WO2024235997A1

WO2024235997A1 - Non-invasive disease detection and monitoring

Info

Publication number: WO2024235997A1
Application number: PCT/EP2024/063285
Authority: WO
Inventors: Tim AITMAN; Martyna ADAMOWICZ; Christelle Robert; Duncan Mclaren
Original assignee: LOTHIAN HEALTH BOARD; University of Edinburgh
Current assignee: LOTHIAN HEALTH BOARD; University of Edinburgh
Priority date: 2023-05-15
Filing date: 2024-05-14
Publication date: 2024-11-21
Anticipated expiration: 2025-11-15
Also published as: GB202307233D0

Abstract

The present invention provides a method of monitoring radiotherapy response in a subject having prostate cancer, the method comprising: a) providing a cell-free DNA (cfDNA)-containing sample obtained from the subject; b) performing DNA sequencing on the cfDNA-containing sample in order to generate methylation sequencing reads; c) assessing methylation levels within the cfDNA at each of a plurality of pre-determined differentially methylated regions (DMRs) to generate a sample methylation profile, wherein each pre-determined DMR is associated with a tissue of origin, thereby detecting the presence of one or more tissues of origin; and d) determining the radiotherapy response of the subject based on the sample methylation profile Also provided are methods of treating a subject with prostate cancer using adaptive radiotherapy.

Description

NON-INVASIVE DISEASE DETECTION AND MONITORING

Field of the Invention

The present invention relates to methods of analysing cell-free DNA (cfDNA)-containing samples from subjects having prostate cancer or a prostate tumour. Methods for monitoring radiotherapy response in a subject having prostate cancer, determining tumour response to radiotherapy and determining nontumour tissue toxicity to radiotherapy are provided.

Background

Around 50% of cancer patients require radiotherapy at some point in their treatment. The probability of treatment success is dependent upon the dose delivered and the relative radio-sensitivity of the tumour. The maximum dose delivered is calculated to ensure that no more than 10% of patients suffer lifechanging late radiotherapy toxicity with an additional 20% reporting toxicity that still has some impact on quality of life. The current "one dose fits all" approach to radiotherapy treatment contrasts with the ethos of precision oncology in other branches of cancer medicine, which increasingly harness the power of genetic and genomic analyses to stratify patient care, monitor tumour response and detect disease progression.

Liquid biopsy - the analysis of cell-free DNA (cfDNA) and other cellular components released from dying or damaged cells into the circulation or other bodily fluids - has utility because the cellular components carry the characteristics of the tissues from which they are derived. Circulating tumour-derived DNA (ctDNA) retains the genetic and epigenetic hallmarks of the originating tumour, including somatic mutation and DNA methylation. Attractive because it is minimally invasive and repeatable, liquid biopsy is gaining traction in early detection and diagnosis, patient stratification, detection of minimal residual disease and prediction of recurrence after primary treatment (1-4).

Analysis of cfDNA has been little applied to management of radiotherapy, especially dosing or scheduling. However, the ability to quantify ctDNA and to define cfDNA tissue-of-origin in real time have the potential both to permit assessment of tumour response to radiation treatment during a course of radiotherapy treatment and, at the same time, to measure the extent of damage to healthy tissues surrounding the targeted tumour. These advances would allow truly personalised adjustment to radiotherapy dose per fraction or number of fractions, based on how the patient is responding to their treatment and would transform how radiotherapy is currently delivered.

Cellular DNA is released into the circulation primarily from dividing, damaged or dying cells undergoing apoptosis and necrosis. In radiotherapy, ionising radiation causes direct and indirect DNA damage predominantly through single and double strand breaks (5). If this damage cannot be corrected by cellular DNA repair mechanisms, cell death will occur by an interplay of apoptosis, necrosis, or mitotic catastrophe (6). In one of the first studies investigating the kinetics of ctDNA release during radiotherapy, Lo et al. quantified Epstein-Barr virus (EBV) ctDNA in patients undergoing radiotherapy for nasopharyngeal cancer. Having observed that 2 out of 10 patients had a transient rise in circulating EBV DNA one week after initiation of radiotherapy, a further 5 patients were recruited, in whom EBV DNA levels were measured daily during the first week. All 5 patients exhibited a transient rise in EBV DNA in the first week of radiotherapy, followed by a fall (7). More recently, in pre-clinical studies, Rostami et al. measured ctDNA release following chemical induction of apoptosis and exposure to radiation in tumour cell line and tumour xenograft mouse models. While there was little immediate ctDNA release after irradiation, a subset of irradiated cell lines demonstrated a rise in ctDNA after 72 - 96 hours. Supporting this timeline, three of five irradiated xenograft mouse models also showed a delayed increase in ctDNA release between 96 - 144 hours after irradiation. These findings led them to propose a model of cell death and DNA release whereby, in certain cell types, mitotic catastrophe predominates in the early response to irradiation exposure, with a peak in apoptosis and cfDNA release later at around 3-6 days following irradiation (6). Similarly, Muhanna et al., studying a model of buccal cancer in rabbits, found an initial rise in ctDNA in the 1-3 days following initiation of radiation therapy followed by a consistent fall (8).

These studies suggest a pattern of release of ctDNA from tumour cells, in which damaged cells undergo a combination of immediate and delayed cell death, with an initial rise of ctDNA release into the circulation following exposure to radiation, and detectable ctDNA concentrations that subsequently fall as tumour bulk decreases. ctDNA in cancer patients usually constitutes a small, frequently tiny proportion of the total circulating cfDNA. In early-stage cancer, ctDNA generally accounts for less than 1% of total cfDNA and only in latestage cancer and in a subset of high-secreting tumours, does the proportion of ctDNA rise above 1% (9- 11). Methods of ctDNA detection must therefore be sensitive and specific to be clinically useful.

Tumour-derived somatic mutation in cfDNA can be detected by targeted or genome-wide sequencing and by multiplex and allele-specific PCR. A range of methods have been developed to maximise sensitivity and specificity including barcoding, amplification, error correction and deep sequencing (10-12). Such methods are necessary because of the low total concentration of cfDNA in the circulation, the often low fraction of cfDNA that is tumour-derived and the molecular heterogeneity across cancer types and of individual tumours. These methods have already been successfully deployed across a range of clinical cancer studies for treatment stratification, assessment of minimal residual disease and detection of disease progression following primary treatment (1-4, 13, 14). In addition, because each tissue in the body, including tumour tissues, carries its own unique DNA methylation profile and these tissue-specific methylation profiles are retained after release of cellular DNA into the circulation, DNA methylation analysis is now recognised as a powerful tool for determining the tissue origin(s) of DNA samples including tumour-derived cfDNA (15-17).

In virus-associated cancers, detection of viral cfDNA by droplet digital PCR (ddPCR) or sequencing has been used to screen for virus-associated tumours in healthy populations and for characterisation of viral status at diagnosis. In longitudinally collected samples post-treatment, cfDNA viral titre has been successfully used as a prognostic marker and early marker of disease progression. Such approaches have shown particular value in EBV-associated nasopharyngeal cancer and human papillomavirus (HPV)-associated oropharyngeal cancers (18-21).

Advances in radiotherapy technology have improved the accuracy of radiation delivery to the tumour and clinical trials have refined dosing schedules with resultant improvements in outcomes, as reviewed elsewhere (5, 22, 23). However, toxicity remains common, in part because of the absence of biomarkers of toxicity, such that once dosing has started, it continues till the end of the schedule unless severe acute symptomatic toxicity occurs. Further, despite known variation in inter-individual and tumour radiosensitivity, manifested at the germline and gene expression level (24-26), genetic methodologies have to date impacted little on cure rates or toxicity (23). Recent advances in understanding the mechanisms and kinetics of ctDNA release from tumours and in the ability to define cfDNA tissue-of- origin suggest that liquid biopsy has the potential to change this (Figure 1).

In EBV-associated nasopharyngeal cancer, Lo et al. defined the kinetics of ctDNA release through measurement of EBV cfDNA (7), leading to the proposal of several groups that measurement of EBV viral load at the midpoint of radiotherapy treatment could be used as a basis for intensification or deintensification of treatment, respectively, in those with an adverse or favourable cfDNA response (18, 20). Similarly, Chera et al. demonstrated that a favourable HPV clearance profile during chemoradiotherapy, defined as >95% clearance of circulating HPV cfDNA from a high HPV baseline copy number, was predictive of disease control in HPV-associated oropharyngeal cancer and could be the basis of trials to investigate de-intensification based on HPV clearance rate (19). The pre-clinical xenograft mouse and rabbit models described above (6, 8), as well as a raft of mostly cross-sectional clinical studies now published, give optimism that the results in virus-associated cancers may also be applicable in non-virus associated cancers (27-30), with the caveat that of many prospective clinical trials in progress, very few have yet progressed to final report (3, 4, 31).

Of perhaps even greater novelty and applicability for radiotherapy would be the development of biomarkers of radiation-induced damage to tissues surrounding the targeted tumour. In the past five years, it has become clear that DNA methylation analysis has this capability. Extensive methylation atlases of healthy tissues and tumour types have been developed with the power to define the tissue composition of cfDNA samples and the ability to detect with high sensitivity the presence of DNA from a wide range of tissues and tumours, with demonstrated or potential application in the context of autoimmune disease, transplant rejection and early stage cancer (2, 15-17). Showing promise for detecting damage to tissues surrounding the targeted tumour, methylation analysis has recently been reported to detect liver-derived cfDNA in patients undergoing radiotherapy treatment for right-sided but not left-sided breast cancer (32).

These proof-of-concept studies offer the prospect of direct tests both for tumour cell death and off-target tissue damage during radiotherapy treatments (9, 33). Such tests could serve as biomarkers for adaptive radiotherapy regimes that guide treatment intensification or de-intensification based on the kinetics of clearance of circulating ctDNA and the presence or absence of cfDNA from healthy tissues surrounding the targeted tumour. However, the variable kinetics of cfDNA release in different cancer types and the currently incomplete knowledge of how different clinical radiotherapy fractionation protocols may impact on cfDNA release indicate that further study will be required to define the optimum timing of cfDNA assessment to impact effectively on radiotherapy outcomes (Figure 2).

While evidence for the value of ctDNA analysis following primary treatment with surgery, chemotherapy and radiotherapy is accumulating rapidly, the use of cfDNA analysis to guide radiotherapy dosing and scheduling during treatment is at an early stage. The existing proof-of-concept studies need confirmation in further clinical studies for example on the kinetics of cfDNA release in different tumour types and with different radiotherapy protocols. Technical advances are also required, such as development of protocols to measure DNA methylation directly in real-time, as may now be achievable by single molecule sequencing (34-36). The emerging evidence suggests that these liquid biopsy methods are sufficiently sensitive to provide clinically useful biomarkers of real-time tumour response and damage to healthy tissues in radiotherapy management. The confirmation of these data, being sought in several labs worldwide, would allow intra-treatment analysis of cfDNA for personalised adaptative radiotherapy scheduling, enabling more effective tumour cell death while minimising healthy tissue toxicity. Such an advance would be transformative for precision management of radiation treatments for cancer.

The present invention has been devised in light of the above considerations.

Summary of the Invention

Broadly, the present inventors have developed an improved method for assessing tumour response to radiotherapy and for assessing non-tumour tissue toxicity to radiotherapy. This is based on the realisation that liquid biopsy analysis of cell-free DNA (cfDNA) can be used to (i) monitor and measure early tumour response to radiotherapy; and (ii) measure damage to tissues in and surrounding the radiotherapy target area. The methods developed by the inventors are based on the concept that disease states such as cancer can be characterised by alterations of epigenetic markers in cfDNA, and additionally that epigenetic markers in cfDNA can be used to identify a tissue of origin. Furthermore, the inventors make use of single-molecule sequencing (such as sequencing performed using an Oxford Nanopore platform) in order to provide a rapid, real-time analysis of a subject’s response to radiotherapy and off-target toxicity. This provides the ability to identify in real-time how a subject is responding to radiotherapy, such as external beam radiotherapy, for example delivered using any currently available platform such as Linear Accelerator, Tomotherapy, MR-Linac or Proton Therapy.

