A randomized phase III trial comparing induction chemotherapy followed by chemoradiotherapy versus chemoradiotherapy alone as treatment of unresectable head and neck cancer

BACKGROUND

Concurrent chemoradiotherapy (CCRT) is the standard treatment for patients with unresectable, nonmetastatic locoregionally advanced squamous-cell carcinoma of the head and neck (LASCCHN). This randomized, open-label, phase III clinical trial compared the efficacy between standard CCRT and two different induction chemotherapy (ICT) regimens followed by CCRT.

PATIENTS AND METHODS

Patients with untreated LASCCHN were randomly assigned to ICT (three cycles), with either docetaxel (Taxotere), cisplatin and 5-fluorouracil (TPF arm) or cisplatin and 5-fluorouracil (PF arm), followed by CCRT [7 weeks of radiotherapy (RT) with cisplatin 100 mg/m(2) on days 1, 22 and 43]; or 7 weeks of CCRT alone. The primary end points were progression-free survival (PFS) and time-to-treatment failure (TTF).

RESULTS

In the intention-to-treat (ITT) population (n = 439), the median PFS times were 14.6 (95% CI, 11.6-20.4), 14.3 (95% CI, 11.8-19.3) and 13.8 months (95% CI, 11.0-17.5) at TPF-CCRT, PF-CCRT and CCRT arms, respectively (log-rank P = 0.56). The median TTF were 7.9 (95% CI, 5.9-11.8), 7.9 (95% CI, 6.5-11.8) and 8.2 months (95% CI, 6.7-12.6) for TPF-CCRT, PF-CCRT and CCRT alone, respectively (log-rank P = 0.90). There were no statistically significant differences for overall survival (OS). Toxic effects from ICT-CCRT were manageable.

CONCLUSION

Overall, this trial failed to show any advantage of ICT-CCRT over CCRT alone in patients with unresectable LASCCHN.

A phase II multicentric study of sunitinib administered as upfront therapy in glioblastoma: A study by the GEINO group

2045

Background: Sunitinib is a multitargeted tyrosine kinase inhibitor that has direct antitumor and antiangiogenesis activity through targeting VEGFR 1-2, PDGFR α-β, c-kit, bFGF, (CSF-1), FLT3 and RET. Experimental studies in GB mice showed angiogenic activity, prolonged survival and synergy with irradiation Methods: Eligible pts were ≤75 years with GB not amenable to resection, PS≥2, MMS ≥25. Treatment was sunitinib 37.5mg daily for 8 w before radiotherapy, during radiotherapy (60Gy, 6 w) and after radiotherapy until disease progression. Primary endpoint was overall response (OR) (RANO criteria) after 8 w of sunitinib treatment before radiotherapy. Secondary endpoints were % pts free of neurological deterioration before radiotherapy, % pts who completed radiotherapy, PFS, OS and 1-year survival. We used a Simon 2-stage design (12 →20) based on OR to calculate the number of patients needed to detect at least 10% of responses with α error of 0.05 and β error of 0.10. Twelve patients were included in the first phase and one response was needed to continue study accrual. Results: 12 pts were enrolled: 7 males, median age 65 years, PS≥1: 8pts, median MMS 26.5. Neurological deficit: 5pts, DXM treatment 10/12, non-EIDS 3/12. Treatment: 5 pts completed 8 w of therapy, 2pts completed 7 w, 4 pts ≤ 6w. Toxicity G3 and 4: diarrhea (1pt), asthenia (2pts), skin and EDDPP (2), mucositis (2), fatal CNS hemorrhage (1), hypercholesterolemia and hypertriglycemia (1). OR was 1/12 SD, 11/12 P; radiological response was SD in 5 pts but 4 pts progressed neurologically. Only 2 pts completed radiotherapy plus sunitinib. The only SD received treatment for 18 w. Median PFS was 7.5w (CI 95% 5.8-9.2), OS 16w (CI 95% 0.5-23.7), 1year OS: 0%. Conclusions: the trial was closed after the first stage due to lack of response. Sunitinib is an inactive agent in glioblastoma.

