Impact of Tumor Location on Nodal

Evaluation for Colon Cancer

Karl Y. Bilimoria, M.D.,1,2  Bryan Palis, M.A.,2  Andrew K. Stewart, M.A.,2 
David J. Bentrem, M.D.,1  Andrew C. Freel, M.D.,2 
ORIGINAL Elin R. Sigurdson, M.D., Ph.D.,3  Mark S. Talamonti, M.D.,4 
CONTRIBUTION Clifford Y. Ko, M.D., M.S., M.S.H.S.2,5
1 Department of Surgery, Feinberg School of Medicine, Northwestern University, Chicago, IL
2 Cancer Programs, American College of Surgeons, Chicago, IL
3 Department of Surgery, Fox Chase Cancer Center, Philadelphia, PA
4 Department of Surgery, Evanston Northwestern Healthcare, Chicago, IL
5 Department of Surgery, University of California, Los Angeles and VA Greater Los Angeles Healthcare System,
Los Angeles, CA

PURPOSE: Adequate lymph node evaluation is important
to stage colon cancers and make adjuvant treatment
decisions. Studies have demonstrated improved survival
when ≥ 12 nodes are examined. Our objective was to
assess differences in the adequacy of nodal evaluation for
right vs. left colon cancers.
METHODS: From the National Cancer Data Base (1998–
2004), 142,009 N0M0 colon cancer patients were identi-
fied. Logistic regression was used to evaluate the number
of nodes examined for right vs. left colectomies. Multi-
variable modeling was used to determine the impact of
examining ≥ 12 nodes on survival.
RESULTS: Of 142,009 patients, 79,444 (56 percent) had
right colectomies, and 62,565 (44 percent) patients had
left colectomies. More nodes were examined during right
colectomies than left (median 12 vs. 8, P<0.0001). When
adjusted for patient, tumor, and hospital factors,
patients undergoing left colectomy were less likely to have
≥ 12 nodes identified (P<0.0001). Patients were more
likely to have ≥ 12 nodes identified for right and left colon
cancers at high-volume hospitals. Survival was better
with examination of ≥ 12 nodes for right and left colon
cancers (P<0.0001).
CONCLUSIONS: Evaluating ≥ 12 nodes for right and left
colon cancers is a feasible, clinically relevant, and modifi-
able factor that will likely improve patient outcomes.
KEY WORDS: Colon neoplasms; Surgery; Colectomy;
National Cancer Data Base; Lymph nodes; Staging.
Presented in part at the meeting of The American Society of Colon and
Rectal Surgeons, St. Louis, Missouri, June 2 to 7, 2007.
Supported by a grant from the National Cancer Institute. Dr. Bilimoria
is supported by the American College of Surgeons, Clinical Scholars in
Residence program, and a research grant from the Department of
Surgery, Feinberg School of Medicine, Northwestern University.
Address of correspondence: Karl Y. Bilimoria, M.D., American College
of Surgeons, National Cancer Data Base, Cancer Programs, 636 N. St.
Clair Street, 25th Floor, Chicago, IL 60611.
154 DOI: 10.1007/s10350-007-9114-2  VOLUME 51: 154–161 (2008)  ©THE ASCRS. 2007
egional lymph node evaluation is critical to accurately
R stage patients with colon cancer.1 Nodal involvement
has been shown to be the most important prognostic
factor affecting long-term survival in patients without
distant metastases.2 Moreover, nodal status guides adju-
vant chemotherapy decisions, and with the demonstration
of highly effective systemic therapies during the last de-
cade, it is paramount to ensure that all appropriate patients
receive adjuvant chemotherapy.2
Numerous studies and a recent structured review have
demonstrated an improvement in disease-free and overall
survival when increasing numbers of lymph nodes are
examined for colon cancer.3–13 Multiple factors contribute
to this effect on survival, but the improvement is likely, in
part, the result of better staging resulting in the increased use
of adjuvant chemotherapy. Alternatively, nodal evaluation
could be a proxy for better surgical technique, which may
improve survival. Studies have attempted to quantify the
minimum number of nodes that need to be evaluated to
declare a patient “node-negative” with a reasonable degree
of certainty. The minimum number of nodes has varied
from 6 to 40; however, the majority of reports and
consensus guidelines have demonstrated that 12 regional
lymph nodes are a reasonable minimum for adequate nodal
examination.1,5,11–17 Moreover, the American College of Sur-
geons, National Comprehensive Cancer Network (NCCN),
and the American Society of Clinical Oncology (ASCO) have
recently adopted the 12 Node Measure for colon cancer with
endorsement from the National Quality Forum (NQF) as
quality surveillance activity.
Differences have been demonstrated in lymph node ex-
amination rates for right compared with left colectomies.18
However, the reason for this finding is poorly understood.
