Sync from GitHub (main)

AGENTS.md

@@ -45,11 +45,13 @@ When making changes to the project, ensure that the following files are updated
 Implemented risk calculators include:
 - **Gail** - Breast cancer risk
 - **Claus** - Breast cancer risk based on family history
+- **BOADICEA** - Breast and ovarian cancer risk (via CanRisk API)
 - **PLCOm2012** - Lung cancer risk
+- **LLPi** - Liverpool Lung Project improved model for lung cancer risk (8.7-year prediction)
 - **CRC-PRO** - Colorectal cancer risk
 - **PCPT** - Prostate cancer risk
 - **Extended PBCG** - Prostate cancer risk (extended model)
-- **BOADICEA** - Breast and ovarian cancer risk (via CanRisk API)
+- **MRAT** - Melanoma risk (5-year prediction)
 - **QCancer** - Multi-site cancer differential
 
 Additional models should follow the interfaces under `src/sentinel/risk_models`.

GEMINI.md

@@ -45,11 +45,13 @@ When making changes to the project, ensure that the following files are updated
 Risk calculators exposed to Gemini-based agents include:
 - **Gail** - Breast cancer risk
 - **Claus** - Breast cancer risk based on family history
+- **BOADICEA** - Breast and ovarian cancer risk (via CanRisk API)
 - **PLCOm2012** - Lung cancer risk
+- **LLPi** - Liverpool Lung Project improved model for lung cancer risk (8.7-year prediction)
 - **CRC-PRO** - Colorectal cancer risk
 - **PCPT** - Prostate cancer risk
 - **Extended PBCG** - Prostate cancer risk (extended model)
-- **BOADICEA** - Breast and ovarian cancer risk (via CanRisk API)
+- **MRAT** - Melanoma risk (5-year prediction)
 - **QCancer** - Multi-site cancer differential
 
 Register additional models in `src/sentinel/risk_models/__init__.py` so they are available system-wide.

README.md

@@ -129,11 +129,16 @@ When making changes to the project, check if the following files should also upd
 
 The assistant currently includes the following built-in risk calculators:
 
-- Gail Model (Breast Cancer)
-- PLCOm2012 (Lung Cancer)
-- CRC-PRO (Colorectal Cancer)
-- PCPT (Prostate Cancer)
-- QCancer (Multi-site cancer differential)
+- **Gail** - Breast cancer risk
+- **Claus** - Breast cancer risk based on family history
+- **BOADICEA** - Breast and ovarian cancer risk (via CanRisk API)
+- **PLCOm2012** - Lung cancer risk
+- **LLPi** - Liverpool Lung Project improved model for lung cancer risk (8.7-year prediction)
+- **CRC-PRO** - Colorectal cancer risk
+- **PCPT** - Prostate cancer risk
+- **Extended PBCG** - Prostate cancer risk (extended model)
+- **MRAT** - Melanoma risk (5-year prediction)
+- **QCancer** - Multi-site cancer differential
 
 ## Generating Documentation
 

src/sentinel/risk_models/__init__.py

@@ -5,6 +5,7 @@ from sentinel.risk_models.claus import ClausRiskModel
 from sentinel.risk_models.crc_pro import CRCProRiskModel
 from sentinel.risk_models.extended_pbcg import ExtendedPBCGRiskModel
 from sentinel.risk_models.gail import GailRiskModel
+from sentinel.risk_models.llpi import LLPiRiskModel
 from sentinel.risk_models.mrat import MRATRiskModel
 from sentinel.risk_models.pcpt import PCPTRiskModel
 from sentinel.risk_models.plcom2012 import PLCOm2012RiskModel
@@ -13,6 +14,7 @@ from sentinel.risk_models.qcancer import QCancerRiskModel
 RISK_MODELS = [
     GailRiskModel,
     PLCOm2012RiskModel,
+    LLPiRiskModel,
     BOADICEARiskModel,
     CRCProRiskModel,
     ExtendedPBCGRiskModel,
@@ -26,6 +28,7 @@ __all__ = [
     "RISK_MODELS",
     "ClausRiskModel",
     "GailRiskModel",
+    "LLPiRiskModel",
     "MRATRiskModel",
     "PLCOm2012RiskModel",
 ]