Impact of antiepileptic drugs on thrombocytopenia in glioblastoma patients treated with standard chemoradiotherapy

Epilepsy in glioblastoma multiforme (GBM) patients is common. Hematological toxicity is a potential side effect of antiepileptic drugs (AEDs) and a frequent limiting-dose effect of temozolomide (TMZ). The aim of the study was to investigate the impact of AEDs on thrombocytopenia in GBM patients treated with radiotherapy and TMZ. A cohort of 101 newly diagnosed GBM patients treated with radiotherapy and TMZ was reviewed. Clinical data, presence of seizures, AEDs use, platelet count, and accumulated TMZ dose were analyzed at each cycle. Thrombocytopenia was operationalized as a continuous platelet count and a dichotomic variable (cut-off <100.000/mm(3)). This cut-off represents the threshold beyond which TMZ treatment is modified. A linear and a probit pooled cross-sectional regression analysis were used to study the impact of age, gender, AEDs, and accumulated TMZ on thrombocytopenia. Impact of AEDs on survival was also analyzed. Thirty-five patients (35%) presented seizures at onset and 18 (27%) during follow-up. Seven (13%) needed two or more AEDs for seizure control. Grade 3-4 thrombocytopenia was found in 8%. Decrease in platelet count was related to accumulated TMZ (p < 0.001), age (p < 0.001), and valproate (p = 0.004). Platelet count <100.000/mm(3) was only associated with accumulated TMZ (p = 0.001). Recursive Partitioning Analysis prognostic class was the only variable with significant impact on survival. Valproate and age had an independent negative effect on total platelet count, although neither had an effect on critical thrombocytopenia (<100.000/mm(3)). Therefore, the systematic withhold of valproate in GBM patients might not be justified. Nevertheless, this negative effect may be taken into account especially in elderly patients.

Does gender matter in glioblastoma?

BACKGROUND The clinical outcome of glioblastoma (GBM) patients who receive radiotherapy alone or with chemotherapy is well established. However, little is known about how many patients do not receive this treatment. We consider it is important to investigate why a proportion of operated patients do not receive further treatment after surgery. METHODS We reviewed all consecutive GBM patients operated on in our hospital between January 2000 and December 2008. RESULTS A total of 216 patients with GBM were identified. Fifty-five (25%) did not receive any treatment after surgery. Univariate analysis showed that factors associated with no further treatment after surgery were older than 60 years (p=0.002), of female gender (p=0.03), had a KPS<70 (p<0.001) and had had a biopsy (p<0.001). Multivariate analysis indicated that age =60 years and KPS <70 were independent predictors of no further treatment after surgery. Gender was not an independent variable. However, women in the whole series were older than 60 years (p=0.01), and they had a worse KPS (p=0.02) and more biopsies (p=0.04) than men. In the whole group, median survival time was 10.4 months for men (n=125) vs. 7.2 months for women (n=91), log rank p<0.04. This difference was not observed in the group that was treated after surgery. CONCLUSIONS One out of four patients could not be treated after surgery. Independent predictors were older age and low KPS. These poor risk variables were more frequent in women and their survival was therefore lower than men in our series.

[Implication of radiological pattern in the prognosis of oligodendroglial tumors: correlation with genetic profile]

INTRODUCTION

1p19q loss of heterozygosity (LOH1p19q) in oligodendroglial tumors has shown to be prognostic of prolonged survival and predictive of therapeutic responsiveness. During the last years, research is actively being directed to the discovery of radiological characteristics related to LOH1p19q. AIMS. To confirm the existence of molecular heterogeneity in oligodendroglial tumors in relation to their anatomic distribution, and to evaluate the correlation between molecular profile and other radiological and clinical characteristics and their prognostic impact.

PATIENTS AND METHODS

Fifty-four patients with oligodendroglial tumors managed according to a previously established protocol were included. Preoperative SE T1, T1 post-gadolinium and T2 magnetic resonance images were reviewed by two independent neuroradiologists, blinded to clinical and molecular information. LOH analysis was assessed from paired tumor-blood DNA acid samples.

RESULTS

LOH1p was highly associated with LOH19q (p < 0.0001), LOH1p (odds ratio, OR = 6.19; 95% confidence interval, 95% CI = 1.66-22.68; p = 0.004), LOH19q (OR = 7.59; 95% CI = 1.84-31.34; p = 0.006) and LOH1p19q (OR = 5.38; 95% CI = 1.51-19.13; p = 0.007) were found to be more frequent in tumors located in the frontal lobe. Frontal location (hazard ratio, HR = 4.499; 95% CI = 1.027-193.708; p = 0.046), ring enhancement (HR = 0.213; 95% CI = 0.065-0.700; p = 0.011) and extent of resection (HR = 9.231; 95% CI = 1.737-49.050; p = 0.009) resulted independent prognostic factors for overall survival in the multivariate analysis.