Failure to evaluate 12 regional lymph nodes is a multifac-
torial problem involving the surgeon, pathologist, patient,
and hospital. In addition, the location of the tumor within
the colon may be a contributing factor. Using a large nation-
ally generalizable sample of N0M0 patients from the National
Cancer Data Base (NCDB), our objectives were 1) to assess
differences in regional lymph node evaluation for right vs. left
colon cancers, 2) to determine the effect of patient and hos-
B ILIMORIA ET AL .: NODAL E VALUATION FOR C OLON C ANCER
pital factors on the adequacy of nodal evaluation (examina-
tion of ≥ 12 nodes) for right vs. left colectomies, and 3) to
determine whether examining ≥ 12 nodes is associated with
improved survival for right and left colon cancers.
PATIENTS AND METHODS
Data Acquisition and Patient Selection
The NCDB is a program of the American College of
Surgeons and the Commission on Cancer (CoC).19 The
NCDB has been collecting data on newly diagnosed cancers
since 1985 and now contains data for more than 20 million
patients from more than 1,440 hospitals. The NCDB collects
data regarding patient demographics, tumor characteristics,
preoperative and postoperative staging, treatment details
(including both adjuvant radiation and chemotherapy),
recurrence, and survival for the purposes of research and
quality improvement. Health systems information and
socioeconomic status are obtained by linking to tertiary
data sources. The NCDB captures approximately 75 percent
of all new cancer diagnoses in the United States annually.
Based on national incidence estimates from the American
Cancer Society, the NCDB captures 67 percent of all new
colon cancers each year.20
From the NCDB (1998–2004), we identified patients
who underwent a colectomy for colon cancer using
International Classification of Disease – Oncology, 2nd
and 3rd editions (ICD-O-2/3) codes for colon cancer
(C18.0, C18.2, C18.3, C18.5, C18.6, C18.7).21 At the time
of this study, 2004 was the most recent year with complete
data on diagnosis, and 1999 was the most recent year with
follow-up data on five-year survival. According to Registry
Operations and Data Standards (ROADS) and Facility
Oncology Registry Data Standards (FORDS) site-specific
procedure coding, patients were limited to those who
underwent a partial or hemicolectomy, specifically exclud-
ing patients who had local procedures (i.e., polypectomy)
or total colectomies.22,23 Right hemicolectomies included
tumors of the cecum, ascending colon, and hepatic flexure.
Left hemicolectomies included descending and sigmoid
colon resections. To minimize confounding, patients were
excluded if they had a histology other than adenocarcino-
ma, appendiceal lesions, transverse colon tumors, nodal
or distant metastases, or underwent neoadjuvant chemo-
therapy. Patients younger than aged 18 years also were
excluded. Rectal cancers were not examined in this study.
Statistical Analysis
Descriptive statistics were calculated for all variables. Con-
tinuous variables were compared by using independent-
sample t-tests. Categorical variables were analyzed by using
the chi-squared test. The Bonferroni correction was used
for multiple comparisons. Medians were compared with the
Kruskal-Wallis and the Mann-Whitney U tests. Adequate
nodal evaluation was defined as having ≥ 12 regional lymph
155
nodes removed and pathologically examined. Multivari-
able logistic regression was used to determine whether
nodal evaluation (≥12 nodes vs. < 12 nodes) differed for
right and left colon cancers when adjusting for potential
confounders including gender, age (<55, 55–65, 65–75,
76–85, >85 years), race (white, black, Asian, Hispanic), me-
dian income quartiles, percent with college degree quartiles,
T stage (1–4), grade (well vs. poorly differentiated), adjuvant
chemotherapy administration, year of diagnosis, and hos-
pital volume quartiles. Odds ratios (OR) with 95 percent
confidence intervals (CI) were generated. The Hosmer-
Lemeshow Goodness of Fit test was assessed to assess the fit
of the model. Procedure volume quartiles were calculated
by ranking all hospitals in order of increasing average
annual hospital volume of colectomies, and whole number
cut points were chosen that most closely sorted patients
into four even groups (quartiles). Since patient-level socio-
economic status is not collected by the NCDB, median
household income and education (percent of patients with
college degrees) were assessed at the zip-code level based on
the patient’s residence at the time of diagnosis by using 2000
United States Census Bureau data.24 Median income and
education level were inserted into the models separately
because there is a high degree of collinearity between them.
Next, logistic regression was used to assess factors associated
with the adequacy of nodal evaluation in right vs. left colon
cancers while adjusting for potential confounding patient,
tumor, and hospital factors (as in logistic regression model
described earlier).
Survival Analysis
Five-year survival rates were calculated as the time from
surgery to death or last contact. Patients diagnosed in
1998 and 1999 were used in the survival analyses because
these patients had at least five years of follow-up data.
Median follow-up was 40 months. Relative overall survival
was calculated by using SPSS, version 14 (Chicago, IL) by
adjusting the observed survival rates for differences in
gender, age, and race/ethnicity using 2000 United States
Census Bureau data.24 Relative survival currently serves as
the best estimate of disease-specific survival using data from
cancer registries. Survival was estimated by the Kaplan-
Meier method and compared by using the log-rank test. Cox
proportional hazards modeling was used to evaluate the
effect of nodal evaluation on survival in left and right colon
cancers while adjusting for potential confounders, including
patient, tumor, treatment, and hospital factors (as in logistic
regression model described earlier). Before analysis, all
variables in the model were examined for collinearity. The
proportional hazards assumptions were confirmed graph-
ically. Hazard ratios (HR) with 95 percent confidence inter-
vals (CI) were generated. Hazard ratios less than 1.0 indicate
a decreased risk of death.