src/sentinel/risk_models/llpi.py

@@ -0,0 +1,230 @@
+"""Lung cancer risk estimation using the Liverpool Lung Project improved (LLPi) model.
+
+This module provides a Python implementation of the LLPi model, which predicts the
+8.7-year probability of developing lung cancer. The model is based on a Cox proportional
+hazards model calibrated to Liverpool data.
+
+The model was published by Marcus et al. (2015) and represents an improvement over
+the original Liverpool Lung Project (LLP) model.
+
+Reference:
+Marcus MW, Chen Y, Raji OY, Duffy SW, Field JK.
+LLPi: Liverpool lung project risk prediction model for lung cancer incidence.
+Cancer Prev Res (Phila) 2015;8:570-5.
+"""
+
+import math
+from typing import Annotated
+
+from pydantic import Field
+
+from sentinel.risk_models.base import RiskModel
+from sentinel.user_input import (
+    CancerType,
+    ChronicCondition,
+    Sex,
+    SmokingStatus,
+    UserInput,
+)
+
+
+class LLPiRiskModel(RiskModel):
+    """Compute 8.7-year absolute lung cancer risk using the LLPi model.
+
+    The LLPi model uses a Cox proportional hazards approach with baseline
+    survival calibrated to Liverpool population data. It requires:
+    - Age
+    - Sex
+    - Years smoked
+    - COPD diagnosis
+    - Prior cancer history
+    - Family history of lung cancer (early vs late onset)
+    """
+
+    # Model coefficients from the R implementation
+    _COEFFICIENTS = {
+        "age": 0.036,
+        "male": 0.391,
+        "smkyears": 0.043,
+        "copd": 0.890,
+        "prior_cancer": 1.044,
+        "fam_cancer_early": 0.521,  # Early onset family history
+        "fam_cancer_late": 0.071,  # Late onset family history
+    }
+
+    # Baseline survival and centering constant from Liverpool data
+    _BASELINE_SURVIVAL = 0.9728386  # Baseline survival at 8.7 years
+    _MEAN_LINEAR_PREDICTOR = 3.556  # Mean linear predictor in Liverpool data
+
+    def __init__(self):
+        super().__init__("llpi")
+
+    REQUIRED_INPUTS: dict[str, tuple[type, bool]] = {
+        "demographics.age_years": (Annotated[int, Field(ge=40, le=85)], True),
+        "demographics.sex": (Sex, True),
+        "lifestyle.smoking.years_smoked": (Annotated[int, Field(ge=0)], True),
+        "personal_medical_history.chronic_conditions": (
+            list,
+            False,
+        ),  # list[ChronicCondition]
+        "personal_medical_history.previous_cancers": (list, False),  # list[CancerType]
+        "family_history": (list, False),  # list[FamilyMemberCancer]
+    }
+
+    def calculate_risk(
+        self,
+        *,
+        age: int,
+        male: int,
+        smkyears: int,
+        copd: int,
+        prior_cancer: int,
+        fam_cancer_onset: int,
+    ) -> float:
+        """Calculate the 8.7-year lung cancer risk using the LLPi formula.
+
+        Args:
+            age: Patient age in years.
+            male: 1 if male, 0 if female.
+            smkyears: Total years smoked.
+            copd: 1 if COPD/emphysema present, 0 otherwise.
+            prior_cancer: 1 if prior cancer history, 0 otherwise.
+            fam_cancer_onset: 0=none, 1=early onset, 2=late onset.
+
+        Returns:
+            Risk as a decimal (0-1).
+        """
+        # Calculate linear predictor
+        linear_predictors = (
+            self._COEFFICIENTS["age"] * age
+            + self._COEFFICIENTS["male"] * male
+            + self._COEFFICIENTS["smkyears"] * smkyears
+            + self._COEFFICIENTS["copd"] * copd
+            + self._COEFFICIENTS["prior_cancer"] * prior_cancer
+        )
+
+        # Add family history component
+        if fam_cancer_onset == 1:  # Early onset
+            linear_predictors += self._COEFFICIENTS["fam_cancer_early"]
+        elif fam_cancer_onset == 2:  # Late onset
+            linear_predictors += self._COEFFICIENTS["fam_cancer_late"]
+
+        # Calculate risk using survival formula
+        # risk = 1 - S0^exp(linear_predictors - mean_linear_predictors)
+        exponent = math.exp(linear_predictors - self._MEAN_LINEAR_PREDICTOR)
