2284 Prognostic Significance of DNA Flow Cytometric Analysis in Patients with Nasopharyngeal Carcinoma Timothy T. C. Yip, Ph.D.1 W. H. Lau, M.B.B.S.1 John K. C. Chan, M.B.B.S.2 Roger K. C. Ngan, M.B.B.S.1 Y. F. Poon, M.B.B.S.1 C. W. Lung, B.Sc.2 T. Y. Lo, B.Sc.1 John H. C. Ho, M.B.B.S., M.D., BACKGROUND. Nasopharyngeal carcinoma (NPC) is a prevalent malignant tumor 3 D.Sc. 1 Department of Clinical Oncology, Queen Elizabeth Hospital, Kowloon, Hong Kong. 2 Department of Pathology, Queen Elizabeth Hospital, Kowloon, Hong Kong. 3 Department of Radiotherapy, Baptist Hospital, Kowloon Tong, Hong Kong. Presented in part in the 17th International Congress of the International Society of Analytical Cytology, Lake Placid, New York, October 16 –21, 1994. Supported by grants provided by the Queen Elizabeth Hospital Cancer Research Fund, Hong Kong Anti-Cancer Society, and Hong Kong Jockey Club. Address for reprints: Timothy T. C. Yip, Ph.D., Radiobiology Unit, Department of Clinical Oncology, Queen Elizabeth Hospital, Room 1305, Block R, 30, Gascoigne Road, Kowloon, Hong Kong. Received September 8, 1997; revision received April 27, 1998; accepted April 27, 1998. © 1998 American Cancer Society among Southern Chinese. Previously, the authors described the prognostic significance of a serum antibody assay to a recombinant Epstein-Barr virus Bam HI–Z replication activator protein (ZEBRA) in NPC patients with long term follow-up. In this study, the authors further reported the use of DNA flow cytometry (DNA-FCM) as an additional technique for determining the prognosis of NPC patients in the same series. METHODS. One hundred and forty-three archival biopsies from 110 NPC patients were deparaffinized and subjected to DNA-FCM analysis. DNA ploidy state and various proliferative indices (PI) of the tumors were correlated with patient survival and frequency of recurrence. RESULTS. Among the biopsies analyzed, 119 were histologically positive NPC and 24 were negative. Fifty-one tumor biopsies that fulfilled the guideline criteria of the DNA Cytometry Consensus Conference were correlated with the clinical manifestations of the patients. Among them, 43 tumors (84%) were DNA diploid and 8 (16%) were aneuploid. Two PI, S-phase fraction (SPF) and proliferation fraction (PF), appear to be potentially useful prognostic indicators. For example, PF in patients who developed locoregional recurrence (15.1%) and distant recurrence (16.4%) after radiation therapy both were significantly higher than PF in patients who were in complete remission (8.2%) (P ⫽ 0.0005 and P ⫽ 0.004, respectively). Significant differences in SPF between patients with distant recurrence (10.6%) and those in remission (5.7%) also was found (P ⫽ 0.005). Using Kaplan-Meier analysis, patients with high PF, high SPF, and aneuploid tumors had significantly poorer 12-year survival rates (35%, 26%, and 28%, respectively) than those patients with low PF, low SPF, and diploid tumors (77%, 67%, and 59%, respectively) (P ⬍ 0.0009, P ⬍ 0.004, and P ⬍ 0.01, respectively). CONCLUSIONS. Determination of tumor PI and DNA ploidy state by DNA-FCM at diagnosis of NPC can be potentially useful in selecting a poor prognostic subgroup of NPC patients. These parameters may enable oncologists to plan for more stringent treatment strategies such as hyperfractionated and accelerated radiation therapy or concomitant chemoradiotherapy for these patients. Cancer 1998;83: 2284 –92. © 1998 American Cancer Society. KEYWORDS: DNA flow cytometry, proliferative fractions, S-phase fraction, DNA ploidy, nasopharyngeal carcinoma, prognostic markers, survival analysis. N asopharyngeal carcinoma (NPC) is a common malignant tumor in Southeast Asia but it is rare in Western countries.1,2 This neoplastic disease is prevalent in two Southern Chinese provinces, Guangdong and Guangxi. In Hong Kong, it is the third most prevalent cancer among males.3 The incidence of disease stands at 20 –27.5 per 100,000 population, which is much higher than that of ⬍ 1 per 100,000 population found among whites in Western countries. Prognostic Role of DNA-FCM in NPC/Yip et al. The diagnosis of NPC relies on endoscopic examination of the nasopharynx of the patients followed by histopathologic examination of nasopharyngeal biopsy.4 Serological tests for IgA antibodies to Epstein-Barr virus (EBV) viral capsid antigen (VCA/IgA) and early antigen (EA/IgA) also are useful as first-line diagnostic tests for NPC.5,6 Radiation therapy (RT) has been the most effective modality of treatment.7–10 However, so far, prognosis of NPC, remains a major problem for oncologists. Clinical staging by Ho’s classification has been shown to be invaluable for predicting the outcome of NPC patients after RT.11–13 However, a fraction of the patients with early stage disease still develop either locoregional or distant recurrence and die of disease.13,14 Conversely, a proportion of patients with intermediate or late stage disease are cured by RT. Therefore, additional parameters have been sought to supplement clinical staging for the prognostication of NPC. Recently, we have shown that an immunoglobulin (IgG) to a recombinant EBV Bam HI–Z replication activator protein (ZEBRA/IgG) can serve as a useful prognostic marker for NPC.15–17 The antibody