Skip to calculator
Experimental screening and nodule research

NTOG Experimental Screening and Nodule Risk Research Hub

Educational browser-only survey and calculators for comparing published screening criteria, person-level lung cancer risk models, nodule malignancy models, growth calculators, and research indices. Entered calculator data is processed locally in the browser and is not transmitted by this page.

Start survey See formulas JSON export

Educational boundary

This static page runs locally in the browser. It compares model outputs, flags missing variables, and exports structured JSON for review.

The page does not send, store, or persist entered data. It does not recommend screening, biopsy, surgery, systemic therapy, radiotherapy, discharge, or surveillance. No patient identifiers are requested or processed.

PLCOm2012 and HUNT estimate person-level future lung cancer risk in ever-smokers. Brock and Mayo estimate nodule-level malignancy probability. The Herder PET-refined calculation is shown as educational-only pending original-table verification. The experimental NTOG emerging-risk questions support research signal documentation and do not alter validated model outputs.

Clinical disclaimer: Use only for education, model comparison, and research planning. Use local validated systems, national guidelines, and multidisciplinary judgement for patient decisions. NTOG scores are not validated and have no clinical action threshold.
Browser-only educational prototype

Survey and calculator

Questions are grouped by subject. NTOG emerging-risk questions are visually separated in green. They feed NTOG research scores and warnings, but never alter the coefficients of PLCOm2012, HUNT, Brock, or Mayo. Herder is displayed as educational-only pending source-table verification.

Privacy: Entered calculator data is processed locally by this static page. This page loads external fonts and CDN assets unless self-hosted.

Governance switches

These switches simulate local governance decisions. Original PLCOm2012 cannot run unless race/ethnicity and education variables are explicitly enabled to reproduce the published model.

Race/ethnicity is used only as a published model variable. It should not be interpreted as a biological risk mechanism.

Demographics and anthropometrics

Used for BMI, PLCOm2012, HUNT, Brock, Mayo, and applicability warnings.

Cigarette smoking and smoke exposure

Pack-years are calculated, but duration, intensity, quit-time, and age-started are preserved separately.

Duration of smoking habit

Respiratory, cancer, and family history

Validated-model fields and clinical context fields are kept in the same subject group.

NTOG research extension

NTOG emerging-risk questions

Green fields feed warnings, research export, and the non-validated NTOG literature-weighted score. They never change PLCOm2012, HUNT, Brock, or Mayo.

CT examination and index nodule

Single-nodule prototype. The JSON structure preserves nodule-level separation.

PET/CT

The educational Herder calculation uses Mayo/Swensen pretest probability and FDG uptake category, and requires original-table verification before clinical or research use. Brock-pretest Herder is not wired in this version.

Basic calculators

BMI
-
kg/m2
Pack-years
-
unknown
Quit-years
-
normalized
VDT
-
days
Growth category: not assessed.

Screening and person-level models

Readiness

Nodule-level models

NTOG experimental scores

Management framework readiness

This lane checks context and data availability only. It does not output Fleischner, BTS, Lung-RADS, NELSON, or ESTI management advice.

Warnings and structured report text

JSON export

Structured export with inputs, derived-variable provenance, model outputs, applicability, formula metadata, thresholds, privacy notes, and dependency graph.

Formulas and score construction

Published model equations are displayed for transparency. PLCOm2012 and HUNT coefficients are inserted pending independent test-case validation. Mayo and Brock are implemented with published/common coefficient forms. Herder remains educational-only pending original-table verification.

Important: The NTOG scores are not validated probability models. Their weights are transparent draft weights. They are not coefficients estimated from a Nordic outcome cohort.

PLCOm2012 original

Person-level 6-year lung cancer risk in ever-smokers. The model first builds a linear predictor from centered clinical and smoking variables, then converts that value to a probability with the logistic function. Original race/ethnicity and education fields require explicit governance opt-in.