CONCLUSIONS

Glioma classification aims to better define patients prognosis. Besides histological and immunohistochemical analyses, molecular information has become of great importance. Our results indicate that the evaluation of some MR features may also be useful. Efforts must be directed toward the use of every available resource at each institution.

Weekly paclitaxel for platin-resistant stage IV head and neck cancer patients

CONCLUSIONS

Weekly paclitaxel may be an active and well tolerated chemotherapy regimen for patients with platin-resistant advanced head and neck cancer.

OBJECTIVES

Weekly paclitaxel should be an active and well tolerated regimen for palliative treatment of platin-resistant patients with recurrent or metastatic carcinoma of the head and neck. We analyzed the antitumor activity and toxicity profile.

PATIENTS AND METHODS

Sixty consecutive patients with advanced head and neck cancer were treated with 1 h infusion of paclitaxel, 80 mg/m(2) weekly, for 6 consecutive weeks. Patients who showed disease response or disease stabilization continued until progression of disease.

RESULTS

A total of 719 doses of paclitaxel were administered to the 60 patients. No complete response was observed. Partial response and stable disease were observed in 26 (43.3%) and 9 (15%) patients, respectively. Median time to tumor progression for patients who responded to therapy was 6.2 months (SD=1.3; 95% CI, 3.7-8.6) and the overall median survival in this group of patients was 8.5 months (SD=1.4; 95% CI, 5.7-11.2). The main toxic effects were leukopenia (26.6%), anemia (43.3%), fatigue (37.4%), alopecia (18.7%), rash/desquamation (13.3%), and thrombophlebitis (6.8%).

Actual proportion of patients (pts) receiving chemotherapy or cetuximab for head and neck squamous carcinoma (HNSC)

e17058

Background: With the new indications of chemotherapy or cetuximab in HNSC, the rate of pts receiving these therapies nowadays is unclear. Methods: This retrospective study identified all consecutive pts with HNSC from January 1, 2006, to December 31, 2007, presented in a multidisciplinary team to decide further treatment in a single institution. ASCO guidelines for larynx preservation were followed to select surgery or chemoradiotherapy (ChRt). We classified the intention-to-treat as palliative, adjuvant or induction therapy. In the last case, always with concomitant radiotherapy (Rt) or prior to concomitant ChRt. Cetuximab was indicated with Rt as induction therapy for pts with problems to receive platin-based chemotherapy. Results: : A total of 350 pts were identified, 320 were male (91%), and 30 female (9%), with mean age 60.4 (range 41–90). Primary tumor was located in glottis (41%), supraglottis (19%), hypopharynx (11%), oropharynx (20%), or mouth (9%). Staging was I (27%), II (22%), III (16%), or IV (35%). Surgery alone was performed in 136 pts (39%) and chemotherapy or cetuximab in 214 other pts (61%). The intention-to-treat was palliative in 69 (32%), adjuvant in 51 (24%), or induction 94(44%) of the pts respectively. Rt plus cetuximab was administered to 31/97 (33%) and Rt plus chemotherapy in 63/97 (67%) pts as induction therapy. During this 2-year period, some pts received both induction/adjuvant and palliative chemotherapy. Conclusions: Chemotherapy or cetuximab is indicated as part of treatment in more than a half of pts with HNSC. Induction therapy is the most frequent indication.

No significant financial relationships to disclose.

Single nucleotide polymorphisms (SNPs) analysis of CYP2C8, GSTT1, GSTP1, MDR1(A), MDR1(B) and ERCC1 as predictor of survival after weekly paclitaxel for relapsed advanced head & neck cancer patients (AHNCP)

