Treatment adaptation based on response to induction chemotherapy in nasopharyngeal carcinoma: An evolving landscape

Patients with nasopharyngeal carcinoma (NPC) represent a distinct group with head and neck cancer. They are often nontobacco users, nonalcohol users, and on average are 10 to 20 years younger than patients with cancers of other head and neck sites. Given good baseline health status and the effectiveness of contemporary treatment,^1-3 patients with NPC typically have long projected life expectancies and commonly develop late treatment effects, such as cranial nerve deficits and dysphagia. Previous efforts in reducing radiation-related toxicity included the use of reduced target doses⁴ and volumes.^{3, 5-8}

In this issue of CA: A Cancer Journal for Clinicians, Tang et al. report the results of their multicenter phase 3 trial of 445 patients with locoregionally advanced NPC, in which patients were randomized to receive either reduced-volume radiotherapy based on the postinduction chemotherapy (post-IC) gross tumor volume (GTV) or standard radiotherapy based on the preinduction (pre-IC) chemotherapy GTV.⁹ The primary end point was locoregional relapse-free survival at 3 years, with a noninferiority margin of 8%. Overall survival, distant metastasis-free survival, failure-free survival, adverse events, and quality of life (QoL) were also reported as secondary end points. The study is well designed, has a large patient cohort, and provides high-quality data exploring this essential question. With a median follow-up of 40.4 months, patients in the post-IC arm had noninferiority in locoregional relapse-free and overall survival as well as lower toxicities and improved QoL compared with patients in the pre-IC arm. This study has important implications for the future of tailored radiotherapy in NPC. Long-term follow-up, however, is necessary to confirm the findings.

The findings of Tang et al. shared similarities with those of another recently published randomized trial.⁸ In that multicenter trial of 212 patients with stage III–VB, locally advanced NPC, the authors demonstrated that treating the post-IC GTV resulted in noninferior locoregional relapse compared with treating the pre-IC GTV, with potentially improved QoL and less late toxicity. Different chemotherapy regimens and schedules were used in the study. Given the results of these two randomized trials demonstrating noninferiority in both locoregional relapse and survival with this de-intensification approach, should the use of the post-IC GTV for intensity-modulated radiotherapy planning be adopted universally? Before we make a conclusion, let us first examine some fundamental questions in NPC treatment.

First, does chemosensitivity equate with radiosensitivity? In these studies, the determination for radiosensitivity was based on chemosensitivity. It is important to recognize that chemosensitivity does not necessarily correlate with radiosensitivity. The molecular mechanisms of chemoresistance and radioresistance differ significantly.¹⁰ Tumors that have a radiographic response to induction chemotherapy may harbor radioresistant clones that ultimately are not included in the GTV. It should also be noted that tumors exhibit heterogeneity and that heterogeneity exists within the tumor of an individual patient and between tumors in different patients. Developing a treatment paradigm based on comprehensive profiling of a patient's clinical plasma Epstein–Barr virus DNA, omics, and imaging features would allow for treatment stratification based on the risk of recurrence. Continued advancements in artificial intelligence may one day replace our current one-size-fits-all approach with a precise and robust risk-stratification strategy that may select candidates more safely for radiation de-intensification.

Second, what constitutes a treatment response? Are anatomic imaging techniques, such as computed tomography (CT) and magnetic resonance imaging (MRI), the best modality to assess treatment response? The authors in the current study defined the nasopharynx GTV in the post-IC arm by the extent of soft tissue involvement on post-IC MRI and the extent of bone involvement on pre-IC MRI. In the pre-IC arm, the nasopharyngeal GTV was defined by the entire extent of pre-IC disease involvement. Positron emission tomography imaging, which distinguishes treated tumor from biologically active tumor in the soft tissues and bones, was not routinely used to define treatment response. Given the imaging methods for assessing response, it is possible that the GTV was delineated generously in the post-IC group, perhaps resulting in overtreatment. The reductions of the GTVs in this study were modest, approximately 10.2 and 4.5 mL for the primary and nodal sites, respectively. The reduction in doses to the organs-at-risk was also modest. For example, the dose difference in the pre-IC and post-IC arms was approximately 3.5Gy for the mean parotid gland dose and 2.4Gy for the maximum dose (Dmax) to the temporal lobe. Standard use of positron emission tomography imaging along CT and MRI both before treatment and after induction chemotherapy may allow for greater integration of volume de-escalation into clinical practice. In the current study, the assessment of chemotherapy response was performed by a central imaging review committee at high-volume hospitals. Can that level of expertise be extrapolated to low-volume facilities? Delineation of tumor targets and assessment of treatment response in NPC require experience and in-depth understanding of skull base anatomy because tumors commonly infiltrate the skull base. Low-volume centers may have more challenges with treatment planning than high-volume centers.

Third, is there a good surgical salvage option for recurrent NPC after chemoradiation failures? There has been a push toward de-intensification of radiation treatment in the field of head and neck oncology, particularly in the treatment of oropharyngeal cancer. Unlike oropharyngeal cancer, which has a high salvage success with surgery for persistent or recurrent disease after chemoradiation, skull base failure in NPC cannot be salvaged with surgery. Re-irradiation of the nasopharynx is associated with significant long-term toxicities, including soft tissue necrosis, osteoradionecrosis, temporal lobe injury, cranial nerve deficits, and trismus. Any treatment de-intensification for NPC should be approached with caution.

Finally, what is the pattern of relapse for NPC? Most local recurrences occur in the GTV,¹¹ lending pause to the question of GTV reduction in a site where radiation therapy accounts for the bulk of disease control. Should the clinical target volume (CTV), which encompasses potential microscopic disease and harbors significantly less tumor burden than the GTV, receive higher priority when designing de-intensified trials for NPC? There is wide variation among radiation oncologists in defining the primary CTV for NPC. Current consensus and guidelines continue to rely heavily on bony landmarks and fixed geometric margins around the GTV for CTV delineation, an approach that was used to define field borders in the conventional two-dimensional and three-dimensional radiation therapy era. Sanford et al. reported outcomes of individualizing primary CTVs based on stepwise patterns of tumor spread in 73 patients at Massachusetts General Hospital.³ Remarkably, there was a reduction of 90 mL in the CTV for both a left-sided T1N0 tumor and a bilateral T4 tumor compared with the national guidelines approach. There was also a significant decrease in doses to most organs at risk. For example, there was 50% reduction in the Dmax of the right optic nerve for the early stage case and 46% reduction in the Dmax of the optic chiasm for the locally advanced case. With this knowledge-based approach, there was no tumor relapse in the CTV after a median follow-up of 90 months.

Reducing the toxicity associated with treatment should be a constant endeavor in our field. The significant adverse effects associated with radiation to the head and neck render this effort even more crucial. Tang et al. have sought to address this issue, providing high-quality phase 3 data that demonstrate promising results in their exploration of reduced GTV radiation in locoregionally advanced NPC. Integration of their approach outside clinical trial setting and in low-volume centers, however, should be exercised with caution.

To maximize trial success, designing clinical trials that minimize risk and maximize benefit is pivotal. More efforts are needed on clinical trials that focus on de-intensifying primary CTV in NPC given the potential for significant reduction in toxicity with minimal risk of local relapse.

The authors disclosed no conflicts of interest.