Max Diehn: Liquid Biopsies and Cancer Detection

Max’s Background and Research

• Max and Peter Atia were classmates in medical school at Stanford
• Max pursued an MD-​​PhD program, splitting his time between medical school and lab research
• Joined Pat Brown’s lab at Stanford, focusing on DNA microarrays 
  • Revolutionary technology at the time, allowing measurement of tens of thousands of genes in one experiment
• Max’s dissertation involved multiple projects, mainly in immunology and oncology 
  • One project focused on T cells and their activation, cataloging genes turned on or off during activation
  • Another project involved isolating RNA stuck to the endoplasmic reticulum inside cells, which was challenging due to RNA’s instability
• Max completed his PhD in about three years and returned to clinical rotations 
  • Was set on doing a residency but unsure in which field
  • Decided to focus on oncology after his father was diagnosed with lymphoma during Max’s undergraduate years
  • Realized there were multiple options within cancer specialties, such as medical oncology, surgical oncology, and radiation oncology
    Born in Germany, Moved to the US
• Born in Munich, Germany 
• Moved to the US at 11 years old
• Grew up on the east coast

Medical School and Residency

• Went to Harvard as an undergrad
• Considered medical oncology, radiation oncology, and surgical oncology
• Chose radiation oncology due to: 
  • More time with patients in clinic
  • Interest in technology aspects
  • Opportunity to make a difference in a field with less molecular-​​level research

Radiation Oncology at Stanford

• One of the first departments of radiation oncology in the US
• Established by Henry Kaplan, who cured Hodgkin’s disease with radiotherapy
• Strong interest in laboratory-​​based research

Interest in Liquid Biopsies

• Did not initially focus on liquid biopsies
• Followed experimental results to the field
• Started research projects based on clinical needs
• Frustration with inability to predict lung cancer recurrence after treatment led to interest in liquid biopsies
  Liquid Biopsy for Cancer Detection
• Traditional imaging methods have limitations in detecting cancer 
  • Hard to see tumors smaller than 1 cm in diameter
  • About a billion cells are undetectable
  • Micrometastatic disease can have millions of cells and still be undetectable
• Protein biomarkers (PSA, CEA, CA 19–9) have been used historically 
  • Shed into the blood by cancer cells
  • Lack specificity, as normal cells can also produce these proteins
• Sensitivity and specificity are important factors in cancer detection 
  • Sensitivity: true positive rate, likelihood of a positive test when the patient has the condition
  • Specificity: true negative rate, likelihood of a negative test when the patient does not have the condition
  • No perfect tests, as increasing sensitivity often decreases specificity
• Liquid biopsy aims to improve cancer detection through blood tests 
  • Motivated by the limitations of imaging methods and protein biomarkers
  • Higher sensitivity and specificity could lead to earlier detection and better treatment outcomes
    Sensitivity and Specificity in Diagnostic Tests
• Sensitivity: the ability of a test to correctly identify patients with a disease 
• Specificity: the ability of a test to correctly identify patients without a disease
• Balancing sensitivity and specificity is crucial for a good diagnostic test
• Pushing one parameter (sensitivity or specificity) often compromises the other

Lung Cancer Overview

• Number one cause of cancer death
• Incidence and mortality rates are decreasing due to reduced smoking rates and improved treatments
• Smoking is the largest risk factor, but not the only one (e.g., pollution, radon gas, genetic factors)

Types of Lung Cancer

• Small cell and non-​​small cell lung cancer
• Non-​​small cell lung cancer includes adenocarcinoma (most common) and squamous cell carcinoma
• Non-​​smokers usually develop adenocarcinoma

Environmental Factors and Lung Cancer

• Particulate matter (PM 2.5) exposure is associated with increased lung cancer risk
• Radon gas exposure is another environmental risk factor
• Secondhand smoke exposure is difficult to quantify but has been linked to increased lung cancer risk in certain professions (e.g., waitresses, flight attendants)
  Lung Cancer Screening and Low Dose CT
• No biomarker to measure exposure for lung cancer screening eligibility 
  • Secondhand smoke not considered
• Screening criteria aims to enrich for highest risk population 
  • At least 0.5–1% risk of developing cancer to be eligible for screening
• Low dose CT has been a major change in lung cancer management in the last 10 years 
  • National Lung Screening Trial showed low dose CT scans significantly reduced lung cancer deaths 
    • Relative risk reduction of about 20%
    • Absolute risk reduction in single digit percent
• Low dose CT scans use much less radiation than traditional CT scans 
  • Reduces risk of causing cancer from radiation exposure
• Medical practitioners weigh the benefits and risks of imaging tests 
  • Should not order imaging if it won’t change patient management

