Quality Control in Sentinel Lymph Node Biopsy in Cervical Cancer

In the current issue, Altgassen et al report on a prospective multicenter study comprising the largest number of cervical cancer patients in a single study undergoing sentinel lymph node (SLN) mapping. The authors analyzed 507 patients between December 1998 and October 2006 and included patients with cervical cancers of all stages. Although the majority of participating centers used both blue dye and technetium-99 for the SLN procedure, the protocol allowed either alone. The SLN and the nodes from the systematic lymphadenectomy were submitted for routine staining (hematoxylin and eosin); no ultrastaging was performed. It is important to note that the detection rate was calculated based on the number of patients with at least one detected SLN. The detection rate of pelvic SLN overall was 88.6% (95% CI, 85.8% to 91.1%) and was higher if a combination of technetium and patent blue was used (93.5%; 95% CI, 90.3% to 96%). Overall sensitivity of pelvic SLN detection was 77.4% (95% CI, 68.2% to 85.0%), which was lower than the predefined noninferiority margin of 90%. The subgroup analyses showed that the sensitivity of SLN detection was higher for patients with tumors 20 mm (90.9%) and with bilateral detection (87.2%). The authors have concluded that systematic lymphadenectomy in patients with cervical cancer should not be omitted at this time. Since the introduction of sentinel node mapping in early-stage melanomas by Morton et al, the SLN procedure has replaced systematic lymphadenectomies in both melanoma and breast cancer populations. More recently, among gynecologic malignancies, SLN mapping with blue dye and lymphoscintigraphy has been studied in both vulvar and cervical cancers. There have been many singleinstitution series in the literature reviewing their experiences with SLN mapping in cervical cancer, demonstrating high detection rates from 86% to 100%. In a recent systematic review, the sentinel node detection rate with the combined technique was 97%, with a sensitivity of 92%. However, the most recent multicenter study in this issue by Altgassen et al did not achieve similar detection and sensitivity rates as those previously published. The authors should be congratulated for accomplishing and advancing the SLN procedure in clinical practice by conducting a large multicenter study. However, this study raises further questions. Why did this promising procedure have less than the expected results? Will a systematic lymphadenectomy ever be omitted in cervical cancer? Will change in practice only be affected by a randomized trial? The results of this study highlight how the success of SLN detection depends on technique, individual surgeon experience, and the patient population chosen for this procedure. This multicenter study is a good example of how excellent results may be diluted if stringent criteria are not widely applied to a general population. The technical aspect of the SLN procedure plays a key role in maximizing results. Quality assurance, including physician training and experience, methods of detection, and central pathology review for both the primary tumor and lymph nodes are essential components to maximize success. For example, the multicenter study of SLN in vulvar cancer by van der Zee et al required each center to have successful experience with sentinel nodes in at least 10 vulvar cancer patients, use both blue dye and technetium, and submit the SLN for pathologic assessment by a protocol that included ultrastaging. Without stringent quality control measures, the overall sensitivity of the procedure can be significantly reduced. Although this study’s strength in patient numbers could only be achieved by including multiple centers, if not all centers had proven experience with the SLN procedure, the decreased detection rate could be a result of the obligatory learning phase, the definition of detection used, and lack of the other quality controls listed earlier. Altgassen et al accrued approximately 600 patients over 7 years from 18 centers. On average, that is less than five patients per year per center and potentially even less per surgeon. Although the ideal number of procedures required in cervix cancer has not been identified, Morton et al suggested a learning phase of at least 30 consecutive patients per center for cutaneous melanoma, whereas De Hullu et al suggested at least 20 consecutive patients for vulvar cancer. Time to perfect the technique, not only with the operative procedure itself, but also in conjunction with nuclear medicine and pathology, is important to consider before dismissing the sentinel node procedure. It would be interesting to know whether the accuracy was higher in the second half of the patients compared with the first half. The main theoretical benefit of the SLN procedure is to reduce the requirement for a complete lymphadenectomy with its associated morbidity in a patient population at low risk for lymph node metastases. In early-stage cervical cancer, the incidence of pelvic lymph node metastasis is approximately 10%; therefore, if the SLN is negative, 90% of patients can avoid a full pelvic lymphadenectomy and the associated morbidity of blood loss, neural injury, lymphocyst, and lymphedema. It is unclear whether more advanced cancers can obtain the same benefit and whether the SLN concept is valid in such patients. Altgassen et al included all stages of cervical cancer (20% of patients had stage IIA or greater). The incidence of pelvic lymph node involvement is approximately 25% to 30% for stage II cervical cancers and higher for subsequent stages. These tumors are also likely to present with adverse primary tumor features such as clinical tumor size more than 3 cm, depth of invasion more than 10 mm, and presence of capillary-lymphatic space involvement. Therefore, the likelihood of recommending adjuvant therapy postoperatively is high, and the usefulness of an SLN procedure in morbidity reduction in this JOURNAL OF CLINICAL ONCOLOGY E D I T O R I A L VOLUME 26 NUMBER 18 JUNE 2

Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.