Original Title: Biopsia de ganglio centinela en carcinoma epidermoide de cavidad oral y labio: efectividad, seguridad y coste-efectividad en el diagnóstico de metástasis ocultas

Authors' objectives: The main objective of the report is to study if SLNB is an effective and safe diagnostic method for the study of occult cervical metastases in oral cavity and lip cancer, compared to cervical lymph node dissection. The secondary objectives include studying the quality of life of the patients and reviewing if the SLNB is cost-effective, compared to the alternatives of cervical dissection and “wait and see”; and assess the organizational impact of the implementation of SLNB in hospitals.

Authors' results and conclusions: RESULTS: From the 394 references initially retrieved in the search for effectiveness studies (July 2020), 21 effectiveness studies of the SLNB in OSCC cT1/ T2N0 and 2 studies on the quality of life of patients were finally selected. Subsequent searches (March and April 2021) allowed us to include another 4 original effectiveness studies and 1 more on quality of life. The search for economic studies retrieved 48 references, from which, finally, 4 were selected. In total, 25 original effectiveness studies, 3 on the quality of life of patients and 4 economic studies have been included. The 6 systematic reviews and meta-analyses and 9 clinical guidelines or consensus documents published since 2016 have also been reviewed. Diagnostic and therapeutic effectiveness studies: Of the 25 studies, 2 were randomized clinical trials and the rest were observational studies. A non-inferiority, controlled and randomized phase III trial, carried out in 16 hospital centers in Japan, with the objective of comparing SLNBguided ND vs END in patients with OSCC cT1/T2N0, in terms of survival, cervical functionality and adverse events associated with both procedures. The presence of ITCs was considered positive for metastases. The OS at 3 years in the SLNB group was 87.9% (lower limit of the 95% CI of 82.4%) which was not lower (12% non-inferiority margin) than that of the ND group (86.6%; lower limit of 95% CI of 80.9%), p for non-inferiority < 0.001. The DFS at 3 years in the SLNB group was 78.7% (lower limit of the 95% CI of 72.1%), not lower than that of the ND group (81.3%; lower limit of the 95% CI of 75.0%), p for non-inferiority < 0.001. Recurrence was observed in 18.2% and 21.6% in the END and SLNB groups, respectively. Recurrences were local and/or regional and no statistical differences were found between the two groups. The rate of FN was 10.5% in the ND group and 15.1% in the SLNB group. Functionality (arm abduction test and neck dysfunction quality of life questionnaire at 1, 3, 6 and 12 months) was significantly better in the SLNB group than in the ND group. Neck stiffness was worse in the first month after ND but improved over time, as did the other functionality items. Adverse events were identified in 14.6% patients in the ND group and in 9% in the SLNB group, with no significant differences between both groups. The second RCT was a randomized, open-label, equivalence trial carried out in 10 centers in France with the aim of studying the equivalence between SLNB vs END. No statistically significant differences were found in the rate of lymph node recurrence between both arms: 10.1% in the ND group vs 9.3% in the SLNB group (p=0.82). Neither were differences found in cervical lymph node RFS, locoregional RFS, DSS or OS between the two groups. Hospital stay length was longer in the ND group (p<0.1) Selfreport questionnaires on neck-shoulder impairment showed significantly worse ratings for the ND group at 2, 4, 6, and 12 months after surgery, but not at 24 months. In the arm abduction test, a significantly lower percentage of patients in the ND group reached 180º without pain or effort at 2, 4, 6 months, but there were no differences at 12 and 24 months. The physiotherapy prescription rate was significantly higher in the ND group at 2, 4 and 6 months, but similar in both groups at 12 and 24 months. The observational studies were 13 prospective and 10 retrospective. In total, they included 2,811 patients with early-stage OSCC (1,556 men and 1,112 women), with a median age of 46 to 65.6 years; 56.3% had tongue cancer, 24.1% had FOM cancer and 3.3% had lip cancer; 49.2% were cT1 and 27.3% were cT2. In most, the SLNB was performed only with a radiotracer, acquiring planar and tomographic images; the histopathological study included IHC and SSS analysis. The follow-up time for these studies ranged from 19 to 92 months. The GCs detection rate ranged between 93.9% and 100%; the median number of GCs detected per patient ranged from 1.9 to 4.8. The percentage of positive SLNs for occult metastases ranged between 11.4% and 50%, although the majority (11 of the 19 that reported this data) presented percentages between 21% and 37%. For these patients, the SLNB allowed a precise staging, increasing the tumor stage by confirming that these patients were N1 and not N0. In total, the Se of the SLNB has been 71.4%-100%, the Sp was 83.3%-100%, the PPV was 60%-100% and the NPV was 90.7%-100%. The FN rate ranged between 0% and 28.6%. A DSS between 85.5% and 96.0% and a DFS of 81% to 95% have been reported. The OS at 3 years (data from 3 studies) in patients treated with SLNB ranged between 82.0% and 89.5% and the OS at 5 years (data from 6 studies) of the SLNB, between 77.3% and 97.5%. Only 6 studies included data on the safety of SLNB. The number of reported complications has been very low. In general, the morbidity associated with ND has been significantly higher. Meta-analysis of 17 studies was performed. Heterogeneity was low for Se and negative LR, and high for Sp and positive LR. The aggregated or summary values were the following: Se of 0.86 (95% CI: 0.82-0.89), Sp of 0.99 (95% CI: 0.96-1.00), positive LR of 86.0 (95% CI: 23.4-315.8), negative LR of 0.14 (95% CI: 0.11-0.18) and DOR of 600 (95% CI: 148-2.433). The SROC curve reached a value of 0.90 (95% CI: 0.81-0.95). Quality of life studies: Three studies of quality of life in health of patients with OSCC cT1/ T2N0 were included, each of them with different measurement tools (EQ-5D-3L, EORTC QLQ-C30 and QLQ-H & N35, and UW-QOL) and with different comparison groups. In one study, two groups of patients were compared (positive SLNB result and negative SLNB result). Various scales (QLQ-C30, QLQ-H & N35) changed significantly. A high level of psychological distress (HADST> 14) was detected at baseline in 35% of patients, in 27% at one week, and in 20% at six-month of follow-up in patients with a negative SLNB, while in patients with a positive result, it was 8% at all times. These differences were not statistically significant. There were also no significant differences in anxiety and depression between the two groups. Only one study compared quality of life outcomes of patients undergoing SLNB (n=15) vs those undergoing END (n=9). The data from the UW-QOL questionnaire showed lower scores in the domains of appearance (p=0.035) and chewing (p=0.041) in the END group compared to the SLNB group; as well as a decrease of approximately 10% in the final mean UW-QOL (p=0.025). The results of the last 7 days, quality of life was classified as "good" in 53.3% in the SLNB group compared to 66.7% in the END group. No patient undergoing END reported an "excellent" quality of life compared to 26.7% in the SLNB group. 40% of patients undergoing SLNB reported that their current quality of life was "excellent" or "very good." In two studies, shoulder disability was also evaluated using the SDQ questionnaire. In the first study, the adjusted mean scores of the SDQ in the patients undergoing watchful waiting, SLNB, SOHND and MRND were: 8.64, 10.57, 18.92 and 33.66, respectively, and the differences were statistically significant. Patients with more shoulder discomfort appeared to have a lower quality of life. In the second study, the objective measurement of the shoulder function for certain movements in a sample of patients, showed no differences between the neck categories (no surgery, SLNB and ND). Economic evaluation studies: Four economic studies were selected, two were cost studies, one was a costeffectiveness study and the other was a cost-utility study. In the most recent cost study, TN patients in the SLNB group were associated to a 42% lower cost than those in the END group. However, patients with pN+ had a 23% higher cost in the SLNB group. SLNB appears to be more cost-effective when the patient showed a good follow up without neck recurrences. In the second cost study, of a cohort of 481 patients with OSCC, the SLNB option turned out to be less expensive (with a weighted mean relative cost ratio of 0.91) compared to the traditional surgical approach, even though 100% of the SLNB were positive. The cost-effectiveness study evaluated 5 diagnostic and treatment strategies for OSCC cT1/T2N0 (END, watchful waiting, PEG followed by ND or watchful waiting, SLNB followed by ND or watchful waiting, and PEG and SLNB (for positive PEG) followed of ND or watchful waiting; a Markov decision analysis model was used. SLNB followed by ND or watchful waiting was the most cost-effective strategy. Compared to END, the incremental cost-effectiveness ratio was € 3,356 per QALY gained. The uncertainty analysis showed that the model was sensitive to variations in the percentage of occult metastases and quality of life. SLNB showed the highest probability (66%) of being cost-effective compared to the other strategies. Finally, the cost-utility study, using a Markov model, compared the following strategies: a) ultrasound-guided FNAC followed by ND or RT if theresult was positive and wait and see when the result was negative, b) SLNB followed by ND or RT when the result was positive and wait and see if the result was negative, c) ultrasound-guided FNAC followed by ND and, if negative, SLNB followed by ND or RT if the result was positive, and d) END. Utilities were measured using the EQ-5D questionnaire. END was the most expensive strategy, with a cost ranged from € 18,007-18,917, depending on the time horizon. The SLNB strategy showed a higher QALY result in a time horizon of 5 or 10 years (3.79 and 6.70 QALY, respectively). However, with a lifetime horizon, the ND strategy had the highest QALY value (12 and 16 QALYs, respectively). CONCLUSIONS: The SLNB for cervical staging of patients with early-stage OSCC is a reliable test, with a similar diagnostic performance to END, with similar results in terms of oncological variables of survival and tumor recurrence. SLNB is a safe technique. The main complications associated with this procedure include functional alterations in the shoulder and neck, and in the arm abduction. Adverse events are very unlikely and, if they occur, of low severity. The evidence generated from randomized clinical trials has confirmed that SLNB associates less morbidity than surgery, although in the long term the differences between the functional alterations from both procedures disappear. Due to its diagnostic effectiveness and low morbidity, SLNB could be used as an alternative to cervical dissection in the clinical practice. SLNB could offer better quality of life results compared to END, but the evidence is no solid enough to drawn rigorous scientific conclusions. The evidence found related to the cost-effectiveness and cost-utility of SLNB in comparison to END is very scarce. However, according to some studies, it appears that using SLNB for staging patients with OSCC cT1/ T2N0 could show some benefit in economic terms.

