Eleven years after Snow had published his observations, inferences, and recommendations for what he deemed to be proper treatment of melanoma, Eve,² in 1903, seconded those recommendations. This is how Eve couched it: ". . . the removal of the nearest chain of lymphatic glands, whether palpably enlarged or not, should never be omitted; for it may be taken as a matter of certainty that in the majority of cases they are infected." It is interesting to note that both Snow and Eve employed the word "infected" for lymph nodes that harbored a metastasis of melanoma, a choice that conveys something pertinent about how the process was conceptualized in those days.

The statement of Eve just quoted and that of Snow quoted previously, inaugurated an era in which radical surgery became the treatment of choice for melanoma and in which END became a popular adjuvant to that theory. Much was written about the value of END for cure of melanoma, and much of it would be invoked as gospel many years later by proponents of SNB. What follows are excerpts from writings of some of the personalities, most of whom have become larger-than-life characters in the history of "modern" surgery, concerning radical surgery for malignant neoplasms in general and for metastatic melanoma in particular.

Handley³ (Figure 1), and later Halsted⁴ (Figure 2) both of whom were especially active early in the twentieth century, advanced the notion that removal of nodes in conjunction with local surgery was essential for patients with primary melanoma of the skin and primary carcinoma of the breast. Both Handley and Halsted claimed to have demonstrated that cancer at first grew centrifugally from the site of the primary neoplasm, and then extended by continuity into lymphatics and through those vascular channels, by continuity, to nodes in which the neoplastic cells proliferated. The idea that neoplastic cells could embolize through lymphatic or blood vessels was dismissed by Handley and Halsted as inconceivable. Both radical mastectomy, devised by Halsted for carcinoma of the breast, and radical node dissection for patients with melanoma, based in part on the work of Handley, were predicated on erroneous concepts of the behavior of malignant neoplasms and, therefore, were designed, wrongly, to remove the cancer in its entirety, in a single sweep, at the primary site and in regional nodes. The rationale for these misguided ideas is expressed in the words of the two surgeons themselves. Halsted, in 1907, told of his results concerning the surgical treatment of carcinoma of the breast thus: "In showing that cancer cells in the blood excite thrombosis, and that the thrombus as it organizes usually destroys or renders them harmless, Goldman and Schmidt seem to have established a fact of primary importance and one which is strongly opposed to the embolic theory as applied to carcinoma . . . We believe with Handley that cancer of the breast in spreading centrifugally preserves in the main continuity with the original growth, and before involving the viscera may become widely diffused along surface planes." Later in the same article, Halsted reinforced that conviction when he said: "There is then a definite, more or less interrupted or quite uninterrupted, connection between the original focus and all the outlying deposits of cancer." Handley came to believe in "continuous growth" of melanoma, rather than "embolic dissemination" of it, in part by virtue of an autopsy he had performed on a woman with disseminated melanoma, the findings of which he presented in 1907 in these lines: " . . . the growth is not invading the tissues in the form of isolated embolic foci scattered at random here and there in the vessels; it is growing along the vessels and choking them up. . . . The process is one of continuous growth of sarcoma along the vessels . . . it is clear that lymphatic permeation is the initial process in the local centrifugal dissemination . . . "

This theme of Handley and Halsted was taken up by Pringle who, in 1908, reported on two patients with disseminated melanoma on whom he had operated, and to whom he returned in follow-up in 1937.⁵ One of the patients had melanoma on a forearm and the other melanoma on a thigh, and both had metastasis of melanoma in regional nodes, yet both were alive 38 and 30 years, respectively, after surgery that had been performed on them in 1898 and 1906, respectively. Based on those apparent successes, Pringle proposed a particular procedure for extirpation of melanoma based on favorable results: " . . . a radical extirpation of this disease . . . including the nearest group of glands . . . should be in one continuous strip . . . The after-history of these two patients—the only two on whom I have operated for melanoma—is good proof of the force of this contention." The advice of Pringle concerning the surgical treatment of melanoma, predicated wholly on his experience with those two patients, was consonant with the principles then being advocated for treatment of melanoma, principles that had been embraced fervently by the entire community of physicians and surgeons. Although the concept that a primary malignant neoplasm spread by growth continuously and centrifugally was flawed woefully, it was not surprising that, in time, even more radical procedures, such as forequarter and hindquarter amputations, would be performed on patients whose regional nodes were affected by metastasis. The belief that such surgery might actually cure patients was rooted in concepts set forth dogmatically by Handley and Halsted, beliefs that would be propagated persistently and passionately for most of the twentieth century.

Adair,⁶ at a symposium dedicated to cancer at the Clinical Congress of the American College of Surgeons in 1935, presented a study of 400 patients with melanoma. In his discussion, he averred that melanoma has potential to spread in three ways: (1) invasion of tissue locally, (2) dissemination widely through the blood vascular system, and (3) dissemination by way of the lymphatic system to the nearest regional nodes. Of this last mechanism he wrote optimistically that " . . . it is in this type of extension that most of our cures come." Adair believed that the melanoma in 70 patients was operable and the rate of survival of those patients for five years was 33%. Although the operations performed were not specified, this is what Adair said about his aspirations for those procedures: "In the removal of the mole containing black pigment as a prophylactic measure lies our chief hope of improving the percentage of patients cured. More and earlier surgery should be employed, rather then less." He promoted the principle of wide and deep excision, END, and other types of radical surgery, no matter how mutilating that surgery might be. That mentality expressed by Adair was the dominant one in the treatment of melanoma for scores of years to come.

In 1945, Pack, et al.⁷ stated tersely the principle utilized by them for surgical treatment of melanoma: " . . . each operation for melanoma [is] to include en masse both the primary and metastatic tumors by the procedure which we refer to as excision and dissection in continuity" (Figure 3). In regard to END, this is what they advised: " . . . elective removal of lymph nodes in the groin [should be performed] even though they are not palpable and there is no clinical evidence of metastases . . . It is our opinion that this dissection should be done routinely for all melanomas of the extremity and genitals." These contentions of Pack and his coworkers were founded, at least in part, on the discovery of metastasis of melanoma in two of seven patients and in five of 10 patients who underwent END of the inguinal and axillary node basins, respectively. In regard to those patients, the collaborators commented thus: " . . . this early operation, although needlessly done in some instances, affords the patient with metastatic melanoma . . . a better and earlier opportunity of cure." The principle of END they endorsed derived from the results of previous studies in which they, themselves, had observed development of metastasis of melanoma to inguinal nodes of patients who had no involvement of those nodes clinically and who had been treated solely by wide excision of the melanoma in the skin. They recommended that END be performed after an interval of six weeks following surgery for the primary neoplasm. In their view, the delay was justified because studies by them showed that, on average, it took 15 months before metastases became apparent clinically.

