In this study, approximately half of p16-negative OPSCCs were detected by GIE, whereas only 4.3% of p16-positive OPSCCs were identified in this manner. Consequently, the proportion of lesions detected at cTis-1 was lower in p16-positive cases, and there were very few cases that could be treated with transoral resection.
The most potential factor in this finding is the accessibility of endoscopic examination. As presented in Table 2, p16-negative OPSCC is more likely to involve risk factors such as histories of alcohol consumption, smoking, ESCC, and HNSCC. Consequently, these patients are categorized as high risk, and they are more frequently subjected to GIE surveillance with image enhancement. Conversely, HPV-related OPSCC is typically diagnosed in younger individuals with less exposure to smoking and alcohol26, which might lead them being perceived as lower risk. As a result, patients with HPV-related OPSCC are less likely to undergo GIE than those with HPV-unrelated OPSCC. Therefore, patients with HPV-related OPSCC are frequently identified at an advanced stage of progression, typically when symptomatic manifestations have become evident. If high-risk cases of HPV-related OPSCC could be identified and actively screened using GIE, it might increase the early detection of HPV-related cancers. Currently, there is no available screening test to detect HPV status in the general population, which makes it impractical to identify high-risk cases. In the future, HPV vaccination history may help stratify the risk of HPV-related OPSCC27, potentially guiding screening strategies and improving early detection.
However, considering that approximately 30% of p16-negative OPSCC without a history of ESCC or HNSCC are detected by GIE, it suggests that factors other than the limited opportunities for undergoing GIE may be responsible for this difference in detection rate. The location of the lesion is also a crucial factor influencing the GI endoscopic detection of OPSCC. Hamada et al. reported that the anterior and lateral walls of the oropharynx are anatomically difficult to observe using GIE28. Consistent with previous reports, our data, as shown in Fig. 2, also indicate that the proportion of lesions detected by GIE was low for anterior wall and lateral wall lesions in p16-negative cancers. Additionally, it has been reported that HPV-related OPSCC is more likely to arise in the anterior and lateral walls of the oropharynx29, and our data showed a similar pattern. Considering these findings, the predilection of p16-positive OPSCC to arise in challenging-to-observe sites makes GIE detection more difficult.
Furthermore, the progression process of HPV-related OPSCC is likely to contribute to its lower GI endoscopic detection rate. These cancers originate from the basal layer of the tonsillar crypt, they are in deeper areas of the tonsils, and the tumors are usually covered by normal mucosa30. Thus, HPV-related OPSCC is not exposed to the luminal surface in the early stage. A tumor growing within the palatine tonsils might be challenging to detect by GIE. In our data, most of the p16-positive OPSCC cases were protruding lesions, but they were probably difficult to detect because they were subepithelial tumor-like smooth elevations and lacked color variation compared to the surrounding tonsillar tissue, making them difficult to identify. In the present study, only the limited number of p16-positive OPSCC were observed with NBI, so the utility of NBI in detection could not be determined. However, in two cases observed using NBI, as presented in Supplemental Figure (3-b and 3-h), microvessels resembling intrapapillary capillary loops were identified. There are case reports suggesting that the use of NBI with laryngoscopy can be effective for detection of HPV-related OPSCC31,32. Although the early detection of p16-positive OPSCC remains challenging, NBI observation might potentially be useful.
This study had some limitations. First, this was a retrospective review of a single center. Second, this study includes cases detected by GIE performed at other institutions. Therefore, it is unclear whether the GIE was performed transnasally or transorally, whether image-enhanced endoscopy was used, or the details of the previous endoscopic history. Subsequently, HPV infection was determined only by p16 immunostaining in this study. In clinical practice, the stage of cancer is diagnosed using the results of p16 immunohistochemistry25. The p16 immunostaining was reported as efficient and highly sensitive for HPV infection, but its specificity was insufficient. In a recent report, 7.0%-13.6% of patients were incorrectly diagnosed as HPV-negative33,34 In situ hybridization is highly specific, and it has been reported to be effective when performed in conjunction with p16 immunostaining35. However, in clinical practice, it is not easy to perform in situ hybridization because of its high cost and complicated procedure. Furthermore, in this study, there was a large difference in the number of detections by GIE between p16-positive and p16-negative cases. Even if 7.0%-13.6% of p16-positive cancers were in fact HPV-negative, we concluded that this result is robust and unlikely to be reversed.