Transoral robotic surgery vs open surgery in head and neck cancer. A systematic review of the literature

Background TORS has become one of the latest surgical alternatives in the treatment of oropharynx squamous cell carcinomas (OPSCC) and has become increasingly accepted by surgeons as a treatment option. Surgical robots were designed for various purposes, such as allowing remote telesurgery, and eliminating human factors like trembling. The study aimed to compare systematic review of the available literature in order to evaluate the safety and efficacy of Transoral Robotic Surgery (TORS) compared with open surgery. Material and Methods We performed a systematic review of the available literature in order to evaluate the safety and effectiveness of TORS compared with open surgery. We compared TORS and open surgery based on 16 outcomes divided in to 3 groups: intra-operative complications, post-operative complications, and functional and oncologic outcomes. An electronic search of observational studies was carried out using the following databases: MEDLINE, EMBASE, Cochrane Central Register of Controlled Trials, Cochrane Oral Health Group Trials Register, and Scielo. Data analysis was carried out in accordance to Preferred Reporting Items for Systematic Reviews and Metanalysis (PRISMA) and the quality of the studies were evaluated using the Newcastle-Ottawa Scale. No language restrictions were imposed. Results From the 4 studies identified (Newcastle-Ottawa Scale mean score 6.5), 371 patients were revised (186 patients were treated with TORS and 185 with conventional surgery). Overall, TORS, when compared with open surgery, appears to have better functional results (less hospital time, decannulation) and fewer intraoperative and post-operative complications. There is no significant difference when assessing the oncological outcomes (positive margins, survival rate) when comparing both techniques. Conclusions TORS has an overall better functional outcome, and less intraoperative and postoperative complications with no difference in positive margins and survival rate when compared with conventional therapy. Key words:Transoral Robotic Surgery, TORS, open surgery, conventional surgery, head and neck cancer, oral cancer.


Introduction
Over the last few years, the incidence of oropharynx squamous cell carcinomas (OPSCC) associated with the Human Papilloma Virus (HPV), is significantly increasing (1)(2)(3). It is expected that if this rate of OPSCC growth continues, then by 2030 there will be at least twice as many annual cases in the U.S.A, compared to present (4). Open surgery still remains the most preferable treatment option in terms of mobility and cosmetic impact (4). Adjuvant radiotherapy (RT) with lymphnode surgery, has been used for more than three decades, but it is associated with multiple complications and reduction in patients' quality of life (QoL) (5,6). Chemoradiotherapy (CRT) is no longer the first-line treatment of choice, due to the demographic changes that have occurred in patients with head and neck squamous cell carcinomas (HNSCC). The type of patient currently diagnosed ranges widely, from the elderly patient with a long history of tobacco consumption and alcohol abuse, to the younger and healthier patient with HPV-associated OPSCC, having no history of substance abuse and thus a better prognosis (1,7). Although patients treated with CRT avoid surgery-associated risks (namely anaesthetic complications, bleeding, postoperative pain, fistula formation, malocclusion, facial deformities and iatrogenic lesions in neurovascular structures), they are not entirely free from complications. For example, late complications such as xerostomia, osteoradionecrosis, mucositis, nephrotoxicity, neurotoxicity, ototoxicity, sepsis, pharyngeal and oesophageal stenosis, fibrosis in muscles involved in speech and swallowing, and the development of a second malignancy have all been reported (2,4,8,9). Therefore, concerns about the toxicity of the treatment and late complications have become even more relevant, since these patients, given their age, potentially have a higher risk of developing late adverse effects (2). Renewed interest in the search for a surgical approach for the treatment of OPSCC led to the introduction of Transoral Laser Microsurgery (TLM). In 2003, Steiner et al. (10) were the pioneers in publishing a series regarding this subject. TLM is a technically-challenging procedure, requiring specific training, a great deal of -Screening process The combination of control terms (MeSH and EM-TREE) and key words were used wherever possible. In addition, a manual search was carried out of the journals related to head and neck oncology to ensure a thorough selection process. We applied no language restrictions. The references in the excluded articles were also checked by seeking studies that met the inclusion criteria. Articles published in the last 6 years, up to December 2019 were reviewed. -Eligibility Criteria The full papers and abstracts identified through the search engines were independently reviewed by the authors (AR, RA & JLL) for inclusion in this systematic review. If there was a disagreement between the reviewers, the authors reviewed the full text before reaching an agreement through discussion. Data extraction was then completed in duplicate by the same independent reviewers. The search strategy and the flow diagram of the article selection are shown in Fig. 1.
