De standaardbehandeling van patiënten met T1-2N0-1 orofarynxcarcinomen is radiotherapie. Transorale robotische chirurgie (Transoral Robotic Surgery, TORS) wordt in aantal ziekenhuizen aangeboden als alternatief. In de Verenigde Staten is TORS geregistreerd voor deze indicatie door de Food and Drugs Administration (FDA). Gezien het feit dat deze aandoening bij een steeds jongere patientenpopulatie vastgesteld wordt door de toename van HPV+ tumoren en gelet de toxiciteit van de radiotherapie, zeker op lange termijn, wordt de behandeling middels TORS al dan niet in combinatie met (een eventueel lagere dosis) radiotherapie intensief onderzocht.

Conclusions

Transoral robotic surgery (TORS, including patients that received adjuvant (chemo)radiotherapy) versus primary radiotherapy (RT, including patients that reveived concurrent chemotherapy and salvage surgery)

Description of studies

Transoral robotic surgery (TORS) versus primary radiotherapy (RT) in Nichols (2019)

Nichols (2019) reported on a multicenter randomized controlled trial with centers recruiting in Canada and Australia to compare TORS with neck dissection (on indication) with primary radiotherapy (RT) as part of the ORATOR-trial. Data and procedures were extracted from the study protocol (Nichols, 2013) and the study report with appendices (Nichols, 2019).

Nichols (2019) included patients who were over 18 years old, had an ECOG-status of 0 to 2, had a histologically confirmed squamous cell carcinoma where the primary site was the oropharynx, had a T1-2 carcinoma and were likely to have negative resection margins at surgery, had N0-2 without extranodal extension (on pre-randomization imaging), the complete blood count/differential was obtained 4 week prior to randomization (with adequate bone marrow, hepatic, and renal function), and the patient was assessed at a multidisciplinary head and neck clinic and presented at a multidisciplinary tumor board prior to randomization. Patients were excluded when there were serious medical comorbidities, there were contraindications to therapy (radiotherapy, chemotherapy, surgery), there was prior history of head and neck cancer within 5 years, the patient had prior head or neck radiation, the patient had metastases, the patient was unable to attend full-course radiotherapy or follow-up visits, the patient had prior malignant disease (unless 5-year disease-free) with the exception of non-melanoma skin cancer, the patient was pregnant or lactating, or when the patient was unwilling or unable to complete the quality of life questionnaires.

A total of 68 patients were recruited and randomly allocated (1:1 allocation, block size 4). Patients in the TORS-group (n=34, 28 males) had a median age of 58.2 years (IQR: 52.6 to 64.5) and the primary tumor located at the tonsil / tonsillar fossa (n=24) or at the base of the tongue (n=10). Patients had an ECOG score of either 0 (n=30) or 1 (n=4), while 6 out of 22 (27%) who were asked drank over 21 drinks per week. Twenty-one (62%) patients undergoing TORS had a smoking history. Clinical T-stage was either T1 (n=17) or T2 (n=17), while clinical N-stage was N0 (n=9), N1 (n=7), or N2 (n=18). There were 30 HPV-positive patients in the TORS-group, as measured with p16 staining. Patients in the radiotherapy-group (n=34, 31 males) had a median age of 60 years (IQR: 53.2 to 65.2) and the primary tumor located at the tonsil / tonsillar fossa (n=26) or at the base of the tongue (n=8). Patients had an ECOG score of 0 (n=30) or 1 (n=4). Eighteen persons were asked about the alcohol-intake, resulting in 1 person (6%) having more than 21 drinks per week. Twenty-eight (82%) patients had a smoking history. In the RT group, the clinical T-stage was T1 (n=13) or T2 (n=21), while the clinical N-stage was N0 (n=12), N1 (n=5), or N2 (n=17). There were 30 HPV-positive patients by p16 staining in the RT-group as well.