titers tested after RT were elevated in NPC patients who later developed distant metastasis. The antibody also was correlated strongly with the survival rate. However, the survival correlation was found only for the antibody titer determined at 10 months after treatment but not at the time of diagnosis. Thus, the initial treatment cannot be tailored accordingly for the poor prognostic group. DNA flow cytometry (DNA-FCM) involves staining of tumor cell DNA content by nucleic acid binding dyes followed by flow cytometric determination of the DNA ploidy state and proliferative indices (PI) of the 17,18 The technique has been shown to be a tumors. useful tool in the diagnosis and prognosis of a variety of malignant tumors.19 –25 In this study, we examined whether DNA-FCM was able to provide prognostic information for NPC in 143 nasopharyngeal biopsies from 110 NPC patients. MATERIALS AND METHODS NPC Patients One hundred and ten patients with newly diagnosed NPC treated at the Department of Clinical Oncology (formerly called Institute of Radiology and Oncology) at the Queen Elizabeth Hospital, Hong Kong between 1978 –1983 were recruited into this study. They were staged clinically according to the classification system of Ho26 and the International Union Against Cancer 27 (UICC). The patients were followed for ⬎ 12 years or until death. Detailed clinical characteristics of each patient such as disease stage; extent of primary dis- 2285 ease; treatment modalities; time, site, and extent of recurrence; modes of salvage treatment; survival periods; and causes of death all were recorded using Foxplus database software (Microsoft Corporation, Redmond, WA) for analysis. Sample Preparation for DNA Flow Cytometric Analysis One hundred and forty-three archival paraffin blocks of nasopharyngeal biopsies from 110 patients were retrieved. Of these, 58 were undifferentiated carcinomas of the nasopharyngeal type (UCNT), 61 were poorly differentiated squamous cell carcinomas (PDSCC), and 24 were histologically negative biopsies.27–30 All positive and negative biopsies were obtained from NPC patients at the time of first presentation. Histologic diagnosis confirmed no malignancy in the 24 negative biopsies. Repeated sections from these biopsies then were subjected to both hematoxylin and eosin (H & E) staining and in situ hybridization (ISH) for Epstein-Barr virus small RNAs (EBERs) (which have been found to be expressed preponderantly in the NPC cells but rarely in the normal cells31,32) prior to DNA-FCM analysis. These sections were determined by one of us (J.K.C.C.) to be devoid of any malignant cells. Moderately or well differentiated squamous cell carcinomas (MDSCC or WDSCC) were rare among Chinese NPC patients. No patient with such histologic types was entered into this study. Consecutive 5-m and 50-m biopsy sections were cut from each block. The tumor cell ratio was assessed from the 5-m H & E stained section by one of us (J.K.C.C.). This ratio also was judged from a consecutive 5-m section after being subjected to ISH labeling for EBERs. These sections were scanned further by a Vidas microscopic imaging system (Kontron Bildanalyze, Munich, Germany) followed by quantitative analysis in a “Medlab” computer program that had been tailor-made for image quantitation of ISH positive cells in the Electronic Engineering Department of Hong Kong Polytechnic University.33 Two to three 50-m biopsy sections were pooled together and processed into nuclear suspension for DNA-FCM analysis by a modified Hedley’s method.34 Briefly, the thick sections were deparaffinized twice in 8 mL of xylene for 10 minutes. They then were rehydrated in 100%, 95%, 70%, and 50% ethanol and then in water for 10 minutes each time. Nuclear suspension was prepared by digesting the tissues in 5 mL of 0.5% pepsin (Sigma Chemical Company, St. Louis, MO) in citrate buffer (pH 6.5) (Sigma Chemical Company) for 2 hours at 37 °C with constant mixing. The resultant nuclear suspension then was filtered through 45-m nylon mesh to remove cell clumps. It was stained according to the supplier’s instructions overnight us- 2286 CANCER December 1, 1998 / Volume 83 / Number 11 ing a DNA preparation kit (Coulter Corporation, Hialeah, FL) that contained propidium iodide, Triton X-100, and ribonuclease A. guidelines drawn up in the DCCC. Only 51 samples that met such stringent criteria were correlated with the clinical outcome of the patients. Flow Cytometric Analysis Statistical Analysis The DNA stained nuclear preparation was analyzed in a Profile II flow cytometry analyzer (Coulter Corporation) that was equipped with a 15-milliwatt air-cooled argon ion laser operating at 488 nanometers (nm) blue light. To achieve optimal running, the flow cytometer was precalibrated with fluorescent DNA-Check Beads (Coulter Corporation) to obtain a percentage of halfpeak coefficient of variation (%HPCV) of ⬍ 2.0 for either forward angle light scatter and red fluorescence at 635 nm. Chick red blood cells, peripheral blood lymphocytes, and cells derived from two paraffin blocks of reactive lymph nodes served as controls to calibrate the fluorescent channel for the normal DNA diploid peak further. This DNA diploid peak also was assessed from the normal cells contained in the