BOT indicates base of tongue; TORS, transoral robotic surgery.
The stage 2 algorithm is used 1 to 2 weeks after the first procedure to determine the necessity of the second surgical operation. BOT indicates base of tongue; TORS, transoral robotic surgery.
Customize your JAMA Network experience by selecting one or more topics from the list below.
Hatten KM, O’Malley BW, Bur AM, et al. Transoral Robotic Surgery–Assisted Endoscopy With Primary Site Detection and Treatment in Occult Mucosal Primaries. JAMA Otolaryngol Head Neck Surg. 2017;143(3):267–273. doi:10.1001/jamaoto.2016.3419
What is the rate of identifying occult mucosal squamous cell carcinomas using the transoral robotic surgery (TORS) technique?
In this retrospective study that included 60 patients, the TORS endoscopic approach successfully identified the occult mucosal cancers in 80% of patients.
Patients presenting with a cervical lymph node metastasis unknown primary benefit from an improved rate of detection and ultimately avoid the need for comprehensive radiation therapy to the entire pharyngeal axis.
Management of cervical lymph node metastasis without a known primary tumor is a diagnostic and treatment challenge for head and neck oncologists. Identification of the occult mucosal primary tumor minimizes the morbidity of treatment.
To analyze the role of transoral robotic surgery (TORS) in facilitating the identification of a primary tumor site for patients presenting with squamous cell carcinoma of unknown primary (CUP). In addition, we assessed treatment deintensification by determining the number of patients who did not undergo definitive radiation therapy and chemotherapy.
Design, Setting, and Participants
In this retrospective case series from January 2011 to September 2015, 60 consecutive patients with squamous cell CUP who underwent TORS-assisted endoscopy and ipsilateral neck dissection were included from an academic medical center and studied to study the rate success rate of TORS identifying occult mucosal malignancy.
Main Outcomes and Measures
Success rate of identifying occult mucosal malignancy; usage of radiation therapy and chemotherapy.
Overall, 60 patients (mean [SD] age, 55.5 [8.9] years) were identified; 48 of the 60 patients (80.0%) had a mucosal primary identified during their TORS-assisted endoscopic procedure. The mean (SD) size of the identified mucosal primary lesions was 1.3 (0.1) cm. All mucosal primaries, when found, originated in the oropharynx including the base of tongue in 28 patients (58%), palatine tonsil in 18 patients (38%), and glossotonsillar sulcus in 2 patients (4%). Among patients in this study, 40 (67%) did not receive chemotherapy, and 15 (25%) did not receive radiation therapy.
Conclusions and Relevance
Advances in transoral surgical techniques have helped identify occult oropharyngeal malignancies that traditionally have been treated with comprehensive radiation to the entire pharyngeal axis. We demonstrate the efficacy of a TORS-assisted approach to identify and surgically treat the primary tumor in patients presenting with CUP. In addition, patients managed with the TORS-assisted endoscopic approach benefit from surgical and pathological triage, which in turn results in deintensification of treatment by eliminating the need for chemotherapy in the majority of patients, as well as avoiding radiation therapy in select patients.
Regional lymph node metastasis from an unknown primary head and neck squamous cell carcinoma has been a diagnostic and management challenge since it was first described by Martin and Morfit in 1944.1 Knowledge of patterns of spread from head and neck mucosal primaries aids in identifying the locations that might harbor these occult malignancie.2 In addition, the availability of modern tools, including high-definition endoscopic instrumentation and radiographic imaging techniques, have significantly improved detection rates. Despite these technological advancements, a primary tumor cannot be identified in 2% to 5% of patients with head and neck cancer.3,4 Identification of these occult malignancies allows patients to avoid prophylactic large-volume mucosal radiation and has been shown to improve long-term survival.5
Traditional diagnostic work-up for patients with a carcinoma of unknown primary (CUP) includes a detailed history, physical exam, tissue sampling of the neck mass, and anatomic radiographic imaging.6 In addition, advances in imaging for CUP with positron emission tomography (PET) complemented with computed tomography (CT) or magnetic resonance imaging (MRI) have improved the detection rates of occult primaries by as much as 30%.7 Preoperative PET-CT and surgical endoscopy leads to the identification of mucosal malignancies in 47% to 59% of patients.8,9 However with PET imaging, there is a substantial risk of false positives as high as 39%.10,11 These persistent limitations have led to the search for alternative detection and treatment methods for patients with CUP. Transoral surgery has been recently advocated as an adjunct to traditional endoscopic techniques to perform excisional biopsies of high-risk subsites in the oropharynx to identify small mucosal malignancies.12-14
Transoral robotic surgery (TORS) offers dexterous instrumentation complemented with enhanced three-dimensional visualization to perform surgery on the oropharynx.15 Augmented visualization during TORS offers a potential technologic advantage in addressing CUP. We report our management of patients presenting with CUP, using TORS to facilitate primary tumor identification and excision of oropharyngeal mucosa at risk of harboring the mucosal primary.