+        risk = 1 - (self._BASELINE_SURVIVAL**exponent)
+
+        return risk
+
+    def _determine_family_cancer_onset(self, user: UserInput) -> int:
+        """Determine family cancer onset category from family history.
+
+        Args:
+            user: The user profile.
+
+        Returns:
+            0 = no family history of lung cancer
+            1 = early onset (diagnosed before age 60)
+            2 = late onset (diagnosed at age 60 or later)
+        """
+        if not user.family_history:
+            return 0
+
+        # Check for lung cancer in first-degree relatives
+        lung_cancer_ages = [
+            member.age_at_diagnosis
+            for member in user.family_history
+            if member.cancer_type == CancerType.LUNG
+            and member.age_at_diagnosis is not None
+        ]
+
+        if not lung_cancer_ages:
+            return 0
+
+        # If any relative had early onset (< 60), classify as early onset
+        if any(age < 60 for age in lung_cancer_ages):
+            return 1
+        else:
+            return 2
+
+    def compute_score(self, user: UserInput) -> str:
+        """Compute the score for the LLPi model.
+
+        Args:
+            user: The user profile.
+
+        Returns:
+            str: Risk percentage as a string or an N/A message if not applicable.
+
+        Raises:
+            ValueError: If required inputs are missing or invalid.
+        """
+        # Validate inputs first
+        is_valid, errors = self.validate_inputs(user)
+        if not is_valid:
+            raise ValueError(f"Invalid inputs for LLPi: {'; '.join(errors)}")
+
+        # Check if the user is a current or former smoker
+        if user.lifestyle.smoking.status == SmokingStatus.NEVER:
+            return "N/A: Model is for current or former smokers only."
+
+        # Check for minimum smoking history
+        if user.lifestyle.smoking.years_smoked == 0:
+            return "N/A: Model requires smoking history."
+
+        # Determine sex (male = 1, female = 0)
+        male = 1 if user.demographics.sex == Sex.MALE else 0
+
+        # Check for COPD
+        copd = (
+            1
+            if ChronicCondition.COPD in user.personal_medical_history.chronic_conditions
+            else 0
+        )
+
+        # Check for prior cancer history
+        prior_cancer = 1 if user.personal_medical_history.previous_cancers else 0
+
+        # Determine family cancer onset category
+        fam_cancer_onset = self._determine_family_cancer_onset(user)
+
+        # Calculate the risk
+        risk = self.calculate_risk(
+            age=user.demographics.age_years,
+            male=male,
+            smkyears=user.lifestyle.smoking.years_smoked,
+            copd=copd,
+            prior_cancer=prior_cancer,
+            fam_cancer_onset=fam_cancer_onset,
+        )
+
+        # Convert to percentage
+        risk_percentage = risk * 100
+
+        return f"{risk_percentage:.2f}%"
+
+    def cancer_type(self) -> str:
+        return "lung"
+
+    def description(self) -> str:
+        return (
+            "The Liverpool Lung Project improved (LLPi) model predicts the 8.7-year "
+            "probability of developing lung cancer for current or former smokers. "
+            "The model uses age, sex, smoking history, COPD status, prior cancer history, "
+            "and family history of lung cancer."
+        )
+
+    def interpretation(self) -> str:
+        return (
+            "The score is the percentage chance of developing lung cancer in the next "
+            "8.7 years. A higher score indicates greater risk. This score is based on "
+            "data from the Liverpool Lung Project and can help identify individuals "
+            "who may benefit from lung cancer screening."
+        )
+
+    def references(self) -> list[str]:
+        return [
+            "Marcus MW, Chen Y, Raji OY, Duffy SW, Field JK. "
+            "LLPi: Liverpool lung project risk prediction model for lung cancer incidence. "
+            "Cancer Prev Res (Phila) 2015;8:570-5."
+        ]