Based on the release of cfDNA from the surrounding normal tissue and tumour cfDNA, and the analysis of methylation of differentially methylated regions (DMRs) within the cfDNA, a clinician will be able to identify which patients require their treatment to be adapted and importantly, how to do this in order to maximise the chances of cure while minimising the chances of life-changing treatment-related toxicity or toxicity that has some impact on quality of life. Importantly, as cfDNA release occurs before the onset of symptoms or side effects, the methods provided herein allow for the treatment schedule of the subject to be adjusted before the onset of any such symptoms or side effects.

Since the probability of being cured by radiotherapy is dependent upon the dose delivered and the relative radiosensitivity of the tumour versus the sensitivity of the surrounding normal tissue, a method than can measure both tumour response and damage to surrounding tissue provides a new paradigm in radiotherapy cancer management.

According to a first aspect, the present invention provides a method of monitoring radiotherapy response in a subject having prostate cancer, the method comprising: a) Providing a cell-free DNA (cfDNA)-containing sample obtained from the subject; b) Performing DNA sequencing on the cfDNA-containing sample in order to generate methylation sequencing reads; c) Assessing methylation levels within the cfDNA at each of a plurality of pre-determined differentially methylated regions (DMRs) to generate a sample methylation profile, wherein each pre-determined DMR is associated with a tissue of origin, thereby detecting the presence of one or more tissues of origin; and d) Determining the radiotherapy response of the subject based on the sample methylation profile.

In some embodiments, determining the radiotherapy response of the subject based on the sample methylation profile comprises comparing the sample methylation profile to a control methylation profile, wherein a significant difference in methylation level between the sample methylation profile and the control methylation profile is indicative of a response in the subject. In some embodiments, the control methylation profile is generated from a first cfDNA-containing sample obtained from the subject prior to receiving radiotherapy and the sample methylation profile is generated from a second cfDNA-containing sample obtained from the subject following at least one dose of radiotherapy.

In some embodiments, determining the radiotherapy response of the subject based on the sample methylation profile comprises determining a sample methylation profile without comparison to a control methylation profile. In such embodiments, the detection of methylation in the sample methylation profile can be used as an absolute measurement which may be indicative of radiotherapy response, for example, non-tumour toxicity to radiotherapy. This provides a rapid and easy to use test for determining radiotherapy response.

The DNA sequencing may comprise single molecule DNA sequencing, bisulphite DNA sequencing or enzymatic methyl-seq. In some embodiments, the DNA sequencing comprises or consists of single molecule DNA sequencing. In some embodiments the single-molecule DNA sequencing comprises nanopore sequencing. For example, the nanopore sequencing may be performed on an Oxford Nanopore platform. In other embodiments the single-molecule DNA sequencing is performed on a PacBio platform.

In some embodiments, the radiotherapy response comprises prostate tumour response to radiotherapy. In some embodiments the one or more tissues of origin comprises prostate tumour tissue. In some embodiments the one or more tissues of origin comprises adjacent non-tumour tissue. The one or more tissues of origin may comprise bladder tissue, small intestine tissue, colon tissue, rectum tissue and/or normal prostate tissue.

In some embodiments, the radiotherapy response comprises non-tumour toxicity to radiotherapy. In some embodiments, the radiotherapy response comprises prostate tumour response to radiotherapy and non- tumour toxicity to radiotherapy.

According to a further aspect, the present invention provides a method of treating a subject with prostate cancer using adaptive radiotherapy, the method comprising administering at least one dose of radiotherapy to the subject, further comprising the method of the first aspect, further comprising adjusting the dose of radiotherapy, altering the radiotherapy field and/or the frequency of radiotherapy.

In some embodiments, adjusting the dose of radiotherapy comprises increasing the dose of radiotherapy, decreasing the dose of radiotherapy, or discontinuing radiotherapy. Adjusting the frequency of radiotherapy may comprise increasing the length of time between doses or decreasing the length of time between doses.

The sample obtained from the subject may be a blood sample.

In some embodiments, the plurality of pre-determined DMRs is selected from the CpG sites of Table 1 or Table 2. The plurality of pre-determined DMRs may comprise at least 5, at least 10, at least 30, at least 50, or at least 100 DMRs.

In some embodiments the plurality of pre-determined DMRs comprises at least 1000, at least 10,000, at least 100,000 or at least 1 x 10⁶ DMRs. In some embodiments, the plurality of pre-determined DMRs comprises at least 2 x 10⁶ DMRs, optionally at least 5 x 10⁶ DMRs, further optionally at least 1 x 10⁷ DMRs.

The invention includes the combination of the aspects and preferred features described except where such a combination is clearly impermissible or expressly avoided.

Summary of the Figures

Embodiments and experiments illustrating the principles of the invention will now be discussed with reference to the accompanying figures in which:

Figure 1. Schematic of how cfDNA methylation assay analysis can be used for treatment personalisation in patients undergoing radiotherapy (RT).

Figure 2. Hypothetical cfDNA release profile during radiotherapy (RT) and post-radiotherapy.

Figure 3. Composition of healthy plasma cfDNA vs. plasma cfDNA from cancer patients. Healthy plasma cfDNA is predominantly from leukocytes (-55%) and erythroblasts (-30%), with -10% derived from the vascular endothelium and the rest from other tissues. Plasma cfDNA from cancer patients comprises a tumoral fraction (from -0.1% to -50% depending on tumour stage and size (<1% in early-stage cancers)).

Figure 4A-4C. (4A) DNA methylation biomarker discovery pipeline. (4B) Methylation data from >1000 publicly available arrays from relevant tissues to identify tissue-specific markers that could detect prostate tumour compared to normal prostate and other components of cfDNA in healthy individuals, and also tissue-specific markers that could detect rectum, colon, small intestine and bladder, which are the major tissues that can suffer damage in RT treatment of prostate cancer. (4C) Multi-dimensional scaling (MDS) plot of publicly available methylation datasets (n=1367 samples) from relevant tissues, based on over 450K CpGs. Samples with methylation data from healthy/normal tissues are represented by their tissue types as follows: bladder (BLADDER), rectum (RECTUM), colon (COLON), liver (LIVER), prostate (PROSTATE), small intestine (SMALLJNT), vascular endothelium (VAS_ENDO), skeletal muscle (SKELMUS) and blood (BLOOD). PR_TUM refers to prostate tumour samples.

Figures 5A-5E. Boxplots showing distributions of methylation beta values of the top 30 most discriminative marker locations in the genomes (CpGs) in each pairwise tissue comparison, based on ranking by absolute median difference for each pairwise comparison. Data was generated in silica from publicly available data for these tissues. The markers are divided into those that were hypermethylated and hypomethylated in the first tissue vs. the second tissue listed in each comparison. The top 30 CpG markers specific to any given pairwise tissue comparison show very distinct ranges of methylation beta values and therefore provide strong discriminatory power between the selected tissue and the major constituents of blood cfDNA. (A-D) Methylation of normal tissues vs. blood; (E) methylation of prostate tumour vs. blood.

Figures 6A-6B. Detection of prostate tumour cfDNA in five PRINTOUT (A) and six PRENOTE (B) patients. The X axis indicates percentage methylation difference between post radiotherapy (RT) (Day X) and pre RT (Day 1) samples across marker sets optimised to detect prostate tumour DNA in plasma cfDNA. Prostate tumour DNA was detected in all 11 patients, at P<0.05 or greater statistical significance. PRINTOUT and PRENOTE are the two ongoing prostate cancer RT studies from which samples for the data are drawn.

Figures 7A-7C. Detection of cfDNA derived from (A) bladder and (B) rectum, tissues adjacent to prostate tumour target area. Bladder-derived cfDNA was detected at P<0.01 in Patient PRL02 (at Days 3 and 4 of RT) and PRI.04 (Day 4 of RT); rectum-derived cfDNA was detected at P<0.01 in Patient PRL03 at Day 5 of RT. Note different colour coding of P-values in A and B. (C) Clinical RTOG scores of genitourinary (GU) and gastrointestinal (Gl) toxicity from Patients 01-05 of PRINTOUT study. Note comparison of Grade 2-3 GU toxicity in Patients PRI.02, 03 and 04, and of Grade 3 Gl toxicity in Patient PRI.03 with detected bladder- and rectum-derived cfDNA from these patients in Figures 7A and 7B.

Figure 8A-8D. Direct methylation sequencing data (Oxford Nanopore Technology) for leukocyte genomic

DNA (gDNA) sheared to the size of cfDNA (A, B, C) and cfDNA (D) samples. (A) Number of CpGs covered at 5X per chromosome (sheared gDNA). (B) Read coverage across chromosome 1 based on non-overlapping 10Kb windows. (C) Methylation percentage based on CpGs covered at 5X on chromosome 1 . (D) Number of CpGs covered at 5X per chromosome (cfDNA).

Figure 9A-9D. Boxplots showing distribution of methylation beta values extracted from WGBS datasets. (A) “Prostate Tumour vs Blood” marker sets tested against WGBS data from healthy cfDNA, white blood cells (WBC), normal prostate (prNormal), prostate tumour (prTumour) which includes both lethal prostate tumour and non-lethal prostate tumour, lethal prostate tumour (prTumourL) and non-lethal prostate tumour (prTumourNL). (B) “Prostate Tumour vs AH” marker sets tested against WGBS data from healthy cfDNA, white blood cells, normal prostate, prostate tumour, lethal prostate tumour and non-lethal prostate tumour. (C) “Prostate Tumour vs Blood” marker sets tested against healthy cfDNA, bladder, colon, prostate, and white blood cells. (D) “Prostate Tumour vs AH” marker set tested against healthy cfDNA, bladder, colon, prostate, and white blood cells.

Figure 10A-10D. Boxplots showing distribution of methylation beta values extracted from WGBS datasets. (A) “Bladder vs Blood” marker sets tested against WGBS data from healthy cfDNA, white blood cells (WBC), normal prostate (prNormal), prostate tumour (prTumour) which includes both lethal prostate tumour and non-lethal prostate tumour, lethal prostate tumour (prTumourL) and non-lethal prostate tumour (prTumourNL). (B) “Bladder vs AH” marker sets tested against WGBS data from healthy cfDNA, white blood cells, normal prostate, prostate tumour, lethal prostate tumour and non-lethal prostate tumour. (C) “Bladder vs Blood” marker sets tested against healthy cfDNA, bladder, colon, prostate, and white blood cells. (D) “Bladder vs All” marker set tested against healthy cfDNA, bladder, colon, prostate, prostate, and white blood cells. (D) “Colon vs All” marker set tested against healthy cfDNA, bladder, colon, prostate, and white blood cells.

Figure 11 A-11 D. Boxplots showing distribution of methylation beta values extracted from WGBS datasets. (A) “Colon vs Blood” marker sets tested against WGBS data from healthy cfDNA, white blood cells (WBC), normal prostate (prNormal), prostate tumour (prTumour) which includes both lethal prostate tumour and non-lethal prostate tumour, lethal prostate tumour (prTumourL) and non-lethal prostate tumour (prTumourNL). (B) “Colon vs AH” marker sets tested against WGBS data from healthy cfDNA, white blood cells, normal prostate, prostate tumour, lethal prostate tumour and non-lethal prostate tumour. (C) “Colon vs Blood” marker sets tested against healthy cfDNA, bladder, colon, prostate, and white blood cells. (D) “Colon vs All” marker set tested against healthy cfDNA, bladder, colon, prostate, and white blood cells.

