Predictive value of potential doubling time in head and neck cancer patients treated by conventional radiotherapy

Int. J. Radiation

Biol.

Phys., Vol. 38, No. 4, pp. 677-683, 1997 Copyright 0 1997 Elsevier Science Inc. Printed in the USA. All rights reserved 0360.3016/97 $17.00 + .OO

PI1 SO360-3016( 97)00066-7

ELSEVIER

l

Oncology

Clinical Investigation PREDICTIVE NECK CANCER

VALUE OF POTENTIAL DOUBLING TIME PATIENTS TREATED BY CONVENTIONAL

BJORN ZACKRISSON, PETRA

IN HEAD AND RADIOTHERAPY

M.D., * HANS GUSTAFSSON, M.D., + ROGER STENLJNG, FLYGARE, M.D.* AND GEORGE D. WILSON, PH.D.§

M.D.,

*

Departmentsof *Oncology, ‘Otorhinolaryngology-Head & Neck Surgery,*Pathology,Umel University, Umei, Sweden, gGrayLaboratory CancerResearchTrust, Mount Vernon Hospital,Northwood,UK Purpose: The goal of this study was to investigate the clinical utility of pretreatment measurements of tumor cell kineticsto predict the outcome of patients with squamous cell carcinoma of the head and neck receiving conventional radiotherapy. Methods and Materials: All patients received between 64 and 70 Gy as 2 Gy fractions, five fractions per week. Cell kinetics were assayed rapidly and quantitatively using flow cytometric evaluation of iododeoxyurldine ( IdUrd) incorporation, in viva, from a biopsy removed several hours after the administration of the DNA precursor to the patient prior to the start of treatment. Results: The measured proliferation parameters were not related to the clinicopathological features of the tumors, emphasizing the independent nature of these parameters. In univariate analysis, nodal involvement was the most important clinical feature of the tumors related to local control followed by T,, DNA aneuploidy, and attainment of complete regression at 6 weeks. Of these only Tpot and nodal status maintained significance in multivarlate analysis, with respect to loco-regional control. In subgroup analysis, T, was able to stratify patients into high or low rate of loco-regional control in node negative patients, in aneuploid tumors and in patients who did achieve complete regression at 6 weeks. For cause specific survival, N-stage was the only parameter that significantly discriminated the prognosis in these patients. Conclusions: The conclusion of this study is that T,, provides clinically important information that can predict patients with a low probability of achieving long-term local control with conventional fractionation. Further improvements to the methodology to address the shortcomings of analyzing diploid tumors may increase the predictive power of the measurement. 0 1997 Elsevler Science Inc. T pot, Head and neck cancer, Radiotherapy.

INTRODUCTION Radiotherapy remains the most important mode of treatment for patients with squamous cell carcinomas of the head and neck (SCC). However, local control is never gained or is subsequently lost in approximately half of the treated patients. Improvements in local tumor control rates should lead to more patient cures and could limit the development of distant metastases(15, 22). The failure of conventional radiotherapy to control more tumors can be attributed to several biological and physical factors. Rapid repopulation of surviving clonogenic tumor cells has received much attention as a possible cause of failure, particularly in treatment scheduleswith long overall time ( 13, 25). If tumors are shown to possessthe capacity for rapid repopulation, acceleration of treatment ( 2 1) or a concomitant boost regime might be more appropriate ( 12, 23 ) .

However, these altered fractionation schedulesmay not be feasible or beneficial in all cases,and this highlights the need to be able to select patients who would benefit most from a modified regime ( 12, 18). These considerations prompted the searchfor a reliable and rapid method to assess human tumor proliferation that did not rely on radioactive DNA precursors.The use of flow cytometric evaluation of the in viuo incorporation of halogenated pyrimidines ( 11, 14) has satisfiedmany of the criteria for a proliferation measurementthat can be used in the clinical setting (24). Over the past few years, several studieshave appearedin the literature reporting the influence of Tpotin both conventional and altered fractionation schedules(4, 5, 8, 9, 16, 17) . The results of thesestudieshave been variable, with someshowing none or borderline significance of T,,, while in others the Tpofmeasurementhas proven to be a strong prognostic parameter.A recurring problem in many of the reported stud-

Reprint requeststo: B. Zacluisson,Departmentof Oncology, UmedUniversity, S-901 85 Umeb,Sweden.

CancerSociety, The Lions CancerResearchFoundationUmei University andCancerResearchCampaign.

