Risk factors for cervical lymph node metastasis in superficial head and neck squamous cell carcinoma

J Med Dent Sci 2015; 62: 19-24 Original Article Risk factors for cervical lymph node metastasis in superficial head and neck squamous cell carcinoma Toru Sasaki, MD 1, 2), Seiji Kishimoto, MD, PhD 1), Kazuyoshi Kawabata, MD 2), Yukiko Sato, DDS, PhD 3) and Tomohiro Tsuchida, MD, PhD 4) 1) Department of Head and Neck Surgery, Tokyo Medical and Dental University Graduate School of Medicine, Tokyo, Japan 2) Division of Head and Neck, Cancer Institute Hospital, Japanese Foundation of Cancer Research, Tokyo, Japan 3) Department of Pathology, Cancer Institute, Japanese Foundation of Cancer Research, Tokyo, Japan 4) Division of Endoscopy, Cancer Institute Hospital, Japanese Foundation of Cancer Research, Tokyo, Japan Introduction: The necessity of transoral surgery for head and neck carcinoma is increasing, but its indications for the treatment of superficial head and neck carcinomas have not yet been established. This study was intended to help establish the standard indications for transoral surgery and additional therapy in patients with superficial head and neck carcinoma. Methods: Sixty-two patients with 83 superficial head and neck carcinoma underwent transoral tumor resection at the Cancer Institute Hospital between June 2006 and September 2011. We measured the tumor size and thickness, examined the gross appearance, permeation of vessels, and droplet infiltration, and analyzed the correlations between each parameter. Results: Sessile type of tumor on gross appearance showed a significantly higher incidence of thickness 1000 µm than the other types. Tumor thickness 1000 µm was associated with higher incidences of permeation of vessels, droplet infiltration, and cervical lymph node metastasis. Conclusions: In superficial head and neck carcinoma, if the endoscopic gross appearance is the sessile type, tumor thickness is likely to be 1000 µm and risk of cervical lymph node Corresponding Author: Toru Sasaki, MD Division of Head and Neck, Cancer Institute Hospital, 3-10-6, Ariake, Koto-ku, Tokyo, 135-8550, Japan. Tel: +81-3-3520-0111 Fax: +81-3-3570-0343 E-mail: tsasaki@jfcr.or.jp Received September 30, 2014;Accepted January 8, 2015 metastasis is increased. Key words: Superficial head and neck carcinoma; transoral surgery; permeation of vessels; droplet infiltration; gross appearance Introduction The conventional therapy for early head and neck carcinoma is radiotherapy. However, some cases can no longer undergo radiotherapy because of previous radiotherapy for other head and neck carcinomas. Furthermore, radiotherapy is associated with problems, such as delayed radiation injury and radiation-induced cancer. Recent innovations in endoscopic devices like narrow-band imaging (NBI) have made it possible to detect superficial head and neck carcinomas. 1-3 Transoral surgery has been undergoing development as a minimally invasive, function-preserving surgery for superficial head and neck carcinoma. Since 2006 at our institute, endoscopic mucosal resection (EMR) has been applied for superficial head and neck carcinoma. The EMR is an effective treatment for superficial esophageal carcinoma, gastric carcinoma, and colorectal carcinoma. In these regions of the digestive tract, the indications for EMR and additional therapy are decided based on the depth of tumor invasion and permeation of vessels (PV). 4-6 However, the indications for transoral surgery for superficial head and neck carcinoma have yet to be established. The present study examined the relationship between

