2018-19 Section 7-Neoplastic and Inflammatory Diseases of the Head and Neck eBook

Otolaryngology–Head and Neck Surgery 156(5)

and ear primary site. 4-6,11-15 In the seventh edition of the American Joint Commission on Cancer (AJCC) staging manual, these features are used to determine T stage, with 2 ‘‘high-risk features’’ qualifying a tumor as a T2. 16 A staging modification proposed by Jambusaria-Pahlajani et al found better stratification of outcomes with an alternative stratification of these same risk factors. 5 The National Comprehensive Cancer Network also recommends consider- ation of tumor border characteristics, primary site, growth rate, neuropathic symptoms, and immunosuppression when classifying tumors as low versus high risk. 17 While factors predictive of recurrence and poor survival are relatively consistent in studies describing patients with early-stage disease treated by dermatologists, there is less consistency for locally advanced tumors treated by head and neck surgeons and radiation oncologists. Most CSCCs treated by head and neck surgeons are recurrent, have sev- eral high-risk features, or have regional disease. For this subset of CSCC patients, several series have demonstrated that PNI, tumor differentiation, diameter, and DOI are also highly predictive of recurrence and poor survival. 7,18-22 However, other series have found that factors such as immunosuppression, parotid and cervical nodal disease, and node size are more strongly associated with poor out- comes. 10,23-26 In our head and neck oncologic surgery practice, we see a large volume of patients with head and neck CSCCs that are either recurrent or deemed too complex for treatment in the office setting by a dermatologist. Given the lack of con- sistency in the literature regarding which factors are predic- tive of recurrence and nodal spread and survival for this group of patients, our objective was to evaluate these ques- tions in our unique patient population. Methods This study was approved by the Institutional Review Board of the University of California–Davis. Patient data were col- lected by querying the electronic medical records of patients seen in the otolaryngology department for International Classification of Diseases, Ninth Revision codes and identi- fying patients with CSCCs of the head and neck treated between January 1998 and June 2014. All patients under- going primary surgery, with or without adjuvant therapy, with curative intent for CSCC were included. Patients were excluded if they had distant metastases at presentation, were treated with palliative intent, and had \ 3 months of follow- up. We collected patient data information (age, sex, and immunologic status) and data regarding tumor characteris- tics (primary site, DOI, diameter, lymphovascular invasion, PNI, presence of regional nodal disease, histologic differen- tiation, adjuvant therapy, and whether tumors were recurrent on presentation). The mean age was 70.4 years; therefore, the age variable was categorized into 2 groups, 70 and . 70. The DOI variable was grouped into 3 categories ( \ 1, 1-2, . 2 cm), as this provided well-balanced group sizes and only 3 patients were below the AJCC staging high-risk threshold of 2 mm. 16 The diameter variable was grouped

into 3 categories ( \ 2, 2-4, . 4 cm) given the AJCC thresh- old of a 2-cm diameter as a high-risk feature. 16 Histologic differentiation was grouped as either poorly differentiated or well to moderately differentiated. Collected outcome data included recurrence, date of recurrence, location of recur- rence, date of last follow-up, date of death, and cause of death. Surgical resection of primary tumors was performed with the intent of achieving negative margins through wide local excision and included radical and craniofacial resection when necessary. Decision to perform a parotidectomy and/or cervi- cal lymphadenectomy was made by the primary surgeon for each case, with patients undergoing therapeutic lymphade- nectomy for clinically evident disease or elective lymphade- nectomy when deemed at high risk for nodal failure. Statistical analyses were performed with SPSS 23.0 (SPSS Inc, Chicago, Illinois). Five-year estimates of disease-free survival (DFS) were made with the Kaplan-Meier method. The associations of patient and tumor characteristics with recurrence and survival were examined with univariate and multivariate cox regressions. Analysis of overall survival and disease-specific survival was limited given incomplete cause of death data and the advanced age of our patient cohort. A multinomial logistic regression was performed to identify fac- tors independently associated with the presence of nodal metastasis at presentation. P \ .05 was used to indicate sta- tistical significance. Results A total of 212 patients met the inclusion criteria, with a mean age of 70.4 years (range, 32-94; median, 73). The majority of patients (87.3%) were men, and the most common primary tumor sites were of the cheek and temple (28.3%) or ear (22.6%; Table 1 ). Mean tumor diameter was 3.65 cm with an average DOI of 1.38 cm. A large propor- tion of tumors were recurrent (52.4%), had PNI (36.8%), or presented with regional nodal disease (29.2%). Only 7 patients (3.3%) were immunocompromised, 3 due to main- tenance of a transplanted organ and 4 with chronic lympho- cytic leukemia. Only 1 patient had a positive margin following surgery, and this patient did not have a recurrence prior to his death, from a ruptured appendix resulting in sepsis and cardiopulmonary arrest. The mean follow-up time was 35 months. There were 67 recurrences (31.6%): 49 local, 18 regional, and 3 distant, with 3 patients presenting with multiple simultaneous sites of recurrence. Of all recurrences, 73% were local, and 61.1% regional failures were in patients who had undergone a therapeutic neck dissection. Five-year Kaplan-Meier esti- mate of DFS was 53.2% for the entire patient cohort ( Figure 1 ). Given the advanced age of our patients and the limited data on cause of death, analysis of overall survival and disease-specific survival were excluded. Eighty-four (39.6%) patients received adjuvant radiation therapy, 11 of which also received chemotherapy. A Cox univariate regression demonstrated that patients with periorbital tumors (hazard ratio [HR] = 3.87, P = .012)

87