histologically negative or NPC biopsies. A DNA aneuploid population is defined as a DNA peak that differs from the DNA diploid peak by at least 10% in DNA content (i.e., with a DNA index of ⱕ 0.9 or ⱖ 1.1). Cellular debris were excluded from analysis by gating in a two-dimensional (2-D) plot of forward angle light scatter against a 90° light side scatter. A large proportion of the cell clumps were excluded by a doublet gate in a 2-D plot of integral and peak fluorescent signals. DNA histograms of the tumor samples without debris gating were assessed using Multicycle DNA-FCM analysis software (Phoenix Flow Systems, San Diego, CA) to determine the S-phase fraction (SPF), G2/M-phase fraction (G2/MF), proliferative fraction (PF) (which is the sum of SPF and G2/MF), cellular debris level, and %HPCV. Guideline criteria drawn up in the DNA Cytometry Consensus Conference (DCCC)35 for tumor paraffin samples were adopted in the data analysis; samples were excluded from correlation for prognosis if they contained ⬍ 20% tumor cells, if the DNA histogram had a %HPCV ⬎ 8.0%, or if cell count was ⬍ 10,000. Of the 119 histologically positive biopsies analyzed from 110 patients, 9 patients required repeated biopsies for confirmation of diagnosis. Although later confirmed to be histologically positive by reviewing the H & E stained slides, five of the nine initial biopsies from these patients had an insufficient tumor cell ratio for DNA analysis, one had insufficient total cell counts, one had unacceptably high %HPCV, and two had an unacceptably high content of cellular debris, most likely due to severe necrosis in the tumors. These 9 biopsies in conjunction with another 59 from the remaining 110 biopsies (i.e., a total of 68 biopsies) were excluded from further analysis according to the PI in the two histologic types and in different clinical stages and sizes of metastatic lymph nodes (LNs) of NPC patients were compared by the Mann-Whitney U test36 whereas the DNA ploidy state in these groups were compared by the Chi-Square (2) test.39 The PI between tumor positive and negative biopsies and between each of the three clinical groups (the distant metastasis [MET] group, the locoregional recurrence [LR-RC] group, and the complete remission [REM] group) also were compared by the Mann-Whitney U test. Survival analysis in patients with different DNA ploidy states or PI was performed using the KaplanMeier method38 and significant differences between these groups were analyzed by the log rank test.39 RESULTS DNA-FCM parameters (namely PF, SPF, G2/MF, and DNA ploidy state) in 51 histologically positive biopsies that met the stringent criteria of the DCCC37 were correlated with tumor histologic types, patients’ clinical stages, and sizes of the NPC metastatic LNs in the neck (Table 1). No significant difference was found in these parameters between the two histologic types (UCNT and PDSCC), Ho’s or UICC stages, and size of the metastatic LNs. Only DNA aneuploid tumors had a significantly higher SPF than the diploid tumors (P ⫽ 0.01). Although there was a gradual increase of DNA aneuploidy frequency from UICC Stage I-IV, this relation was not statistically significant. The three PI in NPC biopsies were compared with those in the histologically negative biopsies. The PF, SPF, and G2/MF of NPC biopsies (mean values: 12.4%, 7.8%, and 4.6%, respectively) were all significantly higher than those of the negative biopsies (mean values: 5.8%, 3.7%, and 2.1%, respectively) (P ⬍ 0.000004, 0.0003, and 0.004, respectively) (Fig. 1). However, the PI of NPC biopsies showed a wide scattering, with values overlapping with those of negative biopsies. The PI then were correlated with the clinical outcome by categorizing the patients into three groups. Groups 1 (MET) and 2 (LR-RC) were comprised of patients who developed distant metastases and locoregional recurrence, respectively, after RT, whereas Group 3 (REM) was comprised of those patients who were in complete remission after treatment. Patients in the REM group generally had low tumor PI (8.2%, 5.7%, and 2.7% for mean PF, SPF, and G2/MF, respectively) (Fig. 2). In contrast, the patients in the 2 recurrence groups (MET and LR-RC) all had Prognostic Role of DNA-FCM in NPC/Yip et al. 2287 TABLE 1 Correlation of Proliferative Indices and DNA Ploidy States with Certain Clinical Parameters in NPC Patients Clinical parameters Histology UCNT PDSCC Ho’s stage I II III IV UICC stage I II III IV Metastatic LN size (cm) 0–6 7–9 ⬎9 DNA ploidy DP AN No. of patients Mean PF (%) ⴞ SD Mean SPF (%) ⴞ SD Mean G2/MF (%) ⴞ SD DNA ploidy ratio (AN/DP) 23a 24a 12.1 ⫾ 6.7 13.2 ⫾ 6.0 7.0 ⫾ 4.3 9.0 ⫾ 5.9 5.2 ⫾ 5.2 4.3 ⫾ 2.6 3/20 5/19 8 11 18 14 10.4 10.4 14.3 12.7 5.3 ⫾ 1.4 6.7 ⫾ 4.2 10.2 ⫾ 6.9 7.0 ⫾ 2.9 5.1 3.7 4.2 5.6 ⫾ 3.1 ⫾ 3.5 ⫾ 2.7 ⫾ 5.9 1/7 0/11 4/14 3/11 8 0 15 28 10.4 ⫾ 4.1 0 13.8 ⫾ 7.0 12.3 ⫾ 6.3 5.3 ⫾ 1.4 0 9.8 ⫾ 7.0 7.5 ⫾ 4.2 5.1 ⫾ 3.1 0 4.0 ⫾ 4.3 4.8 ⫾ 4.1 1/7 0 2/13 5/23 35 8 8 12.1 ⫾ 5.4 12.6 ⫾ 7.6 13.3 ⫾ 8.5 7.7 ⫾ 4.9 9.2 ⫾ 7.5 7.2 ⫾ 2.5 4.5 ⫾ 3.5 3.4 ⫾ 3.6 6.4 ⫾ 6.1 5/30 2/6 1/7 43 8 11.5 ⫾ 5.6 17.1 ⫾ 7.5 6.7b ⫾ 3.4 13.8b ⫾ 8.2 (bDP/AN P ⫽ 0.01) 4.8 ⫾ 4.2 3.3 ⫾ 2.8 / / ⫾ 4.1 ⫾ 4.5 ⫾ 6.5 ⫾ 7.5 NPC: nasopharyngeal carcinoma; PF: proliferative fractions; SD: standard deviation; SPF: S-phase fraction; G2/MF: G2/M-phase fraction; AN: aneuploid; DP: diploid; UCNT: undifferentiated carcinoma of the nasopharyngeal type; PDSCC: poorly differentiated squamous cell carcinoma; UICC: International Union Against Cancer; LN: lymph node. Proliferative fractions, S-phase fraction, and G2/M-phase fraction in different clinical groups were compared using the Mann–Whitney U test. a Of 51 samples analyzed, 23 were undifferentiated carcinoma of the nasopharyngeal type, 24 were poorly differentiated squamous cell carcinoma, and the remaining 4 cases were reported as anaplastic carcinomas (2 cases) and nonkeratinized carcinomas (2 cases); these 4 cases were excluded in the statistical analysis because of their small number. b No statistically significant difference was found in any of the groups with the exception of the S-phase fraction of DNA diploid and aneuploid tumors (P ⫽ 0.01). DNA ploidy states in different clinical groups were compared using the chi-square test. No statistically significant difference was found in any of the groups analyzed. significantly higher PF (16.4% and 15.1%, respectively) than those in the REM group (P ⫽ 0.0005 and 0.004, respectively). A higher tumor proliferation rate in the two recurrence groups also was demonstrated by the higher SPF and G2/MF, although a significant difference was found only between the SPF in the MET and REM groups (P ⫽ 0.005). Setting a cutoff value of ⬎ 11% PF for rapidly proliferating tumors, 78% (18 of 23) of the tumors in the MET and LR-RC groups were proliferating rapidly whereas only 26.3% (5 of 19) of the REM group were fast-growing tumors (Table 2). If the cutoff value was set at 7% SPF, 61% of the patients in the 2 recurrence groups had rapidly proliferating tumors compared with 26.3% in the REM group. Furthermore, patients with recurrence also had a higher frequency of DNA aneuploid tumors (37.5% and 27%, respectively) than those with complete remission (5.2%) (Fig. 3). All three PI (PF, SPF, and G2/MF) were correlated strongly with the survival rate of NPC patients. This was demonstrated by the significantly lower 12-year survival rates in patients with high PF, SPF, and G2/MF (35%, 26%, and 34%, respectively) compared with those with low PF, SPF, and G2/MF (77%, 67%, and 65%, respectively) (p ⬍ 0.0009, ⬍ 0.004, and 0.004, respectively, by the log rank test) (Fig. 4). Furthermore, patients with DNA aneuploid tumors also had a significantly lower 12-year survival rate (28%) than patients with DNA diploid tumors (59%) (P ⬍ 0.01) (Fig. 4). The results of this retrospective analysis demonstrated the importance of tumor PI and ploidy states in determining the outcome of NPC patients after treatment. DISCUSSION Previous studies have shown the applicability of the DNA-FCM technique in the diagnosis and prognosis of tumors.17,18 Various parameters such as DNA ploidy states and PI (such as SPF and PF) have been found to correlate with the survival and recurrence rate for various tumors including carcinomas of the breast, colon, rectum, lung, bladder, ovary, cervix, and the head and neck.19 –25 In the current study, the prognostic significance of DNA-FCM parameters was demonstrated in NPC patients using archival biopsy materials. The recurrence groups more frequently had DNA aneuploid tumors or rapidly proliferating tumors compared with the REM group. Furthermore, using 2288 CANCER December 1, 1998 / Volume 83 / Number 11 TABLE 2 Distribution of Fast Proliferating Tumors in Patients Who Developed Distant Metastasis or Locoregional Recurrence or Who Remained in Complete Remission after Radiation Treatment No. of fast proliferating tumorsa/total no. of tumors analyzed (%) FIGURE 1. Comparison of three proliferative indices (proliferative fractions Category of patients PF > 11% SPF > 7% Locoregional recurrence Distant metastasis All recurrences Remission 6/8 (75%) 12/15 (80%) 18/23 (78%) 5/19 (26.3%) 4/8 (50%) 10/15 (66.7%) 14/23 (61%) 5/19 (26.3%) PF: proliferative fractions; SPF: S-phase fraction. a Fast proliferating tumors were defined arbitrarily as tumors with proliferative fractions ⬎ 11% or S-phase fraction ⬎ 7%. [PF], S-phase fraction [SPF], and G2/M-phase fraction [G2/MF]) in nasopharyngeal carcinoma and histologically negative biopsies. Statistical differences in the indices of the two groups of biopsies were analyzed by the MannWhitney U test. TUM: tumor biopsies; NEG: histologically negative biopsies; SD: standard deviation. FIGURE 2. Comparison of three proliferative indices (proliferative fractions [PF], S-phase fraction [SPF], and G2/M-phase fraction [G2/MF]) in biopsies from nasopharyngeal carcinoma patients who developed distant metastasis (MET) or locoregional recurrence (LR-RC) or who remained in complete remission (REM) after radiation treatment. Statistical differences in the indices of the different groups of biopsies were analyzed by the Mann-Whitney U test. SD: standard deviation. Kaplan-Meier analysis, DNA ploidy state, PF, SPF, and G2/MF all were significantly correlated with the survival rate. Advantages of DNA-FCM for Prognostication We previously reported the prognostic role of serum ZEBRA/IgG antibody in NPC patients.15,16 Titers of this antibody frequently were elevated in patients who developed metastases to the lung and liver 