Acknowledgments-This

work was supported by The Swedish

Accepted for publication 677

26 August 1996.

678

I. I. Radiation

Oncology

l Biology

l

Physics

ies so far is the low numbers of patients in homogeneously treated groups. The reason for this is that studies utilizing halogenated DNA markers for T,, measurements are by definition prospective. The present study has included 89 evaluable patients since February 1989 to November 1994 such that there is a median follow-up of 30 months and a minimum of 12 months.

METHODS

AND MATERIALS

Patients and administration ofldUrd Over the period February 1989 to November 1994, 127 consecutive patients provided informed consent and were given an intravenous infusion of 100 mg IdUrd in 100 ml of 0.9% saline, prepared by the Pharmacy Department, University Hospital, Umei, over a period of less than 10 min. Of these, 7 patients were not squamous cell carcinomas ( SCC ) and were excluded from this study and 12 either received non conventional radiotherapy or surgery alone and were also excluded. In the remaining 108 patients, excision biopsies were taken, usually under general anaesthetic (apart from some oral cavity tumors), at a chosen time after injection; this ranged from 1.8 to 7.8 h, with a median of 4.9 h. No patient has experienced acute or chronic toxicity associated with this procedure. The biopsy was divided in half and one piece was fixed in 4% buffered formaldehyde and embedded in paraffin for routine histopathology, and the other was fixed in 70% ethanol and stored at 4°C for FCM. The study was approved by the Local Ethics Committee of Umei University. Treatment Radiotherapy was given to a total dose of 64 to 70 Gy as 2 Gy fractions given 5 days per week with an overall treatment time of 6.5 to 8 weeks. The radiation qualities were 4 or 6 MV x-rays (some patients also received electrons for the treatment of posterior neck nodes). Only two patients had significant interruptions (10 and 14 days) in their planned treatment due to poor health status. In Tl and T2 tumors, the lesion, with a margin and adjacent neck nodes, were included in the planning target volume. In more advanced disease, all neck nodes were uni- or bilaterally treated prophylactically with 50 Gy. If nodes were clinically involved, they were treated with the same dose as the primary tumor. Postradiotherapy surgery was performed, whenever possible, in cases where local treatment failed. In addition, the Department policy was to routinely excise the site of tumors in the oral cavity and positive neck nodes after radiotherapy. A total of 33 patients underwent surgery as part of the primary treatment. Generally, any surgical procedure took place at 6 weeks after completing radiotherapy. Clinical investigation of tumor status after 6 weeks was used to assess the primary response to radiotherapy; this was frequently accompanied by biopsy or in conjunction with a surgical procedure. The assessment did not include the effect of surgery. Recurrent disease was treated by radical surgery whenever possible.

Volume

38, Number

4, 1997

IdUrd staining and FCM The processing and staining of the tissue has been described in detail elsewhere ( 17). The method is based upon pepsin digestion of tissue fragments into nuclei and staining with a high affinity antibody (IU4, Caltag Laboratories, San Francisco, CA) against iodo- or bromodeoxyuridine after denaturation of DNA with 2 M HCl. Incorporation of IdUrd was detected by attaching a secondary antibody (rabbit antimouse F( ab), fragment FITC conjugate, DAKOPATTS AB, Alvsjii, Sweden) and total DNA content was stained using 10 mg/ml propidium iodide. Flow cytometry was carried out on FACScan (Becton Dickinson, Mountain View, CA) using Lysys II software and equipped with a doublet discrimination module to exclude nuclei doublets, triplets etc.; at least 10,000 events were recorded for each specimen. The profiles were analyzed as previously described (17) to calculate the DNA index, the corrected labeling index (LI) allowing for cell division during the time from injection to biopsy, the duration of S-phase (Ts) using the method of Begg (6) [and the method of Ritter (20) in 5 cases where the labeling period was short], and the potential doubling time (T,,,) using an assumed X value of 0.8. Statistical analysis The cause-specific mortality and disease-free survival from the date of diagnosis was calculated by the KaplanMeier method, and differences in outcome were statistically evaluated by the log rank test. Cutoff values were chosen as the median unless otherwise stated. Independent prognostic significance was established using the Cox’s proportional hazard model. Wilcoxon statistics were used for testing of differences of mean values. Tests of differences between proportions were performed by the chisquare method.

