Preliminary results of a randomized study on therapeutic gain by concurrent chemotherapy for regionally-advanced nasopharyngeal carcinoma: NPC-9901 Trial by the Hong Kong Nasopharyngeal Cancer Study Group

Reviewed by: Stephen Z. Sack, MD PhD
The Abramson Cancer Center of the University of Pennsylvania
Last Modified: January 29, 2006

Authors: Lee AW, Lau WH, Tung SY, Chua DT, Chappell R, Xu L, Siu L, Sze WM, Leung TW, Sham JS, Ngan RK, Law SC, Yau TK, Au JS, O'Sullivan B, Pang ES, O SK, Au GK, Lau JT; Hong Kong Nasopharyngeal Cancer Study Group.
Source: J Clin Oncol. 2005 Oct 1;23(28):6966-75.


Nasopharyngeal carcinoma is separated from most other head and neck cancers due to its different clinical presentation and treatment algorithms. In general, there is a higher incidence of nodes, which are often bilateral, there is an earlier onset of distant metastatic disease, and there is a rather limited role for surgery. Chemotherapy and radiation generally form the basis of treatment. These cancers are more common in the Far East, where they have a slightly different histological pattern. In these endemic areas, WHO Grade III histology comprises almost all cases. In the non-endemic areas, it is far more common to find the keratinizing squamous cell variant (WHO Grade I). The difference between these two histologies has been further demonstrated by PCR for Epstein-Barr virus (EBV), which demonstrates 100% expression for Grade III versus only 14% for Grade I (Hording et al. Eur J Cancer B Oral Oncol., 1993). WHO Grade I carcinomas have a worse prognosis, and thus a potentially greater benefit from more aggressive therapy. Considering histology is also important when reviewing clinical studies. Al-Sarraf's study had a major impact on “standard therapy”, but was composed of an abnormally large number of WHO Grade I patients (37%) (Al-Sarraf et al. JCO, 1998). The distinctive nature of nasopharyngeal cancers in the sphere of head and neck cancers is demonstrated by the separate staging system, which allows for larger nodes. When evaluating this Hong Kong concurrent chemoradiation trial, comparisons to Al-Sarraf et al. and Lin (Lin et al. JCO 2003) are warranted. This paper thus serves to confirm earlier findings while focusing on patients in an endemic area with Grade II/III carcinomas of the nasopharynx.


  • Al-Sarraf et al. JCO, 1998 (aka IGS 0099) compared concurrent chemoradiotherapy (CTRT) vs. radiation (RT) alone for stage III-IVB disease, and was the first trial to demonstrate a significant benefit in overall survival (OS).
  • Lin et al. JCO 2003 confirmed these results with a smaller OS benefit using Cisplatin 80 mg/m 2 with Fluorouracil. 5-yr OS improvement was statistically significant (54% with RT alone versus 72% with CTRT).
  • Lee et al. points out several concerns regarding Al-Sarraf's paper, including:
    • RT alone arm underperformed, with 5-yr OS of only 37% (Lin et al. showed 5-yr OS was 54% in the CTRT arm)
    • High percentage of Grade I patients (37%), thus questionable significance to endemic areas
    • High doses of Cisplatin resulting in poor compliance (63%).
    • Late toxicity data lacking.
  • Al-Sarraf update (PASCO 2001) revealed 5-yr outcome (OS) by histology: WHO Grade I = 37%, Grade II = 55%, and Grade III = 60%.


To confirm the standard treatment recommendation of CTRT for patients with locally advanced nasopharyngeal tumors (excluding WHO Grade I). Thus, to evaluate recommendations for endemic areas with a comparison of treatment toxicities.

Material and Methods

  • 348 eligible T1-4N2-3M0 patients (pts) from Hong Kong or Princess Margaret Hospital in Canada.
  • Eligibility criteria:
    • Biopsy-proven WHO Grade II/III (Grade I or adenocarcinomas excluded)
    • N2/3 disease without “gross evidence of distant mets”
    • Adequate hematologic and renal function for chemotherapy
    • ECOG performance status ≤ 2
    • Age < 70 years
  • Workup:
    • Physical exam
    • Fiberoptic nasopharyngolaryngoscopy (NPL)
    • CT or MRI of nasopharynx
    • Chest x-ray (CXR)
    • Labs (CBC, renal and liver tests, LDH)
  • Radiation:
    • 2 Gy/fraction, 5 days/week, total dose ≥ 66 Gy, with uninvolved neck nodes to ≥ 50 Gy.
    • 2 Dimensional, 3D-conformal or IMRT treatment planning all acceptable
  • Chemotherapy:
    • Cisplatin 100 mg/m 2 IV q3 wks (days 1, 22 and 43)
    • Cisplatin 80 mg/m 2 q4 wks and Fluorouracil 1,000 mg/m 2 /d by 96-hr infusion (days 71, 99 and 127)
  • Follow-up by physical exam and NPL with imaging only as indicated
    • q3 months for first 3 yrs, then every 6 months
  • Median f/u of 2.3 years (range = 0.2 to 5.3 months)
  • Late toxicity defined as occurring > 90 days after completion of RT
  • Endpoints:
    • Failure-free survival (FFS) = time to first failure at any site
    • Progression-free survival (PFS) = time to first failure or death by any cause
    • 2Ú endpoint: Locoregional failure-free rate (LR-FFR) = persistent/recurrent disease in nasopharyngeal or cervical region
    • 2Ú endpoint: Distant failure-free rate (D-FFR) = hematogenous metastasis
    • 2Ú endpoint: OS = death by any cause
    • 2Ú endpoint: Toxicity
  • Statistical Method:
    • With 340 pts, the power was 80% to detect a 15% difference in 5-yr FFS
    • Student's t test used for comparing means of continuous variables
    • Multivariate analyses by Cox proportional hazards regression model to test significance of independent factors.
    • All tests two-sided, and p < 0.5 was considered significant
    • Two planned analyses: one at completion of accrual and another at 5-yr follow-up


