Prolonged response without prolonged chemotherapy: a lesson from PCV chemotherapy in low-grade gliomas

Abstract

Previous studies with temozolomide suggest that a prolonged duration of chemotherapy is important for treating low-grade gliomas (LGGs). PCV (procarbazine, CCNU, vincristine) chemotherapy has demonstrated efficacy in treating LGGs, but this therapy cannot be used for a prolonged period because of the cumulative toxicity. The aim of the present study was to evaluate the impact of first-line PCV chemotherapy on LGGs growth kinetics. The mean tumor diameter (MTD) of 21 LGGs was measured on serial magnetic resonance images before (n=13), during, and after PCV onset (n=21). During PCV treatment, a decrease in the MTD was observed in all patients. After PCV discontinuation, an ongoing decrease in MTD was observed in 20 of the 21 patients. Median duration of the MTD decrease was 3.4 years (range, 0.8–7.7) after PCV onset and 2.7 years (range, 0–7) after the end of PCV treatment with 60% of LGGs, demonstrating an ongoing and prolonged (>2 years) response despite chemotherapy no longer being administered. According to McDonald's criteria, the rates of partial and minor responses were 5% and 38% at the end of PCV but 38% and 42% at the time of maximal MTD decrease, which occurred after a median period of 3.4 years after PCV onset. These results challenge the idea that a prolonged duration of chemotherapy is necessary for treating LGGs and raise the issue of understanding the mechanisms involved in the persistent tumor volume decrease once chemotherapy is terminated.

Low-grade gliomas (LGGs) are progressive tumors characterized radiologically by continuous growth before malignant transformation.¹ Dynamic evaluation of the mean tumor diameter (MTD) has been shown to have a strong prognostic significance² and appears to be a reliable tool for assessing the impact of both chemotherapy and surgery on LGGs growth kinetics.^3,4 In LGGs treated with temozolomide, it has been suggested that a prolonged duration of treatment is important in order to achieve a prolonged response, namely in continuously responding patients.⁴ PCV (procarbazine, CCNU, vincristine) chemotherapy has demonstrated efficacy in treating LGGs, but this therapy cannot be used for a prolonged period because of cumulative toxicity.^5–10 This issue has raised concerns about the potential long-term efficacy of PCV treatment. However, there are no data concerning the kinetics of LGGs growth after PCV treatment. In the present study, we evaluated the kinetics of LGGs growth in a series of 21 patients treated with first-line PCV.

Materials and Methods

We reviewed a series of patients treated for an LGG at the Pierre Wertheimer Neurological Hospital and the Leon Bérard Cancer Center of Lyon between 1994 and 2005. The following inclusion criteria were utilized: histological diagnosis of World Health Organization grade II oligodendroglioma, oligoastrocytoma, or astrocytoma after central review (D.M.); age ≥18 years; Karnofsky performance status ≥40; measurable disease on magnetic resonance imaging (MRI); evidence of progressive disease, either clinically or radiologically; initial treatment with PCV without previous specific treatment of the tumor except surgery; and no suspicion of anaplastic transformation at chemotherapy onset. All patients who underwent a genetic analysis of tumor samples collected for this study signed a written informed consent form.

The treatment protocol consisted of the standard PCV schedule (lomustine 110 mg/m² on day 1, procarbazine 60 mg/m² on days 8–21 and vincristine 1.4 mg/m², maximum 2 mg on days 8 and 29, in 6-week cycles, for a maximum of 6 cycles). After finishing PCV treatment, patients were followed until tumor progression without consolidation radiotherapy. Patients left the study upon anaplastic transformation (histologically proved or suspected when rapidly growing foci of enhanced contrast appeared on imaging) or when tumor progression required another treatment, predominantly radiotherapy.

After PCV onset, at least 4 consecutive MRIs, done on a regular basis until tumor progression, were required to evaluate tumor changes. Before PCV onset, the spontaneous growth of the tumor was analyzed in a subset of patients who had been initially followed prior to PCV initiation and in whom at least 2 successive MRIs were available.

As only printed images were available, tumor volumes were estimated manually by one investigator (M.P.) using the 3-diameter technique (V = D₁ × D₂ × D₃/2), as previously described.¹ Tumor volumes were converted into a mean tumor diameter (MTD = (2 × V)^1/3). To estimate the slope of the growth curve of the MTD over time for each patient under each condition (before, during, and after PCV), we performed linear regressions of the MTD of each patient versus time. Tumor response to PCV was evaluated using modified MacDonald's criteria for LGGs.¹¹ The median duration of MTD decrease and overall survival were calculated according to the Kaplan–Meier technique. Nonparametric Wilcoxon tests were performed to compare the mean MTD slopes during and after PCV treatment.

