Abstract
Background and Aims
Immune-mediated colitis (IMC) is a challenging adverse effect of immune checkpoint inhibitor therapy, often leading to treatment interruption or discontinuation. Current IMC endoscopic scoring systems were adapted from those used for inflammatory bowel disease but fail to fully capture IMC. The newly proposed IMC endoscopic score (IMCES) seeks to help guide prognosis and treatment. We aimed to validate IMCES and explore its association with IMC clinical severity and outcomes.
Methods
This single-center retrospective study included patients with IMC who received immune checkpoint inhibitors and underwent endoscopic evaluation. IMCESs were calculated on the basis of 10 gross endoscopic features, as previously reported. Patients were divided into those with IMCES <4 and IMCES ≥4. Primary end points were selective immunosuppressive therapy (SIT) use, hospitalization, and clinical and endoscopic remission.
Results
We evaluated 807 patients, including the 308 previously used to develop IMCES, with 499 in a validation cohort. We found that IMCES ≥4 was associated with need for steroids and SIT (P < .05). Patients with IMCES ≥4 also tended to be hospitalized more often and longer (P < .05). Including ulceration as a criterion in IMCES strengthened these associations. Clinical symptoms had lower specificities for being associated with SIT use (26.8% for diarrhea, 64.6% for colitis) compared to IMCES ≥4 (91.5%).
Conclusions
IMCES ≥4 had a high specificity for the need for SIT and was associated with worse outcomes. Given the importance of risk stratification in the treatment of IMC to inform the early introduction of SIT, IMCES may be a powerful clinical tool to estimate prognosis and guide management but requires further validation.
Introduction
Immune-mediated colitis (IMC) is a very challenging adverse effect of immune checkpoint inhibitor (ICI) therapy and frequently necessitates treatment interruption or discontinuation. It occurs in approximately 37% of patients who receive ICIs and shares features with various inflammatory colitis.1,2 Because of this, much of the evaluation and management of IMC has been adapted from those used in the treatment of inflammatory bowel disease. Currently several guidelines describe treatment for IMC, with steroids indicated as first-line treatment and selective immunosuppressive therapy (SIT) for treatment of steroid-refractory IMC.3, 4, 5 Infliximab and vedolizumab are the most studied and used agents,6, 7, 8 although growing evidence supports the use of ustekinumab and fecal microbiota transplantation, especially for refractory IMC.9, 10, 11
Although treatment for IMC is extensively discussed in the literature, few studies have explored diagnostic work-up for IMC and its prognostic implications. Levels of fecal lactoferrin and calprotectin have both been shown to be associated with IMC activity with high sensitivity,12,13 with calprotectin level analysis being an effective tool for disease monitoring.13 Determining levels of fecal lactoferrin and calprotectin is often the first step in IMC evaluation because doing so helps determine a patient's need for endoscopic assessment. Multiple endoscopic features are associated with worse disease outcomes. For example, ulcers deeper than 2 mm, ulcers with a surface area larger than 1 cm, or extensive colitis beyond 1 segment are correlated with the need for more-aggressive treatment.12 Endoscopic scores adapted from those used for inflammatory bowel disease have been shown to have prognostic value for IMC, with the Mayo Endoscopic Subscore (MES) and Ulcerative Colitis Endoscopic Index of Severity being similarly associated with the need for SIT.14
There is an important need for standardized endoscopic scoring systems in inflammatory colitis, and IMC in particular, because clinical symptoms may not necessarily capture the full clinical severity, need for specific therapies, and outcomes. In fact, clinical symptoms, have been shown to poorly correlate with disease severity on the basis of endoscopy findings and have low specificity for the need for SIT.14, 15, 16 The MES is an effective tool and is currently the most used endoscopic scoring system for IMC,3 but it fails to capture the full extent of disease manifestations, particularly the aforementioned high-risk endoscopic features as well as the microscopic colitis subtype.3 For these reasons, an IMC-specific endoscopic scoring system that reflects the multitude of IMC features is essential. In a large multicenter study, Wang et al16 proposed the 10-point immune-mediated colitis endoscopic score (IMCES) and showcased its high specificity for being associated with need for SIT, outperforming the MES. However, further validation is needed before IMCES can be widely adopted in current clinical practice.
To this end, our study aimed to validate the IMCES and explore its association with IMC clinical severity and outcomes. We hypothesized that a greater IMCES would be associated with worse IMC severity and outcomes.
Methods
Study design
In this retrospective analysis, we reviewed the medical records of patients who received ICIs and underwent endoscopic evaluation for IMC at a tertiary cancer center between January 2010 and February 2024. Patients suspected of developing IMC were identified by results of lower endoscopy or stool tests for evaluation of gastrointestinal symptoms that developed any time between their first ICI dose to 1 year after their last 1. These data were extracted from patients' electronic health records and subsequently confirmed by manual chart review. We also collected detailed patient data on demographics and baseline factors, endoscopic features, IMC treatments, and clinical outcomes. Inclusion criteria were patients who (1) received ICIs for malignancy, (2) were diagnosed with IMC after ICI therapy, (3) and underwent lower endoscopy (either colonoscopy or flexible sigmoidoscopy) for colitis evaluation. Exclusion criteria were (1) lack of lower endoscopy records, (2) lack of colitis symptoms or evaluation, and (3) non-IMC cause of colitis (ie, infectious, autoimmune, radiation-related or ischemia-related). Patients were divided into 2 cohorts: those with IMCES <4 and those with IMCES ≥4 (IMCES was calculated for each patient as described herein).
Colitis clinical characteristics and outcomes
We collected patients' baseline demographics, oncology-related variables (eg, cancer type and stage and ICI type), and IMC-related variables (eg, diarrhea and colitis duration and grade [as determined by the Common Terminology Criteria for Adverse Events, ie, CTCAE, version 5 found at https://ctep.cancer.gov/protocoldevelopment/electronic_applications/docs/ctcae_v5_quick_reference_5x7.pdf], laboratory results, and treatment type and duration). We also collected data regarding discontinuation of ICI treatment secondary to IMC as well as ICI resumption (successful resumption of ICI with no recurrence of colitis among patients who initially had discontinuation of their treatment). Primary clinical end points included hospitalization and rehospitalization rates, IMC remission rates (as defined as symptom improvement to CTCAE grade 1 or below after initial diagnosis), and symptom duration. Secondary end points included frequency of SIT and use and duration of systemic steroids (both oral and intravenous). The decision was made post hoc to not include endoscopic and histologic remission data because there were an insufficient number of patients with adequate endoscopic follow-up, and remission rates of those who had follow-up were too high to draw any meaningful conclusions.