The logistic regression and Cox models accounted for
clustering of outcomes within hospitals using Intercooled
156 BILIMORIA ET AL .: N ODAL E VALUATION FOR C OLON CANCER

Stata, version 9.0 (College Station, TX). The level of statistical Table 2. Comparison of the differences between right and left
significance was set to P<0.05. All P values reported are colon cancers in median number of nodes examined and the
two-tailed. Statistical analyses were performed using SPSS®, proportion of patients undergoing adequate lymph node
evaluation (≥12 Nodes)
version 14 (SPSS, Inc., Chicago, IL).
Right colon Left colon
cancers cancers P value
RESULTS Median no. of 12 (8–18) 8 (4–13) <0.0001
From the NCDB, we identified 142,009 N0M0 patients nodes examined
Patients with ≥12 41,291 (54) 19,503 (32.6) <0.0001
who underwent colectomy for cancer. Patient character- nodes examined
istics for right vs. left-sided colon cancers are shown in
Data are numbers with interquartile ranges or percentages in parentheses unless
Table 1. Right-sided colon cancers were more frequently otherwise indicated.
seen in females, whereas left-sided lesions were more com-
mon in males (P<0.0001). The median age of diagnosis was
higher for right vs. left colon cancers (75 vs. 70 years, P< and eight nodes for malignancies of the sigmoid colon
0.0001). Patients with left-sided lesions were diagnosed at (P<0.0001; Fig. 2). Cecal and hepatic flexure colon cancers
earlier T stages than patients with right-sided colon cancers had a median of 12 nodes examined, and ascending colon
(P<0.0001). Right-sided tumors demonstrated poor differ- tumors had a median of 13 nodes examined. No nodes
entiation more frequently than left-sided tumors (P<0.0001). were examined in 1,383 (1.8 percent) right colon cancers
Patients with right-sided colon cancers had more nodes and 3,477 (5.8 percent) left colon cancers.
examined than patients with left-sided colon cancers Adequate nodal evaluation was defined as having ≥
(median 12 vs. 8 nodes, P<0.0001; Table 2). From 1998– 12 regional lymph nodes examined. Patients undergoing
1999 to 2003–2004, the median number of nodes exam- left hemicolectomies underwent adequate nodal evalua-
ined increased for right colon cancers (from 11 to 13; P< tion less frequently than those undergoing right hemi-
0.0001) and for left colon cancers (from 7 to 9; P<0.0001; colectomies: 32.6 vs. 54 percent (P<0.0001) (Table 2).
Fig. 1). Because there may be variability in the length of Using a multivariable logistic regression model adjusting
colon removed for left colon cancers, the segments of the for patient, tumor, and hospital characteristics, we eval-
left colon were examined separately: ten nodes for splenic uated the likelihood of having ≥ 12 nodes examined for
flexure lesions, nine nodes for descending colon cancers, right vs. left colon cancers by including the tumor location
as a covariate. Compared with patients with right colon
cancers, those with left-sided lesions were less likely to
Table 1. Patient characteristics by location of cancer within have ≥ 12 nodes examined when adjusted for potential
the colon for 142,009 patients who underwent colectomy for
cancer
FIGURE 1. Change in median number of nodes examined for right
Right colon Left colon compared to left colon cancers over time.
cancers cancers
No. of patients 79,444 62,565 Median Number of
Gender Nodes Examined
Female 44,554 (56.1) 29,868 (47.8) 14 p < 0.0001*
Male 34,863 (43.9) 32,666 (52.2) 1998–1999
Age (yr) 2000–2002
Median (interquartile 75 (66–82) 70 (60–78) 12 2003–2004
range)
Race
10 p < 0.0001*
White 66,607 (83.8) 51,378 (82.1)
Black 7,855 (9.9) 6,217 (9.9)
Asian 1,171 (1.5) 1,553 (2.5) 8
Hispanic 2,645 (3.3) 2344 (3.7)
Other 1,116 (1.5) 1,073 (1.7)
T stage 6
1 11,309 (14.4) 14,449 (23.7)
2 19,533 (24.9) 12,914 (21.2)
3 43,663 (55.7) 30,227 (49.6) 4
4 3,919 (5) 3,356 (5.5)
Grade
Well to moderately 63,514 (83.5) 53,798 (91.7) 2
differentiated
Poorly differentiated 12,531 (16.5) 4,868 (8.3)
Adjuvant chemotherapy 9,596 (12.1) 8,833 (14.1) Right Colon Left Colon
Data are numbers with percentages in parentheses unless otherwise indicated. * Significance comparing any pair of 1998–1999, 2000–2002, or 2003–2004.
