Michael R. Folkert

GPU-based streaming architectures for fast cone-beam CT image reconstruction and demons deformable registration

This paper shows how to significantly accelerate cone-beam CT reconstruction and 3D deformable image registration using the stream-processing model. We describe data-parallel designs for the Feldkamp, Davis and Kress (FDK) reconstruction algorithm, and the demons deformable registration algorithm, suitable for use on a commodity graphics processing unit. The streaming versions of these algorithms are implemented using the Brook programming environment and executed on an NVidia 8800 GPU. Performance results using CT data of a preserved swine lung indicate that the GPU-based implementations of the FDK and demons algorithms achieve a substantial speedup--up to 80 times for FDK and 70 times for demons when compared to an optimized reference implementation on a 2.8 GHz Intel processor. In addition, the accuracy of the GPU-based implementations was found to be excellent. Compared with CPU-based implementations, the RMS differences were less than 0.1 Hounsfield unit for reconstruction and less than 0.1 mm for deformable registration.

Outcomes and Toxicity for Hypofractionated and Single-Fraction Image-Guided Stereotactic Radiosurgery for Sarcomas Metastasizing to the Spine

PURPOSE Conventional radiation treatment (20-40 Gy in 5-20 fractions, 2-5 Gy per fraction) for sarcoma metastatic to the spine provides subtherapeutic doses, resulting in poor durable local control (LC) (50%-77% at 1 year). Hypofractionated (HF) and/or single-fraction (SF) image-guided stereotactic radiosurgery (IG-SRS) may provide a more effective means of managing these lesions. METHODS AND MATERIALS Patients with pathologically proven high-grade sarcoma metastatic to the spine treated with HF and SF IG-SRS were included. LC and overall survival (OS) were analyzed by the use of Kaplan-Meier statistics. Univariate and multivariate analyses were performed by the use of Cox regression with competing-risks analysis; all confidence intervals are 95%. Toxicities were assessed according to Common Terminology Criteria for Adverse Events, version 4.0. RESULTS From May 2005 to November 11, 2012, 88 patients with 120 discrete metastases received HF (3-6 fractions; median dose, 28.5 Gy; n=52, 43.3%) or SF IG-SRS (median dose, 24 Gy; n=68, 56.7%). The median follow-up time was 12.3 months. At 12 months, LC was 87.9% (confidence interval [CI], 81.3%-94.5%), OS was 60.6% (CI, 49.6%-71.6%), and median survival was 16.9 months. SF IG-SRS demonstrated superior LC to HF IG-SRS (12-month LC of 90.8% [CI, 83%-98.6%] vs 84.1% [CI, 72.9%-95.3%] P=.007) and retained significance on multivariate analysis (P=.030, hazard ratio 0.345; CI, 0.132-0.901]. Treatment was well tolerated, with 1% acute grade 3 toxicity, 4.5% chronic grade 3 toxicity, and no grade >3 toxicities. CONCLUSIONS In the largest series of metastatic sarcoma to the spine to date, IG-SRS provides excellent LC in the setting of an aggressive disease with low radiation sensitivity and poor prognosis. Single-fraction IG-SRS is associated with the highest rates of LC with minimal toxicity.

Comparison of Local Recurrence With Conventional and Intensity-Modulated Radiation Therapy for Primary Soft-Tissue Sarcomas of the Extremity

PURPOSE The use of intensity-modulated radiation therapy (IMRT) in the treatment of soft tissue sarcoma (STS) of the extremity is increasing, but no large-scale direct comparison has been reported between conventional external-beam radiation therapy (EBRT) and IMRT. METHODS Between January 1996 and December 2010, 319 consecutive adult patients with primary nonmetastatic extremity STS were treated with limb-sparing surgery and adjuvant radiotherapy (RT) at a single institution. Conventional EBRT was used in 154 patients and IMRT in 165 with similar dosing schedules. Median follow-up time for the cohort was 58 months. RESULTS Treatment groups were comparable in terms of tumor location, histology, tumor size, depth, and use of chemotherapy. Patients treated with IMRT were older (P = .08), had more high-grade lesions (P = .05), close (< 1 mm) or positive margins (P = .04), preoperative radiation (P < .001), and nerve manipulation (P = .04). Median follow-up was 90 months for patients treated with conventional EBRT and 42 months for patients treated with IMRT. On multivariable analysis adjusting for patient age and tumor size, IMRT retained significance as an independent predictor of reduced LR (hazard ratio = 0.46; 95% CI, 0.24 to 0.89; P = .02). CONCLUSION Despite a preponderance of higher-risk features (especially close/positive margin) in the IMRT group, IMRT was associated with significantly reduced local recurrence compared with conventional EBRT for primary STS of the extremity.

