Immune checkpoint inhibitors (ICPIs) represent an efficacious cancer treatment that improves survival in metastatic malignancies [1, 2]. Initially, inhibitors of cytotoxic T-lymphocyte antigen-4 (CTLA-4), followed by inhibitors of the programmed cell death receptor-1 (PD-1) and PD-ligand 1 (PD-L1), showed efficacy and improved survival in patients with melanoma [1–3]. Subsequently, ICPIs demonstrated high effectiveness in the treatment of non-melanoma solid tumors as well, such as non-small cell lung cancer and renal cell carcinoma [4, 5]. Boasting such a positive profile, the indications for ICPI therapy are expected to increase in the near future. Thus, it is critical to be well versed in the adverse effects of these drugs and their optimal management strategies.

Owing to the immunological mechanism of ICPIs, their adverse effects collectively fall under the umbrella of immune-related adverse events (irAEs). Although ICPI-induced irAEs can affect virtually any organ system, those affecting the gastrointestinal (GI) tract are among the most common severe irAEs that lead to ICPI treatment discontinuation [6–8]. Wang et al. reported in their meta-analysis that the incidence of grade 3 and 4 colitis was 9.1% with CTLA-4 monotherapy, 1.3% with PD-1/L1 therapy, and 13.6% with combination therapy [9]. GI-irAEs that affect the lower GI tract present as diarrhea, alone or with additional signs, symptoms, and diagnostic characteristics of colitis. Preliminary evidence on GI-irAEs sheds light on the endoscopic and histological characterization of this entity [10–12]. Additionally, a significant overlap has been found between the endoscopic and histological profile of GI-irAEs and that of inflammatory bowel disease (IBD), where it is used to guide IBD mangement [13]. Therefore, it is necessary to characterize the endoscopic and histological features of GI-irAEs to guide management decisions.

In regards to GI-irAE treatment, the American Society of Clinical Oncology and Society for Immunotherapy of Cancer provide their recommendations based on the Common Terminology Criteria for Adverse Events (CTCAE) grading system [6, 14]; grade 2 irAEs should prompt the initiation of corticosteroid treatment, while, grade 3 and higher indicate the need for hospitalization and consideration of a non-corticosteroid drug such as infliximab. The caveat with current recommendations is their dependence on clinical symptoms only, which may not be the most accurate measure of disease severity in all cases. Also noteworthy is that these guidelines are based on very low level evidence, which consists mainly of expert consensus and very few studies. In addition, the current recommendations do not identify surrogate markers that can aid in prompting additional non-corticosteroid therapy. Data on the appropriate and timely administration of an add-on immunosuppressive drugs such as infliximab or vedolizumab, a monoclonal antibody that has shown efficacy in the treatment of GI-irAE [15], are lacking. Endoscopic and histological findings could provide useful information to help fill the above-mentioned knowledge gaps.

We previously reported a study of 53 patients in which we characterized the endoscopic and histological features of ICPI-induced diarrhea and colitis (IDC) [10]; we found that the presence of ulceration on endoscopy was a surrogate marker for steroid-refractory IDC. However, the above-mentioned knowledge gaps, especially the lack of characterization, which affects disease course and could prompt management decisions, were left unfilled. The aim of this study was to characterize the endoscopic and histological features of IDC and assess their association with clinical characteristics and outcomes to improve upon the currently available guidelines and provide an IDC focused treatment algorithm.

This retrospective study sheds light on the importance of the endoscopic and histological characteristics of IDC and their associations with disease outcomes. We stratified endoscopic findings as high- and low-risk in terms of steroid treatment responsiveness, likewise, histological features as active or no active inflammation. This strategy differs from those documented in previously published studies [15, 16]. The rationale for this strategy stems from the endoscopic and histological overlap between IDC and IBD. It is imperative to identify endoscopic and histological factors that are associated with disease outcomes in a timely manner to pave the way for appropriate treatment recommendations.

Among one of the most notable findings of this study is that the timely performance of endoscopic evaluation (≤ 7 days) decreased the overall duration of symptoms, steroid treatment, as well as hospitalization. Remarkably, the delay in performing endoscopic evaluation (> 30 days) until failure of the first steroid tapering trial led to a delay in initiating guided management, evident by the longer duration from IDC onset to the first infliximab/vedolizumab dose. Although, there was no difference between the two groups in the overall requirement for infliximab/vedolizumab add-on therapy. Subsequently, this resulted in more recurrence and longer duration of steroids. Additionally, the delay in the initiation of infliximab/vedolizumab in the group that had delayed endoscopic evaluation with a resulting inadequate treatment of IDC could explain the finding of a similar proportion of endoscopic and histological findings among the two groups.

Interestingly, patients with active histological inflammation had earlier time to symptom onset, longer duration of symptoms, and higher disease severity. In addition, this correlation was found to have treatment implications, as patients with active histological inflammation required more intravenous corticosteroids and infliximab/vedolizumab infusions due to steroid-refractory disease. The diagnostic correlation between active inflammation and endoscopic features is also worthy of mention, where this trend was also evident in patients with high-risk endoscopic features by the need for more infliximab/vedolizumab infusions. This finding is concordant with that of a previously published series of 92 patients who had developed IDC [11]. In contrast, patients who had immune-mediated diarrhea with normal endoscopy and histology had milder disease course and less recurrence of symptoms.

As the need for immunosuppressive treatment in patients with grade 3 and 4 IDC is obvious, contrariwise to patients with grade 1 IDC, we assessed the value of endoscopy and histology in patients with grade 2 IDC separately, where there is uncertainty about immunosuppression use. Prodigiously, timely endoscopy to guide the early introduction of steroids as well as infliximab or vedolizumab was associated with shorter duration of IDC symptoms and lower recurrence rate. To confirm our observation, the presence of high-risk endoscopic features and active histological inflammation was more frequent in patients who received immunosuppression.

