Clinical characteristics and outcomes of tyrosine kinase inhibitor-related lower GI adverse effects

Abstract

Purpose

A variety of tyrosine kinase inhibitors (TKIs) are currently approved for the treatment of solid tumors and hematological cancers. However, TKIs are often associated with gastrointestinal (GI) adverse effects (AEs), especially diarrhea. Therefore, in the present study, we aimed to describe the clinical features and outcomes of TKI-associated lower GI AEs.

Methods

This was a retrospective single-center cohort study of patients with cancer treated with TKIs from March 2016 to September 2020 who experienced diarrhea without other identifiable causes. Basic and GI AE-related characteristics and outcomes were compared using χ² and Mann–Whitney U tests.

Results

Of 2172 patients who received TKIs over the study period, we included 228 in the final analysis. Of these, 166 (72.8%) had hematological cancers. Besides diarrhea, GI symptoms included nausea (36.4%), vomiting (21.9%), abdominal pain (15.4%), and bleeding (3.1%). Symptoms were typically mild, with 209 patients (91.7%) presenting with Common Terminology Criteria for Adverse Events grade 1–2 diarrhea. Only 5 patients (2.2%) received immunosuppressants for diarrhea treatment, 83 (36.4%) received no treatment, 29 (12.7%) received antibiotics, 101 (44.3%) received supportive antidiarrheal medications, and 17 patients (7.5%) needed TKI dose reduction or cessation of TKI use. When compared with patients with hematological cancers, those with solid tumors had a higher rate of hospitalization (29.0% vs. 7.2%; p < 0.001) and mortality (75.8% vs. 43.4%; p < 0.001) but a lower rate of recurrence of GI AEs (21.0% vs. 42.8%; p = 0.003. Only 15 patients (6.6%) underwent colonoscopy, with normal endoscopic findings in 8 patients (53.3%) and nonulcerative inflammation in 5 patients (33.3%). The inflammation universally involved the left colon. Twelve of the 15 patients who underwent colonoscopy had active colitis. In the hematological cancer group, patients with acute myeloid leukemia had a lower GI AE recurrence rate than did patients with other hematological cancers (7.2% vs. 30.1%; p = 0.001).

Conclusion

Ten percent of cancer patients receiving TKIs experienced lower GI AEs, which were usually mild. Symptoms TKI-related GI adverse effects were nonspecific, often overlapping those of other cancer therapy-related GI AEs. Treatment of GI AEs was largely supportive, with limited roles for antibiotics and immunosuppressants.

Supplementary Information

The online version contains supplementary material available at 10.1007/s00432-022-04316-3.

Introduction

Small-molecule tyrosine kinase inhibitors (SM-TKIs) are agents whose discovery and development have resulted in new, efficacious targeted treatments of many hematological cancers and solid tumors. Since the U.S. Food and Drug Administration first approved imatinib in 2001 for the treatment of chronic myelogenous leukemia, this drug class has expanded to include more than 50 agents approved for the treatment of various cancers (Thomson et al. 2021). Different TKIs target several different families of receptor tyrosine kinases, including epidermal growth factor receptor (EGFR), vascular endothelial growth factor receptor, platelet-derived growth factor receptor, ErbB, human EGFR 2, and non-receptor tyrosine kinases such as mitogen-activated protein kinase kinases involved in mitogen-activated protein pathways (Roskoski 2021). Newer generations of TKIs simultaneously target multiple families of receptors (Ettrich and Seufferlein 2018; Zhao et al. 2020). By inhibiting phosphorylation of tyrosine kinase enzymes, TKIs cause downstream effects on cell growth, cell differentiation, apoptosis, and cell death, thereby targeting new pathways in cancers that previously had few treatment options.

Despite their promising effects in cancer treatment, TKIs also have a broad range of adverse effects (AEs). Many organ systems can be affected, including the skin, gastrointestinal (GI) tract, cardiovascular system, and pulmonary system (Hartmann et al. 2009). We take particular interest in GI toxicity, as studies demonstrated TKI-induced diarrhea to occur in 30–90% of patients treated with TKIs, depending on the specific TKI used (Grothey et al., 2012; Motzer et al. 2007; Rini et al. 2011; Leboulluex et al. 2012; Natale et al. 2011; Lee et al. 2012; Shah et al. 2005; Soria et al. 2018; Wu et al. 2017; Martin et al. 2017; Murthy et al. 2020). Other GI toxic effects of TKIs range from mild diarrhea, nausea, vomiting, and dyspepsia to serious complications such as colonic perforation and severe and life-threatening diarrhea and colitis (Hartmann et al. 2009; Flaig et al., 2008; Hur et al. 2008; Walraven et al., 2011). As described in the literature, TKI-induced diarrhea is generally mild (Common Terminology Criteria for Adverse Events [CTCAE] grade 1–2). However, class-wide GI effects of TKIs are poorly characterized, particularly the endoscopic findings, effective treatments, and outcomes of these effects, including severe AEs and the need for discontinuation of TKI use thus negatively affecting cancer outcome.

