The following key changes were incorporated into protocol amendment 1: • Updated key contacts • Changed phase 3 portion of study to be double-blind and added placebo to the pembrolizumab arm. • Reordered sections to separate phase 1b and phase 3 in the synopsis, and general study procedures sections in order to more clearly delineate which sections are specific to each phase of the study. Additionally, to reduce the number of footnotes, reorganized Section 7 to provide more description to study procedures. • Updated background information for talimogene laherparepvec and pembrolizumab to reflect recent publications or presentations and approvals. • Updated biopsy requirements in inclusion criteria • Removed biodistribution and shedding at select sites in phase 3 as we will have sufficient data from phase 1 • Removed PD-L1 status as a stratification factor • Removed registry study option • Updated contraception language for exclusion criterion and also sections related to pembrolizumab. This is to align with other pembrolizumab protocols. • Added exclusion criteria for active tuberculosis in order to align with other pembrolizumab protocols. • Updated text related to pembrolizumab, rescue medications, dose adjustment, overdose, and supportive care guidelines to align with other pembrolizumab protocols. • Updated pregnancy and lactation reporting language to align with other talimogene laherparepvec and pembrolizumab protocols. • Updated language to clarify the modified response criteria in the appendices (eg, how to evaluate separated lesions, criteria for confirmation of PD). • Added optional photography substudy Additional errors were identified and rectified in the superseding amendment, and administrative errors were corrected.

This protocol was amended to: • Update eligibility criteria triggered by recent safety signal – carotid blowout syndrome, and efficacy signal – lack of benefit in primary refractory patients with progressive disease within 3 months of curative intent multimodality therapy. • Remove progression-free survival as a primary endpoint for phase 3 and make it a secondary endpoint, along with Complete Response Rate (CRR). o The decision to forego of PFS as a dual primary endpoint hinged upon a few factors which included: the recent outcome of CHECKMATE 141 which led to approval of nivolumab based upon OS and also the final approval or pembrolizumab which is dependent upon OS from KEYNOTE 040. Considering the fact that OS is a superior outcome measure of a treatment over PFS especially in a poor prognosis disease and that PFS is not always a surrogate for OS, combined with the precedent set by nivolumab approval on OS, we felt that PFS did not have a truly meaningful role in the assessment of the efficacy of TVEC+pembrolizumab in second line head and neck cancer as well as for regulatory purposes. • Add the use of irRECIST investigator assessment for response assessments and remove RECIST 1.1 central review. • Update QOL/PRO wording and elevation of QLQ C30-3L to secondary endpoint from exploratory. • Update statistical methods to justify the endpoint changes and also to introduce OS IA and futility analyses. • To add more detail around go no go decision from Phase 1b to 3. • Add additional pembrolizumab background information. • Add additional talimogene laherparepvec background information. • Update IP discontinuation/withholding rules. • Update radiographic tumor assessments (sites of disease, spiral CT). • Add additional information for archival tumor tissue. • Add additional information for HPV testing. • Update safety reporting information. • Administrative changes and editorial changes for clarification.

This protocol was amended to: • Add the investigator-assessed RECIST v1.1 secondary tumor response endpoints of objective response rate (ORR) and progress-free survival (PFS) as secondary endpoints and remove CRR per irRECIST as a secondary endpoint. • Replace EQ-5D-3L with EQ-5D-5L to utilize the most recent PRO version. • Revised re-irradiation exclusion criteria. • Replaced modified RECIST v1.1 of 10 maximum lesions with 5 per organ with standard RECIST v1.1 for screening (5 maximum lesions with 2 per organ) to align with the use of standard RECIST v1.1 introduced for key secondary endpoints. • Added RECIST v1.1 assessment in addition to irRECIST assessment for response assessment. • Revised the set of secondary hypotheses to be tested with the Maurer-Bretz procedure to potentially generate more robust efficacy conclusions. Testing of CRR was replaced with ORR. Testing was added for RECIST v1.1 ORR and PFS. The total number of potential hypotheses tested increased from 3 to 5. • The number of events at the OS interim analysis was increased from approximately 255 to 280 to preserve 70% power in the event of a potential treatment lag effect. • Sample size considerations were revised due to the OS interim analysis change and to discuss the power for the revised secondary hypothesis tests. • The OS futility criterion at the interim analysis was changed from a conditional power <10% given a true HR of 0.70 to an observed HR >0.92 considering a potential treatment lag effect. • An audit-based Blind Independent Central Review (BICR) was added to assess the consistency of investigator- and BICR-assessed treatment effects for RECIST v1.1 ORR and PFS. • The definition of the phase 3 primary efficacy and safety analysis sets, primary completion, and end of trial were revised to maintain the study’s statistical considerations. • Updated language on disease related events and reporting procedures per internal Amgen recommendations and to align across program.

This protocol was amended to: • Add language to clarify long term follow-up for subjects in phase 1b due to decision to not proceed to the phase 3 part of the study. • Remove PK and ADA samples from phase 1b portion of protocol per discussion with FDA, EMA, and PDMA (Merck request). • Update End of Study language to align with most recent Amgen template text and to include the definition of Primary Completion and End of Trial for subjects who completed phase 1b. • Include a follow-up analysis for phase 1b. • Add text providing guidance about latex allergies to exclusion criteria no. 223. • Update serious adverse event reporting procedures to align with current safety language. • Update pembrolizumab safety language regarding pregnancy reporting and breastfeeding. • Update Key Sponsor Contacts. • Make administrative and editorial changes.