B ILIMORIA ET AL .: NODAL E VALUATION FOR C OLON C ANCER
Median Number of
Lymph Nodes Examined
14
12
10
8
6 Adequate nodal evaluation of colon cancers allows
4 significant prognostic factor in patients with nonmeta-
2
Cecum Ascending Hepatic Splenic Descending Sigmoid
Flexure Flexure
FIGURE 2. Median number of lymph nodes examined by tumor
location within the colon.
confounding factors (OR, 0.41; 95 percent CI, 0.4–0.42;
P<0.0001).
Using multivariable logistic regression model, we
evaluated factors predicting failure to have ≥ 12 nodes
examined by stratifying the analysis into right vs. left
hemicolectomies (Table 3). Regardless of whether patients
underwent a right or left hemicolectomy, they were less
likely to have ≥ 12 nodes examined if they were male, older
than aged 55 years, had earlier T-stage tumors, had lower
median incomes, or had less education (P<0.05). Race did
not affect the adequacy of nodal evaluation. Patients were
more likely to have ≥ 12 nodes examined if they received
their care at high-volume hospitals as opposed to low-
volume institutions (P<0.0001). The factors that affected
the adequacy of nodal evaluation were the same for left and
right-sided cancers except for tumor grade. Patients with
poorly differentiated tumors of the left colon were more
likely to have ≥ 12 nodes examined, whereas grade did not
affect the likelihood of an adequate lymph node evaluation
for right-sided cancers.
To assess the impact of evaluating ≥ 12 nodes on
outcomes, univariate relative survival was estimated. Sur-
vival was better for patients regardless of the side of the
lesion when ≥ 12 nodes were examined (P<0.0001; Fig. 3).
When adjusted for patient, tumor, treatment, and hospital
characteristics by using a multivariable Cox proportional
hazards model, patients who had ≥ 12 nodes examined
experienced better survival than those with < 12 nodes
examined regardless of whether the patient underwent a
right hemicolectomy (HR, 0.75; 95 percent CI, 0.71–0.8; P<
0.0001) or left hemicolectomy (HR, 0.8; 95 percent CI,
157
0.74–0.86; P<0.0001; Table 4). Males, increasing age, black
race, advanced T stage, and living in a low median income
area were associated with an increased risk of death. Patients
treated at lowest volume hospitals had a modest increased
risk of death compared with patients treated at highest
volume hospitals. Chemotherapy was associated with a
significant reduction in the risk of death in these patients
with N0M0 colon cancer.
DISCUSSION
accurate staging.15 Lymph node involvement is the most
static disease. Although the precise definition of the
number of nodes needed to confidently deem a patient
“node negative” is debatable, consensus guidelines rec-
ommend that ≥ 12 nodes be examined.1,15–17 At best, 50
percent of patients in the United States undergo adequate
lymph node evaluation for colon cancer.6,18 Location of
the cancer within the colon has been associated with
inadequate nodal evaluation; however, the reason for this
is poorly understood.18 Thus, we sought to examine the
adequacy of lymph node evaluation for lesions in the right
compared with the left colon and to determine whether
the 12 node measure could be applied to both left and
right colon malignancies.
Our first objective was to determine differences in the
number of nodes examined after right vs. left hemicolec-
tomies. Our results demonstrate that the median number of
nodes examined was significantly lower for left compared
with right colon cancers (8 vs. 12). Furthermore, although
the median number of nodes examined increased from 1998
to 1999 until 2003 to 2004, differences in nodal evaluation
between right and left colon cancers persisted over time.
Baxter et al. demonstrated that patients with left colon
cancers were 55 percent less likely to have ≥ 12 nodes
examined.18 We found that patients with left-sided colon
cancers were 59 percent less likely to undergo adequate
lymph node evaluation. One important difference is that
our study was able to adjust for hospital volume, whereas
the previous study utilized the National Cancer Institute’s
Surveillance Epidemiology and End Results (SEER) dataset
which does not contain information on hospitals, and thus
they were unable to adjust for differences at the facility level.
Our second objective was to determine whether
patient or hospital factors affected the adequacy of nodal
evaluation for right vs. left colectomies. Baxter et al. found
that those older than aged 51 years and males were less
likely to have ≥ 12 nodes examined. Previous studies have
not examined factors affecting lymph node examination
by the location within the colon; however, we found no
difference for right and left colon cancers in the factors
associated with adequate lymph node evaluation except
for a relatively modest difference by tumor grade.