  • TABLE 1: Two arms are well-balanced
    • Mostly males (72-79%), ECOG performance status mainly 0-1, ~40% treated with 2D techniques only
    • Treatment completed in a mean of 46 days
  • TABLE 2: Compliance with Chemotherapy
    • 94% with 2 or more cycles of concurrent chemotherapy and 81% with 2 or more cycles during adjuvant phase of chemotherapy
  • TABLE 3: Acute Toxicity
    • Mucositis worse with CTRT, while no differences in skin toxicity
    • Chemotherapy-related toxicities as expected
    • Overall, CTRT arm with 71% Grade 3 and 12% Grade 4 toxicity, while RT arm with 52% and 1% respectively.
  • TABLE 4: Patterns of Failure
    • Local failure decreased on CTRT arm from 10.8% to 4.7%, while nodal failure decreased from 8% to 3.5%
    • Distant failure not significantly affected by treatment arm (22.2% decreased to 19.2%)
    • Relapse-free survival at last follow-up increased in CTRT arm (65.9% increased to 73.3%)
  • FIGURE 1: Failure-free Survival
    • Even at 2.3 years (mean follow-up), there is a significant difference in outcome in favor of CTRT.
  • FIGURE 2: A) LOCOREGIONAL failure-free vs. B) DISTANT failure-free rate
    • Demonstrates a benefit in local control without affecting distant control. (Keep in mind, however, that patients were not routinely assessed with imaging for metastatic disease at time of accrual)
  • FIGURE 3: OS currently demonstrates no difference and we need to await longer follow-up
  • TABLE 5: On subset analysis, only the LR-FFR is statistically significant. In the Stage III patients, FFS and PFS are significant.
  • TABLE 6: Multivariate analysis demonstrating that CTRT improves outcome, as does higher dose (and boost to NP), lower stage, and younger age.
  • TABLE 7 and FIGURE 4: Late toxicity worse in CTRT arm.
  • TABLE 8: Nice comparison of various studies.


This study serves to confirm prior research showing that CTRT is one of the most promising strategies to improve tumor control in advanced stage nasopharyngeal carcinoma. The issue of managing potential benefit against excessive toxicity is also addressed, and the authors argue that patients should be informed of these higher risks. This study is interesting in a number of respects. Firstly, the lack of “total body imaging” for initial work-up, and the absence of post-treatment imaging (unless indicated), likely have little bearing on the results. The lack of initial imaging would probably result in some understaging, but this would be expected to be random and thus fairly balanced between the arms. The post-treatment imaging would then primarily serve to detect local recurrence earlier in the follow-up period, and the authors attempt to compensate for this with a stricter follow-up schedule of every 3 months. Once again, as this is a randomized trial, such issues are likely to be balanced between the arms. The change in the staging system since the prior studies were performed is highlighted by the comparatively higher- risk patients that were included (or, more precisely, the lower risk patients that were excluded). Furthermore, this is one of the few studies to incorporate more modern 3D-conformal radiation or IMRT planning techniques. The lack of centralized review of IMRT planning does raise some questions regarding quality control; however, the randomized nature of this trial would likely balance this oversight.

The efficacy of chemotherapy is rather interesting, in that both Chan et al. (IJROBP, 2005) and this study find no benefit with respect to distant disease control. This may point to an inability of chemotherapy to control distant micrometastatic disease in those patients who do manage to obtain a durable local control. The benefit of concurrent chemotherapy with respect to local control is, however, more pronounced and is in concordance with prior studies.

The main issue with this study is the relatively short follow-up. What is admirable are the efforts made at scoring toxicity. Exactly how the toxicity is measured (e.g. do patients undergo formal hearing tests?) is not mentioned, and does limit these conclusions somewhat. Without a difference in overall survival, these results are not yet mature enough for definitive conclusions with respect to balancing toxicity against outcome. Future updates may help to lend guidance in balancing treatment aggressiveness with morbidity.