In 1953, Meyer and Gumport⁸ reported on 105 patients with melanoma and advised, succinctly, what they considered to be proper surgery for that cancer when it was primary: " . . . removal of the primary tumor in continuity with the regional lymph node drainage . . . " The principle of END was invoked as follows: "It is my [H.W.M.] opinion at this time that a regional lymphnode [sic] dissection should be performed even if the lymphnodes [sic] are not clinically palpable." The idea of "prophylactic," i.e., elective, node dissection came directly from the experience of Meyer and Gumport, namely, that at least 40% of nodes not palpable clinically actually harbored melanoma. Their rationale for recommending END was based on the belief that melanoma first was localized to regional nodes from whence neoplastic cells disseminated to other organs. This is how those two surgeons thought about the matter: " . . . this metastatic melanoma may be a focus and source from which malignant cells may spread to other nodes, or may enter the blood stream and cause distant metastases. These metastases, of course will seal the fate of the patient."

At the time that the article by Meyer and Gumport was published, it seems that END had not been accepted universally by physicians as being necessary for management of melanoma and the thinking of opponents of the procedure was summarized by Meyer and Gumport in the lines that follow: " . . . [there is] a school of thought which feels that one should wait until the lymph nodes become clinically palpable. A diagnosis of metastases should then be made by either excision of a node or by aspiration biopsy. If a positive diagnosis is obtained this school advises against regional lymph node dissection. They believe it is better to amputate the regional lymph node area, thereby staying outside of the regional lymph node field." In brief, opponents of END argued on behalf of more extensive surgery and advocated a cure for metastatic melanoma being achievable even after a patient had developed proven metastasis in nodes palpable clinically, the way to achieving cure being to perform disarticulation of a hip, hemipelvectomy, or interscapulo-thoracic amputation (Figure 4). Even though the precise timing of radical surgery was a contentious issue, all those engaged in the debate concurred that a surgical procedure which removed a malignant neoplasm in its entirety, that is, both the primary and metastases from it, would result in cure, but only if the surgery undertaken was sufficiently extensive.

In 1955, Lund and Ihnen⁹ titled an article concerning their experience with 93 melanomas with a poignant query: "Malignant melanoma; clinical and pathologic analysis of 93 cases. Is prophylactic lymph node dissection indicated?" Thirty-two patients had nodes that were thought clinically to be involved by melanoma and 19 of those patients underwent therapeutic node dissection in addition to wide local excision of the cutaneous melanoma. Only one of the 32 patients survived for five years and, paradoxically, the nodes in that particular patient were negative for melanoma. Thirty-eight patients who had no involvement of nodes clinically underwent wide excision of the primary melanoma only. No ENDs were performed in this series of patients. Of the 38 patients with nodes that appeared to be negative for melanoma clinically, 13 survived for five years without any evidence of metastatic melanoma; nine developed regional node metastases and underwent regional node dissection, and of those nine, four survived for five years; 16 patients developed metastases and were dead within five years (details of the course of those patients were not divulged). The authors posited that the 16 patients who did not undergo lymphadenectomy could have benefited from it and this is how they put it: "The place of prophylactic lymph node dissection in the treatment of melanoma is still controversial and unanswered by analysis of this series. It may be said that delay in dissection was not very successful and there is suggestive evidence that earlier lymph node dissections might have been beneficial. On the basis of the available evidence, considering the accepted principles in the treatment of many other tumors, and in view of the relatively low mortality and morbidity attendant upon lymph node dissections, it is our personal opinion that prophylactic dissections are indicated if the regional lymph nodes are in predictable and accessible areas."

In 1968, Cameron,¹⁰ in his discussion of 209 patients with primary melanoma, raised his voice in support of radical surgery for melanoma, his purpose being to counter opposition that existed in some quarters to that kind of extensive extirpation. This is what he wrote: "It is therefore essential that the impression, still held by some, that surgery little influences the course of the disease should be dispelled." For Cameron, early diagnosis and appropriate surgical treatment of melanoma saved lives. This is the treatment he prescribed: "I believe that prophylactic block dissection of the lymphnodes[sic] should be carried out much more often . . . ." He preferred the operation employed by Pringle at the beginning of the century or an amputation, were it acceptable to the patient.

Despite there being no definite evidence that END had any value for a patient, the procedure by 1970 had become highly acceptable to physicians and surgeons for any patient with melanoma who did not have palpable regional nodes or other signs of metastatic melanoma.

It was on the basis of articles just referred to published in the first 70 years of the 20th century that the foundation for radical surgical treatment of melanoma was established. Principles of surgery were constructed on anecdotes and ill-conceived ideas concerning the mechanism whereby metastasis of melanoma developed. Those principles dictated that extensive local surgery and nodal dissection, in continuity if possible, was optimal treatment for patients with nodal metastasis, the procedure coming to be known as "therapeutic node dissection." The very same surgery was recommended if the regional nodes did not seem to contain melanoma, but in this circumstance the procedure was referred to as "END." Resistance to what was in those days conventional thinking, that is, radical surgery for melanoma and metastasis from it, was offered in an editorial in 1970 by Jackson,¹¹ a Canadian surgeon, and captioned, "Persistent fallacies concerning dissemination of malignant melanoma." In that piece Jackson questioned the utility and validity scientifically of END and called attention to the fact that in 1907 Handley did not record spread from a primary melanoma, but rather metastasis from a metastasis of melanoma. For that reason, Jackson argued, any conclusions drawn by Handley from observations about growth of metastasis cannot be assumed to apply to melanoma primary in the skin. Jackson cited other evidence that seemed to give the lie to principles at the time that governed radical surgery for cutaneous melanoma, among the data said to be compelling being that 25% of melanomas spread hematogenously, after bypassing nodes. Moreover, in 6% to 10% of patients with metastases of melanoma but no identifiable primary melanoma, the mode of dissemination was not known. Jackson expressed views contrary to the conventional ones in this comment: "Any survey on the literature on malignant melanoma shows that over the years our thinking on this subject [that is, the treatment of primary cutaneous melanoma] has been unduly influenced by the papers by Handley, Pringle and Pack . . . Surely it is time to shed this simplistic dogma and to approach the management of malignant melanoma on the basis of what is known (or not known) about the natural course of this disease." Jackson was far ahead of his time; what he articulated correctly more than 30 years ago is as unheeded by colleagues today as it was by his contemporaries then.

