Apply the original ordered truth table to a lytic lesion on radiographs, return every raw descriptor, and keep later variants, histology, management and numeric risk separate.
{
"categories": [
{
"outcome_code": "IA",
"pattern": "geographic",
"margin": "regular_lobulated_or_multicentric",
"cortical_penetration": "absent_or_partial",
"sclerotic_rim": "complete",
"expanded_shell": "absent_or_no_more_than_1_cm"
},
{
"outcome_code": "IB",
"pattern": "geographic_without_moth_eaten_margin",
"cortical_penetration": "absent_or_partial",
"triggers": [
"ragged_or_poorly_defined_margin",
"absent_or_incomplete_sclerotic_rim",
"expanded_shell_greater_than_1_cm"
]
},
{
"outcome_code": "IC",
"pattern": "geographic",
"triggers": [
"moth_eaten_margin_no_more_than_1_cm",
"total_cortical_penetration_unless_a_higher_pattern_already_applies"
]
},
{
"outcome_code": "II",
"pattern": "moth_eaten_or_geographic_with_moth_eaten_margin_greater_than_1_cm"
},
{
"outcome_code": "III",
"pattern": "any_permeative_component_anywhere_in_the_lesion"
}
],
"applicability": {
"intended_use": "Estimate relative growth rate and radiographic aggressiveness of an evidently lytic focal bone lesion on conventional radiographs.",
"required_context": [
"patient_age",
"named_bone_and_segment",
"adequate_radiographs_in_available_orthogonal_views",
"lesion_is_evidently_lytic",
"pattern_and_margin_visible",
"cortex_visible",
"comparison_exams_when_available"
],
"outside_scope": [
"predominantly_sclerotic_or_mixed_lesion_without_a_lytic_component",
"histologic_diagnosis",
"malignancy_confirmation",
"tumor_stage",
"pathologic_fracture_score",
"treatment_selection_by_grade_alone",
"calibrated_individual_probability"
],
"nonneoplastic_guard": "The algorithm describes growth morphology, not etiology. Infection, metabolic lesions and other tumor-like processes can appear aggressive; benign lesions can occasionally receive a high grade."
},
"descriptor_dictionary": {
"geographic": "A single or confluent hole in bone. Its margin is subclassified as regular, lobulated, multicentric, ragged or poorly defined, or as a moth-eaten margin no more than or greater than 1 cm.",
"moth_eaten": "Multiple apparently randomly distributed holes lacking uniformity of size.",
"permeative": "Multiple uniformly small holes; any permeative component anywhere in an otherwise geographic lesion activates grade III.",
"cortical_penetration": "Record absent, partial or total. Partial penetration is compatible with IA or IB; total penetration in a geographic lesion activates IC unless the earlier pattern branch already yields II or III.",
"sclerotic_rim": "The rim must be complete to count as present for IA. A faint or incomplete rim is treated by the most aggressive margin feature and supports IB.",
"expanded_shell": "Measure expansion beyond the expected normal cortical contour. No expansion or no more than 1 cm may satisfy IA; more than 1 cm activates IB when the earlier branches have not assigned a higher grade."