Figure 12A-12D. Boxplots showing distribution of methylation beta values extracted from WGBS datasets. (A) “Rectum vs Blood” marker sets tested against WGBS data from healthy cfDNA, white blood cells (WBC), normal prostate (prNormal), prostate tumour (prTumour) which includes both lethal prostate tumour and non-lethal prostate tumour, lethal prostate tumour (prTumourL) and non-lethal prostate tumour (prTumourNL). (B) “Rectum vs AH” marker sets tested against WGBS data from healthy cfDNA, white blood cells, normal prostate, prostate tumour, lethal prostate tumour and non-lethal prostate tumour. (C) “Rectum vs Blood” marker sets tested against healthy cfDNA, bladder, colon, prostate, and white blood cell. (D) “Rectum vs AH” marker set tested against healthy cfDNA, bladder, colon, prostate, and white blood cells.

Figure 13A-13D. Distribution of normalised read counts (y-axis) per read-level methylation ranges (x- axis, from lowest (Bini) to highest (Bini 1) methylation values). HypoM marker was adjusted to 1-hypoM. “Prostate Tumour vs AH” marker set was tested against four control tissues (gDNA libraries) sequenced using Oxford Nanopore Sequencing platform. (A) Bladder control tissue. (B) Leukocyte control tissue. (C) Colon control tissue. (D) Rectum control tissue.

Figure 14A-14D. Distribution of normalised read counts (y-axis) per read-level methylation ranges (x- axis, from lowest (Bini) to highest (Bini 1) methylation values). HypoM marker was adjusted to 1-hypoM. “Colon vs All” marker set was tested against four control tissues (gDNA libraries) sequenced using Oxford Nanopore Sequencing platform. (A) Bladder control tissue. (B) Leukocyte control tissue. (C) Colon control tissue. (D) Rectum control tissue.

Figure 15A-15D. Distribution of normalised read counts (y-axis) per read-level methylation ranges (x- axis, from lowest (Bini) to highest (Bini 1) methylation values). HypoM marker was adjusted to 1-hypoM. “Rectum vs AH” marker set was tested against four control tissues (gDNA libraries) sequenced using Oxford Nanopore Sequencing platform. (A) Bladder control tissue. (B) Leukocyte control tissue. (C) Colon control tissue. (D) Rectum control tissue.

Figure 16A-16D. Distribution of normalised read counts (y-axis) per read-level methylation ranges (x- axis, from lowest (Bini) to highest (Bini 1) methylation values). HypoM marker was adjusted to 1-hypoM. “Bladder vs AH” marker set was tested against four control tissues (gDNA libraries) sequenced using Oxford Nanopore Sequencing platform. (A) Bladder control tissue. (B) Leukocyte control tissue. (C) Colon control tissue. (D) Rectum control tissue.

Figure 17. Single-Molecule Sequencing with Oxford Nanopore (R10, kit14). Showing number of CpGs covered at 5x in plasma cfDNA. Detection of over 24 Million CpGs in plasma cfDNA.

Detailed Description of the Invention

Aspects and embodiments of the present invention will now be discussed with reference to the accompanying figures. Further aspects and embodiments will be apparent to those skilled in the art. All documents mentioned in this text are incorporated herein by reference.

Definitions

To help understand the present application, the meanings of some of the terms as used in the context of the present invention are explained below.

A “sample” as used herein may be a cell or tissue sample, a biological fluid, an extract (e.g., a DNA extract obtained from the subject), from which genomic material can be obtained for genomic analysis. The sample may be a biological fluid sample obtained by liquid biopsy. Any suitable biological fluid sample can be envisaged for the present invention. For example, the biological fluid sample may be a blood sample, plasma sample, serum, lymphatic fluid, synovial fluid, ascites fluid, interstitial or extracellular fluid, cerebrospinal fluid, saliva, mucus, semen, sweat, urine or any other bodily fluids. In some embodiments the biological fluid sample is selected from blood, urine and plasma. Within the context of the present invention, the sample may be a blood sample. The sample may be a cell-free DNA (cfDNA) sample (e.g., a plasma sample). In some embodiments, the cfDNA sample may contain circulating tumour DNA (ctDNA). In some embodiments, the sample is a sample obtained from a subject, such as a human subject. The sample is preferably from a mammal, more preferably from a human. The sample may be one which has been freshly obtained from the subject or may be one which has been processed and/or stored prior to making a determination. Further, the sample may be transported and/or stored, and collection may take place at a location remote from the genomic sequence data acquisition (e.g., sequencing) location, and/or any method steps described herein may take place at a location remote from the sample collection location and/or remote from the genomic data acquisition (e.g., sequencing) location. A plurality of samples may be taken from a single patient e.g., serially during a course of treatment.

In embodiments where the sample is a biological fluid, the sample may have a volume of at least about 10pl, at least about 20 pl, at least about 30pl, at least about 40pl, at least about 50pl, at least about 60pl, at least about 70pl, at least about 80pl, at least about 90pl or at least about 10Opl. For example, in some embodiments, the sample has a volume of at least about 50pl.

As used herein, the term “control methylation profile” will be understood to be an optional reference or baseline methylation profile. This may be generated from a sample obtained from the same subject at a different period in time to when the sample from which the sample methylation profile was generated was obtained from the subject. Alternatively, the control methylation profile may be generated from a sample obtained from a different subject. In some embodiments, the control methylation profile is generated from a first cfDNA-containing sample obtained from the subject prior to receiving radiotherapy and the sample methylation profile is generated from a second cfDNA-containing sample obtained from the subject following at least one dose of radiotherapy. Preferably, the first and second cfDNA-containing samples are each biological fluids, more preferably each blood samples. In some embodiments, a third cfDNA- containing sample is obtained from the subject following a second dose of radiotherapy, and methylation levels within the cfDNA at pre-determined DMRs in the first, second and third cfDNA-containing sample are compared or analysed to determine tumour response and/or non-tumour tissue toxicity. In some embodiments, cfDNA-containing samples are obtained regularly from the subject over the course of the subject’s treatment. In some embodiments, cfDNA-containing samples are obtained for the duration of the subject’s treatment (for example, before and/or after each dose of radiotherapy over the course of treatment). In embodiments comprising at least two cfDNA-containing samples obtained from the same subject, the sample obtained at a later time point may be used to generate the sample methylation profile and the sample obtained at an earlier time point may be used to generate the control methylation profile. Thus, in some embodiments, the sample methylation profile is compared to a plurality of control methylation profiles, each of the plurality of control methylation profiles generated from cfDNA samples each obtained from the subject at a different earlier timepoint.

In some embodiments, the control methylation profile is generated from a plurality of cfDNA-containing samples obtained from subjects having prostate cancer prior to receiving radiotherapy. Alternatively, the control methylation profile may be generated from one or more cfDNA-containing samples obtained from at least one healthy subject. The control methylation profile thus advantageously acts as a baseline or reference to which the sample methylation profile can be compared.

The term “tumour” refers to an abnormal mass of tissue resulting from a benign (non-cancerous) or malignant (cancerous) neoplastic process.

The term “cancer” refers to a disease caused by an uncontrolled division of abnormal cells in a part of the body.

The term “subject” or “patient” refers to all classes of animals, but in particular mammals. The subject may have a prostate tumour or prostate cancer or may be suspected of having a prostate tumour or prostate cancer. In some embodiments, the subject is male. In some embodiments, the subject has an age of at least about 50 years, an age of at least about 55 years, an age of at least about 60 years, an age of at least about 65 years or an age of at least about 70 years. In some embodiments, the subject has an age of less than about 100 years, of less than about 95 years, of less than about 90 years, of less than about 85 years or of less than about 80 years. In some embodiments, the subject has an age of at least about 50 years. In some embodiments, the subject has an age of at least about 60 years. In some embodiments, the subject has an age of at least about 70 years.

The term “healthy” in the context of this application refers to individuals known not to have prostate cancer, or individuals known not to have a prostate tumour.

The term “liquid biopsy” refers to the analysis of cfDNA and other cellular components released from dying or damaged cells into the circulation and other bodily fluids.

The term “cell-free DNA” or “cfDNA” refers to DNA fragments released from dying or damaged cells into the circulation and other bodily fluids. This includes DNA that is freely circulating in the bloodstream; it may be tumour-derived or non-tumour derived.

The term “circulating tumour DNA” or “ctDNA” refers to tumour-derived DNA fragments released into the circulation and other bodily fluids. ctDNA is a type of cfDNA that is tumour-derived.

In the context of the present invention, the term “single-molecule sequencing” refers to DNA sequencing methods which are capable of reading the base sequence directly from individual strands of DNA.

Advantageously, this enables rapid analysis of DNA sequencing, often from relatively small sample sizes which may otherwise be challenging to accurately analyse. Further advantageously, the use, in some embodiments, of single-molecule sequencing enables the real-time generation of methylation results from the subject. This, advantageously, enables the methylation results to be provided to the clinical team during the course of the subject’s radiotherapy, thereby enabling the real-time adaptation of the subject’s radiotherapy dosing or schedule. In some embodiments, the single-molecule sequencing comprises or consists of nanopore sequencing. The nanopore sequencing may be performed using an Oxford Nanopore platform. Alternatively, the single-molecule sequencing may be performed using a PacBio platform.

“Nanopore sequencing” is a single-molecule sequencing technique. The technique enables direct analysis of DNA or RNA fragments. In this context, the term “direct sequencing” refers to the ability of this technique to analyse native DNA without the need for amplification or chemical conversion/labelling (e.g., bisulfite conversion).

The term “DNA methylation” refers to the epigenetic marker involving the covalent transfer of a methyl group to the C-5 position of the cytosine ring of DNA by DNA methyltransferases.

As used herein "treatment" refers to reducing, alleviating or eliminating one or more symptoms of the disease that is being treated, relative to the symptoms prior to treatment. "Prevention" (or prophylaxis) refers to delaying or preventing the onset of the symptoms of the disease. Prevention may be absolute (such that no disease occurs) or may be effective only in some individuals or for a limited amount of time.

“Adaptive radiotherapy” (also referred to as “adaptive radiation therapy”) is a radiotherapy term that involves adjusting the treatment dose, schedule (e.g., the frequency) of doses in response to a particular factor or treatment field during the course of their treatment. As used herein, adaptive radiotherapy may refer to, e.g. decreasing the dose of radiotherapy, decreasing the field of radiotherapy or moving the field of radiotherapy if non-tumour tissue toxicity is detected, increasing the dose of radiotherapy, increasing the field of radiotherapy or moving the field of radiotherapy if the tumour is not responding to treatment or is not responding well enough to treatment, or increasing the dose of radiotherapy if there is not found to be non-tumour tissue toxicity. Adaptive radiotherapy may also result in the discontinuation of treatment, or the use of alternative treatments such as surgery, immunotherapy or chemotherapy.

“Non-tumour tissue toxicity” (also referred to as “off-target toxicity”) as used herein may be assessed by the presence of cfDNA in a sample taken from the patient or subject, such cfDNA being derived from a non-tumour tissue of origin. For example, the presence of bladder-derived cfDNA, colon-derived cfDNA, small intestine-derived cfDNA, rectum-derived cfDNA or healthy prostate cfDNA may indicate the presence of non-tumour tissue toxicity. Non-tumour tissue toxicity may also be assessed by the presence of DNA methylation at a plurality of pre-determined DMRs associated with a particular tissue of origin that is adjacent non-tumour tissue, such as bladder, small intestine, colon, rectum or healthy prostate. The term “radiotherapy” (also called “radiation therapy”) (abbreviated to “RT”) is a treatment where radiation is used to kill cancer cells. In the context of the present invention, in some embodiments the radiotherapy is external beam radiotherapy.