2540

Background: Gene SNPs correlate with survival in cancer patients (pts) treated with chemotherapy (CHM). CYP2C8 and GSTT1, GTSP1 genes are involved in phase 1 and 2 drug cellular metabolisms respectively; MDR1(A) and MDR1(B) are involved in drug membrane transport and ERCC1 in DNA repair Methods: We evaluated the presence of SNPs of these 6 genes and the survival of AHNCP treated with weekly paclitaxel, 80 mg/m2 iv for 6 weeks. Responding pts continue CHM till progression. All pts were cisplatin resistant and no other local therapies were available. We analysed paraffin-embedded biopsies from 47 consecutive AHNCP for SNPs of the mentioned genes. The status of the alleles wild type (wt) or at least 1 SNP was compared with response rate (RR), time to progression (TTP) and overall survival (OS) Results: Of 47 pts, 43 were male and 4 female. The median of age was 57 yr (46–80). RR was 45% (21/47) and the TTP in responders was 5 months of median. OS for all pts was 5.6 months. Wild type vs at least 1 SNPs frequencies according the genes were: CYP2C8 23/24; GSTT1 45/2; GSTP1 36/11; MDR1(A) 21/28; MDR1(B) 13/34; and ERCC1 27/20. OS was significantly better in pts with 2 or more SNPs accumulated (p=0.0455). No other significant differences were observed in RR, TTP or OS in SNPs vs wild type pts. Conclusions: SNPs of CYP2C8, MDR1(A) and MDR1(B) genes were more frequent than wt in our pts. OS was significantly better in pts with 2 or more SNPs accumulated. Paclitaxel provides high rate of responses of short duration in AHNC pts

No significant financial relationships to disclose.

Clinical outcome in patients with intramedullary spinal cord metastases from lung cancer

Intramedullary spinal cord metastases (ISCM) are uncommon and present with rapidly progressing neurological deficits. The objective of this study was to determine the rate, duration of neurological response and survival after radiation therapy. We have retrospectively reviewed the clinical outcome of six cases with a diagnosis of ISCM from primary lung cancer, non-small cell (NSCLC) (n=3) and small cell (SCLC) (n=3). Total radiation dose ranged from 27 Gy/5 fr to 40 Gy/20 fr. Ambulation was preserved in 3 patients and partially recovered in one. Five out of the six patients (83%) showed improvement in neurological signs/symptoms with a mean duration of 17.2 days (max: 40 days; min: 6 days). Median survival time was 5 months (confidence interval (CI) 95%: 0-12) for NSCLC and 5 months (CI 95%: 4-6) for SCLC. Although radiation response rate is high, the interval free of neurological progression is very short. A therapeutic approach should be considered for each individual.

Low dose rate brachytherapy in lip carcinoma

BACKGROUND

Lip cancer is frequently treated with surgery although radiation therapy offers comparable results. The aim of the study was to evaluate the local cure rate in patients with lip carcinoma treated with 192-Ir low dose rate interstitial brachytherapy.

METHODS

Fifty-four patients with a mean age of 70 years (range, 40-90 years) were retrospectively evaluated. The tumour location was the superior lip in 4 (7.4%) and the inferior lip in 50 (92.6%). Tumour stage was T1N0 in 33 patients and T2N0 in 21 patients. The radioactive sources with hypodermic needles in 49 patients (90.7%) and plastic tubes in 5 (9.3%) were placed parallel and equidistant from one another across the tumour volume according to the Paris system rules.

RESULTS

The median dose was 61.5 Gy (range, 60-65 Gy). All patients experienced acute brisk skin and mucositis RTOG grade III around the implanted volume, subsiding within 4-6 weeks after the implant. Local control was achieved in 98% of patients. The mean follow-up was 7 years.

CONCLUSIONS

Low dose rate interstitial brachytherapy with 192-Iridium is a well established and efficacious way to achieve local control of the tumour in lip cancer. It offers the advantage of avoiding surgery in an elderly population.

Prognostic Significance ofO6-Methylguanine-DNA Methyltransferase Determined by Promoter Hypermethylation and Immunohistochemical Expression in Anaplastic Gliomas

PURPOSE

Anaplastic gliomas constitute a heterogeneous group of tumors with different therapeutic responses to adjuvant chemotherapy with alkylating agents. O6-Methylguanine-DNA methyltransferase (MGMT), a DNA repair protein, is one of the implicated factors in glioma chemoresistance. The prognostic value of MGMT remains controversial due in part to the fact that previous published studies included heterogeneous groups of patients with different tumor grades. The aim of this study was to evaluate the prognostic significance of MGMT in patients with anaplastic glioma.

EXPERIMENTAL DESIGN

Ninety-three patients with anaplastic glioma were analyzed for MGMT protein expression by immunohistochemistry. In addition, for those patients from whom a good yield of DNA was obtained (n = 40), MGMT promoter methylation profile was analyzed by methylation-specific PCR. MGMT prognostic significance was evaluated together with other well-known prognostic factors.