},
"ordered_decision_tree": [
{
"priority": 1,
"if": "permeative_destruction_present_anywhere",
"output_code": "III"
},
{
"priority": 2,
"if": "entirely_moth_eaten_pattern_or_geographic_lesion_with_moth_eaten_margin_greater_than_1_cm",
"output_code": "II"
},
{
"priority": 3,
"if": "geographic_lesion_with_moth_eaten_margin_no_more_than_1_cm",
"output_code": "IC"
},
{
"priority": 4,
"if": "geographic_lesion_with_total_cortical_penetration",
"output_code": "IC"
},
{
"priority": 5,
"if": "geographic_regular_lobulated_or_multicentric_margin_and_absent_or_partial_cortical_penetration_and_complete_sclerotic_rim_and_expanded_shell_absent_or_no_more_than_1_cm",
"output_code": "IA"
},
{
"priority": 6,
"if": "remaining_geographic_lesion_without_moth_eaten_margin_and_with_absent_or_partial_cortical_penetration",
"output_code": "IB"
}
],
"boundary_and_pitfall_rules": {
"IA_only_combination": "IA has exactly one descriptor combination: regular, lobulated or multicentric geographic margin; no or partial cortical penetration; complete sclerotic rim; and no or at most 1 cm expanded shell.",
"incomplete_rim": "A rim visible around only part of the lesion is not complete sclerosis and supports IB, not IA.",
"shell_boundary": "Expansion is measured beyond the expected normal contour. Exactly 1 cm remains IA-compatible; greater than 1 cm supports IB when all earlier aggressive-pattern branches are absent.",
"thin_moth_eaten_margin": "A geographic moth-eaten margin no more than 1 cm is IC, not IB or II.",
"wide_moth_eaten_margin": "A geographic moth-eaten margin greater than 1 cm is II even when the dominant component appears geographic.",
"total_cortex": "Total cortical penetration makes an otherwise geographic lesion IC. Do not continue to rim or shell, and do not downgrade a prior II or III pattern.",
"small_permeative_focus": "A small permeative component makes the whole lesion grade III even when most of it is geographic.",
"most_aggressive_feature": "Follow the ordered tree and retain all raw descriptors. Do not average mixed components or let a benign-appearing majority erase a higher-priority component."
},
"acquisition_and_comparison_rules": {
"radiography_role": "Conventional radiography is the central modality for this growth-pattern algorithm. CT can clarify cortex and matrix; MRI is superior for local staging but must not silently replace radiographic descriptor provenance.",
"comparison_rule": "Record interval change and prior dates separately. A changing margin is important but belongs to Madewell or modified Lodwick-Madewell variants, not the original five-category truth table.",
"quality_rule": "If the margin, cortex or shell is obscured by projection, fracture, hardware or anatomic overlap, return an adjacent-grade range or unclassifiable state rather than assuming the missing descriptor is absent."
},
"variant_and_version_boundary": {
"original_1980": "This RadCommons system implements the original IA, IB, IC, II and III algorithm only.",
"Madewell_1981": "Madewell added changing margins on serial studies and combination patterns. Label that variant explicitly.",
"modified_Lodwick_Madewell_2016": "Caracciolo's variant combines original IA and IB as grade I, reassigns original IC to grade II, uses IIIA for changing margins or atypical suspicious combinations, IIIB for original moth-eaten/permeative II-III patterns and IIIC for radiographically occult lesions.",
"noninterchangeability": "Do not silently map original IA-III codes to modified I-IIIA-IIIB-IIIC codes. Neither adjusted variant has been shown superior to the original algorithm."
},
"required_companion_analysis": [
"patient_age_and_clinical_presentation",
"bone_segment_and_medullary_cortical_or_surface_location",
"lesion_size_and_multiplicity",
"matrix_mineralization",
"periosteal_reaction",
"cortical_remodeling_destruction_and_pathologic_fracture",
"soft_tissue_component",
"comparison_stability_or_change",
"known_primary_malignancy_or_systemic_disease"
],
"management_and_cross_system_boundary": {
"grade_only_rule": "No original Lodwick grade independently orders observation, interval imaging, biopsy, surgery or oncology referral. Combine the complete lesion appearance with age, site, symptoms, matrix, periosteal reaction, multiplicity, fracture and clinical context.",
"current_imaging_context": "ACR 2024 guidance recommends radiography for initial detection and characterization of a suspected primary bone tumor. For a radiographically indeterminate or aggressive lesion, MRI or CT is typically an appropriate next study for extent, viability and biopsy or surgical planning.",
"biopsy_planning": "When biopsy is considered, coordinate with the treating orthopedic-oncology team so the tract and definitive surgical plan are not compromised; the grade itself is not a biopsy target-selection algorithm.",
"Bone_RADS_guard": "Bone-RADS is a separate management framework for solitary lesions on CT/MRI, with its own current ACR extensions. It is not an updated numerical translation of original radiographic Lodwick grades.",
"FDG_guard": "FDG uptake overlaps between benign and malignant bone lesions; PET is an adjunct for selected staging or response questions and does not replace morphology or histology."