RESULTS Clinical features Of the 108 patients who initially satisfied the criteria of the study, 7 cases had to be excluded due to little or no tumor being present in the biopsy removed prior to radiotherapy. In a further nine patients, the FCM profiles were not evaluable for either technical reasons or lack of material and three patients were lost to follow-up. This left 89 cases where all the criteria of the study were met. The characteristics of these patients are summarized in Table 1. The patient population did not include any nasopharyngeal, columella, parotid, or lip tumors. All patients were followed up for a minimum of 12 months, with a median of 30 months (range 12-8 1 months ) . Proliferation characteristics The measured flow cytometry parameters are summarized in Table 2. The results are comparable with other data in the literature showing that head and neck tumors are a rapidly proliferating group of tumors with a median

TPOi and outcome

of radiotherapy

Table 1. Clinicopathological

0 B. ZACKRISSON

characteristics

et al.

679

of the 89 patients studied

Parameter 65.0 years (range27-89) No. of patients T2 T3 35 12 N+ NO 34 55 Moderate Poor 19 58 Oropharynx Hypopharynx 19 6

Age at presentation

Median

T-Stage

Tl 7

N-Stage Differentiation

Well 10 Oral Cavity 48

Tumor site

Tpotof only 3.5 days. When the data were further stratified according to ploidy status, significant differences were found in LI and Ts between aneuploid and diploid tumors. In aneuploid tumors, the LI was higher than that found in diploid tumors due primarily to the ability to exclude nonproliferating diploid cells from the profile analysis; a longer Ts was also observed. These two variables are inversely related in the calculation of Tpot, so this resulted in an insignificant difference in Tpat, although there was a trend for shorter values in aneuploid tumors. Proliferation and clinico-pathological features The relationship between the proliferation parameters was studied as a function of clinico-pathological features to establish whether they were dependent or independent variables. Table 3 summarizes the distribution of proliferation parameters amongst the different clinical subtypes. There were no statistically significant associations between any of the proliferation parameters and clinicopathological features emphasizing the independent nature of proliferation information in SCC of the head and neck. Loco-regional control-all patients At the time of last follow-up, the overall local control rate was 52% for all patients. Figure 1 shows the local control data, stratified according to nodal status (Fig. 1A), T-stage (Fig. lB), and differentiation (Fig. 1C). Nodal status, as expected, showed a highly significant influence on outcome of conventional radiotherapy (p < O.OOl), there was a trend for control to be lost asT-stage increased

Table 2. The proliferation

T4 35 Anaplastic 2 Larynx 16

but there was no apparent relationship with differentiation. Of the other clinical features, only the primary response to radiotherapy, measured as status at 6 weeks postradiation therapy, showed any clinical significance (data not shown). Those tumors achieving complete regression (CR) (76%) had a higher probability for loco-regional control than those who had remaining tumor after radiotherapy (p < 0.01). The parametersgenerated by FCM (Fig. 2) showed that there was an effect of DNA ploidy on local control (Fig. 2A) ; aneuploid tumors did significantly worse than diploid (p < 0.05 ) . Of the proliferation parameters,there was a trend for higher LIs to be associated with worse local control, but this did not reach significance (Fig. 2B ), while Ts had no influence on local control (data not shown). Tpotwas the most important parameter with the “fast” (~3.5 days) tumors having greater risk for local recurrence (p < 0.05) than the “slow” (>3.5 days) tumors (Fig. 2C). In multivariate analysis, the influence of N-stage, Tpo,, ploidy and outcome at 6 weeks were tested. Only nodal status (p < 0.001) and TP, (p < 0.05) retained an independent significance. Loco-regional control-subgroup analysis The influence of T,,, on local control within the clinical and biological features that showed the most prognostic significance, in the complete data set, was examined. Table 4 summarizes the data and Fig. 3 shows the local control curves for the significance of Tpocin aneuploid tumors

characteristics

of the studied tumors

Parameter

Time of biopsy (h)

DNA index % aneuploid

Ll (%)

T, 0)

Median Range SD Diploid Aneuploid

4.9 1.8-7.8 1.20

66.3%

8.0 0.8-29.9 6.3 5.2 9.4 p < 0.01

8.1 3.4-33.9 4.6 6.9 9.1

SD represents standard deviation. The last two rows report the proliferation parameters of diploid the values are the median and the statistical difference assessed by Wilcoxon statistics.