20 T. Sasaki et al. J Med Dent Sci each parameter, namely gross appearance of the tumor, 2 tumor size, tumor thickness, and PV. Furthermore, it was intended to help in the establishment of the standard indications for transoral surgery and additional therapy in patients with superficial head and 1 1 Tumor Epithelium Subepithelial layer neck carcinoma. Materials and methods The recent introduction of magnifying gastrointestinal endoscopy with NBI for head and neck region has allowed the detection of superficial carcinoma in the pharynx and larynx. Therefore, superficial head and neck cancers often refer to superficial lesions of pharynx and larynx. In this study, the superficial head and neck cancer was defined as the superficial lesions of the pharynx and larynx. Hence, 62 patients (60 men, 2 women) with a total of 83 lesions underwent transoral tumor resection based on a diagnosis of superficial head and neck squamous cell carcinoma at the Cancer Institute Hospital in Japan between June 2006 and September 2011. They were followed-up for more than one year after the surgery. Among the 83 lesions, 71 were hypopharyngeal carcinoma (pyriform sinus, 58; posterior wall, 9; postcricoid, 4), 6 were oropharyngeal carcinoma (base of tongue, 1; posterior wall, 4; vallecula, 1), and 6 were laryngeal carcinoma (arytenoid, 5; epiglottis, 1) (Table1). Mean age at diagnosis was 65.0 years (range, 41 84 years). Median duration of follow-up was 40.0 months (range, 13 79 months). Tumor thickness and size were measured as shown in Figure1. Median tumor thickness Table 1. Clinical characteristics, treatment profiles, and pathologic status of lesions Abbreviations: EMR-C, endoscopic mucosal resection with cap-fitting endoscope; ESD, endoscopic submucosal dissection; ELPS, endoscopic laryngo-pharyngeal surgery 1 Tumor thickness 2 Tumor size Figure 1. Methods for measuring tumor thickness and size was 500 µm (range, 170 7000 µm). Median tumor size was 15 mm (range, 4 55 mm). All lesions were N0M0. Gross appearance was classified into 3 types according to the General Rules for Clinical Studies on Head and Neck Cancer 5 th Edition: protruding (0-I); flat (0-II); or depressed (0-III). Protruding type was further classified into 2 subtypes: pedunculated type (0-Ip) or sessile type (0-Is). Flat type was also further classified into 3 subtypes: slightly elevated (0-IIa); flat (0-IIb); or slightly depressed (0-IIc). Eleven lesions were type 0-Is, 34 were type 0-IIa, 37 were type 0-IIb, and only 1 was type 0-IIc. No type 0-Ip or type III lesion was encountered (Table 1). The treatment methods comprised of EMR with capfitting endoscope (EMR-C) 7 for 40 lesions, endoscopic submucosal dissection (ESD) 8 for 38 lesions, and endoscopic laryngo-pharyngeal surgery (ELPS) 9 for 5 lesions (Table 1). The EMR-C and ESD are well-known techniques for superficial esophageal, gastric, and colon cancers. These procedures for pharyngeal lesions were done with the patients under general anesthesia in a supine position. In EMRC, a small plastic cap with a snare was attached to the tip of the endoscope. The lesion was suctioned into the cap. The snare closed at the base of the aspirated lesion, and the lesion was then resected electrosurgically. In ESD, mucosal incision was performed on the circumference of the lesion. The lesion was dissected at the submucosal layer. The ELPS is a transoral surgical technique using a curved laryngoscope. Superficial lesions were resected directly with the use of a flexible endoscope. After transoral surgery, each patient underwent both surveillance clinical examination and laryngoscopy at least every 3 months. A CT or ultrasound scan was also 2 1

Risk factors for lymph node metastasis in superficial HNSCC 21 carried out every 6 months to detect any lymph node metastasis. We measured the tumor size, tumor thickness, and PV; wherein, all of which are considered risk factors for lymph node metastasis (LNM). Tumors were measured as shown in Figure 1. We also checked for droplet infiltration (DI), which is a risk factor for LNM from esophageal carcinoma (Figure 2). The DI is defined as an isolated single cancer cell or cluster of cancer cells scattered beyond the tumor margin at the invasive front. 