Oncological outcomes of Crohn's disease‐associated cancers focusing on disease behavior

The overall risk of colorectal cancer in Crohn's disease (CD) is higher than in the general population, and CD‐associated cancer (CDAC) has poorer prognosis than sporadic cancer. Developing treatment strategies for improving the prognosis of CDAC, we evaluated the characteristics of CDAC according to the underlying disease behavior, namely stricturing and penetrating.


| INTRODUC TI ON
One of the many problems surrounding the research field of inflammatory bowel disease is development of cancer against the background of chronic persistent inflammation. Accumulating evidence has shown the increased risk of cancer for patients with Crohn's disease (CD) who have had extensive, long-standing, unresected involvement. [1][2][3][4] CD-associated cancer (CDAC) can develop in diverse sites such as small bowel, colon, rectum, anal canal, and anal fistula, and the prognosis of CDAC is poorer than that of sporadic cancer. 5,6 In terms of exposure to chronic persistent inflammation, recent therapeutic advances represented by biologics decrease the risk of surgery and indicate a potential to increase the number of patients with long-standing CD, implying a possible increase in those with CDAC.
Disease behavior such as stenosis and fistula, which is one of the features of CD, is considered to be one of the phenotypes of chronic persistent inflammation in the local environment.
According to the Montreal classification, CD consists of three components: age at diagnosis, location, and behavior. 7 Changes in lesion location (L1, ileal; L2, colonic; L3, ileocolonic; L4, isolated upper disease) during long-term disease are rare; however, changes in disease behavior (B1, nonstricturing, nonpenetrating; B2, stricturing; B3, penetrating) are often seen over the long-term disease course. 8 Indeed, B1 cases at the time of diagnosis can develop high rates of B2 or B3 disease behavior during the period of disease. 9 The factors responsible for either B2 or B3 complications as a result of chronic persistent inflammation remain unclear, although B2 is thought to occur through fibrosis when regeneration and repair cannot restore normal structure. 10,11 By contrast, B3 is thought to occur directly as a result of active inflammation of the intestinal wall. 12 The causes of both behaviors assume a common process of chronic inflammation, and the phenomenon of migration from B2 to B3 can occur; however, these are classified as different pathological conditions, and the differential course of CD can affect the differential pathology. 13 Considering that disease behavior can represent phenotypes of chronic persistent inflammation, B2 or B3 lesions may affect carcinogenesis, known as the dysplasia-carcinoma sequence. Therefore, we hypothesized that different disease behaviors can provide differing clinical phenotypes of CDAC, which has carcinogenic predisposition against the background of chronic inflammation. In this study, we evaluated the characteristics of CDAC according to the underlying disease behavior and examined the possibility of clinical applications for improving the prognosis of CDAC.

| Patients
A total of 316 CDAC patients who underwent surgery between 1985 and 2019 were enrolled in the project study led by the Japanese Society for Cancer of Colon and Rectum, which collected the data from 39 participating institutions. 14 This multicenter retrospective study was approved with the permission of The University of Tokyo Ethics Review Board (2019220NI-(2)). From the data collected, after excluding those without information on disease behavior, 308 cases were examined (Figure 1). We collected information about patients' baseline characteristics, diagnostic procedures, treatment details, histopathological findings, and long-term oncological outcomes.

| Definition and interpretation of terms
In this study, the disease behavior was basically classified at the time of surgery that led to cancer diagnosis. Stricturing was defined as fibrostenotic features and penetrating as intra-abdominal or perianal fistula, perianal ulcer, inflammatory mass, and/or abscess. Although these behaviors were mainly determined from intra-operative findings, but also comprehensively from preoperative diagnostic studies such as small bowel series, enterographic contrast, as well as magnetic resonance imaging or from postoperative pathological evidence. 15 Histopathological diagnosis was confirmed for enrolled patients based on surgical specimens, including histological differentiation, T stage, venous invasion, lymphatic invasion, node involvements. These were defined based on TNM classification and Japanese Classification of Colorectal, Appendiceal, and Anal Carcinoma. 16 In this study, histological differentiation were classified into two groups, well to moderately differentiated and others.
The "well to moderately differentiated type" includes well differentiated type (tub1) and moderately differentiated type (tub2), on the other hand, "others" includes histological findings of mucinous adenocarcinoma (muc, 67.8%), signet-ring cell carcinoma (sig, 10 Crohn's disease, Crohn's disease-associated cancer, disease behavior, penetrating, stricturing anterior sacral surface and lateral pelvic wall were diagnosed to be local recurrences based on their tendency to enlarge and/or FDG accumulation. Other recurrences such as peritoneal dissemination recurrence, liver metastasis, and lung metastasis were similarly diagnosed by imaging findings.

| Statistical analysis
Comparisons of patients' and clinical characteristics between groups were evaluated using the Kruskal-Wallis test for continuous variables (expressed as median ± interquartile range) and Pearson's chi-squared test or Fisher's exact test for categorical variables, as appropriate. A survival analysis was performed using the Kaplan-Meier method and was compared using the log-rank test. Overall survival (OS) was measured from the date the patient underwent surgery until the date of death resulting from any cause, or until the last known follow-up in patients who were still alive. Relapse-free survival (RFS) was measured from the date the patient underwent surgery to the date of disease recurrence, death from any cause (i.e., cancer-unrelated deaths were not censored), or until last contact with the patient. In survival analysis, we excluded 33 patients because of a lack of follow-up information. Cox's proportional hazards model was used to estimate hazard ratio (HR) for OS and RFS. Clinical variables that were considered for univariate and multivariate analyses, in addition to disease behavior, were previously identified confounding factors with an impact on clinical course of CD: onset age, disease duration, smoking history, ileum lesion, colon lesion, use of 5-aminosalicylic acid, use of steroids, and use of anti-tumor necrosis factorα drugs; and factors with an impact on prognosis in patients with colorectal cancer: sex, age at diagnosis of cancer, histological differentiation (well to moderately differentiated or others), pathological T stage (T1/2 or T3/4), lymphatic invasion (present or absent), venous invasion (present or absent), and lymph node metastasis (present or absent). All p values were two-sided, and p < 0.05 was considered statistically significant. Statistical analyses were performed using JMP Pro 13.1 (SAS Institute).

