Phase I trial of WEE1 inhibition with chemotherapy and radiotherapy as adjuvant treatment, and a window of opportunity trial with cisplatin in patients with head and neck cancer: the WISTERIA trial protocol

Introduction Patients with head and neck squamous cell carcinoma with locally advanced disease often require multimodality treatment with surgery, radiotherapy and/or chemotherapy. Adjuvant radiotherapy with concurrent chemotherapy is offered to patients with high-risk pathological features postsurgery. While cure rates are improved, overall survival remains suboptimal and treatment has a significant negative impact on quality of life. Cell cycle checkpoint kinase inhibition is a promising method to selectively potentiate the therapeutic effects of chemoradiation. Our hypothesis is that combining chemoradiation with a WEE1 inhibitor will affect the biological response to DNA damage caused by cisplatin and radiation, thereby enhancing clinical outcomes, without increased toxicity. This trial explores the associated effect of WEE1 kinase inhibitor adavosertib (AZD1775). Methods and analysis This phase I dose-finding, open-label, multicentre trial aims to determine the highest safe dose of AZD1775 in combination with cisplatin chemotherapy preoperatively (group A) as a window of opportunity trial, and in combination with postoperative cisplatin-based chemoradiation (group B). Modified time-to-event continual reassessment method will determine the recommended dose, recruiting up to 21 patients per group. Primary outcomes are recommended doses with predefined target dose-limiting toxicity probabilities of 25% monitored up to 42 days (group A), and 30% monitored up to 12 weeks (group B). Secondary outcomes are disease-free survival times (groups A and B). Exploratory objectives are evaluation of pharmacodynamic (PD) effects, identification and correlation of potential biomarkers with PD markers of DNA damage, determine rate of resection status and surgical complications for group A; and quality of life in group B. Ethics and dissemination Research Ethics Committee, Edgbaston, West Midlands (REC reference 16/WM/0501) initial approval received on 18/01/2017. Results will be disseminated via peer-reviewed publication and presentation at international conferences. Trial registration number ISRCTN76291951 and NCT03028766.


AbSTRACT
Introduction Patients with head and neck squamous cell carcinoma with locally advanced disease often require multimodality treatment with surgery, radiotherapy and/ or chemotherapy.Adjuvant radiotherapy with concurrent chemotherapy is offered to patients with high-risk pathological features postsurgery.While cure rates are improved, overall survival remains suboptimal and treatment has a significant negative impact on quality of life.Cell cycle checkpoint kinase inhibition is a promising method to selectively potentiate the therapeutic effects of chemoradiation.Our hypothesis is that combining chemoradiation with a WEE1 inhibitor will affect the biological response to DNA damage caused by cisplatin and radiation, thereby enhancing clinical outcomes, without increased toxicity.This trial explores the associated effect of WEE1 kinase inhibitor adavosertib (AZD1775).Methods and analysis This phase I dose-finding, openlabel, multicentre trial aims to determine the highest safe dose of AZD1775 in combination with cisplatin chemotherapy preoperatively (group A) as a window of opportunity trial, and in combination with postoperative cisplatin-based chemoradiation (group B).Modified time-to-event continual reassessment method will determine the recommended dose, recruiting up to 21 patients per group.Primary outcomes are recommended doses with predefined target dose-limiting toxicity probabilities of 25% monitored up to 42 days (group A), and 30% monitored up to 12 weeks (group B).Secondary outcomes are disease-free survival times (groups A and B).Exploratory objectives are evaluation of pharmacodynamic (PD) effects, identification and correlation of potential biomarkers with PD markers of DNA damage, determine rate of resection status and surgical complications for group A; and quality of life in group B. Ethics and dissemination Research Ethics Committee, Edgbaston, West Midlands (REC reference 16/WM/0501) initial approval received on 18/01/2017.Results will be disseminated via peer-reviewed publication and presentation at international conferences.Trial registration number ISRCTN76291951 and NCT03028766.