Loss of heterozygosity (LOH) of chromosomes 1p and 19q by codeletion was assessed by PCR as previously described.⁴

Results

Twenty-one patients fulfilled the eligibility criteria. Their main characteristics are listed in Table 1. Sixteen patients received the intended 6 treatment cycles. Five patients had to discontinue PCV because of hematological toxicity (grade III, n = 3; grade IV, n = 2) after 2, 3 (n = 2), 4, and 5 cycles. The median duration of PCV chemotherapy was 9 months (range, 5–12 months). During PCV treatment, MRIs were performed every 2 cycles (n = 9), every 3 cycles (n = 8), or at the end of the treatment only (n = 4). After PCV discontinuation, MRIs were performed every 4 months (n = 4), every 6 months (n = 15), or every year (n = 2).

Table 1.

Patient characteristics

Sex
Male	n = 8
Female	n = 13
Histology
Grade II oligodendroglioma	n = 15 (70%)
Grade II oligoastrocytoma	n = 4 (20%)
Grade II astrocytoma	n = 2 (10%)
Type of surgery
Biopsy	n = 19 (90%)
Partial resection	n = 2 (10%)
Delay between diagnosis and PCV onset
Median (range)	3.8 years (0.1–13)
Age at PCV onset
Median (range)	47 years (28–60)
Mean tumor diameter at PCV onset
Median (range)	67 mm (40–90)
Karnofsky index at PCV onset
Median (range)	80 (50–100)
Contrast enhancement at PCV onset 1p19q codeletion	n = 3/21 (14%)
Yes	n = 4
No	n = 2

Median MTD growth rate before PCV onset was assessable in 13 patients who had been followed prior to PCV initiation and was 5.5 mm/year (range, 2.2− 21.4 mm/year) (Fig. 1). During PCV treatment, the MTD decreased in all patients, with a median decrease of −10.2 mm/year (range, −1 to −23 mm/year). After PCV discontinuation, a persistent decrease in MTD was observed in 20 of the 21 patients, with a median decrease of −4 mm/year (range, −1.2 to −15.4 mm/year), which was significantly lower than the decrease observed during PCV (P = .0006) (Fig. 2). The median duration of the MTD decrease was 3.4 years (range, 0.8–7.7) after PCV onset and 2.7 years (range, 0–7) after the end of PCV. Median survival of the 21 patients after PCV onset was 7 years.

Fig. 1.

Evolution of the mean tumor diameter (MTD) before, during, and after PCV chemotherapy. The evolution of the MTD of each patient before (n = 13), during (n = 21), after (n = 21), and at progression (n = 14) is shown in blue, green, red, and yellow, respectively. During PCV chemotherapy, a decrease in the MTD (green curves) was observed in all patients. After PCV chemotherapy discontinuation, except in one patient (the patient with the smallest MTD at PCV onset), a persistent decrease in the MTD was observed in all patients (red curves).

Fig. 2.

Examples of the ongoing decrease in LGG volume in 2 patients observed after PCV discontinuation. Evolution of MRI scans in 2 patients: at diagnosis (A and E), before PCV onset (B and F), at the end of PCV treatment (C and G), and at the time of maximal response (D and H), with corresponding evolution of the MTD and MTD growth rate (mm/year). The patient on the left (A, B, C, and D) received only 2 cycles because of hematological toxicity; the patient on the right (E, F, G, and H) received 6 cycles. The time period in gray corresponds to the duration of PCV chemotherapy. On the left curve, a proposed mechanism to explain the persistent volume decrease following the discontinuation of PCV chemotherapy is shown.

In the 4 patients who received less than 5 PCV cycles because of hematological toxicity (1 received 2 cycles, 2 received 3 cycles, and the last received 4 cycles), the duration of the MTD decrease after PCV onset was 3.3, 7.6, 1.5, and 4 years, respectively.

According to the modified McDonald's criteria, the rates of partial and minor responses were 5% and 38% at the end of PCV but 38% and 42% at the time of maximal MTD decrease, which occurred after a median period of 2.7 years after the end of PCV and 3.4 years after PCV onset (Table 2). Because of insufficient material or because specimens had been fixed in Bouin's solution (precluding DNA analysis), LOH could be assessed in only 6 patients. Among the 4 patients with 1p/19q codeletion, 1 patient progressed (with the appearance of a contrast enhancement) just after completing the 6 PCV cycles, but in the 3 others, the MTD was still decreasing 3.8, 4.8, and 7.5 years after PCV onset. In the 2 patients without 1p/19q codeletion, the MTD decrease lasted 2 and 3 years before progression.

Table 2.