An institutional review board–approved retrospective medical record review was performed on patients presenting from January 2011 to December 2015 who underwent TORS-assisted endoscopy as part of their diagnostic evaluation for CUP. For the purposes of this study, CUP was defined as an individual presenting with a squamous cell carcinoma cervical lymph node metastasis and uncertainty regarding the location of the primary cancer following standard of care outpatient clinical and radiologic assessment. Medical records were reviewed to ensure that all patients had a consistent preoperative work-up including office evaluation, fiberoptic nasopharyngolaryngoscopy, and preoperative anatomic imaging including CT and/or MRI that did not reveal a primary malignancy. Each patient was also reviewed at our institutional multidisciplinary tumor board session. All patients in the current study had biopsy-confirmed squamous cell carcinoma diagnosed by either fine-needle aspiration or excisional lymph node biopsy. Expression of p16 positivity and/or oncogenic human papilloma virus (HPV) genotyping was determined by the Department of Pathology and Laboratory Medicine at the Hospital of the University of Pennsylvania.
Patients were excluded if they had a prior history of head and neck cancer or had a mucosal lesion identified during in-office examination. However, individuals were not excluded based on suspicious PET-CT findings if no primary tumor was evident on physical examination due to the high-risk of false-positives detection. The high metabolism of lymphatic tissue and potential heterogenous uptake of 18F-fluorodeoxyglucose in the Waldeyer ring can lead to inaccurate localization using this technique.9
The surgical approach begins with a comprehensive endoscopic evaluation of the upper aerodigestive track including laryngoscopy, bronchoscopy, and esophagoscopy with biopsy of the nasopharynx (Figure 1). The surgeon biopsies any obvious lesions that become evident during the endoscopy. These specimens are sent for frozen section analysis to confirm malignancy. Small mucosal lesions that are not obviously malignant but, owing to coloration or texture appear suspicious on exam, are not biopsied, because frozen section analysis of HPV-related squamous cell carcinomas can be very challenging when lesions are small. Instead, our preference is to remove the entire lesion for permanent pathologic processing. If the frozen section biopsy results are positive for invasive squamous cell carcinoma and the primary is superficial and amenable to a limited TORS resection of the tongue base or glossotonsillar sulcus, then the surgeon proceeds with definitive TORS resection. Immediately following TORS, an ipsilateral neck dissection is performed during the same surgical session while permanent pathology is pending. If the primary tumor is identified in the tonsil, a TORS radical tonsillectomy is not performed concurrently with the planned neck dissection but staged at a later date due to a relatively high incidence of intraoperative pharyngocervical fistula reported in the literature.16
If a primary tumor is not identified during the endoscopic phase of the procedure, the patient undergoes an ipsilateral palatine tonsillectomy. The patient’s oropharynx is then exposed with a Crowe-Davis mouth gag (Storz) for visualization of the palatine tonsils. An extracapsular dissection of the palatine tonsil is performed without violating the pharyngeal constrictor muscle. The tonsil specimen is immediately serially sectioned in the pathology laboratory and grossly evaluated for the presence of submucosal lesions that might be consistent with carcinoma. Frozen section analysis is performed only when a grossly obvious malignancy is noted. If the biopsy results are positive for carcinoma of the palatine tonsil, the search for the CUP is completed and an ipsilateral neck dissection is performed. A TORS radical tonsillectomy is scheduled in a staged fashion as previously described.
If the palatine tonsil tissue is negative for carcinoma by gross visualization or on frozen section, the patient undergoes an ipsilateral lingual tonsillectomy using the da Vinci Surgical System (Intuitive Surgical). In HPV-related squamous cell carcinoma, the contralateral tonsil is not resected due to the extremely low probability of a contralateral primary in patients presenting with unilateral cervical lymph node metastasis. The surgical resection only includes the mucosa, underlying lymphoid tissue, and a very thin layer of muscle from the circumvallate papilla to the vallecula.
