Transoral Robotic Surgery Using the Thulium:YAG Laser:  A Prospective Study FREE

Advancements in surgery, radiotherapy, and chemotherapy during the past 4 decades have made the management of oropharyngeal squamous cell carcinoma (OPSCC) and supraglottic SCC (SGSCC) dynamic and controversial. Regardless of the approach, the dual goals of oncologic management are (1) local, regional, and distant control of disease and (2) preservation of speech and swallow function. Collectively, surgical and nonsurgical treatment strategies have largely demonstrated equivalent survival.^1- 2 Therefore, the management paradigm has focused on improving the quality of life and overall function.¹ Historically, transcervical approaches to the oropharynx and supraglottis resulted in long operative times and poor functional outcomes.³ Although the introduction of transoral laser microsurgery in 1972 offered an endoscopic and minimally invasive alternative to the open approach,⁴ some users found that it was limited by carbon dioxide (CO₂) laser line-of-sight requirements, restrictive and nonintuitive movements, and a limited field of view.³ Recently, transoral robotic surgery (TORS) has overcome these limitations by offering surgeons nonlinear exposure, wristed instrumentation, a 3-dimensional field of view, and tremor reduction.^2- 3,5

Early studies^1- 3,5- 6 using TORS uniformly described the use of monopolar electrocautery (EC) to perform surgical resection. It is well documented that EC results in greater thermal injury and tissue necrosis caused by disseminated electrical injury than does appropriately used CO₂ laser technology.^1,6- 7 Since its introduction, the CO₂ laser has largely replaced EC in endolaryngeal surgery owing to the cutting and hemostatic properties of infrared laser energy.⁸ However, until recently, CO₂ laser technology has been limited to line-of-site applications owing to inflexible fiber technology and was, therefore, not useful for TORS. This prompted investigation into a new laser design: the thulium-doped YAG (Tm:YAG) laser (RevoLix; Lisa Laser Products).⁹ The Tm:YAG laser demonstrates several qualities particularly advantageous for its use in TORS, including a flexible cable, precise cutting ability, and excellent hemostasis. However, the feasibility and safety profile for Tm:YAG laser–assisted TORS is unknown. The primary aim of this study was to assess the feasibility, safety, and functional outcomes associated with use of the Tm:YAG laser during TORS (TY:TORS) by direct comparison with the current gold standard, TORS using EC (EC:TORS), via a prospective matched cohort study.

All the patients were identified through a prospectively maintained database, and data were supplemented by retrospective medical record review. Mayo Clinic Institutional Review Board 10-000192 approval was obtained before study commencement.

Between December 1, 2009, and July 31, 2010, we trialed TY:TORS at a single tertiary academic referral center. Fifteen patients who presented with OPSCC (base of tongue or tonsil) or SGSCC during this time and met the criteria for TORS³ underwent TY:TORS resection of their respective lesion. In addition, patients undergoing EC:TORS for primary surgical resection of OPSCC or SGSCC between March 1, 2007, and August 31, 2010, were used as a participant pool from which matched control subjects were selected. To decrease the risk of confounding variables,^10- 11 patients were matched using the following criteria: (1) tumor location (exact match: base of tongue, tonsil, or supraglottis), (2) clinical T stage according to the American Joint Commission on Cancer (exact match: unknown, 1, 2, 3, or 4), (3) sex (exact match), and (4) age (±20 years). To decrease population bias, each matched control was chosen by searching the prospectively collected TORS database using only the criteria described previously herein to identify the best match for each patient. The reviewing author (K.M.V.A.) was masked to patient clinical course during the selection process. Thirty individuals were selected to match each patient (n = 15) with 2 controls (n = 30).