InTRoduCTIon
Head and neck cancers are the sixth most common cancer worldwide, 1 with 12 000 new cases per year in the UK (Cancer Research UK (CRUK) head and neck cancer statistics).The most common types are head and neck squamous cell carcinom (HNSCC) arising from laryngeal, oral cavity, oropharynx and hypopharyngeal subsites. 2

Strengths and limitations of this study
► The study is one of the first to explore a WEE1 inhibitor in combination with cisplatin and radiotherapy, and the first to do so in head and neck cancer.► Allows collection of tissues for translational research to examine the pharmacodynamic effects of WEE1 and WEE1 in combination with cisplatin.► Uses an efficient time-to-event continual reassessment method design which allows dose assignments to be performed with the flexibility of continual patient accrual or temporary accrual suspension to permit sufficient accumulation of safety information as necessary.► Additional biopsies in the preoperative setting may affect the patient enrolment.► The oral administration of AZD1775 may affect recruitment in postoperative patients if they have difficulty or are unable to swallow.

Open access
Primary surgery is the standard of care for cancers of the oral cavity, and for more advanced cancers of the larynx and hypopharynx, while chemoradiation is the standard approach to oropharyngeal disease.Postoperative radiotherapy (PORT) or postoperative chemoradiation (POCRT) is often recommended in the setting of highrisk features such as a positive margin, multiple involved nodes, extracapsular spread and perineural infiltration.
The role of concomitant chemotherapy was established by the Meta-Analysis of Chemotherapy on Head and Neck Cancer in 2009. 2 This study pooled individual patient data from 93 trials involving 16 485 patients receiving either definitive local therapy alone (surgery or radiotherapy) or local therapy plus chemotherapy (induction, concomitant or adjuvant).Concomitant therapy was assessed in 50 trials involving 9605 patients; chemotherapy decreased the risk of death with an HR of 0.81, translating into a 6.5% overall survival advantage at 5 years as a result of improved disease-free survival.No benefit was seen in patients aged over 70 years.This provides robust evidence that the ability of radiotherapy to secure local control can be enhanced by systemic therapy.
Acute toxicity and long-term sequelae of POCRT have a considerable negative health impact, and despite intensive treatment, 5-year overall survival is suboptimal at 53%. 3 Locoregional relapse is particularly difficult to salvage, and local control is, therefore, closely correlated with overall survival.Morbidity imposed by local recurrence in the absence of distant metastatic disease is considerable and can be extremely challenging to palliate adequately.There is an urgent need to develop novel approaches that achieve improved locoregional disease control and reduce treatment-related morbidity for this patient group.Achieving this may translate into improved overall survival and an enhancement in patient-related outcome measures.