158 BILIMORIA ET AL .: N ODAL E VALUATION FOR C OLON CANCER

Table 3. Logistic regression analysis (stratified by location of the cancer within the colon) of patient and hospital factors associated
with failure to have ≥12 nodes evaluated. Odds ratios less than one indicate a lower likelihood of having ≥12 nodes evaluated
Unadjusted proportion of patients who Adjusted odds ratio
underwent adequate nodal evaluation (95 percent confidence interval)
Right colon Left colon
cancers (percent) cancers (percent) Right colon cancers Left colon cancers
Gender
Female 54.8 33.7 1 (Referent) 1 (Referent)
Male 53 31.7 0.89 (0.87–0.92) 0.91 (0.87–0.94)
Age (yr)
<55 67.8 40.1 1 (Referent) 1 (Referent)
55-65 58.6 32.7 0.7 (0.65–0.75) 0.74 (0.69–0.79)
66-75 52.8 30.7 0.55 (0.51–0.59) 0.65 (0.6–0.7)
76-85 51.2 30.9 0.51 (0.47–0.55) 0.62 (0.58–0.67)
>85 48.2 29.8 0.42 (0.39–0.46) 0.53 (0.48–0.59)
Race
White 53.5 32.4 1 (Referent) 1 (Referent)
Black 55.5 32.7 0.95 (0.89–1.06) 0.93 (0.86–1)
Asian 60.1 36.3 1.15 (0.99–1.33) 1.14 (1–1.29)
Hispanic 55 34.3 0.98 (0.89–1.08) 1.24 (1.06–1.44)
T stage
1 40.9 18.1 1 (Referent) 1 (Referent)
2 50.8 29.7 1.54 (1.34–1.64) 1.94 (1.81–2.08)
3 58.9 40.8 2.11 (2–2.22) 3.22 (3.03–3.42)
4 56 41.2 1.79 (1.57–1.7) 3.15 (2.85–3.49)
Grade
Well to moderately differentiated 53.5 33.3 1 (Referent) 1 (Referent)
Poorly differentiated 59.4 35.4 0.94 (0.87–1.01) 1.17 (1.11–1.23)
Median income
75–99th percentile 56.8 34.7 1 (Referent) 1 (Referent)
50–74th percentile 53.2 31.6 0.9 (0.86–0.94) 0.9 (0.85–0.95)
25–49th percentile 51 31.1 0.81 (0.77–0.85) 0.87 (0.82–0.92)
0–24th percentile 51.5 30.8 0.81 (0.77–0.86) 0.82 (0.76–0.88)
Percent with college degree
75–99th percentile 56.9 34.6 1 (Referent) 1 (Referent)
50–74th percentile 52.4 31.8 0.83 (0.79–0.86) 0.89 (0.82–0.92)
25–49th percentile 51.1 29.9 0.77 (0.74–0.81) 0.73 (0.69–0.78)
0–24th percentile 50.5 30.9 0.73 (0.69–0.78) 0.68 (0.61–0.67)
Hospital volume quartiles
Highest volume 59.7 37 1 (Referent) 1 (Referent)
High volume 56.2 34.1 0.86 (0.82–0.9) 0.87 (0.82–0.92)
Moderate volume 50.9 31 0.8 (0.76–0.84) 0.73 (0.69–0.78)
Low volume 49 28.7 0.76 (0.71–0.8) 0.68 (0.64–0.73)

Adequacy of nodal evaluation was associated with gender,
age, income, and education. The degree of tumor differ-
entiation only affected the adequacy of nodal evaluation for
left-sided tumors. Importantly, we found that patients
treated at high-volume centers were more likely to have an
adequate nodal evaluation than patients seen at low-
volume centers. In a study of 324 patients from the North
Carolina Cancer Registry, Miller et al.25 found that high-
volume hospitals were more likely to examine seven nodes
or more than low-volume centers. Thus, higher rates of
adequate lymph node examination may underlie the better
long-term outcomes at high-volume hospitals.26–29
Our third objective was to determine the impact of
adequate nodal evaluation on survival for right and left
colon cancers. Numerous studies have demonstrated that
increasing numbers of lymph nodes evaluated is associated
with an improvement in survival.4–13 The basis for this
finding is unclear but is likely, in part, a result of better
staging resulting in a higher proportion of node-positive
patients receiving chemotherapy. Adequacy of lymph node
evaluation also may be surrogate for the quality of surgical
technique, perioperative care, and postoperative cancer
surveillance activities. The precise number of nodes that
need to be examined to declare a patient “node negative” is
uncertain. Studies have suggested a wide range for an
adequate nodal evaluation threshold, ranging from 6 to 40.
However, consensus guidelines from the National Compre-
hensive Cancer Network (NCCN), College of American
Pathologists (CAP), National Cancer Institute (NCI), and
the American Joint Committee on Cancer (AJCC) all
B ILIMORIA ET AL .: NODAL E VALUATION FOR C OLON C ANCER 159

a groups, particularly T1 tumors. However, we did not exam-

ine T-stage groups separately because there is little clinical
Cumulative
Survival Rate relevance to such an analysis. It is impossible to know pre-
100 operatively whether the lesion is a T1 tumor, and thus the
12 Node Measure is applicable regardless of T stage. Finally,
80 it is unclear why a small proportion of patients with N0M0
tumors received adjuvant radiation treatment. Radiation
treatment is not associated with survival in these patients,
60
and the inclusion or exclusion of radiation treatment in the
survival models did not affect the hazard ratios for the other
40
covariates in the analysis.
