K.V. Chaung: None. M.Z. Kharouta: None. A.J. Gross: None. G.C. Pereira: None. A.M. Kumar: Stock; AbbVie. D.B. Mansur: None. A.E. Sloan: None. T.R. Hodges: None. S. Choi: None.

The primary treatment modality for non-metastatic soft tissue sarcomas (STS) of the extremity is surgical resection. This can be accomplished with amputation, or limb sparing surgery (LSS) when feasible. Radiation therapy improved survival in patients with tumors that were high grade or undifferentiated and >5cm (high risk) in a prior SEER analysis. The role of chemotherapy remains controversial in STS. The SEER 18 registries database was queried to identify patients age 18 and older with pathologically confirmed STS of the extremity diagnosed in 2004 or later that underwent surgical resection. Patients with Kaposi's sarcoma were excluded as well as those with metastatic disease, <3 months of follow up and if the type of surgery was not specified. The primary endpoint of this study was overall survival, which was determined using the Kaplan and Meier method. Survival curves were subsequently compared using log rank analysis. A total of 8,584 patients were identified that met inclusion criteria, Females comprised 3,862 patients and males 4,722 patients. The average age at diagnosis was 57 (range 18-98). Grade III or undifferentiated tumors were present in 3,860 patients while 4,915 patients had tumors >5cm. The 5 year OS for patients who underwent amputation was 53.4% compared to 77.0% with LSS, corresponding to a HR of 0.40 (p<0.001). Median OS was 68 months with amputation and not reached at 10 years with LSS. Of the 8,122 patients who underwent LSS, 4,369 patients were known to have received radiation therapy and 1,135 were known to have been treated with chemotherapy. Neither radiation therapy nor chemotherapy improved survival in the entire cohort of patients undergoing LSS. In a subset of 2,331 patients with high risk tumors, radiation improved 5 year OS from 51.1% to 59.6% while median OS improved from 71 to 97 months, corresponding to a HR of 0.78 (p=0.002). In the same subset of patients treated with radiation, 689 (37.2%) had neoadjuvant radiation with or without postoperative radiation (PORT) and 1,161 (62.8%) had PORT alone. 5 year OS was 58.5% with PORT and 61.5% with neoadjuvant radiation while median OS was 89 and 97 months respectively, corresponding to a HR of 0.97 (p=.671). For the 720 patients with high risk tumors who were known to have received chemotherapy, the 5 year OS and median OS were 61.7% and 117 months respectively versus 58.5% and 83 months respectively for patients not known to receive chemotherapy, corresponding to a HR of 0.83 (p=0.025). The findings of this analysis support continued use of LSS in surgically eligible patients with STS of the extremity. While neither radiation therapy nor chemotherapy improved survival in all comers treated with LSS, the subset of patients with high risk tumors (high grade/ undifferentiated and >5cm) showed a significant survival benefit from both modalities in addition to surgery. The timing of radiation therapy, neoadjuvant versus PORT did not impact survival in this subset of patients.

HSRT directed for large surgical beds in patients with resected brain metastases improves local control while sparing patients the toxicity associated with whole brain radiation. We review our institutional series to determine factors predictive of local failure. In a total of 38 consecutive patients with brain metastases treated from August 2011-August 2016, 40 surgical beds were treated with HSRT in 3 or 5 fractions. Each surgical bed was treated as a unique data point. All treatments were completed on a robotic radiosurgery platform using the 6D Skull tracking system. Volumetric MRIs from before and after surgery were used for radiation planning. A 2 mm PTV margin was used around the contoured surgical bed +/- resection margins and these were reviewed by the radiation oncologist and neurosurgeon. Lower total doses were prescribed based on proximity to critical structures or if prior radiation treatments were given. Local control in this study is defined as no volumetric MRI evidence of recurrence of tumor within the high dose radiation volume. Statistics were calculated using JMP Pro v12.1. Of the 40 surgical beds analyzed, 23 were from NSCLC, 5 were from breast, 4 from melanoma, 2 from esophagus, and 1 each from SCLC, sarcoma, colon, renal, rectal, and unknown primary. 9 were treated with 3 fractions with median dose 24Gy and 31 were treated with 5 fractions with median dose 27.5Gy using an every other day fractionation. There were no reported grade 3 or higher toxicities. Median follow up was 212 days after completion of radiation. 10 (33%) surgical beds developed local failure with a median time to failure of 148 days. All but 3 patients developed new brain metastases outside of the treated field and were treated with stereotactic radiosurgery, whole brain radiation and/or chemotherapy. 5 patients (13%) developed leptomeningeal disease. With a median follow up of 226 days, 30 Gy/5 fx was associated with the best local control (93%) with only 1 local failure. A lower total dose in 5 fractions (ie 27.5 Gy or 25 Gy) had a local control rate of 70%. For 3 fraction SBRT, local control was 100% using a dose of 27 Gy in 3 fractions (follow up was >600d) and 70% if 24 Gy in 3 fractions was used. A higher total BED (α/β=10) was statistically significant for improved local control (p=0.04) with a threshold BED of 48 or higher associated with better local control. HSRT after surgical resection for brain metastasis is well tolerated and has improved local control with BED>=48 (30Gy/5fx and 27Gy/3fx). Additional study is warranted.

