With imaging improvement, volumetric method theoretically provided more accurate description for gross tumor. In other cancer sites, such as rectal cancer, the extent of tumor volume shrinkage is an established prognostic factor.[17], [18] In small cell lung cancer, response volume also correlated to outcome of definitive RT.[19] In locally advanved nasopharyngeal cancer, a study found a cutoff point (12.6%) of tumor volume reduction rate after neoadjuvant chemotherpay to serve as a prognositc value for disease control, including overall survival.[20]
In HNSCC, lower intra-treatment SUVmax and lower nodal volume on PET/CT has association with improved outcome.[21] Metabolic tumor volume change under PET/CT after first IC cycle has been suggested as an early prognostic indicator of response for following (chemo)radiation.[22] However, the correlation between IC treatment response and following outcomes remains unclear.
In this study we found the retrospective data supporting that TVRR-T has correlation to following curative intent (chemo)radiotherapy. In good responders, 77% of patients had reached complete remission, and also reached a significant superior disease free survival compared to poor responders, while the one dimensional RECIST failed to show correlation to disease control.
As for laryngeal and hypopharyngeal cancer, patients could the benefit from IC to spare morbid surgeries and resulted in satisfactory disease control. Literature focusing on response after IC in HNSCC usually sorted patients into good and poor responders by various measurement, including RECIST or WHO. In some prospective studies, tumor volume reduction rate had taken the role to stratify HNSCC patients.[23],[24] However, there was still no evidence that the response to IC can be used as the indicator of organ preservation[10]
In most larynx preservation prospective trials protocol, for poor responders, defined as stable to progression disease, the patients under go total laryngectomy for an non-inferior comparison to definitive (C)RT. [8, 9, 10, 25, 26] Using our retrospective data, we had collected a curative intent cohort with same protocol after IC for all patients, both good and poor responders alike. Despite locally advanced in general, this cohort resulted in a superior DSF for good responders, reinforcing the rationale that good responders are suitable for organ preservation.
This study also revealed a discordant treatment response between primary tumors and involved lymph nodes in few patients. Some individuals had paradoxical responses to IC, that their primary tumor volume reduced but lymph nodes enlarged. Such finding is coherent with prognostic value that TVRR-T to DFS, while TVRR-N did not. This phenomenon had also been described in some immunotherapy studies, in which tumor shrinkage varied at primary and nodal sites, possibly due to different microenvironment.[27] For poor responders with progression disease, especially in nodal metastases, salvage surgical treatment such as neck lymph node dissection should be considered an option.
We also noticed an overall lower survival in our cohort compared to historical study. In Taiwan, unlike the Western population, patients with non-HPV oropharyngeal cancer still outnumbered the HPV-related counterpart, despite a rising trend in HPV-related group.[28] In our study cohort, up to 92.3% (N=12) were non-HPV-related oropharyngeal cancer, which may result in the poorer outcome, regarding historical study like RTOG 0129, in which study HPV+ and HPV- 3-year locoregional control were 86% and 65%.[29] With majority of patient with smoking (88.1%) and p16-negative (92.3%) risk factors in this cohort, an inferior outcome is understandable.
Other possible cause of inferior disease control in this cohort could be a considerable amount of oral cavity cancer. As a primarily surgical disease, they consisted of 21.1% in our study cohort. These patients reached an overall local control of 21%, 68% never free of disease, and OS of 25%, a much worse treatment outcome compared to PARADIGM trial, in which 18% oral cavity cancer and 45% patients with non-oropharyneal cancer group yielded 3-year PFS 66% [45–80, 95% CI], and OS 73% [52–85, 95% CI][30], non-inferior to oropharyngeal cancer.
Despite the suspicion of oral cavity may contribute to inferior outcome, multivariate Cox regression analysis revealed non-significant hazard to DFS, after adjusting age, IC dose, and oral cavity (versus non oral cavity) cancer as covariates.
In summary, our study showed that TVRR-T after IC is prognostic of DFS in LA-HNSCC patient. As therapeutic and prognostic indicator, combined with future radiomic features, may be a useful reference for clinical judgement.[31] To better justified this treatment response organ preservation paradigm, larger prospective trial is need. Future prospective study design concerning IC could also consider volumetric measurement for treatment response.
The retrospective nature of this study only select patients who are suitable for IC and resulted in a group of patients with a variety of different outcome and treatment modality. This unavoidable heterogeneity may have undermined the survival analysis.
We have chosen docetaxel to represent the intensity of IC dosage in multivariate Cox regression analysis, as in TPF regimen, cisplatin and docetaxel was usually given in 1:1 ratio.
In DFS analysis, we excluded patients who failed to reach initial complete remission, although DFS had remained significantly better for good responders even if never disease free were considered. We believe an initial disease free have better representation for the outcome of definitive treatment, while those with residual tumor could have consideration for surgery.
There were considerable amount of patients who failed to failed to complete definitive (C)RT courses, thirteen patients (n=13/90, 14%), in which eight of them (n=8/13, 61.5%) also failed to reach initial complete remission. After exclude the patients who failed to reach initial complete remission, the completion of RT was still not a significant hazard to DFS in univariate Cox regression analysis.
The dosage of CCRT was not analyzed this retrospective study, despite the there was no significant disease free survival between patients received RT alone and those received CCRT, due to the majority patients received RT alone and a variety cycles was given, possibly a reflection of patient’s performance status. While oncologists usually encourage patient with better performance to receive CCRT, we assumed that IC dose and completion of RT course had sufficed to describe the tolerable intensity of the whole treatment course.
Studies for volumetric measure usually favor MRIs over CT images, for MRIs have better quality to delineate tumor. However, accessibility of CT make CT still be a reliable clinical tool.