Universitätsklinik für Strahlentherapie

BACKGROUND: In advanced prostate cancer (APC), successful drug development as well as advances in imaging and molecular characterisation have resulted in multiple areas where there is lack of evidence or low level of evidence. The Advanced Prostate Cancer Consensus Conference (APCCC) 2017 addressed some of these topics. OBJECTIVE: To present the report of APCCC 2017. DESIGN, SETTING, AND PARTICIPANTS: Ten important areas of controversy in APC management were identified: high-risk localised and locally advanced prostate cancer; "oligometastatic" prostate cancer; castration-naïve and castration-resistant prostate cancer; the role of imaging in APC; osteoclast-targeted therapy; molecular characterisation of blood and tissue; genetic counselling/testing; side effects of systemic treatment(s); global access to prostate cancer drugs. A panel of 60 international prostate cancer experts developed the program and the consensus questions. OUTCOME MEASUREMENTS AND STATISTICAL ANALYSIS: The panel voted publicly but anonymously on 150 predefined questions, which have been developed following a modified Delphi process. RESULTS AND LIMITATIONS: Voting is based on panellist opinion, and thus is not based on a standard literature review or meta-analysis. The outcomes of the voting had varying degrees of support, as reflected in the wording of this article, as well as in the detailed voting results recorded in Supplementary data. CONCLUSIONS: The presented expert voting results can be used for support in areas of management of men with APC where there is no high-level evidence, but individualised treatment decisions should as always be based on all of the data available, including disease extent and location, prior therapies regardless of type, host factors including comorbidities, as well as patient preferences, current and emerging evidence, and logistical and economic constraints. Inclusion of men with APC in clinical trials should be strongly encouraged. Importantly, APCCC 2017 again identified important areas in need of trials specifically designed to address them. PATIENT SUMMARY: The second Advanced Prostate Cancer Consensus Conference APCCC 2017 did provide a forum for discussion and debates on current treatment options for men with advanced prostate cancer. The aim of the conference is to bring the expertise of world experts to care givers around the world who see less patients with prostate cancer. The conference concluded with a discussion and voting of the expert panel on predefined consensus questions, targeting areas of primary clinical relevance. The results of these expert opinion votes are embedded in the clinical context of current treatment of men with advanced prostate cancer and provide a practical guide to clinicians to assist in the discussions with men with prostate cancer as part of a shared and multidisciplinary decision-making process.

This chapter contains sections titled: Introduction Deterministic Approach to Selection Error Threshold For Quasi-species Localization Example of Fitness Landscape and Stationary Populations Conclusion: Review and Outlook Mutation Freqencies and Hamming Distances

Leitlinienprogramm Onkologie der Arbeitsgemeinschaft der Wissenschaftlichen Medizinischen Fachgesellschaften e. V. (AWMF), Deutschen Krebsgesellschaft e. V. und Deutschen Krebshilfe

Background: Anaplastic thyroid carcinoma (ATC) and metastatic poorly differentiated thyroid carcinomas (PDTCs) are rare aggressive malignancies with poor overall survival (OS) despite extensive multimodal therapy. These tumors are highly proliferative, with frequently increased tumor mutational burden (TMB) compared with differentiated thyroid carcinomas, and elevated programmed death ligand 1 (PD-L1) levels. These tumor properties implicate responsiveness to antiangiogenic and antiproliferative multikinase inhibitors such as lenvatinib, and immune checkpoint inhibitors such as pembrolizumab. Patients and Methods: In a retrospective study, we analyzed six patients with metastatic ATC and two patients with PDTC, who received a combination therapy of lenvatinib and pembrolizumab. Lenvatinib was started at 14–24 mg daily and combined with pembrolizumab at a fixed dose of 200 mg every three weeks. Maximum treatment duration with this combination was 40 months, and 3 of 6 ATC patients are still on therapy. Patient tumors were characterized by whole-exome sequencing and PD-L1 expression levels (tumor proportion score [TPS] 1–90%). Results: Best overall response (BOR) within ATCs was 66% complete remissions (4/6 CR), 16% stable disease (1/6 SD), and 16% progressive disease (1/6 PD). BOR within PDTCs was partial remission (PR 2/2). The median progression-free survival was 17.75 months for all patients, and 16.5 months for ATCs, with treatment durations ranging from 1 to 40 months (1, 4, 11, 15, 19, 25, 27, and 40 months). Grade III/IV toxicities developed in 4 of 8 patients, requiring dose reduction/discontinuation of lenvatinib. The median OS was 18.5 months, with three ATC patients being still alive without relapse (40, 27, and 19 months) despite metastatic disease at the time of treatment initiation (UICC and stage IVC). All patients with long-term (>2 years) or complete responses (CRs) had either increased TMB or a PD-L1 TPS >50%. Conclusions: Our results implicate that the combination of lenvatinib and pembrolizumab might be safe and effective in patients with ATC/PDTC and can result in complete and long-term remissions. The combination treatment is now being systematically examined in a phase II clinical trial (Anaplastic Thyroid Carcinoma Lenvatinib Pembrolizumab [ATLEP]) in ATC/PDTC patients.