The utility of laboratory studies is worthy of mention as well. Fecal calprotectin and lactoferrin have been established as useful, cost effective, and non-invasive methods of identifying active inflammation [17]. In our study, calprotectin levels and qualitative lactoferrin correlated with endoscopic and histological findings. Lactoferrin was comparatively more sensitive at detecting histological than endoscopic inflammation. These non-invasive tests provide valuable information about the overall disease status and should be used in practice, particularly before a diagnostic endoscopy, to delineate their value as initial screening tests for the presence of ulcerative inflammation. Calprotectin may be considered for follow-up of disease activity according to recent American Society of Clinical Oncology recommendations [14].

There is substantial debate on the optimal surrogate marker for identifying disease remission in IBD field; it varies from resolution of symptoms to no evidence of disease on endoscopy. It was recently suggested that histological remission is a better surrogate marker of disease remission since patients who do not show any endoscopic evidence of disease tend to have persistent symptomatology and histological inflammation [18]. However, the optimal target on histological or endoscopic remission has not been established.

Although the clinically significant histological characterization of IDC adds to our understanding of this disease entity, achieving histological remission in practice could prove to be cumbersome, requiring longer therapy and follow-up surveillance, which may not be very cost effective. Hence, characterizing endoscopic features, such as the presence of high-risk features that are correlated with a histological profile that has clinical implications, may be an appropriate strategy in the management of IDC. In addition, since endoscopic characterization is correlated with the need for additional immunosuppressant treatment, early diagnostic endoscopic findings should be used to guide treatment strategies, particularly those concerning the need for additional therapy. Focus should also be diverted towards achieving endoscopic remission rather than histological remission, since there is a favorable overlap between the two entities.

The above-mentioned findings lend support to the use of a personalized treatment strategy for IDC. An index diagnostic endoscopy is crucial for characterizing disease features, even if a clinical suspicion of IDC is deemed sufficient to initiate treatment. In our cohort, ~ 10% of patients had isolated right colon/terminal ileum involvement; therefore, full extent colonoscopy should be the preferred procedure to establish IDC diagnosis. Subsequently, according to the location on the first colonoscopy, the type of follow-up procedure can be determined using either flexible sigmoidoscopy or colonoscopy. The knowledge that a patient has active histological inflammation or high-risk endoscopic features should be used to guide timely management decisions, particularly those concerning the initiation of add-on drug such as infliximab or vedolizumab. In addition, it is crucial to identify patients who are at risk of perforation and determine whether early add-on therapy can prevent such serious adverse events.

Only a few studies have been performed on the efficacy of the early introduction of immunosuppressive treatment in the setting of IDC [19, 20]; however, there is ample evidence to support the early addition of an immunosuppression strategy involving infliximab and azathioprine in the setting of IBD, with studies showing higher rates of mucosal healing compared to conventional therapy or a placebo combination [21, 22]. Owing to the similarities in the histological and endoscopic findings of IDC and IBD, such an approach may prove to be efficacious in preventing long disease course, long duration of steroid treatment, symptom recurrence, and subsequent re-hospitalization. We previously reported a trend of a lower duration of corticosteroid therapy in patients who received infliximab therapy compared to those who did not [23]. Future studies should assess whether early initiation of infliximab or vedolizumab can affect the duration of steroid therapy and result in early disease remission.

In addition to the findings mentioned above, the effect of active inflammation on patient outcomes is noteworthy, particularly the statistically significant associations between active inflammation with symptom recurrence and repeat endoscopy. Patients’ perception of outcomes is comparatively short term and is mostly focused on symptom improvement. Hence, recurring symptoms can have an adverse effect on patients’ quality of life, which is an extremely important endpoint to consider in cancer patients. In addition, recurring symptoms can prompt re-hospitalization, and intuitively, they would pave the way for a battery of in-hospital costs, which could prove to be burdensome for the patient as well as the healthcare system. Preventing re-hospitalization in patients with IDC would reduce healthcare costs, particularly since the indications for ICPI are expected to increase in the future. Last, whether endoscopic remission could be used as a surrogate marker that could prompt the resumption of ICPI therapy should be investigated in future studies.

Our study sheds light on multiple clinically significant associations that need to be further validated in future studies. Nonetheless, there are notable limitations to our study. Besides the inherent drawbacks of a retrospective design, our study may have been underpowered in certain subgroups, especially the repeat endoscopy subgroup, limiting our findings. In addition, the limited number of events in our study, particularly colonic perforation and ICU admissions, precluded further analysis. Our cohort could be biased by the fact that usually not all patients undergo endoscopic evaluation, only those with more severe disease. In patients who only underwent flexible sigmoidoscopy, extensive disease or pan colitis could not be ruled out and hence represents a source of potential bias. Last, the treatment strategies used for the management of GI-irAEs were not standardized and were based on the clinical judgement of the treating physician.

In conclusion, endoscopic evidence of high-risk features or presence of active inflammation on histological examination represent important markers of disease severity with clinical implications and should be used to devise GI-irAE-focused treatment algorithms that incorporate a more intricate degree of specificity to improve upon the currently available guidelines. In patients with histological evidence of active inflammation or endoscopic evidence of high-risk features, early initiation of add-on therapy should be given serious consideration to avoid symptom recurrence as well as re-hospitalization, thereby maintaining patients’ quality of life and improving patient outcomes. Fecal calprotectin and lactoferrin assays should be used early in the disease course to establish a trend and delineate the utility of the tests in monitoring disease. Further prospective studies are needed to define the appropriate timing for early as well as conventional combination immunosuppressive therapy. In addition, the utility of resolving active histological inflammation or high-risk endoscopic features requires further validation.