4. Discussion
With the development of treatments, only 10-20% NPC patients would suffer local recurrence after initial treatment29-31. Surgery was one of options for recurrent NPC. In most retrospective studies, surgery has been reported to achieve a similar or better result than re-irradiation with a 5-year LC of 43-74% and a 5-year OS of 47-62%9-13. Moreover, the advantage of surgery was fewer complications and better quality of life. However, it might be due to selected patients. The lesions considered resectable included rT1 disease, rT2-3 with limited parapharyngeal space involvement or disease confined to the base of sphenoid sinus. Others, such as involvement of the internal carotid artery, limited invasion to the clivus, posterior maxillary sinus, pterygoid process and petrous apex, might be resectable, which required the careful judgment by surgeon32. Therefore, there were still many patients not suitable or willing to receive surgical treatment, especially those with late-stage recurrence.
Reirradiation was the most common salvage treatment for recurrent NPC, especially for those unable to receive surgery18. However, previous studies have shown that recurrence usually occurred in high-dose areas with the characteristic of radio-resistance3334. Moreover, when finishing the first course of radiation, changes in the microenvironment such as fibrosis and vascular necrosis might exacerbate radio-resistance35. Reirradiation became a stuff work because of the balance of high dose needed for radio-resistance tumors and dose limited by accumulation of surrounding organs at risks. Previous studies have shown that the recurrent NPC patients with re-irradiation as a salvage treatment, LC and OS at 3 years can reach 70-89% and 46-58%, respectively19-22. However, the side effects of reirradiation remain challenging issues. Virtually most patients with irradiation suffered long-term complications. About 30-70% of patients were likely to develop severe (grade 3-5) complications 202123-25. Moreover, some patients might die of fatal complications, such as necrosis of the temporal lobe necrosis, carotid blowout, teeth occlusion and mucosal ulcer26[23,40,41]. Han and colleagues21reported that the prevalence of advanced toxicity (grade 3-5) of intensity modulated radiotherapy (IMRT)for treatment of recurrent NPC was 70.3%, and that 69 % of patients died of EBRT-related toxicity. Kong and co-workers19reported that 29.3% patients died of radiotherapy-related complications. Of these patients, 23.9% patients died of massive hemorrhage, indicating that massive hemorrhage was the most common cause of death. 75%of patients underwent locally advanced disease. Koutcher and colleagues36 reported an incidence of 73% that grade III or above complications occurred. Teo and co-workers37 reported that the incidence of hearing loss or difficulty in opening mouth was approximately 50-70% after reirradiation.
125I RSI-BT, one of the most common brachytherapy, was often chosen as a salvage treatment for recurrent cancers, such as hypopharyngeal carcinoma38, salivary gland carcinoma39 or other head and neck squamous cell carcinoma40. Because of its sharp dose curve,125I RSI-BT could protect OARs and achieve a higher local dose distribution so as to achieve favorable LC. Many published studies40-46 have reported RSI-BT as a safe and effective treatment for recurrent NPC, with a local control probability at 1- and 3-years of 52-75.2% and 5.3-73%, respectively, and an overall survival probability at 1- and 3-years of 53-84.6% and 6.7-39%, respectively, as well as a accepted toxicities(Table 4).
In this retrospective analysis, LC of RSI-BT as a salvage treatment for recurrent NPC at 1-, 3- and 5-year was 71.3%, 41.9% and 27.9%, respectively and OS at 1-, 3- and 5-year was 57.7%, 23.8% and 11.9%, respectively, which just the same as published studies.
We found that the total times of previous EBRT was a prognostic factor affecting LC (P=0.001) and OS (P=0.012). For patients those received only once EBRT, LC at 1, 3 and 5 years was 93.8%, 58.6% and 58.6%, respectively and OS at 1-, 3- and 5-years was 72.3%, 36.3% and 18.1%, respectively. However, for those received EBRT twice or three times, LC at 1, 3 and 5 years was 42.4%, 25.5% and 0%, respectively and OS at 1- and 3-years was 31.2% and 0%, respectively. The results were probably due to fibrosis, atrophy and necrosis of local tissue, vascular redistribution, and decreased radio-sensitivity after multiple EBRT. Salvage treatment for local recurrent NPC after previous EBRT was a stuff task especially when the patient experienced twice or more courses of EBRT and RSI-BT might be alternative and promising.
We also found that sex was a key factor affecting LC (P=0.037), which was not reported in previous work. It needs to be further investigated to exclude cause of the patient pool’s unbalanced sex ratio males to females and small study cohort.
The use of 3D-PT might improve LC (P=0.078) though the statistically difference was less significant. 3D-PT was creatively designed and introduced into CT-guided RSI technique47. With 3D-PT assistance, RSI-BT may be more accurate and have better doses distribution which close to expected preoperative plan48. It provided a way of RSI-BT to standardization and normalization. An ideal dose distribution may lead to a better LC, but we can only identify the tendency due to our limited case number and still need more data to confirm it.
KPS was a prognostic factor to LC (P=0.033) and might be a prognostic factor to OS (P=0.075) though the statistically difference was less significant, which might have been due to the small study cohort or confounding factors. However, this prognostic factor needed more data to verify.
Furthermore, safety was another key point needed to be paid attention to. Comparing to EBRT, RSI-BT has the advantage of providing a small radius of radiation, high local radiation dose, sharp fall-off of the radiation dose and few radiation effects on adjacent tissues. These features achieve the goal of Precise EBRT and 125I RSI-BT been recommended for treatment of several types of recurrent or relapsed cancer by the National Comprehensive Cancer Network. In our study, only 2 cases (6.5%) suffered severe radiotoxicity: 1 case with of grade 3 skin/mucosal toxicity and another of mandibular osteonecrosis. Besides, eleven patients (41.9%) suffered from late grade 1-2 adverse effects, including ten cases (32.2%) of skin/mucosal toxicities and one case (3.2%) of pain. The prevalence of severe toxic and side effects observed in our study was obviously lower than that of patients receiving reirradiation for recurrent NPC in other studies, and further demonstrated the safety of 125I RSI-BT for treatment of recurrent NPC.