Oral toxicities secondary to TROP2 and HER2-directed antibody drug conjugates: a multicenter retrospective cohort study

Abstract

Purpose

The aim of this multicenter retrospective cohort study was to characterize the incidence, clinical presentation, timing, severity, and management of oral toxicities (OTs) associated with TROP2-directed (datopotamab deruxtecan and sacituzumab govitecan) and HER2-directed (trastuzumab deruxtecan) antibody-drug conjugates in patients with advanced-stage cancers.

Methods

A retrospective medical-records review of 207 patients was conducted to characterize antibody–drug conjugate (ADC)-associated OTs. Patients had been treated for advanced-stage cancers treated with TROP2 (sacituzumab govitecan and datopotamab deruxtecan) and HER2-directed ADCs (trastuzumab deruxtecan) at the Sapienza University–Hospital and Miami Cancer Institute between 2024 and 2025. Multivariate logistic regressions were performed to evaluate any correlation between the ADC type and prevalence, type, severity, time of onset, and time to resolution of OTs.

Results

Overall, 47 patients (22.7%) developed OTs, with a median onset time of 7.5 days (range: 1-358). The OT prevalence was similar between TROP2 (n = 27, 23.5%) and HER2-directed (n = 20, 21.7%) ADCs, however, oral mucositis (OM) was more frequent with TROP2-directed ADCs (14.8% vs 5.4%, P = .04), whereas xerostomia (10.9% vs 6.1%, P = .31) and dysgeusia (9.8% vs 5.2%, P = .28) were more common and more severe with HER2-directed ADCs, although not reaching statistical significance. Patients receiving TROP2-directed ADCs had a 3.02-fold higher-risk of developing OM compared to those receiving HER2-directed ADCs (95% CI: 1.07-8.52, P = .040). In cases of OM, the trajectory of TROP2-associated OM was more acute than with HER-2-associated OM with earlier onset (P = .035) and quicker resolution (P = .045). Management of OTs was provided for 29 (14.0%) patients with the majority of them receiving TROP2-directed ADCs (n = 21, 17.3%). With regard to OTs, patients developing OM were associated with a greater need for management (n = 19/22; P = .001), compared to xerostomia (n = 8/17; P = .471) and dysgeusia (n = 2/15; P = .196). Ultimately, ADC dose reduction (DR) was required in 43 patients (20.8%) and was significantly more frequent with TROP2-directed ADCs (26.1% vs 14.1%, P = .039).

Conclusion

TROP2-directed ADCs are associated with a higher-risk of OM (P = .040) with earlier onset of toxicities (P = .035), and greater DR requirements (P = .039). In contrast, HER2-directed ADCs are associated with more prevalent xerostomia and dysgeusia, a higher overall severity, and later onset but potentially longer duration.

Implication for practice.

This multicenter study describes the prevalence and clinical characteristics of patients with different advanced-stage cancers who developed different oral adverse events following the initiation of sacituzumab govitecan, datopotamab deruxtecan, and trastuzumab deruxtecan. Overall, patients developed 4 types of oral toxicities including oral mucositis (OM), dry mouth/xerostomia, dysgeusia/taste changes, and oral dysesthesia. Above all, OM is a serious yet underreported adverse event, often presenting with extensive, painful, and debilitating oral ulcerations. It significantly impacts patients’ quality of life, leading to dysphagia, malnourishment, and weight loss and most importantly might compromise the patient treatment. Given the novelty and impact of TROP2 and HER2 directed-ADC-associated OM, there has been a rush to implement preventive and therapeutic strategies, most notably, the recommendation of topical steroids based on their success in mitigating mTOR inhibitor-associated stomatitis. However, in our experience, all patients failed to respond to multiple lines of topical steroid therapy. Instead, substantial clinical benefit was observed with oral cryotherapy and dose reduction.

In light of these adverse events, we believe this manuscript holds significant value for the wider oncology community as front-line physicians directly encountering such toxicities. We declare that the manuscript is original, has not been published before, and is not currently being considered for publication elsewhere.

Introduction

Cancer is one of the leading contributors to the global health burden with an estimated worldwide incidence of nearly 20 million new cases, 9.7 million deaths annually and an expected financial burden of approximately USD 25 trillion between 2020 and 2050.^1,2 Projections over the next 30 years not only foresee a yearly increase in new cancer cases but also a rise in the number of patients living with cancer, expected to reach approximately 26 million by 2040.³ One of the major contributors to cancer survivorship is the emergence of novel cancer therapies (ie, immune-checkpoint inhibitors, CAR-T cell therapy) which have revolutionized the clinical course, survival rates, and quality of life of patients with an otherwise incurable disease and who previously would have died within months.^4–6

Antibody–drug conjugates (ADCs), an expanding class of anticancer therapeutics, are emerging as having a significant role in tumor management due to their enhanced target specificity and superior efficacy compared with conventional chemotherapy, as evidenced by their favorable impact on progression-free survival and overall survival across a range of metastatic malignancies.^7,8

Antibody–drug conjugates consist of 3 structural components, a monoclonal antibody directed at a targeted tumor-associated antigen typically overexpressed on malignant cells, a conjugated cytotoxic payload, and a linker with the intent that following binding of the ADC to the tumor cell surface, the complex is internalized, and the payload is released causing cell death. The aim of this process is to induce a targeted cancer cell death while minimizing damage to surrounding normal tissues.⁹ While ADCs are engineered to achieve high precision and tumor selectivity, the presence of targeted antigens on normal tissues, “leakage” of payload agents to impact normal cells and the potential for surface-bound antibody to mediate antibody-dependent cell-mediate cytotoxicity have provided a theoretical explanation for reports of adverse events that are comparable to, or even exceed those observed with unconjugated standard chemotherapy, with adverse events reported in up to 91% of patients across all grades and approximately 46% classified as grade 3 or higher.¹⁰