Thurston,¹² also in 1970, offered reservations about dissection of nodes in patients with melanoma and put it this way: "The literature on prophylactic lymph-node dissection [END] for melanoma is confusing." In his consideration of whether END was more advantageous than dissection of nodes after metastatis of melanoma had become manifest clinically in them, Thurston made this remark: "To date we do not have enough evidence to enable us to choose between these two approaches. Many studies alleged to support prophylactic node dissection were poorly controlled because they did not compare patients who had prophylactic dissection and histologically positive nodes with those who, after treatment of the primary lesion, had therapeutic dissection when node metastases became clinically evident."

By 1970, both Jackson and Thurston had observed, rightly, that the medical literature did not support the then popular belief that END was either indicated or useful for management of a patient with melanoma. Despite the lack of any convincing data on behalf of END, opposition to it in the establishment of community and university hospitals was negligible. In the ensuing 30 years, END was studied intensively and, during that period of time, indications for utilization of it were to change, not once, but often. Nonetheless, the attitude that prevailed was the conventional one — and it appeared to be immutable, no matter how compelling was evidence to the contrary. It is worthwhile to examine critically the conflict and the different claims made about the effectiveness of END.

In 1976, Breslow¹³ published the results of a study he had conducted prospectively of 41 patients with melanoma and added them to those of a study of 97 patients with primary melanoma he had undertaken retrospectively and reported on in 1970.¹⁴ In his analysis, Breslow compared the survival of patients who had undergone wide local excision of the melanoma in conjunction with END with the survival of patients who had only a wide local excision of the neoplasm. Of patients with melanomas thinner than 0.76 mm, 100% survived for five years, irrespective of the treatment given, and of patients with melanomas that measured between 0.75mm and 1.50mm in thickness, 70% survived for five years regardless of the therapy provided. The rate of survival for five years of patients with a melanoma thicker than 1.5mm was 64% if END had been performed in addition to wide local excision, in contrast to 31% if no END had been done. Prior to publication of this study by Breslow (and of other studies similar to it), patients were selected for END¹² based on whether, at the least, cells of their melanoma filled completely the papillary dermis, that is, fulfilled criteria histopathologically for Clarks level III. A patient with a melanoma even thicker than that, such as Clark's level IV (neoplastic melanocytes being in the reticular dermis) or level V (neoplastic melanocytes having entered the subcutaneous fat) surely was destined to undergo END. The results of the study led Breslow to conclude that patients should be selected for END based on the thickness of the melanoma and not on Clark's "level of invasion." This is what he wrote in 1976 about the matter: "It is clear that patients with lesions less than 0.76 mm thick should not be subjected to node dissection while those with lesions greater than 1.5 mm should be so treated. The indication for node dissection for the group with lesions 0.76 mm-1.50 mm thick are still obscure and the treatment will continue to depend on the intuition of the surgeon." Then and now, intuition of a surgeon should not have been a factor in the management of a patient with metastatic melanoma and, futhermore, what seemed to be beneficial, then and now, was illusory.

Recommendations similar to those of Breslow were proposed by Balch, et al.¹⁵ in March 1979 on the basis of their assessment of 384 patients with melanoma. All of the patients had had a wide local excision and, in addition, some had undergone END. The authors confirmed the results published previously by others, namely, that all melanomas thinner than 0.76mm were cured by wide local excision. They found that patients whose melanoma was of intermediate thickness (1.5 to 3.99 mm) had an 82% chance for surviving eight years when both wide excision and END were performed. By contrast, only 25% of patients treated solely with wide local excision lived for eight years. END done for patients whose melanoma was greater than 4.0mm in thickness did not enhance survival. All of this prompted the authors to come to this position: "Elective RND [elective regional node dissection] is not indicated for lesion[sic] < 0.76 mm in thickness, but it is considered for 0.76 to 1.50 mm lesions in selected patients and is employed for virtually all patients with lesions exceeding 1.5 mm in thickness." They also advised END for patients with a melanoma thicker than 4.0 mm, the reasons being that prognosis for those patients could then be determined and adjuvant therapy could then be given in the setting of a clinical trial. In an attempt to explain why END dissection did not benefit that set of patients with a thick melanoma, Balch and his collaborators wrote as follows: "For melanomas exceeding 4.0 mm in thickness, the potential benefits of immediate lymphadenectomy are much less because the incidence of simultaneous metastases at distant sites appear to diminish the beneficial effects of removing any regional metastases."

The ideas of Balch and coworkers about how melanoma metastasizes is conveyed best in their own words as follows: " . . . nodal metastases are more curable in a subclinical stage and that a delay of two or more years waiting for microscopic nodal metastases to enlarge to a clinically detectable size is a critical interval during which regional metastases may disseminate to distant sites." In brief, Balch, et al. believed that if a melanoma less than 4.0 mm thick metastasizes, it tends to reach only as far as regional nodes, which then can be removed and, thereby, cure effected. They claimed that if regional nodes are not removed, those structures become the source for dissemination of metastases.

In August of the same year, Balch, et al. reported,¹⁶ once again, on their experience with the same group of patients, but in slightly modified form. By then, 394 patients had been studied by them, half of whom had had a wide local excision of the melanoma plus an END, whereas the other half had had a wide local excision only. The authors told of eight-year survival of 78% for patients who had undergone END, whereas all patients who had had a wide local excision were dead in eight years. The advantage for survival of patients who had received an END was not apparent if the melanoma itself was thinner than 1.5 mm or thicker than 3.99 mm. In short, the results of this study seemed to indicate how imperative it was that patients with a melanoma between 1.5 mm and 3.99 mm in thickness and without evidence of metastasis clinically should undergo END. That this study was flawed eggregiously and that the concept of how metastases develop as advanced earlier by Balch and associates was inaccurate soon became evident; no other study, before or since, including ones undertaken by Balch, who is a surgeon, has shown such huge disparity in regard to survival and such overwhelming benefit that accrued from END. In fact, no other study has even come close to proclaiming such results.