The mechanisms of TKI-induced diarrhea are not well understood. TKI-related AEs tend to correlate with the specific receptor tyrosine kinase family that is targeted. For example, EGFR-targeted TKIs have a propensity to cause cutaneous AEs, whereas vascular endothelial growth factor receptor-targeted TKIs are more likely to cause vascular AEs such as bleeding, thrombosis, and hypertension (Chen et al. 2018; Kamba and McDonald 2007). However, TKI-induced diarrhea appears to occur with essentially every family of TKI receptor tyrosine kinase targets. Luminal GI AEs appear to be the most common manifestation. The mechanisms are not clearly understood, but possibly include inhibition of GI luminal EGFRs and resultant secretory diarrhea from chloride secretion, direct toxic effects, and increase of GI inflammation (Loriot et al. 2008; Uribe et al., 1996; Lu et al. 2006). In addition, endoscopic and histological findings of TKI-induced GI luminal toxic effects have yet to be studied.

Up to now, studies of TKI-induced diarrhea as a general entity have been lacking. We aimed to provide better characterization of its clinical course, endoscopic and histological findings, and correlation with underlying cancers, treatments, and outcomes. We believed that TKI-induced diarrhea would show endoscopic and histologic signs of inflammation similar to other cancer therapies with direct toxic effects on the GI system. We also predicted that immunosuppressive medication would have a limited role in the treatment of TKI-associated diarrhea due to its poorly understood but different mechanism of action than prior cancer treatments that affect the GI tract. If findings were mild, then further diagnostics and treatments would not be necessary.

Methods

Patient selection

In this retrospective study, consecutive adult patients at a tertiary cancer center who received TKIs for cancer from March 2016 to September 2020 were evaluated. The patient selection flowchart is shown in Fig. S1. Those in whom GI AEs developed between administration of the first TKI dose and 6 months afterward (n = 372) were screened for inclusion. Diarrhea was evaluated using stool microbiology testing and/or lower endoscopy (colonoscopy or flex sigmoidoscopy). Diarrhea owing to any etiology other than TKI use (e.g., infectious, irritable bowel syndrome, inflammatory bowel disease, celiac disease, and non-TKI medications) was excluded, leaving 228 patients in our study population. Approval for this study was obtained from the Institutional Review Board at The University of Texas MD Anderson Cancer Center, and informed consent was waived given the retrospective nature of the study.

Clinical data collection

All patient data were obtained from institutional electronic medical records and pharmacy databases. Basic demographic data, type and stage of cancer, cancer treatment information, GI symptoms, CTCAE grade, and information regarding treatment of the GI AEs were extracted from the medical records. GI AEs were diagnosed by treating physicians based on clinical presentation and exclusion of other causes. When available, endoscopic findings from initial and repeat examinations were noted. Features of endoscopic findings were categorized as ulcers, nonulcerative inflammation (e.g., erythema, friability, erosions, inflammatory exudate, loss of vascular pattern, and edema), or normal. The histological findings were described as normal, acute active colitis (e.g., cryptitis, crypt abscess, apoptosis, eosinophilic infiltration, intraepithelial neutrophil infiltration), chronic active colitis (e.g., crypt architectural distortion, basal lymphoplasmacytosis, Paneth cell metaplasia), or microscopic colitis (e.g., intraepithelial lymphocytic infiltration, subepithelial collagen bands). The endoscopic and histological categorizations of inflammation and colitis were based on our prior endoscopic evaluation of immune checkpoint inhibitor (ICI)-induced colitis (Wang et al. 2018).

Clinical outcomes’ assessment

Clinical outcomes of TKI-related GI AEs included requirement of hospitalization, duration of hospital stay, intensive-care unit admission, clinical remission or response after treatment, recurrent GI symptoms, need for TKI dose reduction or cessation of TKI use, complications related to GI AEs, overall survival, and all-cause mortality. Clinical remission and response were defined as resolution of GI symptoms (CTCAE grade ≤ 1) in the study window and improvement in the severity of GI symptoms (CTCAE grade decreases but still > 1), respectively. Recurrence was defined as resolution of TKI-related GI AEs within our study period, and then recurrence of symptoms again. The follow-up duration was measured from the onset of GI AEs to the date of death or last follow-up examination.