Chemotherapy, radiotherapy and dnA damage response
Cellular DNA damage response (DDR) is central to the preservation of genomic integrity.Cancer development is a multistep process where deregulation of the DDR contributes to phenotypes such as sustained cell proliferation and resistance to cell death.This increase in endogenous DNA damage necessitates the acquisition of compensatory mechanisms if the cell is to avoid death through uncontrolled genomic instability.Importantly, these compensatory changes may constitute a molecular 'Achilles heel' that is vulnerable to therapeutic exploitation with a new generation of targeted agents.POCRT exploits the differential DDR in malignant and normal tissues to eradicate microscopic residual disease.Ionising radiation (IR) generates a variety of biological changes within irradiated tissue, with double-stranded DNA (dsDNA) breaks being therapeutically significant.IR-induced cell death is also immunogenic, with recruitment of immune effectors to the irradiated tumour microenvironment contributing to a successful outcome. 4Platinumbased chemotherapy accentuates IR-induced cell death, in part via the suppression of homologous recombination, the primary repair mechanism for dsDNA damage in irradiated cells. 4Manipulation of the DDR pathway, therefore, represents a rational means by which the therapeutic index of POCRT might be improved.Cell cycle checkpoints are integral components of the DDR, allowing the cell to pause progression through the cell cycle to repair DNA damage.Checkpoints are regulated by a network of phosphorylation cascades.p53, encoded by TP53, is a key regulator of the G 1 /S checkpoint.TP53 mutations, which are seen in 60%-70% of HNSCC cases, 5 are sufficient to impair the function of this checkpoint and thereby create a critical reliance on the later G 2 /M checkpoint.In addition, p53 function can be inactivated by various mechanisms, including somatic and germline mutations as well as polymorphisms. 6 7harmacological abrogation of the G 2 /M checkpoint has been shown to differentially sensitise normal and tumour cells to the effect of DNA damaging agents such as cisplatin and IR. 8 WEE1 kinase and AZd1775 WEE1 kinase is a key regulator of the G 2 /M checkpoint and a promising therapeutic target.It is a serine-threonine kinase involved in phosphorylation and inactivation of cyclin-dependent kinase 1 (CDK1), the only 1 of 14 similar proteins to be indispensable for mitotic entry.Along with a number of other proteins, WEE1 triggers G 2 /M arrest in response to DNA damage.However, inhibition of WEE1 leads to high CDK1 activity, allowing cells to progress through the G 2 /M checkpoint without the opportunity to repair damaged DNA, leading to potentially catastrophic levels of unrepaired DNA damage induction of cell death. 9 10WEE1 also has an effect on CDK2 as its inhibition leads to high CDK2 activity and aberrant DNA replication, resulting in stalled replication forks and DNA double-stranded breaks.WEE1 upregulation is seen in a variety of human cancers and is inversely associated with prognosis in some models. 11 12Two separate kinomic screens in HNSCC identified WEE1 expression as a particularly strong determinant of cell survival, 13 14 indicating that HNSCC may be a fruitful setting in which to investigate the clinical effects of WEE1 inhibition.
Adavosertib (AZD1775) is a potent, selective small molecule inhibitor of WEE1.It has been shown to potentiate the activity of various chemotherapeutic agents in vitro and in vivo.Some studies suggest the sensitising effect is only seen in p53-deficient tumours [14][15][16] although not exclusively. 10It has also been shown to enhance IR-induced cell death in TP53-mutant cell lines.
Coexposure of AZD1775 and cisplatin were found to reduce clonogenic survival, 17 demonstrating this combination therapy has the ability to overcome cisplatin resistance in HNSCC.Similar effects of this compound on radiation-induced cell death have been seen in models of typically radio-resistant cancer, such as pontine glioma, 18 glioblastoma 19 and pancreatic adenocarcinoma. 20Importantly, one study has shown that WEE1 inhibition by Open access AZD1775 sensitises acute myelogenous leukaemia and lung cancer cell lines to cytarabine chemotherapy independently of p53 status, 10 suggesting that p53 mutation as a predictive biomarker for response to WEE1 inhibition may be cancer and/or chemotherapy specific.WEE1 has also been implicated in maintaining genomic stability through stabilisation of replication forks-downregulation reduces replication fork speed during S-phase, generating potentially lethal dsDNA breaks. 21By impacting both cell cycle progression and DNA damage repair, WEE1 inhibition may potentiate cell death in response to chemotherapy and IR.This suggests that there may be an additive effect on clinical outcome in combination with POCRT, as well as potential synergy.
AZD1775 is being tested in many clinical settings including in combination with docetaxel and cisplatin in HNSCC (NCT02508246), 22 with radiotherapy in childhood pontine glioma (NCT01922076), with temozolomide and radiotherapy in glioblastoma (NCT01849146), and with cisplatin and radiotherapy in cervical cancer (NCT01958658).
In summary, the available mechanistic data lend strong support to combining AZD1775 with cisplatin and with POCRT in the clinic.Given that the predictive effect of TP53 mutation on such combinations has yet to be clinically validated, this study does not prospectively stratify patients based on any such clinical or biological characteristic.Biomarker data generated, however, will be important in understanding the mechanism of action and informing future combination studies.