Authors' methods: A systematic review of studies of diagnostic validity, safety and cost-effectiveness of SLNB has been carried out in comparison to the alternatives of cervical dissection and “wait and see”. To evaluate the diagnostic effectiveness, original articles were selected from studies evaluating the diagnostic validity and safety, published since 2016, including case-control studies, cohorts and RCTs, with a sample size greater than 25 patients with OSCC cT1/T2N0, to whom SLNB and/or END have been performed. Systematic reviews, meta-analyses, and HTA reports on this topic were also reviewed. To study the cost-effectiveness and aspects related to the quality of life of the patients, economic evaluation studies (analysis of costs, cost-effectiveness, cost-utility or cost-benefit), narrative reviews and qualitative studies, respectively, were considered. Only studies in English and Spanish were considered. Studies duplicated by subsequent ones, those that did not provide data to generate the 2x2 table or that did not offer other results of therapeutic effectiveness, case studies, conference abstracts, editorials or letters to the editor were excluded. An exhaustive search was carried out in the main sources of scientific information, using free and controlled language and different strategies appropriate to each of them. The selection of studies was carried out independently, between the two researchers. Disagreements arising in this process were resolved by consensus. After reading the title and abstract of the references retrieved in the search, those articles that seemed to meet the inclusion and exclusion criteria were retrieved and after reading the full text, the final selection of articles was made. The main characteristics of the selected studies, the study variables and their results were extracted. The risk of bias of the RCTs was evaluated using the Cochrane tool. Descriptive and narrative analyses of the evidence tables were carried out. Observational studies with a 2x2 table were considered for meta-analysis, using the MIDAS module of STATA v.16. Quantitative synthesis was not performed for RCTs since only two studies with this design were located.

Project Status: Completed

Year Published: 2021

URL for published report: https://redets.sanidad.gob.es/productos/buscarProductos.do?metodo=detalle&id=1036

Requestor: Spanish Ministry of Health

English language abstract: An English language summary is available

Publication Type: Not Assigned

Country: Spain

Organisation Name: Agencia de Evaluacion de Tecnologias Sanitarias

Contact Address: Instituto de Salud "Carlos III", Calle Sinesio Delgado 6, Pabellon 4, 28029 Madrid, Spain. Tel: +34 9 1 822 2005; Fax: +34 9 1 387 7841;

Contact Name: Luis M. Sánchez Gómez

Contact Email: luism.sanchez@isciii.es

Copyright: <p>Agencia de Evaluacion</p>