A novel procedure is proposed to extract T(1), T(2), and relative spin density from the signal time course sampled with a series of TrueFISP images after spin inversion. Generally, the recovery of the magnetization during continuous TrueFISP imaging can be described in good approximation by a three parameter monoexponential function S(t) = S(stst)(1-INV exp(-t/T(*) (1)). This apparent relaxation time T(*) (1) <or= T(1) depends on the flip angle as well as on both T(1) and T(2). Here, it is shown that the ratio T(1)/T(2) can be directly extracted from the inversion factor INV, which describes the relation of the signal value extrapolated to t = 0 and the steady-state signal. Analytical expressions are given for the derivation of T(1), T(2), and relative spin density directly from the fit parameters. Phantom results show excellent agreement with single point reference measurements. In human volunteers T(1), T(2), and spin density maps in agreement with literature values were obtained.

The metabolism and fate of tritiated thymidine were studied in two patients in hemopoietic equilibrium. Plasma clearance of H/sup 3/-thymidine commences in the first circulation time and becomes exponential following apparent equilibrium with total body water. Two components with halftimes of 0.2 minute and 1.0 minute were identified and discussed. The rapid plasma clearance of H/sup 3/-thymidine was associated with incorporation of this compound into newly formed DNA of proliferating cells as early as one minute after injection. Labeling of proliferative cells of the bone marrow was nearly complete within ten minutes after injection, and thereafter the label appeared to remain in these cells or their progeny for their life span, diluted only by successive mitoses. This short availability time of intravenous H/sup 3/-thymidine simplifies the analyses of the time parameters of labeled proliferating cell populations. About one-third of the H/sup 3/-thymidine was catabolized to THO within a few hours after injection. Small amounts of urine nonvolatile H/sup 3/ activity were excreted in the urine during the first day after injection. H/sup 3/- beta aminoisobutyric acid was identified in these urines as a catabolic product of H/ sup 3/-thymidine. Degradation and possible reutilization of the H/sup 3/ label is discussed.more » Severe malnutrition resulted in greater degradation of H/sup 3/ thymidine to THO and urine nonvolatile H/sup 3/ activity. (auth)« less

Complex corrective osteotomies in the lower extremities require precise preoperative planning. Fifty patients (37 male, 13 female) with an average age of 31 years (13 to 61 years) who had suffered fractures of the lower limbs and had been treated by osteosynthetic or conservative methods were studied, using a GE 9800 Quick CT; accurate and reproducible measurements of the angles of torsion of the femur and tibia were obtained. Digital images were produced to standardise the planes of measurement and to measure the length of the limb. The normal extremities of patients older than 18 years showed internal torsion of -20.4 +/- 9 degrees of the femur and external torsion of 33.1 +/- 8 degrees of the tibia. The most important clinical measurement is the intra-individual difference of the torsional angles. Amongst normals this is 4.3 +/- 2.3 degrees in the femur and 6.1 +/- 4.5 degrees in the tibia. Consequently, only angles greater than 9 degrees in the femur and 15 degrees in the tibia should be regarded as abnormal. Radiation exposure was measured by a LiF-thermoluminescence dosimeter on an Alderson phantom. Skin dose was 6.3 +/- 1.2 mGy and gonadal dose for females was 2.5 +/- 0.3 mGy and for males 0.7 +/- 0.1 mGy.