Oral toxicities (OTs) associated with ADCs are also documented with frequency rates ranging from 0.3% to 90.2%, and with oral mucositis (OM; sometimes referred to as “stomatitis”), xerostomia/dry mouth, as well as dysgeusia as the most frequent OTs reported in recent trials (Table S1).^11–13 Among all ADCs, datopotamab deruxtecan (Dato-DXd) and sacituzumab govitecan (SG) 2 types of trophoblast cell surface antigen 2 (TROP2)-directed ADC, and trastuzumab deruxtecan (T-DXd) a type of human epidermal growth factor 2 (HER2)-directed ADC have shown unique toxicity profiles with a high frequency of OTs in recently published trials. Specifically, in the TROPION-PanTumor01 study, 90.2% of patients receiving 6 mg/kg of Dato-DXd intravenously once every 3 weeks experienced OM of any grade, and with 9.8% of them having OM of grade 3 or higher. In the same study, xerostomia was also reported in 12.2% of patients, and dysgeusia occurred at a rate of 2.4%.¹² On the other hand, patients receiving SG and T-DXd for metastatic disease develop lower rates of both OM (11.0% and 41.2%, respectively) and xerostomia/dry mouth (5.9% and 5.5%, respectively), but higher numbers of dysgeusia (9.7% and 13.7%, respectively).^13–16

Despite increasing recognition of ADC-associated OTs, information regarding their clinical presentation, timing, severity, and management remains underrecognized, underreported, and undertreated. This paper describes the clinical features and current management strategies of OTs in patients receiving TROP2 and HER2-directed ADCs for metastatic solid cancers in 2 different cancer centers.

Materials and methods

Study design

This was a multicenter retrospective, observational cohort study of adult patients (≥18 years) who underwent a comprehensive cancer treatment at the Division of Gynecology Oncology, Baptist Health South Florida Miami Cancer Institute (BHSF-MCI), Miami, FL, USA, and at the Breast Oncology Unit, Department of Oncological, Radiological and Pathological Sciences of the Umberto I/Sapienza University Hospital, Rome, and received TROP2 (Dato-DXd or SG) or HER2 (T-DXd) directed ADCs as second-line treatment for their advanced stage and/or recurrent cancer between July 2023 and August 2025. The study population included patients with advanced-stage cancers being treated with ADCs as part of routine standard-of-care clinical practice or through institutional clinical trial participation; this depended on the patient selection, cancer type, and drug availability. This study was reviewed and approved by the institutional review boards of Umberto I/Sapienza University Hospital and the Baptist Health Miami Cancer Institute.

Data collection

Clinical data were extracted from the BHSF-MCI (Cerner, Oracle Health) and Umberto I/Sapienza Hospital (Galileo, Dedalus Italia S.p.A., Firenze, Italy) electronic medical record system with the use of an enterprise data warehouse. The data extraction included sociodemographic information (age, gender, patient-reported race and ethnic group, and tobacco/alcohol use), concomitant medical conditions (ie, hypertension, hyperlipidemia, diabetes, etc.), cancer diagnosis and stage, as well as cancer treatment information prior to initiate ADCs. Data were also collected on ADC type (TROP2 vs HER2-directed ADCs), dosage level of the ADC and regimen, date of ADC start, and types of oral and non-OTs that were developed by each patient during treatment. ADC dose reduction (DR) secondary to oral and non-OTs was also recorded when carried out.

Oral toxicities

For the purpose of the study, data extraction included all OT that was reported in the medical records by the patient’s oncologist, or oral medicine specialist throughout the study timeline. The OT severity and grading (using the Common Terminology Criteria for Adverse Events, CTCAE version 6.0, and the visual analog scale [VAS]),^17,18 onset and duration, anatomic site (for OM only), and therapeutic management were recorded.

Statistical analysis

This analysis compared patients treated with TROP2- and HER2-directed ADCs, with group sizes indicated in table headers and results reported as counts and percentages within each treatment group. The primary focus was on OTs, including OM, xerostomia, dysgeusia, and oral dysesthesia, which were graded according to CTCAE version 6.0.¹⁹ Time to onset was defined as the number of days from the first ADC administration to the initial documented toxicity, whereas duration was calculated from toxicity onset to resolution; cases labeled as ongoing, intermittent, or recorded with non-numeric descriptors were excluded from duration analyses. Symptom severity, including pain, dryness, dysgeusia, and dysesthesia, was assessed using VAS scores and summarized using medians and ranges.

Descriptive statistics were used to characterize the study population and outcomes, with continuous variables such as age, toxicity onset, duration, and symptom scores summarized as medians with ranges and compared between ADC groups using 2-sided Mann–Whitney U tests.

Categorical variables, including toxicity incidence, DRs, anatomic involvement, and treatment categories, were summarized as counts and percentages and compared using 2-sided Fisher’s exact tests. Statistical significance was assessed at a 2-sided alpha level of 0.05, with P-values reported to 3 decimal places and values below .001 denoted accordingly; statistically significant results were marked with an asterisk in the tables. In instances where a group had no observations for a particular endpoint, P-values were reported as not applicable. P-values reported in Tables 2-4 are descriptive markers used to identify between-group differences unlikely to arise from chance alone in this cohort, and are not intended as tests of comparative efficacy given the retrospective, non-randomized design; rows in which a between-arm comparison is not calculable (drug-specific dose categories or empty comparator cells) are denoted by a slash (“/”).

Table 2.

ADC dose, regimen, and oral toxicity rates by ADC class.