In the TORS-group, a surgical robot was used to excise the primary oropharyngeal tumor with the spatula caultry (1cm resection margin). Selective neck dissection was performed at the discretion of the surgeon during surgery or within 2 weeks after surgery. Surgeons had to perform at least 10 TORS excisions prior to enrolling patients in the study. The surgical specimen was sent for frozen section analysis and the excision continued until negative margins were obtained. Adjuvant radiotherapy was given, however could be omitted when there was no extranodal extension, positive margins, pT3-4, nodal disease, or lymphovascular invasion. Thus, radiotherapy was given in the region of positive margins (64 Gy, 30 fractions, 6 weeks), in high risk nodal areas (60 Gy, 30 fractions, 6 weeks), or in low risk nodal areas (54 Gy, 30 fractions, 6 weeks). Concurrent chemotherapy with cisplatin (100mg/m² every 3 weeks in a 3 week cycle) could be administered when there were positive margins or extracapsular extensions. When patients were deemed unfit for cisplatin chemotherapy the doses and schedules could be modified, or cetuximab or weekly carboplatin (AUC 1.5) could be administered at the discretion of the medical oncologist. Out of the 34 patients in the TORS-group, 24 (71%) patients received radiotherapy and 8 (24%) patients received chemotherapy (cisplatin: n=5, carboplatin: n=3). Patients received chemotherapy in a median number of cycles of 6 (IQR: 4.5 to 6). It was unclear whether there were any deviations from the cisplatin schedule (100mg/m² every 3 weeks in a 3 week cycle) at the medical oncologist’s discretion.

In the RT-group, primary radiotherapy was given to the gross tumor and nodes (70 Gy, 35 fractions, 7 weeks). Radiotherapy was further administered in high risk nodal areas (63 Gy, 35 fractions, 7 weeks) and in low risk nodal areas (56 Gy, 35 fractions, 7 weeks). At the treating radiation oncologist’s discretion the radiotherapy regimen could be accelerated or hyperfractionated, where the same dose could be provided in 6 weeks. Concurrent chemotherapy was administered with positive nodes (i.e. N1-2) and omitted when N0. Chemotherapy doses and schedules could be modified for patients who were deemed unfit for cisplatin, or cetuximab or weekly carboplatin (AUC 1.5) could be administered at the discretion of the medical oncologist. Twenty-three (68%) patients received chemotherapy (cisplatin: n=19, carboplatin: n=3, cetuximab: n=1). Patients received chemotherapy in a median number of cycles of 3 (IQR: 2-6). It was unclear whether there were any deviations from the cisplatin schedule (100mg/m2 every 3 weeks in a 3 week cycle) at the medical oncologist’s discretion.Treatment response was evaluated by CT or PET-CT 8 to 12 weeks after completion of radiotherapy. When the CT showed residual nodes over 1 centimeter, patients were recommended salvage surgery. For PET-CT evaluation, residual nodes had to be over 1 centimeter and FDG-avid. Patients were offered surgical salvage when feasible for relapse or progressive disease. Four (12%) patients received salvage surgery, however it was unclear whether nodes were examined by a pathologist to confirm metastases.

Patients were followed-up every 3 months in the first two years and every 6 months in year 3 to 5. Physical examination and adverse event monitoring took place at every visit. Quality of life measurements and a chest x-ray was performed every 6 months, with the exception of the 1-year time point where a CT-scan was taken instead of an x-ray. The RT-group had an additional assessment at 8 to 12 weeks post-radiotherapy to assess the treatment response and the need for salvage surgery. The median follow-up in de study was 27 months (IQR: 20 to 48). Two patients in the RT-group withdrew consent after randomization (no reasons provided).