AIM: The standard treatment for patients with clinically resectable rectal cancer is surgery. Postoperative radiochemotherapy (RCT) is recommended for advanced disease (pT3/4 or pN+). In recent years, encouraging results of pre-operative radiotherapy have been reported. This prospective randomized phase-III-trial (CAO/ARO/AIO-94) compares the efficacy of neoadjuvant RCT to standard postoperative RCT. We report on the design of the study and first results with regard to toxicity of RCT and postoperative morbidity. PATIENTS AND METHODS: Patients with locally advanced operable rectal cancer (uT3/4 or uN+, Mason CS III/IV) were randomly assigned to pre or postoperative RCT: A total dose of 50.4 Gy (single dose 1.8 Gy) was applied to the tumour and the pelvic lymph nodes. 5-FU (1000 mg/m2/d) was administered concomitantly in the 1th and 5th week of radiation as 120 h-continuous infusion. Four additional cycles of 5-FU-chemotherapy (500 mg/m2/d, i.v.-bolus) were applied. RCT was identical in both arms except for a small-volume boost of 5.4 Gy postoperatively. The time interval between RCT and surgery was 4-6 weeks in both arms. Techniques of surgery were standardized and included total mesorectal excision. Primary endpoints of the study are 5-year survival and local and distant control. Secondary endpoints include the rate of curative (R0) resection and sphincter saving procedures, toxicity of RCT, surgical complications and quality of life. RESULTS: As of July 2002, 805 patients were randomized from 26 participating institutions. Acute toxicity (WHO) of RCT was low, with less than 15% of patients experiencing grade 3 or higher toxicity: The principal toxicity was diarrhea, with 12% in the postoperative RCT-arm and 11% in the pre-operative RCT-arm having grade 3-, and 1% in either arm having grade 4-diarrhea. Erythema, nausea and leukopenia were the next common toxicities, with less than 3% of patients in either arm suffering grade 3 or greater leukopenia or nausea. Postoperative complication rates were similar in both arms, with 12% (postop. RCT) and 12% (pre-op. RCT) of patients, respectively, suffering from anastomotic leakage, 3% (postop. RCT) and 3% (pre-op. RCT) from postoperative bleeding, and 6% (postop. RCT) and 4% (pre-op. RCT) from delayed wound healing. CONCLUSION: The patient accrual to the trial is satisfactory. Neoadjuvant RCT is well tolerated and bears no higher risk for postoperative morbidity.

Bei diesem Dokument handelt es sich um die aktualisierte Fassung der 2006 erstmals erstellten S3-Leitlinie zum exokrinen Pankreaskarzinom.

A variety of detectors and procedures for the measurement of small field output factors are discussed in the current literature. Different detectors with or without corrections are recommended. Correction factors are often derived by Monte Carlo methods, where the bias due to approximations in the model is difficult to judge. Over that, results appear to be contradictory in some cases. In this work, output factors were measured for field sizes from 4 mm up to 180 mm side length with different detectors. A simple linear correction for the energy response of solid state detectors is proposed. This led to identical values down to 8 mm field size, as long as the size of the detector is small against the field size. The correction was of the order of a few percent. For a shielded silicon diode it was well below 1%. A physically meaningful function is proposed in order to calculate output factors for arbitrary field sizes with high accuracy.

PURPOSE: One of the major hurdles in enabling personalized medicine is obtaining sufficient patient data to feed into predictive models. Combining data originating from multiple hospitals is difficult because of ethical, legal, political, and administrative barriers associated with data sharing. In order to avoid these issues, a distributed learning approach can be used. Distributed learning is defined as learning from data without the data leaving the hospital. PATIENTS AND METHODS: Clinical data from 287 lung cancer patients, treated with curative intent with chemoradiation (CRT) or radiotherapy (RT) alone were collected from and stored in 5 different medical institutes (123 patients at MAASTRO (Netherlands, Dutch), 24 at Jessa (Belgium, Dutch), 34 at Liege (Belgium, Dutch and French), 48 at Aachen (Germany, German) and 58 at Eindhoven (Netherlands, Dutch)). A Bayesian network model is adapted for distributed learning (watch the animation: http://youtu.be/nQpqMIuHyOk). The model predicts dyspnea, which is a common side effect after radiotherapy treatment of lung cancer. RESULTS: We show that it is possible to use the distributed learning approach to train a Bayesian network model on patient data originating from multiple hospitals without these data leaving the individual hospital. The AUC of the model is 0.61 (95%CI, 0.51-0.70) on a 5-fold cross-validation and ranges from 0.59 to 0.71 on external validation sets. CONCLUSION: Distributed learning can allow the learning of predictive models on data originating from multiple hospitals while avoiding many of the data sharing barriers. Furthermore, the distributed learning approach can be used to extract and employ knowledge from routine patient data from multiple hospitals while being compliant to the various national and European privacy laws.

