Lymphadenectomy of colorectal cancer is a decisive factor for the prognostic and therapeutic staging of the patient. For over 15 years, we have asked ourselves if the minimum number of 12 examined lymph nodes (LNs) was sufficient for the prevention of understaging. The debate is certainly still open if we consider that a limit of 12 LNs is still not the gold standard mainly because the research methodology of the first studies has been criticized. Moreover many authors report that to date both in the United States and Europe the number “12” target is uncommon, not adequate, or accessible only in highly specialised centres. It should however be noted that both the pressing nature of the debate and the dissemination of guidelines have been responsible for a trend that has allowed for a general increase in the number of LNs examined. There are different variables that can affect the retrieval of LNs. Some, like the surgeon, the surgery, and the pathology exam, are without question modifiable; however, other both patient and disease-related variables are non-modifiable and pose the question of whether the minimum number of examined LNs must be individually assigned. The lymph nodal ratio, the sentinel LNs and the study of the biological aspects of the tumor could find valid application in this field in the near future.

In 1998, Curti et al[1] stressed that continually talking about the lymphadenectomy of colorectal cancer makes for incredibly monotonous reading. In fact, even though it has been proven that the excision of lymph nodes (LNs) in colorectal cancer is a crucial measure, in the last decade the problem has mainly shifted its focus to the physical dimensions of the lymph nodal excision and, more specifically, to the number of LNs to be removed. Even if in this area there are precise indications, in reality, they are not always respected due to, above all else, the large number of variables that can interfere with the sampling of the LNs.

All histological staging, even the recent seventh edition of the AJCC[2], has considered the metastatic involvement of the LNs a determining factor for the staging of the colorectal tumor[3-28] as long as examined in sufficient numbers to ensure the “certainty” of a patient’s prognostic classification[13]. Actually, this is not the case, since as we shall see, the lack of reliable data makes the current staging systems inadequate. This lack creates episodes of “stage migration,” which are likely responsible for the 20%-25% of the cases in which a node-negative patient relapses[29-34], as well as for better prognoses for patients staged IIIa than for those staged IIb[31-34].

The correct staging of a patient treated for colorectal cancer is also critical in the planning of adjuvant therapies that certainly, especially for stage III, ensure improved outcomes and may not be prescribed to a patient who has a falsely-judged more “favorable” staging[3,8,10,11,15,17,19,22-24,27,31,33,35-37]. In this context, some authors as well as some organizations such as the American Society of Clinical Oncology (ASCO) and the National Comprehensive Cancer Network (NCCN) recommend adjuvant chemotherapy to patients for whom the LN study proved insufficient[7,21,25,35,38,39].

In addition to its accuracy in staging, the LNs excision also seems to be an independent prognostic factor. Many case-study reviews[3-5,10,11,14,19,20,24-28,35,40-49], particularly in patients with stage II, report a directly proportional relationship between the number of LNs removed and survival. In this regard, it seems appropriate here to refer to the systematic review of Chang et al[24] who report that in 16 of 17 studies the increased survival of patients with stage II colon cancer was associated with increased numbers of LNs evaluated. The most likely explanation is that the higher the number of LNs examined the better select the group of node-negative patients with a better prognosis for which surgery alone should be curative. Other authors[1,21,40,50], however, believe that in patients with more advanced stages, the lymphadenectomy can be therapeutic both by improving tumor clearance by the surgeon and by reducing the metastatic spread through lymphatic drainage. Not all authors agree with this latter view[5,11].

Last but not least, it must be noted that organizations such as the American College of Surgeons, the ASCO, and the National Quality Forum consider the entity of the lymphadenectomy as a way to gauge the quality of a center dealing with this type of pathology[51-55]. Concerning this issue, not all are in agreement mainly because the number of LNs removed may not reflect the quality of the surgeon or the pathologist but, as we shall see later, may be tied to unchangeable factors inherent in the patient or the tumor[6,50].

Many authors[17,27,28,56] claim that in clinical practice there should be no set limit to the number of LNs examined since in addition to survival, as has already been mentioned, there is a direct correlation between the number of examined LNs and the number of LNs with metastasis[4,14,24,28,57-59].

However, in light of this observation, we have to ask ourselves what the minimum number of LNs is, beyond which there is no change in the staging if not within acceptable limits. Therefore, along with McDonald et al[6] we believe that a “ceiling effect may be reached”, above all for the purposes of allowing pathologists to realize the point at which they can feel satisfied with their search no matter how many LNs may be left in the piece removed. In fact, there is no doubt that in the “real world”, pathologists, with the methods that are presently available, do not or simply cannot sample all removed LNs especially when they are small[15].