In 1982, Balch, et al.¹⁷ compared factors in prognosis and survival of patients with melanoma in Alabama, USA, to those of patients in New South Wales, Australia. Incorporated in the study were patients who had had a melanoma and who had been followed for as long as from 1955. It appeared that both groups of patients, when matched for factors in regard to prognosis, had a similar survival overall, which motivated the authors to suggest that the behavior of melanoma in persons living in the USA was no different from the behavior of melanoma in persons in Australia. The results confirmed benefit, similar to that published previously, of END for patients with melanoma who had no evidence of nodal involvement clinically. The authors communicated their sense for benefit of END in these lines: "Patients with thin (< 0.76 mm) and thick melanomas (> = 4.0 mm) did not benefit from elective RLND [elective regional lymph node dissection] while those with immediate thickness (0.76 mm to 3.99 mm) had significantly improved survival rates if their initial surgical management included elective RLND. The benefit of elective lymphadenectomy was particularly striking for patients treated at both institutions with melanomas measuring from 1.5 to 3.99 mm in thickness. For melanomas ranging from 0.76 to 1.50mm thickness [sic], the improved survival rates for elective RLND were primarily in male patients." The inability of END to increase survival overall in patients with primary melanoma greater than 3.99mm in thickness was explained by the authors thus: " . . . patients with thick melanomas (> = 4.0 mm) have a high risk for harboring distant microscopic metastases that negates the curative influence of the regional operation." By this statement, Balch and colleagues sought once again to convey what, in actuality, was a misapprehension, to wit, that if a metastasis of melanoma in a patient whose original melanoma was thinner than 4.0mm, the metastasis may have reached the regional nodes but not spread beyond them, thereby in their minds making the regional metastasis amenable to extirpation surgically.

Milton, et al.,¹⁸ in 1982, wrote about "Prophylactic lymph node dissection in clinical stage I cutaneous malignant melanoma: results of surgical treatment in 1319 patients." The patients all had melanoma without apparent involvement by metastasis of nodes clinically. The patients were divided into two groups according to the treatment they received; the first group had only wide local excision and the second had both wide local excision and END. In both groups, survival for five years was high. For men with lesions of intermediate thickness (in particular 1.6–3.0 mm) who also had END, the rate of survival was higher. For melanomas thicker than 3.0 mm, however, the choice of treatment did not affect survival. In women with thicker lesions, the rate of survival was higher for those who had END.

In 1985, Roses, et al.¹⁹ directed attention to END in an article about "Prognosis of patients with pathologic stage II cutaneous malignant melanoma." Patients with melanoma in nodes suspected clinically and confirmed histopathologically and patients who did not seem to have a metastasis in nodes clinically, but in whom END did reveal metastasis in nodes, were included in the study. The coworkers examined data from 213 consecutive patients. In 157 patients, the metastases were discovered after END and in 56 patients metastasis was suspected clinically and confirmed following dissection of the nodal basin. The five year survival for patients in these two groups was 44% and 21%, respectively, and the 10 year survival was 28% and 12%, respectively, which encouraged the authors to write as follows: "The difference in survival between patients with clinically negative/ histologically positive nodes (clinical Stage 1) and clinically positive/histologically positive nodes (clinical Stage II) was apparent throughout the follow-up period." They concluded that patients who underwent END have a better prognosis than those who did not have the procedure, an inference that was held equally fervently by many other surgeons at that time. Their conclusion, however, is not based on either logic or fact; at the outset, the two sets of patients differed in the extent of disease clinically. One group had palpable regional nodes, an indication, therefore, that neoplastic cells had spread earlier or grew faster in all those patients than in the other group in whom nodes were not palpable. The cohorts, therefore, should not have been compared and the merit of END cannot be judged fairly using data presented in this study.

In 1985, McCarthy, et al.²⁰ published the results of a study of 2347 patients with melanoma whose nodes seemed to be normal by assessment clinically. In large measure the study was prospective, but it was not randomized. The collaborators demonstrated that men with melanomas that measured 1.6 to 3.0 mm in thickness and who underwent END in addition to wide local excision had significantly longer survival overall, that is, 77.5% versus 34.1% for men who had wide local excision only. A major benefit accrued to women if they had a melanoma thicker than 1.5mm on an extremity, even one thicker than 3.0 mm. Women with a melanoma on the trunk, irrespective of the thickness of it, however, did not benefit from END.

In 1988, McCarthy, et al.²¹ reviewed the records of 3171 patients with melanoma, about one third of whom had undergone END, and came to a conclusion that seemed to be different from that set forth previously. In brief, they found that if a melanoma were thicker than 1.6 mm, then END led to a significant decrease in the likelihood of metastasis becoming apparent clinically. No figures concerning survival overall was presented.

In 1990, Crowley and Seigler²² examined the role of END in patients who had a melanoma thicker than 4.0 mm. They assessed the findings in 308 patients, 116 of whom, in addition to wide local excision of the melanoma, had an END. Patients who underwent END did not have a better outcome than those who did not undergo the procedure; their survival overall was not better and the period of time before metastasis became manifest clinically was not at all longer. On the basis of these findings, the authors stated directly that they " . . . do not recommend ELND for patients with thick melanomas because the risk of distant metastases outweighs any benefit of regional node dissection." The results of this study reinforced a conclusion that had been reached previously and had been accepted by other students of the subject, that is, END is of no benefit to patients with a melanoma thicker than 4.0 mm. Another conclusion derived from the study was that patients with a melanoma that measured between 4.0 mm and 6.0 mm in thickness had a poorer prognosis, but not by much (p=0.01) if their nodes were positive after END, whereas patients whose melanoma measured between 6.0 and 10.0 mm in thickness had a poor prognosis, irrespective of whether or not they had positive nodes. This finding was consonant with the possibility that a metastasis may have entered the blood vascular system at the site of the melanoma or that dissemination occurred through lymphatics and cells of melanoma then passed right through nodes without proliferating in them.

In 1993, Drepper, et al.²³ published the results of a study they had undertaken retrospectively of 3616 patients with melanoma who had undergone END. They found that that procedure provided no benefit when all the patients were analyzed as a single group. Following extensive analysis, however, some sets were identified in which patients who had END did better in terms of survival for five years. This particular advantage was seen in men with an axial or acral melanoma whose thickness was between 1.5 and 4.5 mm. The rate of survival for those men improved by 20%. Men with a nonulcerated melanoma benefited from END if the neoplasm was 3–4.0 mm thick or was Clark's level V. There was benefit for women whose melanoma measured somewhere between 2.5 and 5.0mm in thickness, the rate of survival for them improving by about 8–10%. The analysis was performed on five specific sets as follows: thickness of melanoma (three different categories of thickness), gender, anatomic site (categorized as favorable and unfavorable), type of melanoma (four categories), and presence or absence of ulceration.