Endoscopic evaluation
Endoscopy-related data were collected, including endoscopy type, gross description of various features, and time from IMC diagnosis to endoscopy. We calculated the 10-point IMCES by independently reviewing endoscopy reports at the time of IMC diagnosis and tallying all the features listed in the report on the basis of the endoscopic features previously described by Wang et al16 (ie, ulceration, large ulcers, deep ulcers, at least 3 ulcers, erythema, loss of vasculature, friability, erosions, exudate, and extensive inflammation). Representative images for each of these features can be found in the original article.16 The location of colitis was described as the left side of the colon (ie, in the segment of the colon from the left half of the transverse colon to rectum), right side of the colon (ie, in the right half of the transverse colon to the cecum), both, or neither.
Statistical analysis
Statistical analyses were performed using SPSS, version 26.0 (IBM Corp, Armonk, NY, USA). Continuous variables are described by their median and interquartile ranges (IQRs), and categorical variables are described by their frequencies and percentages. The χ2 test with Fisher exact test was used to compare categorical variables. The Mann-Whitney U test was used to compare the distributions of continuous variables. Univariate binary logistic regression was used to explore the association of IMCES with patient clinical characteristics and colitis outcomes. Variables with P < .200 on univariate analysis or those that were clinically relevant on the basis of the authors' clinical expertise were included in the multivariate logistic regression. Finally, specificity was manually calculated using the following formula: . We made the decision not to compared the specificity of the IMCES to the MES as we have already demonstrated this comparison in our previous article.16
Results
Patient demographic characteristics
The patient selection flowchart can be found in Figure 1. A total of 807 patients were included, including 499 patients who underwent endoscopic evaluation after the study by Wang et al16 to develop IMCES. Patients were predominantly of white race (90.2%) and male (58.1%) and had a median age of 64.5 years (IQR, 54.1-71.8 years) (Table 1). Patients most frequently presented with stage IV cancer (71.0%) and an Eastern Clinical Oncology Group performance status of 0 to 1 (88.8%). Melanoma was the most common cancer type (32.6%), followed by genitourinary cancers (25.9%) and lung cancer (14.0%). Most patients received antiprogrammed cell death protein 1/programmed cell death ligand 1 therapy (47.8%) or combination ICI-based therapy (35.7%). Patients were followed for a median 1.6 years (IQR, 0.5-3.5 years) and had an all-cause mortality rate of 43.9%.
Figure 1.
Patient selection flowchart.
Table 1.
Demographic characteristics of patients who developed IMC and underwent endoscopy (n = 807)∗
| Characteristic | n (%) |
|---|---|
| Age at the time of ICI therapy, years, median (IQR) | 64.5 (54.1-71.8) |
| Male sex | 469 (58.1) |
| Race | |
| White | 728 (90.2) |
| Type of ICI | |
| Anti-PD-1/L1 agent | 386 (47.8) |
| Anti-CTLA4 agent | 133 (16.5) |
| Combination | 288 (35.7) |
| Duration of ICI therapy, months, median (IQR) | 0.5 (0.1-1.5) |
| Cancer type | |
| Melanoma | 263 (32.6) |
| Genitourinary | 209 (25.9) |
| Lung | 113 (14.0) |
| Gastrointestinal | 65 (8.1) |
| Head and neck | 43 (5.3) |
| Other | 114 (14.1) |
| Cancer stage | |
| I | 25 (3.1) |
| II | 35 (4.3) |
| III | 133 (16.5) |
| IV | 573 (71.0) |
| ECOG PS | |
| 0 | 337 (41.8) |
| 1 | 379 (47.0) |
| 2-4 | 83 (10.3) |
| All-cause mortality | 354 (43.9) |
| Length of follow-up, median (IQR), years | 1.6 (0.5-3.5) |
CTLA4, Cytotoxic T-lymphocyte antigen 4; ECOG PS, Eastern Clinical Oncology Group performance status; ICI, immune checkpoint inhibitor; IMC, immune-mediated colitis; IQR, interquartile range; PD-1/L1, programmed death 1/programmed death ligand-1.
Some rows in this table do not add up to the total sample size of 807 as the result of missing data.
Colitis clinical and endoscopic features
A summary of colitis clinical and endoscopic characteristics in the overall cohort can be found in Supplementary Table 1 and Supplementary Figure 1, available online at www.igiejournal.org. Patients developed IMC a median of 3.2 months from ICI initiation (IQR, 1.4-8.3 months). Of the 638 patients for whom fecal lactoferrin data were available, 528 (82.8%) had laboratory results positive for fecal lactoferrin around the time of initial evaluation. Almost every patient (94.4%) had diarrhea as a presenting symptom, with abdominal pain (41.3%), blood or mucus in the stool (17.1%), and fever (11.2%) being other common symptoms. In terms of severity, patients frequently presented with grade ≥2 diarrhea (80.5%), and approximately one-half of the patients (49.8%) had grade ≥2 colitis. A small proportion of patients (n = 110, 13.6%) received conservative management with best supportive treatment. The remaining patients received treatment with corticosteroids alone (39.8%), SIT and corticosteroids (66.3%), or fecal microbiota transplantation (7.6%), either as a first- or later-line treatment for IMC. Overall, 39.4% of patients required intravenous corticosteroids, and 67.5% of patients required hospitalization. Patients were hospitalized for a median of 6 days (IQR, 4-9 days). IMC symptoms lasted for a median of 34 days (IQR, 12-76 days), with 93.4% of patients having clinical improvement or remission by end of follow-up. Among patients who had to discontinue ICI treatment, 33.2% were able to resume ICI therapy.
With regard to endoscopic features, most patients had macroscopic evidence of inflammation, and 160 patients (19.8%) had ulcerative inflammation. Common features noted on gross examination can be found in Supplementary Figure 1. Erythema was the most common macroscopic finding (49.9%), followed by loss of vascularity (32.3%) and friability (24.4%). Ulceration was present in 19.8% of patients.