1– 6 months before the clinical diagnosis of distant spread FIGURE 3. Distribution of DNA aneuploid (AN) and diploid (DP) tumors in nasopharyngeal carcinoma patients who developed distant metastasis (MET) or locoregional recurrence (LR-RC) or who remained in complete remission (REM) after radiation treatment. of disease. Posttreatment antibody titers also were correlated significantly with patient survival. However, the titers frequently were not elevated in patients who developed bone metastasis or locoregional recurrence, and only the titers tested at 10 months after RT but not at the time of diagnosis were correlated with survival. In contrast, the DNA-FCM parameters could be more useful clinically because they were correlated with patient survival at the time of diagnosis and the PI also were elevated in a significant proportion of Prognostic Role of DNA-FCM in NPC/Yip et al. FIGURE 4. Survival analysis of nasopharyngeal carcinoma patients in two ploidy states and in high and low proliferative indices (proliferative fractions [PF], S-phase fraction [SPF], and G2/M-phase fraction [G2/MF]) using the Kaplan-Meier method. The statistical difference was analyzed by the log rank test. AN: DNA aneuploid tumor; DP: DNA diploid tumor. patients with either locoregional or distant recurrence, including some bone recurrences. Currently, the most effective treatment modality for NPC is RT, which conventionally is performed by delivering 30 –35 daily fractions to the tumor at approximately 2–2.5 grays (Gy) per fraction for 5 days per week in an overall treatment time of approximately 6 –7 weeks.7–10 Although effective, some patients still develop recurrences even after a full course of such treatment. If the current results can be confirmed by large-scale studies, DNA-FCM analysis could help to select a group of patients with poor prognosis. Intensification of therapy or adjuvant treatment schemes planned for this group hopefully may reduce the recurrence rate and improve survival. Intensified therapy can be achieved by a continuous hyperfractionated and accelerated radiotherapy (CHART) scheme, which aims to complete the whole course of RT within 12 consecutive days (rather than 6 –7 weeks as in conventional RT) by delivering 3 fractions of 1.5 Gy per day,40,41 or an European Organization for Research and Treatment of Cancer (EORTC) hyperfractionated and accelerated radiotherapy scheme, which gives 3 daily fractionated doses of 1.6 Gy each for the first week, followed by a 1-week rest gap and a further 2 weeks of similar treat- 2289 ment.42– 44 Both schemes intend to shorten the overall treatment time to combat the accelerated proliferation rate of tumors during the treatment period.45 Previous analysis has shown that the EORTC accelerated scheme could result in ⬎ 60% of local control in head and neck tumors with a rapid tumor growth rate compared with the conventional scheme, which resulted in ⬍ 30% local control.44 The tumor control rates for the two RT schemes in patients with slowgrowing tumors do not differ from each other significantly. A recent report also confirmed that the accelerated fractionation arm resulted in a 13% gain in locoregional control over the conventional fractionation arm.44 These accelerated schemes potentially can be useful for NPC patients with rapid tumor growth rate. However, they are costly, labor-intensive, and require clinical and allied health staff continuously working overtime. Thus it is more cost-effective to reserve these treatment schemes for the poor prognostic group. Furthermore, both acute and late toxicity are more severe in patients treated by these stringent treatment schemes. A selection using the DNA-FCM technique perhaps could spare the good prognostic group with slowly proliferating tumors from unnecessary side effects induced by these schemes. Furthermore, modification of the protocols in the future such as reducing the 3 daily fractions to 2 with an even longer interfractional period than the 6 hours currently used plus adding booster doses in small irradiation fields also may help to reduce the toxicity. Despite earlier nonrandomized trials showing a higher complete response rate and/or improved survival in NPC patients with adjuvant chemotherapy using platinum regimens compared with historic controls,46 – 49 subsequent prospective randomized trials50,51 (except one52) did not show a survival benefit. Recent concomitant chemoradiotherapy trials also have shown some benefit in improving survival in patients with advanced NPC.53,54 If the benefits can be confirmed in the future by larger randomized trials, the group selected by DNA-FCM to have a poor prognosis also can be given this treatment in the hope of reducing the frequency of distant recurrence and improving survival. However, further study regarding how to fit the chemotherapy within the radiotherapy scheme to prevent unacceptable toxicity and the issue of adding agents to reduce the specific types of toxicity that inevitably occur with such a combination also should be considered. DNA-FCM and Clinical Stages DNA-FCM and the prior ZEBRA/IgG antibody studies were performed in the same series of NPC patients. 