Dose to the Bladder Neck Is the Most Important Predictor for Acute and Late Toxicity After Low-Dose-Rate Prostate Brachytherapy: Implications for Establishing New Dose Constraints for Treatment Planning

PURPOSE To identify an anatomic structure predictive for acute (AUT) and late (LUT) urinary toxicity in patients with prostate cancer treated with low-dose-rate brachytherapy (LDR) with or without external beam radiation therapy (EBRT). METHODS AND MATERIALS From July 2002 to January 2013, 927 patients with prostate cancer (median age, 66 years) underwent LDR brachytherapy with Iodine 125 (n=753) or Palladium 103 (n=174) as definitive treatment (n=478) and as a boost (n=449) followed by supplemental EBRT (median dose, 50.4 Gy). Structures contoured on the computed tomographic (CT) scan on day 0 after implantation included prostate, urethra, bladder, and the bladder neck, defined as 5 mm around the urethra between the catheter balloon and the prostatic urethra. AUT and LUT were assessed with the Common Terminology Criteria for Adverse Events, version4. Clinical and dosimetric factors associated with AUT and LUT were analyzed with Cox regression and receiver operating characteristic analysis to calculate area under the receiver operator curve (ROC) (AUC). RESULTS Grade ≥2 AUT and grade ≥2 LUT occurred in 520 patients (56%) and 154 patients (20%), respectively. No grade 4 toxicities were observed. Bladder neck D2cc retained a significant association with AUT (hazard ratio [HR], 1.03; 95% confidence interval [CI], 1.03-1.04; P<.0001) and LUT (HR, 1.01; 95% CI, 1.00-1.03; P=.014) on multivariable analysis. In a comparison of bladder neck with the standard dosimetric variables by use of ROC analysis (prostate V100 >90%, D90 >100%, V150 >60%, urethra D20 >130%), bladder neck D2cc >50% was shown to have the strongest prognostic power for AUT (AUC, 0.697; P<.0001) and LUT (AUC, 0.620; P<.001). CONCLUSIONS Bladder neck D2cc >50% was the strongest predictor for grade ≥2 AUT and LUT in patients treated with LDR brachytherapy. These data support inclusion of bladder neck constraints into brachytherapy planning to decrease urinary toxicity.

Postoperative pelvic intensity-modulated radiotherapy and concurrent chemotherapy in intermediate- and high-risk cervical cancer.

OBJECTIVE According to national surveys, the use of intensity-modulated radiation therapy (IMRT) in gynecologic cancers is on the rise, yet there is still some reluctance to adopt adjuvant IMRT as standard practice. The purpose of this study is to report a single-institution experience using postoperative pelvic IMRT with concurrent chemotherapy in intermediate- and high-risk early stage cervical cancer. METHODS From 1/2004 to 12/2009, 34 patients underwent radical hysterectomy and pelvic lymph node dissection (28 median nodes were removed) for early stage cervical cancer. Median dose of postoperative pelvic IMRT was 50.4 Gy (range, 45-50.4). All patients received concurrent cisplatin. RESULTS With a median follow-up of 44 months, 3 patients have recurred; 1 vaginal recurrence, 1 regional and distant, and 1 distant. The 3- and 5-year disease-free survival (DFS) was 91.2% (95% CI, 81.4-100%) and overall survival (OS) was 91.1% (95% CI, 81.3-100%). All failures and all deaths were in the high-risk group (n=3/26). There was 32.3% G3-4 hematologic toxicity, 2.9% acute G3 gastrointestinal toxicity, and no acute G3 or higher genitourinary toxicity. There were no chronic G3 or higher toxicities. CONCLUSIONS Oncologic outcomes with postoperative IMRT were very good, with DFS and OS rates of >90% at median follow-up of 44 months, despite a preponderance (76.5%) of high-risk features. Toxicity was minimal even in the setting of an aggressive trimodality approach. Data from this study and emerging data from the Phase II RTOG study (0418) demonstrate the advantages of postoperative IMRT in early stage cervical cancer.

Intraoperative 32P high-dose rate brachytherapy of the dura for recurrent primary and metastatic intracranial and spinal tumors.

