Natural history of superficial head and neck squamous cell carcinoma under scheduled follow-up endoscopic observation with narrow band imaging: retrospective cohort study

Observation using a narrow band imaging (NBI) endoscope with magnified view makes it possible to visualize microvascular irregularities such as abnormalities of the intra-papillary capillary loop (IPCL). Several prospective randomized studies using this imaging technique have shown that the detection rate of superficial squamous cell carcinoma of the larynx and pharynx is enhanced [1‐3].

Almost all superficial head and neck squamous cell carcinoma (HNSCC) can be cured with favorable prognosis by endoscopic resection (ER) or partial resection [4].

The synchronous or metachronous occurrence of esophageal squamous cell carcinoma (ESCC) and other HNSCCs is observed frequently; the former and latter rates are reported to be 14–83 % and 28 %, respectively [5, 6]. It is difficult to decide on a treatment course in such patients with synchronous or metachronous cancers. Generally, treatment for more advanced cancer takes precedence over that for other cancers when multiple synchronous cancers are detected at once. Therefore, treatment for superficial HNSCC is generally planned after concomitant advanced cancers have achieved a cure. Also, superficial HNSCC can be observed alone without any therapeutic intervention in cases where patients have an intolerable physical condition or concomitant further advanced stage of ESCC or HNSCC at other sites. However, there have been few reports regarding the natural history and prognosis of superficial HNSCC; therefore, therapeutic strategies are not well established. The aim of this retrospective study is to investigate the natural history of superficial HNSCC.

To our knowledge, this is the first study examining the natural history of superficial HNSCC. In this study, all lesions were diagnosed with NBI endoscopy and histologically confirmed with findings from biopsy specimens, and most of them progressed naturally. Takemura et al., reporting on the natural history of flat-type brownish lesions 5 mm or less in size in the oropharynx, found that the lesions did not change during 2 years of follow-up. However, pathological diagnosis was not performed and all lesions were described based on endoscopic findings alone [9]. Brownish areas of the pharynx vary from inflammation to invasive SCC. We previously reported the existence of basal cell hyperplasia (BCH), which is recognized as a small brownish area with NBI, but does not fulfill the pathological diagnostic criteria of neoplastic lesions such as SCC or dysplasia [8]. Most BCH in our previous report were 5 mm lesions or smaller. In contrast, most lesions enrolled in the present study were 5 mm or larger when detected. While most lesions were flat type, 85 % of them progressed. Therefore, the clinical course of superficial HNSCC confirmed histologically was different from those of flat-type brownish micro lesions. We believe that a 3 mm or larger superficial HNSCC is a significant lesion that requires careful follow-up or endoscopic intervention.

There are several reports about the depth of tumor invasion. It has been reported that macroscopic classification is related to tumor invasion. We previously reported that submucosal invasion was found in significantly more type 0-I and type 0-IIa lesions than in other macroscopic types. In addition to macroscopic classification, the rate of submucosal invasion increased significantly with larger tumor size [10]. Tateya et al. also reported that all lesions of type 0-I showed submucosal invasion, and 54 % of type 0-IIa lesions showed submucosal or muscular invasion. The ratio of submucosal or muscular invasion in each macroscopic type showed a significant difference [11]. Moreover, Fujii et al. reported there was significant correlation between the microvascular density of pharyngeal SCC and intra-epithelial SCC thickness and submucosal invasion [8]. As mentioned earlier, endoscopic findings can estimate tumor invasion and thickness. However, Taniguchi et al. reported that cervical lymph nodal metastasis of intra-epithelial SCC and submucosal invasive SCC of the pharynx were 0 % (0/77) and 9 % (7/75), respectively. Furthermore, in submucosal invasive SCC, tumor thickness of over 1000 μm was a significant risk factor for nodal metastasis and venous or lymphatic invasion [12]. As stated earlier, there were several reports about the endoscopic findings of submucosal invasive SCC. We decided the treatment plan of superficial HNSCC by those endoscopic findings. In our present study, although there were type 0-IIa lesions, no lesions had other obvious endoscopic findings indicating submucosal invasion at initial diagnosis. Considering these findings, it is suggested that all patients could be cured by ER. However, endoscopic findings, such as tumor enlargement or enhanced thickness, surface irregularity, and protrusions or depressions that suggested submucosal invasion appeared in nine lesions at the earliest at 11 months after diagnosis. Moreover, one progressed lesion treated with surgical resection at 13 months after diagnosis had invaded the submucosal layer, although there were no endoscopic findings during the initial and follow-up examination indicating submucosal invasion. These lesions might have a risk of nodal metastasis, and it might be difficult to cure only with ER or partial resection. This fact suggests that superficial HNSCC should be treated at an early stage, especially before tumor enlargement or appearance of endoscopic findings that suggest submucosal invasion. It is suggested that treatment for superficial HNSCC no larger than 20 mm commence within 1 year, because endoscopic findings indicating submucosal invasion appeared within 1 year after initial diagnosis in this study.

There are several limitations of this study. First, the median follow-up of 20 months might be insufficient to clarify the natural history of superficial HNSCC. Although 17 of the 20 lesions progressed in size during the endoscopic follow-up period, there were no patients who died from progression of superficial HNSCC. The follow-up period might be too short to reveal whether the superficial HNSCC would progress to be a cause of death. However, we believe that superficial HNSCC could be a cause of death because most lesions progressed to submucosal invasion in this study. Furthermore, we divided cohorts into two groups based on whether they progressed or not. We could not identify specific characteristics, such as endoscopic macroscopic type and location of lesions, in the progression group. Also, the follow-up period for the non-progression group was relatively short compared to that of the progression group, and two lesions were treated with ER and the other lesion was exposed to systemic chemotherapy in non-progression group.

Finally, the number of patients and lesions was small in this study. Further study with a large number of cases and a longer follow-up period at a multicenter setting will be required to clarify the natural history of these lesions and to clarify decision criteria for treatment of superficial HNSCC.

This study showed that most superficial HNSCC progressed in size naturally, suggesting they should be treated with less invasive treatment such as ER or partial resection when they are small, if the patient’s situation allows. Whereas, if patients have high comorbidities such as active cancers or a physical intolerability to surgery, these lesions should be carefully followed-up under endoscopic observation.