Incorporating AZd1775 into clinical management
A recent randomised phase III study of neoadjuvant chemotherapy in a Chinese population with locally advanced resectable oral cavity cancer showed that docetaxel, cisplatin and 5-fluorouracil (TPF) treatment generated a response rate of 80%, with an exploratory analysis showing that clinical responders had improved overall survival and local control. 23There was no improvement in overall survival for the study group as a whole compared with the control group without upfront TPF chemotherapy, possibly because patients who would ordinarily have received POCRT received PORT alone in this study.In an earlier study, TP53 mutation conferred increased resistance to preoperative cisplatinbased chemotherapy in oral cancer. 24While neoadjuvant chemotherapy has yet to find a definitive role prior to surgery, these findings raise the possibility that chemosensitisation may further improve outcomes.
The preoperative group A sets out to test this hypothesis by combining 1 week of AZD1775 alone, with a second week of cisplatin (40 mg/m 2 ) with AZD1775.This short course of treatment will minimise any delay between diagnosis and definitive surgery and prevent unacceptable delay for the non-responders from getting standard treatment.If this is a promising combination, the long-term aim is to test AZD1775 in the neoadjuvant setting to downstage the disease through treatment with cisplatin and AZD1775 in the preoperative setting, which may decrease the extent of surgery, and may also reduce the need for PORT/POCRT by reducing the involved margin rate.Many patients with tumours that have highrisk histopathological features develop relapsed disease within the anatomical area treated to a high radiation dose.Chemosensitisation and radiosensitisation via inhibition of the G 2 /M checkpoint by AZD1775 has the potential to improve outcomes with acceptable toxicity.This possibility is the basis of the treatment delivered in the group B cohort.

Wisteria trial objectives
Primary objectives are to determine the recommended dose and safety profile of AZD1775 with cisplatin in the group A preoperative (window-of-opportunity) setting, and with cisplatin/radiotherapy in the group B postoperative setting.Other objectives: obtain information regarding group A and B disease-free survivals; evaluate AZD1775 pharmacodynamic (PD) effects when administered with cisplatin (group A); identify and correlate potential predictive biomarkers with PD markers of DNA damage (group A); determine the rate of R0, R1 or R2 resection status and surgical complication (group A); and group B quality of life (QoL).

Trial design
Wisteria is an open-label, dose-finding, multicentre phase I trial to determine the highest safest dose of AZD1775 in combination with a single dose of cisplatin in the preoperative setting as a window of opportunity trial (group A); and to determine the maximum tolerated dose (MTD) of AZD1775 in combination with cisplatin/radiotherapy in the postoperative setting (group B): both groups recruit and run in parallel (figure 1).The dose-escalation design is based on the modified time-to-event continual reassessment method (TITE-CRM), [25][26][27][28] requiring a maximum of 21 patients per group, and encompassing up to four dose levels of AZD1775.
The MTDs of AZD1775 for both groups are expected to differ given the additional toxicities of radiotherapy in group B. The dose-limiting toxicity (DLT) monitoring periods are defined in table 1. Conservative target DLT rates are selected to minimise the likelihood of compromising individual patient's radical surgery and/or PORT.To maximise recruitment, reduce trial suspension time between cohorts, while balancing safety and optimal patient allocation, screening cohorts of up to five patients will be allowed if the dose has previously been tested.Recruited patients will be allocated to the current recommended dose up to a maximum of five.Patients who become unevaluable may be replaced.
The model will be updated after every two to three evaluable patients, with any subsequent eligible patients (not already receiving treatment) allocated to the latest recommended dose cohort.Subsequent cohorts will be Open access Figure 1 Wisteria trial schema.Group A patients were required to have received a minimum treatment of chemotherapy (cisplatin) and AZD1175 in combination to be DLT evaluable.Group B patients were required to have received a minimum of 2 weeks of treatment (half the total scheduled AZD1775 dose) to be DLT evaluable.DLT, dose-limiting toxicity; PD, pharmacodynamic; TITE-CRM, time-to-event continual reassessment method.assigned a dose level using all the data observed until either-the MTD is determined, the maximum sample size is reached, or the trial is stopped early due to unacceptable DLT levels at the lowest dose.