RESULTS

Fifty-one tumors (54.8%) showed nuclear staining of MGMT. There was a trend towards longer overall survival for those patients with negative MGMT immunostaining (hazard ratio, 1.66; P = 0.066). In a secondary analysis including those patients who actually received chemotherapy (n = 72), the absence of MGMT expression was independently associated with better survival (hazard ratio, 2.12; P = 0.027). MGMT promoter methylation was observed in 50% of the analyzed tumors. No statistical correlation between MGMT expression and MGMT promoter hypermethylation was observed.

CONCLUSIONS

Unlike previous studies, we did not find a correlation between MGMT promoter methylation and survival. However, we observed a correlation between MGMT protein expression and survival in those patients who received chemotherapy thus suggesting that the absence of MGMT expression is a positive predictive marker in patients with anaplastic glioma.

Primary non-Hodgkin's lymphoma of the CNS treated with CHOD/BVAM or BVAM chemotherapy before radiotherapy: long-term survival and prognostic factors

PURPOSE

To assess the long-term survival and prognostic factors associated with the cyclophosphamide, doxorubicin, vincristine, and dexamethasone (CHOD)/carmustine, vincristine, methotrexate, and cytarabine (BVAM) and BVAM chemotherapy regimens followed by cranial radiotherapy in the treatment of primary central nervous system (CNS) non-Hodgkin lymphoma.

METHODS AND MATERIALS

Since 1986, high-dose methotrexate (1.5 g/m(2)), cytarabine, vincristine, and carmustine have been used in the BVAM chemotherapy regimen for primary CNS non-Hodgkin's lymphoma, with one cycle of CHOD given before BVAM in patients <or=70 years of age from 1990 onward.

RESULTS

The median age for the 77 patients treated was 60 years and the median follow-up of surviving patients was 3 years (range, 1.4-15.2 years). The complete response rate after chemotherapy was 62% and after additional radiotherapy was 73%. Multivariate analysis of overall survival showed age (p = 0.004), performance status (p = 0.007), and number of tumors (unifocal disease vs. multifocal disease; p = 0.005) to be statistically significant prognostic factors. Survival decreased with increasing age and performance status score. Using a prediction score giving 1 point for each adverse prognostic factor (age >or=60 years, performance status >or=2, and multifocal and/or meningeal disease [advanced stage]), a score of 0 (8 patients) was associated with a median survival of 55 months, a score of 1 (29 patients) of 41 months, a score of 2 (28 patients) of 32 months, and a score of 3 (12 patients) a median survival of 1 month (p <0.001). The actuarial overall survival for the 35 patients aged <60 years was 32.4% (95% confidence interval, 14.1-50.8%) at 10 years.

CONCLUSION

The Nottingham/Barcelona prediction score, which is similar to the International Prognostic Index, may be useful in comparing survival with different regimens studied in Phase II trials. Patients with primary CNS non-Hodgkin's lymphoma aged <60 years treated with CHOD/BVAM or BVAM followed by radiotherapy have a similar long-term survival to that of patients with large B cell non-Hodgkin's lymphoma at other extranodal sites.

Multidisciplinary approach in advanced cancer of the oral cavity: outcome with neoadjuvant chemotherapy according to intention-to-treat local therapy. A phase II study

OBJECTIVES

To determine outcomes in local-regional control and overall survival in patients with squamous locally advanced cancer of the oral cavity, based on intention-to-treat with neoadjuvant chemotherapy followed by surgery or radiation therapy.

METHODS

Two hundred and four out of 1,089 patients analyzed met the defined criteria. All had squamous cell carcinomas of the oral cavity in stage III or in nonmetastatic stage IV and were selected for surgery or radiation therapy (if located in the tonsils or in the base of the tongue). Chemotherapy was based on cisplatin 120 mg/m(2) i.v. day 1 plus bleomycin 20 mg/m(2) days 1-5 in continuous i.v. perfusion or plus 5-fluorouracil 1,000 mg/m(2) days 1-5 in continuous i.v. perfusion. A total of 418 cycles were given to 204 patients (mean 2.049 per patient). Definitive surgery (n = 73; plus adjuvant radiation therapy) or definitive radiation therapy (n = 131) was performed.