Comparison of colitis clinical features between IMCES <4 and IMCES ≥4
Patients were divided into 2 groups: those with IMCES <4 (n = 651, 80.7%) and those with IMCES ≥4 (n = 156, 19.0%). Compared with patients with IMCES <4, patients with IMCES ≥4 were more likely to have grade ≥2 diarrhea (92.7% vs 82.7%; P = .002) or grade ≥2 colitis (68.5% vs 48.5%; P < .001) (Table 2). Patients with IMCES ≥4 were also more likely to present with abdominal pain (52.0% vs 40.6%; P = .011) and positive lactoferrin levels at baseline (92.3% vs 80.3%; P = .001). IMCES ≥4 was associated with the need for steroids (89.7% vs 73.7%; P < .0001) and SIT (73.5% vs 47.3%; P < .0001), an increased rate of hospitalization for colitis (76.8% vs 65.5%; P = .007), and longer hospitalization duration (median of 7 vs 6 days; P = .016). The aforementioned differences became more pronounced after considering ulceration a mandatory criterion for scores ≥4 (Table 3). No differences in ICI resumption rates were found between both groups, regardless of the presence of ulceration (P > .05). Similar findings also were noted in a subgroup analysis of only the patients not included in the original study by Wang et al16 to develop IMCES (Supplementary Table 2, available online at www.igiejournal.org).
Table 2.
Clinical features in patients with IMC and IMCES <4 or IMCES ≥4 (n = 807)
| Characteristic | n (%) |
P value | |
|---|---|---|---|
| IMCES <4 (n = 651) | IMCES ≥4 (n = 156) | ||
| CTCAE grade diarrhea | .002 | ||
| 1 | 107 (17.3) | 11 (7.3) | |
| ≥2 | 510 (82.7) | 140 (92.7) | |
| CTCAE grade colitis | <.0001 | ||
| 1 | 319 (51.5) | 47 (31.5) | |
| ≥2 | 300 (48.5) | 102 (68.5) | |
| Presenting symptom | |||
| Diarrhea | 610 (97.6) | 152 (100) | .054 |
| Abdominal pain | 254 (40.6) | 79 (52.0) | .011 |
| Fever | 65 (10.4) | 25 (16.4) | .037 |
| Blood and mucus in stool | 99 (15.8) | 39 (25.7) | .005 |
| Baseline calprotectin levels, median (IQR)∗ | 249 (77.8-781.5) | 303 (99.3-682) | .467 |
| Positive lactoferrin levels at baseline | 408 (80.3) | 120 (92.3) | .001 |
| Treatment | |||
| Steroids | 479 (73.7) | 139 (89.7) | <.0001 |
| SIT† | 307 (47.3) | 114 (73.5) | <.0001 |
| FMT | 43 (6.6) | 18 (11.7) | .033 |
| IV steroids needed | 238 (37.5) | 80 (53.3) | <.0001 |
| Duration of steroid treatment, days, median (IQR) | 30 (21-56.7) | 35 (24.2-60) | .110 |
| Multiple SIT agents needed | 58 (18.9) | 17 (14.9) | .343 |
| Number of SIT doses, median (IQR) | 3 (2-4) | 3 (2-4) | .615 |
| Outcomes | |||
| Clinical improvement or remission | 612 (94.6) | 142 (91.6) | .160 |
| Duration of IMC symptoms, days, median (IQR) | 34 (12-76) | 35 (14.2-79.2) | .259 |
| Hospitalization for colitis | 426 (65.5) | 119 (76.8) | .007 |
| Length of hospitalization, days, median (IQR) | 6 (4-9) | 7 (5-11) | .016 |
| Multiple hospitalizations | 171 (40.4) | 51 (42.9) | .634 |
| ICI therapy discontinued | 498 (78.3) | 137 (89.5) | .002 |
| ICI therapy resumed after discontinuation | 172 (34.7) | 39 (28.7) | .189 |
| All-cause mortality | 293 (45.2) | 61 (39.1) | .167 |
| Length of follow-up, years, median (IQR) | 1.5 (0.4-3.3) | 1.6 (0.4-3.5) | .512 |
CTCAE, Common Terminology Criteria for Adverse Events; FMT, fecal microbiota transplantation; ICI, immune checkpoint inhibitor; IMC, immune-mediated colitis; IMCES, immune-mediated colitis endoscopic score; IQR, interquartile range; IV, intravenous; SIT, selective immunosuppressive therapy.
The reference range for calprotectin levels is <50.0 μg/g.
SITs used to treat IMC included infliximab, vedolizumab, and ustekinumab.
Table 3.
Clinical features in patients with IMC with IMCES <4 (with or without ulceration) or IMCES ≥4 with ulceration (n = 731)∗
| Characteristic | n (%) |
P value | |
|---|---|---|---|
| IMCES <4 (n = 651) | IMCES ≥4 (n = 156) | ||
| CTCAE grade diarrhea | .003 | ||
| 1 | 107 (17.3) | 3 (3.9) | |
| ≥2 | 510 (82.7) | 73 (96.1) | |
| CTCAE grade colitis | <.0001 | ||
| 1 | 319 (51.5) | 19 (25) | |
| ≥2 | 300 (48.5) | 57 (75) | |
| Presenting symptoms | |||
| Diarrhea | 610 (97.6) | 80 (100) | .169 |
| Abdominal pain | 254 (40.6) | 43 (55.8) | .011 |
| Fever | 65 (10.4) | 14 (18.2) | .041 |
| Blood and mucus in stool | 99 (15.8) | 24 (31.2) | .001 |
| Baseline calprotectin levels, median (IQR)† | 249 (77.8-781.5) | 340 (143.7-724.7) | .230 |
| Positive lactoferrin levels at baseline | 408 (80.3) | 58 (92.1) | .023 |
| Treatment | |||
| Steroids | 479 (73.7) | 70 (88.6) | .004 |
| SIT‡ | 307 (47.3) | 55 (69.6) | <.0001 |
| FMT | 43 (6.6) | 10 (12.8) | .046 |
| IV steroids needed | 238 (37.5) | 41 (53.2) | .008 |
| Duration of steroid treatment, days, median (IQR) | 30 (21-56.7) | 33 (22.5-61.5) | .370 |
| Multiple SIT agents needed | 58 (18.9) | 9 (16.4) | .657 |
| Number of SIT doses, median (IQR) | 3 (2-4) | 3 (2-5) | .827 |
| Outcomes | |||
| Clinical improvement or remission | 612 (94.6) | 72 (91.1) | .215 |
| Duration of IMC symptoms, days, median (IQR) | 34 (12-76) | 30 (12-69.2) | .939 |
| Hospitalization for colitis | 426 (65.5) | 60 (75.9) | .064 |
| Length of hospitalization, days, median (IQR) | 6 (4-9) | 7 (5-11) | .047 |
| Multiple hospitalizations | 171 (40.4) | 23 (38.3) | .757 |
| ICI therapy discontinued | 498 (78.3) | 71 (91) | .008 |
| ICI therapy resumed after discontinuation | 172 (34.7) | 23 (32.4) | .705 |
| All-cause mortality | 293 (45.2) | 31 (38.8) | .272 |
| Length of follow-up, years, median (IQR) | 1.5 (0.4-3.3) | 1.4 (0.4-3.5) | .953 |
CTCAE, Common Terminology Criteria for Adverse Events; FMT, fecal microbiota transplantation; ICI, immune checkpoint inhibitor; IMC, immune-mediated colitis; IMCES, immune-mediated colitis endoscopic score; IQR interquartile range; IV, intravenous; SIT, selective immunosuppressive therapy.