Patient population, screening and consent
Group A patients are those with biopsy-proven squamous cell carcinoma of the oral cavity, larynx or hypopharynx (having accessible tumours for rebiopsy under local anaesthetic or via ultrasound-guided biopsy), and due to undergo surgery.Patients in group B consist of those already diagnosed with oral cavity, larynx or hypopharynx squamous cell carcinoma, who have undergone surgery and are considered at risk of relapse after surgery (ie, with positive margins and/or nodal extracapsular spread).Patients with cancer of the oropharynx will not be included.Since Human papillomavirus (HPV) is only routinely tested in oropharyngeal squamous cell carcinoma or unknown primary of squamous cell carcinoma in the National Health Service (NHS), HPV status was not determined or requested for the tumours as part of the study.Patients who meet the criteria will be supplied with the patient information sheet.If informed consent (online supplementary appendices 1 and 2) is given, the patient undergoes a full screening evaluation to ensure they satisfy the eligibility criteria prior to registration (boxes 1 and 2).

dose selection
Based on AstraZeneca studies, 29 30 the Wisteria trial uses 3-day dosing, giving AZD1775 weekly to coincide with cisplatin administration thereby potentiating the DNA damage effects.The maximum total dose of AZD1775 1800 mg in combination with 1 dose of 40 mg/m 2 cisplatin for group A is substantially lower than AZD1775 monotherapy dose 2250 mg every 21 days (225 mg two times per day orally over 2.5 days per week for 2 weeks per 21 day cycle) used in Do et al; and although in combination with cisplatin, the highest dose level (dose level 2) is likely to be safe.
2][33] Trial doses and schedule are set out in table 1. AZD1775 is administered orally.The feasibility of the oral administration will be evaluated in postoperative patients.For both groups A and B, cisplatin dose banding is not permitted.BSA to be calculated according to local practice; if weight changes>10% then cisplatin dose should be recalculated.

Trial treatments
AZD1775 (morning dose) should be administered prior to cisplatin, and group B to receive cisplatin 1 hour before radiotherapy; standard supportive treatment, including premedication with antiemetics (excluding aprepitant and fosaprepitant), is allowed according to standard practice guidelines.

Trial outcome measures Primary outcome measures
The recommended doses of AZD1775 (determined by the modified TITE-CRM) for group A and B. ► Group A: the highest safe dose in combination with cisplatin with a predefined target DLT probability of 25% for up to 42 days from start of treatment.► Group B: the MTD in combination with cisplatin/ radiotherapy with a target DLT of 30% for up to 12 weeks from the start of treatment.► Safety profile of AZD1775 in both group A and B.

Secondary outcome measure
Disease-free survival time for both groups and recorded from commencement of treatment to the date of disease recurrence, patient death or end of trial follow-up period.-Haemoglobin ≥100 g/L.
-Aspartate transaminase or alanine aminotransferase ≤2.5 upper limit of normal (ULN).-Total bilirubin ≤1.5 ULN (except for patients with known Gilbert's syndrome).► Male and female participants must agree to take appropriate measures to prevent pregnancy.Contraceptive measures should be used for 2 weeks prior to trial entry, during the trial and for at least 6 months after last receiving treatment.Acceptable methods of contraception include total abstinence (if this is the patient's usual and preferred lifestyle choice), tubal ligation, combined oral, transdermal or intravaginal hormonal contraceptives, medroxyprogesterone injections (eg, Depo-Provera), copper-banded intrauterine devices; hormone impregnated intrauterine systems and vasectomised partners.All methods of contraception (with the exception of total abstinence) should be used in combination with the use of a condom by their male sexual partner for intercourse.In addition to general criteria ► Group A-Accessible tumours for rebiopsy to be taken under local anaesthetic or via ultrasound-guided biopsy.

Radiotherapy
External beam radiation therapy (30 fractions over 6 weeks) to start within 3 months of surgery.Dose levels are 65 Gy for positive margin, 60 Gy to lymph node levels that have been dissected and 54 Gy to elective lymph node areas.A dose of 65 Gy may also be applied to areas of gross extracapsular spread at the discretion of the treating clinician.