Liquid Biopsy for Early Detection of Lung Cancer Recurrence

• Frustration with not being able to diagnose recurrence earlier led to exploration of liquid biopsy
• Two approaches to developing a biomarker for early detection: 
  • Preclinical models using mice, testing hypothesis, and then applying to humans
  • Translational research using human blood samples directly
• Initial funding for the research was provided by the researcher themselves * Startup funds for new faculty members 
  • Used to kickstart research before obtaining grants
• Difficulty in reproducing results in research 
  • Often need large studies to prove something doesn’t work as well as initially reported
  • Limited research resources and time
  • Bias towards positive findings in publication realm
• Importance of reproducing positive results 
  • Ensures robust findings and convincing evidence
• Protein biomarkers and circulating tumor cells (CTCs) in lung cancer research 
  • Protein biomarkers not found to be unique to lung cancer cells
  • CTCs difficult to measure due to low abundance and need for immediate processing
  • Issues with specificity in CTC methods
• Blood sample size for CTC research 
  • Usually 10–20 mL of blood (a few tablespoons)
    Circulating Tumor Cells (CTCs) and Cell-​​Free DNA
• CTCs: cells that have broken away from a primary tumor and entered the bloodstream 
  • Sensitivity is not very good for detecting early-​​stage cancer
  • High levels of CTCs can be a negative prognostic marker, indicating higher risk of recurrence
  • Healthy patients can also have cells that look like CTCs, complicating detection
• Cell-​​free DNA: DNA molecules found in the circulation, outside of cells 
  • Found in the blood plasma
  • Can be used to detect fetal DNA in pregnant mothers
  • Double-​​stranded DNA, about 170 base pairs in length
  • Wrapped around core histones, which protect the DNA from enzymes that break it down

Potential Applications and Limitations

• CTCs are not suitable for cancer screening or predicting disease recurrence in resected patients 
  • May help determine the course of adjuvant therapy in some patients
• Cell-​​free DNA has potential for detecting cancer, but more research and development is needed 
  • Challenges include extracting DNA from cellular compartments and dealing with low levels of cell-​​free DNA in the plasma
  • Patients with certain blood conditions, like beta thalassemia minor, may have reduced plasma volume, making it more difficult to obtain enough cell-​​free DNA for analysis
    Cell-​​Free DNA and Cancer Detection
• Cell-​​free DNA (cfDNA) is found in the blood and is released from cells during cell death 
  • Most cfDNA comes from healthy cells, but a small subset comes from cancer cells
• The amount of cfDNA in the blood is usually low due to enzymes constantly breaking it down 
  • In certain conditions (e.g., advanced cancer, trauma, infection), cfDNA levels can be much higher
• Apoptosis (programmed cell death) is one possible source of cfDNA 
  • Cells chop up their DNA during apoptosis, making it difficult to determine the exact source of cfDNA
• To detect cancer using cfDNA, researchers look for mutations in the DNA that are specific to cancer cells 
  • These mutations serve as markers for the presence of cancer
• Next-​​generation sequencing is used to identify the sequence of DNA bases in millions of molecules 
  • This allows researchers to compare the sequences to the patient’s healthy DNA and look for mutations
• Sequencing the tumor itself can help identify specific mutations to look for in the blood 
  • This method is highly specific and sensitive for detecting the presence of cancer
    Cell-​​Free DNA and Cancer Detection
• Cell-​​free DNA (cfDNA) can be used to detect cancer cells in the body 
• cfDNA is short fragments of DNA (around 170 bases) found in the blood
• 99.9% of cfDNA lines up with germline DNA, but no two segments are the same
• Cancer cells have specific mutations that can be detected in cfDNA

Challenges in Detecting Cancer Mutations

• Only a small percentage of cfDNA comes from cancer cells
• Cancer cells may have only a few mutations in coding regions
• It’s possible to miss cancer mutations in cfDNA due to their rarity

Improving Cancer Detection

• Using next-​​generation sequencing technology to look for multiple mutations at once
• Increases sensitivity by looking for dozens of mutations in parallel
• Does not matter if mutations are coding or non-​​coding, as long as they are present in all cancer cells

Cell-​​Free RNA and Cancer Detection

• Cell-​​free RNA exists but is less stable than cfDNA
• Can be used to measure gene expression in cancer cells
• Complementary to cfDNA, providing additional information about the cancer

Differences Between Cell-​​Free DNA and Circulating Tumor DNA

• Cell-​​free DNA refers to all DNA in circulation, including healthy and cancer cell-​​derived DNA
• Circulating tumor DNA refers to the fraction of cfDNA that comes from cancer cells