Data were collected from the intraoperative, perioperative, and postoperative periods. Intraoperative data included setup time (general anesthesia to incision), total surgical time (incision to closure), estimated blood loss (minimal, ≤150 mL; moderate, 151-300 mL; or high, >300 mL), hemostasis (placement of clips transorally, spot monocautery), primary defect reconstruction, technical complications, pharyngotomy, placement of a nasogastric tube (NGT), and tracheostomy. The time between exit from the operating room and hospital discharge was defined as the perioperative period. Data were collected on level of care required (intensive care unit vs general floor care); pain level, defined by analgesia requirements (level 1, enteral; level 2, intravenous push; or level 3, patient-controlled analgesia); complications before discharge; discharge disposition; and home health requirements. We routinely admit postoperative patients to the general floor. Admission to the intensive care unit is considered for continuous cardiopulmonary monitoring, intravenous blood pressure medications, airway monitoring, and intensive nursing needs. Level of analgesia was defined solely by review of the medication administration records. Our standard of practice is to progress from level 1 to level 3 on the basis of the patient's reported pain control. Postoperative data were collected from hospital discharge to postoperative day (POD) 30. Emergency department visits, readmissions, reoperations, adverse outcomes, and postoperative bleeding were reviewed. Short-term swallow outcomes were determined by evaluating time to NGT removal, time to oral diet, and placement of an NGT or percutaneous endoscopic gastrostomy (PEG) tube between discharge and POD 30. Requirement for a formal dysphagia evaluation was noted, and the dysphagia diagnosis was reported.

The operative setup and surgical technique were identical to those previously reported,³ with the following exceptions: all TY:TORS procedures were conducted using a laser-safe endotracheal tube and laser eye protection for patient and staff. A 5-mm Schertel grasper (EndoWrist; Intuitive Surgical, Inc) was attached to the left robotic arm, and the laser fiber (RevoLix) and introducer (Intuitive Surgical, Inc) were attached to the right robotic arm. Laser resection was performed using a 70-W 2-μm continuous-wave Tm:YAG laser (RevoLix). During all EC:TORS procedures, the 5-mm Schertel grasper was attached to the left robotic arm and the 5-mm spatula cautery was attached to the right. In all cases, neck dissections (NDs) were performed on the same day as TORS.¹² In patients undergoing ND, the wound was inspected for communication, and if present, the pharyngotomy was managed according to the algorithm previously reported.¹²

Intraoperative frozen pathologic specimens were reviewed as previously described.³ The number of attempts per histologically negative margin was reported. A head and neck pathologist then used formalin-fixed paraffin-embedded tissue to evaluate the depth of thermal damage and the impact of thermal char on margin analysis. The extent of thermal damage was measured using an Olympus BX51 microscope rotatable measuring analyzer (Olympus America); after surveying all surface aspects of tissue fragments, the average depth of thermal damage was generated using 3 separate areas of tissue with thermal artifact. Variation in the impact of char on margin analysis was considered and interpreted as follows: if the margin demonstrated thermal artifact that was interpreted as questionable or positive for tumor, the impact of char was considered positive, and if the margin had no thermal artifact or demonstrated thermal artifact that did not affect interpretation of the margin, the impact of char was considered negative.

Comparisons were performed using the independent-samples t test, the χ² test, and the Fisher exact test using commercially available software (JMP; SAS Institute, Inc). P ≤ .05 was considered statistically significant.

Fifteen patients undergoing TY:TORS and 30 receiving EC:TORS met the inclusion criteria and were enrolled. No significant differences were noted in patient characteristics (P >> .05) (Table 1 and Table 2). One matched control (patient 32 in group 2) was 23 years older than the cohort patient but was included based on otherwise ideal site, size, and sex matching. Of the 45 patients included in this study, only 1 presented with recurrent disease. This patient had a history of chemotherapy and radiotherapy 2 years before presentation (Tables 1 and 2). No other patient underwent nonsurgical treatment before surgical intervention.

All the patients successfully underwent TORS without the need for conversion to an open approach. The Feyh-Kastenbauer laryngeal retractor was most commonly used; however, 1 patient in the TY:TORS group (Crowe-Davis) and 2 in the EC:TORS group (Crow-Davis and Kleinsasser) required an alternative retractor to improve exposure. No significant differences were noted in surgical setup time, total operative time, NDs performed, estimated blood loss, clips used transorally, or additional cautery required (P >> .05) (Table 3). We could not separate time for TORS from total operative time based on retrospective review of intraoperative data, and, therefore, we compared total operative time and setup time across participants.

No intraoperative instrument malfunctions were reported. One EC:TORS control was noted to have a thermal burn on the anterior neck skin. This area of skin was adjacent to the cervical ND incision and was satisfactorily excised. An NGT was placed intraoperatively in 53% of patients (8 of 15) and 73% of controls (22 of 30) (P = .20). Four patients in the control group underwent intraoperative tracheostomy (Table 4). Of these, 3 had SGSCC (T2N2c, T2N1, and T2N2a) and 1 had tonsillar SCC (T2N1). No significant intraoperative complications, including death or hemorrhage, occurred.