In the context of the present invention, the term prostate cancer refers to cancer of the prostate. Prostate cancer may include, but not necessarily be limited to, adenocarcinoma of the prostate, transitional cell carcinoma of the prostate, squamous cell carcinoma of the prostate, neuroendocrine prostate cancer, small cell prostate cancer, sarcoma of the prostate or lymphoma of the prostate. In some embodiments the prostate cancer is selected from adenocarcinoma of the prostate, transitional cell carcinoma of the prostate, squamous cell carcinoma of the prostate, neuroendocrine prostate cancer and small cell prostate cancer. Adenocarcinoma of the prostate may comprise acinar adenocarcinoma of the prostate or ductal adenocarcinoma of the prostate.

In the context of the present invention, the term “radiotherapy response” is used to define the response of the subject to radiotherapy. The response may comprise any response in the subject to radiotherapy, for example a side effect to radiotherapy and/or the response of the prostate cancer tumour to radiotherapy. This may, for example, be measured as a reduction in prostate tumour size. The reduction may be a reduction of at least about 10%, at least about 20%, at least about 30%, at least about 40%, or at least about 50% relative to the size of the tumour prior to the subject having radiotherapy. A side effect to radiotherapy may comprise non-tumour toxicity to radiotherapy. Thus, in some embodiments, the radiotherapy response comprises non-tumour toxicity to radiotherapy. In some embodiments, the radiotherapy response comprises prostate tumour response to radiotherapy.

The term “and/or” where used herein is to be taken as specific disclosure of each of the two specified features or components with or without the other. For example, “A and/or B” is to be taken as specific disclosure of each of (i) A, (ii) B and (iii) A and B, just as if each is set out individually herein.

Throughout this specification, including the claims which follow, unless the context requires otherwise, the word “comprise” and “include”, and variations such as “comprises”, “comprising”, and “including” will be understood to imply the inclusion of a stated integer or step or group of integers or steps but not the exclusion of any other integer or step or group of integers or steps.

It must be noted that, as used in the specification and the appended claims, the singular forms “a,” “an,” and “the” include plural referents unless the context clearly dictates otherwise. Ranges may be expressed herein as from “about” one particular value, and/or to “about” another particular value. When such a range is expressed, another embodiment includes from the one particular value and/or to the other particular value. Similarly, when values are expressed as approximations, by the use of the antecedent “about,” it will be understood that the particular value forms another embodiment. The term “about” in relation to a numerical value is optional and means for example +/- 10%.

cfDNA is released into the circulation by damaged and diseased tissues, including cancers. DNA methylation is an epigenetic marker whose profile is specific to each tissue, including cancerous tissues. Methylation analysis of circulating cfDNA can therefore indicate the extent of tissue damage caused by radiation therapy or radiotherapy (RT). Similarly, methylation of circulating ctDNA can indicate the response (or extent of response) to radiotherapy (RT).

In healthy individuals (e.g., those without prostate cancer or a prostate tumour), or in an individual having prostate cancer who has not yet received radiotherapy treatment, there may be minimal or no cfDNA derived from the bladder, small intestine, colon or rectum. Thus, DNA methylation at differentially methylated regions (DMRs) in cfDNA from these tissues may be absent or undetectable in the plasma of these patients.

In an individual having prostate cancer or a prostate tumour who has received at least one dose of radiotherapy, DNA methylation at DMRs within the cfDNA derived from tissues such as the bladder, small intestine, colon, rectum or healthy prostate may be detectable, indicating non-tumour tissue toxicity. In some embodiments, the DNA methylation levels can be used to determine the extent of non-tumour tissue toxicity. In the case of determining non-tumour tissue toxicity, the presence of DNA methylation at DMRs within the cfDNA from non-tumour tissue such as the bladder, small intestine, colon, rectum or healthy prostate may indicate the need to reduce the radiotherapy dose, alter the radiotherapy field or increase the timing between (or decrease the frequency of) radiotherapy doses or fractions.

In an individual having prostate cancer or a prostate tumour who has received at least one dose of radiotherapy, the detection of DNA methylation at DMRs within prostate tumour-derived cfDNA may indicate that the radiotherapy is effective. The absence of detectable prostate tumour-derived cfDNA (or insufficient levels of detectable prostate tumour-derived cfDNA) together with an absence of cfDNA derived from non-tumour tissues may indicate the need to increase the radiotherapy dose or decrease the timing between (or increase the frequency of) radiotherapy doses or fractions.

Thus, provided herein is a method of monitoring radiotherapy response in a subject having prostate cancer, the method comprising a) providing a cell-free DNA (cfDNA)-containing sample obtained from the subject; b) performing DNA sequencing on the cfDNA-containing sample in order to generate methylation sequencing reads; c) assessing methylation levels within the cfDNA at each of a plurality of predetermined differentially methylated regions (DMRs) to generate a sample methylation profile, wherein each pre-determined DMR is associated with a tissue of origin, thereby detecting the presence of one or more tissues of origin; and d) determining the radiotherapy response of the subject based on the sample methylation profile. In some embodiments, determining the radiotherapy response of the subject based on the sample methylation profile comprises comparing the sample methylation profile to a control methylation profile, wherein a significant difference in methylation level between the sample methylation profile and the control methylation profile is indicative of a response in the subject. In some embodiments, the control methylation profile is generated from a first cfDNA-containing sample obtained from the subject prior to receiving radiotherapy and the sample methylation profile is generated from a second cfDNA-containing sample obtained from the subject following at least one dose of radiotherapy.

In some embodiments the DNA sequencing comprises or consists of single-molecule DNA sequencing.

In some embodiments, the radiotherapy response comprises prostate tumour response to radiotherapy. Thus, the present invention also provides a method of determining prostate tumour response to radiotherapy in a subject having prostate cancer, the method comprising steps (a) to (d) as described herein.

Thus, the present invention also provides a method of determining non-tumour tissue toxicity to radiotherapy in a subject having prostate cancer, the method comprising steps (a) to (d).

In some embodiments, the one or more tissues of origin comprise adjacent non-tumour tissue. This may refer to tissues that are in proximity to the prostate and thus may be affected by radiotherapy, such as healthy prostate, the bladder, small intestine, colon and/or rectum.

The methods provided herein may require the pre-selection of DMRs that are associated with a particular tissue of origin. Any of the DMRs listed in Tables 1 and/or 2 may be pre-selected. For example, DMRs may be selected that are associated with the bladder, small intestine, colon or rectum. The presence of hypermethylation or hypomethylation at these DMRs may indicate non-tumour toxicity at a particular tissue of origin. In some embodiments, the pre-determined DMRs are associated with the bladder.

In some embodiments, the pre-determined DMRs are associated with the small intestine.

In some embodiments, the pre-determined DMRs are associated with the colon.

In some embodiments, the pre-determined DMRs are associated with the rectum.

In some embodiments, the pre-determined DMRs are associated with healthy prostate.

In some embodiments, the pre-determined DMRs are associated with prostate tumour, prostate tumour tissue or prostate cancer.

In some embodiments, the pre-determined DMRs are associated with a combination of tissues of origin selected from the healthy prostate, prostate tumour, prostate tumour tissue, small intestine, colon and/or rectum.

In some embodiments, the pre-determined DMRs may be described as one or more “DNA methylation signature(s)” that is/are associated with a particular tissue of origin or plurality of tissues of origin. In some embodiments, the DNA methylation signature is associated with healthy prostate or healthy prostate tissue. In some embodiments, the DNA methylation signature is associated with prostate tumour or cancer. In some embodiments, the DNA methylation signature is associated with the bladder. In some embodiments, the DNA methylation signature is associated with the small intestine. In some embodiments, the DNA methylation signature is associated with the colon. In some embodiments, the DNA methylation signature is associated with the rectum.

In some embodiments, the DMRs are selected from the markers or CpG sites described in Table 1 and/or Table 2 below. Each CpGJD represents a DMR associated with a particular tissue of origin. In some embodiments, the plurality of pre-determined DMRs comprises at least 5, at least 10, at least 30, at least 50 or at least 100 DMRs. In some embodiments the plurality of pre-determined DMRs comprises at least 1000, at least 10,000, at least 100,000 or at least 1 x 10⁶ DMRs. In some embodiments, the plurality of pre-determined DMRs comprises at least 2 x 10⁶ DMRs, optionally at least 5 x 10⁶ DMRs, further optionally at least 1 x 10⁷ DMRs.

In some embodiments, the pre-determined DMRs comprise at least five, optionally at least 10, further optionally at least 20, of the DMRs labelled in Table 1 as “prtum_vsAII” (see column five of Table 1). The inventors have advantageously found that the DMRs categorised in Table 1 as “prtum_vsAII” are particularly suitable to determine the methylation profile of the prostate tumour and hence to determine prostate tumour response to radiotherapy. In some embodiments, the pre-determined DMRs comprise all DMRs labelled in Table 1 as “prtum_vsAII”. In embodiments wherein the radiotherapy response comprises prostate tumour response to radiotherapy, a significant difference in methylation may comprise a significant increase in methylation in the sample methylation profile relative to a control methylation profile. Thus, the sample methylation profile may be hypermethylated, compared to the control methylation profile. In such embodiments, the prostate tumour response may be determined to be high (i.e., the tumour is responding effectively to radiotherapy treatment and is reducing in size). In embodiments wherein the radiotherapy response comprises non-tumour toxicity to radiotherapy, the significant difference in methylation may comprise a significant decrease in methylation in the sample methylation profile relative to a control methylation profile. Thus, the sample methylation profile may be hypomethylated, compared to the control methylation profile. In such embodiments, non-tumour toxicity to radiotherapy may be identified and the radiotherapy response may be altered as described herein.

In some embodiments, the pre-determined DMRs comprise at least five, optionally at least 10, further optionally at least 20, of the DMRs labelled in Table 1 as “bladder_vs_blood” (see column five of Table 1). The inventors have advantageously found that these DMRs are particularly effective at representing potential toxicity of adjacent tissue to the prostate tumour, particularly bladder tissue, and so are particularly suitable to determine non-tumour toxicity to radiotherapy. In some embodiments, the predetermined DMRs comprise all DMRs labelled in Table 1 as “bladder_vs_blood”.

In some embodiments, the pre-determined DMRs comprise at least five, optionally at least 10, further optionally at least 20, of the DMRs labelled in Table 1 as “colon_vs_blood” (see column five of Table 1). The inventors have advantageously found that these DMRs are particularly effective at representing potential toxicity of adjacent tissue to the prostate tumour, particularly colon tissue, and so are particularly suitable to determine non-tumour toxicity to radiotherapy. In some embodiments, the pre-determined DMRs comprise all DMRs labelled in Table 1 as “colon_vs_blood”.

In some embodiments, the pre-determined DMRs comprise at least five, optionally at least 10, further optionally at least 20, of the DMRs labelled in Table 1 as “rectum_vs_blood” (see column five of Table 1). The inventors have advantageously found that these DMRs are particularly effective at representing potential toxicity of adjacent tissue to the prostate tumour, particularly rectum tissue, and so are particularly suitable to determine non-tumour toxicity to radiotherapy. In some embodiments, the predetermined DMRs comprise all DMRs labelled in Table 1 as “rectum_vs_blood”.

In some embodiments, the pre-determined DMRs comprise at least five, optionally at least 10, further optionally at least 20, of the DMRs labelled in Table 1 as “smalllnt_vs_blood” (see column five of Table 1). The inventors have advantageously found that these DMRs are particularly effective at representing potential toxicity of adjacent tissue to the prostate tumour, particularly small intestine tissue, and so are particularly suitable to determine non-tumour toxicity to radiotherapy. In some embodiments, the predetermined DMRs comprise all DMRs labelled in Table 1 as “smalllnt_vs_blood”.