Step 1: code variables education: 1 less than high school, 2 high school graduate, 3 post-high-school training, 4 some college, 5 college graduate, 6 postgraduate/professional COPD, personal_cancer_history, family_history_lung_cancer, current_smoker: yes = 1, no = 0 quit_years = 0 for current smokers Step 2: transform variables age_centered = age - 62 education_centered = education - 4 BMI_centered = BMI - 27 smoking_intensity_transform = (cigarettes_per_day/10)^-1 - 0.4021541613 years_smoked_centered = years_smoked - 27 quit_years_centered = quit_years - 10 Step 3: calculate linear predictor lp = -4.532506 + 0.0778868*(age - 62) - 0.0812744*(education - 4) - 0.0274194*(BMI - 27) + 0.3553063*COPD + 0.4589971*personal_cancer_history + 0.5871850*family_history_lung_cancer + 0.2597431*current_smoker - 1.8226060*((cigarettes_per_day/10)^-1 - 0.4021541613) + 0.0317321*(years_smoked - 27) - 0.0308572*(quit_years - 10) + race_ethnicity_term race_ethnicity_term: White / American Indian / Alaskan Native = 0 Black = 0.3944778 Hispanic = -0.7434744 Asian = -0.4665850 Native Hawaiian / Pacific Islander = 1.0271520 Step 4: convert to 6-year probability risk = exp(lp)/(1+exp(lp))

HUNT Lung Cancer Model, corrected 6-year equation

Person-level 6-year risk in ever-smokers. The corrected equation uses natural logarithms with +1 increments, then converts Xb6 to a 6-year probability.

Step 1: code variables male: male = 1, female = 0 daily_cough: yes = 1, no = 0 quit_years = 0 for current smokers smoke_exposure_hours = daily indoor tobacco-smoke exposure hours Step 2: transform variables age_inverse = (age/100)^(-1) ln_pack_years = ln(pack_years + 1) ln_quit_years = ln(quit_years + 1) ln_BMI = ln(BMI + 1) ln_smoke_exposure = ln(smoke_exposure_hours + 1) Step 3: calculate corrected 6-year predictor Xb6 = 1.18203062 + 0.31573217*male - 1.98496138*(age/100)^(-1) + 1.11994217*ln(pack_years + 1) - 0.04002877*cigarettes_per_day - 0.24019955*ln(quit_years + 1) - 1.70238304*ln(BMI + 1) + 0.08072420*ln(smoke_exposure_hours + 1) + 0.49212668*daily_cough Step 4: convert to 6-year probability risk6 = 1/(1+exp(-Xb6))

Brock / PanCan full model with spiculation, as commonly implemented

Nodule-level probability. The full model combines patient factors, nodule morphology, size transformation, nodule count, and location, then applies the logistic conversion. This prototype prefers maximum diameter, falling back to mean diameter if maximum diameter is unavailable, and reads first-degree relative lung cancer from family history.

Step 1: code variables female, family_history_lung_cancer, emphysema, upper_lobe, spiculation: yes = 1, no = 0 diameter_mm = maximum diameter if available, otherwise mean diameter nodule_type_term: solid = 0 part-solid = 0.377 non-solid / ground-glass = -0.1276 Step 2: transform variables age_centered = age - 62 diameter_transform = (diameter_mm/10)^-0.5 - 1.58113883 nodule_count_centered = nodule_count - 4 Step 3: calculate linear predictor x = -6.7892 + 0.0287*(age - 62) + 0.6011*female + 0.2961*family_history_lung_cancer + 0.2953*emphysema - 5.3854*((diameter_mm/10)^-0.5 - 1.58113883) + nodule_type_term + 0.6581*upper_lobe - 0.0824*(nodule_count - 4) + 0.7729*spiculation Step 4: convert to malignancy probability risk = 1/(1+exp(-x))

Mayo / Swensen pretest probability

Nodule-level pretest probability. This model uses clinical risk factors and nodule imaging features to build a linear predictor, then applies the logistic conversion.

Step 1: code variables smoking: current or former smoker = 1, never smoker = 0 previous_extrathoracic_cancer_gt_5y: yes = 1, no = 0 spiculation: yes = 1, no = 0 upper_lobe: yes = 1, no = 0 diameter_mm = maximum diameter if available, otherwise mean diameter Step 2: calculate linear predictor x = -6.8272 + 0.0391*age + 0.7917*smoking + 1.3388*previous_extrathoracic_cancer_gt_5y + 0.1274*diameter_mm + 1.0407*spiculation + 0.7838*upper_lobe Step 3: convert to malignancy probability risk = 1/(1+exp(-x))

Herder PET-refined probability

Herder refines a pretest nodule probability using FDG-PET uptake. This prototype currently uses Mayo probability directly as the pretest probability input and requires verification against the original model table before clinical or research use.