p < 0.001 and aneuploid

Tprrt (days) 3.5 1.O-58.3 7.3 4.2 3.3 p = 0.61 tumors separately;

I. J. Radiation Oncology l Biology l Physics

680

Table 3. The relationship between proliferation clinicopathological features Clinical

parameter

Node positive Node negative Tl T2 T3 T4 Well

Moderate Poor/anaplastic Oral cavity Oropharynx Wpopharynx Larynx CR at 6 weeks No CR at 6 weeks

LI (%) 8.1 6.8 11.2 7.1 7.3 8.7 7.3 8.2 8.1 7.2 6.8 14.3 8.6

8.1 7.4

T, (h) 7.6 8.7 6.5 7.7 8.2 8.8 7.7 8.3 9.5 8.6 8.8 10.2 6.9 8.6 8.0

parameters and

T,, (davs) 3.4 3.6 3.2 3.7 3.5 3.4 3.5 3.4 4.7 3.4 4.7 2.5 3.0 3.2 3.6

Volume 38, Number 4, 1997

ment schedules and that the pretreatment proliferation characteristics identifies patients at risk of local recurrence if treated over 6-8 weeks. The T,,, proved to be the most important biological characteristic; tumors with a Tpot above the median value of 3.5 days had a local control rate of 7 1% as opposed to 37% observed in tumors with a T,,, less than the median value. Only nodal status was more significant in multivariate analysis. However, stratification of node-negative patients according to TP,, gave a highly significant (p < 0.01) prediction of patients who recurred locally. Two other parameters showed clinical significance; these were the evidence of CR at 6 weeks and DNA aneuploidy. Both of these parameters lost their significance in multivariate analysis, but Tpotwas able to

The values reported are medians; there were no statistical differences in any of the clinical subtypes assessed by Wilcoxon statistics.

(Fig. 3A) and in node negative patients (Fig. 3B). T,, retained the ability to predict local control in the 60 aneuploid tumors but was not discriminatory in the smaller diploid subpopulation. The influence of TPt was highly significant in the 55 node-negative patients, and there was a trend also in the 34 node-positive patients, but this did not reach significance. In those patients who did not achieve complete regression of the tumor at 6 weeks, TPt did not give any further information. However, if complete regression was attained, then Tpot was able to discriminate the group at risk of local recurrence. Several other subgroup analyses were performed. Of particular note was the finding that Tpotdid not discriminate patients who did well or badly with hypopharynx or laryngeal cancers (although numbers were very restricted, 6 and 16 patients, respectively), but was significant for tumors in oral cavity and oropharynx (67 patients) (p < 0.05). The influence of surgery was also investigated. Those patients that underwent surgery (33 patients) showed a trend for T,, to predict outcome but this did not reach statistical significance. The patients who did not undergo surgery (66 patients) showed a statistically significant lower rate of local control when Tpotwas shorter than the median. Cause-specijic survival At the time of last follow-up, the causespecific survival rate was 72%. None of the proliferation parameters were significant, the p-values for LI, Ts, and T,, were 0.46, 0.36, and 0.35, respectively. Only nodal status retained clinical significance in determining survival (p > 0.001) . DISCUSSION This study supports the concept that repopulation of tumor cells is a problem in prolonged radiotherapy treat-

A I - - -

N-negative N-positive

p
I o.oL 0

I

I

10

20

30

I

I

40

50

____----_--

B

g

b B --

.4'l .4

0.0 75 0 6 = n $q

Ilo

2'0

---

Tl T2 T3 T4

~=0.42

3'0

4'0

s'o

___. 1

1.y

-

1 60

----,

---v-m

a .2 I

I

WPll . ..T..

----

,I,

1

Moderate Poor p=O.67

---

I 0.0,

0

. 10

20

I 30

40

. 50

. 60

Time (months) Fig. 1. Local control stratified according to (A) nodal status, (B) T-stage and (C) differentiation.

T,,, and outcome of radiotherapy 0

B. ZACKRISSON

681

et al.

Table 4. The influenceof Tpoton local control in biologicaland clinical subgroups No. of tumors Subgroup

Fast

Slow

p-Value

10

19 25 27 17 35 9

0.60 <0.05

Diploid Aneuploid Node negative Node positive CR at 6 weeks No CR at 6 weeks 0

10

20

30

40

50

60

35 28

17 33

12



Fastand slowrepresentthe numberof tumorsbelow or above the overall medianT,, value of 3.5 days in thesesubgroups. Testsof significancewere madeby the chi-squaremethod.