In some embodiments, the pre-determined DMRs comprise at least five, optionally at least 10, further optionally at least 20, of the DMRs labelled in Table 1 as “bladder_vsAII” (see column five of Table 1). The inventors have advantageously found that these DMRs are particularly effective at representing potential toxicity of adjacent tissue to the prostate tumour, particularly bladder tissue, and so are particularly suitable to determine non-tumour toxicity to radiotherapy. In some embodiments, the predetermined DMRs comprise all DMRs labelled in Table 1 as “bladder_vsAII”. In some embodiments, the pre-determined DMRs comprise at least five, optionally at least 10, further optionally at least 20, of the DMRs labelled in Table 1 as “smallint_vsAH” (see column five of Table 1). The inventors have advantageously found that these DMRs are particularly effective at representing potential toxicity of adjacent tissue to the prostate tumour, particularly small intestine tissue, and so are particularly suitable to determine non-tumour toxicity to radiotherapy. In some embodiments, the predetermined DMRs comprise all DMRs labelled in Table 1 as “smallint_vsAH”.

In some embodiments, the pre-determined DMRs comprise at least five, optionally at least 10, further optionally at least 20, of the DMRs labelled in Table 1 as “colon_vsAII” (see column five of Table 1). The inventors have advantageously found that these DMRs are particularly effective at representing potential toxicity of adjacent tissue to the prostate tumour, particularly colon tissue, and so are particularly suitable to determine non-tumour toxicity to radiotherapy. In some embodiments, the pre-determined DMRs comprise all DMRs labelled in Table 1 as “colon_vsAII”.

In some embodiments, the pre-determined DMRs comprise at least five, optionally at least 10, further optionally at least 20, of the DMRs labelled in Table 1 as “rectum_vsAH” (see column five of Table 1). The inventors have advantageously found that these DMRs are particularly effective at representing potential toxicity of adjacent tissue to the prostate tumour, particularly rectum tissue, and so are particularly suitable to determine non-tumour toxicity to radiotherapy. In some embodiments, the pre-determined DMRs comprise all DMRs labelled in Table 1 as “rectum_vsAII”.

In some embodiments, the pre-determined DMRs are selected from the DMRs labelled in Table 1 as “prtum_vsAII”, “bladder_vs_blood”, “colon_vs_blood”, “rectum_vs_blood” and “smalllnt_vs_blood”. For example, the pre-determined DMRs may comprise at least 5, at least 10, at least 20, at least 30, or at least 100 DMRs labelled in Table 1 as “prtum_vsAII”, “bladder_vs_blood”, “colon_vs_blood”, “rectum_vs_blood” or “smalllnt_vs_blood”. The pre-determined DMRs may comprise at least 1000 DMRs labelled in Table 1 as “prtum_vsAH”, “bladder_vs_blood”, “colon_vs_blood”, “rectum_vs_blood” or “smalllnt_vs_blood”. In some embodiments, the pre-determined DMRs are selected from the DMRs labelled in Table 1 as “prtum_vsAII”, “bladder_vsAII”, “smallint_vsAII”, “colon_vsAII”, and “rectum_vsAII”. For example, the pre-determined DMRs may comprise at least 5, at least 10, at least 20, at least 30, or at least 100 DMRs labelled in Table 1 as “prtum_vsAII”, “bladder_vsAII”, “smallint_vsAII”, “colon_vsAH”, and “rectum_vsAII”.

It will also be appreciated that one or more steps of the methods of the present invention may be computer implemented. For example, the method of monitoring radiotherapy response in a subject having prostate cancer as described herein may be a computer implemented method of monitoring radiotherapy response in a subject having prostate cancer. For example, step c) of the computer implemented method may comprise using a machine learning model to assess methylation levels within the cfDNA at each of a plurality of pre-determined DMRs to generate a sample methylation profile, wherein each pre-determined DMR is associated with a tissue of origin, thereby detecting the presence of one or more tissues of origin. In some embodiments, step d) comprises using a machine learning model to determine the radiotherapy response of the subject based on the sample methylation profile. For example, step d) may comprise using a machine learning model to compare the sample methylation profile to a control methylation profile, wherein a significant difference in methylation level between the sample methylation profile and the control methylation profile is indicative of a response in the subject.

The machine learning model may comprise a regression model, for example a logistic regression model. The machine learning model may comprise a plurality of models, which may otherwise be referred to as an ensemble model. In embodiments comprising a regression model, using the regression model may result in an output which can be used to classify the response in the subject.

Exemplary methods of treatment

Also provided by the present invention is a method of treating a subject with prostate cancer using adaptive radiotherapy, the method comprising administering at least one dose of radiotherapy to the subject, further comprising the steps of the first aspect (the method of monitoring radiotherapy response) described herein, further comprising adjusting the dose of radiotherapy, altering the radiotherapy field and/or the frequency of radiotherapy. In some embodiments, radiotherapy treatment may be discontinued. In some embodiments, alternative treatments may be used, such as surgery, immunotherapy or chemotherapy.

Also provided is a method for the detection and/or staging of prostate cancer. Such a method may be useful for determining the prognosis of a subject. The method may comprise one or more of the methods disclosed herein. For example, such a method may comprise a) providing a cell-free DNA (cfDNA)- containing sample obtained from a subject, b) performing DNA sequencing on the cfDNA-containing sample in order to generate methylation sequencing reads, c) assessing methylation levels within the cfDNA at each of a plurality of pre-determined differentially methylated regions (DMRs) to generate a sample methylation profile, wherein each pre-determined DMR is associated with prostate tumour tissue and/or a tissue of origin, thereby detecting the presence or absence of prostate tumour tissue and/or one or more tissues of origin; and thereby determining whether the subject does, or does not have prostate cancer. In some embodiments, the plurality of pre-determined DMRs comprise at least five, optionally at least 10, further optionally at least 20, of the DMRs labelled in Table 1 as “prtum_vsAII” (see column five of Table 1). In some embodiments, the pre-determined DMRs comprise all DMRs labelled in Table 1 as “prtum_vsAII”. The method may further comprise identifying subjects with lethal or non-lethal disease.

Adaptive radiotherapy

In the methods provided herein, based on the presence or extent of methylation of pre-determined DMRs within cfDNA released by the surrounding normal tissue (i.e. , bladder, small intestine, colon, rectum or healthy prostate) and the presence or extent of methylation of pre-determined DMRs within cfDNA released by the prostate tumour tissue, the radiotherapy treatment for a particular subject can be adapted. In some embodiments, the adaptation is an increase in dose per fraction to increase tumour kill. In some embodiments, the adaptation is a reduction in dose to the surrounding normal tissues (i.e. , bladder, small intestine, colon, rectum or healthy prostate) by reducing the treatment margin adjacent to the tissue at risk. In some embodiments, the adaptation is the replanning of treatment to weight the dose away from the tissue at risk, in effect reducing the dose per fraction and hence total dose administered to the tissue at risk. In some embodiments, the adaption is a reduction in dose to the surrounding normal tissues by changing the radiotherapy field to, for example, exclude more non-tumour tissue.

In some embodiments, the radiotherapy dose or dose per fraction is increased. In some embodiments, the radiotherapy dose or dose per fraction is decreased. In some embodiments, the frequency of radiotherapy doses is decreased. In some embodiments, the frequency of radiotherapy doses is increased. In some embodiments, the time between doses is increased. In some embodiments, the time between doses is decreased.

Generally, radiotherapy dose per fraction schedules can be from 2Gy per day (Conventional Radiotherapy), 3Gy per day (Moderate Hypofractionation) and Stereotactic Body Radiotherapy (SBRT) 7.25-8Gy per day (UltraHypofractionation). Therefore, in some embodiments, a reduction in dose per fraction may comprise less than 2Gy per day, optionally less than 3Gy per day or less than 7Gy per day. Likewise, in some embodiments, an increase in dose per fraction may comprise more than 2Gy per day, optionally more than 3Gy per day, further optionally more than 5GY per day. In some embodiments an increase in dose per fraction may comprise more than 7Gy per day or more than 8Gy per day.

Various forms of radiotherapy are known to the skilled person and encompassed by the present invention. Preferably, the radiotherapy is external beam radiotherapy. External beam radiotherapy can be delivered by various platforms including, but not necessarily limited to Linear Accelerator, Tomotherapy, MR-Linac and Proton Therapy.

The features disclosed in the foregoing description, or in the following claims, or in the accompanying drawings, expressed in their specific forms or in terms of a means for performing the disclosed function, or a method or process for obtaining the disclosed results, as appropriate, may, separately, or in any combination of such features, be utilised for realising the invention in diverse forms thereof.

While the invention has been described in conjunction with the exemplary embodiments described above, many equivalent modifications and variations will be apparent to those skilled in the art when given this disclosure. Accordingly, the exemplary embodiments of the invention set forth above are considered to be illustrative and not limiting. Various changes to the described embodiments may be made without departing from the spirit and scope of the invention.

For the avoidance of any doubt, any theoretical explanations provided herein are provided for the purposes of improving the understanding of a reader. The inventors do not wish to be bound by any of these theoretical explanations.

Any section headings used herein are for organizational purposes only and are not to be construed as limiting the subject matter described. Examples

EXAMPLE 1: Investigating the feasibility of using DNA methylation analysis to define cfDNA tissue-of-

Materials and Methods

As the aim was to discover tissue- and cancer-specific methylation biomarkers for liquid biopsy, we used a three-fold approach. We first compared each tissue of interest against all other tissues from which DNA may be present in the plasma, this is the “X vs all” step and was expected to yield the most specific CpGs. Then we compared each tissue of interest against the haematopoietic samples (leukocytes and erythroblasts), this is the “X vs blood” step and was expected to yield the most sensitive CpGs, as haematopoietic profiles have the most distinct methylation profiles of all tissues analysed. Finally, for tissues that are closely related but need to be distinguished for the purpose of this study, we compared them in a pairwise fashion, this is the “pairwise” step, and was expected to improve the classification accuracy between these tissues. This led to 23 comparisons for the differential methylation analysis.

For each comparison, the M-values were analysed by both the dmpFinder method from the minfi package, and the limma package correcting for dataset as a potential batch effect to determine a robust set of the most differentially methylated CpGs within the training set. We ranked the CpGs by adjusted p-value or q- value and then by median difference in p-value. For each method, we kept the top 3000 CpGs of the list, and then created a list of the CpGs in common between the two methods. We then removed from each comparison any CpG found in another comparison. For each of the 23 comparisons, we retained the top 100 CpGs, except for comparisons that yielded fewer CpGs, as the final set to be analysed on biological samples. We obtained a final list of 2090 CpGs from our differential methylation analysis of public 450k datasets. These are set out in Table 1 , which is presented after Example 2 below.

For the detection of prostate tumour in cfDNA samples from patients in the PRINTOUT and PRENOTE cohorts (Figure 6), the set of markers used was prostate tumour versus all other tissues/blood including in the overall set of comparisons (prtum_vsAII). The methylation signal for prostate tumour detection was refined and optimised by examining and selecting on an individual tumour-specific basis the markers that were lowly methylated (below 1%) at baseline i.e., pre-RT treatment. Regarding the detection of potential toxicity to the surrounding tissues (e.g., bladder, rectum) in Figure 7, the marker sets used were bladder versus blood (bladder_vs_blood) and rectum versus blood (rectum_vs_blood), without any additional selection of a subgroup of markers. Additionally, for each set of markers considered, either for prostate tumour detection or toxicity to surrounding tissues, marker regions were defined as the genomic regions centered at the position of the reference CpG markers and extended by 100 base pairs (bp) on either side of the reference dinucleotide CpGs, resulting in 202bp regions. The methylation signal for each marker region was defined as the average methylation (beta values) across all CpGs located within the 202bp region, excluding any CpG without sufficient read coverage (i.e., a minimum of 5 reads was required). Standard laboratory protocols were used for the remaining methods and are commonly available in the art. For example, standard ONT protocols are readily available and were used in Example 2.