Variable	TROP2 directed-ADCs (N = 115)		HER2 directed-ADCs (N = 92)		P-valuea
	N (%)	OT, n (%)b	N (%)	OT, n (%)b
Overall Ots		27 (23.5)		20 (21.7)	.868
Dosage level
TROP2: 10 mg/kg	108 (93.9)	23 (21.3)	/	/	/
TROP2: All other dose levels	7 (6.1)	4 (3.5)	/	/	/
HER2: 5.4 mg/kg	/	/	78 (84.8)	19 (24.4)	/
HER2: 4.4 mg/kg (reduced)	/	/	12 (13.0)	1(8.3)	/
HER2: All other dose levels	/	/	2 (2.2)	0 (0.0)	/
Standard dose comparisonc	10 mg/kg	23 (21.3)	5.4 mg/kg	19 (24.4)	.723
Dosing regimen
Every 3 weeks	32 (27.8)	7 (21.9)	89 (96.7)	20 (22.5)	1.000
Day 1 and 8 of 21-day cycle	82 (71.3)	19 (23.2)	3 (3.3)	0 (0.0)	1.000
Every 2 weeks	1 (0.9)	1 (100.0)	0 (0.0)	/	/

Abbreviations: ADC, antibody–drug conjugate; HER2, human epidermal growth factor receptor 2; OT, oral toxicity; TROP2, trophoblast cell-surface antigen 2.

^aFisher’s exact test, two-sided. “/” denotes comparisons that are not calculable (drug-specific dose category or empty comparator cell).

^bOT percentages use the number of patients receiving the specified dose or regimen as the denominator.

^cHead-to-head comparison at each drug’s standard licensed dose.

Table 3.

Clinicopathological characteristics of patients with toxicities from ADCs.

	N = 207 (%)	TROP2 (N = 115)	HER2 (N = 92)	P-valuea
Non-OTs
Yes	204 (98.6)	113 (98.3)	91 (98.9)	1.000
No	3 (1.4)	2 (1.7)	1 (1.1)
Median of non-OTs per patient (Range)	2 (1-5)	2 (1-5)	2 (1-5)	.975
Type of non-OTsb
Fatigue	148 (71.5)	83 (72.2)	65 (70.7)	.877
Nausea	82 (39.6)	36 (31.3)	46 (50.0)	.007*
Diarrhea	60 (29.0)	32 (27.8)	28 (30.4)	.758
Neutropenia	61 (29.5)	45 (39.1)	16 (17.4)	<.001*
Neuropathy	27 (13.0)	14 (12.2)	13 (14.1)	.684
Constipation	27 (13.0)	18 (15.7)	9 (9.8)	.299
Poor appetite	14 (6.8)	7 (6.1)	7 (7.6)	.783
Cytopenia	5 (2.4)	3 (2.6)	2 (2.2)	1.000
Cutaneous rash	2 (1.0)	2 (1.7)	0 (0.0)	.504
Dry eyes	3 (1.4)	2 (1.7)	1 (1.1)	1.000
Dry skin	4 (1.9)	3 (2.6)	1 (1.1)	.631
DR
No	164 (79.2)	85 (73.9)	79 (85.9)
Yes	43 (20.8)	30 (26.1)	13 (14.1)	.039*
OTs in patients with DR	N = 43	N = 30	N = 13
Overall	8 (18.6)	6 (20.0)	2 (15.4)	1.000
Xerostomia	5 (11.6)	4 (13.3)	1 (7.7)	1.000
Oral mucositis	4 (9.3)	4 (13.3)	0 (0.0)	.297
Dysgeusia	3 (7.0)	1 (3.3)	2 (15.4)	.213
Months of treatment before DR, median (range)	2 (1-11)	2 (1-11)	1 (1-7)	.390
Percentage of DR, median (range)	25 (10-50)	25 (10-50)	25 (15-50)	.530
Reason for DRc c	N = 43	N = 30	N = 13
Neutropenia	28 (65.1)	20 (66.7)	8 (61.5)	.742
Fatigue	7 (16.3)	3 (10.0)	4 (30.8)	.172
Oral mucositis	4 (9.3)	4 (13.3)	0 (0.0)	1.000
Cytopenia	2 (4.7)	2 (6.7)	0 (0.0)	1.000
Diarrhea	1 (2.3)	1 (3.3)	0 (0.0)	1.000
Nausea	2 (4.7)	1 (3.3)	1 (7.7)	.518
Dry eyes	1 (2.3)	1 (3.3)	0 (0.0)	1.000

Abbreviations: DR, dose reduction; OT, oral toxicity.

^aFisher’s exact test.

^bNumbers add to >N as patients had more than one non-OTs.

^cNumbers add to >N as patients had multiple reasons for DR.

^*Statistically significant (P < .05).

Table 4.

Clinical characterization of ADC-related oral toxicities.