Statistical analysis

Statistical analyses were performed using SPSS software (version 24.0; IBM). Continuous variables were presented as mean (± SD) values or median values with IQRs and compared using the Wilcoxon rank-sum test or Kruskal–Wallis test. Categorical variables were described using frequencies and percentages and compared using Fisher’s exact and χ² tests. Risk factors for diarrhea recurrence and complications were assessed using logistic regression models with odds ratios and 95% CIs. All statistical tests were two-sided, and p values less than 0.05 were considered significant.

Results

Baseline patient characteristics

Baseline demographic characteristics are provided in Table 1. The median age at the time of TKI-associated GI AEs was 59 years, and the majority of the patients were White (67.1%) and male (52.2%). Hematological cancers were the most common cancers in the study population (72.8%); these cancers included acute myeloid leukemia (AML; 36.4%), acute lymphocytic leukemia (9.2%), chronic lymphocytic leukemia (8.8%), and chronic myeloid leukemia (7.0%). The most common solid tumor was lung cancer (12.3%), followed by genitourinary cancer (11.4%). Most of the patients had progressive or stable disease when the GI AEs occurred. We noted concurrent chemotherapy in 167 patients (73.2%) and concurrent ICI use in 22 patients (9.6%). The median time from initiation of treatment with TKIs to GI AE onset was 68 days (IQR, 14–180 days). The median follow-up duration was 6 months (IQR, 2–14 months). The TKI agents are summarized in Table S1 based on different categories.

Table 1.

Baseline demographic characteristics of patients with TKI-related GI AE (n = 228)

Characteristic	No. of patients (%)
Median age at time of TKI-associated GI AE diagnosis, years (IQR)	59.00 (44.75 − 68.00)
Male sex	119 (52.2)
White race	153 (67.1)
Smoking	90 (39.5)
Median Charlson Comorbidity Index, (IQR)	4 (2 − 6)
Cancer type
Melanoma	1 (0.4)
Genitourinary	26 (11.4)
Lung	28 (12.3)
Gastrointestinal/hepatobiliary	5 (2.2)
Head and neck/endocrine	2 (0.9)
Hematological	166 (72.8)
Acute myeloid leukemia	83 (36.4)
Acute lymphocytic leukemia	21 (9.2)
Chronic myeloid leukemia	16 (7.0)
Chronic lymphocytic leukemia	20 (8.8)
Mantle cell lymphoma	5 (2.2)
Diffuse large B-cell lymphoma	4 (1.8)
Myelofibrosis	9 (3.9)
Other hematological cancers*	8 (3.5)
Cancer stage
I	0 (0)
II	1 (0.4)
III	1 (0.4)
IV	60 (26.3)
Cancer staging not available	166 (72.8)
Cancer status at time of TKI-related GI AE onset
Cancer-free remission	12 (5.3)
Partial response	18 (7.9)
Stable disease	95 (41.7)
Progressive disease	103 (45.2)
TKI used
Gilteritinib	40 (17.5)
Ruxolitinib	32 (14.0)
Sorafenib	26 (11.4)
Ibrutinib	23 (10.1)
Cabozantinib	14 (6.1)
Ponatinib	18 (7.9)
Dasatinib	15 (6.6)
Osimertinib	11 (4.8)
Erlotinib	9 (3.9)
Afatinib	9 (3.9)
Lenvatinib	7 (3.1)
Acalabrutinib	4 (1.8)
Other	20 (8.8)
Concurrent non-TKI cancer therapy	167 (73.2)
Median time from initiation of TKI use to GI AE onset, days (IQR)	68 (14 − 180)
Median duration of symptoms, days (IQR)	2 (1 − 3)
Median follow-up duration between GI AE onset and last follow-up examination or death, months (IQR)	6 (2 − 14)

*Consisted of mycosis fungoides, chronic neutrophilic leukemia, rhabdomyosarcoma, polycythemia vera, myelodysplastic syndrome, primary mediastinal large B-cell lymphoma, and chronic myelomonocytic leukemia