Nichols (2019) measured quality of life (QoL) with several validated questionnaires: MD Anderson Dysphagia Index (MDADI, score range: 20 to 100, higher indicates a better QoL), European Organization for Research and Treatment of Cancer Quality of Life of Cancer Patients general (EORTC QLQ-C30, score range: 0 to 100, higher score indicates a worse QoL), Organization for Research and Treatment of Cancer Quality of Life of Cancer Patients Head and Neck (EORTC H&N35, score range: 0 to 100, higher score indicates a worse QoL), Voice Handicap Index (VHI-10, score range: 0 to 40, higher score indicates a better QoL), Neck Dissection Impairment Index (NDII, score range: 0 to 100, higher score indicates a better QoL), Patient Neurotoxicity Questionnaire (PNQ, 2 items on a 5-point Likert scale), and Functional Oral Intake Scale (FOIS, 7-point Likert scale, higher scores indicate a better QoL).

Results

Transoral robotic surgery (TORS) versus primary radiotherapy (RT)

Overall survival

Nichols (2019) reported the 5-year overall survival of both groups, each containing 34 patients (reference group=TORS; log-rank p=0.89; HR=0.83, 95%CI: 0.21 to 8.35). Over the course of 5 years, there were 4 observed deaths in the RT-group (n=3 metastatic disease, n=1 cardiac event outside the hospital) and 5 observed deaths in the TORS group (n=3 metastatic disease, n=1 TORS-related bleeding, n=1 cardiac arrest after alcohol overdose and hypoglycaemia). A total of 27 persons in the TORS-group and 28 persons in the RT-group were censored over the course of 5 years. No reasons for censoring were provided.

Disease-free survival

Nichols (2019) defined the 5-year ‘progression-free survival’ as the time from randomization to either death or recurrence, whichever occurred first. A hazard ratio of 1.07 (95%CI: 0.28 to 4.01; reference group=TORS) and a log-rank p of 0.63 were reported. A total of 26 persons in the TORS-group and 28 persons in the RT-group were censored over the course of 5 years. No reasons for censoring were provided. There were 4 recurrences observed in each group.

Swallowing

Nichols (2019) reported swallowing complications per grade. The relative risk (RR) was not reported, however from these data the RR could be calculated. For grade 1 to 2 dysphagia an RR of 1.00 (95%CI: 0.69 to 1.45, reference group=RT) was reported. Both in the TORS and RT-group 21 grade 1-2 events were observed. Grade 3 dysphagia had 9 events in the TORS-group and 6 events in the RT-group. Here, an RR of 1.50 (95%CI: 0.60 to 3.75) was calculated. When all observed events were combined (TORS: 30 events, RT: 27 events) the RR was 1.11 (95%CI: 0.90 to 1.37). One person in the RT-group used a percutaneous feeding tube at the 1 year time point, compared to none in the TORS-group.

Local control

No studies were included that reported the local control.

Quality of life

Nichols (2019) assessed the 1-year quality of life with several questionnaires, however the quality of life as measured with the MDADI was the primary outcome measure. The MDADI total score differed significantly between groups at 1 year (TORS: 80.1 (SD: 13), RT: 86.9 (SD: 11.4), p=0.042), indicating the RT-group had a better swallowing-related quality of life. Similarly, a statistically significant difference was reported for the MDADI composite score (TORS: 80.2 (SD: 13.1), RT: 86.7 (SD: 11.4), p=0.049), indicating a better swallowing-related quality of life for the RT-group. The total score on the VHI-10 (TORS: 4.5 (SD: 4.3), RT: 4.4 (SD: 4.6), p=0.89) and NDII (TORS: 81.5 (SD: 28.7), RT: 92.3 (SD: 10), p=0.072) showed no significant difference. Summary scores for the EORTC QLQ-C30 and the EORTC H&N35 were not reported. For scores on sub-scales of the questionnaires we refer to Table 1.

Table 1 Overview of the 1-year quality of Life measurements per questionnaire as measured by Nichols (2019)

Trismus

Nichols (2019) reported grade 1-2 trismus in 8 of the 34 patients (23.5%) in the TORS-group and in 1 of the 34 patients (2.9%) in the RT-group. The RR was not reported, however from these data the RR could be calculated. Based on the data an RR of 8.00 (95%CI: 1.06 to 60.5) was calculated. One patient in the TORS-group had a grade 3 trismus complication, therefore the overall RR was 9.00 (95%CI: 1.21 to 67.21).