BACKGROUND Treatment of spinal and intracranial tumors with dural involvement is complicated by radiation tolerance of sensitive structures, especially in the setting of previous treatment. OBJECTIVE To evaluate whether intraoperative brachytherapy with short-range sources allows therapeutic dose delivery without damaging sensitive structures. METHODS The median doses of previous treatment were 3000 cGy (range, 1800-7200 cGy) for 8 patients with primary/recurrent and 17 patients with metastatic spinal tumors and 5040 cGy (range, 1300-6040 cGy) for 5 patients with locally recurrent and 2 patients with metastatic intracranial tumors. Patients underwent gross total or maximal resection of the tumor and were then treated with an intraoperative brachytherapy plaque consisting of a flexible silicone film incorporating P. A dose of 1000 cGy was delivered to a depth of 1 mm; the percent depth dose was less than 1% at 4 mm from the prescription depth. Median postoperative radiation doses of 2700 cGy (range, 1800-3000 cGy) were delivered to 15 spinal tumor patients and 3000 cGy (range, 1800-3000 cGy) to 3 intracranial tumor patients. The median follow-up was 4.4 months (range, 2.6-23.3 months) for spinal tumor patients and 5.3 months (range, 0.7-16.2) for intracranial tumor patients. RESULTS At 6-month follow-up, for all spinal tumor patients, local progression-free survival and overall survival rates were both 83.3% (95% confidence interval [CI]: 62.3%-94.3%); for all intracranial tumor patients, the local progression-free survival rate was 62.5% (95% CI: 23.8%-90.9%) and the overall survival rate was 66.7% (95% CI: 26.7%-92.9%). There were no intraoperative or postoperative complications secondary to radiotherapy. CONCLUSION Use of the P brachytherapy plaque is technically simple and not associated with increased risk of complications, even after multiple radiation courses. Local control rates were more than 80% in patients with proven radiation-resistant spinal disease.

A comparative dosimetric analysis of virtual stereotactic body radiotherapy to high-dose-rate monotherapy for intermediate-risk prostate cancer

PURPOSE Stereotactic body radiotherapy (SBRT) is being used with increasing frequency as definitive treatment of early stage prostate cancer. Much of the justification for its adoption was derived from earlier clinical results using high-dose-rate (HDR) brachytherapy. We determine whether HDRs dosimetry can be achieved by virtual SBRT. METHODS AND MATERIALS Patients with intermediate-risk prostate cancer on a prospective trial evaluating the efficacy of HDR monotherapy treated to dose of 9.5Gy×4 fractions were used for this study. A total of 5 patients were used in this analysis. Virtual SBRT plans were developed to reproduce the planning target volume (PTV) HDR dose distributions. Both normal tissue- and PTV-prioritized plans were generated. RESULTS From the normal tissue-prioritized plan, HDR and virtual SBRT achieved similar PTV V100 (93.8% vs. 93.1%, p=0.20) and V150 (40.3% vs. 42.9%, p=0.69) coverage. However, the PTV V200 was not attainable with SBRT (15.2% vs. 0.0%, p<0.001). The rectal Dmax was significantly lower with HDR (94.2% vs. 99.42%, p=0.05). The rectal D2 cc was also lower (60.8% vs. 71.1%, p=0.07). Difference in D1 cc urethral dose was not significantly different (87.7% vs. 75.2%, p=0.33). Comparing the PTV-prioritized plans, the rectal Dmax (94.2% vs. 111.1%, p=0.05) and mean dose (27.1% vs. 33.3%, p=0.03) were significantly higher using SBRT, and the rectal D2 cc was higher using SBRT (60.8% vs. 81.8%, p=0.07). CONCLUSIONS HDR achieves significantly higher intraprostatic doses while achieving a lower maximum rectal dose compared with our virtual SBRT treatment planning. Future studies should compare clinical outcomes and toxicity between these modalities.

20-Year Experience With Intraoperative High-Dose-Rate Brachytherapy for Pediatric Sarcoma: Outcomes, Toxicity, and Practice Recommendations

PURPOSE To assess outcomes and toxicity of high-dose-rate intraoperative radiation therapy (HDR-IORT) in the management of pediatric sarcoma. METHODS AND MATERIALS Seventy-five pediatric patients underwent HDR-IORT for sarcoma from May 1993 to November 2013. The median age was 9 years old (36 patients were ≤ 6 years old). HDR-IORT was part of initial therapy in 37 patients (49%) and for recurrent disease in 38 patients (51%). Forty-one patients (55%) received HDR-IORT and postoperative external beam RT (PORT), and 22 patients (29%) were previously treated with external beam radiation therapy to the IORT site. Local control (LC), overall survival (OS) and event-free survival (EFS) were estimated using Kaplan-Meier methods. RESULTS At a median follow-up of 7.8 years for surviving patients, 5-year projected rates of LC, EFS, and OS were 63% (95% confidence interval [CI] 50%-76%), 33% (95% CI 21%-45%), and 43% (95% CI 30%-55%), with a median survival of 3.1 years. The 5-year LC, EFS, and OS rates for patients with recurrent disease were 46% (95% CI, 28%-64%), 30% (95% CI, 13%-46%), and 36% (95% CI, 18%-54%). Acute toxicity ≥ grade 3 occurred in 2 (2.5%) treatments; late toxicity ≥ grade 3 occurred in 4 (5.3%) patients 0.3-9.9 years after HDR-IORT. The incidence of toxicity ≥ grade 3 was not associated with HDR-IORT applicator size, HDR-IORT dose, prior RT or PORT, or prior or postoperative chemotherapy, but all toxicity ≥ grade 3 occurred in patients ≤ 6 years treated with HDR-IORT doses ≥ 12 Gy. CONCLUSIONS HDR-IORT is a well-tolerated component of multimodality therapy for pediatric sarcoma, allowing additional local treatment while reducing external beam exposure. Taking clinical considerations into account, doses between 8-12 Gy are appropriate for HDR-IORT in patients ≤ 6 years of age.