Purpose/Objective(s)Appropriate PTV margins for the stomach have not been clearly elucidated, and interfractional variation of the stomach may be significant due to irregularities in shape, volume, and mobility. We investigated the daily variation in gastric position by examining daily megavoltage CT scans in comparison to CT simulation images.Materials/MethodsImages from three patients treated with Tomotherapy for gastric lymphoma were evaluated retrospectively for this study. Patients were simulated with 80 or 100 cc of oral contrast after a 6 hour fast. During therapy, gastric volume was controlled similarly with a 6 hour fast and administration of 80 or 100 cc of water prior to treatment. The stomach volumes were identified and contoured on each daily megavoltage scan, for a total of 41 scans. Each contour was examined by two observers. To determine organ motion variation, daily megavoltage scans were registered to the CT simulation image set and the margins in the r/l, sup/inf, ant/post directions that covered 95% of the daily image motion were computed using a simple grid search with a discrete set of margins ranging from 0 to 2.4 cm in increments of 0.6 cm. In addition, the setup variation was calculated based on the daily shifts using the margin recipe published by van Herk et al., IJROBP, 2000. To estimate total variation for patients not receiving daily image guidance, overall margins were determined by adding organ motion and daily setup variation.ResultsThe maximum margin to the right, left, posterior, anterior, superior, and inferior axes for the line encompassing 95% of daily volume is 0.88cm, 1.75cm, 0.88cm, 1.75cm, 1.75cm, and 0.5cm, respectively. The maximum calculated setup variation is 0.85cm, 0.13cm, 0.05cm, 0.57cm, 0.15cm, and 0.36cm, respectively. Therefore, the overall margin required to cover 95% of the stomach in a patient not receiving daily image guidance is 1.72cm, 1.88cm, 0.92cm, 2.32cm, 1.90cm, and 0.86cm, respectively.ConclusionsThis study demonstrates considerable daily variability in stomach localization. The necessary expansion of the GTV for 95% coverage is direction-dependent and ranges from 0.5cm to 1.75cm with daily image guidance and 0.86cm to 2.32cm without image guidance. While the data suggest margins can be reduced by using image guidance, the required margins are still relatively large. These data should be confirmed with additional patient numbers, but provides guidance for those patients currently undergoing treatment planning for gastric irradiation. Purpose/Objective(s)Appropriate PTV margins for the stomach have not been clearly elucidated, and interfractional variation of the stomach may be significant due to irregularities in shape, volume, and mobility. We investigated the daily variation in gastric position by examining daily megavoltage CT scans in comparison to CT simulation images. Appropriate PTV margins for the stomach have not been clearly elucidated, and interfractional variation of the stomach may be significant due to irregularities in shape, volume, and mobility. We investigated the daily variation in gastric position by examining daily megavoltage CT scans in comparison to CT simulation images. Materials/MethodsImages from three patients treated with Tomotherapy for gastric lymphoma were evaluated retrospectively for this study. Patients were simulated with 80 or 100 cc of oral contrast after a 6 hour fast. During therapy, gastric volume was controlled similarly with a 6 hour fast and administration of 80 or 100 cc of water prior to treatment. The stomach volumes were identified and contoured on each daily megavoltage scan, for a total of 41 scans. Each contour was examined by two observers. To determine organ motion variation, daily megavoltage scans were registered to the CT simulation image set and the margins in the r/l, sup/inf, ant/post directions that covered 95% of the daily image motion were computed using a simple grid search with a discrete set of margins ranging from 0 to 2.4 cm in increments of 0.6 cm. In addition, the setup variation was calculated based on the daily shifts using the margin recipe published by van Herk et al., IJROBP, 2000. To estimate total variation for patients not receiving daily image guidance, overall margins were determined by adding organ motion and daily setup variation. Images from three patients treated with Tomotherapy for gastric lymphoma were evaluated retrospectively for this study. Patients were simulated with 80 or 100 cc of oral contrast after a 6 hour fast. During therapy, gastric volume was controlled similarly with a 6 hour fast and administration of 80 or 100 cc of water prior to treatment. The stomach volumes were identified and contoured on each daily megavoltage scan, for a total of 41 scans. Each contour was examined by two observers. To determine organ motion variation, daily megavoltage scans were registered to the CT simulation image set and the margins in the r/l, sup/inf, ant/post directions that covered 95% of the daily image motion were computed using a simple grid search with a discrete set of margins ranging from 0 to 2.4 cm in increments of 0.6 cm. In addition, the setup variation was calculated based on the daily shifts using the margin recipe published by van Herk et al., IJROBP, 2000. To estimate total variation for patients not receiving daily image guidance, overall margins were determined by adding organ motion and daily setup variation. ResultsThe maximum margin to the right, left, posterior, anterior, superior, and inferior axes for the line encompassing 95% of daily volume is 0.88cm, 1.75cm, 0.88cm, 1.75cm, 1.75cm, and 0.5cm, respectively. The maximum calculated setup variation is 0.85cm, 0.13cm, 0.05cm, 0.57cm, 0.15cm, and 0.36cm, respectively. Therefore, the overall margin required to cover 95% of the stomach in a patient not receiving daily image guidance is 1.72cm, 1.88cm, 0.92cm, 2.32cm, 1.90cm, and 0.86cm, respectively. The maximum margin to the right, left, posterior, anterior, superior, and inferior axes for the line encompassing 95% of daily volume is 0.88cm, 1.75cm, 0.88cm, 1.75cm, 1.75cm, and 0.5cm, respectively. The maximum calculated setup variation is 0.85cm, 0.13cm, 0.05cm, 0.57cm, 0.15cm, and 0.36cm, respectively. Therefore, the overall margin required to cover 95% of the stomach in a patient not receiving daily image guidance is 1.72cm, 1.88cm, 0.92cm, 2.32cm, 1.90cm, and 0.86cm, respectively. ConclusionsThis study demonstrates considerable daily variability in stomach localization. The necessary expansion of the GTV for 95% coverage is direction-dependent and ranges from 0.5cm to 1.75cm with daily image guidance and 0.86cm to 2.32cm without image guidance. While the data suggest margins can be reduced by using image guidance, the required margins are still relatively large. These data should be confirmed with additional patient numbers, but provides guidance for those patients currently undergoing treatment planning for gastric irradiation. This study demonstrates considerable daily variability in stomach localization. The necessary expansion of the GTV for 95% coverage is direction-dependent and ranges from 0.5cm to 1.75cm with daily image guidance and 0.86cm to 2.32cm without image guidance. While the data suggest margins can be reduced by using image guidance, the required margins are still relatively large. These data should be confirmed with additional patient numbers, but provides guidance for those patients currently undergoing treatment planning for gastric irradiation.