AZD1775
AZD1775 PO BID for 3 days on days 2-4 of weeks 1, 2, 4 and 5 (no treatment with AZD1775 during weeks 3 and 6).Dose recommendation will be according to modified TITE-CRM model.
DLT reporting period The minimal reporting period is 56 days (8 weeks) from start of radiotherapy, but patients will be monitored for DLTs up to 84 days (12 weeks), that is, up to 6 weeks from the end of POCRT PK samples Pharmacokinetic samples will be collected pre and post-the fifth dose of AZD1775 on week 1-day 4 (2 samples per patient) DLT, dose-limiting toxicity; PD, pharmacodynamic; PK, pharmacokinetic; POCRT, postoperative chemoradiation; TITE-CRM, time-to-event continual reassessment method.
patient requiring a toxicity-related dose delay of more than 21 days must be discontinued from the study unless there is approval from the chief investigator for the patient to continue.For group A, if any treatment-related grade≥3 toxicity develops in week 1, cisplatin and AZD1775 are discontinued in week 2. For group B, if calculated GFR falls below 60 mL/min, immediately before any cycle of concomitant chemotherapy, weekly cisplatin should be discontinued and consideration given to substitute with weekly carboplatin (Area Under the Curve (AUC)=1.5),according to routine local practice.
In both groups, a full blood count is obtained at the beginning of each week.If haematological toxicity occurs, treatment should be modified as per online supplementary appendix tables 10 and 11.For non-haematological toxicities, treatment should also be modified as per online supplementary appendix tables 12-16.
The selection of PD markers to test are different for group A and group B as group A enables examination of samples pretreatment and post-treatment with AZD1775.

Concomitant medications
All concomitant medications received within 14 days before the first dose of study medication and for 12 weeks after the last dose of study medication should be recorded.Medications may be administered for maintenance of existing conditions prior to study enrolment or for a new condition that develops while on study.The treatments and medications listed in online supplementary appendix 17 are prohibited or to be used with caution while in this study.No other investigational therapy or anticancer agents, other than the study medications, should be given to patients.If such agents are required for a patient, then the patient must first be withdrawn from the study.Live vaccines are not permitted.

discontinuation of investigational medicinal product
Patients should discontinue trial treatment in the following circumstances: intolerable toxicity; confirmed disease recurrence; pregnancy; severe non-compliance to protocol; development of any study-specific criteria for Open access box 3 dose-limiting toxicities (dLT) Patients experiencing any of the following adverse reactions (ARs) during the reporting period will be considered to have experienced a DLT: ► Grade 3 or 4 neutropenia lasting for >7 days despite adequate granulocyte-colony stimulating factor (G-CSF) support.► Grade 4 febrile neutropenia, which includes grade 3 febrile neutropenia accompanied by systemic inflammatory response syndrome/ sepsis or other life-threatening consequences.► A third occurrence of grade 2 or worse neutropenia despite G-CSF support.► Grade 3 or 4 thrombocytopenia lasting ≥7 days.► Any occurrence of grade 4 anaemia.► Any occurrence of grade 4 mucositis.► Any occurrence of grade 3 nausea despite preventative and supportive care according to local practice.Tube feeding is not considered a DLT.► Any occurrence of grade 4 vomiting despite preventative and supportive care according to local practice.Tube feeding is not considered a DLT.► Any occurrence of grade 4 diarrhoea despite preventative and supportive care according to local practice.► Any AR which results in an omission of chemotherapy administration and/or study treatment for >14 days.► A start-of-treatment-to-surgery time of >42 days in group A as a result of a treatment-related toxicity.► Grade 3 mucositis lasting >42 days after the end of treatment in group B. ► An overall treatment time of >49 days from the first day of radiotherapy in group B as a result of a treatment-related toxicity.► Any grade 3 or 4 non-haematological AR (except for fatigue, nausea, vomiting and diarrhoea) for which medical intervention that lasts >7 days is required and the investigators deem that this AR is more severe or prolonged than what would be expected of standard treatment.Tube feeding is not considered a DLT.► Any clinically significant occurrence which investigators within the trial management group agree would place the patient at undue safety risk If the inability to swallow the AZD1775 capsule develops during the course of trial treatment, this event should be noted in the relevant section on the Suspected DLT form and reported immediately.However, this is not considered a DLT in itself.
box 4 Adverse event and serious adverse event definitions

Adverse event
Any untoward medical occurrence in a patient or clinical trial subject administered a medicinal product and which does not necessarily have a causal relationship with this treatment.