B ---I ‘my :----------I

.6”

I

The early promise of Tpotas a predictor of outcome in radiotherapy is still being questioned (2) even though a number of correlative reports have appeared in the literature (4-6, 8, 9, 11, 17). There are limitations, both statistical and methodological, inherent to many of the studies so far published. These include small numbers of patients, variable follow-up times, and the problems associated with the FCM measurement of Tpot(24). In this study we have reported on a group of 89 homogeneously treated patients with a median follow-up of 30 months.

._-____-_--_____-___--

.4’

----.

.2’

L&3.0% Lb8.0% D=o.16

-

L - - - __- - - - ,

.4#’ ----

.2

t 0.01 0

Tpot>B.Sdays

,--_-_----_-_____----__ Tpot>3.5 Tpotz3.5

days days

.6 ’

-:

DCo.05

I 10

I 20

I 30

. 40

I 50

1 60

Time (months) Fig. 2. Local control stratifiedaccordingto (A) DNA aneuploidy, (B) LI and (C) T,,,.

5

EL8 z

.4 ’

1 I--l ---I

1----------1

9L________________

.2’

0.0, 0

A I

10

I

20

30

40

I

50

t

E

further stratify patients into good and bad outcome within the aneuploid tumors and within those patients who did achieve CR at 6 weeks. The fact that the treatment of the patients included surgery in certain tumor types and stages did not seem to bias the results. The inability to reach statistically significant higher risk of local recurrence for those with a short Tporwho underwent surgery may well be due to the small number (33 patients). In this study none of the proliferation parameters showed any statistically significant influence on the cause specific survival. However, the follow-up may be too short to make that type of analyses,because72% of the patients are still alive and too few events of death may have occurred to make this analysis relevant.

---a

,611 .4” .2”

0.0, 0

--7

9,.---------7

Tpob3.5 - - - - Tpots3.5 pco.01

I

10

I

20

I-----------.-------.

days days

B I

30

I

40

I

50

Time (months) Fig. 3. Local control stratifiedaccordingto Tputin (A) aneuploid tumorsand (B ) nodenegativetumors.

682

I. .I. Radiation Oncology 0 Biology 0 Physics

There have been two studies (4, 8) in head and neck cancer that have shown promising initial results but which lost any statistical significance of Tpotin subsequentanalyses (personal communication from Drs. Begg and Bourhis). It has been suggested(2) that this may be due to the effect of the fast tumors dominating the initial analysis: fast tumours are more likely to fail earlier than slow tumors at the same level of cell kill (2). However, Corvo and colleagues originally reported significance of Tpot in only 18 patients with standard radiotherapy (9) and have recently updated this study (10) with 37 patients and a median follow-up of 55 months; the significance of Tpot has increased and become the most important parameter determining local control in multivariate analysis of these conventionally treated patients. The same is true of this present study. In our earlier report ( 17)) 23 patients were assessedwho had received conventional radiotherapy, and the T,, generated by FCM did not achieve significance in predicting outcome in terms of cause-specific mortality. In this study, with 89 patients, we have a significant effect of Tpot(p < 0.05) for loco-regional control in all tumors and in subgroup analysis. A third example of the significance of Tpotincreasing with further analysis is the study of Awwad ( 1) . In a recent analysis on 52 patients receiving radical radiotherapy (conventional and split-course acceleration), a fast Tpot (<4.0 days) predicted patients with worse outcome (p < 0.05); this effect was stronger in the conventional arm alone (p < 0.03) (personal communication) . It would appearthat measurementof cell kinetic parameters may play a major role in selecting patients for more appropriate treatment scheduling, but is it necessary to measure the Tpotor does the LI furnish enough information? In this present study, it was clearly the Tpotthat was the more important parameter. The LI data did not reach significance in predicting loco-regional control. This was also true of the recent update by Corvo ( lo), in which the p-value for Tpotwas 0.0006 in patients treated by conventional fractionation, while the LI was also significant, but at a value of p = 0.04. In the previous studies by Begg et al. (5 ) and Bourhis et al. (8 ) , the Tpotwas as good as or better than LI alone in predicting loco-regional control. This would support the view that it is important to measure the Ts, which can only be achieved by FCM, to be able