Taste

Grade 1 to 2 taste alterations were observed by Nichols (2019) in 17 patients (50%) in the TORS-group, compared to 19 patients (55.9%) in the RT group. The RR was not reported, however from these data the RR could be calculated. An RR of 0.89 (95%CI: 0.57 to 1.40) was calculated. Taste alterations in grade 3 or higher were not prevalent in both groups.

Dryness of the mouth

Grade 1 to 2 dryness of the mouth was reported in 18 of the 34 patients (52.9%) in the TORS-group, compared to 24 out of 34 patients (70.6%) in the RT-group. The RR was not reported, however from these data the RR could be calculated. The calculated RR was 0.75 (95%CI: 0.51 to 1.10). One patient in the RT-group had a grade 3 dryness of the mouth. The RR of all observed events was 0.72 (95%CI: 0.49 to 1.05, TORS: 18 events, RT: 25 events).

Mucositis

Nichols (2019) reported oral mucositis in 8 (23.5%) patients in the TORS-group, compared to 11 (32.4%) patients in the RT-group. The RR was not reported, however from these data the RR could be calculated. From the data, the calculated RR was 0.73 (95%CI: 0.33 to 1.58). Four patients in the RT-group had a grade 3 oral mucositis. The overall RR for oral mucositis was 0.53 (95%CI: 0.26 to 1.09, TORS: 8 events, RT: 15 events).

Level of evidence of the literature

Transoral robotic surgery (TORS) versus primary radiotherapy (RT)

The level of evidence regarding the outcome measure overall survival (crucial) was downgraded by 3 levels because of the number of included patients (imprecision: only 34 patients were included in each study arm. The number of censored participants was large in both arms, providing limited data to the 5-year overall survival); indirectness (1 level for indirectness: about half of the recruited patients had an N2 carcinoma); publication bias could not be assessed. We did not downgrade for a lack of blinding (risk of bias), because overall survival was considered as a ‘hard’ outcome measure.

The level of evidence regarding the outcome measure disease-free survival (crucial) was downgraded by 4 levels because of study limitations (1 level for risk of bias: care providers and outcome assessors were not blinded for the treatment allocation); and the number of included patients (2 levels for imprecision: only 34 patients were included in each study arm. The number of censored participants was large in both arms, providing limited data to the 5-year disease-free survival); indirectness (1 level for indirectness: about half of the recruited patients had an N2 carcinoma); publication bias could not be assessed.

The level of evidence regarding the outcome measure swallowing (crucial) was downgraded by 4 levels because of study limitations (1 level for risk of bias: patients, care providers and outcome assessors were not blinded for the treatment allocation); and the number of included patients (2 levels for imprecision: only 34 patients were included in each study arm); indirectness (1 level for indirectness: about half of the recruited patients had an N2 carcinoma); publication bias could not be assessed.

The level of evidence regarding the outcome measure local control (crucial) was not graded, since there were no studies included that reported this outcome measure.

The level of evidence regarding the outcome quality of life (crucial) was downgraded by 4 levels because of study limitations (1 level for risk of bias: patients, care providers and outcome assessors were not blinded for the treatment allocation); and the number of included patients (2 levels for imprecision: only 34 patients were included in each study arm. Power calculation (using the reported mean difference of 6.8 on de MDADI, SD=12.2, alpha=0.05, beta=0.1, +10% assumed dropout) revealed that 75 participants per arm were needed); indirectness (1 level for indirectness: about half of the recruited patients had an N2 carcinoma); publication bias could not be assessed.

The level of evidence regarding the outcome measure trismus (important) was downgraded by 4 levels because of study limitations (1 level for risk of bias: patients, care providers and outcome assessors were not blinded for the treatment allocation); and the number of included patients (2 levels for imprecision: only 34 patients were included in each study arm); indirectness (1 level for indirectness: about half of the recruited patients had an N2 carcinoma); publication bias could not be assessed.

