Introduction

Nasopharyngeal carcinoma (NPC) is a rare malignancy in most parts of the world, but it is one of the most common types of cancer in Southern China with an annual incidence of 25 to 50 cases per 100,000 persons [1]. Radical radiotherapy (RT) is the main treatment modality for NPC. Over the past 30 years, advances in the RT technique, imaging technology, and chemotherapy have led to a marked improvement in the survival of NPC patients. The preliminary reports of NPC treated with intensity-modulated RT (IMRT) showed excellent results with 3-year overall survival (OS) rates of 83.1% to 94.3% [2–4].

Ionizing radiation, a well-known carcinogenic agent, can induce malignancy [5,6]. With improved survival, the increasing incidences of second malignant neoplasms (SMNs) attributable to RT have become a major concern among long-term survivors. The incidence of SMN after NPC is substantially lesser than those of other head and neck tumors such as oral and laryngeal cancer [7–9]; this can be attributed to the unique predisposing factors of NPC. Because of the rarity of radiation-related SMN in NPC, few studies have addressed the therapeutic management and outcome of this disease. Moreover, the studies that have been performed on this topic are mainly on small series of patients or include a mix of different primary tumor entities, making it difficult to evaluate the optimal treatment strategy and prognostic factors [10–17]. In this article, we report on the analyses of the clinicopathological characteristics, treatment outcomes, and potential prognostic factors of SMN in a large group of NPC cases from endemic regions.

Materials and Methods

Results

Follow-up and Outcome

The median follow-up period for all patients was 26.9 months (range, 0.7–395.7 months). Of the 235 patients who underwent treatment, 63 patients (26.8%) developed recurrent disease (56 local failure and 7 regional failure), and 16 patients (6.8%) developed distant metastasis (7 lung, 7 bone, and 2 liver).The interval between curative treatment and the presentation of recurrence of the disease ranged from 1.8 to 208.0 months, with a median of 10.4 months.

Six patients (2.6%) developed the third primary neoplasms in head and neck. The median interval between the diagnosis of SMN and development of the third neoplasm was 4.7 years (range, 0.8–29.7 years). The anatomical sites of the third tumor were also within the previously irradiated fields i.e., the base of the tongue, neck, soft palate, hard palate, mandible, and maxillary sinus. The histologic types of the third tumor were squamous cell carcinoma in 3 patients, and basal cell carcinoma, osteosarcoma, and leiomyosarcoma in the remaining 3 patients.

At the time of this analysis, 7 patients (3.0%) had been lost to follow-up, 186 patients (79.1%) had died, and 42 patients (17.9%) were still alive, including 28 patients without disease and 14 with disease progression. Median PFS and OS of the 235 patients under treatment were 17.3 months (95% CI, 13.4–21.2 months) and 28.5 months (95% CI, 23.0–34.1 months), respectively. The 5-year PFS and OS rates were 28.2% and 31.9%, respectively (Figure 1).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Figure 1. Overall survival in patients with radiation-related second malignant neoplasm (n = 235).

https://doi.org/10.1371/journal.pone.0084586.g001

Impact of Treatment Modality on Survival

Patients were classified into three groups according to treatment modalities: surgery (151 patients), RT (46 patients, including RT with or without chemotherapy), and chemotherapy (38 patients) groups. Five-year OS rates were 42.9%, 23.7%, and 0% in the surgery, RT, and chemotherapy groups, respectively (P < 0.0001, Figure 2). The median OS was 39.7 months (95% CI, 25.5–53.9 months) in the surgery group, 32.0 months (95% CI, 21.0–42.9 months) in the RT group, and 12.1 months (95% CI, 9.2–14.9 months) in patients who underwent palliative chemotherapy alone.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Figure 2. Overall survival according to treatment modalities of radiation-related second malignant neoplasm (n = 235).

https://doi.org/10.1371/journal.pone.0084586.g002

The survival analysis of the surgery group yielded the following results: survival was significantly better in patients with complete resection than in those with incomplete resection (P < 0.0001); the median OS was 64.1 months (95% CI, 41.3–87.0 months) versus 10.8 months (95% CI, 5.8–15.8 months). For patients who underwent incomplete resection, no significant difference was detected in the survival between patients who underwent surgery alone and those who received combined therapy (P = 0.925).

Discussion

Exposure to ionizing radiation is a known risk factor for cancer. However, not all SMN that arise after RT are induced by irradiation, it is often difficult to differentiate radiation-related SMN from sporadic malignancies. In some types of tumors (lymphomas, breast cancer, etc.), the relative risk of SMN related to RT could be estimated in patients treated by RT versus patients treated by chemotherapy or surgery [6]. Nevertheless, this comparison cannot be performed in NPC patients. Therefore, we used the modified criteria for radiation-related SMN defined by previous reports [18–21].This is a report of the largest series published to date on the clinicopathological features, treatment outcomes, and related prognostic factors of SMN in patients with NPC; this investigation was conducted at a single institution from endemic regions over the past 40 years.

Approximately 10%–20% of patients with head and neck cancer have been reported to develop SMN after definitive therapy [9].The incidence of SMN after RT in patients with NPC ranged from 0.04% to 7% [10–13], significantly less than the high rates associated with other head and neck cancers. In this study, the crude incidence of SMN in NPC was estimated as 0.63%, comparable to the rate of 0.55% reported by Wang et al. [11]. The estimated incidence may be lower than the actual incidence because a proportion of patients were lost to follow-up or had died before the presentation of SMN. Additionally, the follow-up period for patients who received megavoltage X-rays is relatively short.