Of the 45 patients, 1 had a tumor that required piecemeal resection. This patient underwent TY:TORS for a T4N2b tonsillar SCC, which was successfully removed in 2 specimens. Two patients in the TY:TORS group required reconstruction of the primary tumor bed (Table 4). Both had tonsillar SCC requiring either primary closure (T4N1) or sternocleidomastoid muscle flap reconstruction (T2N2a).

The number of intraoperative pharyngotomies (IPs) was analyzed in patients undergoing concomitant ND. Two patients in the TY:TORS group (T2N0 SGSCC and T2N0 tonsillar SCC) did not undergo ND, and, therefore, both they and their control counterparts were excluded from this analysis (TY:TORS: n = 13; EC:TORS: n = 26). Of the patients included, 8% (1 or 15) in the TY:TORS group and 42% (11 of 30) in the EC:TORS group were reported to have an IP, which reached significance (P = .03) (Table 4).

There was no difference between mean number of attempts per negative frozen margin between groups (2.2 for TY:TORS and 2.13 for EC:TORS, P >> .49). Negative margins were ultimately achieved in all participants. Thirteen TY:TORS and 18 EC:TORS formalin-fixed paraffin-embedded specimens were available for review. The mean depth of thermal damage was 0.47 mm for EC:TORS vs 0.21 mm for TY:TORS (P = .06) (Figure 1 and Figure 2). There was no difference between groups regarding thermal artifact effect on margin interpretation (P = .78) (Table 5).

One TY:TORS patient and 2 EC:TORS controls (P >> .05) required admission to the intensive care unit for hypertension due to secretion of inappropriate antidiuretic hormone, hypertension and apnea, and desaturation in the postanesthesia care unit, respectively; all 3 individuals were transferred to general floor care within 24 hours and were discharged from the hospital between PODs 2 and 4. The level of analgesia required by patients in the EC:TORS group was significantly higher than that required by the TY:TORS group (P = .02) (Table 6). In the TY:TORS group, 73% (11 of 15) were tolerating an oral diet before discharge compared with 31% (14 of 30) in the EC:TORS group (P = .12) (Table 6).

No significant differences in complication rates before discharge were observed (P = .26) (Table 6). Complications related to TY:TORS included aspiration pneumonia on POD 3 and difficulty swallowing secretions. Two complications related to EC:TORS included NGT placement for difficulty handling secretions and pharyngotomy recognized in the postanesthesia care unit and successfully managed with immediate NGT placement. One control patient developed a chyle leak after ND. No salivary fistulae or life-threatening complications occurred.

No patient required either intermediate or nursing home care. Home health care was arranged for 3 patients in the control group (10%; P = .54). The mean (SD) hospital stay was 3 (0.8) days for the TY:TORS group and 3.47 (1.5) days for the EC:TORS group (P = .18).

One TY:TORS patient presented to the emergency department with pain. Six control patients were evaluated in the emergency department: 1 for altered mental status and 5 for postoperative bleeding (mean [SD] PODs: 8.6 [4.2]). These controls were taken to the operating room for hemostasis and were admitted for observation (P = .28). No significant difference was noted for emergency department evaluation, readmission, reoperation, or postoperative bleeding (P >> .05).

The mean (SD) time to NGT removal was 13.25 (11.8) days for TY:TORS vs 24.59 (43) days for EC:TORS (P = .27). In the TY:TORS group, 1 patient (7%) required PEG tube placement after initially tolerating an oral diet. Similarly, in the control group, 2 patients (7%) required placement of an NGT in the postoperative period and 4 required a PEG tube (13%). Three of the 4 patients who underwent intraoperative tracheotomy were successfully decannulated on PODs 2, 3, and 21. The fourth patient (T2N2c SGSCC) died 4 months postoperatively after chemoradiotherapy; this patient never achieved decannulation, and his cause of death is unknown as his family declined autopsy.