Volume 38, Number 4. 1997

to calculate the Tpot. The duration of S-phaseusually provides no independent information in terms of predicting outcome, but it is a heterogeneous parameter and is apparently related to DNA aneuploidy. There is a trend, in most studies, for a longer Ts to be associatedwith aneuploid tumors and a shorter Ts with diploid tumors. Therefore, the ranking of tumors using LI alone can be quite different from that when Tpot is used due to the counterbalancing relationship between LI and Ts in calculating Tpot. A recognized problem with FCM is the uncertainty of measuring proliferation in diploid tumors because of the inability to discriminate normal cells from neoplastic cells. It was interesting to note from this study that ploidy status did have significance in delineating loco-regional control (Fig. 2). This has not been the case with most other studies of ploidy in head and neck cancer. Of more interest was the finding that, within the diploid or aneuploid subgroups, Tpotwas only able to discriminate outcome in the aneuploid tumors (Table 4). This would also support the concept that proliferation measurementsin diploid tumors are flawed. A significant percentage of tumors may have their Tpotvalues underestimated using FCM alone. There are methods to distinguish normal from neoplastic cells in diploid tumors using either an FCM approach (3, 19) or immunohistochemical recognition (7). We are currently assessingthe latter option in this series of patients. In keeping with the other studies quoted, none of the proliferation parameters were able to predict the overall survival of patients treated with conventional radiotherapy. It is well recognized that other factors determine the overall survival of patients, such as alcohol and tobacco abuse, poor general health status, and second primary tumors, which are almost certainly independent of cell proliferation characteristics. The overall conclusion of this study is that Tpot measurement is a predictive parameter for patients who have a high risk of loco-regional loss of control with conventional fractionation schedules. These patients might benefit from an altered regime such as acceleration or concomitant boost. Further improvements in the methodology including sample processing, data analysis, and accounting for contaminating normal cells in diploid tumors should improve the efficacy of the technique and establish proliferation measurementsas a routine predictive assay.

REFERENCES 1. Awwad, H. Unconventional fractionation studies and T,,, correlation. Semin. Radiat. Oncol. 2:62-66; 1992. 2. Begg, A. C. The clinical status of Trot as a predictor? or why no tempest in the T,,,! Int. J. Radiat. Oncol. Biol. Phys. 32: 1539- 1541; 1995. 3. Begg, A. C.; Hofland, I. Cell kinetic analysis of mixed populations using three-color flow cytometry. Cytometry 12:144-151; 1993. 4. Begg, A. C.: Hofland. I.; Moonen, L.: Bartelink, H.; Schraub, S.; Bontempts, P.; Le Fir, R.; Van den Bogaert, W.; Caspers, R.; van Glabbeke, M.; Horiot, J.-C. The predictive value of

cell kinetic measurements in a European trial of accelerated fractionation in advanced head and neck tumors: an interim report. Int. J. Radiat. Oncol. Biol. Phys. 19:1449-1453; 1990. 5. Begg, A. C.; Hofland, I.; van Glabbeke, M.; Bartelink, H.; Horiot, J.-C. Predictive value of potential doubling time for radiotherapy of head and neck tumor patients: Results from the EORTC co-operative trial 2285 1. Semin. Radiat. Oncol. 2~22-25; 1992. 6. Begg, A. C.: McNally, N. J.; Shrieve, D. C.; Karcher, H. A. A method to measure the DNA synthesis and potential dou-

T, and outcome of radiotherapy

7.

8.

9.

10.

11.

12. 13. 14. 1.5.