The level of evidence regarding the outcome measure taste (important) was downgraded by 4 levels because of study limitations (1 level for risk of bias: patients, care providers and outcome assessors were not blinded for the treatment allocation); and the number of included patients (2 levels for imprecision: only 34 patients were included in each study arm); indirectness (1 level for indirectness: about half of the recruited patients had an N2 carcinoma); publication bias could not be assessed.

The level of evidence regarding the outcome measure dryness of the mouth (important) was downgraded by 4 levels because of study limitations (1 level for risk of bias: patients, care providers and outcome assessors were not blinded for the treatment allocation); and the number of included patients (2 levels for imprecision: only 34 patients were included in each study arm); indirectness (1 level for indirectness: about half of the recruited patients had an N2 carcinoma); publication bias could not be assessed.

The level of evidence regarding the outcome measure mucositis (important) was downgraded by 4 levels because of study limitations (1 level for risk of bias: patients, care providers and outcome assessors were not blinded for the treatment allocation); and the number of included patients (2 levels for imprecision: only 34 patients were included in each study arm); indirectness (1 level for indirectness: about half of the recruited patients had an N2 carcinoma); publication bias could not be assessed.

A systematic review of the literature was performed to answer the following question:

What are the (un)beneficial effects of TORS (with or without transoral microsurgery, and with or without adjuvant radiotherapy) on overall survival, disease-free survival, local control and quality of life, e.g. swallowing, trismus, taste, dryness of the mouth, and mucositis in patients with a T1-2N0-1 oropharyngeal carcinoma, when compared to primary radiotherapy.

P: patients with a T1-2N0-1 oropharyngeal carcinoma;

I: transoral robotic surgery with or without transoral microsurgery and with or without adjuvant radiotherapy;

C: primary radiotherapy;

O: overall survival, disease-free survival, local control, swallowing (complaints), trismus, taste, dryness of the mouth, mucositis, quality of life.

Relevant outcome measures

The guideline development group considered overall survival, disease-free survival, local control, swallowing, and quality of life (as measured with the MD Anderson Dysphagia Index (MDADI)) as a critical outcome measure for decision making; and trismus, dryness of the mouth, and mucositis as an important outcome measure for decision making.

A priori, the working group did not define the outcome measures listed above but used the definitions used in the studies.

The working group defined a minimal clinically relevant difference as (in line with “NVMO-commissie ter Beoordeling van Oncologische Middelen (BOM)”):

• > 5% difference or more >3% and HR <0.7 in overall survival.
• HR< 0.7 for progression free survival.

And, in case of absence of a clinically relevant difference in overall survival or progression free survival:

• A minimal clinically important difference of 10 points on the quality of life instrument EORTC QLQ-C30 (in line with Mehanna, 2019) or a difference of a similar magnitude on other quality of life instruments.
• Statistically significant less complications/adverse events.
• Statistically significant better functional outcomes, work participation.

Search and select (Methods)

The databases Medline (via OVID) and Embase (via Embase.com) were searched with relevant search terms until 6^th of January 2019 for systematic reviews and randomized controlled trials. The detailed search strategy is depicted under the tab Methods. The systematic literature search resulted in 106 hits. Studies or systematic reviews were selected based on the following criteria: patients had a T1-2N0-1 oropharyngeal carcinoma, transoral robotic surgery was compared with primary radiotherapy, at least one of the outcomes of interest was reported (overall survival, disease-free survival, local control, swallowing (complaints), trismus, dryness of the mouth, mucositis, quality of life), the study design was a randomized controlled trial or the systematic review contained randomized controlled trials, reports had to be written in English or Dutch. Twenty-one studies were initially selected based on title and abstract screening. After reading the full text, 20 studies were excluded (see the table with reasons for exclusion under the tab Evidence tables) and 1 randomized controlled trial was included.

Results

One randomized controlled trial (Nichols, 2019) was included in the analysis of the literature. Important study characteristics and results are summarized in the evidence tables. The assessment of the risk of bias is summarized in the risk of bias tables (under the tab Evidence tables).