Results

To demonstrate the feasibility of using DNA methylation analysis to define cfDNA tissue-of-origin, publicly available methylation data was mined to identify DNA methylation markers that could distinguish cfDNA derived from normal prostate and prostate cancer, and from the tissues surrounding the prostate, to identify the tissue(s) of origin of this cfDNA (Figures 4B and 4C). Building on data mining methods developed for analysis of breast cancer cfDNA (de Proce et al. Development of methylation-based biomarkers for breast cancer detection by model training and validation and synthetic cell-free DNA; bioRxiv doi 10.1101/2022.02.11.480085 (2002)) (37), a panel of 2,751 tissue-of-origin markers for analysis of cfDNA in prostate cancer patients were identified from publicly available Illumina methylation microarrays and scientific publications.

The 2,751 DNA methylation markers are set out in Tables 1 and 2, which are presented after the Example 2 below. Each DNA methylation marker is identified by a “CpGJD”, which refers to standardised Illumina nomenclature. The person skilled in the art can locate any of the markers by searching for their CpGJD in public databases. The CpGJD identifiers are consistent across genome assemblies. From the 2,751 markers identified in Tables 1 and 2 below, it is possible to derive a DNA methylation signature for each tissue of origin that can be used to identify this particular tissue in a cfDNA-containing sample obtained from a subject.

The top 30 markers from each pairwise tissue comparison were tested for their ability to discriminate tissues by in silico analysis, including the ability to detect DNA from prostate cancer tissue and from normal prostate, bladder, rectum and colon. The markers showed very distinct ranges of methylation values for these tissues, providing strong discriminatory power between DNA from tissues sampled and blood-derived DNA, and between these tissues and prostate cancer DNA. These markers were highly discriminatory for distinguishing DNA derived from prostate cancer, normal prostate, rectum, colon, and bladder from the normal blood components of cfDNA in healthy people (Figure 5A-5E).

These markers were then tested on cfDNA samples from prostate cancer patients undergoing radiotherapy treatment. Patient samples were obtained from patients taking part in ongoing trials of radiotherapy dosing and scheduling: PRENOTE (BioResource Ethics REF 15-ES-094) and PRINTOUT (NCT04081428). The results showed, first, that, by Day 5 of radiotherapy treatment, prostate tumour- derived cfDNA could be detected in the plasma of all 11 of the patients (Figure 6). This indicates the treatment (i.e., radiotherapy) response. Second, the results show that in a subset of the 11 patients, cfDNA released during radiotherapy from tissues surrounding the targeted tumour could be detected (Figure 7) (Bladder cfDNA detected in 3/3 patients with Grade 2+ GI/GU toxicity (Figure 7A) and rectum- derived cfDNA was detected in 1/2 patients with Grade 3 GU toxicity (Figure 7B)). This indicates treatment (i.e., radiotherapy) toxicity. Overall, these results demonstrate that the identified methylation biomarkers can identify DNA tissue-of- origin and therefore can provide a sensitive quantitative measure, in individual patients, of RT-induced damage to the tumour and surrounding tissues.

The 2,751 markers used to generate the results presented in Figures 5-7 were derived from publicly available Illumina methylation microarrays, which represent up to 850,000 locations (CpG sites) in the genome. While this appears to be a large number of locations, the number is limited by the Illumina microarray technology and is only around 3% of the 28.8 million total CpG sites in the genome from which such methylation data can be generated. By harnessing the power of defining markers from all 28.8 million CpG sites in whole-genome analyses, the sensitivity and specificity for detecting and measuring radiation-induced death of tumour tissue and damage to surrounding tissues during radiation treatment will be substantially increased. A single-molecule sequencing approach will allow for whole-genome methylation analysis where data can be generated early in a course of radiotherapy treatment.

EXAMPLE 2: Use of ONT platform to read methylation from genomic DNA

The inventors confirmed whether the ONT platform could be used to read methylation from genomic DNA. The protocol used was that described in Lau et al. (36). Leukocyte genomic DNA was used.

The results are shown in Figure 8. The data confirms (i) that the ONT platform can be used to read methylation from genomic DNA at 5.9 million CpGs across the genome, almost 10-fold that which can be achieved using Illumina microarrays (Figure 8A), (ii) even distribution of 5x sequence coverage of hundreds of thousands of CpGs across chromosomes (other than repetitive centromeric regions) (Figures 8B and 8C), and (iii) methylation calls at over 250,000 CpGs from plasma cfDNA (Figure 8D).

EXAMPLE 3: Testing methylation marker sets against Whole Genome Bisulfite Seguence data

The markers identified in Example 1 (see Table 1) were tested for their ability to identify specific tissues.

Materials and Methods

Whole-Genome Bisulfite Sequencing

Publicly available whole-genome bisulfite sequencing (WGBS) datasets were extracted for a set of reference healthy tissues and cfDNA samples. WGBS data from plasma cfDNA samples of 23 healthy adult individuals were extracted from Fox-Fisher et al. Elife 2021 .

A set of WGBS data was extracted from the DNA methylation atlas of normal human cell types (Loyfer et al. Nature. (2023) 613:355-364), including methylation data for 5 bladder, 12 colon and 4 prostate tissue samples. WGBS datasets from prostate tumour tissue samples were extracted from Pidsley et al. Clin. Transl. Med. (2022) 12(1):e1030, and accounted for 15 prostate cancer patients’ samples following radical prostatectomy (RP). The RP tissue samples were further classified as originating from patients with lethal (n=7) or non-lethal (n=8) disease (median follow-up 19.5 years). In addition, the WGBS data corresponding to the prostate tissue adjacent to the tumour was collected for a subset of 4 prostate cancer patients.

Data analysis

The methylation beta values were extracted from the publicly available WGBS datasets described above and averaged, for each marker, across all samples for each tissue type considered. For each marker set specific to the detection of a given tissue type, all markers covered by at least three reads were considered for the comparison of their methylation profiles across tissue types. Within each marker set, the markers were stratified by their TCGA-derived status, i.e. either hyperM or hypoM. The hyperM TCGA-derived markers have a higher methylation beta value in their respective target tissue compared to the background. The hypoM TCGA-derived markers have a lower methylation beta value in their respective target tissue compared to the background. The software wgbstools developed by Loyfer et al. (2023) was used to add CpG indexes to .bed files specifying the targeted marker regions (wgbstools convert) as well as to convert WGBS .beta files into tables (wgbstools beta_to_table). The distribution of methylation beta values is represented using boxplots for each tissue type.

Results

The identified tissue methylation marker sets identified in Example 1 (see also Table 1) were demonstrated to be highly specific for individual tissues.

Prostate tumour vs Blood and Prostate Tumour vs All Marker Sets

The expression of “Prostate tumour vs Blood” and “Prostate tumour vs All” markers was analysed in WGBS samples from healthy cfDNA, white blood cell, bladder, colon, and prostate samples (Loyfer et al. 2023), and in WGBS samples from normal prostate, and prostate tumour (lethal tumour and non-lethal tumour samples) (Pidsley et al. 2022), as described in Example 3 Materials and Methods.

As shown in Figure 9A, the “Prostate tumour vs Blood” markers were able to identify all prostate tissues (normal, tumour, lethal tumour, and non-lethal tumour), and distinguish these prostate tissues from healthy cfDNA and white blood cell samples. This marker set was shown not to differentiate between different prostate tumour types. These results show that eligible markers from the “Prostate tumour vs Blood” markers may be useful to denote the absence of DNA derived from prostate tumour or other tissues. For example, there may arise a situation where none of the markers in the “Prostate tumour vs. All” marker set are sufficiently covered (i.e., do not meet a minimum read coverage, such as at least one read or any other minimum number of reads (>0)). Therefore, in this situation, the markers may not be eligible for methylation analysis. However, if in this situation any of the markers in the “Prostate tumour vs Blood” markers are sufficiently covered, then these markers may be used as follows: if none of the eligible markers in the “prostate tumour vs. blood” show any elevated methylation profile compared to a healthy cfDNA baseline (i.e. if there is no detection of any tissue type including bladder, colon, rectum, normal prostate and prostate tumour), these eligible markers (prostate tumour vs. blood) would indicate the absence of detection of prostate tumour (as well as the absence of the other tissue types). As the “Prostate tumour vs Blood” markers are not specific to their target tissue, detection of an elevated methylation profile could be due to the presence of DNA derived from any tissue (or a combination of tissues), and therefore this marker set can be used as a negative control (i.e. showing the absence of detection of prostate tumour in the context referred above).

As shown in Figure 9B, the “Prostate tumour vs AH” markers show strong specificity to the prostate tumour tissues, being able to distinguish prostate tumour, lethal tumour, and non-lethal tumour samples from cfDNA, white blood cell, and normal prostate samples. This marker set was also shown to be able to discriminate between lethal and non-lethal prostate tumours.

As shown in Figure 9C, the “Prostate tumour vs Blood” markers were found to not be specific to the target tissue, i.e. other tissue types were detected using these markers. These other tissues included normal prostate, colon, and bladder samples.

As shown in Figure 9D, the “Prostate tumour vs AH” markers did not detect any of the healthy tissues tested (bladder, cfDNA, colon, normal prostate, or white blood cells). This demonstrates the high specificity of this marker set for prostate tumour tissue.

Bladder vs Blood and Bladder vs All Marker Sets

The expression of “Bladder vs Blood” and “Bladder vs AH” markers was analysed in WGBS samples from healthy cfDNA, white blood cell samples, and WGBS samples from normal prostate, and prostate tumour (lethal tumour and non-lethal tumour samples), as described in Example 3 Materials and Methods.

Similarly to the result seen in Figure 9C with the “Prostate tumour vs Blood” markers, the “Bladder vs Blood” markers were found to not be specific to the target tissue (bladder), as shown in Figures 10A and 10C. As with the “Prostate tumour vs Blood” markers, the “Bladder vs Blood” marker set may be useful to denote the absence of DNA derived from certain tissues.

The “Bladder vs AH” marker set was found to be highly specific for the target tissue (bladder) over other tissue samples. As shown in Figure 10B, the “Bladder vs AH” markers did not detect any prostate tissue when tested in normal prostate, and prostate tumour (lethal and non-lethal tumours) samples. In particular, the HyperM markers indicated a strong absence of detection of prostate DNA. Furthermore, the results shown in Figure 10D demonstrate that the “Bladder vs All” markers specifically detect bladder- derived DNA when tested against cfDNA, white blood cells, colon and healthy prostate samples. Again, the HyperM markers show a greater specificity to bladder tissue.

These results further validate the markers’ specificity to their target tissue.

Colon vs Blood and Colon vs All Marker Sets

The expression of “Colon vs Blood” and “Colon vs AH” markers was analysed in WGBS samples from healthy cfDNA, white blood cell samples, and WGBS samples from normal prostate, and prostate tumour (lethal tumour and non-lethal tumour samples), as described in Example 3 Materials and Methods. Similarly to the results seen in with the “Prostate tumour vs Blood” markers and the “Bladder vs Blood” markers, the “Colon vs Blood” markers were found to not be specific to the target tissue (colon), as shown in Figures 11 A and 11 C. As with the “Prostate tumour vs Blood” and “Bladder vs Blood” markers, the “Colon vs Blood” marker set may be useful to denote the absence of DNA derived from certain tissues.

The “Colon vs All” marker set was found to be highly specific for the target tissue (colon) over other tissue samples. As shown in Figure 11 B, the “Colon vs AH” markers did not detect any prostate tissue when tested against normal prostate, and prostate tumour (lethal and non-lethal tumours) samples. Furthermore, the results shown in Figure 11 D demonstrate that the “Colon vs AH” markers specifically detect colon-derived DNA when tested against cfDNA, white blood cells, bladder and healthy prostate samples. Both HypoM and HyperM markers show high specificity to colon tissue. However, similarly to the results from the “Bladder vs AH” marker set, the HyperM markers from the “Colon vs AH” marker set show a greater specificity to colon tissue.