Characteristics of TKI-related GI AEs

The characteristics of the study patients’ TKI-related GI AEs are listed in Table 2. Overall 10.5% of patients treated with TKI in the study period developed GI AEs. Symptoms were typically mild, with 209 patients (91.7%) presenting with CTCAE grade 1–2 diarrhea. Besides diarrhea, GI symptoms included nausea (n = 83 [36.4%]), vomiting (n = 50 [21.9%]), abdominal pain (n = 35 [15.4%]), and bleeding (n = 7 [3.1%]). Whereas only 5 patients (2.2%) received immunosuppressants for diarrhea treatment, 83 patients (36.4%) received no treatment, 29 patients (12.7%) received antibiotics, 101 patients (44.3%) received supportive antidiarrheal medications (e.g., loperamide), and 17 patients (7.5%) needed TKI dose reduction or cessation of TKI use. When compared with the hematological cancer group, patients with solid tumors more often had prior exposure to TKIs (19.4% vs. 7.8%; p = 0.013) and history of ICI-related AEs (6.5% vs. 0; p < 0.001). GI AEs led to more frequent TKI dose reduction or cessation of TKI use (17.7% vs. 3.6%; p < 0.001) and a higher rate of hospitalization (29.0% vs. 7.2%; p < 0.001) but a lower rate of GI AE recurrence (21.0% vs. 42.8%; p = 0.003) in patients with solid tumors than in those with hematological cancers. The overall cancer-related mortality rate was higher in the solid tumor group than in the hematological cancer group (75.8% vs. 43.4%; p < 0.001).

Table 2.

TKI-related GI AE characteristics (n = 228)

Characteristic	No. of patients (%)		P value
	Hematological cancers, n = 166	Solid tumors, n = 62
TKI used, no. (%)			< 0.001
Gilteritinib	40 (24.1)	0 (0)
Ruxolitinib	33 (19.8)	0 (0)
Sorafenib	26 (15.7)	0 (0)
Ibrutinib	23 (13.9)	0 (0)
Cabozantinib	0 (0)	14 (22.6)
Ponatinib	20 (12.0)	0 (0)
Dasatinib	15 (9.0)	0 (0)
Osimertinib	0 (0)	11 (17.7)
Erlotinib	0 (0)	9 (14.5)
Afatinib	0 (0)	9 (14.5)
Lenvatinib	0 (0)	7 (11.3)
Acalabrutinib	4 (2.4)	0 (0)
Other	5 (3.0)	12 (19.3)
History of previous TKI use	13 (7.8)	12 (19.4)	0.013
Previous irAEs related to ICI use*	0 (0)	4 (6.5)	< 0.001
Clinical symptoms†			0.932
Diarrhea	166 (100.0)	62 (100.0)
Nausea	59 (35.5)	24 (38.7)
Vomiting	34 (20.5)	16 (25.8)
Abdominal pain	24 (14.5)	11 (17.7)
Bleeding	5 (3.0)	2 (3.2)
Highest grade diarrhea			0.654
1–2	153 (92.2)	56 (90.3)
3–4	13 (7.8)	6 (9.7)
Highest grade colitis			0.385
1–2	164 (98.8)	62 (100)
3–4	2 (1.2)	0 (0)
Treatment of TKI-related GI AEs+			0.606
No treatment	56 (33.7)	27 (43.5)
Loperamide	74 (44.6)	27 (43.5)
Antibiotics	22 (13.3)	7 (11.3)
Immunosuppressants	4 (2.4)	1 (1.6)
Other§	26 (15.7)	5 (8.1)
Required TKI dose reduction or cessation of TKI use owing to GI AEs	6 (3.6)	11 (17.7)	< 0.001
Outcomes of TKI AEs
Hospitalization	12 (7.2)	18 (29.0)	< 0.001
Median hospitalization duration, days (IQR)	8 (4.75–13.25)	5 (4–7)	0.107
ICU admission	2 (1.2)	2 (3.2)	1.000
Recurrent TKI-related GI AEs	71 (42.8)	13 (21.0)	0.003
Final GI AE outcome at last follow-up			0.821
Clinical remission	150 (90.4)	56 (90.3)
Clinical response	15 (9.0)	6 (9.7)
Persistent symptoms	1 (0.6)	0 (0)
Mortality	72 (43.4)	47 (75.8)	< 0.001
Cause of death			0.495
Progression of cancer	63 (38)	43 (69.4)
Cancer-related complications	9 (5.4)	4 (6.5)
Median follow-up duration in months (IQR), n = 127	7.0 (2 − 15.8)	2.0 (1 − 9.8)	< 0.001

ICU intensive-care unit, irAEs immune-related AEs

*All four irAEs were ICI-related colitis

⁺Multiple symptoms in 107 patients

^‡Multiple treatments in 22 patients

^§Other treatments included diphenoxlate/atropine (Lomotil), hyoscyamine, tincture of opium, cholestyramine, and intravenous administration of fluids

Endoscopic and histological features of GI AEs

The study patients’ endoscopic and histological findings are summarized in Table 3. Only 15 patients (6.6%) underwent colonoscopy, with normal endoscopic findings in 8 patients (53.3%) and nonulcerative inflammation in 5 patients (33.3%); the remaining 2 patients had ulcers. The distribution of nonulcerative inflammation universally involved the left colon. Histology demonstrated acute colitis in 12 patients (80%). Of note, three patients (20%) had normal endoscopic findings.

