Number of lymph nodes retrieved

The most commonly described benefit of using lymph node revealing solutions is the pure increase in the numbers of lymph nodes harvested, many of which are of a smaller size than might be identified by manual dissection. Studies have shown a variable increase in harvested lymph nodes. In one study, a mean harvest of 76.4 and 73.7 lymph nodes was seen after application of alcohol in colonic and rectal resections, respectively.19 In the same study, a secondary manual dissection identified a mean of 18.1 and 21.2 lymph nodes, respectively, but the authors did not clarify whether both sets of dissections were performed by the same individuals.19 If manual dissections had been carried out by less experienced individuals then it is possible that this may have also affected the numbers of nodes harvested.

Metastatic incidence and upstaging

Metastatic incidence refers to the proportion of lymph nodes which contain tumour deposits. A decrease in metastatic incidence after the use of lymph node revealing solutions has been reported.17 ,18 ,22 ,25 ,27 ,31 ,38 ,40–42 Saleki and Haeri40 attributed significance to this finding, stating it to be due to the overall greater number of lymph nodes harvested after secondary dissection. In contrast, five studies showed an increase in metastatic incidence,29 ,32 ,36 ,37 ,39 but not always with significance.37 ,39

Upstaging refers to an upwards change in pathological staging, which may then alter patient treatment if there is a shift from node-negative (pN0) to node-positive (pN1 or pN2). This is because node-positive patients receive chemotherapy, while node-negative patients may not.45

Nine studies reported upstaging after the use of lymph node revealing solutions,17 ,23 ,25 ,27 ,31 ,32 ,40–42 ranging from 2.4% to 33% (table 2).31 ,41 Six of these claimed the finding to be significant, in that upstaging from Dukes’ B to Dukes’ C was reported, prompting adjuvant therapy.17 ,25 ,27 ,32 ,40 ,41 However, this may not have been a correct assumption because most of these studies had questionable underlying primary manual dissection practice with fewer than the recommended minimum of 12 lymph nodes found on average (range 2.94–7.3).17 ,25 ,32 ,40 ,41 These studies were therefore more likely to identify upstaging once a lymph node revealing solution had been applied. It is likely that upstaging would have been insignificant, or not present at all, had there been optimal primary manual dissection. In one study by Koren et al,41 there was upstaging in 10 cases, and a further eight cases had the staging changed from Nx to N0, suggesting an underlying deficit in primary manual dissection technique. The case upstaged by Brown et al27 was a soft tissue metastasis which the authors suggested may have been artefactual. The evidence in the literature is therefore questionable.

Does lymph node size matter?

Multiple studies have demonstrated smaller sized lymph nodes after lymph node revealing solutions are used (table 3).3 ,18 ,19 ,22 ,25 ,36 ,39 ,40 ,42 ,44 Some of the more recent studies using GEWF have assessed and attributed statistical significance to this.36 ,39 ,40 ,42 Brown et al27 found that 83% of additional lymph nodes were ≤2 mm in size. Where GEWF is used this may be due to the white colour of lymph nodes which facilitates detection.39 There is ongoing debate regarding the clinical significance of CRC metastases in small lymph nodes. Dhar et al46 concluded that metastatic lymph node size is a strong prognostic variable in CRC, using two sample log rank testing to demonstrate that the prognostic impact decreased when lymph nodes were more than 10 mm in diameter. Dhar et al46 did concede that their findings needed to be confirmed with a larger study before clinical application. In another recent study, Märkl et al concluded that ‘minute lymph nodes [<1 mm] have virtually no role in correct histopathological lymph node staging’.47 They did however agree that the detection of relatively small lymph nodes (1–5 mm) was an important factor for exact lymph node staging and was prognostically relevant, with an association between a high number of harvested lymph nodes and a favourable outcome in colon carcinoma.47

It is important to consider whether finding a greater number of smaller lymph nodes has the potential to change patient management. If the only significant finding is a greater number of smaller tumour-free lymph nodes, then the patient will remain node-negative and there will be no change in treatment. There will be no benefit to the patient but there will be a cost to the laboratory, both in terms of increased turnaround times and finances.

If metastases are prevalent in larger lymph nodes (ie, >5 mm), then they should be identified by manual dissection, providing the dissector is adequately experienced. If this is the case, then one might argue that the use of lymph node revealing solutions is not necessary. It may be that education is as important a tool as is the use of adjunct chemicals, but currently there remains a lack of evidence to prove or disprove this.