Purpose: Though radiation therapy is generally considered the most effective treatment for unresectable pilocytic astrocytomas in children, there are few data to support this claim. To examine the efficacy of radiation therapy for pediatric pilocytic astrocytomas, we retrospectively reviewed the experience at our institution.Methods and Materials: Thirty-five patients 18 years old or younger with unresectable tumors and without evidence of neurofibromatosis have been treated since 1982. Patients were treated with local radiation fields to a median dose of 54 Gy. Six patients were treated with radiosurgery to a median dose of 15.5 Gy. Five patients were treated with initial chemotherapy and irradiated after progression.Results: All patients were alive after a median follow-up of 5.0 years. However, progression-free survival was 68.7%. None of 11 infratentorial tumors progressed compared with 6 of 20 supratentorial tumors. A trend toward improved progression-free survival was seen with radiosurgery (80%) compared with external beam alone (66%), but this difference did not reach statistical significance. Eight of the 9 patients progressing after therapy did so within the irradiated volume.Conclusions: Although the survival of these children is excellent, almost one third of patients have progressive disease after definitive radiotherapy. Improvements in tumor control are needed in this patient population, and the optimal therapy has not been fully defined. Prospective trials comparing initial chemotherapy to radiation therapy are warranted. (C) 2011 Elsevier Inc.