The articles that were included in this systematic review met the following inclusion criteria: retrospective or prospective clinical observational studies, with 50 or more human subjects and the clinical results of TORS in head and neck cancer were reported. Consequently, several factors such as study design, number of patients included in the latest follow-up assessment, test/control groups, tumour size and TNM staging were studied. Furthermore, in order to address the aim of this study more comprehensibly, other technique-related parameters were also extracted, including operating time, blood loss, postoperative complications (e.g. Infection, bleeding, and respiratory tract oedema), functional outcomes (e.g. duration of in-patient stay, and swallowing function) and oncological outcomes (average survival and disease-free average). By contrast, case reports or case series, studies comparing TORS with Chemoradiotherapy (CRT), reviews, author debates, letters to the editor, pre-clinical studies, clinical studies involving animals or with deficient information were excluded.
-Quality of studies The Newcastle-Ottawa Scale (NOS) was used to assess the quality of non-randomised studies (Case Controls & Cohorts Studies) by 5 reviewers (AR, XRL, AMR, EJS, JLL). Using a scoring system with stars, each study was judged on three broad aspects: the selection of study groups, the homogeneity of the compared groups and the methodology of the results obtained (Table 1). If there was a disagreement between the reviewers, the authors would reach a consensus and decide the overall score.
-Results of the systematic review The total number of patients studied was 371 (305 men and 66 women). 186 patients were treated with TORS and 185 with conventional surgery. With regard to the tumour staging, there were 118 cases of T1, 194 at T2, 41 at T3 and 18 at T4. Regarding TNM staging, there were a total of 50 cases of TNM I, 60 cases of TNM II, 50 cases of TNM III and 211 cases of TNM IV ( Table 2). Two of the articles (6,14) referred to the amount of blood lost during surgery, with a finding of a difference of over 200ml between the control group (open surgery) and the study group (TORS). 3 of the 4 articles (6,9,14) also discussed the operating times and a notable difference can be seen between the shortest times in the study group than in the control group. However, it should be mentioned that in the article by Lee et al. (9) there is a subgroup in the test group which had much longer intraoperative times due to the need to perform mandibulectomy (Table 3). With regards to postoperative fistulas, only two articles discussed these. The two studies had a very similar percentage of fistulas: 6% (6) and 7.69% (14). However, in neither of the two studies do the fistulas occur in the TORS group, but in the open surgery group ( Table 3). As regards the functional results, it can be seen that TORS obtained significant results compared with open surgery ( Table 4). The hospitalisation time varies a lot between the various studies, but among them all, the TORS group reveals a better result; 3.8 days (6), 26.1 days (14) and 14.6 days (9) against 8.0 days (6), 43.4 days (14) and 24.6 days (9), respectively. Tracheotomy cannot be assessed properly since not all study groups followed the same criteria -some studies performed it in advance (14) and others did not (6,9). However, in decannulation we observed that the TORS group obtained much better results. Park et al. (14) took an average of 7.2 days to decannulate TORS patients and 15 days in open-surgery patients. Likewise, Lee et al. (9) took an average of 5 days in the TORS group and 13.2 in the control group. We can also observe that patients   Author (      With regards to the oncological results, 3 (3,9,14) out of the 4 articles (3,6,9,14) do not show any significant results in terms of disease-free and survival time, and the differences between the test and control groups were    (Table 4).
-Quality of Studies The Newcastle-Ottawa Scale (NOS) was used to analyse the quality of observational studies and scores of between 6 and 7 were obtained (Table 1).

Discussion
The nature of treating head and neck cancer raises several ethical reasons when conducting randomised clinical trials. Consequently, evidence for TORS and open surgery to treat head and neck cancer is limited. Assessing the quality of observational studies allows us to address some of previous limitations. Although, we acknowledge the limitations commonly seen at cohort and case control studies selected for this study.
There are several methods to address the quality of evidence for observational studies such as ROBINS, Newcastle-Ottawa Scale (NOS), "Downs and Black" and "Scottish Intercollegiate Guidelines Network's methodology checklist". Although ROBINS has a has been highlighted, for its good validity (20), when assessing surgical records NOS is easy to apply (21), and has been widely used due to recommendations from the Cochrane Collaboration (22). Newcastle-Ottawa Scale (NOS) has been used in other systematic reviews to assess TORS2 but has not yet been compared with open surgery. In order to provide similar methodology, as well as reveal the advantages previously mentioned, we agreed to use NOS. However, we do acknowledge the limitations of NOS, when it is compared with ROBINS. Treating HNSCC patients with surgery, instead of CRT, means they avoid the typical complications of RT and CT; xerostomia, osteoradionecrosis, mucositis, sepsis and trismus (2,4,9,23). A recent study (24) reported an overall survival rate of 84.9% and a low rate of xerostomia in 442 patients who received intense modulated radiotherapy. However, Forastiere et al. (25) reported that toxicity increased by combining chemotherapy and radiotherapy, causing chronic lesions and severe fibrosis in the pharyngeal mucosa and masticatory mucosa. Whilst treating patients with surgery prevents them from experiencing complications of CRT, they are not free from complications. Some previous studies indicate that almost 50% of patients who undergo open surgery for cancers of the head and neck develop postoperative complications or problems with the healing of wounds (26). Surgical complications of TORS include: formation of fistulas, postoperative haemorrhages, as well as surgical infections, pneumonia, nerve damage and lesions of the teeth (27). Two studies conducted at the Mayo Clinic (Rochester, MN, USA.) claim that their series of case studies resulted in 6% formation of fistulas (28,29). These patients underwent lymphadenectomy at the same time as TORS surgery. Fistulas were resolved with antibiotic treatment. Chia et al. (30) in a retrospective survey only had 0.2% of cases with fistulas in patients who received TORS, which is very close to the results of our review. As mentioned previously, only two studies (6,14) provided data regarding fistulas, but all instances occurred in cases of open surgery. None of the cases treated with TORS developed fistulas. Blood loss, prior chemotherapy and type of surgery are risk factors for postoperative infections and fistulas (31).