Outcomes and Prognostic Factors in Women With 1 to 3 Breast Cancer Brain Metastases Treated With Definitive Stereotactic Radiosurgery

BACKGROUND With the continuing increase in the use of definitive stereotactic radiosurgery (SRS) for patients with limited brain metastases (BM), clinicians need more specific prognostic tools. We investigated clinical predictors of outcomes in patients with limited breast cancer BM treated with SRS alone. METHODS AND MATERIALS We identified 136 patients with breast cancer and 1-3 BM who underwent definitive SRS for 186 BM between 2000 and 2012. The Kaplan-Meier method was used to assess overall survival (OS), regional failure (RF), and local failure (LF). Associations between clinical factors and outcomes were tested using Cox regression. A point scoring system was used to stratify patients based on OS, and the predictive power was tested with concordance probability estimate (CPE). RESULTS The median OS was 17.6 months. The 12-month RF and LF rates were 45% and 10%, respectively. On multivariate analysis, >1 lesion (hazard ratio [HR] = 1.6, P=.02), triple-negative (TN) disease (HR=2.0, P=.006), and active extracranial disease (ED) (HR=2.7, P<.0001) were significantly associated with worse OS. The point score system was defined using proportional simplification of the multivariate Cox proportional hazards regression function. The median OS for patients with 3.0-4.0 points (n=37), 4.5-5.5 points (n=28), 6.0-6.5 points (n=37), and 8-8.5 points (n=34) were 9.2, 15.6, 25.1, and 45.1 months, respectively (P<.0001, CPE = 0.72). Active ED (HR=2.4, P=.0007) was significantly associated with RF. Higher risk for LF was significantly associated with larger BM size (HR=3.1, P=.0001). CONCLUSION Patients with >1 BM, active ED, and TN had the highest risk of death after SRS. Active ED is an important prognostic factor for OS and intracranial control.

Endoluminal high-dose-rate brachytherapy for early stage and recurrent esophageal cancer in medically inoperable patients.

PURPOSE The management of superficial primary and recurrent esophageal cancer (EC) in medically inoperable patients is complex. Endoluminal high-dose-rate (HDR) brachytherapy has shown mixed results in terms of toxicity and local control. In this study, we examined the outcomes and toxicities in a set of patients with superficial primary and recurrent EC treated with a consistent HDR technique. METHODS AND MATERIALS Between 8/2008 and 7/2011, 14 patients were treated with HDR intraluminal brachytherapy, 10 (71.4%) with recurrent disease, and 4 (28.6%) with previously unirradiated lesions. Patients received three weekly fractions to a median dose of 12 Gy (range, 10-15 Gy); dose was prescribed to 7-mm median depth with mucosal dose limited to 8-10 Gy using a 12-14-mm applicator. RESULTS Median followup was 15.4 months. Overall freedom from failure (OFFF) and overall survival (OS) at 18 months were 30.8% (95% confidence interval [CI]: 5.2, 56.4) and 72.7% (95% CI: 45.3, 100), respectively. For patients with recurrent disease, OFFF and OS at 18 months were 11.1% (95% CI: 0, 32.1) and 55.6% (95% CI: 15.4, 95.8), respectively. For patients with previously unirradiated disease, OFFF and OS at 18 months were 75.0% (95% CI: 31.6, 100) and 100.0%, respectively. Eight (57.1%) patients had Grade 1 acute adverse effects; 6 (42.9%) patients had chronic Grade 1 adverse effects; 1 (7.1%) patient developed Grade 2 stricture. Grade 3 tracheoesophageal fistula occurred in 1 (7.1%) patient. One patient died before completion of treatment of unrelated causes. CONCLUSIONS HDR endoluminal brachytherapy is a well-tolerated treatment for superficial primary and recurrent EC in medically inoperable patients.