To develop a comprehensive program for modulated electron radiotherapy (MERT). All steps; Monte Carlo (MC) based treatment planning, energy and segment optimization, delivery using the photon multileaf collimators, and monitoring during treatment, are necessary. Comparisons with conventional (CRT) and photon IMRT plans will be conducted. We previously demonstrated the multi-leaf collimation (MLC) system for modulating photons is ideal for MERT due to the systems' efficiency and safety. Multiple electron segments were combined to deliver predictable, conformal dose distributions. Monte Carlo calculations were used; BEAMnrc for phase space file generation, MCSim for MC dose calculations, and CERR for energy/segment size/dose weighting optimization using custom GUIs. Once segments were devised, MLC leaf delivery instructions were appropriated using Shaper. Dosimetric validation was performed with film and ionization chambers for Varian IX accelerator delivered electrons. Phantom targets including imbedded heterogeneities (bone, lung) were planned with delivery validated by measurements. Clinical cases (post-mastectomy chest wall and cutaneous lymphoma of the scalp) previously treated by CRT or IMRT methods were planned for MERT. Comparison of isodose and DVHs was conducted in CERR for MERT, IMRT (Pinnacle or Tomotherapy) and/or CRT plans. Treatment execution at 70cm SSD, necessary for MERT, was simulated to ensure safety and efficiency. Service mode operation is currently required to deliver MERT with the photon MLC. AlignRT (VisionRT, UK), a video based surface imaging system, was tested for daily localization and monitoring. Dosimetric validation for all available energies in the phantom studies was successful (2 mm/3% agreement) for segment sizes ranging from 10 to 100 mm. The MERT chest wall cases were planned and optimized for a single table and gantry position, 9 segments, and 4 energies. Phantom delivery time was 3.5′. Planning provided homogeneous coverage (± 10%) to the chest wall and IM nodes, without consequence of excessive high dose regions, as with IMRT and CRT. Dose to lung and heart were lower for MERT plans, typically less than half. Dose to the contralateral breast using MERT was less than 10% of the dose received using CRT or IMRT. The VisionRT system worked for the MERT technique (short SSD). Scalp planning was complex, requiring 2 table and 3 gantry angles, 2 energies, and 11 segments. Planning provided homogeneous coverage (± 9%), with complete sparing of brain tissue. We have demonstrated conformal distributions for chest wall and scalp with MERT, superior to CRT or IMRT. Planning, optimization, delivery, and verification for MERT were achieved. IRB approval for MERT with photon MLCs is being pursued.

Fractionated TBI (FTBI) followed by allogeneic hematopoietic SCT results in donor engraftment and improves survival in children with high-risk hematologic malignancies. However, acute toxicities (skin, lung and mucosa) are common after FTBI. Late complications include cataracts, endocrine dysfunction, sterility and impaired neurodevelopment. Instead of FTBI, we used low-dose single fraction TBI (550 cGy) with CY as transplant conditioning for pediatric hematologic malignancies. GVHD prophylaxis included CYA and short-course MTX; methylprednisolone was added for unrelated donor transplants. A total of 55 children in first (40%) or second remission and beyond (60%) underwent transplantation from BM (65%) or peripheral blood; 62% from unrelated donors; 22% were mismatched. Median follow-up was 18.5 months (1–68). Overall survival and disease-free survival at 1 year were 60 and 47%, respectively. Acute toxicities included grade 3–4 mucositis (18%), invasive infections (11%), multiorgan failure/shock (11%), hemolytic anemia (7%), veno-occlusive disease (4%) and renal failure (4%). TRM was 11% at 100 days. Non-relapse mortality was 6% thereafter. Graft rejection occurred in 2%. Three patients (5%) died of GVHD. The regimen was well tolerated even in heavily pretreated children and supported donor cell engraftment; long-term follow up is in progress.

Our intent is to determine the accuracy of on-board kV imaging (kVi) by comparing its suggested patient positioning to that obtained using an infrared optical camera (SOC), having an established accuracy of <0.6 mm and 0.1°. Twenty cranial stereotactic treatments were reviewed. A CT-based (spiral-axial, 512 × 512 pixels, 350 mm FOV, 1.5 mm spacing, 0.6 pitch) radiotherapy treatment plan was created for frameless SOC localization for each patient in a thermoplastic mask. Two orthogonal digitally reconstructed radiographs (DRRs) were computed in the anterior-posterior (A/P) and the right-left (R/L) directions. Patients were initially positioned using the daily calibrated SOC system and were optically monitored for the duration of the treatment. Localization with SOC was the benchmark for position comparisons. A/P and R/L kVi immediately followed without moving the patient. The kVi transferred to a record and verify (RNV) system for registration to corresponding DRRs. Registration of the image pairs was achieved using 5 corresponding anatomical landmarks in paired images. The following were noted: shift to agreement in orthogonal directions; rotation in coronal and sagittal planes; the landmark residual errors (REs); mean, standard deviation, minimum and maximum distance to agreement for landmarks. Analysis resulted in the following [key: mean (median)]: superior/inferior (S/I) shift: 0.3 (0.3) mm; A/P shift: 0.4 (0.4) mm; R/L shift: 1.0 (0.9) mm; coronal plane rotation: 0.3 (0.1)°; sagittal plane rotation: 1.2 (1.1)°. Maximum in-plane rotation was 2.7° for one patient. Shifts were found up to 2.8 mm. Differences ranged from 0.1 to 1.9 mm for S/I shifts between the AP and the lateral views for the same treatment. Three of these exceeded 1.0 mm. REs for the 5 landmark correlation were ∼1 mm but ranged from 0.53 to 1.76 mm. Plots of the shift to agreement vs. REs showed little correlation. Errors in quadrature had values with mean (median) of 1.46 (1.33) mm, ranging up to 3 mm. Positional errors in any one direction or rotation were typically small (∼1 mm or less), below clinical and technical tolerances on the linac, and confirmed accurate agreement of kVi with SOC. Outliers existed, however, implying that errors can accumulate during the course of data manipulation to generate DRRs, registration of kVi with DRRs and transfer of the data through the various computer workstations: and small patient motions. DRR resolution is a probable source of error even though the scans were spaced 1.5 mm. Slice spacing of 1 mm would be desirable, but some treatment planning systems preclude the resulting large number of axial scans. REs of 1 mm corroborate good correlation technique Although patient position was continuously monitored with the SOC, small motions before taking the kVi could account for some observed error. This was evident since three S/I shifts were significantly different between successive A/P and R/L kVi, ranging from 1.2 to 1.9 mm. Patient data exhibited small rotations typically <1° with one sagittal rotation of 2.7°. Of note is that sagittal plane rotational error is 4.6 times greater compared to coronal. The implication is that the immobilization head mask restricts yaw rotation better than pitch.

