NCT01507428 · OUTCOME

FDG SUVmax @ Baseline FDG PET/CT, to Predict Lung Cancer Cause-specific Survival and Optimal Threshold (ECOG-ACRIN)

Positron Emission Tomography and Computed Tomography in Guiding Radiation Therapy for Stage III Non-small Cell Lung Cancer · Source last updated 2026-04-21

Results reported by the registry submitting party. Population, time horizon, units and missing values must be read together.

Measure: MEAN
Unit: SUV
Interval / dispersion: Standard Deviation
Time frame: Baseline to up to 5 years

What was measured

SUVmax, as measured at Baseline, will be assessed in its ability to predict Lung Cancer Cause-specific Survival (LCCS) and an Optimal Threshold (OT) for differentiating responders from non-responders will be identified LCCS defined as NO Lung Cancer Cause-specific disease 5 years post registration (Responders) OT defined as the SUVmax value corresponding to the maximum Youden index

Analysis population: Number analyzed represents the denominator for the estimated value The sum of rows = Overall Number of Participants Analyzed

Groups in this outcome

Both Arms

Patients undergo chemoradiotherapy as defined for the arm of the study to which they were randomized. 18F-Fluoromisonidazole: Undergo FMISO PET/CT prior to therapy for patients enrolling at institutions participating in the FMISO correlative study

Arm I (Standard Chemoradiotherapy)

Patients undergo standard radiotherapy QD 5 days a week for 30 fractions. Patients also receive paclitaxel IV over 1 hour and carboplatin IV over 30 minutes once weekly for 6 weeks. Patients undergo FMISO-PET imaging at baseline 18F-Fluoromisonidazole: Undergo FMISO PET/CT prior to therapy for patients enrolling at institutions participating in the FMISO correlative study Carboplatin: Given IV External Beam Radiation Therapy: Undergo radiotherapy

Arm II (Experimental Chemoradiotherapy)

Patients undergo an individualized dose of image-guided radiotherapy QD 5 days a week for 30 fractions and undergo 18 F FDG-PET/CT between fractions 18 and 19. Based on the scan results, patients undergo individualized adaptive radiotherapy for the final 9 fractions. Patients also receive paclitaxel and carboplatin as in Arm I. 18F-Fluoromisonidazole: Participants enrolled at institutions participating in the FMISO correlative study undergo FMISO-PET prior to treatment Carboplatin: Given IV Computed Tomography: Undergo FDG PET/CT Computed Tomography: Undergo FMISO PET/CT (Correlative studies) Fludeoxyglucose F-18: Undergo FDG PET/CT Image-Guided Adaptive Radiation Therapy: Undergo individualized adaptive radiotherapy Laboratory Biomarker Analysis: Correlative studies Paclitaxel: Given IV Positron Emission Tomography: Undergo FDG PET/CT Positron Emission Tomography: Undergo FMISO PET/CT (Correlative studies)

Group identifiers and denominators belong to this outcome only. They may differ in another outcome or in safety reporting.

Analysis denominator · Participants

Group	Source count
Both Arms	114
Arm I (Standard Chemoradiotherapy)	40
Arm II (Experimental Chemoradiotherapy)	74

Reported measurements

Responder(No Event)

Analysis denominator · Participants

Group	Source count
Both Arms	51
Arm I (Standard Chemoradiotherapy)	17
Arm II (Experimental Chemoradiotherapy)	34

Values in SUV · Interval/dispersion: Standard Deviation
Group	Value	Spread	Lower	Upper	Comment
Both Arms	19.66	9.078	Not reported	Not reported	Not reported
Arm I (Standard Chemoradiotherapy)	20.55	7.846	Not reported	Not reported	Not reported
Arm II (Experimental Chemoradiotherapy)	19.22	9.717	Not reported	Not reported	Not reported

Non-Responder

Analysis denominator · Participants

Group	Source count
Both Arms	63
Arm I (Standard Chemoradiotherapy)	23
Arm II (Experimental Chemoradiotherapy)	40

