Pii: S1010-7940(99)00174-8

Surgeons have always been interested in anatomic structures that can be visualized, palpated, dissected and resected. Because lymph nodes have all of these qualitative characteristics, they have attracted the interest of surgeons since the late 19th century. In fact, oncologic signi®cance has been attributed to them and resection has been advocated and even mandated as being an essential component for the treatment of patients with carcinoma. Unfortunately, this theory was predicated more upon intuition than scienti®c evidence. Nevertheless, as we enter the 21st century, surgeons still remain adamant in believing that lymph node resection can bene®t and even cure patients with cancer. The Barrier Theory, postulated by 19th century surgeons, maintains that lymph ̄ows in a consistent, predictable sequential manner from nodal station to nodal station where malignant cells are impeded or detained. In reality, lymph ̄ows in a haphazard, bizarre, unpredictable manner skipping many nodal stations. Malignant cells pass through lymph nodes as if they were a sieve, and because of in®nite lymphaticovenous communications, they can easily enter the vascular system achieving diffuse and distant dissemination. Radical lymphadenectomy is not an innocuous procedure since it requires a major dissection that interrupts and damages neurogenic, vascular and lymphatic structures in the mediastinum. Neurogenic interruption can cause pulmonary vascular spasm that physiologically reduces vascular volume simulating cor pulmonale and can increase right heart burden. Cardiac output can be decreased and arrythmias incited. A concurrent lobectomy, which anatomically removes a large vascular volume, contributes further to these detrimental occurrences. Bleeding requiring transfusion is not uncommon. Administration of blood can have adverse affects upon survival for patients with cancer, possibly, by suppressing the immune system. Radical lymphadenectomy requires a massive dissection that injures tissue causing the release of cytokines, growth factors, etc. which are thought to enhance tumor growth [1]. Finally, the dissection must end somewhere in the periphery leaving open, lymphatic channels that can pump malignant cells onto healing, reparative tissue. This makes a superb substrate for the growth of tumor. Our current staging system is inaccurate, inconsistent and unreliable because it is more concerned with anatomy and morphology than the pathophysiology of the tumor. Numbering the lymph node stations is comparable to numbering the bones for metastases and trying to imply that one bone has a worse prognosis than another bone. Yet, we do this with lymph nodes as if it were scienti®c dogma. One malignant cell in any lymph node connotes metastasis and a very poor prognosis. Current pathological evaluation of lymph nodes is gross and inaccurate. Only a small portion of the node is evaluated microscopically leaving most of the specimen unexamined. Hematoxylin and Eosin (H and E) staining is crude, insensitive and will frequently fail to diagnose single malignant cells [2]. Reverse transcriptase polymerase chain reaction is much more sensitive and can identify malignant cells when H and E is negative. Unfortunately, it is still used very infrequently. In 1988, Naruke published a provocative study implying better outcomes when radical lymphadenectomy was included with resection of the lesion for patients with lung cancer [3]. He compared patients who could be resected to those who were unresectable because of advanced disease. Criticisms have been raised questioning the validity or fairness in comparing two such dissimilar cohorts of patients. Notably, eight surgical deaths occurred in patients with Stage I disease. Certainly, they would have had a better outcome without radical lymphadenectomy. Recently, Dr Naruke's criteria for radical lymphadenectomy have become more stringent limiting its application to a smaller number of patients. The commentary, by McKneally, of the 1988 publication concluded that ``we must accept the verdict of not proven for the therapeutic ef®cacy of systemic lymphadenectomy.'' I believe that statement is even more valid in 1998. Martini reported a 5-year survival of 30% following resection of N2 nodes [4]. Of 706 patients with N2 disease only 151 could be completely resected and only 45 survived for 5 years. Interestingly, some of these patients had occult disease which could be diagnosed only post-operatively by European Journal of Cardio-thoracic Surgery 16 (Suppl. 1) (1999) S11±S12