2290 CANCER December 1, 1998 / Volume 83 / Number 11 ZEBRA/IgG appeared to be tumor load-associated because the antibody titers were correlated with clinical stage and the size of the metastatic cervical LNs.1 NPC cells contained abundant EBV. EBV load and, subsequently, ZEBRA/IgG antibody response gradually could accumulate as the tumor mass of NPC increases with stage or size of the metastatic LNs. However, PI did not correlate significantly with either one of these clinical parameters (Table 1). The tumor proliferation rate usually starts off exponentially in the initial phase of tumor development but gradually slows down as the size of the primary tumor reaches a substantially high level.55 This can be due to overgrowth of the tumor cells over the surrounding blood vessels, impairing the diffusion of nutrients and oxygen.56,57 Comparison with Previous Studies Flow cytometric analysis previously has been performed in Australian NPC patients.58 Unfortunately, various PI were not reported. The investigators did not find any significant difference in 5-year survival between the patients with DNA diploid and aneuploid tumors. These results differ from ours, most likely due to the difference in the histology of the tumors analyzed. In our study, all 51 Chinese NPC patients analyzed had either UCNT (n ⫽ 25) or PDSCC (n ⫽ 26). Among the 55 Australian patients studied by Costello et al.,58 65% had keratinizing squamous cell carcinoma (SCC) and only 35% had either UCNT or nonkeratinizing carcinoma (NKC) (NKC being equivalent to PDSCC). SCC is substantially different from UCNT or PDSCC (NKC) biologically. The latter two types usually are more radio-sensitive and hence generally have better prognosis after RT.59 In addition, the latter more frequently express EBV DNA, RNA, and a variety of viral latent antigens and in a greater quantity as well.60 In flow cytometric analyses of various head and neck tumors, PDSCC generally has a higher DNA aneuploidy rate than SCC (WDSCC or MDSCC).61– 63 Although survival analysis between the two ploidy types also was performed in individual UCNT (i.e., lymphoepithelioma), NKC, and SCC in the Australian study, the number of patients after further stratification was too small for statistical significance to be achieved. Future Approaches Despite the prognostic significance of the DNA-FCM study, it has certain drawbacks. Among 119 tumor biopsies investigated in the current study, ⬎ 50% had to be excluded from the prognostic analysis according to the guideline criteria set in the DCCC,35 such as exceptionally high %HPCV, scanty number of tumor cells, or unacceptably high cellular debris. These criteria can improve the accuracy and consistency of DNA-FCM study substantially. In this study, the major reason for excluding the samples was insufficient tumor cells. The greatest problem with the study of NPC is the frequent presence of large numbers of intratumoral and peritumoral lymphoid cells resulting in a low tumor cell ratio in the biopsy samples. This problem potentially can be solved by enriching the carcinoma cells using anticytokeratin antibody or EBER probes,32,33 which can enable more samples to be analyzed and enhance the accuracy of DNA-FCM analysis. We currently are embarking on a prospective study to increase the sample size to ⬎ 800 NPC patients so that the prognostic role of DNA-FCM parameters in determining individual clinical stage, especially the early disease stages, can be analyzed further. Multivariate analysis also can be performed in such a large study to compare the prognostic significance of these parameters with other potentially useful laboratory prognostic markers. It is anticipated that in the near future a number of laboratory prognostic indicators could be combined for the prognosis of NPC to enable one marker to supplement the shortcomings of the other. REFERENCES 1. Ho JHC. Nasopharyngeal carcinoma. Adv Cancer Res 1972; 15:57–92. 2. Huang DP. Epidemiology and aetiology. In: van Hasselt CA, Gibb AG, editors. Nasopharyngeal carcinoma. Hong Kong: Chinese University Press, 1991:23–35. 3. Hong Kong Cancer Registry. Cancer incidence and mortality in Hong Kong 1992. Hong Kong: Hospital Authority Hong Kong, 1996:12. 4. Wei WI, Sham JS, Zong YS, Choy D, Ng MH. The efficacy of fiberoptic endoscopic examination and biopsy in the detection of early nasopharyngeal carcinoma. Cancer 1991;67: 3127–30. 5. Henle G, Henle W. Epstein-Barr virus specific IgA serum antibody as an outstanding feature of nasopharyngeal carcinoma. Int J Cancer 1976;17:1–7. 6. Ho JHC, Ng MH, Kwan HC, Chau JCW. Epstein-Barr virus specific IgA and IgG serum antibodies in nasopharyngeal carcinoma. Br J Cancer 1976;34:655–9. 7. Ho JHC, Lau WH, Fong M. Treatment of nasopharyngeal carcinoma (NPC). In: Grundmann E, Krueger GRF, Ablashi DV, editors. Nasopharyngeal carcinoma, cancer campaign. Volume 5. New York: Gustav Fischer Verlag, 1981:279 – 85. 8. Ho JHC. Nasopharynx. In: Halnan KE, Boak JL, Crowther D, von Essen CE, Orr JS, Peckham MJ, editors. Treatment of cancer. London: Chapman and Hall, 1982:249 – 67. 9. Mesic JB, Fletcher GH, Geopfert H. Megavoltage irradiation of epithelial tumors of the nasopharynx. Int J Radiat Oncol Biol Phys 1981;7:447–53. 