Accelerated and Hyperfractionated Radiotherapy - treatment of head and neck cancer
An area of current interest in the treatment of head and neck cancer is to use “nonstandard” radiation treatment regimens to improve the therapeutic ratio between tumor control and normal tissue damage. Standard once-a-day treatment approaches developed empirically, and while they are convenient from the viewpoint of operating a radiation oncology department, there is no reason to think that they cannot be improved upon. There are currently two basic approaches that are being investigated in an attempt to find a more optimal radiation treatment fractionation schema.

Hyperfractionation refers to using smaller fraction sizes, multiple daily treatments, a higher total dose of radiation, and a total treatment time that is about the same duration as for conventional radiotherapy. The basic idea with this approach is to take advantage of the difference in the shapes of the cell survival curves between tumors and late-responding tissues in order to deliver a higher dose of radiation without increasing the late effects of treatment. A typical example for head and neck cancer would be to give 120 cGy per fraction (Fx) twice daily (bid) to a total dose of 8,160 cGy. The interval between fractions must be approximately 4.5 to 6 h to allow for repair of damage to normal tissues.

Accelerated fractionation refers to using a fraction about the same size (or perhaps slightly smaller) as in conventional fractionation, multiple daily treatments, a shorter overall treatment time, and a total dose about the same (or perhaps slightly less) than given in the conventional radiation schema.

The basic idea with this approach is to overcome the effects of tumor repopulation by shortening the overall time. Theoretically, this should improve tumor control for the same radiation dose without increasing the overall late effects. Following Ang, the various accelerated fractionation schemas can be classified into three categories: A, a short, intensive course; B, a split course; and C, a concomitant boost. The continuous hyperfractionated accelerated radiotherapy (CHART) regimen, which consists of giving 150 cGy/Fx three times a day on 12 consecutive days to a total dose of 5,400 cGy, is a prototype regimen in category A. The approach of Wang and colleagues (160 cGy/Fx bid to a total dose of 6,720 cGy with a 2-week break after 3,840 cGy) is a prototype regimen in category B. The concomitant-boost approach of Ang and colleagues (180 cGy/Fx qd to 5,400 cGy, with an additional daily treatment of 150 cGy/Fx to the final target volume being delivered during the final 12 days of therapy) is a prototype regimen in category C.

Model calculations show that which of these approaches would be expected to produce the best results is critically dependent on the radiobiologic parameters of the tumor (eg, a/b, and the tumor proliferative properties [eg, TD and F]). In clinical practice, even tumors arising in a particular site and having the same basic histology can exhibit a wide variation in the values for these parameters. In the absence of predictive assays that would allow the clinician to individualize the treatment for a given patient, large-scale clinical trials are necessary to compare one approach to another for the treatment of specific classes of tumors.

In the United States, early work using hyperfractionation to treat locally advanced head and neck cancer took place at the University of Florida. Doses of up to 8,011 cGy were given, using 120 cGy bid. Compared to historical controls, there was better locoregional control, a greater degree of acute mucositis, and equivalent late effects. The first large phase III trials took place outside the United States. Datta and colleagues reported on 176 patients treated in India who were randomized to standard fractionation to 6,600 cGy versus 120 cGy bid to 7,920 Gy. At 2 years, locoregional control was 63% on the hyperfractionation arm versus 33% on the standard arm (p < .001). Pinto and colleagues reported on 98 patients treated in Brazil who were randomized to either 6,600 cGy via standard fractionation or 110 cGy bid to 7,040 cGy. Their patients all had stages III and IV carcinomas of the oropharynx and so represented a reasonably homogeneous population. Locoregional control was 84% on the hyperfractionation arm versus 64% on the standard arm (p = .02). Horiot and colleagues reported on a European Organization for Research and Treatment of Cancer (EORTC) study involving 356 patients with oropharyngeal tumors who were randomized to either 7,000 cGy standard fractionation or 115 cGy bid to 8,050 cGy. Locoregional control was 40% on the standard arm versus 59% on the hyperfractionation arm (p = .02). In all of these trials, the acute mucositis was more severe on the hyperfractionation arm, but the late effects were similar.

One of the most aggressive of the accelerated treatment regimens is the CHART approach. Doses of 150 cGy/Fx are given three times a day on 12 consecutive days to a total dose of 5,400 cGy. An early analysis of a phase I/II study showed more severe early effects and four cases of radiation myelitis in patients whose spinal cords received 4,500 to 4,800 cGy.535 This high incidence of myelitis cannot be satisfactorily explained on the basis of incomplete repair of nerve tissue between fractions. However, when the regimen was taken into a phase III trial, the cord dose was limited to 4,000 cGy. The randomized trial showed no difference in locoregional control or surviva,l but there were no additional cases of myelitis with this lower dose.

In 1991 the Radiation Therapy Oncology Group (RTOG) began a randomized phase III trial testing four different radiation fractionation schemas for patients with inoperable squamous cell tumors of the head and neck. Standard fractionation at 200 cGy/Fx to 7,000 cGy was the control arm. Another arm was hyperfractionation radiotherapy at 120 cGy/Fx bid to a total dose of 8,160 cGy, which was determined to be an acceptable dose on a prior dose-searching study. The remaining two arms were categories B and C variants of an accelerated fractionation schema. One of these arms was the split-course regimen of Wang and colleagues (described above) and the other was the concomitant-boost regimen of Ang and colleagues (also described above). The study closed in 1997 with 1,073 patients being evaluable for analysis.

Analysis of the three altered fractionation regimens was done in comparison with the standard fractionation arm. There was no improvement in locoregional control with the accelerated/split course, but there was a statistically significant benefit with the hyperfractionated (p = .045) and concomitant-boost (p = .05) regimens. There was also a trend toward improved disease-free survivals, which did not reach statistical significance. As might be expected, with death due to locoregional failure being only about 50% of the total deaths, all the arms were equivalent in terms of absolute survival. The incidence of grade 3 or greater (RTOG/EORTC scoring scheme) acute toxicity was worse on all three altered fractionation arms, with the differences being statistically significant. Only the concomitant-boost regimen had significantly worse late effects when compared to standard fractionation. However, this may be an artifact of the definition of “late effects” as those that are present 90 days or longer following treatment. The “acute effects” on the concomitant-boost arm persisted longer than on the other arms, and hence the difference in scoring. If a longer cutoff time is chosen, then the late-effects spectra are the same on all four arms.

Fraction size may be important in determining late effects. Jen and colleagues have reported an increased risk of temporal lobe necrosis in patients with nasopharyngeal tumors who were treated with 160 cGy bid instead of 120 cGy bid. In both fractionation schemas, the time interval between same day fractions was kept at 6 h. In the 120 cGy fraction group, the total tumor dose was higher at 8,000 cGy while in the 160 cGy fraction group, the dose was in the range of 6,840 to 7,640 cGy. The portal configuration was the same in each group. None of the 70 patients treated with the 120 cGy fractions developed temporal lobe necrosis whereas 3 of 11 (27%) of patients treated with 160 cGy fractions developed symptomatic temporal lobe necrosis. The estimated doses to the temporal lobes were, respectively, 6,000 to 7,440 cGy on the 120 cGy bid arm and 4,480 to 6,700 cGy on the 160 cGy bid arm.