LAUSR

In their recent study, Ali et al reported on the genomic profiles of 190 nasopharyngeal carcinoma (NPC) cases with different subtypes: 20 nasopharyngeal adenocarcinoma cases, 62 nasopharyngeal squamous cell carcinoma (NPSCC) cases, and 108 nasopharyngeal undifferentiated carcinoma (NPUC) cases. However, with regard to their study, we believe several issues are of concern. First, in the absence of any pathological data presented, the real subtype classification defined in the study by Ali et al introduced great doubts regarding the related genomic profiles described. The histological classification of NPC described was different from the updated World Health Organization classification. According to the current World Health Organization classification, squamous cell carcinoma (SCC) of the nasopharynx is classified into nonkeratinizing and keratinizing subtypes, as well as basaloid SCC. Nonkeratinizing tumors are subcategorized further as undifferentiated or differentiated. The distinct nasopharyngeal papillary adenocarcinoma tumor is very rare, and represents <1% of all NPC cases. Furthermore, Epstein-Barr virus (EBV) positivity is a key characteristics of nonkeratinizing NPC (undifferentiated or differentiated), and is present in 95% to 100% of nonkeratinizing NPCs in endemic areas. Thus, without pathological details stated in the article, the relatively low rate of EBV positivity reported (48.2%) was unexplained. The apparent lack of information regarding the source (endemic or nonendemic region), clinical staging (stage I to IV), and disease status (primary, local recurrent, or metastatic tumors) hinders us from making an accurate interpretation of the genome findings reported. Second, the authentic NPC origins of the sequenced tumors should be clearly addressed. For example, Table 1 in the study by Ali et al indicated that only 6 of 108 NPUC tumors originally were obtained from the “nasopharynx and paranasal sinuses.” Tumors arising from the nasopharynx are different from those arising from the paranasal sinuses. Furthermore, the remaining majority of NPUC tumors (102 cases) appeared to be largely obtained from the nasal cavity rather than the nasopharynx. In fact, from the tumor sites shown by the authors in Table 1, it is highly possible that tumors from other head and neck regions, instead of the nasopharynx, were included in their genome profiling. This is supported further by the “unusual” NPUC genomic profile described in their study: a very high rate of the isocitrate dehydrogenase 2 (IDH2) mutation (15.7%), which also was found across all 3 NPC subtypes defined in the study. In fact, to the best of our knowledge, all 4 NPC wholeexome sequencing studies published from 2014 through 2017 consistently demonstrated that an IDH2 (or even isocitrate dehydrogenase 1, IDH1) mutation event is absolutely rare in NPC. From our recently published genome study of nonkeratinizing NPC, which to the best of our knowledge is the largest published to date (>100 cases), there was only 1 somatic IDH2 mutation identified. The other 3 NPC genome studies reported no IDH2 mutations. Thus, the unconventional “combined site” defined by Ali et al as “nasopharynx and paranasal sinuses” may suggest a possible inclusion of sinonasal carcinoma for sequencing, which is known to harbor a high rate of IDH2 mutations. Third, the genomic profile of NPSCC in the 62 cases reported in the study by Ali et al appears to be largely similar to that of head and neck SCC. The tumor sites listed in Table 1 under the NPSCC subtype indeed included various head and neck regions (ear, head and neck, larynx, mouth, lymph nodes, nasal cavity, etc). In our opinion, the lack of accurate histological classification and evidence diminishes the clinical value of their findings regarding both major and rare NPC subtypes. Lastly, for the detection of latent EBV infection in patients with NPC, Epstein-Barr virus small RNA (EBER) in situ hybridization is the gold standard test, with a high sensitivity and specificity. This assay is performed routinely for the diagnosis of NPC by pathology laboratories in endemic regions (eg, Singapore and Hong Kong). Because EBV latent infection is a well-known characteristic of nonkeratinizing NPC in endemic regions, we believe that the EBV status, as determined by EBER in situ hybridization, should be performed carefully for all cases presented in the study by Ali et al. This is particularly important because to our knowledge none of the NPC genome studies published to date have had enough power in terms of case number to compare the genomic