Table 3.

Endoscopic features of TKI-related GI AEs

Characteristic	No. of patients (%)
Colonoscopy performed for diarrhea (n = 228)	15 (6.6)
Colonoscopy findings (n = 15)
Ulcer	2 (13.3)
Non-ulcerative inflammation	5 (33.3)
Normal	8 (53.3)
Distribution of inflammation (n = 7)
Both left and right colon	4 (57.1)
Left colon alone	3 (42.9)
Right colon alone	0 (0)
Isolated ileum	0 (0)
Pattern of inflammation (n = 7)
Diffuse	1 (14.3)
Segmental	2 (28.6)
Patchy	4 (57.1)
Histological features (n = 15)
Acute active colitis	9 (60.0)
Chronic active colitis	3 (20.0)
Normal	3 (20.0)
Repeat endoscopy after treatment (n = 15)	3 (20.0)
Final endoscopic remission achieved (n = 3)	3 (100.0)
Final histological remission achieved (n = 3)	3 (100.0)

Comparison of TKI-related GI AEs

When we compared the TKI-related GI AEs in patients with and without previous ICI use, we found that previous ICI exposure was not associated with a more severe colitis disease course (e.g., higher grade of diarrhea or colitis, different clinical presentation pattern, requirement of hospital admission, more aggressive medical treatment) than was lack of exposure (Table 4). However, patients with prior ICI exposure had TKI dose reduction or cessation of TKI use owing to GI AEs more frequently than did those who did not receive ICIs (p = 0.056). A majority of patients in both groups had mild GI symptoms and also experienced clinical response and remission with a short symptom duration and management.

Table 4.

TKI-related GI AEs in patients with and without previous ICI use

Characteristic	No. of patients (%)
	Previous ICI use (n = 23)	No previous ICI use (n = 205)	P value
Cancer type			< 0.001
Solid tumor	14 (60.9)	48 (23.4)
Hematological cancer	9 (39.1)	157 (76.6)
History of previous TKI use without GI AEs	4 (17.4)	21 (10.2)	0.298
Clinical symptoms			0.005
Diarrhea	23 (100.0)	205 (100.0)
Nausea	12 (52.2)	71 (34.6)
Vomiting	12 (52.2)	38 (18.5)
Abdominal pain	3 (13.0)	32 (15.6)
Bleeding	0 (0)	7 (3.4)
Highest grade of diarrhea			0.947
1–2	21 (91.3)	188 (91.7)
3–4	2 (8.7)	17 (8.3)
Highest grade of colitis			0.634
1–2	23 (100.0)	203 (99.0)
3–4	0 (0)	2 (1.0)
Treatment of TKI-related GI AEs			0.414
No treatment	10 (43.5)	73 (35.6)
Loperamide	7 (30.4)	94 (45.9)
Antibiotics	5 (21.7)	24 (11.7)
Immunosuppressants	1 (4.3)	4 (2.0)
Other*	2 (8.7)	29 (14.1)
Required TKI dose reduction or cessation of TKI use owing to GI AEs	4 (17.4)	13 (6.3)	0.056
Outcomes
Hospitalization	5 (21.7)	25 (12.2)	0.199
Median hospitalization duration, days (IQR)	6 (5–7)	5 (4–12)	0.826
ICU admission	0 (0)	4 (2.0)	0.499
Recurrent TKI-related GI AEs	6 (26.1)	78 (38.0)	0.259
Final GI AE outcome at last follow-up			0.343
Clinical remission	19 (82.6)	187 (91.2)
Clinical response	4 (17.4)	17 (8.3)
Persistent symptoms	0 (0)	1 (0.5)
Mortality	19 (82.6)	100 (48.8)	0.002
Cause of death			0.096
Progression of cancer	19 (82.6)	87 (42.4)
Cancer-related complications	0 (0)	13 (6.3)
Median overall survival duration, months (IQR)	4 (2–10)	6 (2–15)	0.516

ICU intensive-care unit

*Other treatments included Lomotil, hyoscyamine, tincture of opium, cholestyramine, and intravenous administration of fluids

TKI-related GI AEs stratified by CTCAE grade of diarrhea are summarized in Table S2. Patients with grade 2–4 diarrhea were more likely to have previous AEs related to ICI use (4.5% vs. 0.6%; p = 0.074), higher grades of colitis (p = 0.006), and higher rates of hospitalization (30.3% vs. 6.1%; p < 0.001) than were patients with grade 1 diarrhea. Furthermore, a higher proportion of the patients with grade 2–4 diarrhea needed TKI dose reduction or cessation of TKI use (15.1% vs. 4.3%; p = 0.005).