P<0.0001
0 1 2 3 4 5
Years
b
Table 4. Multivariable Cox proportional hazards model
Cumulative
Survival Rate
assessing the effect on survival of having ≥12 nodes evaluated
for colon cancer (compared with <12 nodes) for right vs. left
100
colon cancers
Hazard ratio
80
(95 percent confidence interval)
Right Left
60
Nodes examined
≥12 0.8 (0.74–0.86) 0.75 (0.71–0.8)
40 <12 1 (Referent) 1 (Referent)
P<0.0001 Gender
Male 1.24 (1.18–1.31) 1.19 (1.12–1.27)
0 1 2 3 4 5 Female 1 (Referent) 1 (Referent)
Years from Diagnosis Age (yr)
<55 1 (Referent) 1 (Referent)
FIGURE 3. Five-year relative survival after resection for colon cancer 55–65 1.72 (1.43–2.06) 1.47 (1.25–1.74)
based on whether ≥ 12 nodes or < 12 nodes were examined. 66–75 2.9 (2.46–3.41) 2.82 (2.43–3.28)
a. Right-sided colon cancers. b. Left-sided colon cancers. 76–85 5.08 (4.33–5.95) 4.82 (4.16–5.57)
>85 9.21 (7.83–10.84) 9.1 (7.8–10.63)
Race/ethnicity
White 1 (Referent) 1 (Referent)
recommend that ≥ 12 nodes be examined because this is a Black 1.21 (1.11–1.33) 1.31 (1.18–1.44)
feasible and effective threshold.1,15–17 Our results demon- Hispanic 0.88 (0.68–1.14) 0.86 (0.7–1.04)
Asian 0.79 (0.67–0.94) 0.71 (0.55–0.92)
strate that survival is improved for left and right colon Other 0.89 (0.69–1.16) 0.82 (0.6–1.11)
cancers with examination of ≥ 12 nodes. Moreover, the T stage
magnitude of the benefit (20–25 percent decreased risk of 1 1 (Referent) 1 (Referent)
death) is statistically similar for right and left colon cancers. 2 1.09 (0.98–1.2) 1.43 (1.28–1.61)
Thus, examination of 12 nodes is important to adequately 3 1.4 (1.28–1.53) 2.07 (1.88–2.29)
4 2.9 (2.57–3.28) 3.7 (3.24–4.23)
stage malignancies and improve outcomes of both right and Median income
left colon cancers. 75–99th percentile 1 (Referent) 1 (Referent)
This study has some potential limitations. First, the 50–74th percentile 1.07 (1–1.14) 1.04 (0.96–1.12)
study population was limited to N0M0 patients to avoid the 25–49th percentile 1.08 (1–1.16) 1.03 (0.94–1.12)
confounding effect of more extensive pathologic examina- 0–24th percentile 1.14 (1.05–1.24) 1.18 (1.07–1.29)
Hospital volume
tion after an involved or suspicious lymph node is detected. quartiles
Moreover, adjuvant treatments have less of a role in N0M0 Highest 1 (Referent) 1 (Referent)
disease; thus, the potential confounding effect of these High 1.1 (1.02–1.18) 0.97 (0.89–1.06)
therapies on survival analyses is avoided. Second, hospitals Moderate 1.06 (0.98–1.14) 1.03 (0.94–1.12)
that report to the NCDB must meet the criteria set forth by Low 1.11 (1.03–1.19) 1.05 (0.96–1.14)
Adjuvant
the CoC; thus, low-volume hospitals reporting to the NCDB chemotherapy
may exhibit a higher level of oncologic specialization than Administered 0.73 (0.66–0.81) 0.6 (0.59–0.73)
typical low-volume hospitals in the United States that do not None 1 (Referent) 1 (Referent)
report to the NCDB. Third, some have suggested that the 12 Hazard ratios less than 1.0 indicate a lower risk of death with examination of ≥12
Node Measure may not be appropriate for all T-stage nodes.
160
There are widespread efforts underway to develop
surgical quality measures that assess quality at the hospital
level. The American College of Surgeons, NCCN, and ASCO
have developed and harmonized a quality surveillance
measure that requires examination of ≥ 12 nodes for
nonmetastatic colon cancer. Moreover, the National Quality
Forum (NQF) has recently endorsed this 12 Node Measure.
Currently this is only meant to be a surveillance measure,
and there is no associated accountability at the hospital or
physician level. To assess hospital concordance with these
widely accepted recommendations, regional lymph node
evaluation will be examined at the hospital level for the
∼1,440 facilities reporting to the National Cancer Data Base.
The process is meant to increase awareness of the impor-
tance of adequate nodal evaluation, allow comparison of
hospital performance, and further our understanding of the
importance of nodal evaluation for colon cancer. Partici-
pating hospitals will be able to compare their performance
to the other facilities. Importantly, hospitals with outlying
nodal evaluation rates can be identified and notified in
hopes of increasing lymph node evaluation rates at that
facility.
There has been uncertainty among surgeons regarding
the applicability and equity of this quality measure for left
colon cancers. Some have suggested that fewer nodes are
present in the left colon mesentery than on the right and that
12 nodes may not be important or feasible for left-sided
colon cancers. Similarly, hospitals have suggested that nodal
counts may not be a fair measure by which to compare
facilities because some centers perform more left hemicolec-
tomies, which are typically associated with lower nodal
counts; however, we found that there was minimal variability
in the ratio of left-to-right colon cancers by hospital volume.