Dieses Dokument wurde zum persnlichen Gebrauch heruntergeladen. Vervielfltigung nur mit Zustimmung des Verlages.

Stereotactic radiotherapy with its forms of intracranial stereotactic radiosurgery (SRS), intracranial fractionated stereotactic radiotherapy (FSRT) and stereotactic body radiotherapy (SBRT) is today a guideline-recommended treatment for malignant or benign tumors as well as neurological or vascular functional disorders. The working groups for radiosurgery and stereotactic radiotherapy of the German Society for Radiation Oncology (DEGRO) and for physics and technology in stereotactic radiotherapy of the German Society for Medical Physics (DGMP) have established a consensus statement about the definition and minimal quality requirements for stereotactic radiotherapy to achieve best clinical outcome and treatment quality in the implementation into routine clinical practice.

Abstract Studying generalized Onsager models the effect of boundary conditions (finite volume and potential cut-off), used in machine simulations, on the structure of polar systems is examined critically. It is found that deviations from the infinite-system structure stem primarily from the truncation of the potential, which is by no means equivalent to a finite volume, as assumed so far. Intended originally to model computer-generated R-dependent Kirkwood g-factors, continuum theory also predicts correctly the qualitative shape of h Δ- and hD -curves, reported in [10] for various geometries. The present analysis enables, for the first time, a physical understanding of the influence of the cut-off. It turns out that the mean electrostatic energy is only slightly affected and that the asymptotic value of Gk (R) profits from a cancellation of errors. An improved relation is given for the volume dependence of <M 2>. Within the framework of our models one can also understand rigorously the origin of Barker's reaction field.

This abridged version is based on an interdisciplinary guideline which corresponds to development stage 3 (S3) of guidelines according to the classification of the Association of the Scientific Medical Societies in Germany (AWMF). The guideline development process is characterized by the combination of formal evidence-search, formal consensus, logic (algorithms), decision and outcome analysis, and interdisciplinary development with the cooperation of 15 German and Austrian medical societies. [Table 1] shows the relationship between levels of evidence, consensus, and resulting recommendation grades of the recommendations of this guideline. The recommendation grades A – D are added to the recommendations in the abridged version. For the evidence levels, see the full version [1].

OBJECTIVE: The aim of our study was to analyze patterns of care and to identify prognostic factors in patients at least 18 years of age who received radiotherapy for malignant pineal parenchymal tumors. METHODS: In a multicenter, retrospective study, we analyzed data for 37 previously published cases and 64 patients treated at the participating institutions. RESULTS: A total of 56 patients received postoperative radiotherapy, and 45 patients received primary radiotherapy. Chemotherapy was administered to 34 patients. The median follow-up period was 38 months, and median overall survival was 100 months. The variables that significantly influenced overall survival were the extent of disease (localized versus disseminated; P = 0.0002), differentiation (pineal parenchymal tumor of intermediate differentiation versus pineoblastoma; P = 0.001), and residual disease (> or = 50% versus < 50% reduction in size; P < 0.0001). In a multivariate analysis, the parameters turned out to be independent risk factors. The median survival in patients with local or spinal failure was 15 months. Local control was better in older patients (> or = 32 yr versus < 32 yr; P = 0.02). Spinal control was more successful in patients with pineal parenchymal tumors of intermediate differentiation than it was in patients with pineoblastomas (P = 0.03). Nine of 45 treatment failures occurred later than 5 years after treatment. CONCLUSION: Stage, histological characteristics, and response are independent risk factors in adults with malignant pineal parenchymal tumors. Late relapses are common.