RESULTS

One hundred thirty-five out of 204 (66%) patients were chemotherapy responders, 16% complete and 50% partial. One hundred ninety-four patients (95%) completed 2 courses of chemotherapy. After neoadjuvant chemotherapy, 34 out of 46 patients considered inoperable initially (74%) obtained a disease-free status with surgery. Eighty-three percent of surgical patients obtained a disease-free status (initial tumor control) versus 72% of radiation therapy patients. Disease-free survival rates at 5 years were 26 and 22%, respectively. A better prognosis was observed in stage III over IV (p = 0.02); primary tumor in the retromolar trigone, palate or buccal mucosa over tongue, tonsil or floor of the mouth (p = 0.0085); negative cervical nodes over positive (p = 0.0186); responders to chemotherapy over nonresponders (p = 0.0003); and adjuvant postsurgical radiation therapy (p = 0.0013). Causes of death were relapses in local area (86%), regional nodes (10.5%) or distant metastases (3.5%). Eleven patients (5%) died of a second primary. The main toxic effects were vomiting in 9% of patients and hemolytic-uremic syndrome in 3% of the patients treated with bleomycin.

CONCLUSIONS

In locally advanced squamous cell carcinoma of the oral cavity, neoadjuvant chemotherapy induces a high response rate that may facilitate definitive surgery or radiotherapy. In this study, patients have an acceptable long-term survival.

Pre-irradiation semi-intensive chemotherapy with carboplatin and cyclophosphamide in malignant glioma: a phase II study

We undertook a phase II trial in 17 patients with malignant glioma and large measurable disease to assess response rate and survival with pre-irradiation chemotherapy, using higher doses than standard, trying to improve the outcome. Patients characteristics were: male/female 10/7, age 49 (range 23-59), median Karnofsky index 90% (range 70-100), glioblastoma multiforme/anaplastic astrocytoma 14/3. Treatment consisted of 2 cycles of carboplatin 200 mg/m(2) days 1-3 (or AUC x 8, total dose) plus cyclophosphamide 1000 mg/m(2) days 1-3. One partial response (6.5%) and two stabilizations (13.5%) were observed after pre-irradiation chemotherapy. Twelve out of 15 patients (80%) progressed after chemotherapy. Median survival time was 7.6 months and the survival at 1 year was 33%. Main toxicity was hematologic in the first cycle: neutropenia grade 4 in 100%; thrombocytopenia grade 4 in 73% and grade 3 in 27%; anemia grade 3 in 7%; in the second cycle: neutropenia and thrombocytopenia grade 4 in 100% and anemia grade 3 in 50%). No toxic death was related to treatment. This regimen showed limited activity in malignant glioma with large residual disease after surgery or biopsy.

Importance of radiotherapy in the outcome of patients with primary CNS lymphoma: an analysis of the CHOD/BVAM regimen followed by two different radiotherapy treatments

PURPOSE

To assess the effect of a reduced dose of radiotherapy (RT) in patients with primary CNS lymphoma (PCNSL) responding to the cyclophosphamide, doxorubicin, vincristine, and dexamethasone (CHOD)/carmustine, vincristine, methotrexate, and cytarabine (BVAM) regimen.

PATIENTS AND METHODS

Patients received one cycle of CHOD and two of BVAM. In the first trial, all 31 patients received 45-Gy whole-brain RT (CHOD/BVAM I). In the second, with 26 patients, RT dose was reduced to 30.6 Gy if there was a complete response (CR) after chemotherapy (CHOD/BVAM II).

RESULTS

Age, performance status, and chemotherapy received were similar in both protocols. CR rate at the end of all treatment was 68% for CHOD/BVAM I and 77% and for CHOD/BVAM II. Treatment modality was the only predictor of relapse, with 3-year relapse risks of 29% and 70% for CHOD/BVAM I and II, respectively. This was specifically important in the 25 patients less than 60 years old (3-year relapse risk, 25% v 83%; P =.01). The 5-year overall survival (OS) was 36%. Age (< 60 v > or = 60 years) was the only predictor for OS in the multivariate analysis (relative risk, 2.1; 95% confidence interval, 1.4 to 2.8). RT dose was the only predictor of OS in patients younger than 60 years old who achieved CR at the end of all treatment (3-year OS, 92% v 60% for patients receiving 45 or 30.6 Gy, respectively; P =.04).

CONCLUSION

Reduction of the RT dose from 45 Gy to 30.6 Gy in patients younger than 60 years old with PCNSL who achieved CR resulted in an increased risk of relapse and lower OS.