10. Huang SC, Chu GL. Nasopharyngeal cancer: study II. Int J Radiat Oncol Biol Phys 1981;7:713– 6. Prognostic Role of DNA-FCM in NPC/Yip et al. 11. Ho JHC. Stage classification of nasopharyngeal carcinoma: a review. In: De-The G, Ito Y, editors. Nasopharyngeal carcinoma: etiology and control. Lyon: International Agency for Research on Cancer Science Publication, 1978:99 –114. 12. Ho JHC. Stage classification of nasopharyngeal carcinoma: a review. In: De-The G, Ito Y, eds. Nasopharyngeal carcinoma: etiology and control. Lyon: International Agency for Research on Cancer Science Publication, 1978:99 –114. 13. Teo PM, Leung SF, Yu P, Tsao SY, Foo W, Shiu W. A comparison of Ho’s, International Union Against Cancer and American Joint Committee stage classifications for nasopharyngeal carcinoma. Cancer 1991;67:434 –9. 14. Lee AWM, Poon YF, Foo W, Law SCK, Cheung FK, Chan DKK, et al. Retrospective analysis of 5037 patients with nasopharyngeal carcinoma treated during 1976 –1985: overall survival and patterns of failure. Int J Radiat Oncol Biol Phys 1992;23:261–70. 15. Yip TTC, Ngan RKC, Lau WH, Poon YF, Joab I, Cochet C, et al. A possible prognostic role of immunoglobulin-G antibody against recombinant Epstein-Barr virus BZLF-1 transactivator protein ZEBRA in patients with nasopharyngeal carcinoma. Cancer 1994;74:2414 –24. 16. Yip TTC, Lau WH, Ngan RKC, Poon YF, Joab I, Cochet C, et al. Role of Epstein-Barr virus serology in the prognosis of nasopharyngeal carcinoma: the present and the future. Epstein-Barr Virus Rep 1996;3:25–33. 17. Merkel DE, Dressler LG, McGuire WL. Flow cytometry, cellular DNA content and prognosis in human malignancy. J Clin Oncol 1987;5:1690 –703. 18. Merkel DE, McGuire WL. Ploidy, proliferative activity and prognosis: DNA flow cytometry of solid tumors. Cancer 1990;65:1194 –205. 19. Witzig TE, Ingle JN, Schaid DJ, Wold LE, Barlow JF, Gonochoroff NJ, et al. DNA ploidy and percent S phase as prognostic factors in node-positive breast cancer: results from patients enrolled in two prospective randomized trials. J Clin Oncol 1993;11:351–9. 20. Ensley JF, Maciorowski Z. Clinical applications of DNA content parameters in patients with squamous cell carcinomas of the head and neck. Semin Oncol 1994;21:330 –9. 21. Jones DJ, Moore M, Schofield PF. Refining the prognostic significance of DNA ploidy status in colorectal cancer: a prospective flow cytometric study. Int J Cancer 1988;41:206 – 10. 22. Isobe H, Miyamoto H, Shimizu T, Haneda H, Hashimoto M, Inoue K, et al. Prognostic and therapeutic significance of the flow cytometric nuclear DNA content in non-small cell carcinoma. Cancer 1990;65:1391–5. 23. Lopez-Beltran A, Croghan GA, Croghan I, Matilla A, Gaeta JF. Prognostic factors in bladder cancer: a pathologic, immunohistochemical and DNA flow-cytometric study. Am J Clin Pathol 1994;102:109 –14. 24. Braly PS, Klevecz RR. Flow cytometric evaluation of ovarian cancer. Cancer 1993;71:1621– 8. 25. Lutgens LCHW, Schutte B, Jong JMA, Thunnissen FBJM. DNA content as prognostic factor in cervix carcinoma stage IB-III treated with radiotherapy. Gynecol Oncol 1994;54:275– 81. 26. Lee AW, Foo W, Poon YF, Lew CK, Chan DK, O SK, et al. Staging of nasopharyngeal carcinoma: evaluation of N-staging by Ho and UICC/AJCC systems. Union Internationale Contre le Cancer, American Joint Committee for Cancer. Clin Oncol (R Coll Radiol) 1996;8:146 –54. 27. Shanmugaratnam K, Chan SH, de-The G, Goh JEH, Khor TH, 28. 29. 30. 31. 32. 33. 34. 35. 36. 37. 38. 39. 40. 41. 42. 43. 2291 Simons MJ, et al. Histopathology of nasopharyngeal carcinoma. Correlations with epidemiology, survival rates and other biologic characteristics. Cancer 1979;44:1029 – 44. Krueger GRF, Kottaridis SD, Wolf H, Ablashi DV, Sesterhenn K, Bertram G. Histological types of nasopharyngeal carcinomas as compared to EBV serology. Anticancer Res 1981;1: 187–94. Hsu HC, Chen CL, Hsu MM, Lynn TC, Tu SM, Huang SC. Pathology of nasopharyngeal carcinoma. Proposal of a new histologic classification correlated with prognosis. Cancer 1987;59:945–51. McGuire LJ, Lee JCK. The histopathologic diagnosis of nasopharyngeal carcinoma. Ear Nose Throat J 1990;69:229 –36. Wu TC, Mann RB, Epstein JI, MacMahon E, Lee WA, Charache P, et al. Abundant expression of EBER1 small nuclear RNA in nasopharyngeal carcinoma: a morphologically distinctive target for detection of Epstein-Barr virus in formalin-fixed paraffin-embedded carcinoma specimens. Am J Pathol 1991;138:1461–9. Yip TTC, Chan JKC, Poon YF, Lau WH, Ngan RKC, Ma VWS. Usefulness of in situ hybridization (ISH) for Epstein-Barr virus encoded RNAs (EBERs) in assisting the histopathological diagnosis of nasopharyngeal carcinoma (NPC) [abstract]. Int J Surg Pathol 1995;2:259. Yip TTC, Chan JKC, Poon YF, Lau WH, Ngan RKC, Ma VWS. Delineation of various factors affecting the sensitivity and specificity of in situ hybridization (ISH) for Epstein-Barr virus encoded RNAs (EBERs) [abstract]. Int J Surg Pathol 1995;2:381. Hedley DW. Flow cytometry using paraffin-embedded tissue: five years on. Cytometry 1989;10:229 – 41. Shankey TV, Rabinovitch PS, Bagwell B, Bauer KD, Duque RE, Hedley DW, et al. Guideline for implementation of clinical DNA cytometry. Cytometry 1993;14:472–7. Whitney DR. A bivariate extension of the U statistic. Ann Math Stat 1951;22:274 – 82. Samuels ML. Statistics for the life sciences, 1st edition. San Francisco and London; Collier