Background
In 2010, an estimated 51,370 deaths from colorectal cancer (CRC) are expected to have occurred, accounting for 9% of all cancer deaths in the USA [
1]. For CRC patients, the stage of the disease predicts long-term survival and is weighed in designing treatments [
2]. The acquisition of a single positive lymph node (LN) identifies Stage III patients, and the prognosis worsens as the number of involved LNs increases [
3]. These patients are characterized by a high recurrence rate but may be benefitted by adjuvant chemotherapy [
4‐
6]. Currently, due to conflicting results from clinical trials and population-based studies, the role of adjuvant chemotherapy for Stage II patients remains controversial [
6]. Some investigators, however, recommend chemotherapy for all high-risk Stage II CRC patients, including those with inferior LN recoveries and with peritoneal involvement, extramural vascular invasion, tumor perforation and/or tumor obstruction [
3,
7].
LN involvement is the key factor that determines the stage and prognosis for CRCs [
8]. Nevertheless, LN positivity alone does not identify all patients with a poor prognosis, as 20 to 40% of patients with Stage II (LN-negative) disease die of their cancers [
9,
10]. In population-based studies, the percentages of CRCs in Stages II and III are approximately 40% and 30%, with 5-year, cancer-specific survival rates ranging between 50-80% and 30-60%, respectively [
11,
12]. The proportion of Stage III tumors may, however, be higher than reported because of missing LN metastases due to inadequate examination and resulting under-staging [
13]. The suggested minimum number of LNs to be examined to stage these patients has ranged between 6 and 20 [
11,
14‐
18]. The World Congress of Gastroenterology proposed examination of a minimum of 12 LNs for classification of tumors as Stage II [
19]. In the USA, the American Joint Committee on Cancer (AJCC), the American Society of Clinical Oncology (ASCO) and the College of American Pathologists (CAP), American College of Surgeons (ACoS), Commission on Cancer (CoC), and the National Comprehensive Cancer Network (NCCN) have also recommended examination of at least 12 LNs to assign Stage II disease [
8,
20,
21].
Several institutional and population-based studies showed a survival benefit associated with increasing numbers of LNs examined from Stage II and Stage III CRC patients [
22‐
25]. The origin of the database (Surveillance Epidemiology and End Results, SEER, versus NCCN) also influenced the findings of LN examination on patient prognosis [
17]. In one investigation, increased numbers of LNs examined was associated with improvements in overall survival and relapse-free survival for Stage II but not in Stage III patients [
26]. Goldstein [
27] reported that the predictive probability of finding positive LNs increased with increasing numbers of LNs examined. In contrast, Bui et al [
22] and Wong et al [
15] found no substantial increase in LN positivity with increased numbers of examined LNs.
The recommendation by AJCC, ASCO, ACoS-CoC, CAP, and NCCN, that examination of ≥12 LNs is sufficient to stage a patient with CRC would seem to end the debate yet anecdotal evidence suggests these recommendations may not be followed. To determine how many LNs should be examined from Stage II and III patients with colon cancer, we evaluated a consecutive retrospective cohort and assessed cancer-specific mortality and recurrence. We also attempted to derive a minimum number of LNs needed to stage patients appropriately and thus to minimize under-staging.
Discussion
For both Stage II and III colon cancer patients, increased numbers of LNs retrieved were associated with reduced risk of recurrence and improved cancer-specific survival. Identification of ≥20 LNs correlated significantly with reduced risk of recurrence and mortality of Stage II but not Stage III patients. For Stage III patients, collection of six additional LNs resulted in identification of one additional positive LN, and the probability of finding patients with pN2 nodal stage increased with increasing numbers of LNs examined.
LN involvement determines the pathologic stage and forms the basis for selection of patients for adjuvant therapy [
8]. Although Stage II disease, which has a relatively good prognosis, is characterized by the absence of LN involvement, about one third of these patients experience recurrences, due either to missed micro-metastases or to aberrant drainage of LNs beyond the field of resection, leading to under-staging [
3,
9,
12]. Furthermore, for Stage III patients, inadequate LN recognition is associated with a poorer prognosis [
15,
22], and increased LN collection has a favorable effect on prognosis [
29‐
31]. Thus, diligent searches for LNs are needed for accurate assessment of nodal status and for correct assignment of stage.
The AJCC, ASCO, CAP, ACoS-CoC, and NCCN have recommended that a minimum of 12 LNs be examined in order to rule out metastases via the lymphatic system to nodal tissues [
8,
20,
21]. This recommendation, however, is not widely practiced. Only 58% of those in the SEER database had ≥12 LNs harvested [
17], and the NCCN database review documented a 60% failure rate in achieving resection of 12 LNs among various USA hospitals [
18]. Various minimum numbers of LNs harvested (range: 6 to 40) from colon resections, have been suggested for adequate staging of colon cancer patients [
15,
22,
26,
32].
An increased LN examination confers a survival benefit, especially for stage II disease [
14,
27,
32‐
34]. In the current investigation, however, examination of 12 LNs showed no significant survival benefit. By disease stage, there was a 55% and 31% reduced risk of cancer-specific mortality for Stage II and III patients, respectively, for those with ≥ 20 LNs examined. The 5-year survival of Stage II cases was 54.9%, whereas the survival of those who had ≥ 9 LNs examined after surgery was 79.9% [
35]. The 5-year survival of Stage II patients who had ≤ 8 LNs examined was similar to that for Stage III patients (51.8%). Sixteen of 17 studies of Stage II and 4 of 6 studies of Stage III showed improved patient survival with increased number of LNs examined [
16]. In contrast, for Stage III patients, the number of LNs examined did not serve as a prognosticator [
36]. The demonstration of increased mortality associated with the examination of ≤ 6 LNs, compared to > 6, especially in Stage II patients, is in concordance with other reports [
16,
24,
27].