The reason for examination of fewer nodes on the left is
unclear but is certainly multifactorial. First, the left colon
mesentery may anatomically contain fewer lymph nodes,
but there is no evidence for this, and a cadaveric study may
help to further understand this issue. Second, it may be that
a formal right hemicolectomy is a more standard operation
and is performed regularly for right-sided lesions, whereas
the length of colon resected for lesions of the left colon is
more variable. In examination of five common surgical
textbooks, we found that the resection margins were
standard for a right hemicolectomy, and a formal right
hemicolectomy was uniformly recommended for malig-
nancies of the cecum, ascending colon, and hepatic
flexure.30–34 However, for left colon cancers, the extent of
colectomy recommended varied from segmental resection
to a formal left hemicolectomy. In this study, we found that
there was variability in nodal examination rates depending
on the lesion location within the left colon. Finally, it may be
that pathologic examination is inadequate; however, this
likely is not the case because both right and left-sided lesions
would be equally affected. If fewer than 12 nodes are found
in a pathology specimen, the pathologist should be
BILIMORIA ET AL .: N ODAL E VALUATION FOR C OLON CANCER
encouraged to search for more nodes and, if necessary, use
advanced techniques, such as fat clearance.35
Despite increases in the adequacy of nodal evaluation for
colon cancer during the past few years, the disparity in nodal
examination rates for left compared with right colon cancer
begs the question of whether 12 nodes is an appropriate
threshold for malignancies of both the right and left colon.
However, single institutions and entire geographic areas have
been able to drastically improve nodal examination rates for
colon cancer. The M.D. Anderson Cancer Center was able to
increase the proportion of patients having ≥ 12 nodes
examined from approximately 60 percent in 1998 to
approximately 95 percent in 2005 by increasing multidisci-
plinary institutional awareness (personal communication
from Dr. George J. Chang to CYK, May 2007). Similarly,
emphasizing the importance of nodal evaluation through an
intensive educational program has resulted in significant
improvement in nodal evaluation in the Canadian province
of Ontario (personal communication from Dr. Andrew J.
Smith to CYK, , May 2007). A campaign to educate surgeons
and pathologists regarding colorectal nodal evaluation across
Ontario resulted in a more than threefold increase in the
proportion of patients who had ≥ 12 nodes examined (from
23 to 77 percent), and the majority of patients who had fewer
than 12 nodes examined were those who underwent
neoadjuvant therapy for rectal cancer. If a large institution
or entire province can improve lymph node examination
rates, then it is likely that 12 nodes is a reasonable benchmark
for right and left colon cancers. We found that nodal
evaluation rates have increased for both left- and right-sided
cancers. Thus, increasing lymph node examination rates is
possible at the local, regional, and national level, and
increasing awareness of the importance of nodal evaluation
may be all that is required.
CONCLUSIONS
To improve cancer care in the United States, a standard
benchmark for nodal evaluation must be established for
colon cancer. This benchmark should be used for quality
surveillance and improvement and not for assessing or
judging provider performance. There has been some
disagreement regarding the importance of nodal evaluation;
however, the preponderance of evidence suggests that
ensuring adequate regional nodal evaluation is an important
measure of quality for colon cancer surgery. Nodal counts
for left-sided cancers are still lower than for right-sided
malignancies despite the fact that examination rates have
increased. Patient, surgeon, pathologist, and hospital factors
contribute to the adequacy of nodal evaluation. The results
of our study demonstrate that adequate nodal evaluation is
associated with improved survival for right and left colon
cancers. Thus, evaluating ≥ 12 nodes for right and left colon
cancers is a feasible, clinically relevant, and modifiable factor
that will likely improve patient outcomes.
B ILIMORIA ET AL .: NODAL E VALUATION FOR C OLON C ANCER
REFERENCES
1. National Comprehensive Cancer Network (NCCN). Avail-
able at: http://www.nccn.org/professionals/quality_measures/
default.asp. Accessed April 26, 2007.
2. Chau I, Cunningham D. Adjuvant therapy in colon cancer:
what, when and how? Ann Oncol 2006;17:1347–59.
3. Chang GJ, Rodriguez-Bigas MA, Skibber JM, Moyer VA.
Lymph node evaluation and survival after curative resec-
tion of colon cancer: systematic review. J Natl Cancer Inst
2007;99:433–41.
4. Berger AC, Sigurdson ER, LeVoyer T, et al. Colon cancer
survival is associated with decreasing ratio of metastatic to
examined lymph nodes. J Clin Oncol 2005;23:8706–12.
5. Caplin S, Cerottini JP, Bosman FT, et al. For patients with
Dukes’ B (TNM Stage II) colorectal carcinoma, examination
of six or fewer lymph nodes is related to poor prognosis.
Cancer 1998;83:666–72.
6. Chen SL, Bilchik AJ. More extensive nodal dissection
improves survival for stages I to III of colon cancer: a
population-based study. Ann Surg 2006;244:602–10.
7. Cianchi F, Palomba A, Boddi V, et al. Lymph node re-
covery from colorectal tumor specimens: recommendation
for a minimum number of lymph nodes to be examined.
World J Surg 2002;26:384–9.
