Intraglandular Off-the-Shelf Allogeneic Mesenchymal Stem Cell Treatment in Patients with Radiation-Induced Xerostomia: A Safety Study (MESRIX-II)

Abstract No effective therapy exists for the most common long-term side effect of radiation therapy for head and neck cancer (HNC)—xerostomia. The objective was to evaluate safety and provide proof of concept for efficacy of allogeneic adipose tissue-derived mesenchymal stem/stromal cells (AT-MSCs) injected into the major salivary glands of irradiated patients. This open-label, first-in-human, phase 1b, and single-center trial was conducted with repeated measurements days 0, 1, 5, and 30 and 4 months. Eligible patients with objective and subjective signs of radiation-induced salivary gland damage after treatment of oropharyngeal squamous cell carcinoma stages I-II (UICC 8) were enrolled. Twenty-five million cryopreserved AT-MSCs were injected into each submandibular and 50 million AT-MSCs into each parotid gland. Data were collected on adverse events, unstimulated and stimulated whole saliva (UWS and SWS) flow rates and saliva composition, patient-reported outcomes (EORTC QLQ-H&N35 and Xerostomia Questionnaire [XQ]), blood samples and salivary gland scintigraphy. Data were analyzed using repeated measures linear mixed models. Ten patients (7 men, 3 women, 59.5 years [range: 45-70]) were treated in 4 glands. No treatment-related serious adverse events occurred. During 4 months, UWS flow rate increased from 0.13 mL/minute at baseline to 0.18 mL/minute with a change of 0.06 (P = .0009) mL/minute. SWS flow rate increased from 0.66 mL/minute at baseline to 0.75 mL/minute with a change of 0.09 (P = .017) mL/minute. XQ summary score decreased by 22.6 units (P = .0004), EORTC QLQ-H&N35 dry mouth domains decreased by 26.7 (P = .0013), sticky saliva 23.3 (P = .0015), and swallowing 10.0 (P = .0016). Our trial suggests treatment of the major salivary glands with allogenic AT-MSCs is safe, warranting confirmation in larger trials.


ASCs
Adipose-derived MSCs CSCC_ASC The purpose of the current study is to assess the efficacy and safety of the injection of allogeneic adipose tissue-derived mesenchymal stromal/stem cells (MSCs) from healthy donors} on radiation-induced salivary gland hypofunction and xerostomia in patients with previous oropharyngeal cancer. The project can potentially help to develop a clinically relevant treatment option for the growing number of patients suffering from xerostomia after irradiation.

Study Objective(s)
The study objectives are to: • to evaluate the safety and tolerability of treatment with allogeneic adipose tissuederived MSCs transplanted into both submandibular and parotid glands in patients with radiation-induced hyposalivation and xerostomia.
• to evaluate the feasibility and efficacy of the treatment with allogeneic adipose tissuederived MSCs transplanted into both submandibular and parotid glands assessed by change in salivary flow after one and four months compared to baseline in patients with radiation-induced hyposalivation and xerostomia.

Trial Design
The study is an investigator-initiated, prospective, single-centre, uncontrolled, open-label trial with repeated measurements performed to investigate the safety and apparent efficacy of allogeneic adipose tissue-derived mesenchymal stem cells (ASCs) as a treatment for radiationinduced hyposalivation and xerostomia in a few highly motivated patients with previous oropharyngeal cancer patients. The intention-to-treat population will represent 10 patients with severe complications and reduced quality of life after radiotherapy. The study will evaluate the safety and efficacy of allogeneic adipose-tissue-derived MSCs study drug, CSCC(50). Because the allogeneic ASC product has not been tested in human salivary glands, this study is designed as an uncontrolled, open-label Phase I study.
The first 5 patients received the study drug with 5% DMSO and the last 5 patients received the study drug with DMSO 10%. The first two participants in each DMSO-group (patient 1-2 and patient 6-7) will at the intervention (baseline) have the ASCs transplanted into both submandibular glands and the right parotid gland, come again the following day checkup and blood samples. After five days the first two patients (patient 1-2 and 6-7) returned for checkup and have the ACSCs injected into the left parotid gland and left submandibular gland.

N/A.
Intervention Model: Single group Assignment Masking: None

Sample Size and Power
The sample size for this study was not determined by formal statistical methods but was based on extent and availability of data (i.e. feasibility).

Timing of Final Analysis
The final analysis will be performed after last patient, last follow-visit (LPLV).

Confidence intervals and P Values
All 95% confidence intervals (95%CI's) and P values will be two sided. We will not apply explicit adjustments for multiplicity, rather we will analyze and interpret these exploratory data with caution. Thus, we will disclose that our exploratory findings might reveal a false discovery (i.e. possibly having a false-positive inference); i.e. recalling that the significance level represents the risk of making a false discovery in any individual test.