There was no difference in overall postoperative complications (P = .91) (Table 7). Complications related to TY:TORS included aspiration, reflux/hoarseness, velopharyngeal insufficiency, and muscle spasm/jaw pain. Complications in the TY:TORS group related to ND included lymphedema, marginal mandibular neuropathy, accessory neuropathy, temporary Horner syndrome, and muscle spasm/jaw pain. Complications related to EC:TORS included 5 postoperative oropharyngeal bleeds, severe dysphagia requiring placement of a PEG tube, velopharyngeal insufficiency, and supraglottic stenosis. Complications related to ND in the control group included accessory neuropathy, marginal mandibular nerve weakness, right-sided vocal cord paresis, wound infection, and lymphedema.

Functional swallow was evaluated on or before POD 30 (Table 8). Nine TY:TORS patients (60%) and 20 EC:TORS patients (67%) underwent formal swallow evaluation by a speech pathologist (P = .75). No significant difference was noted between dysphagia diagnoses (P = . 43). Patients swallowing in the immediate postoperative period, not formally evaluated, were considered normal (Table 6).

The da Vinci S surgical system (Intuitive Surgical Inc) averages $1.5 million per robot. The Tm:YAG laser, recently purchased by the Mayo Clinic, cost $62 500. The institutional cost of each da Vinci cutting accessory is given in Table 9.

The Tm:YAG laser is a diode-pumped solid-state laser with a YAG rod in which thulium ions are directly excited by high-power laser diodes.^9,14 Compared with its predecessor, the holmium laser, the diodes in the thulium laser cause narrowband excitation, which limits excessive heat generation inside the crystal, increases power efficiency, and eliminates the need for special electrical installation.¹⁴ The Tm:YAG laser produces a continuous-wave beam with a wavelength of 2013 nm (2 μm),^9,14 allowing for smooth incision and vaporization of tissue.¹⁴ In addition, the shorter wavelength and the continuous-wave output results in a depth of penetration of only 0.25 mm, decreasing the risk of serious tissue damage.^14- 15 This wavelength allows for the use of clear safety glasses, which do not produce color distortion for the surgeon or assistant.¹⁵

Similar to the CO₂ laser, the target chromophore of the Tm:YAG laser is water. Due to its inherent temperature stability, water retains its absorption properties when heated by the laser beam to its boiling point, which coincides with tissue vaporization.^14- 15 According to Teichmann et al,¹⁴ tissue remaining after each pass of the laser is, therefore, covered by a coagulated seam with ample water for efficient absorption during successive laser passes. This not only results in excellent hemostasis but also provides the surgeon with a clear surgical view while decreasing the risk of serious collateral soft-tissue damage.^14- 15 Finally, laser energy can be delivered through a flexible glass fiber (0.365-0.550 mm), which allows for application to urologic surgery, office-based endolaryngeal surgery, and, now, TORS.^9,14- 15

In 2006, Zeitels et al⁹ first reported use of the Tm:YAG laser in endolaryngeal surgery. They found that it resulted in a 50% greater thermal damage zone than that produced by the CO₂ laser but a lesser thermal damage zone than EC.¹⁶ They noted that additional EC was unnecessary with the Tm:YAG laser during either ablation of bulky exophytic disease or extensive resection in the deep compartments of the larynx.⁹ They concluded that the hemostatic properties of the Tm:YAG laser were superior to those of the CO₂ laser.⁹ In a retrospective review of 443 laryngotracheal cases treated by unsedated office-based laser surgery, Koufman et al¹⁷ compared pulsed-dye, CO₂, and Tm:YAG lasers and concluded that the Tm:YAG laser is a useful choice for ablative procedures requiring excellent hemostasis.

These early basic science and clinical studies suggest that the Tm:YAG laser may be superior to other cutting and ablating lasers regarding hemostasis while simultaneously avoiding the extensive tissue necrosis inherent to EC. We compared the feasibility, safety, and 30-day functional outcome of TY:TORS for surgical management of OPSCC and SGSCC and conducted a prospective matched cohort study comparing this cutting and ablating tool with EC, the current standard of care for TORS.