In sum, these studies, in general, showed benefit for patients who underwent END, especially if the melanoma was thinner than 4.0 mm. It was thought that the reason END proved to be successful in those patients was that the melanoma had metastasized to regional nodes, but not beyond them. The same case was not made for melanomas thicker than 4.0 mm; it was said that a melanoma thicker than 4.0 mm most likely would have metastasized beyond regional nodes. For melanomas between 0.75 and 1.5 mm, no definite claims about the worth of END were made. None of those studies were randomized and prospective, and, therefore, none could stand the test of what scientifically is a valid study.

The first prospective and randomized study of END was performed by Veronesi, et al.^24,25 and reported on by them in 1980 and again in 1982. Those collaborators, surgeons mostly, assessed 553 patients with melanoma of the limbs whose regional nodes seemed to be free of metastasis clinically. The study was carried out between September 1967 and January 1974 under the auspices of the W.H.O. Collaborating Centres for Evaluation of Methods of the Diagnosis and Treatment of Melanoma. Two hundred and eighty-six patients were just observed following wide local excision of their melanoma, whereas, in addition, 267 patients underwent END. After five and 10 years, overall survival was identical for both groups of patients. Analysis of sets showed that women had better survival overall than men, but that END did not actually improve survival in women. Analyses of other sets, such as Clark's "histogenetic" type, Clark's level, and Breslow thickness, revealed that whether or not patients had an END, they fared the same. In short, END provided no advantage in terms of survival for patients. Because all of the melanomas were situated on a limb and the lymphatic drainage of the skin of limbs is predictable with considerable accuracy, the design of this study likely eliminated any chance that the results would be influenced by lymph that flowed to nodes located at a site not included in the dissection of regional nodes. In conclusion, Veronesi, et al. stated that " . . . delayed node dissection is as effective as the immediate dissection in stage I melanoma of the extremities . . . , " that is, patients with melanoma had the same survival overall irrespective of whether or not they had an END or a therapeutic node dissection at the time metastases manifested themselves clinically.

In 1978 and again in 1986, Sim, et al.^26,27 reported on 171 patients with melanoma who were studied prospectively. The patients underwent wide local excision of the melanoma and were randomized in three groups; 63 simply were observed and not given any additional therapy, 56 had an END three months after the melanoma had been excised, and 54 underwent END immediately after removal of the melanoma. In sixty-eight patients, the melanoma was positioned on the trunk and in the others the neoplasm was on an extremity. Patients whose melanoma was at the midline of the trunk and those whose melanoma was situated directly above a zone that housed nodes were excluded from the study. This precaution was taken in order to avoid including a region in which the flow of lymph could not be predicted accurately, a situation that would obviate a deleterious effect on results that might derive from having dissected the wrong nodal basin. In 1976, the authors were forced to this conclusion: "None of these regimens differed significantly from the others in its effect on length of survival or interval to metastasis." In 1986, the same authors reiterated that conclusion in these words: " . . . immediate lymphadenectomy seems to be of no significant benefit to the patient." Their study, as well as the one performed by Veronesi, et al., enabled definitive statements to be made about the worth (or lack of it) of END. But the story did not end there, as it should have; more retrospective studies, replete with flaws, were yet to be published in favor of and against the value END.

In brief, the studies by the groups of Sim and of Veronesi, both of them prospective and randomized, compelled those authors to conclude that END had no merit.

In 1987, Karakousis, et al.²⁸ published the results of a study engaged retrospectively of patients with melanoma, some of whom had undergone both wide local excision of the primary neoplasm and END. They found no difference in survival between patients who had had END and those who had not. A surprise to them was that patients who were treated for their melanoma more recently, that is, between 1976 and 1980, survived longer, irrespective of the treatment they received, than those who were managed between 1965 and 1970. The authors postulated that it was increased awareness in the recent past, that is, in the period between 1976 and 1980, of the implications of melanoma that made physicians more alert to diagnosis of it, thereby, leading to diagnosis of melanoma earlier and, as a consequence, to survival longer.

In 1994, Slingluff, et al.²⁹ inspected the charts of 4682 patients with melanoma. Of those, 3550 had no evidence of metastasis clinically. END was performed in 911 of those patients, of whom 214 (23%) had nodal metastasis. Some time after END had been completed, 71 patients whose nodes were negative manifested metastasis of melanoma in the nodal basin that had been dissected previously. The overall survival of all patients who underwent END was no different from those who did not. Analysis of sets showed that patients with melanoma less than 0.76 mm thick on an extremity had a slightly better survival if they had had END, whereas patients with melanoma thicker than 4.0 mm on the head and neck survived for a shorter time if they had undergone END. This galvanized the authors to the view that "Stratified by Breslow thickness and primary site, no significant improval in survival was observed with ELND."

In 1996 and again in 2000, Balch and coworkers^30,31 told of their experience in a study that was randomized and prospective. In brief, that study showed that overall no benefit accrued to patients who received an END. Analysis of sets did reveal some benefit, however. This analysis seemed to resuscitate END and for that reason, it will be examined thoroughly later in this work.

Cascinelli, et al.³², in 1998, shared the results of a study by them that was randomized and prospective, and undertaken between 1982 and 1989 under the auspices of the WHO Melanoma program. They investigated 240 patients with truncal melanoma that measured 1.5 mm or more in thickness. The patients were divided into two groups as follows: One of 122 patients whose melanoma was excised widely in conjunction with END and a second of 180 patients whose melanoma also was excised widely but who then merely were observed for development of metastasis. The regional nodes removed were selected on the basis of attempting to predict clinically the route that metastasis would likely take. Patients with melanoma at sites from which flow of lymph could not be predicted with confidence clinically were excluded from the study until, at a later time, lymphoscintigraphy became available (the authors did not make clear, however, when, exactly, lymphoscintigraphy was introduced). Five-year survival did not differ significantly between the two groups, being 61.7% for patients with END and 51.3% for those who only were observed. Cascinelli and coworkers made this statement by way of attempt at synthesis: "Our results from this randomized trial confirm the inefficacy of elective regional node dissection as routine treatment in all melanoma patients with a primary melanoma of the trunk thicker than 1.5 mm (p=0.09)." In what the authors called a "secondary outcome of the study," 36 patients who did not undergo END and during the course of the period of observation developed manifestations of metastasis clinically, were found to have a poorer outcome than 27 patients who underwent END that yielded positive nodes. The five-year survival was 48.2% versus 26.6%, with a level of significance of 0.04. On the basis of these findings, the authors advised that "Sentinel node biopsy may become a tool to identify patients with occult node metastases, who could then undergo node dissection."