To develop modulated electron radiotherapy (MERT) programs for two challenging treatment sites, chest-wall (post mastectomy) and scalp (cutaneous lymphoma). Techniques using modulated photons, matched electron beams, or brachytherapy have been met with technical and/or clinical complications. We previously demonstrated the multi-leaf collimation (MLC) system for modulating photons is ideal for MERT due to the control systems' efficiency and safety. Multiple electron segments can be combined to deliver predictable, conformal dose distributions. Commercial planning systems do not support calculations for MERT, therefore Monte Carlo (MC) calculations were used. BEAMnrc generated phase-space files, and MCSIM, the dose calculations. Dosimetric validation was performed with film and chambers for Varian IX accelerator delivered electrons. Phantom targets were planned with forward optimization, with subsequent delivery validated by measurements. Clinical cases previously treated by conventional methods were planned for MERT. Execution of treatment was simulated to ensure safety and efficiency. Validation for all available energies was successful (3 mm/3% agreement) for 70 cm SSD, necessary for MERT, and segments ranging from 10 to 100 mm. To reduce variance 109 histories were required. The chest wall cases were planned with a single table and gantry position, 8 segments, and 2 energies. Phantom delivery time was only 3.5′. Planning provided homogeneous coverage (±10%) to the chest wall and IM nodes. Dose to lung, heart, and contralateral breast were low (<5%). Scalp planning was more complex, requiring 2 table & 3 gantry angles, 2 energies, and 11 segments. Planning provided homogeneous coverage (±9%), with complete sparing of brain tissue. Phantom delivery was completed in 16′ (Figures). We have demonstrated conformal distributions for these sites with MERT. DVH data will demonstrate superiority over modulated photons and other complex techniques.View Large Image Figure ViewerDownload Hi-res image Download (PPT)

Irradiation of the axillary nodes is often indicated in the clinical radiotherapy of patients with lymphoma and breast cancer. The relative location of the axillary nodes and the humeral head has historically been estimated based on lymphangiographic data. Since CT offers more precise definition, the axillary nodal region in relation to the humeral head was systematically studied in 61 patients undergoing CT simulation for breast conservation therapy. Differences in arm position (degree of abduction) significantly affected the location of the axillary nodal region in relation to the humeral head. With the arm abducted beyond 55 degrees, humeral head blocking would result in blocking of some of the axillary nodes.

Seventy-three patients were evaluated for the changes in pain relief, numbness and paresthesias after Gamma Knife radiosurgery to a maximum dose of 76–87 Gy for trigeminal neuralgia. Patients experienced pain relief as follows: 59% attained complete pain relief without prior surgery (33% with prior surgery); 25% achieved ≧50% pain reduction (28% with prior surgery); 11% of surgery patients obtained minor pain relief, and 16% of patients without surgery had no relief (28% with prior surgery). Level of pain decreased rapidly within 6 weeks after radiosurgery. Numbness/paresthesias developed slowly over the first 12–15 months. Bothersome levels were experienced by 15% of the patients without prior surgery (22% with prior surgery). Comparison of the occurrence of numbness/paresthesias, with respect to prior surgery, was not statistically significant. Only 2% of all patients had persistently bothersome side effects. In conclusion, radiosurgery is an effective treatment of trigeminal neuralgia, especially for those patients not having prior surgery.