tests/test_risk_models/test_llpi_model.py

@@ -0,0 +1,786 @@
+"""Tests for the Liverpool Lung Project improved (LLPi) Risk Model.
+
+Ground truth values are calculated from the reference R implementation
+in lcmodels/R/lcmodels.R (risk.llpi function).
+"""
+
+import pytest
+
+from sentinel.risk_models.llpi import LLPiRiskModel
+from sentinel.user_input import (
+    Anthropometrics,
+    CancerType,
+    ChronicCondition,
+    Demographics,
+    FamilyMemberCancer,
+    FamilyRelation,
+    FamilySide,
+    Lifestyle,
+    PersonalMedicalHistory,
+    RelationshipDegree,
+    Sex,
+    SmokingHistory,
+    SmokingStatus,
+    UserInput,
+)
+
+# Ground truth test cases generated from R implementation
+GROUND_TRUTH_CASES = [
+    {
+        "name": "low_risk_young_female",
+        "input": UserInput(
+            demographics=Demographics(
+                age_years=50,
+                sex=Sex.FEMALE,
+                anthropometrics=Anthropometrics(height_cm=165.0, weight_kg=70.0),
+            ),
+            lifestyle=Lifestyle(
+                smoking=SmokingHistory(
+                    status=SmokingStatus.FORMER,
+                    years_smoked=10,
+                    cigarettes_per_day=10,
+                    years_since_quit=5,
+                )
+            ),
+            personal_medical_history=PersonalMedicalHistory(
+                chronic_conditions=[],
+                previous_cancers=[],
+            ),
+            family_history=[],
+        ),
+        "expected": 0.73,
+    },
+    {
+        "name": "moderate_risk_male",
+        "input": UserInput(
+            demographics=Demographics(
+                age_years=60,
+                sex=Sex.MALE,
+                anthropometrics=Anthropometrics(height_cm=175.0, weight_kg=80.0),
+            ),
+            lifestyle=Lifestyle(
+                smoking=SmokingHistory(
+                    status=SmokingStatus.CURRENT,
+                    years_smoked=25,
+                    cigarettes_per_day=20,
+                    years_since_quit=None,
+                )
+            ),
+            personal_medical_history=PersonalMedicalHistory(
+                chronic_conditions=[],
+                previous_cancers=[],
+            ),
+            family_history=[],
+        ),
+        "expected": 2.91,
+    },
+    {
+        "name": "high_risk_copd",
+        "input": UserInput(
+            demographics=Demographics(
+                age_years=70,
+                sex=Sex.MALE,
+                anthropometrics=Anthropometrics(height_cm=175.0, weight_kg=75.0),
+            ),
+            lifestyle=Lifestyle(
+                smoking=SmokingHistory(
+                    status=SmokingStatus.FORMER,
+                    years_smoked=45,
+                    cigarettes_per_day=30,
+                    years_since_quit=2,
+                )
+            ),
+            personal_medical_history=PersonalMedicalHistory(
+                chronic_conditions=[ChronicCondition.COPD],
+                previous_cancers=[],
+            ),
+            family_history=[],
+        ),
+        "expected": 21.62,
+    },
+    {
+        "name": "high_risk_prior_cancer",
+        "input": UserInput(
+            demographics=Demographics(
+                age_years=65,
+                sex=Sex.MALE,
+                anthropometrics=Anthropometrics(height_cm=175.0, weight_kg=78.0),
+            ),
+            lifestyle=Lifestyle(
+                smoking=SmokingHistory(
+                    status=SmokingStatus.FORMER,
+                    years_smoked=35,
+                    cigarettes_per_day=25,
+                    years_since_quit=5,
+                )
+            ),
+            personal_medical_history=PersonalMedicalHistory(
+                chronic_conditions=[],
+                previous_cancers=[CancerType.PROSTATE],
+            ),
+            family_history=[],
+        ),
+        "expected": 14.31,
+    },
+    {
+        "name": "early_onset_family_history",
+        "input": UserInput(
+            demographics=Demographics(
+                age_years=55,
+                sex=Sex.FEMALE,
+                anthropometrics=Anthropometrics(height_cm=165.0, weight_kg=68.0),
+            ),
+            lifestyle=Lifestyle(
+                smoking=SmokingHistory(
+                    status=SmokingStatus.CURRENT,
+                    years_smoked=20,
+                    cigarettes_per_day=15,
+                    years_since_quit=None,
+                )
+            ),
+            personal_medical_history=PersonalMedicalHistory(
+                chronic_conditions=[],
+                previous_cancers=[],
+            ),
+            family_history=[
+                FamilyMemberCancer(
+                    relation=FamilyRelation.MOTHER,
+                    side=FamilySide.MATERNAL,
+                    degree=RelationshipDegree.FIRST,
+                    cancer_type=CancerType.LUNG,
+                    age_at_diagnosis=55,  # Early onset (< 60)
+                )
+            ],
+        ),
+        "expected": 2.24,
+    },
+    {
+        "name": "late_onset_family_history",
+        "input": UserInput(
+            demographics=Demographics(
+                age_years=58,
+                sex=Sex.MALE,
+                anthropometrics=Anthropometrics(height_cm=178.0, weight_kg=82.0),
+            ),
+            lifestyle=Lifestyle(
+                smoking=SmokingHistory(
+                    status=SmokingStatus.FORMER,
+                    years_smoked=30,
+                    cigarettes_per_day=20,
+                    years_since_quit=3,
+                )