Feasibility indicators of particular interest were tumor exposure, hemostasis, operative time, and ability to achieve negative margins. No patient had the surgical plan altered or aborted owing to limited exposure, and all the tumors were removed en bloc except 1 (a large T4 tonsillar SCC). This success reinforces the importance of careful preoperative tumor selection and operator experience.³ Estimated blood loss was reported as minimal (<150 mL) for 87% of TY:TORS patients (13 of 15) and 63% of EC:TORS controls (19 of 30), which is consistent with other studies.^3,5,18- 19 Theoretically, the deeper and wider thermal damage zone associated with EC might provide superior hemostasis to Tm:YAG; however, in this study, vascular control was equivalent between groups. In addition, although setup time and operative time are difficult factors to use as measures of surgical efficacy, they are important to consider when introducing a new technique into the current medical environment.³ The Mayo Clinic's initial experience with TORS for oropharyngeal lesions was presented in 2009 and included patients from March 2007 through January 2008.³ That study found that although setup time decreased significantly in the first 10 cases, operative times for primary tumor resection remained fairly constant among the 45 patients.³ Time was affected primarily by tumor stage (increased time for higher stages).³ When reviewing the present cases, only 3 control patients were included who underwent surgery in 2007. This initial experience, coupled with the finding that there was no difference between setup time and overall operative time, suggests that there was no significant temporal bias between study arms based on learning curve. These results further suggest that the transition to TY:TORS from EC:TORS was relatively smooth for the surgeon and ancillary staff.

Finally, with the ultimate goal of achieving a tumor-free margin while minimizing damage to normal tissue, the ability to interpret surgical margins accurately without requiring additional tissue due to thermal artifact is an important concept. In this study, the choice of cutting tool did not affect the number of attempts required to achieve a negative margin, and there was no difference in the number of new margins requested by the pathologist due to thermal artifact. However, there was a trend toward deeper mean thermal damage zones in specimens removed via EC:TORS (0.47 mm vs 0.21 mm, P = .06), which is consistent with the literature. This analysis was limited by specimen availability and the subjectivity inherent to retrospectively reviewing surgical margins and warrants further investigation.

Safety was evaluated based on clinical course and complications. Intraoperative complications were relatively rare, consistent with the TORS literature.⁵ The incidence of IP in patients undergoing concomitant ND was significantly higher in the EC:TORS group (42%; 8 tonsil and 3 base of tongue) compared with the TY:TORS group (8%; 1 tonsil) (P = .03). The incidence in the control group was similar to the 29% to 40% previously described in patients undergoing EC:TORS for OPSCC.^3,12 No patient developed a postoperative orocutaneous fistula. In a recent study,¹² the risk of fistula after IP was found to be only 1.4%, and all the patients were successfully managed with conservative therapy with no delay in adjuvant therapy. Despite the relatively low morbidity associated with IP and its sequelae, attempts to further decrease its occurrence are warranted. We hypothesize that decreased collateral thermal damage and improved tissue visualization associated with the use of TY:TORS contributed to this finding. In addition, the relatively small spot size of the laser in TY:TORS may allow for more precise cutting and preservation of the buccopharyngeal fascia, consequently limiting the number of pharyngotomies occurring during tumor resection.

Postoperative complications were considered minor and were consistent between study arms and with the TORS literature.^3,5 One patient in the TY:TORS group developed aspiration pneumonia, which did not delay hospital discharge or predict poor outcome. Although hospital length of stay has been largely institution and surgeon dependent,⁵ the present data are similar to those previously reported in the literature. Mean (SD) length of hospital stay in this study was 3 (0.8) days for TY:TORS and 3.47 (1.5) days for EC:TORS. In a study by Bordeaux et al,¹⁸ the mean hospital stay was 2.7 days. In a series of 45 patients with OPSCC, the mean length of stay was 3.8 days³; the main reason for prolonged hospital stay in this study was management of neck drain wounds.³ The safety of TY:TORS was further supported by the fact that in the 30-day postoperative period, there were no significant or life-threatening complications. Although there was no statistically significant difference in the incidence of postoperative bleeding, 5 patients in the EC:TORS group and none in the TY:TORS group required reoperation for this purpose. Postoperative mucosal bleeding is generally considered a minor complication⁵; however, it requires readmission and usually a return to the operative theater, placing the patient at additional risk.