Values in SUV · Interval/dispersion: Standard Deviation
Group	Value	Spread	Lower	Upper	Comment
Both Arms	17.42	9.021	Not reported	Not reported	Not reported
Arm I (Standard Chemoradiotherapy)	14.57	7.901	Not reported	Not reported	Not reported
Arm II (Experimental Chemoradiotherapy)	19.06	9.306	Not reported	Not reported	Not reported

Source statistical analyses

Group order, estimate direction, methods and comments are retained. No treatment ranking is inferred.

ci Lower Limit
0.4855
ci Num Sides
TWO_SIDED
ci Pct Value
95
ci Upper Limit
0.7000
group Description
Accuracy of SUVmax @ Baseline to Predict LCCS at year 5 as measured by the Area under the ROC curve (AUC/c-statistic) and Optimal Threshold separating responders and non-responders as measured by the Youden index
group Ids
OG000
non Inferiority Comment
The AUC/C-statistic is a measure of discrimination between responders and non-responders
non Inferiority Type
OTHER
param Type
AUC
param Value
0.5927
ci Num Sides
TWO_SIDED
group Description
Maximum Youden index
group Ids
OG000
non Inferiority Comment
Maximum Youden index is the point value corresponding tothe optimum threshold for distributional separation. Youden index (max( sensitivity + specificity-1)) over all SUVmax @ Baseline
non Inferiority Type
OTHER
param Type
Youdin index
param Value
0.2390
ci Num Sides
TWO_SIDED
group Description
The optimal SUVmax @ Baseline threshold for discriminating Positive and Negative cases as determined by the value having the maximum Youden index
group Ids
OG000
non Inferiority Comment
SUVmax @ Baseline @ Max(Youden index)
non Inferiority Type
OTHER
param Type
Optimal SUVmax @ Baseline threshold
param Value
17.3000
ci Lower Limit
0.5464
ci Num Sides
TWO_SIDED
ci Pct Value
95
ci Upper Limit
0.9063
group Description
Accuracy of SUVmax @ Baseline to Predict LCCS at year 5 as measured by the Area under the ROC curve (AUC/c-statistic) and Optimal Threshold separating responders and non-responders as measured by the Youden index
group Ids
OG001
non Inferiority Comment
The AUC/C-statistic is a measure of discrimination between responders and non-responders
non Inferiority Type
OTHER
param Type
AUC
param Value
0.7263
ci Num Sides
TWO_SIDED
group Description
Maximum Youden index
group Ids
OG001
non Inferiority Comment
Maximum Youden index is the point value corresponding tothe optimum threshold for distributional separation. Youden index (max( sensitivity + specificity-1)) over all SUVmax @ Baseline
non Inferiority Type
OTHER
param Type
Youdin index
param Value
0.5320
ci Num Sides
TWO_SIDED
group Description
The optimal SUVmax @ Baseline threshold for discriminating Positive and Negative cases as determined by the value having the maximum Youden index
group Ids
OG001
non Inferiority Comment
SUVmax @ Baseline @ Max(Youden index)
non Inferiority Type
OTHER
param Type
Optimal SUVmax @ Baseline threshold
param Value
18.2700
ci Lower Limit
0.3592
ci Num Sides
TWO_SIDED
ci Pct Value
95
ci Upper Limit
0.6298
group Description
Accuracy of SUVmax @ Baseline to Predict LCCS at year 5 as measured by the Area under the ROC curve (AUC/c-statistic) and Optimal Threshold separating responders and non-responders as measured by the Youden index
group Ids
OG002
non Inferiority Comment
The AUC/C-statistic is a measure of discrimination between responders and non-responders
non Inferiority Type
OTHER
param Type
AUC
param Value
0.4945
ci Num Sides
TWO_SIDED
group Description
Maximum Youden index
group Ids
OG002
non Inferiority Comment
Maximum Youden index is the point value corresponding tothe optimum threshold for distributional separation. Youden index (max( sensitivity + specificity-1)) over all SUVmax @ Baseline
non Inferiority Type
OTHER
param Type
Youdin index
param Value
0.0794
ci Num Sides
TWO_SIDED
group Description
The optimal SUVmax @ Baseline threshold for discriminating Positive and Negative cases as determined by the value having the maximum Youden index
group Ids
OG002
non Inferiority Comment
SUVmax @ Baseline @ Max(Youden index)
non Inferiority Type
OTHER
param Type
Optimal SUVmax @ Baseline threshold
param Value
17.3000