The median latency of 9.5 years in our series is consistent with the data from other studies, which ranged from 7.6 to 17.0 years [10–18]. Kuttesch et al. [23] reported that the administration of chemotherapy may shorten the latency between RT for the first cancer and subsequent presentation of SMN. This purported relationship between chemotherapy and latency was confirmed by our results. In addition, our results are consistent with previous studies suggesting that there may be a shorter latency interval in patients who receive higher radiation dose. Chen et al. [7] reported that the risk of sarcoma increased significantly after a latency period of some years and continued to increase after 10 years. We also found that nearly half of SMN occurred after a latency of more than 10 years, which indicated that a careful and adequately long-term follow-up protocol should be designed for patients with NPC. Additionally, the long-term follow-up should be maintained after curative therapy for SMN because a third neoplasm may occur over several years.

Amemiya et al. [19] reviewed the records of 1,358 patients treated by RT for head and neck SCC (HNSCC) and found that 25 patients developed SMN, including 18 tongue cancer and 3 maxillary sinus cancer. Scélo et al. [24] reported that tongue cancer significantly increased after radiation of NPC in a pooled analysis of 13 cancer registries (standardized incidence ratio = 11.1). In the study by Teshima et al. [25], the most common sites of SMN after NPC were oral cavity and pharynx. Consistent with previously published data, we noted that oral cancer was the most common type of SMN; particularly, malignancies in the tongue and gingiva accounted for approximately 40% of all SMN. Although a series of organs were exposed to high or intermediate dose of radiation, the incidences of developing SMN at different sites varied significantly. For example, SMN in thyroid, salivary gland, and brain were relatively rare. The most likely reason for this phenomenon is that the sensitivity of different organs to radio-induced carcinogenesis varies to a large extent [5].

The treatment outcome of SMN in head and neck was less satisfactory than that of patients with de novo primary head and neck cancer. The 5-year OS rate of 31.9% in our study was similar to that reported by Jones et al. [26], who found a 5-year OS rate of 31% in patients with second cancer following HNSCC. In a recent report from Taiwan, the 5-year OS rate was 30%, with the median survival of 1.7 years for 712 cases of SMN in NPC patients [7]. Their survival outcome is slightly lower than ours, primarily because the researchers included SMN of various regions, such as liver and lung cancers.

Our results indicated that SMN in the subtype of sarcoma was one of the independent prognostic factors associated with poor survival. Survival of sarcoma in our series approached the lower end of the previously reported wide range (range, 8%–60%) [27–29].The poor survival of patients with sarcoma can be attributed to several factors: pathologically aggressive features; difficulty of performing resection in the head and neck region; high potential for local recurrence and metastasis; insensitivity to chemotherapy and RT. The detailed report of prognostic factors for patients with sarcomas had been presented in the previous paper [30].

Despite the limited details of treatment modalities for SMN in the published reports, the important role of surgery has been documented in many reports [27–29]. In the current study, the 5-year OS rates of 55.5% obtained in the early stage subgroup are encouraging. Multivariate analysis revealed that surgical resection was one of the independent prognostic factors for better survival, which was in agreement with reports from Brady et al. [27] and Cha et al. [28]. Surgery for SMN in the head and neck region is rather challenging because of the difficulty of resection, high risk of surgical complications in a highly irradiated area, and the close proximity of the tumor to critical structures, such as carotid artery and skull base. However, compared with other treatment modalities, complete resection with a negative surgical margin apparently offers the best chance for long-term survival. With the median OS of only 10.8 months, the survival of the incomplete resection group was very poor. The current analysis showed that, in the incomplete resection group, survival after combined therapy was similar to that after surgery alone. However, the limited number of patients underwent incomplete resection did not permit us to reach a definitive conclusion in this regard.

Although surgery is preferable to RT as the treatment modality for SMN, often, resection cannot be performed in a substantial proportion of patients with locally advanced tumors. For unresectable SMN, treatment of chemotherapy alone is a palliative option with no long-term survivors in our series. The use of full-dose reirradiation for SMN is debatable. Theoretically, the tumor in the previously irradiated site could be insensitive or sometimes resistant to RT and chemotherapy, owing to tissue fibrosis after the first curative irradiation. Furthermore, the accumulation of a high radiation dose could result in serious organ injury. Nevertheless, the practice of reirradiation for second prmary head and neck cancer has been reported as feasible in recent years, particularly HNSCC [31–33]. As shown in our study, the survival of patients in the RT group was significantly better than that of patients who underwent chemotherapy alone, particularly in the carcinoma subgroup. In cases of patients with locally advanced and inoperable SMN, reirradiation with or without chemotherapy is the only possibility to offer the promise of long-term survival. Therefore, on the basis of histologic type of SMN, performance status, age, location, latency, tumor bulk, and pre-existing organ dysfunction, repeated RT should be strongly considered for treatment of a select proportion of patients. This being a long-term retrospective study with implying limitations, we did not deeply analyze the late complications of reirradiation in long-term survivors. Therefore, prospective studies with larger samples need to be performed to confirm this finding. Instead of two-dimensional RT technique, patients undergoing reirradiation should be treated with conformal techniques such as IMRT and proton-beam RT to limit the radiation dose to critical structures and to minimize the risk of complications.

In recent years, external RT technologies have evolved significantly, and IMRT has been increasingly used in the treatment of head and neck cancers. There is concern that IMRT might increase SMN risks because of the greater volume of normal tissues exposed to low-level dose of radiation [34]. We have used IMRT for treatment of a series of NPC patients at our institution since February 2001. At the time of this analysis, since only 1 patient of them had developed SMN, the effect of IMRT could not be assessed. Apparently, a longer follow-up and a larger series of patients will be necessary to confirm whether IMRT actually increases the incidence of SMN.