With tumor recurrence as the outcome for Stage II patients, all higher categories of LN collections showed a decreased risk, but only the ≥ 20 category approached significance, with a 67% decreased risk of recurrence within 2 years after surgery. In contrast, for Stage III patients, there was no relationship between increasing numbers of LNs examined with colon cancer recurrence. A low risk of recurrence for patients with ≥ 14 LNs examined compared to smaller numbers was reported earlier [
32]. There was a significant difference in recurrence with the number of LNs examined. Further, for pN
1 and pN
2 patients, disease-free survival improved as more LNs were removed, but there was no such association for node-negative patients. The impact of LN ratio (ratio of tumor-infiltrated nodes to total number of harvested LNs) on 3-year, disease-free survival was more prominent for patients with > 12 LNs examined [
37].
In general, examination of an increased number of LNs results in greater chances of identifying LN metastases, thus minimizing under-staging [
14,
27,
32‐
34]. In our investigation of patients with Stage III disease, for each additional LN collected, there was 19% increased probability of finding a positive LN. Thus, collection of six additional LNs resulted in finding one additional positive LN. In a mathematical model, the predictive probability of identifying single LN metastases was 0.25 if 12 LNs were examined and 0.46 if 18 LNs were examined [
27]. Since the probability of LN positivity increases as the number examined increases, there is no minimum number that reliably stages all patients [
27,
33]. Higher LN counts, however, do not always correlate with increased rates of nodal positivity [
22].
The accuracy of staging depends on multiple factors, including those that are modifiable (e.g., surgeon and pathologist) and un-modifiable (e.g., age, obesity, and socioeconomic status of the patient and anatomic location of the tumor) [
16,
38,
39]. Pathologists encounter challenges in adequate LN retrieval. In Stage II colon cancer, the age of the patient, tumor size, specimen length, use of a structured pathology template, and academic status of the hospital are predictors of LN collection [
40]. Up to 70% of metastases are found in LNs that are < 5 mm in diameter and hence likely to be missed on routine visualization or palpation [
41]. Another challenge for pathologists relates to micro-metastases or isolated tumor cells that are missed in routine histological examinations. Although immunohistochemistry and polymerase chain reactions to identify cytokeratin and carcinoembryonic antigen [
42] have been used to highlight malignant cells, the prognostic significance of LNs containing such micro-metastases is uncertain [
35]. Targeted LN examination by mapping of the most proximal LN (sentinel LN) improves the staging accuracy for colon cancer [
43,
44]. In LN mapping, however, there are inconsistencies [
45‐
47] that may be attributable to inadequate standardization, training, and interpretation of micro-metastases and to skip metastases [
43,
48].
Survival in colon cancer is influenced by the presence of positive LNs and by the total number of positive LNs [
32]. For Stage III tumors, the AJCC sub-classifies nodal staging into pN
1 and pN
2, based on the presence of ≥4 positive LNs [
21]. In our analysis, the probability of having pN
2 patients increased from 18% to 43% as the number of LNs examined increased from < 7 to ≥ 20. Similarly, there was increased disease-free survival as more LNs were examined from pN
1 and pN
2 patients [
32]. The probability of missing a positive LN was 29.7%, 20.0%, and 13.6% when five, eight, and twelve LNs, respectively, were examined [
49]. For node-positive patients, increased numbers of LN examination correlated with a lower LN ratio, which was associated with a better prognosis [
37,
50]. Although most of these investigations involved large sample sizes, the cutoff values differed; thus, further investigations were warranted.
Conclusions
In summary, the mandatory 12 LNs examination recommended by different agencies (AJCC, ASCO, NCCN, etc.) did not demonstrate a significantly low risk of recurrence or survival benefit. Moreover, collection of ≥7 or ≥12 LNs had similar outcomes. Hence, a minimum of 6 LNs should be examined for adequate staging of Stage II and III colon cancer patients. Collection of ≥ 20 LNs, however, was associated with reduced risk of recurrence and improved survival for Stage II but not for Stage III colon cancer patients. Also, there is an improved survival with increased numbers of LNs harvested from Stage II and Stage III patients regardless of the cutoff points used. For Stage III tumors, every six additional LNs harvested resulted in identification of a positive LN. The probability of finding a pN2 patient increased with increasing numbers of LNs collected. Thus, to minimize stage misclassification and to aid in therapeutic decisions for colon cancer patients, the surgeons should perform more extensive lymphadenectomies and the pathologists should screen the surgical specimens diligently and examine as many LNs as possible. Furthermore, the findings from institutional studies, like ours, relate to the population of the serving area they represent. Thus, there may be geographic differences which can be addressed in future studies and minimized when one follows uniform treatment and pathology protocols.
Financial and non-financial competing interests
The authors declare that they have no competing interests.
Authors' contributions
CKS involved in conception, design, data collection, data assembly, data analysis, data interpretation, and manuscript writing. RBH involved in conception, design, data collection, data assembly, data analysis, data interpretation, and manuscript writing. NCJ involved in conception, design, data collection, data assembly, data analysis, data interpretation, and manuscript writing. VRK involved in data collection, data assembly, data analysis, and data interpretation. BZ involved in data analysis and data interpretation. JAP involved in provision of study patients, data collection, data assembly, data analysis, data interpretation, and manuscript writing. HLB involved in data collection, data assembly, data analysis, data interpretation, and manuscript writing. WEG involved in data analysis, data interpretation, and manuscript writing. IEE involved in data analysis and data interpretation. GPS involved in data analysis, data interpretation and manuscript writing. UM involved in administrative support, conception, design, provision of study patients, data collection, data assembly, data analysis, data interpretation, and manuscript writing. All authors read and approved the final manuscript.