Purpose: To evaluate the dose distribution outside of a cyst instilled with phosphorous-32 (P-32, an electron emitter with a short effective range of 2-8 mm and average energy of 0.69 MeV, used to treat cystic craniopharyngioma) as a function of cyst size with and without plating (migration and adhesion of P-32 to the cyst surface). Methods and Materials: A cystic craniopharyngioma treated with instillation of P-32 was approximated by a sphere of uniformly distributed and plated chromic P-32 colloid. The percent depth dose was calculated along a radial position vector exterior to the sphere with a three-dimensional convolution integral and a dose point kernel. Results: The percent depth dose variation of surface or volume source external to a family of spheres was plotted. Complex cyst geometry is amenable to evaluation by approximation with simple spheres. Error estimates are calculated for the dose outside of truncated sphere segments. Plating might occur and raise the dose outside the cyst by more than a factor of 5.0. This has the potential to cause damage to adjacent tissues, including the optic chiasm. Conclusion: Clinicians are faced with a number of treatment options for cystic craniopharyngioma, including intracystic instillation of colloid P-32. Unfortunately, plating might occur and potentially damage adjacent normal tissues. It is recommended that the propensity for a craniopharyngioma to plate be evaluated before full treatment, especially after previous treatment. (c) 2005 Elsevier Inc.

Purpose: To retrospectively determine the long-term outcome of intracranial ependymoma patients treated with surgery and postoperative radiation therapy.Methods and Materials: Sixty patients were treated at our institution between 1964 and 2000. Forty patients had World Health Organization Grade II ependymoma, and 20 patients had Grade III ependymoma. The median patient age was 10.7 years. The majority of patients were male (55%), had infratentorial tumors (80%), and had subtotal resections (72%). Postoperative radiation therapy was delivered to all patients to a median total dose of 50.4 Gy. Craniospinal radiation therapy was used in the earlier era in only 12 patients (20%).Results: The median follow-up of surviving patients was 12.5 years. The 5-year and 10-year disease-free survival rates for all patients were 58.4% and 49.5%, respectively. The 5-year and 10-year overall survival rates for all patients were 71.2% and 55.0%, respectively. Supratentorial tumor location was independently associated with a worse disease-free survival. Subtotal resection and supratentorial location predicted a worse overall survival, but this failed to reach statistical significance. No statistically significant effect on prognosis was observed with tumor grade, patient age, or radiation dose or volume.Conclusion: Our long-term follow-up indicates that half of ependymoma patients will have disease recurrences, indicating the need for more effective treatments. (C) 2005 Elsevier Inc.

To describe the ipsilateral breast tumor recurrence (IBTR) rate in women with invasive breast cancers treated with standard breast conservation therapy (BCT) that would have been eligible for accelerated partial breast irradiation. A database of 1360 women treated with standard BCT (defined as a lumpectomy, axillary lymph node assessment, and adjuvant whole breast irradiation) at a single institution from 1976 to 1997 was reviewed to generate a cohort of patients who met pathologic criteria for accelerated partial breast irradiation (APBI). To be included in the cohort, the primary tumor was required to be of an invasive non-lobular histology with a tumor size ≤ 3 cm; the axilla was required to be pathologically negative via sentinel lymph node biopsy or axillary lymph node dissection (minimum of six sampled nodes); and the final surgical margins on the lumpectomy had to be negative by ≥ 2 mm. The Kaplan-Meier method was used to describe the IBTR. There were 919 patients that met the above pathologic criteria for APBI. Median age was 58 years with a median follow-up of 5.1 years. Tumor size was pT1 in 83% with 37% having a tumor size of ≤1 cm. Estrogen receptors were classified as positive in 55%. Adjuvant chemotherapy was administered to 32% of patients while 43% received adjuvant hormonal therapy. No patient was treated with neoadjuvant chemotherapy or neoadjuvant hormonal therapy. The five-year IBTR was 4% while the ten-year IBTR was 9%. Whole breast radiation therapy results in a very low ipsilateral breast tumor recurrence rate at long follow-up in cancers that may be deemed eligible for partial breast treatment. Studies on the various methods of accelerated partial breast irradiation must match this high standard to be considered as viable treatment options for women with early stage breast cancer

The efficacy of post-operative radiation therapy in breast cancer management has been evident for several decades. In recent years there has been increasing interest in treating the internal mammary lymph nodes. In an attempt to simplify their localization for treatment planning purposes, the position of the internal mammary lymph nodes was systematically studied in 65 breast cancer patients undergoing CT simulation.