In light of this, the number of LNs to be sampled still varies widely even though it has been discussed for over 20 years. In fact, since 1990, at the World Congress of Gastroenterology in Sydney, 12 was established as the minimum standard of LNs to be examined since this number would allow for a correct diagnosis of N0 in 90% of cases[6,21,51,60].

This number, referred to as “magic”[50], was later included in many guidelines and has been endorsed by a large number of United States and European organizations[7,9,17-19,36,40,50,51,61,62].

In this regard, what Stocchi et al[41] have recently reported seems paradigmatic. He claims that, considering only patients treated for stage II colon cancer, the examination of at least 12 LNs is associated with an improvement in results; this improvement reduces if a smaller sample of LNs get examined, but it does not increase with a larger sample of LNs.

Other data reported by Nelson et al[51], Norwood et al[19], and Lee et al[58] show data compatible with Stocchi’s theory. Nelson et al[51] report that by examining 12 LNs, the lymph nodal positivity is correctly identified in 90% of patients; Norwood et al[19] claim that only when the number of LNs is < 12 there is a reduction in the survival rate; finally, Lee et al[58] reports that the examination of a number of LNs ≥ 12 increases the chance of diagnosing positive LNs by 30%.

In light of this data, the number 12 indeed seems correct, but this is not the case of course if in 2012 Fingerhut[4] still asked himself, “Why all the fuss?”

The debate is certainly still open if we consider that a limit of 12 LNs is, as of today, still not the gold standard mainly because the research methodology of the first studies[40,63,64], which do not go beyond a level of III or IV and a grade C recommendation, has been criticized[6,51].

In fact, in the literature there is no uniformity in determining what the minimum lymph nodal sampling is to allow for a greater diagnosis of positive LNs, a different staging that justifies adjuvant chemotherapies, or, ultimately, a better survival rate (Table 1).

This confusion is also documented by McDonald et al[6] who, citing 10 observational studies that analyzed more than 43000 patients, points out that not only is there no agreement on what the LN cut-off point should be, but that in a wide range of LNs examined (between 6 and 21) the actual cut-off point fluctuated. This range is similar to the one reported by Valsecchi et al[21] (between 6 and 17) and lower than the one reported by Noura et al[42] (between 6 and 40).

This variety of data leads us to ask what our main objective is when we examine LNs? If the goal is to “certify” a node-negative patient, then evaluating a number of LNs equal to 12 or perhaps even higher is likely to be necessary; if instead metastatic LNs are identified, it is then possible to “settle” for a smaller number of LNs which, according to some authors[46], are easier to identify as they are more visible and palpable. However, even in this respect “everything and the opposite of everything” can be said if considering what is reported by some authors[35,60,77] who claim that about 50% of enlarged LNs are negative or only an expression of a vigorous immune response, while 45%-78% of metastatic LNs have a less than 5 mm diameter.

Gelos et al[78], however, focuse on yet another aspect, arguing that in patients with a malignancy at an earlier stage which can lead to a lower immune response, it is likely that we can settle for a sample of less than 12 LNs.

The fact that even today the number “12” target is “uncommon, “not adequate,” or accessible only in highly specialized centers[21,41,50] is demonstrated by the fact that in the United States in 2001[4], the number of 12 sampled LNs was reached for only 44% of patients and that the target of patients increased to 75% in only 38% of hospitals in 2004-05[69], 15 years after Sydney. Moreover, once again in the United States, reports published between 2005 and 2010 revealed that, despite the “dense forest of articles”[79] lymphadenectomy was still considered inadequate in 48% to 63% of cases[79,80].

Similar experiences are also reported in Europe; in fact, in Germany in 2009, the Dutch Surgical Colorectal Audit[77] reported that in 73% of colon cancers and 58% of rectum cancers, the number of LNs examined was ≤ 10; in England, Johnson et al[81] and Mitchell et al[82] have recently pointed out that the limit of 12 LNs was not reached between 33% and 50% of colorectal cancer cases.

It should however be noted that both the pressing nature of the debate and the dissemination of guidelines have been responsible for a trend that, over the years, has allowed for a general increase in the number of LNs examined, thus enabling the U.S. to increase the number of hospitals that reach the target of 12 LNs from 15% in 1995-96 to 38% in 2004-05[69] and in 2005-2008[52] reach the figure of 92% albeit only in centers that are members of the NCCN and are thus made up of top institutions[69]. This, however, is probably not the case in smaller hospitals[9,21].

Ideally, the surgeon should remove all the LNs pertaining to the tumor and the pathologist should sample and examine them thoroughly. However, even if this were carried out, all authors agree that there would be “unmodifiable” factors patient-related and disease-related that could make a node-count problematic (Table 2).