| Patients' characteristics before diagnosis of cancer
Patients' characteristics and clinical courses before diagnosis of cancer are shown in Table 1. There was no association between the clinical course of CDAC patients and the presence of stricturing behavior. Similarly, there was no association between clinical course and the presence of penetrating behavior. However, it has the significant relationship between location of cancer and disease behavior, and between type of surgery and disease behavior, respectively.
Reflecting the large number of cases with lesions in anus and anal canal, abdominoperineal resection was often selected as the operative procedure.

| Difference in Postoperative Information for CDAC Patients According to CD Behavior
Postoperative information, which included pathological findings and recurrence type, are shown in Table 2 Penetrating behavior was associated with poorly differentiated cancer (p = 0.02) and high T stage (p = 0.03). Interestingly, the recurrence type in patients with CD-associated cancer was shown to be different for each behavior. Specifically, stricturing behavior was associated with recurrence of peritoneal dissemination (p = 0.04), whereas penetrating behavior was associated with local recurrence (p = 0.008).

| Oncological outcome of CDAC patients according to CD behavior
We next investigated the impact of CD behavior on the oncological outcomes. Time-to-event analysis showed that disease behavior affected the oncological outcome of CDAC patients ( Figure 2).
Although stricturing behavior was not associated with OS and RFS, penetrating behavior was identified as bringing about a poor oncological outcome for both OS and RFS (OS, p = 0.02; RFS, p = 0.002; log-rank test). F I G U R E 1 A total of 316 CDAC patients who underwent surgery between 1985 and 2019 were enrolled in the project study. From the data collected, 308 cases, excluding those without information on disease behavior, were examined.

| Assessment of disease behaviors with both stricture and penetration
The previous evaluations in this study were based on the presence or absence of stricturing or penetrating behavior. We next examined this cohort subdivided without duplication, categorized according to the  Figure 3). It is noteworthy that the Kaplan-Meier curve of B2 + B3-CDAC was located almost between that of B2-CDAC and B3-CDAC. Furthermore, the association between age-related factors and behaviors showed that onset age of CD in B2 + B3-CDAC was significantly younger than for B1-CDAC, and disease duration of CD in B2 + B3-CDAC tended to be longer than that of B1-CDAC ( Figure 3). Notably, in terms of medians, CDAC patients with B2 + B3 but not B3 had the youngest age at onset of CD, longest disease duration, and youngest age at diagnosis of cancer.  and penetrating behavior was suggested to be an independent disease phenotype.

| DISCUSS ION
The main result of our study is that different disease behaviors can provide different clinical phenotypes of CDAC. We speculate that differences in the aspect of inflammation led to differences in the carcinogenic process. It is possible that the difference in inflammation that leads to fibrosis in B2 10,11

TA B L E 3 (Continued)
assumption of the presence of cancer. Nevertheless, stricturing-CDAC patients were often not operated on with malignancy in mind, and the results suggest high malignancy of penetrating CDAC, which had a poor prognosis even though it was operated on with presumed malignancy. These results were consistent with our hypothesis that differential disease behavior of CD can affect differing pathological conditions and bring about distinct phenotypes of these cancers.
Practical classification of CDAC seemed to conflict with the conventional Montreal classification because this classification is mainly for CD as a benign disease and does not assume malignancy.
However, this classification is considered to be useful for understanding the pathogenesis of CD and examination of our hypothesis.
When the classification was not based on the presence or absence of stricturing and penetrating behavior but on the nonoverlapping classification of B1, B2, B3, and B2 + B3 as in the Paris criteria, 17 each of the disease types showed time-dependent characteristics.
The tendency of B1, B2, B2 + B3, and B3 in that order toward poor prognosis suggests that the B2 + B3 phenotype is a migration stage from B2 to B3, but in fact B2 + B3 cases are those with the longest disease duration, suggesting that they may be independent disease phenotypes ( Figure 4). Although this study focused on disease behavior, location of cancer is also an important factor (Tables S1 and   S2 poor outcome ( Figure S1). However, in those with anal CDAC who were well-known as having poor outcome, 14 the only significant difference in outcome by behavior is in the RFS analysis evaluated by penetrating or not ( Figure S2). A classification system that takes into account not only the concept of CD as a benign disease but also the oncological prognosis of CDAC may be useful in determining future treatment strategies.
Limitations of this study include the following. First, the study was a retrospective study. Second, our data collection spans a long In conclusion, our study highlights the different characteristics of CDAC according to the underlying disease behavior, elucidating the poor prognosis of CDAC patients with penetrating (B3) behavior.
As the number of CDAC cases is expected to increase in the future, physicians may contribute to improved prognosis by developing a treatment plan for each case based on the individual's disease behavior.

AUTH O R CO NTR I B UTI O N S
Conception and design, acquisition of data, and analysis and inter-

F I G U R E 4
The tendency of B1, B2, B2 + B3, and B3 toward poor prognosis in that order suggests that the B2 + B3 phenotype is a migration stage from B2 to B3, but in fact B2 + B3 cases represent the group with the longest disease duration, suggesting that they may be independent disease phenotypes.