The TKI-related GI AEs in patients with AML are compared with those in patients with other hematological cancers in Table S3. Only AML patients received gilteritinib (48.2%), and a much larger proportion of AML patients than other hematological cancer patients received sorafenib (30.1% vs. 1.2%). Also, AML patients had a lower rate of recurrent TKI-related GI AEs (7.2% vs. 30.1%; p = 0.001), whereas the two groups had equal rates of clinical remission at last follow-up examination (90.4%). We found no other significant differences in TKI-related GI AEs between the two groups.

More details of the GI AEs related characteristics are included in Tables S4 and S5 based on category of TKIs and cancer types. When stratified by class, each class of SM-TKI behaves quite similarly in terms of GI AEs. The vast majority of all classes cause low-grade CTCAE grade 1–2 diarrhea and colitis, and no particular class required more aggressive treatments. Most cases across classes resolved with no treatment or loperamide. The EGFR sub-class required dose reduction or cessation (22.6%) more often than other classes of TKIs. GI AEs had similar rates of clinical remission and response. In addition, when stratified by cancer type, mild-grade diarrhea and colitis also predominate across all cancer types and most cases across all cancer groups resolved with no treatment or supportive treatment with loperamide. GU, lung, and other (melanoma, head and neck, GI) cancers, comprising the solid tumor cancers, all had higher rates of hospitalization than the hematologic malignancies. AML and ALL had highest rates of recurrent TKI-related GI toxicity among cancer groups.

Discussion

SM-TKIs represent an important and expanding class of cancer therapeutics for many advanced cancers that have proven successful. At the same time, their AE profile must be recognized and managed effectively. GI AEs are among the most common of all AEs of SM-TKIs, but their impact on outcome of patients with cancer treated with TKIs has not been well studied. The present study is the first to characterize GI AEs after treatment with TKIs on a class-wide scale, examine endoscopic findings in patients with these AEs, and describe the disease course and clinical outcomes of TKI-related GI AEs. We found that 10.5% of the study patients had TKI-related GI AEs, with the most common being mild diarrhea. Supportive treatment (antidiarrheal and hydration) is effective at providing symptom control in the majority of cases TKI-related GI AEs, with immunosuppressive medications or antibiotics usually not needed.

Mechanistically, receptor tyrosine kinases play several roles in many different pathways and affect many different organ systems. This is somewhat different from the roles of ICIs, which are among of the most commonly used cancer therapeutics and trigger overstimulation of the immune system, leading to immune-mediated damage of the GI tract (Som et al. 2019). The propensity of TKIs to cause intraluminal GI AEs is still not well understood. Researchers found that diarrhea associated with SM-TKI toxicity is secretory. One proposed mechanism of SM-TKI toxicity is implicated to be EGFR expression in normal gastrointestinal mucosa and its role in regulation of chloride secretion (Loriot et al. 2008; Uribe et al., 1996). Another proposed mechanism of SM-TKI toxicity is direct damage to intestinal mucosa. This is supported by the dose-dependent effects of SM-TKI toxicity on GI luminal AEs seen in previous studies (Lu et al. 2006; Van Glabekke et al. 2006; Demetri et al. 2002). Our endoscopic and histological findings of inflammation in the colon support the theory of direct intraluminal GI toxicity of SM-TKIs. An increase in the number of inflammatory mediators (Kanazawa et al. 2006) or toxic drug metabolites (Boussios et al. 2012) may also play a role GI AEs. Further study is needed to better understand the pathophysiology of GI toxicity related to TKIs.