10, 11 This represents a similar pathological feature to tumor budding in colorectal adenocarcinoma. primarily due to superficial head and neck carcinoma was noted. Eight patients died due to other malignancies, and 1 patient died due to other disease without malignancy (Table 1). Figure 3 shows the correlations between tumor size and thickness and PV. Comparing the tumor thickness <1000 µm and 1000 µm, a significant difference was recognized in relation to the incidence of PV. On the other hand, no significant difference in the incidence of PV was seen in correlation to the tumor size (Table 2). A B Figure 2. Histological findings of droplet infiltration (DI) (arrow) Small cancer nests are found at the invasive front (Hematoxylin-eosin stain, original magnification A: X100, B: X400). Statistical analysis Statistical analysis was performed using Fisher s exact test in cross tables. A p value <.05 was considered statistically significant. All statistical calculations were performed using Ekuseru-Toukei 2010 (Social Survey Research Information Co., Ltd.). This study protocol was reviewed and approved by the ethics committee of our hospital. (IRB receipt number: 2013-1133) Figure 3. Correlations between each parameter, namely gross appearance of the tumor, tumor size, tumor thickness, PV and DI Positive results for PV are shown as squares and positive results for DI as diamonds and protruding-type (0-I) cases on gross appearance are shown as asterisk. Negative results for each parameter and flat type (0-II) cases are shown as white circles. Table 2. Correlations between pathological findings and tumor thickness or size Incidence of PV is significantly increased when tumor thickness is 1000 µm. No significant association is apparent between incidence of PV and tumor size. Similarly, incidence of DI is significantly increased when tumor thickness is 1000 µm. No significant association is apparent between incidence of DI and tumor size. Results On pathological examination, PV and DI were positive in 6 and 12 lesions, respectively (Table 1). The results of treatment were as follows: local recurrence occurred in 8 lesions, regional recurrence in the form of LNM in 3 patients, and no distant metastases in all patients. All local recurrence cases were horizontal marginal recurrence. No mortality

22 T. Sasaki et al. J Med Dent Sci The DI-positive cases tended to be 1000 µm thick (Figure 3). In like manner, tumors 1000 µm thick showed a significantly higher incidence of DI. However, no significant association was apparent between the incidence of DI and tumor size (Table 2). Figure 3 shows the relations between tumor size and thickness and gross appearance. All protruding lesions (0-I) were 1000 μm thick. There was a significant difference in the rate of PV between protruding lesions (0-I) and flat lesions (0-II) (Table 3). Postoperative LNM occurred in 3 cases. The rate of LNM was 3.6%. Tumor thickness in these three cases was 1000 µm. The rate of LNM for tumors 1000 μm thick was 17.6%. Tumors 1000 µm thick showed a significantly higher incidence of LNM. Furthermore, lesions with a 0-I gross appearance showed a significantly higher incidence of LNM than 0-II lesions (Table 4). Among the 3 cases of LNM, two cases were PVpositive. One case was PV-negative but DI-positive Table 3. Correlations between gross appearance and tumor thickness or size Frequencies of both tumor thickness 1000 µm and PV are significantly increased in protruding-type (0-I) lesions. Table 4. Correlations between lymph node metastasis (LNM) and tumor thickness or gross appearance Incidence of LNM is significantly increased in case of tumors 1000 µm thick and PV-positive cases. Table 5. Clinicopathologic characteristics of cases with cervical lymph node metastasis All cases were sessile type (0-Is) in gross appearance, 1000 µm thick, and positive for PV and/or DI. (Table 5). Neck dissection was performed with all LNM cases showing multiple metastases. As adjuvant therapy, two cases underwent chemotherapy with S-1, and 1 case received radiotherapy. All patients remain alive without carcinoma (Table 5). Discussion Indications for transoral surgery and standards for additional therapy after transoral surgery need to be established in the near future. Currently, one of the most important issues is the risk of LNM. With superficial esophageal carcinoma, the risk of LNM is based on tumor depth. Even in superficial head and neck carcinoma, some previous studies have argued that LNM correlates better with depth of invasion than 12, 13 with thickness. However, no lamina muscularis mucosae is present on the pharyngeal mucosa and mucosal thickness varies widely, so measurement of the tumor depth in pathological study is difficult. A virtual mucosal line needs to be constructed before measuring the tumor depth; however, setting up such a line reproducibly is difficult. Particularly in superficial carcinoma, since the measured value of tumor depth is small, slight differences in setup may exert major influences on the result. We therefore chose to measure tumor thickness in this study. Ultrasound endoscopy is considered useful to measure the depth of tumor invasion before EMR in esophageal carcinoma. 