Purpose/Objective: An examination of patterns of failure in patients with medulloblastoma who received optimal radiation therapy. Materials/Methods: Radiation therapy records of 95 medulloblastoma patients, ages 1 to 48 years (median 10), treated at Mallinckrodt from 1959 to 1999 were reviewed. Median follow-up was 5.2 years. Chang T-staging was 5% T1, 45% T2, 15% T3a, 21% T3b, and 7% T4. M-stage was performed in 47% of patients: 52% M0, 5% M1, 18% M2, 25% M3. Gross total resection was achieved in 46% and 54% underwent subtotal resection or biopsy only. The median dose to the posterior fossa (PF) was 50.28Gy, the whole brain (WB) was 39.60Gy, and the spinal axis (SA) was 35Gy. Fraction sizes ranged from 0.10Gy to 2.50Gy (median 1.80Gy PF, 1.70Gy WB, 1.66Gy SA). Chemotherapy was administered in 29% of patients. Results: The 5-year overall survival (OS) and disease free survivals were 68% and 63%, respectively. There were 41 treatment failures, some occurring at more than one site. Sites of any failure were local in 29 (71%) neuroaxis in 16 (39%), and distant in 24 (58%). Sites of first failure were local in 26 (63%), neuraxis in 12 (29%), and distant in 7 (17%). The 5 year local PF control was 20%, 30% and 90% for patients receiving PF doses of <45Gy, 45–53.99Gy, or ≥54Gy, respectively (p=0.05). In optimally treated patients (n=15), defined as receiving PF doses >=50Gy, and SA doses ≥30Gy (with or without chemotherapy), sites of first failure were local in 9 (56%), neuroaxis in 3 (19%), and distant in 3 (19%), compared with 15 (63%), 3 (12%), and 3 (12%), in sub-optimally treated patients (n=21), respectively. In patients who received ≥50Gy to the PF, sites of first failure were local in 9 (43%), neuroaxis in 7 (33%), and distant in 3 (14%), compared with 14 (70%), 1 (5%), and 3 (15%) in those receiving <50Gy. In patients who received ≥30Gy to the SA, sites of first failure were local in 20 (59%), neuroaxis in 5 (15%), and distant in 6 (18%), compared with 3 (50%), 2 (33%), and 0 in those receiving <30Gy. Conclusions: The posterior fossa remains the dominant site of failure but as PF dose increases, the rate of local failure decreases. As radiation therapy quality improves, patterns of failure appear to shift from the posterior fossa and neuroaxis to the neuroaxis and distant sites. Analyses such as this will help oncologists to better tailor treatment to sites at risk.

Purpose/Objective: To implement intensity modulated radiation therapy (IMRT) in the treatment of breast cancer we investigated two factors: 1) The quantification of breast/chest wall movement both interfraction (set-up variation), and intrafraction (breathing motion) using electronic portal imaging, and 2) the radiation dose distribution delivered to whole breast, biopsy cavity, regional nodes, and uninvolved normal tissue using a commercial IMRT system. Materials/Methods: Electronic portal images were taken for 12 patients during tangential treatments. For each patient an average of 8 images per field were acquired during a single fraction. This was repeated for at least 10 fractions, yielding a total of 2931 images. To analyze this large number of images, an algorithm was developed using the Canny method. Images were overlaid in-frame and the anterior-posterior motion was measured perpendicular to the beam central axis. The vertical motion of the diaphragm was also measured. The PTV and normal tissue planning volume for the breast/regional lymphatics were created based on these intrafraction data. Using a commercial IMRT system, plans were created for two patients, one that included the breast only as target volume, and one that included breast and regional lymph nodes. Three treatment plans for each case were generated with two-, three-, and four-fields. Intensity modulated tangential or near tangential 6MV beams were used to minimize low dose scatter to normal tissue, especially lungs. Target coverage and normal tissue dose were compared using dose volume histograms. Results: The extent of interfraction variation is depicted in the composite of all images for a representative patient seen in figure on left. From the composite, the maximum chest wall and breast variation can be measured to be up to 2 cm. The intrafractional movement is depicted in the composite of all images seen in figure on right. The maximum chest wall and breast movement can be measured to be up to 1 cm. The diaphragm motion reached up to 1.5 cm. The 3 and 4 field IMRT plans demonstrated excellent dose distribution with the elimination of multiple fields and junctions, and kept lung dose-volume parameters below our published institutional limits only when internal mammary nodes were not included as a target. Conclusions: The magnitude of interfractional setup error was significantly greater than the intrafractional variation. The magnitude of intrafractional motion was much less than we expected. Interfraction variation indicates a need for improved patient positioning, either with better immobilization or image guided patient setup technology. Intrafraction movement (primarily breathing motion) must also be taken into account in the implementation of IMRT in the treatment of breast cancer patients. IMRT allows elimination of junctions and fewer fields to be used to treat breast cancer.