+            ),
+            personal_medical_history=PersonalMedicalHistory(
+                chronic_conditions=[],
+                previous_cancers=[],
+            ),
+            family_history=[
+                FamilyMemberCancer(
+                    relation=FamilyRelation.FATHER,
+                    side=FamilySide.PATERNAL,
+                    degree=RelationshipDegree.FIRST,
+                    cancer_type=CancerType.LUNG,
+                    age_at_diagnosis=65,  # Late onset (>= 60)
+                )
+            ],
+        ),
+        "expected": 3.59,
+    },
+    {
+        "name": "multiple_risk_factors",
+        "input": UserInput(
+            demographics=Demographics(
+                age_years=68,
+                sex=Sex.MALE,
+                anthropometrics=Anthropometrics(height_cm=172.0, weight_kg=76.0),
+            ),
+            lifestyle=Lifestyle(
+                smoking=SmokingHistory(
+                    status=SmokingStatus.CURRENT,
+                    years_smoked=40,
+                    cigarettes_per_day=30,
+                    years_since_quit=None,
+                )
+            ),
+            personal_medical_history=PersonalMedicalHistory(
+                chronic_conditions=[ChronicCondition.COPD],
+                previous_cancers=[CancerType.PROSTATE],
+            ),
+            family_history=[
+                FamilyMemberCancer(
+                    relation=FamilyRelation.MOTHER,
+                    side=FamilySide.MATERNAL,
+                    degree=RelationshipDegree.FIRST,
+                    cancer_type=CancerType.LUNG,
+                    age_at_diagnosis=58,  # Early onset
+                )
+            ],
+        ),
+        "expected": 58.29,
+    },
+    {
+        "name": "female_copd_family_history",
+        "input": UserInput(
+            demographics=Demographics(
+                age_years=62,
+                sex=Sex.FEMALE,
+                anthropometrics=Anthropometrics(height_cm=162.0, weight_kg=65.0),
+            ),
+            lifestyle=Lifestyle(
+                smoking=SmokingHistory(
+                    status=SmokingStatus.FORMER,
+                    years_smoked=28,
+                    cigarettes_per_day=18,
+                    years_since_quit=4,
+                )
+            ),
+            personal_medical_history=PersonalMedicalHistory(
+                chronic_conditions=[ChronicCondition.COPD],
+                previous_cancers=[],
+            ),
+            family_history=[
+                FamilyMemberCancer(
+                    relation=FamilyRelation.FATHER,
+                    side=FamilySide.PATERNAL,
+                    degree=RelationshipDegree.FIRST,
+                    cancer_type=CancerType.LUNG,
+                    age_at_diagnosis=70,  # Late onset
+                )
+            ],
+        ),
+        "expected": 6.19,
+    },
+    {
+        "name": "minimal_smoking_no_risk",
+        "input": UserInput(
+            demographics=Demographics(
+                age_years=52,
+                sex=Sex.FEMALE,
+                anthropometrics=Anthropometrics(height_cm=160.0, weight_kg=62.0),
+            ),
+            lifestyle=Lifestyle(
+                smoking=SmokingHistory(
+                    status=SmokingStatus.FORMER,
+                    years_smoked=5,
+                    cigarettes_per_day=8,
+                    years_since_quit=10,
+                )
+            ),
+            personal_medical_history=PersonalMedicalHistory(
+                chronic_conditions=[],
+                previous_cancers=[],
+            ),
+            family_history=[],
+        ),
+        "expected": 0.63,
+    },
+    {
+        "name": "heavy_smoker_female",
+        "input": UserInput(
+            demographics=Demographics(
+                age_years=58,
+                sex=Sex.FEMALE,
+                anthropometrics=Anthropometrics(height_cm=168.0, weight_kg=72.0),
+            ),
+            lifestyle=Lifestyle(
+                smoking=SmokingHistory(
+                    status=SmokingStatus.CURRENT,
+                    years_smoked=38,
+                    cigarettes_per_day=25,
+                    years_since_quit=None,
+                )
+            ),
+            personal_medical_history=PersonalMedicalHistory(
+                chronic_conditions=[],
+                previous_cancers=[],
+            ),
+            family_history=[],
+        ),
+        "expected": 3.20,
+    },
+]
+
+
+class TestLLPiModel:
+    """Test suite for LLPiRiskModel."""
+
+    def setup_method(self):
+        """Initialize LLPiRiskModel instance for testing."""
+        self.model = LLPiRiskModel()
+
+    @pytest.mark.parametrize("case", GROUND_TRUTH_CASES, ids=lambda x: x["name"])
+    def test_ground_truth_validation(self, case):
+        """Test against calculated ground truth results from R implementation.
+
+        Args:
+            case: Parameterized ground truth case dict.
+        """
+        user = case["input"]
+        score_str = self.model.compute_score(user)
+        calculated = float(score_str.rstrip("%"))
+        expected = case["expected"]
+
+        # Using tight tolerance since these are calculated values
+        assert calculated == pytest.approx(expected, abs=0.01)
+
+    def test_never_smoker_handling(self):
+        """Test that never smokers receive N/A response."""
+        never_smoker = UserInput(
+            demographics=Demographics(
+                age_years=55,
+                sex=Sex.MALE,
+                anthropometrics=Anthropometrics(height_cm=175.0, weight_kg=80.0),