Macmillan Publishers, 1989. Breslow N. Statistical methods for censored survival data. Environ Health Perspective 1979;32:182–92. Peto R, Pike MC, Armitage P, Breslow NE, Cox DR, Howard SV, et al. Design and analysis of randomized clinical trials requiring prolonged observation of each patient. Br J Cancer 1977;35:1–39. Saunders MI, Dische S, Hong A, Grosch EJ, Fermont DC, Ashford RF, et al. Continuous hyperfractionated accelerated radiotherapy in locally advanced carcinoma of the head and neck region. Int J Radiat Oncol Biol Phys 1989;17:1287–93. Saunders MI, Dische S. Continuous, hyperfractionated, accelerated radiotherapy (CHART) in non small-cell carcinoma of the bronchus. Int J Radiat Oncol Biol Phys 1990;19: 1211–5. Begg AC, Hofland I, Van Glabekke M, Bartlelink H, Horiot JC. Predictive value of potential doubling time for radiotherapy of head and neck tumour patients: results from the EORTC Cooperative Trial 22851. Semin Radiat Oncol 1992;2:22–5. Horiot JC, Bontemps P, Begg AC, Le Fur R, van den Bogaert W, Bolla M, et al. Hyperfractionated and accelerated radiotherapy in head and neck cancers: results of the EORTC trials and impact on clinical practice. Bull Cancer Radiother 1996;83:314 –20. 2292 CANCER December 1, 1998 / Volume 83 / Number 11 44. Horiot JC, Bontemps P, van den Bogaert W, Le Fur R, van den Weijngaert D, Bolla M, et al. Accelerated fractionation (AF) compared to conventional fractionation (CF) improves loco-regional control in the radiotherapy of advanced head and neck cancers: results of the EORTC 22851 randomized trial. Radiother Oncol 1997;44:111–21. 45. Hall EJ, editor. Radiobiology for the radiologist. 4th edition. Philadelphia: J.B. Lippincott Co., 1994:211–30. 46. Azli N, Armand JP, Rahal M, Wibault P, Boussen H, Eschwege F, et al. Alternating chemo-radiotherapy with cisplatin and 5-fluorouracil plus bleomycin by continuous infusion for locally advanced undifferentiated carcinoma of the nasopharyngeal type. Eur J Cancer 1992;28A:1792–7. 47. Tsujii H, Kamada T, Tsuji H, Takamura A, Matsuoka Y, Usubuchi H, et al. Improved results in the treatment of nasopharyngeal carcinoma using combined radiotherapy and chemotherapy. Cancer 1989;63:1668 –72. 48. Dimery IW, Peters LJ, Goepfert H, Morrison WH, Byers RM, Guillory C, et al. Effectiveness of combined induction chemotherapy and radiotherapy in advanced nasopharyngeal carcinoma. J Clin Oncol 1993;11:1919 –28. 49. Al-Sarraf M, Pajak TF, Cooper JS, Mohiuddin M, Herskovic A, Ager PJ. Chemo-radiotherapy in patients with locally advanced nasopharyngeal carcinoma: a Radiation Therapy Oncology Group Study. J Clin Oncol 1990;8:1342–51. 50. Chan ATC, Teo PML, Leung WT, Leung TWT, Leung SF, Lee WY, et al. A prospective randomized study of chemotherapy adjunctive to definitive radiotherapy in advanced nasopharyngeal carcinoma. Int J Radiat Oncol Biol Phys 1995;33:569 –77. 51. Rossi A, Molinari R, Boracchi P, Del Vechio M, Marubini E, Nava M, et al. Adjuvant chemotherapy with vincristine, cyclophosphamide, and doxorubicin after radiotherapy in local-regional nasopharyngeal cancer: results of a 4-year multicentre randomized study. J Clin Oncol 1988;6:1401–10. 52. Cvitkovic E. Neoadjuvant chemotherapy with epirubicin, cisplatin, bleomycin (BEC) in undifferentiated nasopharyngeal cancer: preliminary results of an international phase III trial. Proc Am Soc Clin Oncol 1994;13:283. 53. Cheng SH, Liu TW, Jian JJ, Tsai SY, Hao SP, Huang CH, et al. Concomitant chemotherapy and radiotherapy for locally advanced nasopharyngeal carcinoma. Cancer J Sci Am 1997; 3:100 – 6. 54. Lin JC, Jan JS, Hsu CY. Pilot study of concurrent chemotherapy and radiotherapy for stage IV nasopharyngeal cancer. Am J Clin Oncol 1997;20:6 –10. 55. Steel GG. Tumor growth kinetics. In: Steel GG, Adams GE, Horwich A, editors. The biological basis of radiotherapy. 3rd edition. Amsterdam: Elsevier Science Publishers, 1995:8 –13. 56. Tannock IF. The relation between cell proliferation and the vascular system in a transplanted mouse mammary tumor. Br J Cancer 1968;22:258 –73. 57. Vaupel P, Kallinowski F, Okunieff P. Blood flow, oxygen and nutrient supply, and metabolic micro-environment of human tumors: a review. Cancer Res 1989;49:6449 – 65. 58. Costello F, Mason BR, Collins RJ, Kearsley JH. A clinical and flow cytometric analysis of patients with nasopharyngeal cancer. Cancer 1990;66:1789 –95. 59. Reddy SP, Raslan WF, Gooneratne S, Kathuria S, Marks JE. Prognostic significance of keratinization in nasopharyngeal carcinoma. Am J Otolaryngol 1995;16:103– 8. 60. Liebowitz D. Nasopharyngeal carcinoma: the Epstein-Barr virus association. Semin Oncol 1994;21:376 – 81. 61. Kokal WA, Gardine RL, Sheibani K, Zak IW, Beatty JD, Riihimaki DU, et al. Tumor DNA content as a prognostic indicator in squamous cell carcinoma of the head and neck region. Am J Surg 1988;156:276 – 80. 62. Hemmer J, Kriedler J. Flow cytometric DNA ploidy analysis of squamous cell carcinoma of the oral cavity. Comparison with clinical staging and histologic grading. Cancer 1990;66: 317–20. 63. Rua S, Comino A, Fruttero A, Cera G, Semeria C, Lanzillotta L, et al. Relationship between histologic features, DNA flow cytometry, and clinical behavior of squamous cell carcinomas of the larynx. Cancer 1991;67:141–9.