Complete source fields

analyses

ci Lower Limit
0.4855
ci Num Sides
TWO_SIDED
ci Pct Value
95
ci Upper Limit
0.7000
group Description
Accuracy of SUVmax @ Baseline to Predict LCCS at year 5 as measured by the Area under the ROC curve (AUC/c-statistic) and Optimal Threshold separating responders and non-responders as measured by the Youden index
group Ids
OG000
non Inferiority Comment
The AUC/C-statistic is a measure of discrimination between responders and non-responders
non Inferiority Type
OTHER
param Type
AUC
param Value
0.5927
ci Num Sides
TWO_SIDED
group Description
Maximum Youden index
group Ids
OG000
non Inferiority Comment
Maximum Youden index is the point value corresponding tothe optimum threshold for distributional separation. Youden index (max( sensitivity + specificity-1)) over all SUVmax @ Baseline
non Inferiority Type
OTHER
param Type
Youdin index
param Value
0.2390
ci Num Sides
TWO_SIDED
group Description
The optimal SUVmax @ Baseline threshold for discriminating Positive and Negative cases as determined by the value having the maximum Youden index
group Ids
OG000
non Inferiority Comment
SUVmax @ Baseline @ Max(Youden index)
non Inferiority Type
OTHER
param Type
Optimal SUVmax @ Baseline threshold
param Value
17.3000
ci Lower Limit
0.5464
ci Num Sides
TWO_SIDED
ci Pct Value
95
ci Upper Limit
0.9063
group Description
Accuracy of SUVmax @ Baseline to Predict LCCS at year 5 as measured by the Area under the ROC curve (AUC/c-statistic) and Optimal Threshold separating responders and non-responders as measured by the Youden index
group Ids
OG001
non Inferiority Comment
The AUC/C-statistic is a measure of discrimination between responders and non-responders
non Inferiority Type
OTHER
param Type
AUC
param Value
0.7263
ci Num Sides
TWO_SIDED
group Description
Maximum Youden index
group Ids
OG001
non Inferiority Comment
Maximum Youden index is the point value corresponding tothe optimum threshold for distributional separation. Youden index (max( sensitivity + specificity-1)) over all SUVmax @ Baseline
non Inferiority Type
OTHER
param Type
Youdin index
param Value
0.5320
ci Num Sides
TWO_SIDED
group Description
The optimal SUVmax @ Baseline threshold for discriminating Positive and Negative cases as determined by the value having the maximum Youden index
group Ids
OG001
non Inferiority Comment
SUVmax @ Baseline @ Max(Youden index)
non Inferiority Type
OTHER
param Type
Optimal SUVmax @ Baseline threshold
param Value
18.2700
ci Lower Limit
0.3592
ci Num Sides
TWO_SIDED
ci Pct Value
95
ci Upper Limit
0.6298
group Description
Accuracy of SUVmax @ Baseline to Predict LCCS at year 5 as measured by the Area under the ROC curve (AUC/c-statistic) and Optimal Threshold separating responders and non-responders as measured by the Youden index
group Ids
OG002
non Inferiority Comment
The AUC/C-statistic is a measure of discrimination between responders and non-responders
non Inferiority Type
OTHER
param Type
AUC
param Value
0.4945
ci Num Sides
TWO_SIDED
group Description
Maximum Youden index
group Ids
OG002
non Inferiority Comment
Maximum Youden index is the point value corresponding tothe optimum threshold for distributional separation. Youden index (max( sensitivity + specificity-1)) over all SUVmax @ Baseline
non Inferiority Type
OTHER
param Type
Youdin index
param Value
0.0794
ci Num Sides
TWO_SIDED
group Description
The optimal SUVmax @ Baseline threshold for discriminating Positive and Negative cases as determined by the value having the maximum Youden index
group Ids
OG002
non Inferiority Comment
SUVmax @ Baseline @ Max(Youden index)
non Inferiority Type
OTHER
param Type
Optimal SUVmax @ Baseline threshold
param Value
17.3000