Although GI AEs are common side effects of many cancer therapies, researchers have only described GI AEs related to TKI use in limited case studies of select TKIs. Systemic studies of GI AEs for the entire TKI class administered for different cancer indications are lacking. The clinical presentations of ICI-related GI AEs, which are very common, overlap those of TKI-related GI AEs, including diarrhea, abdominal pain, nausea, and vomiting. However, TKI-associated GI AEs have a lower incidence rate (10.5%) and typically have a milder grade. Patients with ICI-related colitis frequently need immunosuppressive therapy for grade 2 or higher diarrhea (Som et al. 2019; Shieh et al. 2021), whereas those with TKI-related diarrhea rarely need immunosuppressive therapy. Loperamide and supportive care appear to be sufficient to address most cases of TKI-related diarrhea. Congruently, we noted lower rates of discontinuation of treatment in patients treated with TKIs than in those treated with ICIs. Notably, our cohort included 22 patients who received a combination of ICIs and SM-TKIs. Only one reported study has examined the risk of GI toxicity of a combination of ICIs and TKIs (Alessandro et al. 2021). In a meta-analysis of four phase 3 trials, authors found a higher risk of all-grade diarrhea, higher rate of grade 3–4 diarrhea, and higher incidence of decreased appetite in patients given ICI and SM-TKI combinations than in those given SM-TKIs alone. The data on our cohort contradict this finding, as patients given combination therapy with ICIs did not experience higher grade diarrhea than did those given monotherapy. However, that could have resulted from a multitude of factors, including the small sample size of patients taking both ICI and TKI treatments and the presence of other confounders, such as different tumor types, disease stages, and other co-morbidities. Nonetheless, this highlights the need to better understand the intricacies of combination therapy with ICI and SM-TKIs, especially considering their growing use for cancer treatment (Dougan et al. 2021; Rini et al. 2019; Motzer et al. 2018, 2021; Motzer 2019).

Different classes of TKIs are known to cause specific adverse effects, though across different TKI classes, we observed similar characteristics in terms of GI AEs grade, treatment, and outcome, although the EGFR sub-class required dose adjustment or cessation more often than other classes of TKIs probably related to the proposed mechanism of direct mucosal damage through EGFR receptors in the luminal GI tract (Loriot et al. 2008; Uribe et al., 1996). Similar to our findings, dasatinib and erlotinib when used in combination were found to rarely cause high-grade diarrhea (Haura 2010). This was also consistent with findings from the EGFR family in which erlotinib was only rarely found to cause severe GI AEs and no patients enrolled in the trial required dose reductions (Mehik 2008). Similarly, VEGFR TKIs had mild-grade symptoms reported in Santoni’s study (Santoni et al., 2014).

With regard to the evaluation of mucosal inflammation, only two articles in the literature mention the use of endoscopy in cases of SM-TKI GI AEs (Boers-Sonderen et al. 2016; Saito et al. 2014); in one of these studies, researchers used endoscopy for research purposes only. In our cohort, TKI-related colitis was not frequently evaluated via endoscopy, and endoscopic findings were normal in the majority of these cases. Those with abnormal findings had nonspecific findings, such as non-ulcerous inflammation. Endoscopy has far more utility in cases of ICI colitis. Not only is it recommended for all suspected cases of ICI colitis, the spectrum of findings is much broader than in TKI-related diarrhea (Wang et al. 2018), with most patients having signs of acute active colitis and chronic inflammation upon histological evaluation. Owing to the low yield of endoscopic findings, the minimal impact of endoscopic findings on treatment strategies, and, especially, the fact that most cases are mild and require only supportive treatment, endoscopy may be deferred when TKI-related colitis is most suspected. This contrasts with cases of ICI-related colitis, in which diarrhea and colitis often can be severe and more often require therapy with immunosuppressants (Som et al. 2019). Given the high response rate with antidiarrheal treatment and lack of follow-up endoscopy, recurrence of TKI-related GI AEs is not frequently encountered, suggesting a relatively short and mild disease course for these effects.

In terms of comparison of GI AE across different cancer types, mild-grade diarrhea and colitis were predominate presentations. Santoni’s study reported that patients being treated for RCC had slightly higher rates of GI AEs than those treated for non-RCC malignancies with mostly mild diarrhea symptom which was confirmed in our study (Santoni et al., 2014). Treatment of NSCLC with TKIs required dose reduction in 25% of cases in our study due to GI AEs, which is in concordance with prior findings (Kassem 2019).

No investigators directly compared TKI-related GI AEs in solid tumor and hematological cancer patients before the present study. The literature supports that TKIs used to specifically treat hematological cancers cause lower rates of and milder diarrhea. TKIs commonly used to treat hematological cancers, such as gilteritinib, ruxolitinib, and ibrutinib, are known to cause diarrhea at rates of about 30% and severe diarrhea at rates lower than 5% (Perl et al. 2017; Verstovsek et al. 2017; Paydas 2019). TKIs used to treat solid tumors, on the other hand, such as cabozantinib, osimertinib, and erlotinib, which were some of the most frequently used TKIs in our study, have caused diarrhea in up to 90% of patients and severe diarrhea in less than 15% of cases (Elisei et al. 2013; Soria et al. 2018; Kelly et al. 2015; Rugo et al. 2019). We found that patients with solid tumors more frequently needed TKI dose reduction because of GI AEs (17.7% vs. 3.6%) and hospitalization (29.0% vs. 7.2%) than did those with hematological cancers. This may be attributable to sub-class differences in the agents used in the two patient groups. TKIs used to treat hematological cancers more often target BCR-ABL, whereas TKIs used to treat solid tumors more often target vascular endothelial growth factor receptor or EGFR. Specific mechanisms that may explain the increased incidence of GI AEs with use of EGFR-targeted TKIs include inhibition of GI epithelial growth and healing. The mechanisms of GI AEs caused by BCR-ABL– and vascular endothelial growth factor receptor-targeted TKIs are not as well understood.