Seventy-six patients had an IMCES ≥4 but did not have ulceration and were therefore excluded from this analysis.
The reference range for calprotectin levels is <50.0 μg/g.
SITs used to treat IMC included infliximab, vedolizumab, and ustekinumab.
Comparison of endoscopic characteristics
A comparison of endoscopic and histologic features between patients with IMCES <4 and those with IMCES ≥4 (in the subgroup of patients not included in the Wang et al16 study) can be found in Supplementary Table 3, available online at www.igiejournal.org. All gross endoscopic features were more commonly seen in patients with IMCES ≥4 (P < .0001), except for large ulceration (P = .056). Ulcers, in particular, were more commonly seen among patients with IMCES ≥4 compared with patients with IMCES <4 (51.1% vs 10.7%; P < .0001). There was no significant difference in the distribution of inflammation between the 2 groups, with the descending colon being the most common site of inflammation in approximately 75% of patients in both groups, followed by the rectum and ascending colon (about 60% each; P > .05). Compared with patients with IMCES ≥4, patients with IMCES <4 tended to have greater rates of normal findings (17.4% vs 1.2%; P < .0001) and microscopic colitis (9.6% vs 0%; P < .0001) on histologic evaluation.
Specificity analysis
A comparison of the specificity of diarrhea severity, colitis severity, and IMCES ≥4 for SIT use among the subgroup of patients not included in the Wang et al16 study can be found in Table 4. In this subgroup, we found that CTCAE grade ≥2 diarrhea and CTCAE grade ≥2 colitis had specificities of 26.9% and 64.6%, respectively. IMCES ≥4 had a high specificity of 91.6% for SIT use, which increased to 93.8% after considering ulceration a mandatory criterion. Flexible sigmoidoscopy achieved a similarly high specificity of 95%, albeit with a sample size of only 20. When including the overall cohort, similar findings were observed (Supplementary Table 4, available online at www.igiejournal.org).
Table 4.
IMCES specificity among patients not included in the original study by Wang et al16 to develop IMCES (n = 499)
| Characteristic | SIT treatment | No SIT treatment | Specificity |
|---|---|---|---|
| Colonoscopy and flexible sigmoidoscopy | |||
| IMCES < 4 | 214 | 195 | NA |
| ICMES ≥ 4 | 68 | 18 | 91.6% |
| IMCES ≥ 4 with ulcers | 31 | 13 | 93.8% |
| Flexible sigmoidoscopy | |||
| IMCES < 4 | 14 | 19 | NA |
| ICMES ≥ 4 | 7 | 1 | 95% |
| IMCES ≥ 4 with ulcers | 4 | 0 | 100% |
| Clinical symptoms | |||
| CTCAE < 2 diarrhea | 25 | 53 | NA |
| CTCAE ≥ 2 diarrhea | 239 | 145 | 26.9% |
| CTCAE < 2 colitis | 123 | 128 | NA |
| CTCAE ≥ 2 colitis | 137 | 70 | 64.6% |
CTCAE, Common Terminology Criteria for Adverse Events; IMCES, immune-mediated colitis endoscopic score; NA, not applicable; SIT, selective immunosuppressive therapy.
Univariate and multivariate logistic regression
Associations between IMCES ≥4 and various end points in univariate analysis can be found in Supplementary Table 5, available online at www.igiejournal.org. IMCES ≥4 was significantly associated with SIT use (odds ratio, 3.5; confidence interval, 2.0-6.0; P < .0001). This was also true in multivariate analysis (Table 5).
Table 5.
Multivariate analysis of variables associated with IMCES ≥4 (n = 807)
| Variable | OR (CI) | P value |
|---|---|---|
| Melanoma vs nonmelanoma | 1.2 (0.7-2.0) | .477 |
| CTCAE grade ≥2 vs grade <2 colitis | 1.8 (1.1-3.1) | .019 |
| CTCAE grade≥ 2 vs grade <2 diarrhea | 1.7 (0.6-4.2) | .258 |
| Steroid use vs no steroid use | 1.0 (0.4-2.4) | .856 |
| SIT therapy vs no SIT therapy | 2.3 (1.1-4.5) | .013 |
| Hospitalization vs no hospitalization | 8 (0.5-1.5) | .624 |
| ICI therapy discontinuation vs. ICI therapy continuation | 2.4 (0.9-5.3) | .079 |
CI, Confidence interval; CTCAE, Common Terminology Criteria for Adverse Events; ICI, immune checkpoint inhibitor; IMCES, Immune-Mediated Colitis Endoscopy Score; OR, odds ratio; SIT, selective immunosuppressive therapy.
Discussion
Endoscopic scoring systems to assess disease severity in IMC have yet to be developed, and current clinical practice relies on such systems used to evaluate inflammatory bowel disease. However, inflammatory bowel disease systems fail to account for the full spectrum of macroscopic manifestations of IMC. Building on previous studies, Wang et al16 proposed the IMCES as a tailored approach to gauge IMC severity and stratify patients by their need for SIT. The current study aimed to validate the proposed model using a retrospective cohort. We found that IMCES ≥4 was associated with the need for SIT use and with worse outcomes, including a greater need for aggressive treatments as well as more frequent and longer hospitalizations. These results suggest that the IMCES could be a useful clinical tool in the management of IMC to help estimate prognosis and help guide treatment.