Response rates according to McDonald's criteria at the end of PCV treatment and at the time of maximal mean tumor diameter (MTD) decrease

	Response rates at the end of PCV treatment (PR/MR/SD)	Response rates at maximal MTD decrease (PR/MR/SD)	Time to maximal MTD decrease after PCV onset (median)
All patients (n = 21)	5%/38%/57%	38%/42%/20%	3.4 years (0.8–7.7)
Oligodendrogliomas (n = 15)	7%/40%/53%	47%/33%/20%	3.2 years (0.8–4.8)
Oligoastrocytomas (n = 4)	0%/50%/50%	25%/75%/0%	5 years (2–7.7)
Astrocytomas (n = 2)	0%/0%/100%	0%/50%/50%	NA (3–7.5)
1p19q codeletion (n = 4)	0%/50%/50%	25%/50%/25%	NR (0.8–7.5)
No 1p19q codeletion (n = 2)	0%/100%/0%	100%/0%/0%	NA (2–3)

Abbreviations: PR, partial response; MR, minor response; SD, stable disease; NA, not available because of the small sample size; NR, not reached.

Discussion

Several studies have demonstrated that PCV chemotherapy is an effective regimen for the treatment of LGGs, but a dynamic evaluation of its impact on LGG growth kinetics has not been made. The present study has several limitations. It is based on a small number of patients, 1p19q codeletion status was available in only 28% of the patients, and because of its retrospective design, imaging was not performed on the same regular interval in all patients. However, it provides the first evaluation of LGGs growth kinetics during and after PCV treatment. During PCV treatment, we observed a decrease in the MTD in all patients in a manner very similar to that observed after temozolomide onset and with a similar slope of MTD decrease (−10.2 mm/year in this study and −9.2 mm/year with temozolomide).⁴ No MTD regrowth was observed during PCV treatment. These results are in agreement with the very low rates of progression that were previously reported during treatment with PCV (1 of 16 patients⁸ and 1 of 28 patients⁵) but conflict with the reported rate of MTD regrowth during temozolomide treatment (one-third of the patients after a median period of 1 year).⁴ More strikingly, after PCV discontinuation, we observed an ongoing and prolonged (>2 years) MTD decrease in the majority of patients (60%). This is in marked contrast to the evolution of the MTD reported after temozolomide discontinuation.⁴ Indeed, when temozolomide was discontinued in the absence of progression (after a median period of 18 cycles), 60% of tumors resumed their growth within 1 year.⁴ Thus, this study demonstrates that patients treated with PCV can achieve prolonged and persistent responses, despite a shorter duration of treatment, than those treated with temozolomide. These results challenge the idea that a prolonged duration of chemotherapy is necessary for treating LGGs. Strikingly, 3 of the 4 patients who received 4 or fewer PCV cycles also achieved long-lasting responses. Because of the low percentage of patients (28%) for whom 1p/19q status was available, we could not assess whether 1p/19q codeletion was significantly associated with a longer duration of MTD decrease, as reported in patients treated with temozolomide.⁴ This is, however, very likely, as in 3 of the 4 patients with 1p/19q codeletion, the MTD was still decreasing more than 3.5 years after PCV onset.

Though it was previously mentioned in some patients,^7,8 the ongoing and prolonged decrease in LGGs volume following the PCV chemotherapy discontinuation was not recognized as a common phenomenon until now. This observation, however, has important consequences on response assessment and might in part explain the different response rates previously reported with PCV chemotherapy (from 29% up to 77%).^5–10 As shown here, the rate of partial response was only 5% at the end of PCV but 38% at the time of maximal response, which occurred after a median period of 3.4 years after PCV onset.

Whether PCV should be preferred over temozolomide for treating LGGs cannot be assessed from the present study. Furthermore, even if it was acceptable in this series, PCV toxicity has been well established.¹² Further studies are needed to validate prospectively the results of the present study and to assess whether responses are more prolonged with PCV than with temozolomide. It also remains to be seen whether chemosensitive LGGs, namely LGGs with 1p/19q codeletion, might achieve prolonged responses whatever the chemotherapy regimen, even with shorter durations of chemotherapy, which would have the advantage of reducing the risk of long-term toxicity.

Finally, the mechanisms underlying the persistent and prolonged MTD decrease in LGGs once chemotherapy is terminated are intriguing. We are not aware of another example of a tumor achieving a persistent ongoing volume decrease once chemotherapy is stopped. One hypothesis explaining these observations is that LGGs would consist of 2 main types of tumor cells: one subpopulation of proliferating cells and one population of nonproliferating daughter cells, which would undergo apoptosis after a determined period. In the absence of treatment, the growth rate of the proliferating cells would be higher than the apoptosis rate of the daughter cells. Chemotherapy would act by producing durable inhibition of the proliferating subpopulation. Thus, the natural apoptosis rate of the tumor would be unmasked, explaining why a persistent tumor volume decrease can be observed despite chemotherapy no longer being administered.