+            ),
+            lifestyle=Lifestyle(
+                smoking=SmokingHistory(
+                    status=SmokingStatus.NEVER,
+                    cigarettes_per_day=0,
+                    years_smoked=0,
+                    years_since_quit=None,
+                )
+            ),
+            personal_medical_history=PersonalMedicalHistory(),
+            family_history=[],
+        )
+
+        score = self.model.compute_score(never_smoker)
+        assert score == "N/A: Model is for current or former smokers only."
+
+    def test_no_smoking_history(self):
+        """Test handling when years_smoked is 0."""
+        user = UserInput(
+            demographics=Demographics(
+                age_years=55,
+                sex=Sex.MALE,
+                anthropometrics=Anthropometrics(height_cm=175.0, weight_kg=80.0),
+            ),
+            lifestyle=Lifestyle(
+                smoking=SmokingHistory(
+                    status=SmokingStatus.CURRENT,
+                    cigarettes_per_day=10,
+                    years_smoked=0,
+                    years_since_quit=None,
+                )
+            ),
+            personal_medical_history=PersonalMedicalHistory(),
+            family_history=[],
+        )
+
+        score = self.model.compute_score(user)
+        assert "Model requires smoking history" in score
+
+    def test_age_validation(self):
+        """Test age validation (40-85 range)."""
+        young_user = UserInput(
+            demographics=Demographics(
+                age_years=35,
+                sex=Sex.MALE,
+                anthropometrics=Anthropometrics(height_cm=175.0, weight_kg=80.0),
+            ),
+            lifestyle=Lifestyle(
+                smoking=SmokingHistory(
+                    status=SmokingStatus.CURRENT,
+                    cigarettes_per_day=20,
+                    years_smoked=15,
+                    years_since_quit=None,
+                )
+            ),
+            personal_medical_history=PersonalMedicalHistory(),
+            family_history=[],
+        )
+
+        with pytest.raises(ValueError, match=r"Invalid inputs for LLPi:"):
+            self.model.compute_score(young_user)
+
+        old_user = UserInput(
+            demographics=Demographics(
+                age_years=90,
+                sex=Sex.MALE,
+                anthropometrics=Anthropometrics(height_cm=175.0, weight_kg=80.0),
+            ),
+            lifestyle=Lifestyle(
+                smoking=SmokingHistory(
+                    status=SmokingStatus.FORMER,
+                    cigarettes_per_day=20,
+                    years_smoked=30,
+                    years_since_quit=10,
+                )
+            ),
+            personal_medical_history=PersonalMedicalHistory(),
+            family_history=[],
+        )
+
+        with pytest.raises(ValueError, match=r"Invalid inputs for LLPi:"):
+            self.model.compute_score(old_user)
+
+    def test_copd_detection(self):
+        """Test COPD detection from chronic conditions."""
+        user = UserInput(
+            demographics=Demographics(
+                age_years=60,
+                sex=Sex.MALE,
+                anthropometrics=Anthropometrics(height_cm=175.0, weight_kg=80.0),
+            ),
+            lifestyle=Lifestyle(
+                smoking=SmokingHistory(
+                    status=SmokingStatus.CURRENT,
+                    cigarettes_per_day=20,
+                    years_smoked=25,
+                    years_since_quit=None,
+                )
+            ),
+            personal_medical_history=PersonalMedicalHistory(
+                chronic_conditions=[ChronicCondition.COPD, ChronicCondition.DIABETES],
+            ),
+            family_history=[],
+        )
+
+        score = self.model.compute_score(user)
+        assert "%" in score
+
+        # Risk should be higher with COPD
+        risk_with_copd = float(score.rstrip("%"))
+
+        # Compare to same user without COPD
+        user_no_copd = UserInput(
+            demographics=Demographics(
+                age_years=60,
+                sex=Sex.MALE,
+                anthropometrics=Anthropometrics(height_cm=175.0, weight_kg=80.0),
+            ),
+            lifestyle=Lifestyle(
+                smoking=SmokingHistory(
+                    status=SmokingStatus.CURRENT,
+                    cigarettes_per_day=20,
+                    years_smoked=25,
+                    years_since_quit=None,
+                )
+            ),
+            personal_medical_history=PersonalMedicalHistory(
+                chronic_conditions=[ChronicCondition.DIABETES],
+            ),
+            family_history=[],
+        )
+
+        score_no_copd = self.model.compute_score(user_no_copd)
+        risk_no_copd = float(score_no_copd.rstrip("%"))
+
+        assert risk_with_copd > risk_no_copd
+
+    def test_prior_cancer_detection(self):
+        """Test prior cancer history detection."""
+        user = UserInput(
+            demographics=Demographics(
+                age_years=60,
+                sex=Sex.MALE,
+                anthropometrics=Anthropometrics(height_cm=175.0, weight_kg=80.0),
+            ),
+            lifestyle=Lifestyle(
+                smoking=SmokingHistory(
+                    status=SmokingStatus.CURRENT,
+                    cigarettes_per_day=20,
+                    years_smoked=25,
+                    years_since_quit=None,
+                )
+            ),
+            personal_medical_history=PersonalMedicalHistory(
+                chronic_conditions=[],
+                previous_cancers=[CancerType.COLORECTAL, CancerType.PROSTATE],