BACKGROUND: In 5 consecutive pediatric and adolescent Hodgkin's disease trials DAL-HD since 1978 the invasive diagnostic procedures and the radiotherapy have gradually been reduced and chemotherapy modified to minimize toxicity and the risk of late effects. Since 1982 the overall survival increased up to 95%. In this trial the possibility of reducing local radiation doses to 20 Gy in patients with good response to chemotherapy and omitting radiotherapy totally for patients with complete remission after chemotherapy was tested. PATIENTS AND METHODS: Over a period of 6 years, from August 1995 to July 2001, 1018 children and adolescents with Hodgkin's disease from Germany, Austria,Switzerland, the Netherlands, Sweden, Norway and Denmark were enrolled in this trial. The chemotherapy was equivalent to previous trial DAL-HD 90. The treatment group (TG) 1 (stages I and IIA) received 2 cycles OPPA for girls and 2 cycles OEPA for boys, TG2 (stages IIEA, IIB, IIIA) and TG3 (stages IIEB, IIIEA, IIIB, IV) received additional 2 or 4 cycles COPP respectively. In contrast to trial DAL-HD 90 boys in stage IIIB and IIIEB received OPPA instead of OEPA. The initial staging as well as the restaging for evaluating tumor volume reduction after chemotherapy was reviewed by the study center. Radiotherapy was planned accordingly: patients with complete remission after chemotherapy were not irradiated (21.9%); all other patients received local radiotherapy to the initially involved sites, depending on the tu-mor response. Patients with a partial remission of> 75 tumor regression were irradiated with 20 Gy (50AX), partial remission of< 75% with 30 Gy (4.1 %), and residual masses of > 50 ml were boosted up to 35 Gy (20.2 %). RESULTS: 36 tumor progressions and 49 relapses occurred over a period of 7 1/2 years (median followup 3 years, data deadline 12/19/02). Kaplan-Meier-analysis after 5 years showed a probability for event-free survival (pEFS) for all patients of 0.88 and for overall survival (pOS) of 0.97. For the total group the pDFS (disease free survival) was lower in 222 non irradiated patients than in the 758 irradiated patients (0.88 vs. 0.92,p - 0.049). But there was a difference between the individual treatment groups. In TG 1 there was no difference between nonirradiated and irradiated patients (0.97 vs. 0.94) and the non-ir-radiated patients showed a better trend. In TG 2, and in TG 2 and TG 3 combined, the pDFS was significantly worse for non irradiated patients in comparison with the irradiated patients (TG2:0.78 vs. 0.92; TG 2 +3:0.79 vs. 0.91). Compared to former DAL-HD trials the pOS stayed stable despite therapy reduction. CONCLUSIONS: A reduction of radiotherapy to 20 Gy for patients in all stages with good response to chemotherapy is possible without deterioration of the results. The omission of radiotherapy for patients in complete remission after chemotherapy is recommended only for patients in early stages (TG1). In future trials the possibility of a wider selection for chemotherapy alone for this group needs to be evaluated. In intermediate (TG2) and advanced (TG3) stages omission of radiotherapy for patients incomplete remission results in a lower pEFS, but the pOS is not significantly reduced. Only with knowledge of the long term effects of today's therapy we can give a satisfactory answer to the question whether in future trials the primary aim should be pEFS as high as possible due to front-line-therapy or reduction of late effects.

BACKGROUND: The treatment of Wilms Tumor is integrated into clinical trials since the 1970's. In contrast to the National Wilms Tumor Study Group (NWTSG) the SIOP trials and studies largely focus on the issue of preoperative therapy to facilitate surgery of a shrunken tumor and to treat metastasis as early as possible. PATIENTS AND METHODS: In the SIOP 93-01/GPOH trial and study 1 020 patients with a newly diagnosed renal tumor were registered. 847 of them had a histological proven Wilms Tumor, of whom 637 were unilateral localized, and 173 tumors had an other histology [40 congenital mesoblastic nephroma (CMN), 51 clear cell sarcoma (CCSK), 24 rhabdoid tumor (RTK) and 58 other tumors]. Preoperative chemotherapy in benign tumors was given to 1.3 % of the patients. The main objective of the trial was the randomized question, if the postoperative two drug chemotherapy for stage I in intermediate risk or anaplasia can be reduced from conventional 3 courses to an experimental 1 course without loss of efficacy. RESULTS: 519 patients with unilateral nonmetastatic Wilms did receive preoperative chemotherapy. The histology in this group of patients was of intermediate risk in 469 (90 %) patients, 14 (3 %) tumors were low risk and 36 (7 %) high risk. The stage distribution of the tumors was stage I in 315 (61 %), stage II N- in 126 (24 %), stage II N+ in 25 (5 %) and stage III in 36 (7 %) patients. In 17 (3 %) patients the tumor stage remained unclear. Tumor volume was measured in 487 patients before and in 402 after preoperative chemotherapy. The median tumor volume did shrink from 353 to 126 ml. The amount of volume reduction depends on the histological subtype. The event free survival (EFS) after 5 years was 91 % for all patients with unilateral Wilms tumor without distant metastasis. Randomisation was done in 43.7 % for stage I patients and there was no difference in EFS for both treatment arms (90 versus 91 %). The EFS is identical for patients with stage I and II N- (0.92), as well as for stage II N+ and III (0.82). The tumor volume after chemotherapy is a prognostic factor for intermediate risk tumors with the exception of epithelial and stromal predominant tumors. These two subtypes often present as large tumors, they do not shrink during preoperative chemotherapy but they still have an excellent prognosis. On the other hand the prognosis of patients with blastemal predominant subtype after preoperative chemotherapy is worse than in any other patient group of intermediate risk tumors. There are less blastemal predominant tumors compared to primary surgery, but they are chemotherapeutic resistant selected by the preoperative chemotherapy. CONCLUSION: Patients with unilateral Wilms tumor without metastasis have an excellent prognosis. The post-operative chemotherapy in stage I can be reduced to 4 weeks without worsening treatment outcome. The reduction of the tumor volume could be identified as a helpful marker for stratification of post-operative treatment. Post-chemotherapy blastemal predominant subtype of Wilms tumor has to be classified as high risk tumor. Focal anaplasia has a better prognosis than diffuse anaplasia and will be classified as intermediate risk tumor.