	Total (N = 207)	TROP2 (N = 121)	HER2 (N = 92)	P-valuea
Oral toxicities (OT)	47 (22.7)	27 (23.5)	20 (21.7)	.868
Median OT/patient (range)	1.0 (1-3)	1.0 (1-3)	1.0 (1-3)	.754
Median severity (range)b	6 (1-10)	5 (1-10)	6 (2-10)	.616
Median onset days after ADCs initiation (range)	32 (1-358)	29 (1-358)	65.5 (1-337)	.883
Any treatment provided	29 (14.0)	21 (17.3)	8 (8.7)	.069
Oral mucositis (OM)	22 (10.6)	17 (14.8)	5 (5.4)	.040*
Severity of OM
Median pain score, VAS (range)	6 (1-10)	6 (1-10)	6 (5-6)	.892
Grading of OMc
Grade 0	0 (0.0)	0 (0.0)	0 (0.0)	1.000
Grade 1	3 (1.4)	2 (1.7)	1 (1.1)	1.000
Grade 2	13 (6.3)	11 (9.6)	2 (2.2)	.041*
Grade 3	1 (0.5)	1 (0.9)	0 (0.0)	1.000
n/a	5 (2.4)	3 (2.6)	2 (2.2)	1.000
Onset time of OM (median (range) days)	28 (1-337)	14 (1-332)	119 (1-337)	.355
Duration of OM median (range) days	8.5 (4-30)	17 (4-30)	8.5 (7-0.14)	1.000
Anatomic site of OM
Tongue	9 (4.3)	9 (7.8)	0 (0.0)	.005*
Buccal mucosa	2 (1.0)	1 (0.9)	1 (1.1)	1.000
Gingiva	2 (1.0)	1 (0.9)	1 (1.1)	1.000
n/a	9 (4.3)	6 (5.2)	3 (3.3)	.734
Treatment for OMb
Any treatment provided	19 (9.2)	18 (14.8)	1 (1.1)	.001*
Magic mouthwash	7 (3.4)	7 (6.1)	0 (0.0)	.018*
Topical corticosteroids	6 (2.9)	5 (4.3)	1 (1.1)	.229
Dexamethasone 0.01 mg/mL oral solution	5 (2.4)	5 (4.3)	0 (0.0)	.067
Clobetasol 0.05% oral gel	2 (1.0)	2 (1.7)	0 (0.0)	.504
Intralesional TA 40 mg/mL	2 (1.0)	1 (0.9)	1 (1.1)	1.000
Oral cryotherapy	4 (1.9)	4 (3.3)	0 (0.0)	1.000
Baking soda	2 (1.0)	1 (0.9)	1 (1.1)	1.000
ADC dose reduction	4 (1.9)	4 (3.3)	0 (0.0)	.300
Chlorhexidine 0.2% oral solution	1 (0.5)	1 (0.9)	0 (0.0)	1.000
Xerostomia	17 (8.2)	7 (6.1)	10 (10.9)	.308
Severity of xerostomia
Median dryness score, VAS (range)	3.5 (2-8)	2 (2-8)	5 (1-8)	1.000
Onset time of xerostomia—median (range) days	89 (1-358)	58 (1-358)	139.5 (1-326)	.587
Duration of xerostomia—median (range) days	14 (7-21)	14 (14-14)	14 (7-21)	1.000
Treatmentb
Any treatment provided	8 (3.9)	3 (2.6)	5 (5.4)	.471
ADC dose reduction	5 (2.4)	4 (3.3)	1 (1.1)	.393
Improve hydration	4 (1.9)	3 (2.6)	1 (1.1)	.262
OTC salivary substitutes	2 (1.0)	0 (0.0)	2 (2.2)	.497
Baking soda	2 (1.0)	0 (0.0)	2 (2.2)	.196
Dysgeusia	15 (7.2)	6 (5.2)	9 (9.8)	.282
Severity of dysgeusia
Median dysgeusia score, VAS (range)	7 (3-10)	6 (4-10)	7 (3-10)	1.000
Onset time of dysgeusia—median (range) days	23 (1-45)	23 (1-45)	N/A	N/A
Duration of dysgeusia—median (range) days	1.5 (1-2)	1 (1-1)	2 (2-2)	1.000
Treatmentb
Any treatment provided	2 (1.0)	0 (0.0)	2 (2.2)	.196
ADC dose reduction	3 (1.4)	1 (0.9)	2 (2.2)	.372
Dietary adjustments	1 (6.7)	0 (0.0)	1 (1.1)	1.000
Zinc supplements	1 (6.7)	0 (0.0)	1 (1.1)	1.000
Oral dysesthesia	4 (1.9)	2 (1.7)	2 (2.2)	1.000
Severity of oral dysesthesia
Median dysesthesia score, VAS (range)	3 (2-5)	3 (3-3)	4 (2-5)	1.000
Onset time of oral dysesthesia—median (range) days	46 (1-66)	40 (1-61)	66 (66-66)	.400
Duration of oral dysesthesia—median (range) days	6 (5-7)	6 (5-7)	N/A	N/A

Abbreviations: ADC, antibody–drug conjugate; OT, oral toxicity.

^aFisher’s exact test.

^bNumbers add to >N as patients had more than one anatomic site involvement for OM, underwent multiple treatment approaches for the OTs. OT were assessed using the US National Institutes of Health, Common Terminology Criteria for Adverse Events (CTCAE) v6.0.

^cThis data represents the median days from the initiation date of the ADC to the date when the OT was reported on medical records for the first time.

^*Statistically significant (P < .05).

Missing data were handled using pairwise deletion, whereby patients without valid data for a specific endpoint were excluded from analyses of that endpoint, and no imputation was performed. In some analyses, category totals exceeded the number of patients because multiple toxicities, anatomic sites, or reasons for DR could be recorded for a single individual; these instances were clarified in table footnotes. Dose reduction was evaluated as a binary outcome and compared between groups using Fisher’s exact tests, while the time on treatment prior to DR and the magnitude of DR were summarized using medians and ranges and compared with Mann–Whitney U tests. Reasons for DR were analyzed categorically, allowing for multiple responses per patient.

Given the exploratory nature of this study, statistical comparisons were performed to identify potential differences in OT patterns between ADC classes rather than to establish causal relationships or definitive comparative toxicity profiles. Specifically, prespecified multivariable logistic regression models were applied to binary endpoints, including the occurrence of any OT as well as specific toxicities such as OM, xerostomia, and dysgeusia. Covariates included ADC group, age greater than 65 years, stage IV disease at the time of ADC initiation, and any reported alcohol or tobacco use. Results from these models are presented as odds ratios with 95% confidence intervals and Wald P-values, without automated variable selection or interaction testing. All analyses were conducted using Python, with data processing performed in pandas, statistical testing in SciPy, and regression modeling in Statsmodels. Continuous outcomes are reported as medians with ranges, percentages are rounded to one decimal place, and P-values were not adjusted for multiple comparisons.

Results

Patients’ characteristics

A total of 207 patients were included in the study. The cohort was predominantly female (98.1%) and Hispanic (74.4%), with a median age of 62 years (range: 32-93 years). Most patients had at least one chronic comorbidity, including hypertension (n = 88, 42.5%), hyperlipidemia (n = 81, 39.1%), and depression or anxiety (n = 37, 17.9%). The majority of patients reported never using tobacco (n = 174, 84.1%) or alcohol (n = 161, 77.8%) (Table 1).

