.: Discloses no conflicts. B.B.: Discloses no conflicts. P.M.: Discloses no conflicts. A.K.S.: Discloses no conflicts. R.R.: Discloses no conflicts. K.A.: Discloses no conflicts. K.G.: Discloses no conflicts. S.P.A.: Discloses no conflicts. K.N.B.: Discloses no conflicts. D.W.M.: Discloses no conflicts. B.A.K.: Discloses no conflicts. R.J.L.: Discloses no conflicts. J.B.K.: Research funding and intellectual property licensed to Exact Sciences. N.C.-P.: Research funding from Fujifilm, Cook Endoscopy, and Alexion Pharma.

Data Availability

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

Ethical Considerations

This work has been approved by the ethical committees at Mayo Clinic–Rochester and subjects gave prior research authorization.

Article Guarantor

N.C.-P. is the article guarantor and critically reviewed the manuscript for important intellectual content and accepts full responsibility for the conduct of the study.

Conference Presentation

This work was previously presented at the American College of Gastroenterology Annual Meeting and Postgraduate Course in Vancouver, Canada, in 2023 and Digestive Disease Week in Washington, DC, in 2024.