There are numerous studies and I have highlighted the most pertinent aspects of the molecular landscape to bring you to speed.

In one word-potential. Molecular amalgamation of the studies in clinic is time consuming, expensive and not standardised.

Most important aspect of this paper? 10 gene signature which predicts radiosenstivity. Although, this has been picked up from various tumour cells lines which brings us to its own set of uncertainties.

The excerpts are below:

Page 1

Heterogeneous tumour biology within a particular disease population, it is reasonable to hypothesize that there are sub- populations of patients for whom the therapeutic index of radiotherapy is not sufficiently optimize

Unique aspects of the tumor habitat, radiosensitivity, and microenvironment may allow an even more sophisticated method of adaptive treatment planning

However, radiation oncologists have traditionally delivered homogenous doses for any given disease and stage despite underlying biology following a Gaussian distribution.

A recent area of focus in the field of personalized oncology is the use of clinical-genomic biomarkers

Page 2

RTOG 1016 evaluated this question by comparing radiotherapy and cetuximab to radiotherapy and cisplatin for HPV- associated oropharyngeal disease. This study revealed an inferior outcome with cetuximab vs cisplatin

Currently there are minimal data on the role of EGFR status in respect to personalized dose or volumes for primary or nodal disease.

Regional lymph nodes have historically been treated to 70 Gy for gross disease, 63 Gy for high-risk regions concerning for microscopic disease, and 56 Gy for low-risk regions— regardless of primary histology.

As viral DNA incorporates into the host genome, the overexpression of viral-associated oncoproteins E6 and E7 promotes the proteolysis of the tumor suppressor genes p53 and Rb, respectively. With impaired DNA repair function, HPV- associated cancers have a reduced ability to withstand and overcome the double-stranded DNA damage from radiotherapy.

HPV status has been incorporated into the recent implementation of the American Joint Commission of Cancer staging system (eighth edition) for oropharyngeal cancers, but its final impact on the clinical decision-making process remains a topic of active investigation

Additionally, presence of HPV16DNA has predicted improved rates of LRC following postoperative chemoradiation

EGFR is found to be upregulated in a significant portion of HNSCC,14 and overexpression is concerning for more aggressive disease due to its association with decreased loco-regional control

Importantly, de- escalation with elective dosing as low as 40 Gy.34,35 and 36 Gy32 was accomplished without compromising disease control.

In addition to radiation de-escalation, the NRG trial HN-002 is evaluating the possible exclusion of concurrent chemotherapy.

Page 3

Akervall et al proposed a panel of biomarkers to predict radioresistance of HNSCC. Originally identified in 38 patients and subsequently validated in another 86 patients, several markers were correlated with decreased relapse free survival (YAP1, BCL2) and CSS (YAP-1, VEGF, CLAUDIN-4) as well as synergistic radioresistance (YAP-1 and c-MET)

An interim report of the first 279 patients identified 6 frequently mutated genes that have been determined to key signal pathways for HNSCC carcinogenesis: TP53, NOTCH1, CDKN2A, PIK3CA, HRAS, and PTEN.

HPV-positive tumors were associated with upregulation of PIK3CA, loss of TRAF3, and amplification of E2F1, while smoking associated HPV-negative disease was associated with loss of TP53 and CDKN2A

Perhaps the most specific evidence supporting the use of CTC for lymph node personalization comes from Hristozova et al, who found a significant correlation of N2b or greater nodal stage of HNSCC with CTCs frequency.

Page 4

Similar findings were shown in a study of patients treated with definitive radiotherapy with subsequent studies demonstrating a potential role for this biomarker in selecting candidates with very low-risk or low-risk disease for active surveillance following initial biopsy

Decipher 22-gene signature was established in an effort to better predict distant metastasis-free survival in patients with localized disease following definitive treatment

The targets involved in elective nodal irradiation for this disease site generally include the obturator, internal and external iliac lymph nodes, with doses typically prescribed to 45 Gy in 25 fractions

The predictive utility was further improved in combination with a validated model predicting outcomes in the postoperative setting.

The Postoperative Radiotherapy Outcome Score (POR- TOS) is a 24-gene signature score which is inversely related to the 10-year rate of distant metastases in the postprostatectomy setting (high PORTOS = lower incidence of metastases).

In an effort to decrease the number of patients undergoing biopsy, a diagnostic 4k score comprised of 4 kallikrein markers, digital rectal exam, and age was established to identify aggressive disease

The OncotypeDx Genomic Prostate Score is a 17-gene assay for patients with either very-low risk or low risk disease and a life expectancy greater than 10 years who have undergone biopsy.