Table 1.

Sociodemographic information of patients undergoing ADC therapy.

	N = 207 (%)	TROP2 (N = 115)	HER2 (N = 92)	P-valuea
Age, median (range)	62 (32-93)	62 (32-88)	62 (34-93)	.658
Gender, n (%)				1.000
Female	203 (98.1)	113 (98.3)	90 (97.8)
Male	4 (1.9)	2 (1.7)	2 (2.2)
Race, n (%)				.204
Hispanic	154 (74.4)	89 (77.4)	65 (70.7)
Caucasian	46 (22.2)	21 (18.3)	25 (27.2)
African American	3 (1.4)	3 (2.6)	0 (0.0)
Asian	4 (1.9)	2 (1.7)	2 (2.2)
Past medical historyb				.534
Hypertension	88 (42.5)	46 (40.0)	42 (45.7)
HLD	81 (39.1)	41 (35.7)	40 (43.5)
Depression/anxiety	37 (17.9)	20 (17.4)	17 (18.5)
Osteoporosis	37 (17.9)	20 (17.4)	17 (18.5)
Hypothyroidism	26 (12.6)	21 (18.3)	5 (5.4)
CVD	24 (11.6)	10 (8.7)	14 (15.2)
DMII	24 (11.6)	14 (12.2)	10 (10.9)
GERD	19 (9.2)	11 (9.6)	8 (8.7)
CKD	4 (1.9)	3 (2.6)	1 (1.1)
Chronic gastritis	3 (1.4)	2 (1.7)	1 (1.1)
DMI	2 (1.0)	2 (1.7)	0 (0.0)
Cancer type, n (%)				.086
Breast	199 (96.1)	111 (96.5)	88 (95.7)
Uterus	3 (1.4)	3 (2.6)	0 (0.0)
Lung	2 (1.0)	0 (0.0)	2 (2.2)
Gastric	2 (1.0)	0 (0.0)	2 (2.2)
Bladder	1 (0.5)	1 (0.9)	0 (0.0)
Cancer stage, n (%)				.175
I	4 (1.9)	2 (1.7)	2 (2.2)
II	9 (4.3)	8 (7.0)	1 (1.1)
III	19 (9.2)	12 (10.4)	7 (7.6)
IV	175 (84.5)	93 (80.9)	82 (89.1)
Cancer treatment (before ADC)				.495
Surgery + RT + chemo	80 (38.6)	44 (38.3)	36 (39.1)
Chemo + Immuno	54 (26.1)	25 (21.7)	29 (31.5)
Surgery + RT + chemo + Immuno	33 (15.9)	20 (17.4)	13 (14.1)
Surgery + chemo	30 (14.5)	20 (17.4)	10 (10.9)
RT + chemo + Immuno	4 (1.9)	3 (2.6)	1 (1.1)
Chemo alone	4 (1.9)	2 (1.7)	2 (2.2)
Tobacco use				.538
Never	174 (84.1)	98 (85.2)	76 (82.6)
Former	27 (13.0)	15 (13.0)	12 (13.0)
Current	6 (2.9)	2 (1.7)	4 (4.3)
Alcohol consumption				.922
Never	161 (77.8)	89 (77.4)	72 (78.3)
Current	38 (18.4)	21 (18.3)	17 (18.5)
Former	8 (3.9)	5 (4.3)	3 (3.3)

Abbreviations: ADC, antibody–drug conjugate; Chemo, chemotherapy; CVD, cardiovascular diseases; DMI, diabetes mellitus I; DMII, diabetes mellitus II; HLD, hyperlipidemia; RT, radiation therapy.

^aFisher’s exact test.

^bNumbers add to >N as patients had multiple conditions.

Cancer characteristics and treatment

Of study patients, the overwhelming majority were being treated for breast cancer (n = 199, 96.1%), followed by uterine cancer (n = 3, 1.4%), lung cancer (n = 2, 1.0%), gastric cancer (n = 2, 1.0%), and bladder cancer (n = 1, 0.5%). At the time of ADC initiation, most patients had stage IV disease (n = 175, 84.5%), followed by stage III (n = 19, 9.2%), stage II (n = 9, 4.3%), and stage I disease (n = 4, 1.9%).

Regarding cancer management, the majority of patients received multimodal therapy, including surgical resection followed by chemoradiation (n = 80, 38.6%), or standard chemotherapy and immunotherapy (n = 54, 26.1%). The remaining patients received various combinations of these modalities (Table 1). All patients received ADCs as second-line therapy following recurrence and/or disease progression after standard treatment.

Overall, 115 patients (55.5%) received TROP2-directed ADCs, including Dato-DXd (n = 4, 1.9%) and SG (n = 111, 53.6%), administered intravenously at 10 mg/kg on days 1 and 8 of a 21-day cycle. The remaining 92 patients (44.5%) received T-DXd intravenously at 5.4 mg/kg every 3 weeks (Table 2).

Toxicity profiles of patients

Overall, 204 patients (98.6%) developed at least one toxicity following ADC initiation, with a median of 2 toxicities per patient (range: 1-5). The most frequently reported toxicities were fatigue (n = 148, 71.5%), nausea (n = 82, 39.6%), and neutropenia (n = 61, 29.5%). Nausea was more prevalent among patients receiving T-DXd (P = .007), whereas neutropenia occurred more frequently in patients receiving TROP2-directed ADCs (P < .001) (Table 3).

Treatment regimens were adjusted based on toxicity profiles. Overall, 43 patients (20.8%) required ADC DRs of 20%-25% due to oral and non-OTs, with DRs occurring more frequently in patients receiving TROP2-directed ADCs (P = .039; Table 3). Neutropenia was the most common reason for DR (n = 28, 13.5%), followed by fatigue (n = 7, 3.4%), OM (n = 2, 1.0%), and other causes (Table 3).