The attitude of physicians to END has varied greatly since the procedure was undertaken initially about 100 years ago. The content of articles published about the subject prior to 1984 did not lend itself to deriving conclusions definitively about whether patients with melanoma benefited from END. Only 43% of American surgeons performed END for a melanoma between 1.5mm and 3.0mm in thickness.³³

Cady,³⁴ in September, 1984, sought to capture, definitively, the essence of studies that dealt with END and his opinion was communicated as follows: "Only two trials have been conducted that randomly allocated patients between prophylactic lymph node dissection and observation, and neither trial demonstrated a survival rate difference . . . These trials have been criticized for methodologic reasons . . ." Cady was referring to the trials performed by Veronesi, et al.^24,25 and by Sim, et al.^26,27 Of studies that were not randomized, he wrote thus: "Balch, et al. described patients from both Alabama and Australia, with significant advantages from elective regional lymph node dissection only in 27% of patients with intermediate-thickness lesions [that is, between 1.5 and 4.0mm in thickness] . . . [the study] demonstrated differences that were so large . . . that conceptually the advantage could not be postulated to result from the lymph node dissection alone . . . Other retrospective studies have failed to substantiate the assumptions of Balch, et al. and have not demonstrated survival rate differences between patients whose regional lymph nodes were initially observed and those patients whose regional lymph nodes were initially resected. Thus, although the final answer about the actual lack of advantage of prophylactic regional lymph node resection cannot be given yet in melanoma . . . " Despite the disclaimer, a concept was emerging clearly. Results of studies performed without randomization and, for the most part, retrospectively, were not in synchrony with one another, yet results of the two prospective studies were unequivocal; patients who underwent END derived no benefit from it.

Later in 1993, Scott and McKay³⁵ returned to the subject of END, focusing their attention on melanoma of the limbs, the reason for that being given in this way: " . . . limb primaries are common, lymphatic drainage is predictable and major studies have concentrated on this aspect." The authors, on review of the literature, stated their opinion thus: "ELND cannot be undertaken without risk of significant morbidity from lymphoedema[sic] and wound infection. Many patients undergoing elective lymphadenectomy will not need this procedure. Any advantage from ELND is likely to be small, even in selected subgroups. It is, therefore, reasonable to pursue a policy of therapeutic rather than elective lymphadenectomy. There is no evidence from the trials discussed that the removal of healthy nodes confers any advantage or disadvantage to the patient with malignant melanoma."

By the last decade of the 20th century, by virtue of the studies performed by Veronesi, et al.,^24,25 Sim, et al.,^26,27 Cascinelli, et al.³² and even Balch, et al.,^30,31 it appeared that sufficient, scientifically valid evidence was available to ring the knell for END. But the study performed by Balch and co-workers^30,31 in what was to be the era of sentinel lymph node biopsy (SNB) gave new hope, albeit illogically, to those who still believed in END and even to those who were losing faith!

In 1996 and again in 2000, Balch, et al.^30,31 published what are the only articles in the medical literature in which a prospective randomized study claims to demonstrate that benefit was obtained from END, although only in specific sets of patients. The article in 1996 was presented as a study still in progress and the article in 2000 as a study then completed, thereby enabling the authors to remark that "The data set is now matured to make definitive conclusions because the relapse rates for all stages have plateaued." The results provided in the articles came from a randomized trial conducted at several different institutions and that involved 740 patients with melanoma measuring 1–4.0 mm in thickness and no evidence of metastasis clinically. The goal of the study was twofold: (1) to determine whether END improves the rate of survival for patients with melanoma 1–4.0 mm in thickness compared to patients observed clinically only, and (2) to define sets of patients with melanoma who most likely would profit from END. Unlike some of the studies conducted previously, this one utilized lymphoscintigraphy performed preoperatively for the purpose of identifying all nodal basins at risk for metastases from a melanoma situated on the trunk, but not on the extremities or the head and neck. All nodal basins deemed to be at risk for metastasis were dissected. Statistical analysis demonstrated no difference in survival overall at both five and 10 years for patients who had had END and those who were observed only clinically for any sign of nodal involvement by metastasis of melanoma. A better and statistically significant survival rate was claimed for certain sets of patients who underwent END when the melanoma had the following attributes:

The coworkers, themselves, discounted all results pertaining to age and they did that in these words: "Although a survival benefit of ELND was demonstrated among patients 60 years and younger in this study, this clinical feature was not a prospectively stratified variable, so the conclusions about any surgical benefit are not as conclusive." The manner in which sets were analyzed by Balch, et al. gives the impression of an attempt to lobby on behalf of END in the face of three randomized and prospective studies that showed, conclusively, that END was not advantageous to patients.

In brief, the work of Balch and colleagues indicated that rate of survival for patients with melanoma that measured 1–4.0 mm in thickness is not improved significantly by END. Analysis of sets did seem to show that END benefits patients whose melanoma was not ulcerated or measured between 1 and 2.0 mm, or a melanoma situated on an extremity. Balch, et al. made these comments in regard to the study: " . . . prognostic factors (especially tumor thickness and ulceration) can prospectively identify melanoma patients at sufficiently high risk for occult regional metastases to justify surgical excision of their regional lymph nodes, and a sufficiently low risk for distant occult metastases to have a therapeutic benefit." In this statement, Balch, et al. reiterated their long-held belief that melanoma may metastasize only as far as nodes and no further, at least for a short period of time. Moreover, if any case can be made for END, the results of Balch and associates indicate that END would be appropriate only for those patients whose melanoma measures between 1 and 2.0 mm, is not ulcerated, or occurs on a limb. Balch and coworkers went on to aver that "Of these three groups [melanomas without ulceration, melanomas measuring between 1 and 2.0 mm in thickness and melanomas occurring on the limbs], we believe the most significant for clinical decision making is the presence or absence of melanoma ulceration." That contention is illogical; ulceration of melanoma may be the consequence of trauma and not an inherent characteristic of the neoplasm.

The initial report by Balch, et al.³⁰ on this cohort of patients was published in 1996 along with responses to it that had been solicited, at the 116th annual meeting of the American Surgical Association in Phoenix, Arizona, from colleagues considered to be expert about the subject of END. Because the results of both communications, that is, those published in 1996 and 2000, are so similar, the reactions to the initial report in 1996 is applicable equally to the final report published in 2000. For that reason, the dialogue between those expert physicians and Balch are quoted, in part, now.

