lynch_syndrome

Gene-Specific Variation in Colorectal Cancer Surveillance Strategies for Lynch Syndrome

Input (Data folder)

• model_inputs.xlsx: contains several sheets with model inputs
  • MLH1, MSH2, MSH6, and PMS2: cumulative risk of colon cancer by age for each gene
  • Sporadic and Survival are not used ?????
  • CRC Survival: probability of cancer mortality by stage; used in get_cancer_death_prob() of probability_functions.py
  • Stage Dists: probabilities of each cancer stage at diagnosis, by screening interval; used to set normal/adenoma -> cancer probabilities in create_t_matrix() of lynch_simulator.py

Code

lynch_presets.py

• genes: list of genes to be analyzed (MLH1, MSH2, MSH6, and PMS2)
• csy_protocols: colonoscopy protocols
  • nono: natural history (no colonoscopy screening)
  • current: current screening recommendation (annual screening starting at age 25)
  • new: new screening strategies (varying start age in 5-year increments between 25 and 40 years, and surveillance intervals of between 1 and 5 years)
• ALL_STATES: dictionary containing all states of the model
  • mutation: everyone starts here (with one of the genetic mutations)
  • nono: healthy state for running the natural history model
  • current: healthy state for running the current recommended screening schedule
  • new: healthy state for running the new screening schedules
• CONNECTIVITY: dictionary indicating possible transitions between states
  • No transitions between cancer stages
  • The adenoma state can be skipped because it's not always detected before cancer development
• risk_ratios: ?????????????????? risk of what? based on screening inverval?
• run_type: class containing information about a run
  • Input: screening interval, gene, gender, start age (default 25)
  • risk_ratio: based on screening interval
  • interval: screening interval (1-5 years)
  • gene
  • gender
  • start_age: defaults to 25
  • guidelines: one of the csy_protocols described above (nono, current, or new)
• risk_adn_male_data:

lynch_simulatory.py

create_t_matrix() function:

• Input: instance of run_type class (run_spec), time (defaults to ps.time), start age (defaults to ps.START_AGE)
• this_gender: contains parameters specific to gender; instance of gender_obj class of gender.py
• names: flipped version of the ALL_STATES dictionary (state names are dictionary keys)
• Set natural history risk of colon cancer by age using cumul_prob_to_annual() of probability_functions.py (nodes_nono, risk_probs_nono, risk_rates_nono)
• If not a natural history run:
  • Set risk of colon cancer by age as above, but using a risk ratio dependent on screening interval (nodes, risk_probs, risk_rates)
  • Set

Output