Quality of evidence

The greatest challenge in assessing the true value of lymph node revealing solutions in CRC surrounds the quality of the existing evidence. The majority of existing studies are open to at least one type of bias which may invalidate the conclusions. Different types of bias which may have affected the existing studies are summarised in box 1.

Underlying primary dissection practice

Many of the studies did not achieve the recommended targets during primary manual dissection,17 ,22 ,25 ,29 ,32 ,35 ,36 ,38 ,39–42 with the mean number of lymph nodes harvested ranging from 2.722 to 21.2.19 Kelder et al39 only found a mean of 5.0 lymph nodes by primary manual dissection in 117 colonic specimens, even though their study was relatively recent. The highest number of lymph nodes found in any specimen in their study was only 17,39 which was lower than the average number found by primary manual dissection in a number of other studies.19 ,23 ,27 ,30 ,37 In the study by Schmitz-Moormann et al,22 routine primary dissection yielded a mean nodal count of 2.7, and failed to identify any nodes in six out of the 75 cases. This issue is supported by a number of studies where the importance of enthusiasm and skill of both pathologist and surgeon is noted because it directly affects the quality of the specimen and subsequent nodal harvest.18 ,35 ,42 Gregurek and Wu37 found that educating pathologists in appropriate primary manual dissection practice gave more powerful results than the use of lymph node revealing solutions; however, there was potential bias in their study (box 1). Additionally, failing to consider the experience of dissectors may also introduce sampling bias, perhaps via the involvement of inexperienced dissectors who might miss smaller lymph nodes in comparison with dissectors who are highly experienced in manual dissection. It was often unclear in the case control studies who performed the secondary dissection.17–24 26–29 31–43 The exceptions to this were the studies by Jass et al30 and Vogel et al,25 where secondary dissection was performed by the first author or one of three pathologists not aware of the outcome of the primary dissection, respectively. Only one of the studies included true randomisation of specimens into study groups.38 Gregurek and Wu37 claimed that cases were alternately enrolled into study and control groups; however, pathologists were given the opportunity to change this, which weakened their study design.

Blinding

Studies involving GEWF will always have an immediate detection bias, caused by an inability to use blinding. Iversen et al38 described GEWF as having ‘its own characteristic macroscopic appearance, which is impossible to hide’. This could then either consciously or unconsciously give dissectors the ability to alter their practice which could skew any potentially significant findings. Newell et al36 admit to this limitation, commenting that ‘those pathologists using the standard technique would likely examine pericolic fat more thoroughly’.

Time and cost

The most rapid treatments took 6 h to complete and all used GEWF.40–42 In contrast, the longest treatment using a combination of alcohol and xylene took a minimum of 3 weeks.32 Unsurprisingly, many of the more lengthy treatments have been associated with multistep studies, where more than one chemical has been used in the lymph node revealing solution.27 ,28 ,32 Many studies taking a day or less of additional time to harvest lymph nodes used GEWF.35–38 40–42 With the need to modernise National Health Service (NHS) histopathology departments,48–51 it is unsurprising that focus appears to be shifting towards GEWF, which acts as a fixative while also clearing fat in a shorter period of time than other lymph node revealing solutions. In the 21st century, focusing on lengthy techniques cannot be justified.48–51 Even if lengthy multistep techniques are deemed to provide significant findings, it would be inappropriate to substantially delay reporting. In order to maintain and improve turnaround times, a quicker and more effective method of fat clearance is required if it is to be used routinely. As well as adding a diagnostic delay, older lymph node revealing solutions are also said to be expensive.42

Toxicity

Many older studies used noxious substances, most notably the aromatic hydrocarbon xylene.17–22 ,28 ,30–34 Xylene was once ubiquitous in histopathology laboratories as a clearing agent used routinely in processing and staining. Laboratories are now seeking to eliminate the use of xylene in processing,52–54 due to its known carcinogenic potential.52 This has been facilitated by the introduction of xylene-free processing technology.55 ,56 As a result of this, only one study in the last 10 years has included xylene.27 In contrast, there have been a number of recent studies assessing the use of GEWF, which is seen as a better lymph node revealing solution than its predecessors because it is safe, cheap, easy to prepare and handle,3 ,35 ,36 ,38–41 ,43 ,44 odourless, can be used with standard ventilation, and has no adverse effect on routine special stains or immunohistochemistry.36