Endoscopy plays a pivotal role in visualizing gastrointestinal disease and is a cornerstone in the evaluation and surveillance of inflammatory colitis. A common issue with endoscopic assessment is its inherent subjectivity and potential for interrater variability, which can be impacted by individual endoscopist training and experience. Scoring systems provide a standardized framework to report endoscopic findings that can help avoid such subjectivity and increase endoscopic reproducibility and reliability.17 These systems can be validated and are frequently associated with disease outcomes. For instance, among patients with Crohn’s disease, the Simple Endoscopic Score for Crohn’s Disease has been associated with biologic-refractory disease,18 recurrence after colectomy,19 and endoscopic remission.20 Similarly, among patients with ulcerative colitis, the Ulcerative Colitis Endoscopic Index of Severity scoring system has been found to predict biologic-refractory disease,21 the need for colectomy,22 and clinical remission.23 The strength and utility of these systems lie in their specificity for their respective disease. The MES is the most recommended and widely used endoscopic score for ulcerative colitis, in large part due to its simplicity.24 That said, it is outperformed by the more detailed and specific Ulcerative Colitis Endoscopic Index of Severity.22, 23, 24 A similar trend can be seen for IMC, with the IMCES, compared with the MES, being a more sensitive predictor of SIT use.16 The MES is currently the most pervasive scoring system used for IMC3,14,25 but has serious limitations (ie, its failure to capture the full extent of IMC disease manifestations12,26,27 and its omission of key features associated with disease outcomes12). For this reason, a more-specific scoring system tailored to IMC is needed to aid in the risk stratification of patients with regard to disease course and outcomes.
The most recent clinical guidelines for IMC acknowledge 2 scoring systems for the endoscopic evaluation of IMC: the MES and the MD Anderson score.3 Endoscopic evaluation is particularly important in this disease; timely endoscopy has been associated with shorter duration of symptoms and steroid use, as well as earlier SIT introduction,12 and specific features seen on endoscopy have been associated with disease outcomes.12,15,26 In particular, high-risk features, which are overlooked in the prevailing MES, such as ulcers deeper than 2 mm or larger than 1 cm in surface area or extensive colitis, were associated with steroid-refractory disease and the need for SIT.12,26 The IMCES is an adaptation of the MD Anderson score that includes the aforementioned high-risk features as well as elements adopted from various other inflammatory bowel disease scoring systems, offering the first and only IMC-specific scoring system to date.16 On the one hand, IMCES ≥4 was associated with the need for SIT, which is an important finding, as early introduction of SIT has been associated with favorable outcomes.28 On the other hand, clinical symptom severity by CTCAE criteria has a very low specificity for SIT use, and it frequently underestimates actual disease severity.14, 15, 16 For this reason, the IMCES shows great promise for its potential clinical utility in practice. Our study, with its larger sample size, helps validate these findings, showing a consistently high specificity of IMCES for predicting SIT use, starkly in contrast with the specificity of clinical symptoms, and demonstrating an association between IMCES and disease outcomes.
There are several limitations inherent to the retrospective single-center study design. The accuracy of the data collected was restricted to what was documented in patients' medical charts. Important information on endoscopic features or treatments and outcomes could have been missed (either in the review process or because such information was not captured in our records, especially if treatment was done elsewhere). Approximately 400 patients with colitis who were initially identified were excluded owing to inadequate or total lack of documentation of endoscopic findings, which may have affected our results. Similarly, there was no central or blinded reading of endoscopy reports or tapes by independent scorers, which may have limited the interrater reliability of the available documentation. That said, as this was a single-center study, we anticipate more-consistent reporting patterns than those in a multicenter study. Given the low number of patients who underwent flexible sigmoidoscopy, we were unable to calculate an accurate specificity of the IMCES for SIT use among this group. Our findings may, therefore, not be generalizable to patients who did not undergo full colonoscopy. Moreover, selection bias and confounding factors can affect the results because we did not assess cancer-specific outcomes or quality-of-life end points. Finally, the ideal study design for the validation of a proposed score would be a prospective study. Although this was unfeasible because of time and resource constraints, our large sample size partially makes up for this shortcoming.
Conclusions
Our study to validate the IMCES scoring system for IMC showed that IMCES ≥4 had a high specificity for the need for SIT and was associated with worse outcomes. Previous research has shown that early SIT introduction correlates with favorable outcomes for patients with IMC, and a scoring system tailored to IMC risk stratification is crucial. The IMCES may be a powerful tool in this endeavor, both aiming to help identify patients who may need more aggressive treatment earlier and as a prognostic tool. Further validation in prospective studies is needed before this score can be widely adopted.
Ethics approval and consent
The MD Anderson Institutional Review Board (PA18-0472) granted ethics approval and a waiver of patient consent for the current study. The data sets generated and analyzed during the current study are available from the corresponding author upon reasonable request.
Disclosure
The following authors disclosed financial relationships: P. Grivas: Consultant for AbbVie, AstraZeneca, Asieris Pharmaceuticals, Astellas Pharma, Bicycle Therapeutics, Bristol-Myers Squibb, CG Oncology, Daiichi Sankyo, Fresenius Kabi, Gilead, ImmunityBio, Janssen, Merck KGaA, MSD, Pfizer, Roche, Strata Oncology, and Replimune. All consulting occurred within the past 2 years but is unrelated to the current work. The research institution has also received research funding from Acrivon Therapeutics, ALX Oncology, Bristol-Myers Squibb, Merck KGaA, MSD, Genentech, Gilead, and QED Therapeutics. Y. Wang: Consultant for Thornhill, Sanarentero, Sorriso, and Beigene and received research funding from 3-D Matrix and Janssen. All of these conflicts are unrelated to the current work. All other authors disclosed no financial relationships.
Acknowledgments
We thank Ashli Nguyen-Villarreal, Associate Scientific Editor, and Erica Goodoff, Senior Scientific Editor, in the Research Medical Library at The University of Texas MD Anderson Cancer Center for editing this article.
Supplementary data
Supplemental Figure 1.
Distribution of immune-mediated colitis endoscopic features.
Supplemental Table 1.