8. Cserni G, Vinh-Hung V, Burzykowski T. Is there a minimum
number of lymph nodes that should be histologically assessed
for a reliable nodal staging of T3N0M0 colorectal carcinomas?
J Surg Oncol 2002;81:63–9.
9. Goldstein NS. Lymph node recoveries from 2427 pT3
colorectal resection specimens spanning 45 years: recom-
mendations for a minimum number of recovered lymph
nodes based on predictive probabilities. Am J Surg Pathol
2002;26:179–89.
10. Johnson PM, Porter GA, Ricciardi R, Baxter NN. Increas-
ing negative lymph node count is independently associated
with improved long-term survival in stage IIIB and IIIC
colon cancer. J Clin Oncol 2006;24:3570–5.
11. Le Voyer TE, Sigurdson ER, Hanlon AL, et al. Colon cancer
survival is associated with increasing number of lymph
nodes analyzed: a secondary survey of intergroup trial INT-
0089. J Clin Oncol 2003;21:2912–9.
12. Prandi M, Lionetto R, Bini A, et al. Prognostic evaluation
of stage B colon cancer patients is improved by an adequate
lymphadenectomy: results of a secondary analysis of a large
scale adjuvant trial. Ann Surg 2002;235:458–63.
13. Swanson RS, Compton CC, Stewart AK, Bland KI. The
prognosis of T3N0 colon cancer is dependent on the number
of lymph nodes examined. Ann Surg Oncol 2003;10:65–71.
14. Joseph NE, Sigurdson ER, Hanlon AL, et al. Accuracy of
determining nodal negativity in colorectal cancer on the
basis of the number of nodes retrieved on resection. Ann
Surg Oncol 2003;10:213–8.
15. AJCC Cancer Staging Manual. 6th ed. Chicago: Springer, 2002.
16. Hammond ME, Fitzgibbons PL, Compton CC, et al.
College of American Pathologists Conference XXXV: solid
161
tumor prognostic factors-which, how and so what? Sum-
mary document and recommendations for implementa-
tion. Cancer Committee and Conference Participants. Arch
Pathol Lab Med 2000;124:958–65.
17. Nelson H, Petrelli N, Carlin A, et al. Guidelines 2000 for
colon and rectal cancer surgery. J Natl Cancer Inst
2001;93:583–96.
18. Baxter NN, Virnig DJ, Rothenberger DA, et al. Lymph
node evaluation in colorectal cancer patients: a population-
based study. J Natl Cancer Inst 2005;97:219–25.
19. Winchester DP, Stewart AK, Bura C, Jones RS. The
National Cancer Data Base: a clinical surveillance and
quality improvement tool. J Surg Oncol 2004;85:1–3.
20. Jemal A, Siegel R, Ward E, et al. Cancer statistics, 2007. CA
Cancer J Clin 2007;57:43–66.
21. International Classification of Disease for Oncology. 3rd
ed. Geneva: World Health Organization, 2000.
22. Standards of the Commission on Cancer Volume II:
Registry Operations and Data Standards. Vol. II: Commis-
sion on Cancer, 1998.
23. Facility Oncology Registry Data Standards. Chicago: Com-
mission on Cancer, 2004.
24. United States Census Bureau. Census 1990. Available at:
http://www.census.gov/main/www/cen1990.html. Accessed
January 21, 2007.
25. Miller EA, Woosley J, Martin CF, Sandler RS. Hospital-to-
hospital variation in lymph node detection after colorectal
resection. Cancer 2004;101:1065–71.
26. Begg CB, Cramer LD, Hoskins WJ, Brennan MF. Impact of
hospital volume on operative mortality for major cancer
surgery. JAMA 1998;280:1747–51.
27. Dimick JB, Welch HG, Birkmeyer JD. Surgical mortality as
an indicator of hospital quality: the problem with small
sample size. JAMA 2004;292:847–51.
28. Schrag D, Cramer LD, Bach PB, et al. Influence of hospital
procedure volume on outcomes following surgery for colon
cancer. JAMA 2000;284:3028–35.
29. Rogers SO, Jr., Wolf RE, Zaslavsky AM, et al. Relation of
surgeon and hospital volume to processes and outcomes of
colorectal cancer surgery. Ann Surg 2006;244:1003–11.
30. American College of Surgeons., WebMD (Firm). ACS
surgery: principles & practice. New York: WebMD Corp.,
2002:v.
31. Brunicardi FC, Schwartz SI. Schwartz’s principles of surgery.
New York: McGraw-Hill Health Pub. Division, 2005.
32. Corman ML. Colon and rectal surgery. 5th ed. Philadel-
phia: Lippincott Williams & Wilkins, 2005.
33. Greenfield LJ, Mulholland MW. Surgery: scientific princi-
ples and practice. 3rd ed. Philadelphia: Lippincott Williams
& Wilkins, 2001.
34. Townsend CM, Sabiston DC. Sabiston textbook of surgery:
the biological basis of modern surgical practice. 17th ed.
Philadelphia: Saunders, 2004.
35. Scott KW, Grace RH. Detection of lymph node metastases
in colorectal carcinoma before and after fat clearance. Br J
Surg 1989;76:1165–7.