+            ),
+            family_history=[],
+        )
+
+        score = self.model.compute_score(user)
+        assert "%" in score
+
+        # Risk should be higher with prior cancer
+        risk_with_cancer = float(score.rstrip("%"))
+
+        # Compare to same user without prior cancer
+        user_no_cancer = UserInput(
+            demographics=Demographics(
+                age_years=60,
+                sex=Sex.MALE,
+                anthropometrics=Anthropometrics(height_cm=175.0, weight_kg=80.0),
+            ),
+            lifestyle=Lifestyle(
+                smoking=SmokingHistory(
+                    status=SmokingStatus.CURRENT,
+                    cigarettes_per_day=20,
+                    years_smoked=25,
+                    years_since_quit=None,
+                )
+            ),
+            personal_medical_history=PersonalMedicalHistory(
+                chronic_conditions=[],
+                previous_cancers=[],
+            ),
+            family_history=[],
+        )
+
+        score_no_cancer = self.model.compute_score(user_no_cancer)
+        risk_no_cancer = float(score_no_cancer.rstrip("%"))
+
+        assert risk_with_cancer > risk_no_cancer
+
+    def test_family_history_early_vs_late(self):
+        """Test that early onset family history has higher coefficient than late onset."""
+        base_user_data = {
+            "demographics": Demographics(
+                age_years=60,
+                sex=Sex.MALE,
+                anthropometrics=Anthropometrics(height_cm=175.0, weight_kg=80.0),
+            ),
+            "lifestyle": Lifestyle(
+                smoking=SmokingHistory(
+                    status=SmokingStatus.CURRENT,
+                    cigarettes_per_day=20,
+                    years_smoked=25,
+                    years_since_quit=None,
+                )
+            ),
+            "personal_medical_history": PersonalMedicalHistory(),
+        }
+
+        # Early onset family history
+        user_early = UserInput(
+            **base_user_data,
+            family_history=[
+                FamilyMemberCancer(
+                    relation=FamilyRelation.MOTHER,
+                    side=FamilySide.MATERNAL,
+                    degree=RelationshipDegree.FIRST,
+                    cancer_type=CancerType.LUNG,
+                    age_at_diagnosis=55,  # Early onset
+                )
+            ],
+        )
+
+        # Late onset family history
+        user_late = UserInput(
+            **base_user_data,
+            family_history=[
+                FamilyMemberCancer(
+                    relation=FamilyRelation.FATHER,
+                    side=FamilySide.PATERNAL,
+                    degree=RelationshipDegree.FIRST,
+                    cancer_type=CancerType.LUNG,
+                    age_at_diagnosis=70,  # Late onset
+                )
+            ],
+        )
+
+        # No family history
+        user_none = UserInput(**base_user_data, family_history=[])
+
+        risk_early = float(self.model.compute_score(user_early).rstrip("%"))
+        risk_late = float(self.model.compute_score(user_late).rstrip("%"))
+        risk_none = float(self.model.compute_score(user_none).rstrip("%"))
+
+        # Early onset should confer higher risk than late onset
+        assert risk_early > risk_late > risk_none
+
+    def test_family_history_non_lung_cancer_ignored(self):
+        """Test that non-lung cancer family history is ignored."""
+        user_lung = UserInput(
+            demographics=Demographics(
+                age_years=60,
+                sex=Sex.MALE,
+                anthropometrics=Anthropometrics(height_cm=175.0, weight_kg=80.0),
+            ),
+            lifestyle=Lifestyle(
+                smoking=SmokingHistory(
+                    status=SmokingStatus.CURRENT,
+                    cigarettes_per_day=20,
+                    years_smoked=25,
+                    years_since_quit=None,
+                )
+            ),
+            personal_medical_history=PersonalMedicalHistory(),
+            family_history=[
+                FamilyMemberCancer(
+                    relation=FamilyRelation.MOTHER,
+                    side=FamilySide.MATERNAL,
+                    degree=RelationshipDegree.FIRST,
+                    cancer_type=CancerType.LUNG,
+                    age_at_diagnosis=65,
+                )
+            ],
+        )
+
+        user_breast = UserInput(
+            demographics=Demographics(
+                age_years=60,
+                sex=Sex.MALE,
+                anthropometrics=Anthropometrics(height_cm=175.0, weight_kg=80.0),
+            ),
+            lifestyle=Lifestyle(
+                smoking=SmokingHistory(
+                    status=SmokingStatus.CURRENT,
+                    cigarettes_per_day=20,
+                    years_smoked=25,
+                    years_since_quit=None,
+                )
+            ),
+            personal_medical_history=PersonalMedicalHistory(),
+            family_history=[
+                FamilyMemberCancer(
+                    relation=FamilyRelation.MOTHER,
+                    side=FamilySide.MATERNAL,
+                    degree=RelationshipDegree.FIRST,
+                    cancer_type=CancerType.BREAST,
+                    age_at_diagnosis=65,
+                )
+            ],
+        )
+
+        user_none = UserInput(
+            demographics=Demographics(
+                age_years=60,
+                sex=Sex.MALE,
+                anthropometrics=Anthropometrics(height_cm=175.0, weight_kg=80.0),
+            ),