Postoperative haemorrhage is a potentially life-threatening complication in patients treated with TORS. In the same study by Chia et al. (30) it is reported that, although the mortality rate after TORS was low (0.3%), haemorrhage was the causative factor in of all cases. Doazan et al. (32) reported 2 deaths because of postoperative hemorrhage (1.6%). This rate is higher than what was reported for the series of Mandal et al. (33) with 0.9% of severe hemorrhage that led to death. A total of 3.1% of postoperative haemorrhages were reported which closely approximates the analysed study by Park et al. (14), which has a haemorrhage rate of 3.3%, which are values that approximate a very recent published review, with a rate of 2.4% (34). However, some articles record values close to 6% (35) or even above 10%, such as the analysed study by White et al.,with 11% or the multi-centre study by Aubry et al. (36) who reported18.5% of postoperative haemorrhages. As expected, Asher et al. (35) found that larger tumours have a greater bleeding risk than smaller tumours. Anticoagulant therapy and/or anti-platelet therapy was the only other risk factor identified in postoperative haemorrhages (35,36). Based on these results, additional data and larger samples are needed to obtain more accurate conclusions with regards to postoperative haemorrhages. However, given the results of this review, intraoperative bleeding is much less when using TORS than open surgery; but in postoperative bleeding there is not much difference between the two types of surgery and there is even a risk of death due to postoperative bleeding (30). While the surgical robot is an excellent tool to improve visualisation and instrumentation in the upper digestive tract, unfortunately, TORS can be less ideal for controlling haemorrhages after transoral robotic surgery. This reflection led Asher et al. (35) to perform preventive ligatures of the lingual artery. Improvements are being attempted in this respect, such as: the use of intraoperative ultrasound images in real time to increase the 3D visualisation of the robotic system. Ultrasound imaging may allow the identification of the tumour boundaries and neurovascular structures, thus improving tumour resection and reducing morbidity (36). TORS provides the ability to carry out, under a 3D-HD view enlarged on multiple planes, en-bloc tumour resections, allowing the margins to be assessed more accurately during resections. The greater degree of instrumentation freedom, means that the safety margins of resections have results very similar to those of conventional open surgery, but with a much lower cost in terms of surgical morbidity (37). Our systematic review confirms this, since out of the 4 articles analysed, 3 showed no significant results regarding safety margins.
Only the study by White et al. (6) reported significant results, resulting in 9% positive margins in the TORS group and 29% in the control group. The results on the safety margins are likely to have a direct proportional relationship to the overall survival rates of the disease and disease-free patients, since only the same study by White et al. (6) presents statistically significant results in these two sections. However, we must also note that the great difference between the results of White et al. (6) and those of the other 3 studies (3,9,14) on these two indices. The other 3 studies (3,9,14) (40) concluded that TORS cost are 24-35% less than IMRT, but that one of the reasons why TORS costs could increase is the adjuvant therapy as de Almeida et al. (38) explains. Spellman et al. (40) concluded that TORS is highly cost-effectiveness is early stage oropharyngeal cancer.
One of the limitations of the analysed studies is that they are rather heterogeneous with one another:

Conslusions
In conclusion, TORS appears to be a promising technique with fewer complications compared with open surgery. Its effectiveness depends on the location of the tumour, for which more specific studies of each type of cancer are needed. The skill of the surgical team also has an effect, since the training curve is relevant. TORS obtains much better functional results when compared with Open Surgery therapeutic techniques for treating patients with HNSCC. Although the oncological outcomes of TORS are good, they do not appear better than those of open surgery. We suggest that more high quality observational studies, to understand the effectiveness of TORS when compared with Open Surgery, are needed. The studies will need to use good statistical power and standardised outcome measures to obtain actuate results.