To evaluate functional outcome, we looked to pain control and swallow function. Postoperative pain and analgesic requirements were higher in patients undergoing EC:TORS than in those undergoing TY:TORS (P = .02). This, may be due to the decreased collateral thermal damage, finer tissue cuts, and improved visualization associated with TY:TORS. This is relevant, as postoperative pain is closely associated with functional outcome and quality of life.²⁰ In a study published by Kehlet²⁰ in 1997, postoperative pain was found to impair organ function and delay mobilization and overall recovery. Furthermore, immobilization was found to increase the risk of thromboembolic and pulmonary complications and to increase fatigue, hypoxemia, and muscle loss.²⁰ Despite the fact that there were no significant differences between complications, overall length of stay, or functional outcomes between these 2 groups, this finding warrants further investigation as we try to identify strategies to improve functional outcomes in this patient population.

The function most at risk from laryngeal and oropharyngeal disease progression or treatment is swallowing³; therefore, we evaluated time to oral diet, gastrostomy tube dependence, and dysphagia diagnoses to evaluate procedure impact. In the TY:TORS group, 73% (11 of 15) compared with 47% (14 of 30) in the EC:TORS group were tolerating an oral diet before hospital discharge, and 93% (14 of 15) compared with 80% (24 of 30), respectively, were tolerating an oral diet by 30 days. Improvement in swallowing in the first 30 days after TORS, as demonstrated by both groups, is consistent with previously published TORS data.³ In a study by Skoner et al,²¹ 65% of patients resumed an oral diet after surgery but before radiotherapy. Moore et al³ reported that 88.9% of patients in a series of 45 patients with OPSCC undergoing EC:TORS were tolerating an oral diet within 4 weeks of surgery.

To further evaluate swallowing, other researchers have suggested that a reasonable surrogate is gastrostomy tube dependence.⁵ In this study, 53% of study patients (8 of 15) and 73% of controls (22 of 30) had an NGT placed intraoperatively. The average number of days to NGT removal was 13.3 and 24.6 days, respectively. Finally, we evaluated swallowing function based on informal and formal dysphagia studies. Although no difference was seen between the 2 groups, all 4 patients diagnosed as having severe dysphagia had SGSCC primary tumors. Of these, 1 patient was dependent on an NGT at 30 days but improved sufficiently to be diagnosed as having mild oropharyngeal dysphagia on POD 39 and successfully transitioned to a full oral diet. The remaining 3 patients were PEG tube dependent at 30 days. Of the 9 patients diagnosed as having moderate dysphagia, only 2 required a PEG tube. Three patients were discharged with an NGT, which was removed on PODs 9, 15, and 18. The remaining patients were provided exercises and swallowing modifications and ultimately tolerated an oral diet. No patient diagnosed as having normal swallow function or mild dysphagia required a PEG within 30 days of surgery. Data from this study support the conclusion that TORS is associated with a relatively low rate of gastrostomy tube dependence^3,5 and that TY:TORS offers functional outcomes equivalent to those of EC:TORS.

Despite these encouraging data, this study is not without limitations. It is subject to confounding and population bias and to limitations inherent to retrospective data collection and small sample size. We chose matching criteria a priori to create similar distributions of confounding variables.¹⁰ Bloom et al¹⁰ described the confounding phenomena as highly contingent on the proposed causal pathway between the risk factor (cutting device) and the outcome of interest (feasibility, safety, and functional outcome). We, therefore, chose to match based on tumor size, tumor location, age, and sex, as these factors were considered by the authors to most strongly bias the outcomes of interest in this study. In addition, by blinding the reviewing author to clinical outcomes during selection of the matched cohort, we strove to decrease population bias between these 2 groups. By matching 2:1, we attempted to increase the efficiency of the analysis. Finally, there are also limitations to the use of TY:TORS itself. It is considerably more expensive per use than is EC:TORS, with a startup cost of $62 500 for each laser and a total of $330 per use vs $110 per use for EC. In addition, regardless of the instrument used, sound technique and oncologic principles must remain an absolute priority.

In conclusion, the use of Tm:YAG during TORS for resection of OPSCC and SGSCC seems comparably safe to EC and may offer several distinct clinical advantages. The use of TY:TORS resulted in fewer IPs and less postoperative pain than did EC:TORS, which may be due to decreased collateral thermal damage, improved visualization, and finer cutting with the thulium laser. Future study is necessary to further validate these findings and to compare flexible CO₂ laser (CO₂:TORS) with TY:TORS and EC:TORS.