A landmark study of advanced endometrial cancer, GOG 258, was published in the New England Journal of Medicine this summer.1 This clinical trial compared the use of carboplatin and paclitaxel chemotherapy with a combination of chemotherapy with external beam radiation, exploring the notion of “more is better.” The results of the trial revealed that the “more” (chemotherapy with external beam radiation) was no better than chemotherapy alone with respect to overall survival. These results have challenged a creeping dogma in gynecologic oncology, which has seen many providers embrace combination therapy, particularly for patients with stage III (node-positive) endometrial cancer, a group of patients who made up approximately three-quarters of GOG 258’s study population. Many have been left searching for justification of their early adoption of combination therapy before the results of a trial such as this were available. For me it also raised a slightly different question: In the light of these results, what IS the role of para-aortic lymphadenectomy in the staging of endometrial cancers? If radiation to the nodal basins is no longer part of adjuvant therapy, then
It was in the 1980s that the removal of clinically normal para-aortic lymph nodes (those residing between the renal and proximal common iliac vessels) became a part of surgical staging. This practice was endorsed by the International Federation of Gynecology and Obstetrics (FIGO) and the Gynecologic Oncology Group (GOG) surgical committee in response to findings that 11% of women with clinical stage I endometrial cancer had microscopic lymph node metastases which were discovered only with routine pathologic evaluation of these tissues. Among those with pelvic lymph node metastases (stage IIIC disease), approximately one-third also harbored disease in para-aortic nodal regions.2 Among all patients with endometrial cancer, including those with low-grade disease, only a small fraction (approximately 2%) have isolated para-aortic lymph nodes (positive para-aortic nodes, but negative pelvic nodes). However, among patients with deeply invasive higher-grade tumors, the likelihood of discovering isolated para-aortic metastases is higher at approximately 16%.3 Therefore, the dominant pattern of lymph node metastases and lymphatic dissemination of endometrial cancer appears to be via the parametrial channels laterally towards the pelvic basins, and then sequentially to the para-aortic regions. The direct lymphatic pathway to the para-aortic basins from the uterine fundus through the adnexal lymphatics misses the pelvic regions altogether and may seen logical, but actually is observed fairly infrequently.4
Over the subsequent decades, there have been debates and schools of thought regarding what is the optimal degree of lymphatic dissection for endometrial cancer staging. Some advocated for full pelvic and infrarenal para-aortic nodal dissections in all patients, including even those in the lowest risk for metastases. Others advocated for a more limited, inframesenteric para-aortic nodal dissection, although the origins of such a distinction appear to be largely arbitrary. The inferior mesenteric artery is not a physiologic landmark for lymphatic pathways, and approximately half of para-aortic metastases are located above the level of the inferior mesenteric artery. This limited sampling may have been preferred because of relative ease of dissection rather than diagnostic or therapeutic efficacy.
As the population became more obese, making para-aortic nodal dissections less feasible, and laparoscopic staging became accepted as the standard of care in endometrial cancer staging, there was a further push towards limiting the scope of lymphadenectomy. Selective algorithms, such as the so-called “Mayo clinic criteria,” were widely adopted. In this approach, gynecologic oncologists perform full pelvic and infrarenal para-aortic lymphadenectomies but only in the presence of a high-risk uterine feature such as tumor size greater than 2 cm, deep myometrial invasion, or grade 3 histology.3 While this reduced the number of para-aortic dissections being performed, it did not eliminate them, as approximately 40% of patients with endometrial cancer meet at least one of those criteria.
At this same time, we learned something else critical about the benefits, or lack thereof, of lymphadenectomy. Two landmark surgical-staging trials were published in 2009 which randomly assigned women to hysterectomy with lymphadenectomy or hysterectomy alone, and found no survival benefit for lymphadenectomy.5,6 While these trial results initially were met with noisy backlash, particularly from those who had legitimate concerns regarding study design and conclusions that reach beyond the scope of this column, ultimately their findings (that there is no therapeutic benefit to surgically removing clinically normal lymph nodes) has become largely accepted. The subsequent findings of the Laparoscopic Approach to Cancer of the Endometrium (LACE) trial further support this, as there was no difference in survival found between patients who were randomly assigned to laparoscopic versus open staging for endometrial cancer, even despite a significantly lower rate of lymphadenectomy among the laparoscopic arm.7
SLN biopsy, in which the specific nodes which drain the uterus are selectively removed, represents the most recent development in lymph node assessment for endometrial cancer. On average, only three lymph nodes are removed per patient, and para-aortic nodes infrequently are removed, because it is rare that lymphatic pathways drain directly into the aortic basins after cervical injection. Yet despite this more limited dissection of lymph nodes, especially para-aortic, with SLN biopsy, surgeons still observe similar rates of IIIC disease, compared with full lymphadenectomy, suggesting that the presence or absence of lymphatic metastases still is able to be adequately determined. SLN biopsy misses only 3% of lymphatic disease.8 What is of particular interest to practitioners of the SLN approach is that “atypical” pathways are discovered approximately 20% of the time, and nodes are harvested from locations such as the presacral space or medial to the internal iliac vessels. These nodes are in locations previously overlooked by even the most comprehensive pelvic and para-aortic lymphadenectomy. Therefore, while the para-aortic nodes may not be systematically removed with SLN biopsy, new and arguably more relevant regions are interrogated, which might explain its equivalent diagnostic virtue.
With this evolution in surgical-staging practice, what we have come to recognize is that the role of lymph node assessment is predominantly, if not exclusively, diagnostic. It can help us determine which patients are at risk for distant relapse and therefore candidates for systemic therapy (chemotherapy), versus those whose risk is predominantly of local relapse and can be adequately treated with local therapies alone, such as vaginal radiation. This brings us to the results of GOG 258. If defining the specific and complete extent of lymph node metastases (as if that was ever truly what surgeons were doing) is no longer necessary to guide the prescription and extent of external beam radiation, then lymph node dissection need only inform us of whether or not there are nodal metastases, not specifically the location of those nodal metastases. The prescription of chemotherapy is the same whether the disease is limited to the pelvic nodes or also includes the para-aortic nodes. While GOG 258 discovered more para-aortic failures among the chemotherapy-alone group, suggesting there may be some therapeutic role of radiation in preventing this, it should be noted that these para-aortic relapses did not negatively impact relapse-free survival, and these patients still can presumably be salvaged with external beam radiation to the site of para-aortic relapse.
It would seem logical that the results of GOG 258 further limit the potential role of para-aortic lymphadenectomy in women with clinical stage I disease. Perhaps para-aortic dissection can be limited to women who are at highest risk for isolated para-aortic disease, such as those with deeply invasive high-grade tumors not successfully mapped with the highly targeted sentinel node biopsy technique? Most clinicians look forward to an era in which we no longer rely on crude dissections of disease-free tissue just to prove they are disease free, but instead can utilize more sophisticated diagnostic methods to recognize disseminated disease.
Dr. Rossi is assistant professor in the division of gynecologic oncology at the University of North Carolina at Chapel Hill. Email her at.