Serious adverse event
Any untoward medical occurrence or effect that at any dose may: ► Result in death.► Is life threatening.► Requires hospitalisation or prolongation of existing in-patient hospitalisation.► Results in persistent or significant disability or incapacity.► Is a congenital anomaly/birth defect.► Is considered medically significant by the clinical investigator.
discontinuation and investigator decision, for example, if the patient requires a prohibited concomitant medication.
Safety monitoring: adverse events and serious adverse events All medical occurrences which meet the definition of an adverse event (AE) or serious AE (SAE) (as defined in box 4) should be reported.This includes abnormal laboratory findings of grade 3 and above only.AEs will be monitored for and reported from date of informed consent until the 12-week follow-up visit for group A and until the 12-month follow-up visit for group B. Investigators should report SAEs within 24 hours of first knowledge of the event, until 12-week follow-up visit.After this time, expedited reporting is no longer required and should be reported as an AE.Any poststudy serious unexpected serious adverse reactions (SUSARs) should be reported within 24 hours.An independent Safety Committee will review all SAEs.Fatal or life-threatening SUSARs will be reported to the Medicines and Healthcare products Regulatory Agency (MHRA) and Research Ethics Committee (REC) within 7 days.Detailed follow-up information will be provided within an additional 8 days.All other events categorised as SUSARs will be reported within 15 days.The MHRA and REC will be notified immediately if a significant safety issue is identified during the course of the trial.All SAEs relating to AZD1775 will be reported to AstraZeneca within 24 hours of notification.data collection and monitoring Data will be collected via a set of forms capturing details of eligibility, baseline characteristics, treatment and outcome details.This trial will use an electronic remote data capture system.SAE reporting and notification of pregnancy will be paper based.All missing and ambiguous data will be queried.In all cases, it remains the responsibility of the investigator to ensure that data are accurate.Details regarding data collection for group A and B are given in online supplementary appendix tables 7 and 8.The Investigator will permit trial-related monitoring, audits, ethical review and regulatory inspection(s) at their site, providing direct access to source data/ documents.

Statistical methodology
The AZD1775 dose is allocated to patients according to a modified Bayesian TITE-CRM, using an empiric dosetoxicity model, F(x,β) given as: where model parameter β is assumed random and follows a normal distribution and will be estimated by its posterior mean. 27At the point of model update, patients who have started treatment but have not experienced DLT (and have completed their full DLT assessment) will be included in the probability calculation with a weight equal to the proportion of the full DLT assessment period they have completed.Patients who experience

Open access
DLT within the assessment period or complete the full assessment period are assigned full weight.The model also enables inclusion of partial patient information that cannot be formally evaluated for DLTs due to withdrawal, treatment discontinuation or death, which are unrelated to treatment, within the DLT assessment period.
Table 1 displays the four dose levels of AZD1775 for group A and group B. For both groups, dosing starts at level 0 and allows for possible escalation to two higher levels, or deescalation to a lower dose, as recommended by the TITE-CRM, without skipping untried doses in escalation.The TITE-CRM model aims to recommend the next dose with estimated DLT probability closest to the target DLT level, taking into account the practical considerations as detailed below:

Analysis of outcome measures
The primary outcome measures are the recommended doses of AZD1775 in group A and group B, and the safety profile of the combination therapy.To be DLT evaluable, group A patients must have received AZD1775 and cisplatin doses scheduled up to and including day 8; group B patients must have received at least the first 2 weeks of treatment.The recommended doses of AZD1775 will be those that have an estimated DLT rate closest to the target DLT rate: 25% for group A; 30% for group B. DLT rates and corresponding 90% probability intervals will be reported.Safety will be monitored in patients treated with at least one dose of the trial treatment until end of follow-up.All adverse reactions, AEs, SAEs, suspected unexpected SUSARs, deaths, deviations and withdrawals experienced by trial patients, throughout the trial's duration, will be reported separately for group A and group B. Disease-free survival in group A and group B patients will be reported separately.Patients will be followed up for 12 months after which, time-to-event outcomes will be assessed using the method of Kaplan and Meier.Median disease-free survival with corresponding 95% CIs will be reported where appropriate.Further analysis will be performed separately on subsets of group A and B patients who have completed at least 90% of AZD1775 trial scheduled treatments.
Tertiary outcome measures will be explored and reported descriptively using basic descriptive statistics where appropriate: ► PD effects of AZD1775 and correlation with TP53 mutation status: Expression levels of parameterschanges over time and correlation with TP53 mutation status will be reported.► PK effects of AZD1775: Blood samples for PK analyses are collected (as scheduled in online supplementary appendix tables 7 and 8).Circulatory levels of AZD1775 and changes over time will be reported.► Complete pathological response rate for group A: Pathology data will be reported along with the results of the formalin-fixed paraffin-embedded (FFPE) tissue samples.Response rate will be reported as the number of evaluable group A patients showing a response (numerator) divided by the total number of group A evaluable patients (denominator).► Positive resection margin status in group A: This will be reported as the number and percentage of group A patients showing positive resection margins.► Surgical complications in group A: This will be reported by category of surgical complication with severity and given as numbers and percentages.► QoL in group B: QoL data are gathered over time and will be reported as changes over time and plotted.
In addition to the descriptive analyses, the repeated measures over time data may be modelled, if appropriate, with a linear mixed effects model (taking subject correlation into account) using linear or more flexible models.Interim analyses will be performed once each cohort of patients has been recruited and assessed within the defined assessment timeframe of DLTs.In each meeting, the safety committee will be presented with recruitment and safety data, together with a statistical report for next dose recommendation.Additional meetings might be convened if late-onset DLTs were observed.The safety committee will decide whether to progress to the recommended dose as indicated by the modified TITE-CRM model.The final analyses will be conducted in two parts: (1) analyses of the primary outcome measures of MTDs for groups A and B, to be carried out approximately 3 months after the safety review of the last cohort in each group and (2) longer-term outcomes-including secondary and tertiary outcome measures and updated safety data, to be carried out 1 year after the end of trial.

Table 1
Details of the Wisteria TITE-CRM trial design for groups A and B DLT, dose-limiting toxicity; MDT, multidisciplinary team; TITE-CRM, time-to-event continual reassessment method.
Histologically confirmed diagnosis of oral, laryngeal or hypopharyngeal squamous cell carcinoma.► Multidisciplinary team (MDT) recommendation for surgical resection with curative intent.► Eastern Cooperative Oncology Group performance status 0/131.
► ► Normal liver and bone marrow function:

Table 2
Any previous treatment for the same cancer, or previous head and neck malignancy, apart from laser excision of carcinoma in situ, with minimal residual functional deficit or registration and treatment in group A prior to surgery.► Patients with cancer of the oropharynx or non-primary cancer will not be included.► Any metastatic disease from any primary site.Clinical judgement by the investigator that the patient should not participate in the study.► Known hypersensitivity to the study drugs or active substances or excipients of the preparations.Predefined DLTs have been specified by the Clinical Investigators (box 3), and recorded toxicity profiles for the Wisteria treatments are given in online supplementary appendix table 9. Toxicity is assessed using the National Cancer Institute Common Terminology Criteria for Adverse Events V.4.0 37 unless otherwise specified.Any Trial summaries of group A and group B ► Group B-High-risk histopathological features after surgical resection, that is, nodal extra-capsular spread and/or tissue resection margin <1 mm as agreed at MDT. ► Group B-Patients who have previously registered to group A can be considered for inclusion in group B.box 2 Exclusion criteria-all patients ► ► Unstable ischaemic heart disease (myocardial Infarction within 6 months prior to trial entry or angina requiring the use of nitrates greater than once weekly).► Patients who have a history of torsades de pointes (unless all risk factors that contributed to torsades have been corrected).► Active gastrointestinal disease that might limit absorption of study drug, for example, coeliac disease, Crohn's disease, ulcerative colitis, pancreatic insufficiency.► ► Patients with any contraindications to cisplatin use.► ► Pregnant or breastfeeding patients.► Stop early due to safety concerns if there is sufficient evidence that the posterior probability of DLT at the lowest dose is greater than the target DLT rate, implying that the lowest dose is too toxic.► Stop early if sufficient patients have been allocated to the current MTD (and remains the recommended dose level for the next cohort if the trial continues) before the full recruitment of 21 patients.► The minimum DLT period is set at 30 days for group A and 8 weeks for group B. This means that partial infor- ►