TORS lingual tonsillectomy is performed using the Feyh-Kastenbauer-Weinstein-O’Malley retractor to provide exposure of the tongue base. The da Vinci 0-degree 11-mm endoscope and 2 instrument arms mounted with a 5-mm Maryland dissector and a 5-mm monopolar spatula cautery are introduced into the oropharynx. A horizontal mucosal cut is first made just posterior to the retractor blade and carried to the depth of the junction of lymphoid tissue and muscle. Midline and lateral cuts are made to the same depth and the dissection is carried to the vallecula. The specimen is then evaluated in the pathology laboratory at which time the superior, inferior, medial, lateral and deep margins are inked with different color inks and serially sectioned to look grossly for any suspicious lesions. If no obvious lesion is found frozen section analysis is deliberately avoided because, in our opinion, many of the tongue base primaries are microscopic, and it is better to assess these small lesions with serial sectioning of the entire specimen as a permanent section. If a primary cancer is grossly detected in the tongue base and margins appear widely negative, no frozen sections are performed. However, if any mucosal margins appear grossly positive or close, an adjacent mucosal margin is resected at that time. When deep margins appear close or positive, then definitive surgical resection is deferred to a later date to avoid the risk of creating continuity between the primary site and neck and to avoid delayed fistula. An additional benefit of performing the neck dissection approximately a week prior to the definitive TORS resection is that the surgeon is able to ligate the ipsilateral lingual, facial, and superior laryngeal arteries which has been suggested to decrease bleeding during and after the TORS primary tumor resection.17
As described, all patients in whom the primary is located in the tonsillar fossa return to the operating room within 10 days of the TORS-assisted endoscopy and neck dissection and undergo a TORS radical tonsillectomy.18 If a tongue base primary is found and there is a microscopically close (<2 mm) or positive margin, then the patient returns for TORS (Figure 2). The patient then undergoes wide mucosal reresection of the mucosal margin that was close or positive on the initial resection. If the tongue base primary is identified and had a positive deep margin, our preference is to perform a TORS tongue base hemiglossectomy, which is our standard approach for almost all tongue base resections.19
Since the vast majority of cancers are found on the side ipsilateral to the neck metastasis, our current approach is to refrain from bilateral resection of the lingual and palatine tonsils in order to avoid additional morbidity in the majority of patients. In patients where no ipsilateral cancer is found, the patient returns for contralateral palatine tonsillectomy. If no primary is identified, this is immediately followed by a limited contralateral lingual tonsillectomy. Secondary procedures are planned as described above if permanent section margins are close or positive. Lastly, in our practice the contralateral neck dissection is only considered in patients who would potentially benefit from single modality surgical therapy with a single nodal metastasis and no evidence of extracapsular spread that would warrant adjuvant therapies.
Between January 2011 and December 2015, 60 consecutive patients met the criteria for this study and underwent TORS-assisted endoscopy in search of an occult primary. The patient demographics for the study are presented in Table 1. The majority of these patients were male (n = 55 [91.7%]) with a mean (range) age of 55.5 (42-69) years. A history of tobacco use was present in the majority of patients (n = 36 [60%]).
Prior to the planned TORS-assisted endoscopy, tissue sampling of the cervical lymph node metastasis was performed in all subjects. Squamous cell carcinoma was identified by fine-needle aspirate or an excisional biopsy (n = 26 [43%]) in those patients who had an inconclusive fine-needle aspirate or had an excisional biopsy performed elsewhere. The pathology of biopsies performed outside of our institution (n = 34 [57%]) was reviewed at the Hospital of the University of Pennsylvania for diagnostic confirmation. Presence of HPV was assessed by either DNA extraction of cytology specimens testing for the presence of high-risk HPV serotypes and/or immunostaining for p16 protein, a surrogate marker for HPV-related oropharyngeal carcinoma.20,21 Fifty-five (92%) of the enrolled patients were positive for at least 1 of the 2 markers for HPV-related squamous cell carcinoma.
The type of radiologic studies available prior to TORS-assisted endoscopy varied because the majority of patients presented to our institution after cross-sectional anatomic imaging had been obtained by an outside facility. Contrast-enhanced CT was the most common imaging modality (n = 44 patients); CT scanning suggested the location of the primary mucosal tumor in only 5 patients (12%), with 2 CT scans accurately predicting the primary tumor location. Fourteen patients underwent MRI with contrast; 4 patients (29%) had findings suggestive of a possible primary and of these 4 patients, 3 were ultimately in the location of the primary malignancy. All but 1 patient had a PET-CT performed prior to endoscopy, and 32 out of 59 (54%) failed to identify a discrete hypermetabolic primary mucosal lesion. Of the remaining 27 patients who underwent PET-CT and appeared to have identified a potential mucosal primary, 9 incorrectly identified the primary, yielding an overall false-positive rate of PET-CT in this series of 33%.