14 However, it cannot be used for the pharynx and larynx because these structures are complex and show bone and cartilage frameworks. Determining tumor thickness preoperatively is therefore difficult. According to the General Rules for Clinical Studies on Head and Neck Cancer 5 th Edition, if the tumor thickness is greater than 1000 µm, the frequency of lymph-vascular invasion is high. Therefore, we compared the tumor thickness <1000 µm and 1000 µm. This study revealed that when tumor thickness is 1000 µm, the incidences of PV and DI are significantly increased. Because PV and DI are increased, the risk

Risk factors for lymph node metastasis in superficial HNSCC 23 of LNM is also increased when tumor thickness is 1000 µm. A tumor 1000 µm thick is therefore considered a risk factor for LNM. On the other hand, no relationship was found between the tumor size and incidence of PV or DI. Hence, the risk of LNM cannot be estimated based on tumor size. In previous studies, pt classification has not shown any significant 13, 15 correlation with LNM. Tumor thickness is a pathological finding and thus is not valuable for preoperative treatment planning. The present study revealed a correlation between endoscopic gross appearance and tumor thickness. Type 0-I is a definitely protruding lesion diagnosed as superficial cancer based on the findings of size, height, and a relatively narrow basis. Type 0-Is is a lesion without a peduncle, and generally, the horizontal spread of the base is more conspicuous than the height. Type 0-II is a lesion without definite protrusion or depression. Type 0-IIa is a lesion with a slight elevation up to about 1mm in height. Type 0-IIb is a lesion without macroscopic elevation or depression. 15 Gross appearance is determined by the endoscopist according to endoscopic findings. Type 0-Is lesions tend to be 1000 μm thick, and the incidences of PV and LNM are significantly increased. When endoscopic gross examination shows a 0-Is type tumor, the patient should be considered to be at high risk for LNM. According to the results of this study, the risk of LNM was able to be estimated from the gross appearance before surgery. This finding poses a very big merit in making a treatment plan. Taniguchi et al. reported that tumor thickness >1000 µm might serve as the most predictive factor for nodal involvement, and that elevated type (0-I or 0-IIa) in endoscopic morphology showed a close relationship to lymph node involvement. 15 The incidence of cervical lymph node metastasis was 17.6% for tumors 1000 µm thick. Previous studies on superficial esophageal carcinoma showed that the rates of LNM were 0 9% for m3 and 8 32% for sm1. 16-21 These lesions are considered as indications for EMR without lymphadenectomy. Head and neck carcinomas 1000 μm thick carry a risk of LNM at the same level as m3 and sm1 esophageal carcinoma and are therefore considered to represent an indication for transoral surgery. The prognosis for superficial head and neck carcinoma is good. However, the sessile-type (0-Is) tumor carries a risk of LNM. Careful monitoring for LNM is necessary for patients with this type of tumor. Given the present results, adjuvant therapy may not be necessary even for tumors 1000 µm thick. No patients died of superficial head and neck carcinoma, and all cases with LNM appeared to be salvaged by neck dissection and postoperative therapy. Close follow-up led to avoidance of surplus treatment. Conclusion This study revealed that in superficial head and neck carcinoma, the sessile-type (0-Is) lesions in endoscopic gross appearance tend to be 1000 μm thick, and the incidences of PV and DI are significantly increased. Therefore, the sessile-type tumor carries a risk of LNM and should be carefully followed-up. In the future, it is necessary to determine the cutoff point of tumor thickness that requires elective neck dissection. Further accumulation of cases is necessary to clarify the factors associated with high metastatic risk and to establish the standards for additional therapy for superficial head and neck carcinoma. References 1 Muto M, Nakane M, Katada C, et al. Squamous cell carcinoma in situ at oropharyngeal and hypopharyngeal mucosal sites. Cancer 2004; 101: 1375-1381. 2 Watanabe A, Tsujie H, Taniguchi M, et al. Laryngoscopic detection of pharyngeal carcinoma in situ with narrowband imaging. Laryngoscope 2006; 116: 650-654. 3 Ugumori T, Muto M, Hayashi R, et al. Prospective study of early detection of pharyngeal superficial carcinoma with the narrowband imaging laryngoscope. Head Neck 2009; 31: 189-194. 4 Japanese Gastric Cancer Association. Japanese gastric cancer treatment guidelines 2010(ver.3). Gastric Cancer. 2011; 14: 113-123. 5 Tamegai Y, Fukunaga Y, Chino A, et al. Progress of Endoscopic Treatment for Early colorectal cancers J Jpn Soc Coloproctal 2012; 65: 800-807 [in Japanese with English abstract] 6 Momma K, Yoshida M, Fijiwara J, et al. Treatment Strategy of m3 and sm1 Esophageal Cancer Based on the Clinical Results of Endoscopic Resection Associated with Adjuvant Treatments. Stmach and Intestine 2006; 41: 1447-1458 [in Japanese with English abstract] 7 Suzuki H, Saito Y, Oda I, et al. Feasibility of endoscopic mucosal resection for superficial pharyngeal cancer: a minimally invasive treatment. Endoscopy. 2010; 42:1-7. 8 Iizuka T, Kikuchi D, Hoteya S, et al. Endoscopic submucosal dissection for treatment of mesopharyngeal and hypopharyngeal carcinomas. Endoscopy. 2009; 41: 113-117. 9 Satou Y, Omori T, Tagawa M. Treatment of superficial carcinoma in the hypopharynx [in Japanese with English

24 T. Sasaki et al. J Med Dent Sci abstract]. Nihon Jibiinkoka Gakkai Kaiho. 2006; 109: 581-586. 10 Egashira H, Yanagisawa A, Kato Y. Predictive factors for lymph node metastasis in esophageal squamous cell carcinomas contacting or penetrating the muscularis mucosae: the utility of droplet infiltration. Esophagus. 2006; 3: 47-52. 11 Teramoto H, Koike M, Tanaka C, et al. Tumor budding as a useful marker in T1-stage squamous cell carcinoma of the esophagus. J Surg Oncol. 2013; 108: 42-46. 12 Ambrosch P, Kron M, Fisher G, Brinck U. Micrometastases in carcinoma of upper aerodigestive tract: detection, risk of metastasizing, and prognostic value of depth of invasion. Head Neck. 1995; 17: 473-479. 13 Tomifuji M, Imanishi Y, Araki K, et al. Tumor depth as a predictor of lymph node metastasis of supraglottic and hypopharyngeal cancers. Ann Surg Oncol. 2011; 18: 490-496. 14 Lowe VJ, Booya F, Fletcher JG, et al. Comparison of positron emission tomography, computed tomography, and endoscopic ultrasound in the initial staging of patients with esophageal cancer. Mol Imaging Biol. 2005; 7: 422-430. 15 Taniguchi M, Watanabe A, Tsujie H, et al. Predictors of cervical lymph node involvement in patients with pharyngeal carcinoma undergoing endoscopic mucosal resection. Auris Nasus Larynx. 2011; 38:710-717. 16 Japanese classification of esophageal cancer. Japan Esophageal Society, 10 th ed. Tokyo: Kanehara; 2008. P67-70. 17 Shimada H, Nabeya Y, Matsubara H, et al. Prediction of lymph node status in patients with superficial esophageal carcinoma: analysis of 160 surgically resected cancers. Am J Surg. 2006; 191: 250-254. 18 Endo M, Yoshino K, Kawano T, Nagai K, Inoue H. Clinicopathologic analysis of lymph node metastasis in surgically resected superficial cancer of the thoracic esophagus. Dis Esophagus. 2000; 13: 125-129. 19 Fujita H, Sueyoshi S, Yamana H, et al. Optimum treatment strategy for superficial esophageal cancer: endoscopic mucosal resection versus radical esophagectomy. World J Surg. 2001; 25: 424-431. 20 Ancona E, Rampado S, Cassaro M, et al. Prediction of lymph node status in superficial esophageal carcinoma. Ann Surg Oncol. 2008; 15: 3278-3288. 21 Akutsu Y, Uesato M, Shuto K, et al. The overall prevalence of metastasis in T1 esophageal squamous cell carcinoma: a retrospective analysis of 295 patients. Ann Surg. 2013; 257: 1032-1038.