Using Methylation Patterns for Cancer Detection

• Methylation patterns on DNA can be informative about the origin of the DNA
• Different cells have different methylation patterns based on their tissue of origin
• Methylation profiles can be used to identify tissue of origin and potentially screen for cancer
  Sensitivity and Practicality in Cancer Detection
• Mutation-​​based methods are more sensitive than methylation-​​based methods 
  • New method: 100 times more sensitive, can detect 1 in a million
  • Methylation-​​based assays: sensitivity around 0.1% (1 in 1000)
• Grail: pan-​​screen test using methylation patterns of cell-​​free DNA 
  • Sensitivity for all stages: 50% with 99% specificity
  • Sensitivity for stage one: 20%
• Importance of breaking down sensitivity by stage 
  • Stage one lung cancer sensitivity in Grail data: 5% or less
• High specificity and sensitivity may not significantly change pretest probability 
  • Example: 50% sensitivity, 99% specificity, 1% prevalence 
    • Negative predictive value: 99.5%
    • Positive predictive value: 40%
• Need for studies proving cancer-​​specific survival benefits of tests 
  • Currently not on the roadmap due to expense and time
• Stage one lung cancer heterogeneity 
  • Type 1: cancer cells only in the lung, can be cured by surgery
  • Type 2: microscopic cells in other areas, not curable by surgery
• Screening tests must catch a significant fraction of Type 1 stage one cancers to be useful
  ctDNA Assays and Cancer Detection
• ctDNA assays can detect microscopic deposits of cancer cells spread throughout the body 
• Randomized studies needed to prove the effectiveness of ctDNA assays in detecting early-​​stage cancer
• Ethical concerns about withholding potentially life-​​saving tests from patients
• FDA-​​approved liquid biopsies (e.g., Garden) used for identifying mutations in patients with advanced disease 
  • Helps determine appropriate treatments based on specific mutations
• Tests like Grail not yet FDA-​​approved but permitted for use in CLIA-​​compliant labs 
  • Less regulated, allowing for quicker availability to patients and providers

Future of Cancer Detection and Treatment

• The “Holy Grail” of cancer detection: blood tests with high sensitivity and specificity for early-​​stage cancer 
  • Could lead to earlier detection and treatment, improving survival rates
• Current ctDNA assays can detect minimal residual disease (microscopic cells remaining after treatment) 
  • High positive predictive value for cancer recurrence
• Clinical trials underway to use ctDNA assays to guide adjuvant therapy for early-​​stage cancer patients 
  • Aim to improve outcomes by targeting residual cancer cells before they become clinically detectable
    Cancer Detection and Treatment
• Importance of early detection in cancer treatment 
  • Less tumor burden, less heterogeneity, better outcomes
• Liquid biopsies and ctDNA (circulating tumor DNA) tests 
  • Can help identify high-​​risk patients who need adjuvant therapy
  • Useful in cancers where a subset of patients develop recurrence
  • Active studies in colorectal, breast, and lung cancer
  • Challenges in prostate and breast cancer due to low levels of ctDNA
• Potential future uses of liquid biopsies 
  • Repeated testing to catch recurrence early and only give adjuvant therapy when needed
  • Avoid over-​​treatment by only treating patients with evidence of remaining cancer cells
  • Annual cancer screening for early detection and treatment

Lung Cancer Screening with ctDNA

• Lung CLiP (Cancer Likelihood in Plasma) method 
  • Sequences plasma and white blood cell DNA
  • Subtracts mutations found in white blood cells to focus on cancer-​​derived mutations
  • Uses machine learning to analyze remaining mutations and other factors (e.g., cell-​​free DNA molecule length)
  • Aims to identify lung cancer early for more effective treatment
    Dye in Tissue and Blood Exposure
• Dye in tissue takes time to get into the blood 
• Exposure to enzymes may be a problem
• Histones and chromatin configuration may contribute

Machine Learning Model for Mutation Detection

• Model considers:
• Presence of mutation
• Gene mutation is in
• Cell-​​free DNA fragment length
• Whether mutation is caused by smoking
• Model outputs probability of blood sample being from a patient with lung cancer

Combining Mutation Detection with Methylation

• Possibility of combining mutation detection with methylation for better results
• Combination of methods still in early research phase
• Need to develop an affordable assay for large-​​scale use

Theoretical Limits of Sensitivity and Specificity

• Ideal sensitivity and specificity for early-​​stage cancer detection:
• 80% sensitivity
• 99.5% specificity
• Current tests have lower sensitivity and specificity
• Goal is to improve liquid biopsy tests to reach these levels

Challenges in Developing Liquid Biopsy Tests

• Need to prove that tests decrease cancer-​​specific death
• Need to compare tests to existing screening methods
• Practical considerations, such as access to imaging facilities and radiation exposure concerns

Efforts to Improve Liquid Biopsy Tests

• Companies and researchers working to improve tests
• Easier to change tests in CLIA environment than under FDA approval
• No clear solution yet, but promising ideas being explored

NIH Funding for Liquid Biopsy Research

• Amount of funding for liquid biopsy research has increased dramatically
• Shift in focus from circulating tumor cells to liquid biopsy in recent years
  NIH Study Section and Liquid Biopsy Research
• NIH study sections involve scoring grants for research funding 
  • Requires reading and commenting on dozens of grants
  • Serves as a crucial part of the research funding system
• Cancer biomarker-​​focused study sections see a significant increase in liquid biopsy work since 2012 
  • High interest from NCI and NIH
  • Recognized value in pushing forward liquid biopsy research
• Interviewee, Max, provided valuable insights into the landscape and history of liquid biopsy research 
  • Both interviewer and listeners likely gained new knowledge from the discussion