The mucosal malignancy was identified in 48 of the 60 (80%) patients using the surgical protocol described above. The mean (SD) [range] diameter of the primary tumor found during the endoscopic surgery was 1.3 (0.1) [0.2-4.2] cm. All tumors were histologically confirmed to be squamous cell carcinoma and each identified in the oropharynx. Of these 48 cancers, 28 primary tumors (58%) were identified in the base of tongue, 18 cancers (38%) were identified in the palatine tonsils, and 2 were located in the glossotonsillar sulcus (Table 2).
All patients underwent ipsilateral neck dissection in addition to the TORS-assisted endoscopy. Descriptive features of nodal pathology are described in Table 2. Extracapsular spread (ECS) of lymph node metastases was assessed based on pathologic assessment of neck dissection specimens and was present in 18 patients (30%) and absent in 38 patients (63%); ECS status was not available in the remaining 4 patients due to prior excisional lymph node biopsy at outside institutions where no comment was made regarding ECS.
All patients tolerated an oral diet prior to their hospital discharge following the TORS-assisted endoscopy. In addition, all but 1 patient went on to receive definitive treatment for head and neck squamous cell carcinoma within the University of Pennsylvania Health System. The 1 patient who did not ultimately receive definitive treatment following TORS assisted endoscopy was attributable to a perioperative death. The patient experienced an acute cardiopulmonary event while ambulating 3 days following secondary TORS hemiglossectomy and died.
Eight patients (13%) experienced surgery-related complications. The most common complication was a postoperative oropharyngeal bleed, seen in 3 patients (5%). Of these 3 patients, 2 returned to the operating room for cauterization. The remaining complications were related to the cervical lymphadenectomy performed at the time of surgery (Table 3). No pharyngocervical fistulas were encountered using the staged TORS-assisted endoscopic surgery.
The management of the head and neck unknown primary has changed significantly from the original description of a heterogeneous grouping of tumors that required comprehensive pharyngeal mucosal radiation and concurrent chemotherapy to all the tissue felt to be at risk.3,22 The vast majority of occult malignancies are now found to originate in the cryptic lymphoid tissue of the oropharynx.23 CUP malignancies are commonly identified as occult oropharyngeal squamous cell carcinomas, attributable to the oncogenic potential of HPV.24,25 Human papillomavirus–related oropharyngeal squamous cell carcinomas are associated with an improved disease-specific survival and overall survival when compared with non-HPV–related squamous cell carcinomas.26,27 Despite HPV oropharyngeal cancers having excellent oncologic outcomes and presenting frequently as CUP, detection rates still fail to identify 40% of mucosal primaries in this setting.28 We have developed a unique treatment algorithm for CUP that is tailored to improve detection rates of the primary cancer and potentially avoid comprehensive chemoradiotherapy to the pharyngeal axis.
Detection methods for identifying the CUP have evolved significantly over time since the approach described by Martin and Morfit in 1944,1 with the introduction of CT, MRI, and PET scans as screening tools. However, anatomic imaging has demonstrated limited value in our series, with CT accurately identifying the primary in 2 out of 44 scans (5%). Surprisingly, the largest primary was 4.2 cm and was not identified on CT imaging. PET scanning continues to show a high rate of false positives related to the elevated metabolic activity in the Waldeyer ring.29 In our series, only 18 of the 27 positive PET-CT scans were able to accurately identify the correct subsite, resulting in a false positive rate of 33.3% (9of 27). The high rate of false positives confirms that additional diagnostic tools are needed to assist in finding the occult malignancies.
Surgical techniques, including the palatine tonsillectomy, have proven to be more effective for identifying an unknown primary tumor when compared with deep biopsies.30 However, only recently has an excisional biopsy of the lingual tonsils been a feasible surgical option in the diagnostic work-up for the CUP. The novel approach of using TORS to identify the CUP was first introduced in 2011 by Abuzaid.31 Additional studies have described the diagnostic success of transoral surgery to identify occult primaries, ranging from 72% to 94%,12,32,33 yet there has been limited evidence that the surgical technique has reduced the need for therapeutic chemotherapy and radiation therapy, as has been shown in this study.