Small-field dosimetry used in advance treatment technologies poses challenges due to loss of lateral charged particle equilibrium (LCPE), occlusion of the primary photon source, and the limited choice of suitable radiation detectors. These challenges greatly influence dosimetric accuracy. Many high-profile radiation incidents have demonstrated a poor understanding of appropriate methodology for small-field dosimetry. These incidents are a cause for concern because the use of small fields in various specialized radiation treatment techniques continues to grow rapidly. Reference and relative dosimetry in small and composite fields are the subject of the International Atomic Energy Agency (IAEA) dosimetry code of practice that has been published as TRS-483 and an AAPM summary publication (IAEA TRS 483; Dosimetry of small static fields used in external beam radiotherapy: An IAEA/AAPM International Code of Practice for reference and relative dose determination, Technical Report Series No. 483; Palmans et al., Med Phys 45(11):e1123, 2018). The charge of AAPM task group 155 (TG-155) is to summarize current knowledge on small-field dosimetry and to provide recommendations of best practices for relative dose determination in small megavoltage photon beams. An overview of the issue of LCPE and the changes in photon beam perturbations with decreasing field size is provided. Recommendations are included on appropriate detector systems and measurement methodologies. Existing published data on dosimetric parameters in small photon fields (e.g., percentage depth dose, tissue phantom ratio/tissue maximum ratio, off-axis ratios, and field output factors) together with the necessary perturbation corrections for various detectors are reviewed. A discussion on errors and an uncertainty analysis in measurements is provided. The design of beam models in treatment planning systems to simulate small fields necessitates special attention on the influence of the primary beam source and collimating devices in the computation of energy fluence and dose. The general requirements for fluence and dose calculation engines suitable for modeling dose in small fields are reviewed. Implementations in commercial treatment planning systems vary widely, and the aims of this report are to provide insight for the medical physicist and guidance to developers of beams models for radiotherapy treatment planning systems.

OBJECTIVE: To analyze the role of radiosurgery alone in patients with brain metastases. There were three specific study goals: 1) to determine whether survival of patients selected for this treatment approach can be predicted successfully by use of the recursive partitioning analysis classification defined by the Radiation Therapy Oncology Group; 2) to evaluate local control; and 3) to identify risk factors of cerebral failure. METHODS: A total of 101 patients with Karnofsky Performance Scale scores of at least 50 and up to three brain metastases, each 3 cm or less in maximum diameter, were treated with radiosurgery alone. Survival, local control, distant brain freedom from progression (FFP), and overall brain FFP were evaluated according the method of Kaplan and Meier. Risk factors for survival and overall brain FFP were analyzed using the Cox model. RESULTS: Median survival was 13.4 months, 9.3 months, and 1.5 months for patients in recursive partitioning analysis Classes 1, 2, and 3, respectively (P < 0.0001). At 1 year, local control, distant brain FFP, and overall brain FFP were 91, 53, and 51%, respectively. An interval greater than 2 years between diagnosis of the primary tumor and diagnosis of brain metastases and the presence of a single brain metastasis were associated with significantly higher overall brain FFP. CONCLUSION: Recursive partitioning analysis classification successfully predicted survival. Radiosurgery alone yielded high local control. Overall brain FFP was highest in patients with an interval greater than 2 years between primary diagnosis and diagnosis of a single brain metastasis.