Purpose/Objective: To determine the outcome of intracranial ependymoma patients treated with surgery and post-operative radiation therapy (RT). Materials/Methods: Between 1964 and 2000, 63 patients with intracranial ependymoma (n=44) or intracranial anaplastic ependymoma (n=19) were treated with surgery and post-operative RT with curative intent. There were 56% male patients, and 78% of tumors were infratentorial. The extent of resection included gross total (n=15), subtotal (n=44), biopsy only (n=1), and unknown (n=3). No patient had evidence of craniospinal seeding or other disseminated disease at the time of RT. The median RT dose was 50.4 Gy. In the earlier years of the study 13 patients received prophylactic RT to the craniospinal axis. However, treatment philosophy evolved over the time period and the majority of patients (n=50) were treated to local fields. Results: The median follow-up of surviving patients is 13.0 years (range 1-36). The 5 and 10 year overall survival (OS) for all patients is 66.0% and 53.1%, respectively. The 5 and 10 year disease free survival (DFS) for all patients is 56.8% and 48.2%, respectively. Those patients with anaplastic ependymoma have a 5 and 10 year overall survival of 48.9% and 20.4%, respectively, compared to 76.1% and 64.5% for patients with ependymoma (p=0.02). The use of craniospinal RT is associated with worse DFS at 5 (44.9% vs 59.9%) and at 10 years (26.9% vs 54.0%) compared to local field RT, but this difference did not reach statistical significance (p=0.16). Of the 33 patients with recurrent disease, pattern of failure information is available on 28. Of these patients, 26 (93%) failed within sites of previous disease and within the RT field. Multivariate analysis was utilized to determine the impact of tumor histology, extent of resection, volume of RT, location of tumor, and patient age on outcome. Only anaplastic histology was found to adversely influence OS (p=0.008). Both anaplastic histology (p=0.01) and supratentorial tumor location (p=0.03) adversely influenced DFS. Conclusions: While approximately two thirds of patients with ependymoma can achieve long-term cure with surgery and post-operative RT, patients with anaplastic ependymoma have significantly worse prognosis. In this retrospective study, prophylactic craniospinal RT did not improve outcome, and the primary pattern of failure was within the local RT field. For patients with poor prognostic factors, especially anaplastic histology, additional therapies and/or RT dose escalation should be investigated.

Purpose: To investigate the impact of radiographic parameter and radiation technique on the volumetric dose of lung and heart for intact breast radiation.Methods and Materials: Forty patients, with both two-dimensional (2D) and computed tomographic (CT) simulations were enrolled in the study. Central lung distance (CLD), maximal heart distance (MHD), and maximal heart length (MHL) were measured under virtual simulation. Four plans were compared for each patient. Plan A used a traditional 2D tangential setup. Plan B used clinical target volume (CTV) based three-dimensional (3D) planning. Both plans C and D used a combination of a medial breast field with shallow tangents. Plan D is a further modification of plan C.Results: Under the traditional tangential setup, the mean ipsilateral lung dose and volume at 20, 30, and 40 Gy correlated linearly with CLD (R = 0.85similar to0.91). The mean ipsilateral lung dose (Gy) approximated 4 times the CLD value (cm), whereas the percentage volume (%) of ipsilateral lung at 20, 30, and 40 Gy was about 10 times the CLD (cm). The mean heart dose and percentage volume at 20, 30, and 40 Gy correlated with MHD (R = 0.76similar to0.80) and MHL (R = 0.65-0.75). The mean heart dose (Gy) approximated 3 times the MHD value (cm), and the percentage volume (%) of the heart at 10, 20, 30, and 40 Gy was about 6 times MHD (cm). Radiation technique impacted lung and heart dose. The 3D tangential plan (plan B) failed to reduce the volumetric dose of lung and heart from that of the 2D plan (plan A). The medial breast techniques (plans C and D) significantly decreased the volume of lung and heart receiving high doses (30 and 40 Gy). Plan D further decreased the 20 Gy volumes. By use of the medial breast technique, the lung and heart dose were not impacted by original CLD and MHD/MHL. Therefore, the improvement from the tangential technique was more remarkable for patients with CLD greater than or equal to 3.0 cm (p < 0.001).Conclusions: The CLD and MHD impact the volumetric dose of lung and heart. The application of 3D planning for tangential breast irradiation does not decrease heart and lung dose. Adding a medial breast port significantly decreases percentage volume (PV) of lung and heart receiving high doses, especially when the CLD is excessive. (C) 2002 Elsevier Science Inc.