IMC characteristics, treatment, outcomes, and endoscopic features in the overall cohort (n = 807)∗
| Characteristic | n (%) |
|---|---|
| Interval between ICI initiation and development of IMC, mo, median (IQR) | 3.2 (1.4-8.3) |
| Symptoms | |
| Diarrhea | 762 (94.4) |
| Abdominal pain | 333 (41.3) |
| Blood or mucus in stool | 138 (17.1) |
| Fever | 90 (11.2) |
| Clinical improvement or remission | 754 (93.4) |
| Days of symptoms, median (IQR) | 34 (12-76) |
| Required hospitalization | 545 (67.5) |
| Length of hospitalization, days, median (IQR) | 6 (4-9) |
| ICI discontinuation | 635 (78.7) |
| ICI resumption, n = 635 | 211 (33.2) |
| Highest CTCAE grade for diarrhea | |
| 1 | 118 (14.6) |
| 2-4 | 650 (80.5) |
| Highest CTCAE grade for colitis | |
| 1 | 366 (45.4) |
| 2-4 | 402 (49.8) |
| Positive lactoferrin levels at baseline, n = 638 | 528 (65.4) |
| Baseline calprotectin levels,† median (IQR), n = 608 | 273 (80.7-745) |
| Treatment | |
| Supportive alone‡ | 110 (13.6) |
| Corticosteroids | 618 (76.6) |
| IV corticosteroids | 318 (39.4) |
| Corticosteroids alone, n = 618 | 246 (39.8) |
| Corticosteroids and SIT, n = 618 | 410 (66.3) |
| Duration of corticosteroid therapy, days, median (IQR) | 32 (22-57) |
| SIT, n = 421 | 421 (52.2) |
| Infliximab alone | 117 (27.8) |
| Vedolizumab alone | 229 (54.4) |
| Both infliximab and vedolizumab# | 57 (13.5) |
| Ustekinumab | 18 (4.2) |
| Number of SIT doses, median (IQR) | 3 (2-4) |
| FMT | 61 (7.6) |
| Endoscopic features | |
| Erythema | 403 (49.9) |
| Loss of vascularity | 261 (32.3) |
| Friability | 197 (24.4) |
| Erosions | 143 (17.7) |
| Extensive inflammation | 123 (15.2) |
| Exudate | 57 (7.1) |
| Ulceration | 160 (19.8) |
| Large ulcers, >1 cm wide | 22 (2.7) |
| Deep ulcers, >2 mm deep | 19 (2.4) |
| ≥3 ulcers | 35 (4.3) |
| Histologic features, n = 740 | |
| Normal | 114 (15.4) |
| Acute inflammation | 383 (51.7) |
| Chronic inflammation | 175 (23.6) |
| Microscopic inflammation | 68 (9.1) |
| IMCES <4 | 651 (80.7) |
| IMCES ≥4 | 156 (19.3) |
CTCAE, Common Terminology Criteria for Adverse Events; FMT, fecal microbiota transplantation; ICI, immune checkpoint inhibitor; IMC, immune-mediated colitis; IMCES, Immune-Mediated Colitis Endoscopic Score; IQR, interquartile range; IV, intravenous; SIT, selective immunosuppressive therapy.
Some patients had missing data, so they were excluded from some analyses, resulting in smaller sample sizes.
The reference range for calprotectin levels is <50.0 μg/g.
Supportive treatments include loperamide, diphenoxylate-atropine, IV hydration, and mesalamine.
SITs used to treat IMC included infliximab, vedolizumab, and ustekinumab.
Supplemental Table 2.
Colitis clinical features for patients with IMCES <4 and patients with IMCES ≥4 in the validation cohort who were not included in the Wang et al16 study to develop IMCES (ie, a validation cohort, n = 499)
| Characteristic | n (%) |
P value | |
|---|---|---|---|
| IMCES <4 (n = 411) | IMCES ≥4 (n = 88) | ||
| CTCAE grade diarrhea | .005 | ||
| 1 | 72 (18.9) | 6 (7.2) | |
| ≥2 | 308 (81.1) | 77 (92.8) | |
| CTCAE grade colitis | .001 | ||
| 1 | 222 (58.6) | 31 (38.3) | |
| ≥2 | 157 (41.1) | 50 (61.7) | |
| Presenting symptoms | |||
| Diarrhea | 378 (98.2) | 84 (100) | .249 |
| Abdominal pain | 155 (40.3) | 43 (51.2) | .044 |
| Fever | 27 (7) | 10 (11.9) | .103 |
| Blood or mucus in stool | 59 (15.3) | 20 (23.8) | .046 |
| Baseline calprotectin levels, median∗ (IQR) | 277 (83.7-936) | 323 (103.4-711) | .423 |
| Positive lactoferrin levels at baseline | 298 (82.1) | 76 (91.6) | .021 |
| Treatment | |||
| Steroids | 304 (74.1) | 77 (88.5) | .002 |
| SIT† | 214 (52.3) | 69 (79.3) | <.0001 |
| FMT | 38 (9.3) | 14 (16.1) | .050 |
| IV steroids needed | 151 (37.8) | 44 (53) | .008 |
| Duration of steroid treatment, days, median (IQR) | 29 (21-51) | 34.5 (24.7-54) | .157 |
| Multiple SIT agents needed | 38 (17.8) | 11 (15.9) | .443 |
| Number of SIT doses, median (IQR) | 3 (2-4) | 3 (2-4) | .474 |
| Outcomes | |||
| Clinical improvement or remission | 387 (94.6) | 82 (94.3) | .528 |
| Duration of IMC symptoms, days, median (IQR) | 36 (12.5-80) | 43.5 (14-102.5) | .330 |
| Hospitalization for colitis | 267 (65.1) | 63 (72.4) | .118 |
| Length of hospitalization, days, median (IQR) | 6 (4-9) | 7 (5-11) | .096 |
| Multiple hospitalizations | 83 (31.4) | 20 (31.7) | .537 |
| ICI therapy discontinued | 328 (80.4) | 81 (94.2) | .001 |
| ICI therapy resumed after discontinuation | 114 (34.8) | 20 (24.7) | .053 |
| Mortality from any cause | 175 (42.7) | 28 (31.8) | .038 |
| Length of follow-up, median (IQR), years | 1.3 (0.4-2.7) | 1.3 (0.4-2.7) | .945 |
CTCAE, Common Terminology Criteria for Adverse Events; FMT, fecal microbiota transplantation; ICI, immune checkpoint inhibitor; IMC, immune-mediated colitis; IMCES, Immune-Mediated Colitis Endoscopic Score; IQR, interquartile range; IV, intravenous; SIT, selective immunosuppressive therapy.
The reference range for calprotectin levels is <50.0 μg/g.
SITs used to treat colitis include infliximab, vedolizumab, and ustekinumab.
Supplemental Table 3.