Taken together, PORTOS and Decipher scores are valuable tools which can be used in combination with traditional risk stratifications for individualized selection to adjuvant rather than salvage radiotherapy.

While this is suggested as a biomarker for selecting patients suitable for active surveillance, the Genomic Prostate Score has been slow to assimilate into routine clinical practice.

Increase of the cell cycle progression score was independently associated with biochemical recurrence (hazard ratio [HR] 1.77, 95% confidence interval [CI] 1.40-2.22, P < 0.0001) after prostatectomy, as well as time to de

Page 5

Practice involving elective irradiation of levels I and II of the axilla varies but is generally considered following surgical staging when there is evidence of extracapsular extension, greater than 33% positivity of sampled lymph nodes, orno axillary lymphnode dissection in the presence of a positive sentinel lymph node biopsy

The Oncotype Dx Recurrence Score (RS) was shown to be an accurate predictor for the risk of distant recurrence and OS at 10 years independent of age and tumor size in patients with lymph node negative, ER  tumors treated with tamoxifen.

Radiation sensitivity signature (RSS) to identify patients who would benefit from adjuvant radiotherapy

AJCC eighth edition staging guidelines incorporate the Oncotype DX RS as staging modifiers

Subsequent meta-analysis (N= 5456) of breast and prostate cancer patients found a significant association between an ATM SNP and normal tissue toxicity

Page 6

Due to concerns for toxicity and marginal regional control benefit with ENI, image-guided selective nodal irradiation (SNI) of involved nodal fields was evaluated as an alternative

Based on their prior work, Vinogradskiy et al conducted a retrospective analysis of SNPs as biomarkers predicting for radiation pneumonitis.

Using their model, they found that an individualized mean lung dose based on SNPs data led to a prescription change >5 Gy in 59% of patients. The model also would have predicted a lower prescribed dose in nearly all (96%) patients who developed pneumonitis. Other promising biomarkers in NSCLC include single-gene mutations characteristic of more radioresistant tumors, such as KEAP1 and NRF2.

Increased PD-L1 expression as measured by immunohistochemistry has been consistently shown to enrich for selection of responders to anti- PD-1 therapy.Tumor mutational burden has been found to predict response to anti-PD1 therapy in NSCLC126

Efforts toward personalizing treatment for advanced NSCLC have manifested in predictive models for patients treated with chemoradiation,as well as targeted biologic agents toward EGFR mutation and ALK gene rearrangement.

Additional molecular targets including ROS1, BRAF, MEK, Trk1/2/3, and others have available targeted therapies which can be utilized for delivery of personalized, biologically targeted therapy

Page 8

Using a systems biology approach, this 10-gene molecular signature was trained to predict the inherent radiosensitivity of a tumor based on the cellular survival at 2 Gy (SF2) after radiotherapy in 48 human cancer cell lines

The 10 genes identified were selected from the top 500 genes predicting radiosensitivity in various primary tumors, including melanoma, breast, prostate, renal, colon, NSCLC, pancreas, glioblastoma, and leukemia

RSI has been further validated in patient cohorts in rectal, esophageal, prostate, glioblastoma, pancreas, lung, head and neck, and breast cancer primaries.

Page 9

GARD was predictive of distant metastasis-free survival in patients receiving adjuvant therapy for breast cancer. In light of the current high rates of local control in breast cancer, it is reasonable to hypothesize from this data that there is a substantial population of patients who may be candidates for dose personalization.

Combining RSI with other genomic, clinical, and image-based prognostic markers offers important prospects for optimizing treatment through biologic guidance of both treatment fields and dose fractionation

Since RSI is based on SF2, it was hypothesized that RSI could be integrated into the biologically effective dose to develop a clinically actionable metric based on the intrinsic genomic radiosensitivity/radioresistance of tumors. Use of the genomic adjusted radiation dose (GARD) represents an application through which radiotherapy can be employed as a component of precision medicine.

RSI analyses for 1362 primary colon and 704 metastatic sites of disease were performed, revealing a higher incidence of radioresistant samples in metastatic (64%) compared to primary tumors (54%, P =0.01). They were able to categorize disease sites in descending order of radioresistance as follows: ovary (0.48), abdomen (0.47), liver (0.43), brain (0.42), lung (0.32), and lymph nodes (0.31), P < 0.0001, where an RSI score of <0.375 was considered radiosensitive

It is noteworthy that among the sites of colorectal metastasis, lymph nodes were shown to have the lowest RSI values (low RSI, more radiosensitive). This evidence allows us to consider differential lymph node radiotherapy doses to personalize treatments

Page 10

GARD score distribution was widely variable in disease sites where regularly prescribed subclinical (perioperative) doses are approximately 45 Gy