Improving colon cancer Improving colon cancer prognosis using Index of prognosis using Index of Metastasis: Metastasis: A SEER analysis A SEER analysis George J. Chang, Chung-Yuan Hu, Miguel A. Rodriguez-Bigas, John M. Skibber The University of Texas, M.D. Anderson Cancer Center Houston, Texas Abstract 4019
17
Embed
Improving colon cancer prognosis using Index of Metastasis: A SEER analysis George J. Chang, Chung-Yuan Hu, Miguel A. Rodriguez-Bigas, John M. Skibber.
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
Improving colon cancer Improving colon cancer prognosis using Index of prognosis using Index of
George J. Chang, Chung-Yuan Hu, Miguel A. Rodriguez-Bigas, John M. Skibber
The University of Texas, M.D. Anderson Cancer CenterHouston, Texas
Abstract 4019
Disclosure Slide
Background
• Colon cancer affects approximately 107,000 new patients each year and is the second leading cause of cancer death among men and women.
• Survival and staging in localized colon cancer is related to the number of lymph nodes containing metastasis.• Related to the total number of lymph nodes evaluated.
• Fewer than one-half of cases of resected colon cancer have had an adequate LN evaluation based on a benchmark of 12 or more lymph nodes.
Limitations of AJCC 6th Edition• American Joint Committee on Cancer (AJCC)
TNM staging does not account for the cases with only few lymph nodes.
• Survival within an AJCC/TNM stage can range widely due to the heterogeneous nature of the sub-stage cohort and variations in surgical and pathologic quality.
Objective
To evaluate the predictive value of the Index of Metastasis (IM) for survival
among Stage III colon cancer patients and to define the boundaries of this
relationship.
Methods—Data Source and Patient Characteristics• Surveillance Epidemiology and End Results
(SEER 17) Program of the NCI• Patients diagnosed 1988 to 2005,
Methods—Statistical Analysis• Patients were stratified by the total number of LN (totLN)
evaluated and by IM quartile within each T stage category (T1-2, T3, or T4).
• Stratified Kaplan-Meier analysis was performed to compare cancer specific survival (CSS).
• Log-rank cut-point analysis was performed to determine degree of sensitivity of AJCC sub-staging and IM staging to the total number of lymph nodes evaluated.
• Data were analyzed using STATA Intercooled v.10.0 (rel. 2007, College Station, TX).
EthnicityWhite (incl. Hispanic)Black Asian (incl. Pac Island)Unknown or others
36,217 (80.5)4,698 (10.5)3,968 (8.8)
107 (0.2)
Median positive LN (IQR) 2 (1-4)
Median total LN (IQR) 12 (8-18)
Median Index of Metastasis (IQR)
.22 (.11-.45)
Total N= 44,990
Distribution of cases by TN stage and IM quartile
T1-2(N=4,513)
IM T3(N=31,575)
IM T4(N=8,877)
IM
IM Q1 1,134 8,442 2268
IM Q2 1,135 0.091 7,699 0.111 2251 0.133
IM Q3 1,221 0.167 7,654 0.222 2146 0.286
IM Q4 1,023 0.333 7,780 0.429 2212 0.563
N1 3,873 21,059 5099
N2 640 10,516 3778
IM Quartile distribution specific to each T stage group
Survival by AJCC Stage varies widely within each stage group and is affected by the total number LN examined
GROUP
5-yr Cancer-Related Survival (%)Stratified by totLN quartile and N stage
T1-2 (N=4,513) T3 (31,575) T4 (N=8,877)
<7 All* >15 <9 All >18 <9 All >18
N1 82.9 85.0 89.3 61.4 68.3 76.7 41.7 50.2 60.7
N2 82.1 71.9 69.6 41.6 47.5 52.0 24.5 29.0 35.2
Median Total LN examined and IQR = 9 (6, 15) for T1-2, 12 (8, 18) for T3, and 12(8, 18) for T4 tumors.* Total LN examined quartile 1 (0-25th percentile) specific to each T stage group (1-2, 3, or 4). † Entire cohort of patient within T stage group.‡ Total LN examined quartile 4 (75th-100th percentile) specific to each T stage group (1-2, 3, or 4).
Cancer-specific survival by total number LN examined quartile
0.00
0.25
0.50
0.75
1.00
0 50 100 150 200analysis time
totLN Q1 totLN Q2totLN Q3 totLN Q4
Kaplan-Meier survival estimates AJCC IIIB
73.3
56.1
0.00
0.25
0.50
0.75
1.00
0 50 100 150 200analysis time
totLN Q1 totLN Q2totLN Q3 totLN Q4
Kaplan-Meier survival estimates AJCC IIIC
48.7
39.0
T-stage stratified IM accurately predicts survival
GROUP
5-yr Cancer-Related Survival (%)Stratified by Total Number LN Evaluated (totLN) and IM quartile
Median Total LN examined and IQR = 9 (6, 15) for T1-2, 12 (8, 18) for T3, and 12(8, 18) for T4 tumors* Total LN examined quartile 1 (0-25th percentile) specific to each T stage group (1-2, 3, or 4). † Entire cohort of patient within T stage group.‡ Total LN examined quartile 4 (75th-100th percentile) specific to each T stage group (1-2, 3, or 4).
CSS among T3 patients by IM quartile
0.00
0.25
0.50
0.75
1.00
0 50 100 150 200analysis time
totLN Q1 totLN Q2totLN Q3 totLN Q4
Kaplan-Meier survival estimates T3 IM Q2
0.00
0.25
0.50
0.75
1.00
0 50 100 150 200analysis time
totLN Q1 totLN Q2totLN Q3 totLN Q4
Kaplan-Meier survival estimates T3 IM Q3
0.00
0.25
0.50
0.75
1.00
0 50 100 150 200analysis time
totLN Q1 totLN Q2totLN Q3 totLN Q4
Kaplan-Meier survival estimates T3 IM Q1
0.00
0.25
0.50
0.75
1.00
0 50 100 150 200analysis time
totLN Q1 totLN Q2totLN Q3 totLN Q4
Kaplan-Meier survival estimates T3 IM Q4
5 yr
AJCC6 IIIB
AJCC6 IIIC
25% of patients N156% of patients N1
Cut point analysis shows AJCC stage is highly sensitive to totLN
Stage IIIA Stage IIIB Stage IIIC IMQ1 for T1-2 IMQ2 for T1-2
IMQ3 for T1-2 IMQ4 for T1-2 IMQ1 for T3 IMQ2 for T3 IMQ3 for T3
IMQ4 for T3 IMQ1 for T4 IMQ2 for T4 IMQ3 for T4 IMQ4 for T4
IIIB
IIIC
IM Q4 for T3
Summary of results
• 5-year CSS for each AJCC stage group was dependent upon the total number of LN examined.
• Range of 5-year CSS within a single TN stage group varies based on the total number of LN examined.
• AJCC staging incorrectly predicts prognosis when few LN are evaluated
Summary of results
• Cut point analysis demonstrated survival prognosis by AJCC stage is very sensitive to the total number of LN examined.
• IM reliably predicts CSS and is not sensitive to the total number of LN examined.
Conclusions
• AJCC 6th ed. staging poorly predicts prognosis in an individual patient due in part to variability in the total number of LN examined at resection.
• When stratified by T stage, index of metastasis more accurately predicts survival and except at the highest IM levels, is not influenced by the number of LN examined.