Treatment deintensification for CUP is the primary goal for our TORS approach to this diagnostic dilemma. In this series of CUP, 54 of the 60 patients (90%) presented with stage IV squamous cell carcinoma, for which the National Comprehensive Cancer Network guidelines recommends chemotherapy and radiation to “putative mucosal primary sites.”34 This treatment regimen of concurrent chemoradiotherapy with intensity-modulate radiation therapy has excellent 2-year overall survival rate of 92%. However, the impact on swallow function after total mucosal chemoradiotherapy results in the majority of patients developing esophageal strictures and 50% of patients ultimately undergo esophageal dilations.35 Our treatment goal was to minimize comprehensive chemoradiotherapy and associated treatment morbidity. In the current study, 40 of the 60 patients (67%) did not receive chemotherapy as they had their primary oropharyngeal cancer resected with negative margins and no evidence of extracapsular spread in cervical lymph node metastasis. Chemotherapy was indicated in 18 patients with evidence of extracapsular spread on pathologic review of cervical lymph node metastasis,36 as well as 2 patients without an identifiable primary following TORS surgical session and a nondiscernable ECS status. In addition, 15 patients (25%) were treated with surgery alone. Patients with low-volume neck disease are ideal for the TORS-assisted approach to identify occult mucosal malignancies that may potentially be treated with unimodality therapy. Six patients had N1 nodal disease, and 9 patients had N2a and elected to forego radiation therapy after consultation with radiation oncology. Of note, the treating radiation therapist recommended adjuvant radiation therapy for N2a disease; however, also offered the alternative of observation for select patients with a single lymph node metastasis.
The safety profile for the robotic portion of the unknown primary has resulted in a 5% complication rate, manifested as a postoperative oropharyngeal bleed in each case. Despite the TORS surgical approach resecting significantly more oropharyngeal tissue compared with the traditional palatine tonsillectomy, the rate of postoperative hemorrhage is comparable to the 4.6% bleed rate and 3% reoperation rate reported with adult tonsillectomy.37,38
Despite the strengths of this study, which represents, to our knowledge, the largest reported experience to date using the TORS-assisted approach to CUP, there are significant limitations associated with its retrospective design. In addition, there are no documented long-term oncologic outcomes for patients with CUP treated with the TORS-assisted approach. It is, however, a reasonable assumption that patients in whom an HPV–related primary site cancer was identified would experience at least comparable oncologic outcomes to those previously reported in whom the primary malignancy was known and treated with TORS, which for 114 consecutive patients treated at the University of Pennsylvania was a 2-year, 3.3% local-regional failure rate and 8.4% distant metastatic rate.39
Our TORS-assisted approach to CUP is a potentially valuable tool for identifying and thereby deintensifying treatment for patients initially presenting with CUP. This novel surgical approach improves identification rates of occult mucosal malignancies compared to traditional endoscopic and radiographic approaches. In addition, this approach has reduced the need for chemotherapy as well as radiation in select patients who would otherwise be treated with comprehensive chemoradiotherapy to all at-risk mucosal sites.
Corresponding Author: Kyle M. Hatten, MD, Department of Otorhinolaryngology–Head & Neck Surgery, University of Maryland School of Medicine, 16 S Eutaw St, Ste 500, Baltimore, MD 21230 (email@example.com).
Accepted for Publication: September 10, 2016.
Published Online: December 8, 2016. doi:10.1001/jamaoto.2016.3419
Author Contributions: Dr Hatten had full access to all of the data in the study and takes responsibility for the integrity of the data and accuracy of the data analysis. Drs Hatten, O’Malley, and Weinstein contributed equally. Drs O’Malley Jr, Hatten, and Weinstein are cofirst authors.
Concept and design: Hatten, Bur, Patel, Rassekh, Cannady, Chalian, Hodnett, Lukens, Weinstein.
Acquisition, analysis, or interpretation of data: Hatten, O'Malley, Bur, Patel, Rassekh, Newman, Cannady, Chalian, Hodnett, Lin, Cohen, Bauml, Montone, Livolsi, Weinstein.
Drafting of the manuscript: Hatten, Patel, Rassekh.
Critical revision of the manuscript for important intellectual content: All Authors.
Statistical analysis: Hatten.
Administrative, technical, or material support: O'Malley, Newman, Chalian, Cohen, Bauml, Weinstein.
Study supervision: Bur, Rassekh, Newman, Cannady, Chalian, Lin, Montone, Weinstein.
No additional contributions: Patel, Hodnett, Lukens, Livolsi.
Conflict of Interest Disclosures: All authors have completed and submitted the ICMJE Form for Disclosure of Potential Conflicts of Interest. Drs Weinstein and O’Malley receive royalties from Olympus Corporation. Dr Patel received compensation serving on the Advisory Board for AstraZeneca and proctors for Intuitive Surgical. No other conflicts are reported.
Previous Presentation: This study was presented at the American Head & Neck Society Ninth International Conference on Head and Neck Cancer; July 16-20, 2016; Seattle, Washington.
Create a personal account or sign in to: