[Relationship between morphological characteristics and prognosis of non-nasopharyneal EBV-associated carcinoma].

ABSTRACT

Objective: To analyze the pathological features and their influence on the clinical outcome of non-nasopharyngeal EBV-associated carcinomas. Methods: One hundred and twenty cases of non-nasopharyngeal EBV-associated carcinoma confirmed by in situ hybridization were identified at Zhejiang Cancer Hospital from January 1, 2006 to May 1, 2018, and the clinicopathological data were collected and analyzed using Kaplan-Meier survival analysis, Cox univariate and multivariate analysis. Results: One hundred and twenty cases were involved in the study; the male to female ratio was 1∶1; patients' age range was 24 to 89 years (median 50 years). The primary sites were large parotid glands (62 cases), lung(26 cases), stomach(15 cases), and others (oral, oropharynx, larynx, cervix, liver; totally 17cases). Non-nasopharyngeal EBV-associated cancer could be divided into two histological types according to the amount of interstitial lymphocytes type Ⅰ was "lymphoepithelial-like carcinoma" and rich in stromal lymphocytes; type Ⅱ lacked lymphocytic infiltration. Ninety-eight primary tumor samples could be classified morphologically 43 cases were as type Ⅰ and 55 cases as typeⅡ; the distribution of type Ⅰ was 57.4% (27/47) in large parotid glands, 20.8% (5/24) in lung, 4/13 in stomach, and 7/14 in other sites. Complete treatment and survival data were obtained for 114 patients. According to the TNM staging criteria of WHO, 52 patients were at early stages (Ⅰ-Ⅱ) and 62 were at advanced stages (Ⅲ-Ⅳ); 102 patients underwent surgery. Seventy-four patients received adjuvant chemotherapy before or after surgery, and 52 patients received local radiotherapy. Kaplan-Meier survival analysis showed that patients with type Ⅱ EBV-associated carcinoma had a worse prognosis than patients with type Ⅰtumors (P=0.010 2). In addition, vascular invasion(P=0.021 8),neural recidivism(P=0.000 1),advanced stage(P=0.017 1),lymph node metastasis (P=0.005 0) and chemotherapy (P=0.013 2) were poor prognostic factors; female patients had better survival than male (P=0.028 4). Cox multivariate regression analysis found that lymph node metastasis (95%CI 1.489-13.830, P=0.007 6) and neural recidivism (95%CI 1.228-6.544, P=0.014 7) were independent adverse prognostic factors. Cox multivariate regression analysis after stratification by site revealed that radiotherapy was a preferable prognostic factor for EBV-associated carcinoma of the large salivary glands (95%CI 0.003-0.569, P=0.016 8). Conclusion: EBV associated carcinoma can be divided into two types, for which type Ⅰ was with abundant interstitial lymphocytes and type Ⅱ was lack of interstitial lymphocytes. TypeⅡ EBV-associated carcinoma has a worse prognosis than type Ⅰ. Radiation therapy can prolong the survival time of patients with primary EBV-associated carcinoma of large salivary glands.