Characteristics of patients with OTs

A total of 47 patients (22.7%) developed OT following ADC initiation. Of these, 27 patients (23.5%) were treated with TROP2-directed ADCs, and 20 patients (21.7%) were treated with T-DXd.

Across all OTs, patients experienced a median of one toxicity (range: 1-3), with a median onset of 32 days (range: 1-358 days) and a median symptom severity score of 6 (range: 1-10; VAS). When stratified by ADC class, patients receiving TROP2-directed ADCs developed OTs earlier (median onset: 29 days; range: 1-358 days) than those receiving T-DXd (median onset: 65.5 days; range: 1-337 days). Conversely, patients receiving T-DXd reported greater symptom severity (median severity score: 6; range: 1-10) compared with those receiving TROP2-directed ADCs (median severity score: 5; range: 1-10) (Table 3). Management of OTs was provided to 29 (14.0%) patients with the majority of them receiving TROP2-directed ADCs (n = 21, 17.3%) while the remaining 8 (8.7%) receiving HER2-directed ADCs. When the management of the 2 cohorts was considered with regard to the OTs, patients receiving TROP2-directed ADCs were mostly treated for OM (n = 18/21), whereas those receiving HER2-directed ADCs were treated for xerostomia (n = 5/8) and dysgeusia (n = 2/2).

Oral mucositis

Oral mucositis was the most frequent OT, affecting 22 patients (10.6%). Oral mucositis occurred more frequently in patients receiving TROP2-directed ADCs (n = 17, 14.8%) compared with those receiving T-DXd (n = 5, 5.4%; P = .040). Most cases were moderate in severity (CTCAE grade 2; n = 13, 6.3%), with a median pain score of 6 (range: 1-10). Grade 2 OM was significantly more common in the TROP2-directed ADC group (n = 11, 9.6%; P = .041).

Lesions most frequently involved the tongue (n = 9, 4.3%; P = .005). More than half of OM cases developed within 1-60 days of ADC initiation (n = 12, 5.8%), with the majority of early-onset cases occurring in the TROP2-directed ADC group (P = .035). Most patients experienced substantial improvement within 30 days (n = 9, 4.3%), particularly those receiving TROP2-directed ADCs (n = 8, 7.0%; P = .045).

Management of OM was implemented in 14 (6.8%) patients, predominantly in the TROP2-directed ADC cohort (n = 13; P = .004). Interventions included magic mouthwash (n = 7, 3.4%; P = 0.018), topical corticosteroids (n = 6, 2.9%), oral cryotherapy (n = 4, 1.9%), baking soda rinses (n = 2, 1.0%), and chlorhexidine 0.2% mouth rinses (n = 1, 0.5%) (Table 4).

Xerostomia (dry mouth)

Xerostomia was the second most prevalent OT, affecting 17 (8.2%) patients, most of whom were treated with T-DXd (n = 10, 10.9%). The median dryness severity score was 3.5 (range: 2-8), with more severe symptoms observed in patients receiving HER2-directed ADCs.

Most cases of xerostomia developed more than 60 days after ADC initiation (n = 9, 4.3%), particularly among patients receiving T-DXd (n = 6, 6.5%). Complete symptom resolution occurred in 11 patients, most within 30 days (n = 9, 4.3%), whereas 2 patients improved within 60 days. Persistent xerostomia was reported in 3 patients (1.4%). Therapeutic management was implemented in 8 patients (3.9%), primarily in the T-DXd cohort (n = 5, 5.4%). Interventions included hydration counseling, over-the-counter salivary substitutes, and baking soda mouth rinses (Table 4).

Dysgeusia

Dysgeusia developed in 15 (7.2%) patients, most frequently among those receiving T-DXd (n = 9, 9.8%). The median severity score was 7 (range: 3-10). Taste alterations developed within 1 week of ADC initiation in 6 patients (2.9%), while 5 patients (2.4%) experienced symptom onset more than 60 days after treatment initiation.

Management was implemented in 2 patients (1.0%), both treated with T-DXd, and included dietary modifications and zinc supplementation. Most patients experienced symptom improvement within 30 days (n = 5, 2.4%; P = .016). Two patients receiving SG reported persistent dysgeusia, and one reported intermittent symptoms (Table 4).

Oral dysesthesia

Oral dysesthesia was reported in 4 patients (1.9%), including 2 treated with T-DXd and 2 with SG. The median severity score was 3 (range: 2-5). Symptom onset occurred within 1 week of ADC initiation in 1 patient and within 60 days in another; onset timing was “not reported” in the remaining cases, due to missing information. Symptoms resolved within 1 week in both SG-treated patients. No therapeutic interventions were implemented for oral dysesthesia (Table 4).

Multivariate analysis of risk factors for OTs

Multivariable logistic regression analyses were performed to identify factors associated with the development of OTs. Covariates included age (>65 years), cancer stage (stage IV vs. lower stages), tobacco and alcohol use, and ADC type. Across all models, the only significant association identified was a higher risk of OM in patients receiving TROP2-directed ADCs, with a 3.02-fold increased odds compared with those receiving HER2-directed ADCs (95% CI: 1.07-8.52; P = .040) (Table 5).

Table 5.

Multivariate analysis of risk factors for oral toxicities.