Another possible explanation for these findings is that TKIs used to treat solid tumors are usually indicated for advanced late-stage or metastatic cancer. For example, cabozantinib is used to treat advanced hepatocellular carcinoma, renal cell carcinoma, and medullary thyroid cancer. Sorafenib is used to treat advanced or metastatic hepatocellular carcinoma and renal cell carcinoma. In contrast, imatinib, dasatinib, and nilotinib are usually first-line therapeutics for chronic myeloid leukemia and Philadelphia chromosome-positive acute lymphocytic leukemia (Loriot et al. 2008). This may explain the overall poorer response to treatment, including increased GI AE incidence, hospitalization, and mortality, in patients receiving TKIs for solid tumors given the more aggressive underlying disease. Among our patients with hematological cancers, the AML subgroup had lower rates of recurrent GI AEs than those with other hematologic malignancies. The reason for this observation is unclear. The use of specific TKIs for the treatment of AML (sorafenib and gilteritinib) that are not used for treatment of other hematological cancers may play a role. This suggests that AML patients have better tolerance of TKI-related GI AEs than do patients with other hematological cancers. Further studies are needed to elucidate sub-class differences in TKI-related GI AEs.

Our study had limitations given its retrospective and single-center nature. In our chart review, information on TKI-related AEs, including their severity and duration, may have been missed if the information was not accurately documented. Clinical decisions regarding initiation, cessation, and resumption of treatment with TKIs and decisions regarding the role of endoscopy were made at the discretion of the patients’ treating physicians, which could have limited our sample size, patient selection, and analysis of outcomes. The relatively small sample size also limited our ability to perform subgroup analysis. The exclusion criteria may have introduced selection bias and have the potential to affect the results of our study. Cases of diarrhea lacking a negative stool microbiology result or endoscopic findings of other causes of diarrhea were excluded for the purpose of limiting confounders. However, eligible patients who may have had true TKI-related diarrhea may have been excluded. In addition, the inclusion of patients on concurrent TKI and ICI or chemotherapy regimens could have introduced confounders to our study; however, given the prevalence of combination therapies in practice, this may serve as reference for the current cancer care.

Conclusion

SM-TKIs are well-tolerated therapeutics for various types of cancer. We found that 10% of cancer patients given SM-TKIs experienced lower GI AEs, most commonly low-grade diarrhea. Treatment of diarrhea was largely supportive, with a limited role for endoscopy or treatment with antibiotics or immunosuppressants. Further large-scale studies are warranted for clarification of TKI-specific GI AEs and their disease course and outcome.

Supplementary Information

Below is the link to the electronic supplementary material.

Abbreviations

AE

Adverse effect

CTCAE

Common Terminology Criteria for Adverse Events

EGFR

Epidermal growth factor receptor

GI

Gastrointestinal

ICI

Immune checkpoint inhibitor

SM-TKI

Small-molecule tyrosine kinase inhibitor

Author contributions

YW and AT: were the senior authors of the manuscript; they developed the concept, designed the study, interpreted the results, preserved data accuracy and integrity at all stages, were accountable for all aspects of the study, were in charge of the overall direction and planning of the study, and contributed to writing of the manuscript with input from all authors. CL and RA collected the data for the study, conducted and interpreted the data analysis, and wrote the manuscript. MS: contributed to the data analysis. NS, MA, AS, OA, and PO critically revised the final version of the manuscript. All authors read and approved the final manuscript.

Funding

This study was not supported by any funding.

Availability of data and materials

The data sets used and analyzed in this study are available from the corresponding author upon reasonable request.

Declarations

Competing interests

The authors declare no competing interests.

Conflict of interest

The authors declare that they have no conflicts of interest.

Ethics approval and consent to participate

The ethics approval for this study was granted by the MD Anderson Institutional Review Board (PA18-0472). Patient consent was waived for this study.

Consent for publication

This study was granted a waiver of patient consent.