12-phase flow (12)

1. 1FRAME
   Eligibility gate (the prior selector): asymptomatic adult age 50-80, ≥20 pack-yr smoking history, currently smoking OR quit <15 yr, with sufficient life expectancy AND willingness/ability for curative-intent lung surgery. Below any threshold the false-positive cascade + overdiagnosis harm exceeds expected mortality benefit → STOP-screening. Excludes symptomatic patients (diagnosis, not screening) and prior-lung-cancer survivorship surveillance (USPSTF 2021 Grade B PMID 33687470)
   inputs: age, smoking_history_pack_years, smoking_quit_interval_years, life_expectancy_estimate, curative_surgery_candidacy
   advance: Asymptomatic, USPSTF-eligible (or PLCOm2012-expanded), adequate life expectancy + curative-surgery candidacy confirmed
2. 2ENTRY
   Trigger: routine wellness visit in an eligible smoker, overdue/never-screened gap closure, PLCOm2012 risk-model expansion, Lung-RADS-driven recall, or former-smoker quit-interval re-check (USPSTF 2021)
   advance: Entry trigger captured
3. 3CONTEXT
   Establish/refine the pre-test prevalence prior: pack-years, quit interval, age, COPD/emphysema, family hx lung cancer, occupational/radon exposure, prior LDCT date + Lung-RADS category, pregnancy, prior-cancer status. This phase SETS the Bayesian prior — the USPSTF-vs-PLCOm2012 eligibility decision flows from it (USPSTF 2021; Tammemägi NEJM 2013 PMID 23425165)
   inputs: prior_ldct_date_and_lung_rads, copd_or_emphysema, family_hx_lung_cancer, occupational_radon_asbestos_exposure, pregnancy_status, prior_lung_cancer_survivor
   advance: Pre-test prevalence tier assignable (categorical USPSTF or continuous PLCOm2012)
4. 4RED_FLAGS
   Symptomatic patient is NOT a screening candidate: hemoptysis, unexplained weight loss, or a known/suspected mass means the pre-test probability is already diagnostic — exit the screening engine and route to the diagnostic pathway onc.lung-cancer.core.v1 (USPSTF 2021 — screening applies only to the asymptomatic)
   inputs: pulmonary_red_flag_symptoms
   advance: Red-flag symptoms excluded (asymptomatic confirmed) OR routed to diagnostic engine
5. 5INITIAL_WORKUP
   Order annual LDCT — the test with its fixed characteristics (NLST initial-round sensitivity 93.8% / specificity 73.4%, Church NEJM 2013 PMID 23697514); the order IS the "intervention". Document shared decision-making AND tobacco cessation counseling (CMS coverage requirement) (USPSTF 2021 PMID 33687470; NLST PMID 21714641)
   inputs: smoking_history_pack_years
   actions: workup.lung_screening_ldct
   advance: LDCT ordered with shared decision + tobacco counseling documented
6. 6BRANCHING_WORKUP
   Positive-screen Bayesian follow-up — the Lung-RADS v2022 category IS the likelihood ratio applied to the prior: 1/2 (<1%) → return to annual; 3 (~1-2%) → 6-month LDCT; 4A (5-15%) → 3-month LDCT or PET/CT; 4B/4X (>15%) → tissue sampling / PET / multidisciplinary review. Category-by-category PPV drives the next test (ACR Lung-RADS v2022)
   inputs: prior_ldct_date_and_lung_rads
   actions: workup.lung_nodule_followup
   advance: Every positive screen has a Lung-RADS-PPV-appropriate follow-up route assigned
7. 7DIFFERENTIAL
   True-positive (early-stage lung cancer, stage-shift 158 vs 70 stage I in NLST) vs false-positive cascade (benign granuloma, intrapulmonary lymph node — 96.4% of positives in NLST) vs incidental findings (emphysema, coronary calcium, adrenal/thyroid nodule, fibrotic ILD). The false-positive cascade + overdiagnosis are encoded as harm DATA, not prose (NLST PMID 21714641; Gould AJRCCM 2015 PMID 26214244)
   advance: Each finding classified true-positive vs false-positive-cascade vs incidental
8. 8RISK_STRATIFICATION
   The eligibility model IS the prior estimator: USPSTF 2021 categorical (binary age/pack-yr/quit) vs PLCOm2012 6-yr risk ≥1.3-1.7% continuous. PLCOm2012 sensitivity 83.0% vs USPSTF 71.1% at matched specificity (~62.9% vs 62.7%), 41.3% fewer cancers missed — risk-model selection is the superior prior, expanding eligibility to higher-yield candidates the categorical rule misses (Tammemägi NEJM 2013 PMID 23425165; Ten Haaf PLoS Med 2017 PMID 28376113)
   inputs: smoking_history_pack_years, family_hx_lung_cancer, copd_or_emphysema
   actions: calc.plco_m2012
   advance: Prior tier + eligibility/action threshold assigned (USPSTF categorical or PLCOm2012)
9. 9TREATMENT
   The screen IS the intervention; document shared decision-making + tobacco cessation counseling (CMS requires both for LDCT coverage). Integrate smoking-cessation pharmacotherapy at every screen — the LDCT mortality benefit is conditional on cessation; cessation alone is 3-5x the benefit of early detection in the NLST cohort (USPSTF 2021 PMID 33687470)
   inputs: smoking_history_pack_years
   advance: LDCT ordered, shared decision + tobacco counseling documented, cessation pharmacotherapy offered to active smokers
10. 10DISPOSITION
    All screening is ambulatory; positive screens route to the specific diagnostic pathway: Lung-RADS 4B/4X / new or growing nodule → pulmonology / thoracic surgery / oncology, routing to onc.lung-cancer.core.v1; indeterminate Lung-RADS 3/4A stays in workup.lung_nodule_followup until resolved or biopsied (ACR Lung-RADS v2022; USPSTF 2021)
    advance: Disposition + specialist routing documented
11. 11MONITORING
    Interval + recall: annual LDCT while eligible age 50-80 (USPSTF 2021); Lung-RADS-driven nodule recall (6-mo for category 3, 3-mo for 4A); re-check quit interval and life-expectancy/curative-surgery candidacy at each visit (the prior and the benefit horizon are time-dependent) (USPSTF 2021 PMID 33687470; ACR Lung-RADS v2022)
    inputs: smoking_quit_interval_years, life_expectancy_estimate
    actions: workup.lung_nodule_followup
    advance: Next-screen date set + recall scheduled
12. 12FOLLOWUP
    STOP-screening logic (the deprescribing-equivalent): discontinue when the patient has not smoked for ≥15 yr, OR develops a health problem substantially limiting life expectancy, OR is unwilling/unable to undergo curative lung surgery (USPSTF 2021). Patient education on the false-positive rate (NLST 96.4% of positives benign over 3 rounds), the NNS (~320 to prevent 1 lung-cancer death), overdiagnosis, and the value of NOT screening when harm dominates (USPSTF 2021 PMID 33687470; NLST PMID 21714641)
    inputs: smoking_quit_interval_years, life_expectancy_estimate, curative_surgery_candidacy
    advance: Stop-screening decision OR continued-annual-interval education delivered and documented