Endoscopic features for patients with IMCES <4 and patients with IMCES ≥4 in the validation cohort (n = 499)
| Characteristic | n (%) |
P value | |
|---|---|---|---|
| IMCES <4 (n = 411) | IMCES ≥4 (n = 88) | ||
| Endoscopic features | |||
| Erythema | 155 (37.7) | 88 (100) | <.0001 |
| Loss of vascularity | 87 (21.2) | 80 (90.9) | <.0001 |
| Friability | 41 (10) | 76 (86.4) | <.0001 |
| Erosions | 16 (3.9) | 63 (71.6) | <.0001 |
| Extensive inflammation | 9 (2.2) | 22 (25) | <.0001 |
| Exudate | 21 (5.1) | 43 (48.9) | <.0001 |
| Ulceration | 44 (10.7) | 45 (51.1) | <.0001 |
| Large ulcers, >1 cm wide | 5 (1.2) | 4 (4.5) | .056 |
| Deep ulcers, >2 mm deep | 4 (1) | 5 (5.7) | .011 |
| ≥3 ulcers | 6 (1.5) | 9 (10.2) | <.0001 |
| Inflammation location | |||
| Ileum | 68 (28.5) | 14 (22.6) | .224 |
| Ascending colon | 144 (60.8) | 25 (48.1) | .064 |
| Transverse colon | 111 (46.4) | 31 (50) | .360 |
| Descending colon | 188 (78.7) | 45 (72.6) | .196 |
| Rectum | 138 (57.7) | 38 (61.3) | .361 |
| Histologic features | |||
| Normal | 67 (17.4) | 1 (1.2) | <.0001 |
| Acute inflammation | 202 (52.5) | 51 (60.7) | .105 |
| Chronic inflammation | 79 (20.5) | 32 (38.1) | .001 |
| Microscopic inflammation | 37 (9.6) | 0 (0) | <.0001 |
IMCES, Immune-Mediated Colitis Endoscopic Score.
Supplemental Table 4.
IMCES specificity for SIT use (n = 807)
| Variables | IMCES <4 | IMCES ≥4 | IMCES ≥4 with ulcers |
|---|---|---|---|
| SIT treatment | 307 | 113 | 54 |
| No SIT treatment | 342 | 41 | 24 |
| Specificity | NA | 89.3% | 93.4% |
| CTCAE <2 diarrhea | CTCAE ≥2 diarrhea | ||
| SIT treatment | 35 | 367 | NA |
| No SIT treatment | 83 | 282 | NA |
| Specificity | NA | 22.7% | NA |
| CTCAE <2 colitis | CTCAE ≥2 colitis | ||
| SIT treatment | 175 | 223 | NA |
| No SIT treatment | 189 | 179 | NA |
| Specificity | NA | 51.3% | NA |
CTCAE; Common Terminology Criteria for Adverse Events; IMCES, Immune-Mediated Colitis Endoscopic Score; SIT, selective immunosuppressive therapy; NA, not applicable.
Supplemental Table 5.
Univariate analysis of variables associated with IMCES ≥4 in the validation cohort (n = 499)
| Variable | OR (CI) | P value |
|---|---|---|
| Female vs male | 1.1 (0.7-1.8) | .574 |
| ECOG PS 0 vs 1-5 | 1.3 (0.8-2.1) | .225 |
| Melanoma vs nonmelanoma | 1.4 (0.8-2.2) | .180 |
| Cancer stages III-IV vs I-II | 0.7 (0.3-1.4) | .338 |
| CTCAE grade ≥2 vs grade <2 colitis | 2.2 (1.4-3.7) | .001 |
| CTCAE grade ≥2 vs grade <2 diarrhea | 0.3 (0.1-0.8) | .013 |
| Steroid use vs no steroid use | 2.7 (1.3-5.3) | .005 |
| SIT vs no SIT | 3.5 (2.0-6.0) | <.0001 |
| FMT vs no FMT | 1.8 (0.9-3.6) | .062 |
| Hospitalized vs not hospitalized | 1.4 (0.8-2.3) | .192 |
| Multiple hospitalizations vs no or single hospitalization | 1.0 (0.5-1.8) | .962 |
| ICI therapy discontinued vs ICI therapy not discontinued | 3.9 (1.5-10.0) | .004 |
| ICI therapy resumed vs ICI therapy not resumed | 0.6 (0.3-1.0) | .086 |
| Remission vs no remission | 0.9 (0.3-2.5) | .891 |
CI, Confidence interval; CTCAE, Common Terminology Criteria for Adverse Events; ECOG PS, Eastern Cooperative Oncology Group performance status; FMT, fecal microbiota transplantation; ICI, immune checkpoint inhibitor; IMCES, Immune-Mediated Colitis Endoscopy Score; OR, odds ratio; SIT, selective immunosuppressive therapy.