Adherence and Protocol Deviations
All will be described

Full Analysis Set:
The analyses will be based on the Intention to Treat (ITT) population, i.e., based on the Full Analysis Set. This principle asserts the effect of the treatment that was planned, rather than the actual treatment given. Accordingly, participants included in the treatment group will be followed up, assessed and analyzed as members of that group, irrespective of their adherence to the planned course of treatment (i.e., independent of withdrawals, cross-over phenomena, and missing data in general).
Repeated Measurements Using Mixed Models: Our analyses will be based on the ITT population, including all included patients with available data at baseline. Missing data will be handled indirectly and statistically modeled using repeated-measures linear mixed models (see below).
These models will be valid if data are 'Missing at Random' (MAR): i.e. "Any systematic difference between the missing values and the observed values can be explained by differences in observed data". Contrasts between different time points (i.e. within group analyses) will be estimated based on repeated-measures analysis mixed linear models (i.e., at t= 120 days from baseline considered the primary endpoint assessment).

As Observed Analysis Set:
For these analyses we will assume that the mechanism causing missing data may depend neither on observed data nor on the missing data; we refer to these analyses as being valid if data is missing completely at random (MCAR). We will report these with (basic) descriptive statistics (i.e. Medians and Interquartile ranges). If the MCAR assumption is valid, even an incomplete dataset will be representative for the entire dataset.
These patients could potentially be fundamentally different from those who were completing the entire 120 days with repeated visits. The characteristics that differ between patients who received the intended therapy and were assessed at the scheduled visits, and those who did not could easily influence whether the MESRIX intervention is a potentially doable treatment strategy.

Per Protocol Analysis Set:
The set of data generated by the subset of participants who complied with the protocol sufficiently to ensure that these data would be likely to exhibit the effects of treatment, according to the conceptual biomedical model. Compliance covers such considerations as exposure to treatment, availability of measurements and absence of major protocol violations.

Screening Data
Information of the screening, eligibility and recruitment will be described in the paper and illustrated in the modified CONSORT flow diagram

Eligibility
Patients who had radiation-induced xerostomia and salivary gland dysfunction following radiotherapy for an oropharyngeal squamous cell carcinoma (OPSCC) and a minimum of two of follow-up without relapse were eligible for the trial. The full list of criteria is listed below. 10. Any other disease/condition judged by the investigator to be grounds for exclusion Controls 10 age-, sex-, race-and education matched controls were included to compare saliva samples.
The controls had to fulfill the same inclusion and exclusion criteria as the patients apart from xerostomia and salivary gland dysfunction and previous OPSCC and radiotherapy.

Recruitment
The study participants were recruited through the ENT Department of Rigshospitalet and Oncology Departments at Rigshospitalet and Herlev Hospital.

Withdrawal/follow-up
The trial will be discontinued for individual participants in cases where they wish to be withdrawn from the current protocol or in exceptional circumstances where it is impossible to complete the experiment. Likewise, extraordinary events that render the project incomplete in its entirety will lead to withdrawal for all ongoing trials participants.

Baseline Patient Characteristics
Patients with prior p16+ or p16-oropharyngeal squamous cell carcinoma with radiation-induced xerostomia and dysfunction of the major saliva glands. Baseline characteristics will be presented in a table in the study publication.

Efficacy: Salivary gland function
Assessed by salivary gland 99mTc scintigraphy (performed at baseline and after 120 days from baseline)

Efficacy: Saliva composition analysis
Saliva composition of inorganic ions (performed at baseline and after 120 days)

Analysis Methods
The objectives of repeated measures designs are to make inferences about the expected values of the observations, that is, about the means of the populations from which participants are sampled. This objective is achieved by taking into account treatment and time effects in the model. Data will be analyzed using SAS, with the particular outcome variable (Yi), using a multilevel repeated measures random effects model with participants as the random effect factor based on a restricted maximum likelihood (REML) model.
For the continuous outcomes, these will be the response (dependent) variable, unstimulated and stimulated whole saliva flow rate and time (days; 5 levels) will be included as fixed effect covariates, as well as the interaction between treatment group and time; Patient ID will be handled as a random effects factor. This statistical model will hold all between-time comparisons for all assessment points up to 120 days from baseline (including baseline) and allows for evaluation of the average effect, as well as the trajectory over time from baseline to 120 days follow-up.
For the purpose of sensitivity, we will analyze the patients with missing data "as observed" comparing differences between baseline and 4-months Wilcoxon ranks test presented as Medians with interquartile ranges.

Missing Data and Outliers
The detection of outliers can be an important problem in model building, inference and analysis of a regression model like the linear mixed models we will apply. Since the presence of outliers can lead to biased estimation, misspecification of the model and inappropriate predictions.
We will apply model diagnostics (incl 'Cooks D'), in order to evaluate the distribution of studentized residuals and detect the presence of outliers and influential points; using studentized residuals and Cooks distance for detecting outliers in Y-direction plotted against the predicted estimates.

Strata and Covariates
Not applicable.

Harms
This will be assessed as elaborated on as part of the primary protocol; measuring Serious adverse event related to the intervention, Serious Adverse events, not related, and patient-reported adverse events.        Study design