+            lifestyle=Lifestyle(
+                smoking=SmokingHistory(
+                    status=SmokingStatus.CURRENT,
+                    cigarettes_per_day=20,
+                    years_smoked=25,
+                    years_since_quit=None,
+                )
+            ),
+            personal_medical_history=PersonalMedicalHistory(),
+            family_history=[],
+        )
+
+        risk_lung = float(self.model.compute_score(user_lung).rstrip("%"))
+        risk_breast = float(self.model.compute_score(user_breast).rstrip("%"))
+        risk_none = float(self.model.compute_score(user_none).rstrip("%"))
+
+        # Lung cancer family history should increase risk
+        assert risk_lung > risk_none
+        # Breast cancer family history should not affect risk
+        assert risk_breast == pytest.approx(risk_none, abs=0.01)
+
+    def test_sex_difference(self):
+        """Test that males have higher risk than females (all else equal)."""
+        user_male = UserInput(
+            demographics=Demographics(
+                age_years=60,
+                sex=Sex.MALE,
+                anthropometrics=Anthropometrics(height_cm=175.0, weight_kg=80.0),
+            ),
+            lifestyle=Lifestyle(
+                smoking=SmokingHistory(
+                    status=SmokingStatus.CURRENT,
+                    cigarettes_per_day=20,
+                    years_smoked=25,
+                    years_since_quit=None,
+                )
+            ),
+            personal_medical_history=PersonalMedicalHistory(),
+            family_history=[],
+        )
+
+        user_female = UserInput(
+            demographics=Demographics(
+                age_years=60,
+                sex=Sex.FEMALE,
+                anthropometrics=Anthropometrics(height_cm=175.0, weight_kg=80.0),
+            ),
+            lifestyle=Lifestyle(
+                smoking=SmokingHistory(
+                    status=SmokingStatus.CURRENT,
+                    cigarettes_per_day=20,
+                    years_smoked=25,
+                    years_since_quit=None,
+                )
+            ),
+            personal_medical_history=PersonalMedicalHistory(),
+            family_history=[],
+        )
+
+        risk_male = float(self.model.compute_score(user_male).rstrip("%"))
+        risk_female = float(self.model.compute_score(user_female).rstrip("%"))
+
+        # Males should have higher risk (positive coefficient for male)
+        assert risk_male > risk_female
+
+    def test_model_metadata(self):
+        """Test model metadata methods."""
+        assert self.model.name == "llpi"
+        assert self.model.cancer_type() == "lung"
+        assert "LLPi" in self.model.description()
+        assert "8.7-year" in self.model.description()
+        assert "percentage chance" in self.model.interpretation()
+        assert isinstance(self.model.references(), list)
+        assert len(self.model.references()) > 0
+        assert "Marcus" in self.model.references()[0]
+
+    def test_calculate_risk_directly(self):
+        """Test the calculate_risk method directly with known inputs."""
+        # Test case: moderate_risk_male from ground truth
+        risk = self.model.calculate_risk(
+            age=60,
+            male=1,
+            smkyears=25,
+            copd=0,
+            prior_cancer=0,
+            fam_cancer_onset=0,
+        )
+
+        # Should match ground truth: 2.91%
+        assert risk * 100 == pytest.approx(2.91, abs=0.01)
+
+    def test_increasing_age_increases_risk(self):
+        """Test that increasing age increases risk (positive age coefficient)."""
+        risks = []
+        for age in [50, 60, 70]:
+            user = UserInput(
+                demographics=Demographics(
+                    age_years=age,
+                    sex=Sex.MALE,
+                    anthropometrics=Anthropometrics(height_cm=175.0, weight_kg=80.0),
+                ),
+                lifestyle=Lifestyle(
+                    smoking=SmokingHistory(
+                        status=SmokingStatus.CURRENT,
+                        cigarettes_per_day=20,
+                        years_smoked=25,
+                        years_since_quit=None,
+                    )
+                ),
+                personal_medical_history=PersonalMedicalHistory(),
+                family_history=[],
+            )
+            risk = float(self.model.compute_score(user).rstrip("%"))
+            risks.append(risk)
+
+        # Risk should increase with age
+        assert risks[0] < risks[1] < risks[2]
+
+    def test_increasing_smoking_years_increases_risk(self):
+        """Test that longer smoking history increases risk."""
+        risks = []
+        for years in [10, 25, 40]:
+            user = UserInput(
+                demographics=Demographics(
+                    age_years=60,
+                    sex=Sex.MALE,
+                    anthropometrics=Anthropometrics(height_cm=175.0, weight_kg=80.0),
+                ),
+                lifestyle=Lifestyle(
+                    smoking=SmokingHistory(
+                        status=SmokingStatus.CURRENT,
+                        cigarettes_per_day=20,
+                        years_smoked=years,
+                        years_since_quit=None,
+                    )
+                ),
+                personal_medical_history=PersonalMedicalHistory(),
+                family_history=[],
+            )
+            risk = float(self.model.compute_score(user).rstrip("%"))
+            risks.append(risk)
+
+        # Risk should increase with more smoking years
+        assert risks[0] < risks[1] < risks[2]