Variable	Odds ratio (95% CI)	P-value
Any oral toxicity
TROP2 vs HER2	1.10 (0.57-2.13)	.868
Age >65 years	1.04 (0.54-2.03)	1.000
Cancer stage IV	1.71 (0.62-4.71)	.365
Alcohol use (any vs none)	1.95 (0.94-4.04)	.076
Tobacco use (any vs none)	0.72 (0.28-1.87)	.651
Oral mucositis
TROP2 vs HER2	3.02 (1.07-8.52)	.040*
Age >65 years	0.86 (0.34-2.14)	.821
Cancer stage IV	0.80 (0.25-2.55)	.755
Alcohol use (any vs none)	1.99 (0.88-4.48)	.111
Tobacco use (any vs none)	1.21 (0.46-3.21)	.795
Xerostomia
TROP2 vs HER2	0.53 (0.19-1.46)	.308
Age >65 years	1.07 (0.39-2.94)	1.000
Cancer stage IV	1.41 (0.31-6.47)	1.000
Alcohol use (any vs none)	1.57 (0.64-3.87)	.327
Tobacco use (any vs none)	0.91 (0.29-2.81)	1
Dysgeusia
TROP2 vs HER2	0.51 (0.17-1.48)	.282
Age >65 years	1.36 (0.48-3.92)	.592
Cancer stage IV	6.28 (0.37-107.58)	.135
Alcohol use (any vs none)	1.30 (0.39-4.29)	.747
Tobacco use (any vs none)	2.04 (0.61-6.86)	.268

^*Statistically significant (P < .05).

Discussion

Antibody–drug conjugates are reshaping cancer care; however, despite their precise design to maximize efficacy and minimize toxicity, most administered ADC molecules are catabolized in normal tissues, resulting in unique toxicities and adverse events in over 90% of treated patients, some of which occur in the oral cavity mucosa.^20,21 The toxicity profile of ADCs is largely divided into off-target/off-tumor effects (unrelated to the intended antigen and resulting from cytotoxic effects due to premature payload deconjugation, also known as Bystander Killing Effect or BKE) and on-target/off-tumor effects, which occur when the monoclonal antibody binds to antigens expressed in nonmalignant tissues.^22–24 In contrast, a plausible mechanism underlying ADC-associated OTs, involves on-target/off-tumor effects, given the expression of TROP2 and HER2 in the normal oral and oropharyngeal epithelium and salivary glands. This expression may permit local ADC internalization, leading to payload-mediated cytotoxicity, including OM, and salivary gland dysfunction.^22,25,26 However, on-target/off-tumor toxicity may not fully explain the development of oral adverse events. Recent studies suggest that extracellular linker cleavage and premature payload release may occur within the tumor microenvironment and normal tissues, resulting in local exposure to topoisomerase-I inhibitor payloads independently of target expression. Such bystander effects could potentially contribute to epithelial and salivary gland injury within the oral cavity. Since all ADCs evaluated in the present study utilize deruxtecan- or SN-38-based topoisomerase-I inhibitor payloads, it is possible O are at least partially related to payload-specific mucosal toxicity in addition to target-mediated mechanisms.^27–29

In this multicenter retrospective study, we comprehensively characterized OTs associated with TROP2-directed (SG and Dato-DXd) and HER2-directed (T-DXd) ADCs in patients with advanced or metastatic cancer. The overall prevalence of OTs was 22.7%, with OM occurring in 10.6% of patients, xerostomia in 8.2%, dysgeusia in 7.2%, and oral dysesthesia in 1.9%.

The prevalence of OTs was similar between TROP2- and HER2-directed ADCs (23.5% vs. 21.7%); however, OM occurred more frequently in patients receiving TROP2-directed ADCs compared with HER2-directed ADCs (14.8% vs. 5.4%, P = 0.040). In recently published trials, rates of OM (any grade) vary substantially, ranging from 3.2% to 90.2%, depending on the ADC agent, dosage, and therapeutic regimen. Overall, the highest rates (48.1%-90.2%) have been reported in patients receiving Dato-DXd intravenously at 6 mg/kg once every 3 weeks, with OM described as one of the major adverse events alongside nausea, neutropenia, and fatigue.^12,30–32

Our observed OM frequency among patients receiving SG (14.8%) is consistent with results from the ASCENT-03 and IMMU-132-01 trials that demonstrated OM frequencies (any grade) of 15%, with 1% and 0% experiencing grade ≥3 OM, respectively.^33,34 Similar OM rates have been reported with T-DXd, ranging from 7.8% to 15.6%^35–37, although another study reported a higher prevalence of 35.3%, with 2.0% of patients developing grade ≥3 OM.¹³

Data regarding the onset and resolution of OM remain limited. In the TROPION-PanTumor01 trial, OM (also reported as “stomatitis”) onset occurred earlier at higher Dato-DXd doses, with median onset ranging from 15 (range, 2-205), 11 (range, 1-259), and 8 (range, 2-78) days at doses of 4, 6, and 8 mg/kg, respectively; median time to resolution was not reported.³⁸ Similarly, in the TROPION-Breast01 trial, the median onset time of OM was 22 days, with a median time to resolution of 37 days.³⁹ Our findings were comparable, although OM developed earlier in patients receiving TROP2-directed ADCs and resolved more rapidly overall, particularly in patients receiving T-DXd. In contrast to the TROPION-Breast01 trial, the time to resolution of OM in our cohort was shorter overall. Nevertheless, patients treated with TROP2-directed ADCs demonstrated a longer median time to resolution compared with those receiving T-DXd (17 days vs. 8.5 days, respectively).

In contrast to OM, ADC-induced xerostomia, dysgeusia, and oral dysesthesia remain underreported and poorly characterized in the literature, with no prior studies have reported on oral dysesthesia associated with ADCs. The highest rates of xerostomia/dry mouth were observed in patients receiving Dato-DXd in the TROPION-PanTumor01 trial, with an overall prevalence of 12.2%, although no patients experienced grade ≥3 xerostomia. In the same study, 2.4% of patients reported taste changes/dysgeusia; however, neither onset time nor time to resolution or clinical characteristics were reported.¹² Dry mouth was also reported in patients receiving SG, with frequencies ranging from 3.2% to 6.7%, based on results publicly available on ClinicalTrials.gov (NCT03725761, NCT03901339, NCT05113966) but not yet published as peer-reviewed manuscripts. Patients receiving SG also demonstrated similar frequencies of taste changes, ranging from 4.7% to 11%.^34,40

Similar results on dry mouth/xerostomia were also found in patients receiving T-DXd with numbers ranging from 4.6%,⁴¹ 5.5%,⁴² 7%,⁴³ and 8.6% (NCT04539938), and with dysgeusia from 2.5% to 13.7%.^13,44 Again, no data were reported regarding onset time, time to resolution, or clinical characteristics of xerostomia and dysgeusia. In our cohort, the prevalence of xerostomia and dysgeusia was comparable (8.2% and 7.2%, respectively); however, differences emerged between patients receiving TROP2- and HER-directed ADCs, with patients treated with T-DXd demonstrating a higher prevalence of both xerostomia (10.9% vs. 6.1%) and dysgeusia (9.8% vs. 5.2%).