},
"risk_and_reliability": {
"direction": "Higher original grade indicates faster inferred growth and generally greater malignancy concern, but no grade proves benignity or malignancy.",
"modified_variant_cohort": "In a retrospective 183-lesion study of the modified Lodwick-Madewell system, 76 of 81 grade-I lesions (94 percent) were benign and 39 of 48 grade-III lesions (81 percent) were malignant. These figures belong to the modified three-tier variant and cannot be assigned as IA-, IB- or original-III bedside probabilities.",
"current_reliability": "A 2025 study of 48 radiolucent-lesion case sets and 20 readers found only 39 percent interobserver agreement for original Lodwick (Fleiss kappa 0.23; ordinal alpha 0.54) and intraobserver kappa 0.45. Highest and lowest grades were more reproducible than intermediate grades.",
"implication": "Return the descriptors and uncertainty with the grade. Reader disagreement and cohort selection preclude a precise individual malignancy probability."
},
"agent_output_contract": [
"original_Lodwick_IA_IB_IC_II_III_or_explicit_uncertainty",
"radiographic_views_and_technical_adequacy",
"destruction_pattern_and_geographic_margin_subtype",
"cortical_penetration_absent_partial_or_total",
"complete_incomplete_or_absent_sclerotic_rim",
"expanded_shell_measurement_relative_to_expected_contour",
"all_companion_lesion_features_and_comparison_change",
"variant_name_if_not_original_1980",
"no_histology_management_or_numeric_risk_claim_from_grade_alone"
],
"missing_input_behavior": [
"If the lesion is not evidently lytic or the radiographic pattern cannot be assessed, return not_applicable_or_unclassifiable rather than assigning IA.",
"If cortex is not visible, do not assume absent penetration; retain the grade range supported by the known pattern and margin.",
"If the rim is incomplete or equivocal, do not call it complete for IA; report IB or IA-versus-IB according to confidence.",
"If shell expansion is near 1 cm without a reliable normal contour, retain IA-versus-IB and report the measurement uncertainty.",
"If mixed moth-eaten or permeative foci are suspected but not resolved, preserve IC-versus-II or II-versus-III uncertainty and recommend adequate characterization.",
"If a modified Lodwick-Madewell code is requested, do not output the original code set without first naming and applying that separate variant."
],
"supporting_sources": [
{
"role": "original_algorithm",
"citation": "Lodwick et al. Radiology. 1980;134:577-583",
"doi": "10.1148/radiology.134.3.6928321",
"pmid": "6928321"
},
{
"role": "original_truth_table_decision_tree",
"citation": "Benndorf et al. Skeletal Radiol. 2022;51:737-745",
"doi": "10.1007/s00256-021-03868-8",
"pmid": "34302499",
"pmcid": "PMC8854272"
},
{
"role": "modified_variant_and_malignancy_cohort",
"citation": "Caracciolo et al. AJR. 2016;207:150-156",
"doi": "10.2214/AJR.15.14368",
"pmid": "27070373"
},
{
"role": "current_reliability",
"citation": "Willenbring et al. BMC Musculoskelet Disord. 2025;26:649",
"doi": "10.1186/s12891-025-08815-5",
"pmid": "40616039",
"pmcid": "PMC12228233"
},
{
"role": "current_imaging_workup",
"citation": "ACR Appropriateness Criteria Suspected Primary Bone Tumors: 2024 Update",
"doi": "10.1016/j.jacr.2025.02.020",
"pmid": "40409893"
},
{
"role": "separate_incidental_CT_MRI_framework",
"citation": "Chang et al. Skeletal Radiol. 2022;51:1743-1764",
"doi": "10.1007/s00256-022-04022-8",
"pmcid": "PMC9283187"
}
],
"source_locator": "Lodwick et al. 1980, DOI 10.1148/radiology.134.3.6928321, original truth table; Benndorf et al. 2022, PMC8854272, Table 1, Figure 1, Pitfalls/Table 2 and Subsequent works for the exact ordered algorithm and variant boundary; Caracciolo et al. 2016 abstract for modified-variant cohort composition and malignancy associations; Willenbring et al. 2025 results for inter- and intraobserver agreement; ACR Suspected Primary Bone Tumors 2024 Update, variants 1-3 for current imaging-workup context."
}