Rectum vs Blood and Rectum vs All Marker Sets

The expression of “Rectum vs Blood” and “Rectum vs AH” markers was analysed in WGBS samples from healthy cfDNA, white blood cell samples, and WGBS samples from normal prostate, and prostate tumour (lethal tumour and non-lethal tumour samples), as described in Example 3 Materials and Methods.

Similarly to the results seen previously, the “Rectum vs Blood” markers were found to not be specific to the target tissue (rectum), as shown in Figures 12A and 12C. Due to a lack of rectum WGBS tissue samples in the Atlas (Loyfer et al. 2023), colon was used as a proxy target tissue. When the “Rectum vs Blood” marker set was tested against healthy tissues, colon-derived DNA was detected against cfDNA and white blood cell samples. However, other tissues were also detected (both bladder and prostate). As described previously, this marker set may be useful to denote the absence of DNA derived from certain tissues.

The “Rectum vs All” marker set was found to be highly specific for the target tissue (rectum) over other tissue samples. As shown in Figure 12B, the “Rectum vs AH” markers did not detect any prostate tissue when tested against normal prostate, and prostate tumour (lethal and non-lethal tumours) samples. Furthermore, the results shown in Figure 12D demonstrate that the “Rectum vs AH” markers specifically detect colon-derived DNA (colon used as a proxy for rectum target tissue due to the lack of rectum WGBS tissue samples in the Atlas) when tested against cfDNA, white blood cells, bladder, and healthy prostate samples. Both HypoM and HyperM markers show high specificity to colon/rectum tissue.

Conclusions

Taken together, these results demonstrate the utility of the identified marker sets for identifying target tissues of interest. The results also demonstrate the high specificity of the “Target tissue vs AH” marker sets for identifying the target tissue. The “Prostate Tumour vs AH” marker set is highly specific for prostate tumour samples, and able to distinguish DNA derived from prostate tumour samples from all other tissues tested, including healthy prostate samples. Furthermore, the “Prostate Tumour vs AH” marker set is able to distinguish between lethal and non-lethal prostate tumour samples, demonstrating the high sensitivity and specificity of this marker set.

Each of the other “Target tissue vs AH” marker sets also demonstrates extremely high specificity for the target tissue, allowing highly sensitive discrimination between detecting DNA derived from different tissues.

EXAMPLE 4: Testing Marker Sets on Control Tissue Sequenced using Oxford Nanopore Technology Platform

Materials and Methods

Single-molecule sequencing with Oxford Nanopore Technology

Four control tissues (gDNA libraries) were sequenced using Oxford Nanopore sequencing on a PromethlON platform at Edinburgh Genomics (R10, kit 14 was used). Each sequenced library contains a pool of three gDNA samples from the same control tissue, originating from three healthy individuals.

Data Analysis

The four control tissues (gDNA libraries) sequenced with Oxford Nanopore Technology were processed using the basecaller Dorado for canonical and 5-methylcytosine (5mC) base calling (i.e. referred to as methylation calling). All reads with a minimum read mean quality score of 8 were considered for methylation data analysis. All reads were aligned to the reference human genome GRCh38. For each marker region, the read-level methylation values (referred to as alpha values) were calculated for all reads aligned to the marker region and were based on the fraction of the reads overlapping the marker region. For a given read, the read-level methylation value corresponds to the average methylation across all CpGs within the read that overlap a given marker region. Alpha values were grouped into 11 bins corresponding to distinct ranges of read-level methylation values with the following methylation percentages in each bin: bin1 [0%-5%[, bin2 [5%-15%[, bin3 [15%-25%[, bin4 [25%-35%[, bin5 [35%- 45%[, bin6 [45%-55%[, bin7 [55%-65%[, bin8 [65%-75%[, bin9 [75%-85%[, bin10 [85%-95%[, bin11 [95%-100%], where each interval includes the first value and excludes the second, except for the final interval (bin11) where both values are included. The alpha values (x) associated with reads aligned to hypoM markers were converted to 1-x, so that all TCGA-derived markers indicate the detection of their respective target tissue with an elevated methylation profile compared to the background. Each read is assigned to an alpha value bin based on its read-level methylation value. The total number of reads in each alpha value bin is normalised by the total number of aligned reads across all marker regions. The distribution of normalised read counts per alpha value bin is represented using histograms, for each of the marker set specific to the detection of DNA derived from prostate tumour (prtum_vsAII; figures 13A-13D), bladder (bladder_vs_all; figures 16A-16D), colon (colon_vs_all; figures 14A-14D), and rectum (rectum_vs_all; figures 15A-15D) against a set of four Nanopore-sequenced control tissue samples as described in Example 4 ‘Materials and Methods’. The detection of a target tissue relating to any of the organs at risk of side effects from radiotherapy (e.g. large bowel) is demonstrated with a significant number of reads associated with read-level methylation profiles ranging from 55% to 100% in the specific target tissues (e.g. colon/rectum, as in Figures 14C, 14D, 15C and 15D). In comparison, the absence of significant detection of reads associated with these methylation profiles (between 55% and 100%) in negative control tissues (e.g. bladder, leukocytes) offers additional evidence supporting the high specificity of the designated marker sets for their respective target tissues. The high specificity of the ‘prtum_vsAH’ marker set in detecting DNA derived from prostate tumour, demonstrated based on fifteen prostate tumour tissues samples (WGBS datasets), is further evidenced by the absence of detection of DNA derived from any of the organs at risk based on the Nanopore-sequenced negative control tissue samples (figures 13A-13D).

Results

Using the improved Oxford Nanopore sequencing platform (R10, kit 14), the inventors demonstrated that over 24 million CpG markers can be detected in cfDNA (Figure 17).

The marker sets were tested against the gDNA libraries of four control tissues (bladder, colon, leukocyte, and rectum) processed using Oxford Nanopore Technologies, and analysed as described in Example 4 Materials and Methods.

Figures 13A to 13D show the results of the “Prostate Tumour vs All” marker set. None of the healthy tissues were detected with this marker set, as demonstrated by a significant absence of reads with readlevel methylation of 55% to 100% (Bin7 to Bini 1).

These results provide further evidence to support the high specificity of this marker set to the prostate tumour tissue.

The “Colon vs All” marker set was shown to be specific for detecting DNA derived from the large bowel. Figure 14C and 14D respectively show that the “Colon vs All” marker set specifically detected colon- derived DNA in the colon control tissue sample, and rectum-derived DNA in the rectum control tissue sample.

The specificity of this marker set to detect DNA derived from large bowel (colon, rectum) is further demonstrated with the absence of a significant detection of DNA derived from negative control tissues (bladder and leukocytes), as shown in Figures 14A and 14B.

Similarly, the “Rectum vs AH” marker set was found to be specific for detecting DNA from the large bowel (colon and rectum). Figures 15D and 15C respectively show that the “Rectum vs AH” market set specifically detected rectum-derived DNA in rectum control tissue, and colon-derived DNA in the colon control tissue. The specificity of this marker set to detect DNA derived from the large bowel (colon, rectum) is further demonstrated with the absence of a significant detection of DNA derived from negative control tissues (bladder and leukocytes), as shown in Figure 15A and 15B.

Figures 16A to 16D show the results of the “Bladder vs All” marker set. As seen in Figure 16A, this marker set detected bladder-derived DNA in bladder control tissue. This marker set also detects colon- and rectum-derived DNA.

These results show that the “Bladder vs All” marker set is not specific to the bladder, although the detection signal is strongest in the bladder tissue. Interestingly, specific ranges of read-level methylation values (in particular Bin7 & Bi n 10) show unique signal to the bladder tissue, compared to colon, leukocyte, or rectum samples.

Conclusions

These results again demonstrate the high specificity of the “Target tissue vs AH” marker sets in discriminating between DNA tissue of origin. The “Rectum vs AH” and “Colon vs AH” marker sets have been demonstrated to be highly specific for large bowel tissues. The “Prostate Tumour vs AH” marker sets have also been demonstrated to show extremely high specificity for prostate tumour tissue.

The “Bladder vs AH” marker set, particularly the hyperM markers, discriminate DNA derived from bladder tissue against other tissues and healthy cfDNA, as demonstrated with WGBS data. As indicated in Figures 16A-16D, the “Bladder vs AH” marker set shows a stronger detection of DNA derived from bladder tissue compared to other tissues. A subset of markers from the “Bladder vs AH” marker set might demonstrate specificity to the target tissue in ONT data when refining the markers in association with specific alpha value bins. Indeed, some reads show unique read-level methylation profiles (i.e. , bins 7 and 10) that are exclusively present in the bladder control tissue. These unique read-level methylation profiles might be sufficient to confirm the presence of DNA derived from bladder.

Taken together, these results further demonstrate that the marker sets identified in Example 1 can be used to identify cfDNA derived from tissues relating to different organs, enabling the detection of nontumour tissue toxicity following chemotherapy.

The column “medDiff for each CpG and for given pairwise tissue comparison (tissue1_vs_tissue 2) indicates the marker’s directionality (medDiff > 0 = hypermethylated CpG in tissue 1 versus tissue 2) (medDiff < 0 = hypomethylated CpG in tissue 1 versus tissue 2)

“AH” = all other tissues included in the overall set of comparisons; for example, “colon_vsAH” = colon versus [blood, bladder, small intestine, rectum, prostate normal and prostate tumour]. bladder_vsAII — bladder vs All smallint_vsAII — small intestine vs All colon_vsAII — colon vs All prtum_vsAII — prostate tumour vs All prostall_vsAII — prostate tumour + normal vs All rectum_vsAII — rectum vs All prtum_vs_prostate — prostate tumour vs prostate normal blad_vs_Rect — bladder vs rectum blad_vs_Colon — bladder vs colon blad_vs_Smalllnt — bladder vs small intestine blad_vs_ProstAII — bladder vs prostate tumour + normal rect_vs_ProstAII — rectum vs prostate tumour + normal colon_vs_ProstAII — colon vs prostate tumour + normal smallint_vs_ProstAII — small intestine vs prostate tumour + normal Rectum_vs_Colon — rectum vs colon

Colon_vs_Smalllnt — colon vs small intestine Rectum_vs_Smalllnt — rectum vs small intestine prostate_vs_blood — prostate normal vs blood. prTum_vs_blood — prostate tumour vs blood bladder_vs_blood — bladder vs blood Colon_vs_blood — colon vs blood rectum_vs_blood — rectum vs blood smalllnt_vs_blood — small intestine vs blood

“Position_hg 19” refers to the location of the Cytosine in the CpG dinucleotide on the Watson (forward) strand. For a CpG located on a “+” (forward strand): position indicates the location of the Cytosine in the CpG on the forward strand. For a CpG located on a (reverse strand) position indicates the location of the Guanine in the CpG located on the reverse strand, i.e., the location of the corresponding Cytosine on the Watson (forward) strand.

All positions refer to hg19 (GRCh37) genome build, using 1 -based coordinate format.

Table 2: Literature-derived CpGs extracted from scientific publication (n = 661)

References

A number of publications are cited above in order to more fully describe and disclose the invention and the state of the art to which the invention pertains. Full citations for these references are provided below. The entirety of each of these references is incorporated herein.

1 . Heitzer E, Haque IS, Roberts CES, Speicher MR. Current and future perspectives of liquid biopsies in genomics-driven oncology. Nat Rev Genet. 2019;20(2):71-88.

2. Cescon DW, Bratman SV, Chan SM, Siu LL. Circulating tumor DNA and liquid biopsy in oncology. Nat Cancer. 2020;1 (3):276-90.

3. Ignatiadis M, Sledge GW, Jeffrey SS. Liquid biopsy enters the clinic - implementation issues and future challenges. Nat Rev Clin Oncol. 2021 ;18(5):297-312.

4. Tie J, Cohen JD, Lahouel K, Lo SN, Wang Y, Kosmider S, et al. Circulating Tumor DNA Analysis Guiding Adjuvant Therapy in Stage II Colon Cancer. N Engl J Med. 2022;386(24):2261-72.