Step 1: start with Mayo pretest probability mayo_probability = Mayo / Swensen probability as a 0-1 value Step 2: code FDG-PET uptake indicators absent/no uptake: faint = 0, moderate = 0, intense = 0 faint/discrete: faint = 1, moderate = 0, intense = 0 moderate: faint = 0, moderate = 1, intense = 0 intense/high: faint = 0, moderate = 0, intense = 1 Step 3: calculate PET-refined predictor x = -4.739 + 3.691*mayo_probability + 2.322*faint_uptake + 4.617*moderate_uptake + 4.771*intense_uptake Step 4: convert to PET-refined probability risk = 1/(1+exp(-x))

Volume doubling time

Requires current and prior volume, CT dates, and same measurement method.

volume_change_ratio = current_volume / prior_volume days_between = daysBetween(prior_date, current_date) vdt = days_between * ln(2) / ln(volume_change_ratio)
NTOG experimental screening model v0.3

ntog_experimental_screening_score

This score is a non-validated NTOG draft research index combining age, tobacco architecture, respiratory/fibrosis phenotype, environmental/occupational exposures, and host/family factors. Range 0-100.

A = age component, max 15 T = tobacco architecture component, max 25 R = respiratory and fibrosis phenotype component, max 20 X = environmental and occupational exposure component, max 20 H = host, family, prior thoracic RT, and immunosuppression component, max 20 Point bands: A: <40y=0, 40-49y=4, 50-54y=7, 55-59y=9, 60-64y=11, 65-69y=13, >=70y=15 T: pack-years up to 8, duration up to 7, intensity up to 4, current/recent quit up to 4, cap 25 R: IPF 8, non-IPF ILD 5, CPFE 6, emphysema 4, COPD 3, cough/infection-pattern points, cap 20 X: radon up to 5, occupational exposures up to 10, passive/indoor smoke and non-cigarette inhaled exposures, cap 20 H: first-degree/family history, prior thoracic RT, prior cancer, transplant, HIV, steroids, biologics/DMARDs, cap 20 Research signal bands: <20 low, 20-39 intermediate, 40-59 elevated, 60-79 high, >=80 very high. No clinical screening threshold is defined.
NTOG post-CT research score

NTOG post-CT research score

This post-CT score is a non-validated, literature-weighted research composite. Missing domains are not imputed; score is normalized by available weight.

Weights if available: P = person-model component, max 20 points, max of PLCOm2012 and HUNT percent risk N = nodule-model component, max 35 points, max of calculated Brock, Mayo, and verified Herder percent risk G = growth component, max 15 points, VDT or simple growth category S = smoking-architecture component, max 10 points, duration, intensity, current smoking, recent quitting E = NTOG emerging-risk component, max 20 points, ILD/CPFE, thoracic RT, immunosuppression, TB/NTM/bronchiectasis, radon, occupational exposures, autoimmune disease, COVID imaging context, diabetes, never-smoker context W_available = sum of component weights that are calculable. Missing validated domains are not imputed. Point bands: P: <0.5%=2, 0.5-0.99%=4, 1.0-1.49%=8, 1.5-1.99%=12, 2.0-3.99%=16, >=4%=20 N: <1%=1, 1-4.99%=4, 5-9.99%=10, 10-29.99%=18, 30-64.99%=28, >=65%=35 G: VDT <100d=15, 100-399d=13, 400-599d=10, categorical growing/new=8, stable/decreasing/resolved=0 S: duration up to 5, intensity up to 2, current smoking 2, quit within 5 years 1, cap 10 E: summed literature modifier points, cap 20

Source notes

  • PLCOm2012: Tammemagi MC et al. N Engl J Med. 2013;368:728-736.
  • HUNT Lung Cancer Model: Markaki M et al. EBioMedicine. 2018;31:36-46. Corrigendum EBioMedicine. 2022;82:104187.
  • Brock / PanCan: McWilliams A et al. N Engl J Med. 2013;369:910-919.
  • Mayo / Swensen: Swensen SJ et al. Arch Intern Med. 1997;157:849-855.
  • Herder: Herder GJ et al. Chest. 2005;128:2490-2496.
  • Emerging-risk inclusion: NCI and CDC lung cancer risk-factor summaries, IARC occupational carcinogen evidence, never-smoker lung cancer literature, IPF/lung cancer systematic review links.