Hiram C. Polk, Jr. raised the question of complications associated with END as follows: "1) What was the acute morbidity associated with the node dissection . . . ? 2) An occasional to rare patient will die after this operation. That obviously needs to be defined. . . ." And Murray F. Brennan followed that same line of thought in these words: " . . . you have shown us that there is no overall survival benefit, that the benefit is to the subgroup under the age of 60 . . . help us by telling about the morbidity of the procedure of elective dissection so that we can make decisions for our own patients . . . "

Critique by us of the ideas of Balch, et al.: In reply to the comments by Polk and Brennan, Balch told of an incidence of 21% and 11% for wound infection or wound disruptions in femoral and axillary node dissections, respectively, and 20% for lymphedema. He did not address the question of acute morbidity and mortality related to the procedure of END itself. Curiously, pathetically little has been written about the matter of the morbidity of the procedure, even though complications of it are well known to surgeons, some of whom consider the risks associated with END to outweigh possible benefit from it. That position deserves commendation because, in reality, END has no benefit whatsoever for a patient. In a study by Sim and coworkers,²⁷ published in 1986, complications of END were detailed as follows: 32 of 171 patients developed lymphedema, which was marked in five patients; one patient, adhesive capsulitis of the shoulder; 15 patients, seromas, 12 in the groin and three in the axilla; seven patients, delayed healing of the incision in the inguinal region; one patient, serosanguineous drainage from the incision site and one patient a hematoma, and one patient necrosis of a skin flap and another an infected wound, both patients requiring surgical revision of the site. In brief, END is not an innocuous procedure and, that being the case, should not be undertaken unless there is compelling evidence that it bestows benefit—and there is none.

Brennan aired his concern thus: "You made no mention of whether or not the benefit was to those patients who had positive lymph nodes." Implicit in the statement is the fact that END could only benefit patients with metastasis restricted to a specific nodal basin and that, in actuality, the procedure is redundant in those patients who do not have metastasis. Balch answered as follows: " . . . the nodal incidence increases from 17% for 1-to-2- thickness groups to 37% for 3- to 4-mm groups and was higher in patients with tumor ulceration."

Critique by us of the ideas of Balch, et al.: Balch informed of the rate of metastasis to nodes, but did not respond to the matter of whether END benefited patients with metastatic melanoma, although the results of the study implies that it does confer benefit. Nonetheless, in both the analysis of 1996 and 2000, no comparison is made of survival overall for patients with metastasis only, even though that data surely was available to Balch and his coworkers. Is that because patients with metastasis have the same survival overall irrespective of whether or not, and when, they undergo lymphadenectomy? The evidence provided by the randomized studies of Veronesi, et al., Sim, et al. and even Balch, et al. suggests that it would be the same.

Donald L. Morton: " . . . [We performed] a prospective review of ELND . . . We found that the 5-year and 10-year survival rates were 9% and 13% higher, respectively, for patients whose primary melanomas were 1.2 to 3.5 mm; patients with thinner or thicker lesions did not benefit. Overall, males undergoing ELND had a 17% and 19% survival benefit at 5 and 10 years, respectively; males under 60 had a 14% and 15% survival benefit, respectively. Both males and females over 60 years old benefited from ELND, but females under 60 did not benefit."

Critique by us of the ideas of Balch, et al.: It is instructive to compare and contrast the results of Morton with those of Balch, et al., the latter who endorsed benefit from END to patients whose melanoma was 1–2.0 mm thick, not ulcerated, or present on an extremity. According to Balch, et al., benefit derived, too, from END for patients who were younger than 60 years. No difference in gender was commented on by them. These results so disparate from one another in regard to outcome following analysis of multiple sets, is seen also in studies undertaken retrospectively, for example, those performed by Milton, et al.¹⁸ and Balch, et al.^15,16,17 Analysis of sets may yield conclusions that are plain wrong as Balch, et al. admitted to in 1996 in the discussion of their findings, and that surely is one of the reasons such different outcomes are reported on in studies. Moreover, Morton did not indicate whether the data he presented rose to a level of significance statistically. Balch, et al. were only able to demonstrate statistical significance in their study by doing an analysis of sets. Even then, significance is not compelling.

Blake Cady: "In order for results to be accepted . . . they must be biologically plausible. The only biologically plausible explanation for why lymph node dissection improves survival is to assume the lymph node metastases are generating foci for further systemic metastases not already generated by the primary melanoma . . . My questions are: If this represents a biological phenomenon, why doesn't it work . . . greater than 2 or 3.0 mm in thickness?" In response Balch gave this explanation: "Theoretically, there is a window of time in the biology of melanoma where patients have isolated regional metastasis, which if removed surgically, may prevent further dissemination of the disease. However, as the disease progresses it reaches a point where distant metastases have seeded from either regional metastasis or from the primary melanoma that negates the benefit of the regional operation."

Critique by us of the ideas of Balch, et al.: The concept that a melanoma metastasizes to a single node, is confined to that node for a time, and that node then becomes the sole source of additional metastases boggles because it is so simplistic, to say nothing of being dead wrong. Cells of melanoma may pass through or bypass nodes, just as do other malignant cells,³⁶ inflammatory cells such as macrophages, and lymphocytes, erythrocytes, bacteria,³⁴ and foreign material.^{37, 38}

William C. Wood noted that of 325 patients, approximately 45 patients whose nodes were positive underwent END and approximately 65 patients with positive nodes were just observed. He asked this question of Balch: " . . . Do you not believe that this study is underpowered to address the very question toward which it was targeted?" Balch replied that it took seven years to accumulate the number of patients necessary for the study to qualify as being valid statistically. He claimed that the design of the study and the analysis of the data confirmed that, in fact, the study was valid. Exploring further the theme relating to the significance of those findings, Roger S. Foster made the following observation: "The biologic behavior of melanoma has always been puzzling, and I fear it has become more puzzling if we accept at face value the current subset analysis of these patients." Balch did not respond to that particular observation. In the same spirit as Wood and Foster, Cady commented as follows: " . . . If you torture data long enough, they will confess to anything." For us, all of this calls to mind the maxim attributed to Mark Twain to the effect that "There are three kinds of lies: lies, damn lies, and statistics."