Overall, 53.2% of patients in our OT patients (25/47) required clinical management, with OM accounting for the greatest management burden; when management needs were analyzed by ADC-agent, patients with TROP2-targeting agents associated OM demonstrated a significantly greater need for clinical intervention compared with those treated with T-DXd (P = 0.001).

Given their novelty, current management strategies for ADC-related OTs are limited to OM and largely determined by clinical evidence from oral adverse events (AEs) associated with targeted and chemotherapeutic agents (eg, mTOR inhibitor-related stomatitis and immune checkpoint inhibitor-associated oral AEs), following the Multinational Association of Supportive Care in Cancer/International Society of Oral Oncology (MASCC/ISOO) guidelines.^45,46 Recommended strategies for ADC-induced OM include optimal oral hygiene practices, supportive and palliative care with antiseptic or palliative mouth rinses, 1% lidocaine gels or rinses, oral moisturizers, oral cryotherapy, and dexamethasone 0.1 mg/mL oral solution, whereas there is poor to no information on management strategies for ADC-induced dry mouth, dysgeusia, and oral dysesthesia given their limited data.^11,47,48 In our cohort, patients with OM were managed using a multidisciplinary approach incorporating magic mouthwash, baking soda and salt rinses, chlorhexidine 0.2% oral solution, and topical corticosteroids such as dexamethasone 0.1 mg/mL oral solution, clobetasol 0.05% oral gel, and fluocinonide 0.05% oral gel, and intralesional triamcinolone acetonide injections (40 mg/mL), however, topical corticosteroids did not result in improvement and prevention of OM lesions. Factors that demonstrated a statistically significant positive impact on OM outcomes included “any treatment provided” (P = .001) and the use of magic mouthwash (P = .018). Additionally, oral cryotherapy was associated with symptomatic benefit in a subset of patients (n = 4, 8.5%) with TROP2-directed ADCs and appeared to reduce the development of new ulcerative lesions. Cryotherapy was delivered using a self-contained, patient-applied cryotherapy device for OM prevention; patients were instructed to place the device in a standard freezer for at least 6 hours prior to use and to apply it for 15 minutes before, during, and after ADC infusion, as well as nightly when feasible.^49,50

This study has several limitations that should be considered. First, its retrospective design may have led to underreporting of the overall prevalence of OTs (potentially explaining the partial discrepancy with the rates reported in clinical trials) and to inaccurate characterization of ADC-related OTs. Moreover, the exploratory nature of the analyses involving OTs and the comparisons between ADC classes may have introduced bias.

Another important limitation is the predominance of patients with breast cancer, as patient identification was performed across 2 large institutional divisions treating patients receiving ADCs. This may have limited the generalizability of our findings to other tumor types.

Considering the range of adverse events patients may experience following ADCs (including severe and potentially life-threatening conditions) it is reasonable that mild, non-disruptive and non-painful OTs (eg, dysgeusia and xerostomia) may be overlooked by patients and therefore not reported to clinicians unless specifically asked. Additionally, most patients were evaluated by medical oncologists and physician assistants, rather than oral medicine specialists, potentially limiting diagnostic accuracy. Moreover, most patients had undergone prior multimodal treatments, and some adverse events (ie, peripheral neuropathy, dysgeusia and dysesthesia) may have been present at the initiation of ADC therapy, potentially introducing a bias by worsening ADC-related oral side effects. Finally, although interstitial lung disease is a clinically important adverse event associated particularly with trastuzumab deruxtecan, the primary objective of the present study was characterization of OTs; therefore, only the most frequent non-oral adverse events are reported.

In this multicenter retrospective study, TROP2-directed ADCs were associated with a significantly higher risk of OM, greater need for clinical intervention, and increased rates of DR compared with HER2-directed ADCs. In contrast, HER2-directed ADCs were more frequently associated with xerostomia and dysgeusia, later onset of OTs, and comparable or shorter symptom duration.

These findings underscore the importance of proactive oral assessment and early supportive care in patients receiving ADC therapy. Larger prospective studies are warranted to further define the clinical course, quality-of-life impact, and optimal management strategies for ADC-associated OTs.

Author contributions

Paolo J. Fantozzi (Conceptualization, Data curation, Methodology, Validation, Visualization, Writing—original draft, Writing—review & editing), Stephen Sonis (Supervision, Validation, Visualization, Writing—review & editing), Andrea Botticelli (Supervision, Writing—review & editing), Simone Scagnoli (Investigation), Monica Verrico (Investigation), Giulia Bianchini (Investigation), Maria Vittoria Bonomo (Investigation), Chiara Bonadonna (Investigation), Mathilde Casagrande (Investigation), Lucia Borghetti (Investigation), Gianluca Tenore (Investigation), Sankalp Das (Data curation, Formal analysis), John P. Diaz (Investigation), Umberto Romeo (Investigation), and Alessandro Villa (Conceptualization, Data curation, Investigation, Validation, Visualization, Writing—review & editing)

Supplementary material

Supplementary material is available at The Oncologist online.

Funding

None declared.

Conflicts of interest

None declared.

Data availability

The data underlying this article will be shared on reasonable request to the corresponding author.