References
- 1.Tran A.N., Wang M., Hundt M., et al. Immune checkpoint inhibitor-associated diarrhea and colitis: a systematic review and meta-analysis of observational studies. J Immunother. 2021;44:325–334. doi: 10.1097/CJI.0000000000000383. [DOI] [PubMed] [Google Scholar]
- 2.Nielsen D.L., Bogh Juhl C., Chen I.M., et al. Immune checkpoint inhibitor-induced diarrhea and colitis: incidence and management. A systematic review and meta-analysis. Cancer Treat Rev. 2022;109 doi: 10.1016/j.ctrv.2022.102440. [DOI] [PubMed] [Google Scholar]
- 3.Dougan M., Wang Y., Rubio-Tapia A., et al. AGA clinical practice update on diagnosis and management of immune checkpoint inhibitor colitis and hepatitis: expert review. Gastroenterology. 2021;160:1384–1393. doi: 10.1053/j.gastro.2020.08.063. [DOI] [PubMed] [Google Scholar]
- 4.Brahmer J.R., Abu-Sbeih H., Ascierto P.A., et al. Society for Immunotherapy of Cancer (SITC) clinical practice guideline on immune checkpoint inhibitor-related adverse events. J Immunother Cancer. 2021;9 doi: 10.1136/jitc-2021-002435. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Thompson J.A., Schneider B.J., Brahmer J., et al. Management of immunotherapy-related toxicities, version 1.2019. J Natl Compr Canc Netw. 2019;17:255–289. doi: 10.6004/jnccn.2019.0013. [DOI] [PubMed] [Google Scholar]
- 6.Abu-Sbeih H., Ali F.S., Alsaadi D., et al. Outcomes of vedolizumab therapy in patients with immune checkpoint inhibitor-induced colitis: a multi-center study. J Immunother Cancer. 2018;6:142. doi: 10.1186/s40425-018-0461-4. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Dahl E.K., Abed O.K., Kjeldsen J., et al. Safety and efficacy of infliximab and corticosteroid therapy in checkpoint inhibitor-induced colitis. Aliment Pharmacol Ther. 2022;56:1370–1382. doi: 10.1111/apt.17201. [DOI] [PubMed] [Google Scholar]
- 8.Zou F., Faleck D., Thomas A., et al. Efficacy and safety of vedolizumab and infliximab treatment for immune-mediated diarrhea and colitis in patients with cancer: a two-center observational study. J Immunother Cancer. 2021;9 doi: 10.1136/jitc-2021-003277. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Shirwaikar Thomas A., Lee S.E., Shatila M., et al. IL12/23 blockade for refractory immune-mediated colitis: 2-center experience. Am J Gastroenterol. 2023;118:1679–1683. doi: 10.14309/ajg.0000000000002332. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Wang Y., Varatharajalu K., Shatila M., et al. First-line treatment of fecal microbiota transplantation for immune-mediated colitis [abstract] J Clin Oncol. 2023;41(suppl):2510. [Google Scholar]
- 11.Wang Y., Varatharajalu K., Shatila M., et al. Effect of fecal transplantation on patients’ reported outcome after immune checkpoint inhibitor colitis [abstract] J Clin Oncol. 2023;41(16_suppl):2645. [Google Scholar]
- 12.Abu-Sbeih H., Ali F.S., Luo W., et al. Importance of endoscopic and histological evaluation in the management of immune checkpoint inhibitor-induced colitis. J Immunother Cancer. 2018;6:95. doi: 10.1186/s40425-018-0411-1. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Zou F., Wang X., Oliva I.C.G., et al. Fecal calprotectin concentration to assess endoscopic and histologic remission in patients with cancer with immune-mediated diarrhea and colitis. J Immunother Cancer. 2021;9 doi: 10.1136/jitc-2020-002058. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Cheung V.T.F., Gupta T., Brown A.O., et al. Immune checkpoint inhibitor-related colitis assessment and prognosis: can IBD scoring point the way? Br J Cancer. 2020;123:207–215. doi: 10.1038/s41416-020-0882-y. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Mooradian M.J., Wang D.Y., Coromilas A., et al. Mucosal inflammation predicts response to systemic steroids in immune checkpoint inhibitor colitis. J Immunother Cancer. 2020;8 doi: 10.1136/jitc-2019-000451. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Wang Y., Abu-Sbeih F., Tang T., et al. Novel endoscopic scoring system for immune mediated colitis: a multicenter retrospective study of 674 patients. Gastrointest Endosc. 2024;100:273–282.e4. doi: 10.1016/j.gie.2024.01.024. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Tontini G.E., Bisschops R., Neumann H. Endoscopic scoring systems for inflammatory bowel disease: pros and cons. Expert Rev Gastroenterol Hepatol. 2014;8:543–554. doi: 10.1586/17474124.2014.899899. [DOI] [PubMed] [Google Scholar]
- 18.Li L., Chen R., Zhang Y., et al. A novel model based on serum biomarkers to predict primary non-response to infliximab in Crohn's disease. Front Immunol. 2021;12 doi: 10.3389/fimmu.2021.646673. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Narula N., Wong E.C.L., Dulai P.S., et al. The performance of the Rutgeerts score, SES-CD, and MM-SES-CD for prediction of postoperative clinical recurrence in Crohn's disease. Inflamm Bowel Dis. 2023;29:716–725. doi: 10.1093/ibd/izac130. [DOI] [PubMed] [Google Scholar]
- 20.Narula N., Wong E.C.L., Colombel J.-F., et al. Predicting endoscopic remission in Crohn's disease by the modified multiplier SES-CD (MM-SES-CD) Gut. 2022;71:1078–1087. doi: 10.1136/gutjnl-2020-323799. [DOI] [PubMed] [Google Scholar]
- 21.Di Ruscio M., Variola A., Vernia F., et al. Role of Ulcerative Colitis Endoscopic Index of Severity (UCEIS) versus Mayo Endoscopic Subscore (MES) in predicting patients' response to biological therapy and the need for colectomy. Digestion. 2021;102:534–545. doi: 10.1159/000509512. [DOI] [PubMed] [Google Scholar]
- 22.Xie T., Zhang T., Ding C., et al. Ulcerative Colitis Endoscopic Index of Severity (UCEIS) versus Mayo Endoscopic Score (MES) in guiding the need for colectomy in patients with acute severe colitis. Gastroenterol Rep (Oxf) 2018;6:38–44. doi: 10.1093/gastro/gox016. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Arai M., Naganuma M., Sugimoto S., et al. The ulcerative colitis endoscopic index of severity is useful to predict Medium-to long-term prognosis in ulcerative colitis patients with clinical remission. J Crohns Colitis. 2016;10:1303–1309. doi: 10.1093/ecco-jcc/jjw104. [DOI] [PubMed] [Google Scholar]
- 24.Chen H., Wu L., Wang M., et al. Use of the ulcerative colitis endoscopic index of severity and Mayo Endoscopic Score for predicting the therapeutic effect of mesalazine in patients with ulcerative colitis. Laparosc Endosc Robot Surg. 2021;4:33–39. [Google Scholar]
- 25.Vaziri H., Turshudzhyan A., Vecchio E. Immunotherapy-induced colitis: a comprehensive review of epidemiology, clinical presentation, diagnostic workup, and management plan. J Clin Gastroenterol. 2022;56:555–564. doi: 10.1097/MCG.0000000000001705. [DOI] [PubMed] [Google Scholar]
- 26.Wang Y., Abu-Sbeih H., Mao E., et al. Endoscopic and histologic features of immune checkpoint inhibitor-related colitis. Inflamm Bowel Dis. 2018;24:1695–1705. doi: 10.1093/ibd/izy104. [DOI] [PubMed] [Google Scholar]
- 27.Choi K., Abu-Sbeih H., Samdani R., et al. Can immune checkpoint inhibitors induce microscopic colitis or a brand new entity? Inflamm Bowel Dis. 2019;25:385–393. doi: 10.1093/ibd/izy240. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28.Abu-Sbeih H., Ali F.S., Wang X., et al. Early introduction of selective immunosuppressive therapy associated with favorable clinical outcomes in patients with immune checkpoint inhibitor-induced colitis. J Immunother Cancer. 2019;7:93. doi: 10.1186/s40425-019-0577-1. [DOI] [PMC free article] [PubMed] [Google Scholar]