In our opinion, all the variables, modifiable and unmodifiable, that can affect lymph nodal sampling should be examined so as to make the work of both the surgeon and of the pathologist more efficient.

The seventh edition of the AJCC classification[2], as mentioned previously, subdivides patients treated for colorectal cancer into prognostic categories according to the number of metastatic LNs. The accuracy of the staging is, however, influenced by the number of retrieved LNs that must be ≥ 12. The lymph nodal ratio (LNR)[49], which is the relationship between positive nodes divided by the total number of retrieved nodes, is in our opinion, justified mainly because it means not having to reach the so called “magic number.” The LNR prognostic validity could in fact be effective also in cases of reduced lymph nodal sampling[12,16,45,114]. However, not all authors who have written on the subject agree[6,43,49]. The LNR, independent of the number of LNs sampled, is also justified since, taken with the AJCC classification, it would allow us to sub-divide, according to the risk involved, stage III patients reducing the excessive prognostic heterogeneity[12,16,45].

In light of this, reviewing and taking into consideration the work of Bamboat et al[8], Qiu et al[12], Song et al[13], and Greenberg et al[114] in 2011, the LNR seems to be an independent prognostic factor in colorectal cancer, superior to the classification based only on N stage (number of positive modes). In fact, based on the LNR analysis, Greenberg et al[114] himself state that the survival rate of stage III patients with favorable LNR is similar to that of stage II patients.

Conversely, Noura et al[42] only one year before published an interesting and somewhat more cautious editorial. In fact, the author reported that even though the LNR seemed to be a more reliable prognostic factor, its validity, in actuality, could not be completely agreed upon. In fact, clinical records were very different, randomized and multi-centric studies were lacking, and, most importantly, a uniformly valid cut-off was missing.

One thing is certain, given the importance of both the lymph nodal sampling and of the evidence of the lymph nodal metastasis, it is unthinkable that a pathologist could stop after “having found” the first neoplastic LN[78]. However, it is not exactly clear what the “ceiling effect” is even in this case.

Despite the fact that numerous authors have expressed their opinions on the number of LNs sampled, it can be gathered that the number is between 6 and 40[42]. Therefore, only in light of this wide range, should we all refer to the minimum number of LNs (now the obsessive “12”) which can allow us to avoid the so-called “Will Roger’s phenomenon”[115] responsible for understaging.

Even if the surgeon and pathologist, as variables in the equation, could improve simply by standardizing surgical technique and by increasing the amount of time dedicated to this procedure, the other, more important variables[22,31,114], namely patient and cancer-related, are not as easily modifiable. It is with these latter two variables in mind that we still pose the question whether it is possible, as we hope, to establish a universally valid cut-off node for all patients or whether it should instead be varied according to individual cases[6,78].

Today, a valid perspective is still necessary for the identification of the sentinel LNs (at least 3)[116]. This approach is based on the idea that the lymphatic flow originating from the tumor occurs “step by step”[35,117] and the purpose of this technique in colorectal cancer would be not so much to modify the size of the resection, as has happened with other diseases, but for its “potential” effects on improving the staging[29,31,40,41] since it would allow for more involved and expensive techniques on only a few LNs[29,30,32-34] which would reveal micrometastases or isolated tumor cells.

The identification of the sentinel LNs, actually still remains a controversy among those authors who consider the mesenteric lymph drainage, especially in the rectum, too complicated[32,40], and the majority of authors who, on the other hand, maintain that an aberrant lymphatic drainage occurs only in a small percentage of cases[27,29,33,34].

As certainly interesting, the biological aspects of the tumor still remain the subject of speculation. Some authors suggest that reduced survival is not necessarily due to an inappropriate dissection performed by the surgeon and the pathologist, but may be linked to a cancer that is quite virulent and is hence responsible for a low immune response from the patient[35,40,57,117]. Still, some other authors[11,40,41,46,47,118,119] maintain that an elevated sampling could be determined by a vigorous immune response; this, in turn, is determined by the molecular instability of the tumor, which manifests itself as a high rate of “neo-antigens” and therefore causes a more limited neoplastic progression. Not coincidentally, these malignancies are located in the right colon[41,44,118], where, as mentioned, more LNs are found.

As has already been pointed out, the obsession with the number 12 has its origins in studies which lack clear statistical evidence. Just as Curti et al[1] asserted that, as of 1998, not even a single prospective study had been published, authors still today are calling for prospective controlled studies that are, without question, difficult to predict both for a number of ethical reasons and for the sheer volume of clinical records. Hence, obtaining reliable data that would allow us to go beyond this obsession with the number 12 will not be easy.