bling time from a single sample. Cytometry 6620-626; 1985. Bennett, M. H.; Wilson, G. D.; Dische, S.; Saunders, M. I.; Martindale, C. A.; Robinson, B. M.; O’Halloran, A. E.; Leslie, M. D.; Laing, J. H. E. Tumour proliferation assessed by combined histological and flow cytometric analysis: Implications for therapy in squamous cell carcinoma in the head and neck. Br. J. Cancer 65:870-878; 1992. Bourhis, J.; Wilson G. D.; Wibault, P.; Bosq, J.; Chavaudra, N.; Janot, F.: Luboinski, B.; Eschwege, F.; Malaise, E. P. In viva measurement of the potential doubling time by flow cytometry in oropharyngeal cancer treated by conventional radiotherapy. Int. J. Radiat. Oncol. Biol. Phys. 26:793-799; 1993. Corvo, R.; Giaretti, W.; Sanguinetti, G.; Geido, E.; Orecchia, R.; Barra, S.; Margarino, G.; Bacigalupo, A.; Vitale, V. Potential doubling time in head and neck tumors treated by primary radiotherapy: Preliminary evidence for a prognostic significance in local control. Int. J. Radiat. Oncol. Biol. Phys. 27:1165-1172; 1993. Corvo, R.; Giaretti, W.; Sanguineti, G.; Geido, E.; Orecchi, R.; Guenzi, M.; Margarino, G.; Bacigalupo, A.; Garaventa, G.; Barbieri, M.; Vitale, V. In vivo cell kinetics in head and neck squamous cell carcinomas predicts local control and helps guide radiotherapy regimen. J. Clin. Oncol. 13: 18431850; 1995. Dolbeare, F. A.; Gratzner, H. G.; Pallavicini, M. G.; Gray, .I. W. Flow cytometric measurement of total DNA content and incorporated bromodeoxyuridine. Proc. Sot. Natl. Acad. Sci. 80:5573-5577; 1983. Fowler, J. F. How worthwhile are short schedules in radiotherapy? Radiother. Oncol. 18: 165 - 18 1; 1990. Fowler, J. F.; Lindstrom, M. Loss of local control with prolongation in radiotherapy. Int. J. Radiat. Oncol. Biol. Phys. 23:457-467; 1992. Gratzner, H. G. Monoclonal antibody to 5-bromo and 5iododeoxyuridine: A new reagent for detection of DNA replication. Science 218:474-475; 1982. Leibel, S. A.; Scott, C. B.; Mohiuddin, M.; Martial, V. A.; Coia, L. R.; Davis, L. W.; Fuks, Z. The effect of local-re-

l

16.

17.

18.

19. 20.

21.

22. 23.

24.

25.

B.

ZACKRISSON

er al.

683

gionai control on distant metastatic dissemination in carcinoma of the head and neck: Results of an analysis from the RTOG head and neck database. Int. J. Radiat. Oncol. Biol. Phys. 21:549-556; 1991. Lochrin, D.; Wilson G. D.; McNally, N. J.; Dische, S.; Saunders, M. I. Tumor cell kinetics, local tumor control and accelerated radiotherapy: A preliminary report. Int. J. Radiat. Oncol. Biol. Phys. 24:87-91; 1992. Nylander, K.; Anneroth, G.; Gustafsson, H.; Roos, G.; Stenling, R.; Zackrisson, B. Cell kinetics of head and neck squamous cell carcinomas: Prognostic implications. Acta Oncol. 33:23-28; 1994. Peters, L. J.; Ang, K. K.; Thames, H. D. Accelerated fractionation in the treatment of head and neck cancer: A critical comparison of different strategies. Acta Oncol. 27: 185- 194; 1988. Ritter, M. A.; Albrechtson, J. L.; Kim, Y. J.; Kinsella, T. J. Tumor cell kinetics in mixed cell populations. Cytometry 16:118-128; 1994. Ritter, M.; Fowler, J. F.; Kim, Y .; Lindstrom. M. J.; Kinsella, T. J. Single biopsy, tumor kinetic analyses: A comparison of methods and extension to shorter sampling intervals. Int. J. Radiat. Oncol. Biol. Phys. 23:81 I-820; 1992. Saunders, M. I.; Dische, S.; Grosch, E.; Fermont, D. C.; Ashford, R. F. U.; Maher, J.; Makepeace A. R. Experience with CHART. Int. J. Radiat. Oncol. Biol. Phys. 21:X71-878; 1991. Suit, H. D. Potential for improving the survival rate for the cancer patient by increasing the efficacy of treatment of the primary lesion. Cancer 50:1227-1234; 1982. Thames, H. D.; Peters, L. J.; Withers, H. R.: Fletcher, G. H. Accelerated fractionation vs. hyperfractionation: Rationales for several treatments per day. Int. J. Radiat. Oncol. Biol. Phys. 9:127-138; 1983. Wilson, G. D. Limitations of the bromodeoxyuridine technique for measurement of tumor proliferation. In: BeckBornholdt, H. P., ed. Current topics in clinical radiobiology of tumors. Berlin: Springer Verlag; 1993:27-43. Withers, H. R.; Taylor, J. M. G.; Maciejewski, B. The hazard of accelerated tumor clonogen repopulation during radiotherapy. Acta Oncol. 27: 13 1- 146: 1988.