5. De Ruysscher D, Niedermann G, Burnet NG, Siva S, Lee AWM, Hegi-Johnson F. Radiotherapy toxicity. Nat Rev Dis Primers. 2019;5(1):13.

6. Rostami A, Lambie M, Yu CW, Stambolic V, Waldron JN, Bratman SV. Senescence, Necrosis, and Apoptosis Govern Circulating Cell-free DNA Release Kinetics. Cell Rep. 2020;31 (13):107830.

7. Lo YM, Leung SF, Chan LY, Chan AT, Lo KW, Johnson PJ, et al. Kinetics of plasma Epstein-Barr virus DNA during radiation therapy for nasopharyngeal carcinoma. Cancer Res. 2000;60(9):2351-5.

8. Muhanna N, Eu D, Chan HHL, Douglas C, Townson JL, Di Grappa MA, et al. Cell-free DNA and circulating tumor cell kinetics in a pre-clinical head and neck Cancer model undergoing radiation therapy. BMC Cancer. 2021 ;21 (1):1075.

9. De Michino S, Aparnathi M, Rostami A, Lok BH, Bratman SV. The Utility of Liquid Biopsies in Radiation Oncology. Int J Radiat Oncol Biol Phys. 2020;107(5):873-86.

10. Phallen J, Sausen M, Adleff V, Leal A, Hruban C, White J, et al. Direct detection of early-stage cancers using circulating tumor DNA. Sci Transl Med. 2017;9(403). 11. Newman AM, Bratman SV, To J, Wynne JF, Eclov NC, Modlin LA, et al. An ultrasensitive method for quantitating circulating tumor DNA with broad patient coverage. Nat Med. 2014;20(5):548-54.

12. Newman AM, Lovejoy AF, Klass DM, Kurtz DM, Chabon JJ, Scherer F, et al. Integrated digital error suppression for improved detection of circulating tumor DNA. Nat Biotechnol. 2016;34(5):547-55.

13. Sanz-Garcia E, Zhao E, Bratman SV, Siu LL. Monitoring and adapting cancer treatment using circulating tumor DNA kinetics: Current research, opportunities, and challenges. Sci Adv. 2022;8(4):eabi8618.

14. Cohen JD, Li L, Wang Y, Thoburn C, Afsari B, Danilova L, et al. Detection and localization of surgically resectable cancers with a multi-analyte blood test. Science. 2018;359(6378):926-30.

15. Lehmann-Werman R, Neiman D, Zemmour H, Moss J, Magenheim J, Vaknin-Dembinsky A, et al. Identification of tissue-specific cell death using methylation patterns of circulating DNA. Proc Natl Acad Sci U S A. 2016;113(13):E1826-34.

16. Moss J, Magenheim J, Neiman D, Zemmour H, Loyfer N, Korach A, et al. Comprehensive human cell-type methylation atlas reveals origins of circulating cell-free DNA in health and disease. Nat Commun. 2018;9(1):5068.

17. Loyfer N, Magenheim J, Peretz A, Cann G, Bredno J, Klochendler A, et al. A human DNA methylation atlas reveals principles of cell type-specific methylation and identifies thousands of cell typespecific regulatory elements. bioRxiv. 2022;https://doi.org/10.1101/2022.01 .24.477547.

18. Leung SF, Chan KC, Ma BB, Hui EP, Mo F, Chow KC, et al. Plasma Epstein-Barr viral DNA load at midpoint of radiotherapy course predicts outcome in advanced-stage nasopharyngeal carcinoma. Ann Oncol. 2014;25(6):1204-8.

19. Chera BS, Kumar S, Beaty BT, Marron D, Jefferys S, Green R, et al. Rapid Clearance Profile of Plasma Circulating Tumor HPV Type 16 DNA during Chemoradiotherapy Correlates with Disease Control in HPV-Associated Oropharyngeal Cancer. Clin Cancer Res. 2019;25(15):4682-90.

20. Lv J, Chen Y, Zhou G, Qi Z, Tan KRL, Wang H, et al. Liquid biopsy tracking during sequential chemo-radiotherapy identifies distinct prognostic phenotypes in nasopharyngeal carcinoma. Nat Commun. 2019;10(1):3941 .

21. Warlow SJ, Adamowicz M, Thomson JP, Wescott RA, Robert C, Carey LM, et al. Longitudinal measurement of HPV copy number in cell-free DNA is associated with patient outcomes in HPV-positive oropharyngeal cancer. Eur J Surg Oncol. 2022;48(6):1224-34.

22. Citrin DE. Recent Developments in Radiotherapy. N Engl J Med. 2017;377(22):2200-1 .

23. Keener A. Precision Oncology: Rethink, Aim and Fire. Nature. 2020;585:S10-S2.

24. Scott JG, Berglund A, Schell MJ, Mihaylov I, Fulp WJ, Yue B, et al. A genome-based model for adjusting radiotherapy dose (GARD): a retrospective, cohort-based study. Lancet Oncol. 2017;18(2):202- 11.

25. Kerns SL, Fachal L, Dorling L, Barnett GC, Baran A, Peterson DR, et al. Radiogenomics Consortium Genome-Wide Association Study Meta-Analysis of Late Toxicity After Prostate Cancer Radiotherapy. J Natl Cancer Inst. 2020;112(2):179-90.

26. Schack LMH, Naderi E, Fachal L, Dorling L, Luccarini C, Dunning AM, et al. A genome-wide association study of radiotherapy induced toxicity in head and neck cancer patients identifies a susceptibility locus associated with mucositis. Br J Cancer. 2022;126(7):1082-90. 27. Christensen E, Birkenkamp-Demtroder K, Sethi H, Shchegrova S, Salari R, Nordentoft I, et al. Early Detection of Metastatic Relapse and Monitoring of Therapeutic Efficacy by Ultra-Deep Sequencing of Plasma Cell-Free DNA in Patients With Urothelial Bladder Carcinoma. J Clin Oncol. 2019;37(18):1547- 57.

28. Garcia-Murillas I, Schiavon G, Weigelt B, Ng C, Hrebien S, Cutts RJ, et al. Mutation tracking in circulating tumor DNA predicts relapse in early breast cancer. Sci Transl Med. 2015;7(302):302ra133.

29. Kurtz DM, Scherer F, Jin MC, Soo J, Craig AFM, Esfahani MS, et al. Circulating Tumor DNA Measurements As Early Outcome Predictors in Diffuse Large B-Cell Lymphoma. J Clin Oncol. 2018;36(28):2845-53.

30. Tie J, Wang Y, Tomasetti C, Li L, Springer S, Kinde I, et al. Circulating tumor DNA analysis detects minimal residual disease and predicts recurrence in patients with stage II colon cancer. Sci Transl Med. 2016;8(346):346ra92.

31. Dasari A, Morris VK, Allegra CJ, Atreya C, Benson AB, 3rd, Boland P, et al. ctDNA applications and integration in colorectal cancer: an NCI Colon and Rectal-Anal Task Forces whitepaper. Nat Rev Clin Oncol. 2020;17(12):757-70.

32. Barefoot ME, Loyfer N, Kiliti AJ, Schmidt MO, Shabi-Porat S, Jain S, et al. Cell-free, methylated DNA in blood samples reveals tissue-specific, cellular damage from radiation treatment. bioRxiv. 2022;https://doi.org/10.1101/2022.04.12.487966.

33. Blomain ES, Moding EJ. Liquid Biopsies for Molecular Biology-Based Radiotherapy. Int J Mol Sci. 2021 ;22(20):DOI: 10.3390/ijms222011267.

34. Choy LYL, Peng W, Jiang P, Cheng SH, Yu SCY, Shang H, et al. Single-molecule Sequencing Enables Long Cell-free DNA Detection and Direct Methylation Analysis for Cancer Patients. Clin Chem. 2022.

35. Katsman E, Orlanski S, Martignano F, Fox-Fisher I, Shemer R, Dor Y, et al. Detecting cell-of- origin and cancer-specific methylation features of cell-free DNA from Nanopore sequencing. Genome Biol. 2022;23(1):158.

36. Lau BT, Almeda A, Schauer M, McNamara M, Bai X, Meng Q, et al. Single molecule methylation profiles of cell-free DNA in cancer with nanopore sequencing. bioRxiv. 2022;doi 10.1101/2022.06.22.497080.

37. de Proce et al. Development of methylation-based biomarkers for breast cancer detection by model training and validation and synthetic cell-free DNA (2002); bioRxiv doi 10.1101/2022.02.11.480085.

For standard molecular biology techniques, see Sambrook, J., Russel, D.W. Molecular Cloning, A

Laboratory Manual. 3 ed. 2001 , Cold Spring Harbor, New York: Cold Spring Harbor Laboratory Press

Claims

Claims:

1 . A method of monitoring radiotherapy response in a subject having prostate cancer, the method comprising: a. Providing a cell-free DNA (cfDNA)-containing sample obtained from the subject; b. Performing DNA sequencing on the cfDNA-containing sample in order to generate methylation sequencing reads; c. Assessing methylation levels within the cfDNA at each of a plurality of pre-determined differentially methylated regions (DMRs) to generate a sample methylation profile, wherein each pre-determined DMR is associated with a tissue of origin, thereby detecting the presence of one or more tissues of origin; and d. Determining the radiotherapy response of the subject based on the sample methylation profile.

2. The method of claim 1 , wherein the radiotherapy response comprises prostate tumour response to radiotherapy.

3. The method of claim 2, wherein the one or more tissues of origin comprises prostate tumour tissue.

4. The method of any one of claims 1 to 3, wherein the radiotherapy response comprises non-tumour toxicity to radiotherapy.

5. The method of claim 4, wherein the one or more tissues of origin comprises adjacent non-tumour tissue.

6. The method of claim 4 or 5, wherein the one or more tissues of origin comprises bladder tissue, small intestine tissue, colon tissue, rectum tissue and/or normal prostate tissue.

7. The method of any one of the preceding claims, wherein determining the radiotherapy response of the subject in step d comprises comparing the sample methylation profile to a control methylation profile, wherein a significant difference in methylation level between the sample methylation profile and the control methylation profile is indicative of a response in the subject.

8. The method of claim 7, wherein the control methylation profile is generated from a first cfDNA- containing sample obtained from the subject prior to receiving radiotherapy and the sample methylation profile is generated from a second cfDNA-containing sample obtained from the subject following at least one dose of radiotherapy.

9. A method of treating a subject with prostate cancer using adaptive radiotherapy, the method comprising administering at least one dose of radiotherapy to the subject, further comprising the steps of any of claims 1 to 8, further comprising adjusting the dose of radiotherapy, altering the radiotherapy field and/or the frequency of radiotherapy.

10. The method according to claim 9, wherein adjusting the dose of radiotherapy comprises increasing the dose of radiotherapy, decreasing the dose of radiotherapy, or discontinuing radiotherapy.

11. The method of claim 9, wherein adjusting the frequency of radiotherapy comprises increasing the length of time between doses or decreasing the length of time between doses.

12. The method of any one of the preceding claims, wherein the DNA sequencing comprises singlemolecule DNA sequencing.

13. The method of claim 12, wherein the single-molecule sequencing comprises nanopore sequencing.

14. The method of any one of claims 1 to 13, wherein the sample obtained from the subject is a blood sample.

15. The method of any one of claims 1 to 14, wherein the plurality of pre-determined DMRs is selected from the CpG sites of Table 1 or Table 2.

16. The method of any one of claims 1 to 15, wherein the plurality of pre-determined DMRs comprises at least 5, at least 10, at least 30, at least 50, or at least 100 DMRs.

17. The method of claim 16, wherein the plurality of pre-determined DMRs comprises at least 1000, at least 10,000, at least 100,000 or at least 1 x 10⁶ DMRs.