classes

categories
measurements
group Id
OG000
spread
9.078
value
19.66
group Id
OG001
spread
7.846
value
20.55
group Id
OG002
spread
9.717
value
19.22
denoms
counts
group Id
OG000
value
51
group Id
OG001
value
17
group Id
OG002
value
34
units
Participants
title
Responder(No Event)
categories
measurements
group Id
OG000
spread
9.021
value
17.42
group Id
OG001
spread
7.901
value
14.57
group Id
OG002
spread
9.306
value
19.06
denoms
counts
group Id
OG000
value
63
group Id
OG001
value
23
group Id
OG002
value
40
units
Participants
title
Non-Responder

denoms

counts
group Id
OG000
value
114
group Id
OG001
value
40
group Id
OG002
value
74
units
Participants

description

dispersion Type

Standard Deviation

groups

description
Patients undergo chemoradiotherapy as defined for the arm of the study to which they were randomized. 18F-Fluoromisonidazole: Undergo FMISO PET/CT prior to therapy for patients enrolling at institutions participating in the FMISO correlative study
id
OG000
title
Both Arms
description
Patients undergo standard radiotherapy QD 5 days a week for 30 fractions. Patients also receive paclitaxel IV over 1 hour and carboplatin IV over 30 minutes once weekly for 6 weeks. Patients undergo FMISO-PET imaging at baseline 18F-Fluoromisonidazole: Undergo FMISO PET/CT prior to therapy for patients enrolling at institutions participating in the FMISO correlative study Carboplatin: Given IV External Beam Radiation Therapy: Undergo radiotherapy
id
OG001
title
Arm I (Standard Chemoradiotherapy)
description
Patients undergo an individualized dose of image-guided radiotherapy QD 5 days a week for 30 fractions and undergo 18 F FDG-PET/CT between fractions 18 and 19. Based on the scan results, patients undergo individualized adaptive radiotherapy for the final 9 fractions. Patients also receive paclitaxel and carboplatin as in Arm I. 18F-Fluoromisonidazole: Participants enrolled at institutions participating in the FMISO correlative study undergo FMISO-PET prior to treatment Carboplatin: Given IV Computed Tomography: Undergo FDG PET/CT Computed Tomography: Undergo FMISO PET/CT (Correlative studies) Fludeoxyglucose F-18: Undergo FDG PET/CT Image-Guided Adaptive Radiation Therapy: Undergo individualized adaptive radiotherapy Laboratory Biomarker Analysis: Correlative studies Paclitaxel: Given IV Positron Emission Tomography: Undergo FDG PET/CT Positron Emission Tomography: Undergo FMISO PET/CT (Correlative studies)
id
OG002
title
Arm II (Experimental Chemoradiotherapy)

param Type

MEAN

population Description

Number analyzed represents the denominator for the estimated value The sum of rows = Overall Number of Participants Analyzed

reporting Status

POSTED

time Frame

Baseline to up to 5 years

title

FDG SUVmax @ Baseline FDG PET/CT, to Predict Lung Cancer Cause-specific Survival and Optimal Threshold (ECOG-ACRIN)

type

SECONDARY

unit Of Measure

SUV

Download exact source JSON → · Snapshot ctgov-results-191f4516d4760045304b24a0

Preserved source evidence · Not medical advice