The questions posed to Balch by those reputed to be expert in the field call into question seriously the legitimacy of END for treatment effectively of a patient with metastatic melanoma. Morbidity from the procedure was an issue to boot, and complications episodically were undeniable. In sum, the most recent study of Balch, et al. must be assessed in the context of other randomized studies that showed, repeatedly, that END has no rightful place in the treatment of melanoma, that it cannot effect cure of a patient with metastasis of melanoma, and it has no measurable influence on longevity. Veronesi, et al.,²⁵ in 1982, had taken on themselves an analysis of sets, which included melanomas of different thickness, similar to the one undertaken by Balch and coworkers in 1996 and 2000, but the results of that study indicated that there was no advantage for survival in any set of patients. Rather than the findings in the study by Balch, et al. compelling to a conclusion that END has benefit for patients with metastasis of melanoma, they do just the reverse.

In 2001, Reintgen and associates,³⁹ in what seems to be a deliberate attempt to negate the results of the prospective and randomized studies performed by Sim and Veronesi and their collaborators and to bolster those of Balch, et al., trumpeted the following: "The old trials of elective lymph node dissection (ELND) (WHO trials, #1, 13, Mayo Clinic Trial) are un-interpretable in light of what we know about the adequate power of trials and the importance of defining the cutaneous anatomy with the use of lymphatic mapping principles. None of these old trials involved pre-operative lymphoscintigraphy to define basins at risk for metastases, and thus the wrong surgery was performed in up to 32% of the patients. The only trial that incorporated the lymphatic mapping principles in the design of the trial was the Balch Intergroup study . . . "

In fact Balch, et al. had performed lymphoscintigraphy prior to the END on only 152 patients, all of whom had melanoma on the trunk; 227 patients had melanoma on an extremity, the head, or the neck, and they were subjected to END in the absence of a lymphoscintigram. In brief, only 40% of patients who underwent END had a nodal basin defined by lymphoscintigraphy. If the number "32%" invoked by Reintgen, et al. in their statement regarding selection of a nodal basin for END is correct, then 73 patients, that is, 19% of patients in the study of Balch, et al. may have had a dissection of the wrong nodal basin, which is undermining to any analysis. If that truly is the case, then Reintgen, et al., in a spirit of equitableness should characterize that work, too, as "un-interpretable." Moreover, if lymphoscintigraphy is requisite for a valid study, then the study of Balch, et al. reinforces the idea that END delivers no benefit to a patient with metastatic melanoma because, in that very study, patients with melanoma on the trunk who had lymphoscintigraphy were not benefited by END.

In 2002, Fisher⁴⁰ analyzed retrospectively 1444 patients with melanoma of the head and neck, and found that no benefit accrued from END. Some patients were just observed, whereas others had dissection of nodes, either electively or once signs clinically of what likely was metastasis appeared. Not only was there no benefit from END for patients whose nodes were positive for melanoma, but the outcome for them actually was worse than that for patients who were spared END. The author interpreted the findings in this way: "Data from this paper does[sic] not support the thesis that ELND has a positive impact on survival but suggests the contrary, i.e., survival seemingly is better in patients undergoing DLND [dissection of nodes at the time that metastasis becomes apparent clinically] as opposed to ELND with occult disease even when clinical parameters are stratified and multivariate analysis is used." The outcome referred to was significant statistically, the p value being equal to 0.01.

The studies alluded to thus far in this essay represent a selection, judiciously, of a huge body of literature that has been given to the subject of END. After all the verbage, the sophistry, and the promotion of positions that have associated with them secondary gain, the crucial question remains: Does END serve any purpose whatsoever for of a patient with metastatic melanoma? In our judgment the answer is a resounding "No!" and the reason for that conclusion follows here.

The concept of END being essential to treatment of melanoma was set forth originally by Snow in the 19th century and was given impetus by Halsted and Handley early in the 20th century. Despite the utter lack of proof of its effectiveness, END continued for decades to be practiced as a valuable operation, but the basis for it was devoid of logic and rationality. The aim of END was to extirpate nodes in the event that they might harbor a metastasis of melanoma. Although the results of many published studies were said to indicate that END confers an advantage in regard to survival of patients, those studies were not randomized and prospective and, therefore, were meaningless because they were biased. Nonetheless, those studies lent credence to END and support to the necessity of performing it. To this day, some oncologists and surgeons still advocate END for treatment of patients whose melanoma measures between 1 and 4.0 mm in thickness. That recommendation is derived from the conclusions drawn from studies by Breslow^13,14 and by Balch and co-workers,¹⁵ persons deemed to be "important" in the field. Despite evidence to the contrary, in their recent publication, Balch, et al.^30,31 advocate END as being of use only in patients whose melanoma measures 1 to 4.0 mm in thickness and is not ulcerated.

Intuition is the domain of psychics, diviners, and fortune-tellers. Physicians are supposed to be scientists who think in a critical, incisive, logical fashion and who make judgments in the best interests of patients. The evidence now available from studies of END is incontrovertible: All of them that are randomized and prospective^{24,25,26,27,30,31,32} show no benefit whatsoever for survival overall. In actuality, END should have been jettisoned as early as the 1970's when it was known beyond doubt that no evidence for benefit from it could be marshalled¹¹ on behalf of it. Eggermont⁴¹ got it right in the last year of the 20th century when he said this directly and unflinchingly: "Routine ELND is overtreatment in 80% of the patient population and is even without convincing proof of impact on the survival of the remaining 20%. Thus ELND must be abandoned." He is as right now as he was then; there was never justification for END, either in rational theory or in real practice.

Despite the fact that after more than 50 years no convincing evidence has been generated in favor of efficacy of END, the subject of SNB awakened interest anew in END, albeit with a minor modification, namely, the patients now selected for lymphadenectomy must have a proven metastasis, not simply a risk for metastasis. If END were to have any benefit to patients, it could only have been to those whose nodes harbored cells of melanoma. But that it did not benefit them is apparent from a reading, critically, of the literature about the subject. Advocates of the more recent modification of END, that is, nodal basin dissection consequent to a positive SNB, insist, nonetheless, that by removing nodes that actually contain cells of melanoma, survival is enhanced and even cure will be accomplished. The reason given for that contention is the claim that cells of melanoma may exist only in a nodal basin, without any spread of those cells beyond the nodes; by extirpating the nodes, the neoplastic cells confined to them are prevented from traveling to distant sites. The surmise is wrong patently; no metastasis of melanoma ever is restricted to a single node or even to a single nodal basin. A metastasis, by definition, signifies "distant" spread always.